2021 Book FibreReinforcedConcreteImprove

RILEM Bookseries
Pedro Serna
Aitor Llano-Torre
José R. Martí-Vargas
Juan Navarro-Gregori Editors
Fibre Reinforced
Concrete:
Improvements
and Innovations
RILEM-fib International Symposium
on FRC (BEFIB) in 2020
Fibre Reinforced Concrete: Improvements
and Innovations
RILEM BOOKSERIES
Volume 30
RILEM, The International Union of Laboratories and Experts in Construction

Materials, Systems and Structures, founded in 1947, is a non-governmental scientific
association whose goal is to contribute to progress in the construction sciences,
techniques and industries, essentially by means of the communication it fosters
between research and practice. RILEM’s focus is on construction materials and their
use in building and civil engineering structures, covering all phases of the building
process from manufacture to use and recycling of materials. More information on
RILEM and its previous publications can be found on www.RILEM.net.
Indexed in SCOPUS, Google Scholar and SpringerLink.
More information about this series at http://www.springer.com/series/8781

Pedro Serna Aitor Llano-Torre
• •
José R. Martí-Vargas Juan Navarro-Gregori

•
Editors
Fibre Reinforced Concrete:

Improvements
and Innovations
RILEM-fib International Symposium
on FRC (BEFIB) in 2020
123
Editors
Pedro Serna Aitor Llano-Torre
ICITECH ICITECH
Universitat Politècnica de València Universitat Politècnica de València
Valencia, Spain Valencia, Spain
José R. Martí-Vargas Juan Navarro-Gregori

ICITECH ICITECH
Universitat Politècnica de València Universitat Politècnica de València
Valencia, Spain Valencia, Spain
ISSN 2211-0844 ISSN 2211-0852 (electronic)

RILEM Bookseries
ISBN 978-3-030-58481-8 ISBN 978-3-030-58482-5 (eBook)
https://doi.org/10.1007/978-3-030-58482-5
© RILEM 2021
No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written
permission from the Publisher, with the exception of any material supplied specifically for the purpose of
being entered and executed on a computer system, for exclusive use by the purchaser of the work.
Permission for use must always be obtained from the owner of the copyright: RILEM.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
2020, a weird year.

I always considered that writing the preface for the BEFIB2020 Proceedings
would be an interesting moment with a great responsibility to summarise in a few
words what was expected from the most important symposium on Fibre Reinforced
Concrete, which we can only enjoy once every four years. Unfortunately, the
sudden appearance of the COVID-19 represented a drastic change in all approaches.
Although we considered at the beginning that we could overcome this crisis, finally
it was not possible to ensure optimal safety conditions and we decided to postpone
the symposium for one year.
As Organising Committee, we were aware of the effort already done in both
writing and reviewing the scientific papers: a total of 153 papers were accepted
from the 252 submitted abstracts. At this point, we considered that awaiting one
additional year would have represented an excessive delay to maintain the scientific
novelty. Therefore, to ensure the quality level of the conference and to take
advantage of the great work done by the authors and the scientific committee during
the peer review process, the BEFIB2020 Organising Committee decided to publish
an online edition of the BEFIB2020 proceedings considering all the accepted full
papers and make a new call for abstract for the next BEFIB2021 symposium
edition.
However, we assumed that one of the main objectives when submitting a paper
for a conference is the oral presentation of the work to our colleagues and discuss
about them in person. Therefore, considering the postponement, the Organising
Committee gave the choice to the authors to withdraw or publish their contribution
in the BEFIB2020 online publication without oral presentation. Consequently,
some authors preferred to wait for another publishing opportunity to submit their
work and decided to withdraw their contribution for this publication.
This book entitled “Fibre Reinforced Concrete: Improvements and Innovations”
is the result of the peer review process performed during 2020 and includes a total
of 101 papers representing the contemporary topics of interest for the FRC
researchers and practitioners. Fresh and hardened concrete properties are covered
v
vi Preface
from rheology and early-age properties and technological aspects to the mechanical
properties and quality control.
But also, it can be observed an important effort on studying analytical and
numerical models and a tendency to structural design, codes and standards covering
interesting case studies and FRC structural applications which foreshadow a good
future for the FRC.
Durability and long-term properties and special FRC like UHPFRC or other
smart FRC, textile concretes are still challenging topics that cover some needed
advances or new concepts for the continuous progress.
I would like to acknowledge the contribution of all the authors and their com-
prehension, as well as the great work done by the members of the scientific com-
mittee to guarantee high-quality publications. Without you this work could not have
been done. I would also like to thank the support of my colleagues from the
Organising Committee in the management, structuring and finalising of this pub-
lication, especially Aitor Llano-Torre for his constant and meticulous dedication
and the collaboration of José R. Martí-Vargas.
I would like also to acknowledge the sponsor companies KrampeHarex and
SCE, FIBRAFLEX, ArcelorMittal Fibres, BASF Construction Solutions and
RIMSA Metal Technology, for their support and confidence in our event. I wish we
could enjoy a great even with their support next year in the BEFIB2021 edition.
Finally, I want to finish these words with my most alive desire to beat the
COVID-19 and to achieve the objective of our congress, which is none other than to
bring together all the people who work at the FRC in the same room. Please, be
careful, stay safe and work hard since I would like to welcome all of you next year
in Valencia for the BEFIB 2021.
I hope that the quality of this work content will cover the interest of the readers
and the authors objectives.
Thanks to all of you for your great contribution.
Pedro Serna
BEFIB2020 Chairman
Organisation
Committees
Organising Committee
Pedro Serna (Chairman), Spain

Aitor Llano-Torre (Secretariat), Spain
José R. Martí-Vargas, Spain
Juan Navarro-Gregori, Spain
Scientific Committee
M. A. Aiello, Italy E. Cuenca, Italy

A. Aguado, Spain F. Dehn, Germany
C. Aldea, Canada A. De La Fuente, Spain
S. Al-Toubat, UAE E. Denarie, Switzerland
S. Austin, UK M. di Prisco, Italy
G. L. Balázs, Hungary Y. Ding, China
N. Banthia, Canada A. Fantilli, Italy
B. Barragan, France L. Ferrara, Italy
J. A. O. Barros, Portugal A. Figueiredo, Brazil
E. S. Bernard, Australia S. Foster, Australia
A. Bettencourt Ribeiro, Portugal J. Gálvez, Spain
S. Billington, USA E. Garcia-Taengua, UK
J. Bolander, USA R. Gettu, India
P. Borges, Brazil G. M. Giaccio, Argentina
W. P. Boshoff, South Africa S. Grunewald, Netherlands
N. Buratti, Italy P. Kabele, Czech Republic
S. H. P. Cavalaro, UK T. Kanda, Japan
J. P. Charron, Canada T. Kanstad, Norway
A. Conforti, Italy I. Khan, Saudi Arabia
vii
viii Organisation
K. Kobayashi, Japan K. A. Rieder, Germany

M. Konsta-Gdoutos, Greece M. Roig-Flores, Spain
M. Kunieda, Japan P. Rossi, France
V. Li, USA E. Schlangen, Netherlands
Y. M. Lim, South Korea P. Serna (Chair), Spain
A. Llano-Torre, Spain S. Shah, USA
I. Löfgren, Sweden F. Silva, Brazil
K. Lundgren, Sweden V. Slowik, Germany
P. Lura, Switzerland S. Soleimani-Dashtaki, Canada
J. R. Marti-Vargas, Spain L. Sorelli, Canada
B. Massicotte, Canada H. Stang, Denmark
C. Mazzotti, Italy J. Sustersic, Slovenia
A. Meda, Italy G. Tiberti, Italy
V. Mechtcherine, Germany R. D. Toledo Filho, Brazil
F. Minelli, Italy F. Toutlemonde, France
B. Mobasher, USA L. Vandewalle, Belgium
C. Molins, Spain G. Van Zijl, South Africa
A. Naaman, USA J. Walraven, Netherlands
J. Navarro-Gregori, Spain Y. Yao, China
B. Nematollahi, Australia D.-Y. Yoo, South Korea
G. Parra-Montesinos, USA Y.-S. Yoon, South Korea
G. Plizzari, Italy C. Zanotti, Canada
D. Redaelli, Switzerland R. Zerbino, Argentina
J. Resplendino, France
Institutions
Organised by
UPV ICITECH
Universitat Politècnica de València Institute of Concrete Science and
Technology
Organisation ix
Supported By
RILEM fib ACI

International Union The International Amercian Concrete Institute
of Laboratories and Experts Federation for
in Construction Materials, Structural
Systems and Structures Concrete
Sponsors
Gold Sponsor
x Organisation
Bronze Sponsors
Contents
Rheology and Early-Age Properties

Influence of Different Fibre Types on the Rheology of Strain
Hardening Cementitious Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Hassan Baloch, Steffen Grünewald, Karel Lesage, and Stijn Matthys
Using Fiber Reinforced Concrete to Control Early-Age Shrinkage
in Replacement Concrete Pavement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Nakin Suksawang and Daniel Yohannes
Early Age Shrinkage Crack Distribution in Concrete Plates
Reinforced with Different Steel Fibre Types . . . . . . . . . . . . . . . . . . . . . . 24
Sébastien Wolf, Simon Cleven, and Oldrich Vlasák
Technological Aspects
Influence of Synthetic Fibres on Seismic Resistance of Reinforced
Concrete Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
E. Stefan Bernard
Development and Mechanical Characterization of Dry Fiber-
reinforced Concrete for Prefabricated Prestressed Beams . . . . . . . . . . . 49
Kamyar Bagherinejad Shahrbijari, Suman Saha, Joaquim A. O. Barros,
Isabel B. Valente, Salvador Dias, and João Leite
Simulation of Fibre Orientation in Self-compacting Concrete:
Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Thomas Bauwens, Steffen Grünewald, and Geert De Schutter
Mix Design and Properties of Self-compacting Fibrous Concrete . . . . . . 75
Rafael R. Polvere, Ana R. L. Pires, Sidiclei Formagini,
and Andrés B. Cheung
xi
xii Contents
Aligned Interlayer Fibre Reinforcement for Digital Fabrication

with Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Lukas Gebhard, Jaime Mata-Falcón, Tomislav Markić,
and Walter Kaufmann
Mixture Proportioning of Steel Fibre Reinforced Self-compacting
Concrete Based on the Compressible Packaging Method: Comparison
with ACI 237R-07 and RILEM TC 174-SCC Recommendations . . . . . . 99
M. G. Cardoso, R. M. Lameiras, T. T Oliveira, F. B. Santana,
and V. M. S. Capuzzo
Evaluation the Yield and Ultimate Strain of FRC in Compression . . . . 111
Salam Wtaife, Nakin Suksawang, and Ahmed Alsabbagh
Electromagnetic Shielding Characteristics of High Performance
Fiber Reinforced Cementitious Composites . . . . . . . . . . . . . . . . . . . . . . 123
Namkon Lee, Sungwook Kim, and Gijoon Park
Mechanical Properties
The Manufacture of Fiber Cement Blocks Using Chemical
and Thermomechanical Pulps and Rice Husk Ash . . . . . . . . . . . . . . . . . 133
Javad Torkaman
Post-Fire Flexural Tensile Strength of Macro Synthetic Fibre
Reinforced Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
Olivia Mirza, Brendan Kirkland, Kurt Bogart, and Todd Clarke
Experimental Investigation on the Cyclic Behaviour of Steel Fibre
Reinforced Concrete Under Bending . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Maure De Smedt, Rutger Vrijdaghs, Els Verstrynge, Kristof De Wilder,
and Lucie Vandewalle
Effect of Test Setups on the Shear Transfer Capacity Across Cracks
in FRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Alejandro Giraldo Soto and Walter Kaufmann
Bearable Local Stress of High-Strength SFRC . . . . . . . . . . . . . . . . . . . . 176
Sven Plückelmann, Rolf Breitenbücher, Mario Smarslik, and Peter Mark
Impact Response of Different Classes of Fibre Reinforced Concrete . . . 189
Juan C. Vivas, Raúl L. Zerbino, María C. Torrijos,
and Graciela M. Giaccio
An Experimental Study on the Fatigue Failure Mechanisms
of Pre–damaged Steel Fibre Reinforced Concrete at a Single
Fibre Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Humaira Fataar, Riaan Combrinck, and William P. Boshoff
Contents xiii
Development of an HPFRC for Use in Flat Slabs . . . . . . . . . . . . . . . . . . 209

Julia Blazy, Sandra Nunes, Carlos Sousa, and Mário Pimentel
Influence of the Steel Fibres on the Tension and Shear Resistance
of Anchoring with Anchor Channels and Channel Bolts Cast
in Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Mazen Ayoubi, Christoph Mahrenholtz, and Wilhelm Nell
Fiber Reinforced Concrete After Elevated Temperatures:
Techniques of Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
Ronney Rodrigues Agra, Ramoel Serafini,
and Antonio Domingues de Figueiredo
Influence of the Curing Temperatures on the Mechanical Properties
of Hemp Fibre-Reinforced Alkali-Activated Mortars . . . . . . . . . . . . . . . 245
Bojan Poletanovic, Gergely Nemeth, and Ildiko Merta
Equivalence Between Flexural Toughness and Energy Absorption
Capacity of FRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
Sergio Carmona and Climent Molins
Alkali Resistant (AR) Glass Fibre Influence on Glass Fibre Reinforced
Concrete (GRC) Flexural Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
S. Guzlena and G. Sakale
Fiber Reinforced Concrete Crack Opening Evaluation Using Digital
Image Correlation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
Kaio Cézar da Silva Oliveira, Gabriela Silva Dias,
Isadora Queiroz Freire de Carvalho,
Wandersson Bruno Alcides de Morais Silva, Danilo José Pereira Freitas,
Christiano Augusto Ferrario Várady Filho,
and Aline da Silva Ramos Barboza
Effect of Distribution and Orientation of Fibers on the Post-cracking
Behavior of Steel Fiber Reinforced Self-compacting Concrete in Small
Thickness Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Néstor Fabián Acosta Medina, Rodrigo de Melo Lameiras,
Ana Carolina Parapinski dos Santos, and Fábio Luiz Willrich
Ductility of the Four-Year-Old Steel Fibre Reinforced Concrete . . . . . . 290
Jakob Šušteršič, Rok Ercegovič, David Polanec, and Andrej Zajc
Sensitivity of the Flexural Performance of Glass and Synthetic FRC
to Fibre Dosage and Water/Cement Ratio . . . . . . . . . . . . . . . . . . . . . . . 301
Razan H. Al Marahla and Emilio Garcia-Taengua
Bond Between Steel Reinforcement Bars and Fiber Reinforced
Cement-Based Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Margareth S. Magalhães, Paulo José B. Teixeira,
and Maria Elizabeth N. Tavares
xiv Contents
An Experimental Study of the Influence of Moderate Temperatures

on the Behavior of Macrosynthetic Fiber Reinforced Concrete . . . . . . . 322
Marta Caballero-Jorna, Marta Roig-Flores, and Pedro Serna
Post-cracking Behaviour of Glass Fibre Reinforced Concrete
with Recycled Aggregates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
Brecht Vandevyvere, Lucie Vandewalle, Els Verstrynge, and Jiabin Li
Long-Term Properties
A Computational Sectional Approach for the Flexural Creep Behavior
of Cracked FRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
Rutger Vrijdaghs, Marco di Prisco, and Lucie Vandewalle
Shrinkage of Steel-Fibre-Reinforced Lightweight Concrete . . . . . . . . . . 359
Hasanain K. Al-Naimi and Ali A. Abbas
Time Dependent Deflection of FRC Members Under Sustained Axial
and Flexural Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
Murray Watts, Ali Amin, R. Ian Gilbert, and Walter Kaufmann
Influence of the Residual Tensile Strength on the Factor
for Quasi-permanent Value of a Variable Action w2 . . . . . . . . . . . . . . . 380
Darko Nakov, Goran Markovski, Toni Arangjelovski, and Peter Mark
Compressive and Tensile Creep and Shrinkage of Synthetic FRC:
Experimental Results and Comparison to Codes . . . . . . . . . . . . . . . . . . 392
Razan H. Al Marahla and Emilio Garcia-Taengua
Creep in FRC – From Material Properties to Composite Behavior . . . . 402
Martin Hunger, Jürgen Bokern, Simon Cleven, and Rutger Vrijdaghs
Durability
Morphology of Corrosion of Metallic Fibers in Aggressive Media . . . . . 417
Carmen Andrade and Miguel A. Sanjuán
Effects of Fibres on the Flexural Behaviour of Sound and Damaged
RC Beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
Raúl L. Zerbino, María C. Torrijos, Graciela M. Giaccio,
and Antonio Conforti
Fiber Reinforced Concrete Elements Exposed
to Accelerated Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
Camelia Negrutiu, Ioan Sosa, Bogdan Heghes, Oana Gherman,
and Horia Constantinescu
Effect of Corroded Steel Fibers on Mechanical Behavior
of Steel Fiber Reinforced Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
Minoru Kunieda, Masaki Tsutsui, and Le V. Tri
Contents xv
Self-healing of Fibre Reinforced Concrete Containing an Expansive

Agent in Different Exposure Conditions . . . . . . . . . . . . . . . . . . . . . . . . . 453
K.-S. Lauch, C. Desmettre, and J.-P. Charron
Characterisation of Strain-Hardening Cementitious Composite
(SHCC) Under Cyclic Loading Conditions for Self-healing
Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 466
Zixuan Tang, Chrysoula Litina, and Abir Al-Tabbaa
Corrosion Pattern and Mechanical Behaviour of Corroded Rebars
in Cracked Plain and Fibre Reinforced Concrete . . . . . . . . . . . . . . . . . . 477
E. Chen, Carlos G. Berrocal, Ingemar Löfgren, and Karin Lundgren
Evaluation of the Self-healing Capability of Ultra-High-Performance
Fiber-Reinforced Concrete with Nano-Particles and Crystalline
Admixtures by Means of Permeability . . . . . . . . . . . . . . . . . . . . . . . . . . 489
Hesam Doostkami, Marta Roig-Flores, Alberto Negrini,
Eduardo J. Mezquida-Alcaraz, and Pedro Serna
Analytical and Numerical Models

Material Characterisation for Nonlinear Finite Element
Analysis (NLFEA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
P. J. van der Aa and A. A. van den Bos
Comparison Between the Cracking Process of Reinforced Concrete
and Fibres Reinforced Concrete Railway Tracks by Using Non-linear
Finite Elements Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515
Jean-Louis Tailhan, Pierre Rossi, and Thierry Sedran
Experimental/Computational-Based Determination of Material
Parameters for Nonlinear Simulation of UHPFRC . . . . . . . . . . . . . . . . . 527
David Lehký, Martin Lipowczan, Drahomír Novák, Radomír Pukl,
and Milad Hafezolghorani
Mechanical Response of High Strength Fibre Reinforced Concrete
Under Extreme Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536
B. Luccioni, F. Isla, F. Fiengo, R. Codina, D. Ambrosini, J.C. Vivas,
Raúl L. Zerbino, Graciela M. Giaccio, and María C. Torrijos
Numerical Damage Modelling of Macro-synthetic Fibre
Dayani Kahagala Hewage, Christophe Camille, Olivia Mirza,
Fidelis Mashiri, Brendan Kirkland, and Todd Clarke
Finite Element Analysis of Ultra High Performance Fibre Reinforced
Concrete Beams Using Microplane Modelling . . . . . . . . . . . . . . . . . . . . 558
William. Wilson and Tomas O’Flaherty
xvi Contents
Machine Learning Prediction of Flexural Behavior of UHPFRC . . . . . . 570

Joaquín Abellán-García, Jaime A. Fernández-Gómez,
Nancy Torres-Castellanos, and Andrés M. Núñez-López
Study of Dimensioning Aspects of FRC Based on the Beam Flexion
Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584
Iva E. Pereira Lima and Aline S. Ramos Barboza
Load-Carrying Capacity of SFRC Suspended Slabs with Different
Support Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 596
Olugbenga B. Soyemi and Ali A. Abbas
Discrete Element Simulation of the Fresh State Steel Fiber Reinforced
Self-compacting Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 610
A. Najari, A. Blanco, A. de la Fuente, and S. H. P. Cavalaro
Crack Width Simulation and Nonlinear Finite Element Analysis
of Bursting and Spalling Stresses in Precast FRC Tunnel Segments
Under TBM Thrust Jack Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 621
Mehdi Bakhshi and Verya Nasri
Finite Element Modelling of UHPFRC Flexural-
Reinforced Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
Eduardo J. Mezquida-Alcaraz, Juan Navarro-Gregori, and Pedro Serna
Punching Shear Response of RC Slab-Column Connections
Strengthened with UHPFRC - Finite Element Investigation . . . . . . . . . . 651
Demewoz W. Menna and Aikaterini S. Genikomsou
Numerical Evaluation the Effect of Specimen Thickness
on Fibre Orientation in Self-consolidating Engineered
Cementitious Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 661
Hai Tran Thanh, Jianchun Li, and Y. X. Zhang
Numerical Modelling of Fiber-Reinforced Concrete
Shear-Critical Beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670
Santiago Talavera-Sánchez, Juan Navarro-Gregori, Francisco Ortiz-Navas,
and Pedro Serna
Experimental Analysis of Crack Development of an UHPC Wall
Element Under Shear Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 681
V. Příbramský, M. Kopálová, and L. Dlouhý
Assessment of the Shear Behaviour of Fibre Reinforced Concrete
Through Numerical Modelling of Shear-Friction Theory . . . . . . . . . . . . 693
Álvaro Picazo, Marcos G. Alberti, Alejandro Enfedaque,
and Jaime C. Gálvez
Contents xvii
Modeling the Compressive Behavior of Steel Fiber Reinforced

Concrete Under High Strain Rate Loads . . . . . . . . . . . . . . . . . . . . . . . . 703
Honeyeh Ramezansefat, Mohammadali Rezazadeh, Joaquim A. O. Barros,
Isabel B. Valente, and Mohammad Bakhshi
Structural Design
Post-cracking Strength Classification of Macro-synthetic Fibre
Reinforced Concrete for Sleeper Application . . . . . . . . . . . . . . . . . . . . . 717
Christophe Camille, Dayani Kahagala Hewage, Olivia Mirza,
Fidelis Mashiri, Brendan Kirkland, and Todd Clarke
Structural Behaviour of Steel-Fibre-Reinforced
Lightweight Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 730
Experimental Analysis of Beams Produced in Self-compacting
Concrete Reinforced with Different Contents of Steel Fibres . . . . . . . . . 745
Ana R. L. Pires, Rafael R. Polvere, Sidiclei Formagini,
Innovation in Durable Segments for CSO Tunnels . . . . . . . . . . . . . . . . . 757
Ralf Winterberg, Michael R. Garbeth, and Brian Glynn
Incorporation of Rate-Dependent Fracture Properties in the Design
of Precast Concrete Tunnel Segment with Hybrid Reinforcement . . . . . 770
Stefie J. Stephen and Ravindra Gettu
Codes and Standards

Developments and Standardisation of Flowable Concrete Reinforced
with Fibres for Structural Design, Update of fib TG 4.3 . . . . . . . . . . . . 779
Steffen Grünewald, Liberato Ferrara, and Frank Dehn
Assesment of Codal Provision for SFRC Beam in Minimum Shear . . . . 791
Kranti Jain and Bichitra S. Negi
Laboratory Investigations on the Installation of Fasteners in Fiber
Panagiotis Spyridis, Lars Walter, Julia Dreier, and Dirk Biermann
Quality Control
Applications of Statistical Process Control in the Evaluation
of QC Test Data for Residual Strength of FRC Samples of Tunnel
Lining Segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 815
Chidchanok Pleesudjai, Devansh Patel, Mehdi Bakhshi, Verya Nasri,
and Barzin Mobasher
xviii Contents
Using Decades of Data to Rethink Proportioning and Optimisation of

FRC Mixes: The OptiFRC Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 827
Emilio Garcia-Taengua
Case Studies: Structural and Industrial Applications

Fire Resistance of Steel Fibre Reinforced Concrete Elevated
Suspended Slabs: ISO Fire Tests and Conclusions for Design . . . . . . . . 841
Xavier Destrée, Andrejs Krasnikovs, and Sébastien Wolf
Structural Behavior of a Traditional Concrete and Hollow Tiles Mixed
Floor Reinforced with HPFRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 852
D. Sirtoli, P. Riva, and P. Girardello
Pedestrian Bridge over Las Vegas Avenue in Medellín. First Latin
American Infrastructure in UHPFRC . . . . . . . . . . . . . . . . . . . . . . . . . . 864
Joaquín Abellán-García, Andrés M. Núñez-López,
and Samuel E. Arango-Campo
First Experimental Full-Scale Elevated FRSCC Slab
in South America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 873
Luis Segura-Castillo, Diego Figueredo, Iliana Rodríguez,
and Nicolás García
Masonry Walls Strengthened with Fiber Reinforced Concrete
Subjected to Blast Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883
Salah Altoubat, Abdul Saboor Karzad, Moussa Leblouba,
Mohamed Maalej, and Pierre Estephane
Smart FRCs
Interfacial Bond Quality in Functionally Graded Concretes
Incorporating Steel Fibres and Recycled Aggregates . . . . . . . . . . . . . . . 897
Ricardo Chan, Isaac Galobardes, and Charles K. S. Moy
Towards Rebar Substitution by Fibres – Tailored Supercritical
Fibre Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 908
Katharina Look, Peter Heek, and Peter Mark
A Constitutive Model for Steel-Fibre-Reinforced Lightweight
Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 920
UV-C Treatment to Functionalize the Surfaces of Pet and PP Fibers
for Use in Cementitious Composites. Adherence Evaluation . . . . . . . . . 938
María E. Fernández, María E. Pereira, Fernando Petrone, Claudia Chocca,
and Gemma Rodríguez
Contents xix
Potential of Using Recycled Carbon Fibers as Reinforcing Material for

Fiber Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 949
Magdalena Kimm, Amna Sabir, Thomas Gries, and Piyada Suwanpinij
Textile Reinforced Concrete (TRC)

Development of Textile Reinforced UHPC with Reduced
Steel Fiber Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 963
Mengchao Zhai, Yiming Yao, Jingquan Wang, and Barzin Mobasher
Reinforcement of Concrete with Glass Multifilament Yarns:
Effect of the Impregnation on the Yarn Pull-Out Behaviour . . . . . . . . . 971
A. -C. Slama, J. -L. Gallias, and B. Fiorio
Influence of Fibres Impregnation on the Tensile Response of Flax
Textile Reinforced Mortar Composite Systems . . . . . . . . . . . . . . . . . . . . 983
Giuseppe Ferrara, Marco Pepe, Enzo Martinelli,
and Romildo D. Tolêdo Filho
Experimental Investigation of Mechanical Properties of Smart Textile
Reinforced Concrete Pipes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 991
Gozdem Dittel, Michelle Wangler, Bastian Maiworm, and Thomas Gries
UHPFRC, SHCC and ECC

Full-Scale Construction Test for Improvement of RC Void Slab
Bridges Using UHPFRC – Part 1: Experimental Test Plan . . . . . . . . . . 1003
Tohru Makita, Yuji Watanabe, Shuji Yanai, and Hirokazu Kitagawa
Full-Scale Construction Test for Improvement of RC Void Slab
Bridges Using UHPFRC – Part 2: Test Results . . . . . . . . . . . . . . . . . . . 1012
Yuji Watanabe, Shuji Yanai, Tohru Makita, and Hirokazu Kitagawa
Influence of Fiber Type on the Tensile Behavior of High-Strength
Strain-Hardening Cement-Based Composites (HS-SHCC) During
and After Exposure to Elevated Temperatures . . . . . . . . . . . . . . . . . . . 1022
Iurie Curosu, Sarah Burk, Marco Liebscher, and Viktor Mechtcherine
Tensile and Compressive Performance of High-Strength Engineered
Cementitious Composites (ECC) with Seawater and Sea-Sand . . . . . . . . 1034
Jing Yu, Bo-Tao Huang, Jia-Qi Wu, Jian-Guo Dai,
and Christopher K. Y. Leung
Effect of Fiber Content Variation in Plastic Hinge Region
of Reinforced UHPC Flexural Members . . . . . . . . . . . . . . . . . . . . . . . . . 1042
Mandeep Pokhrel, Yi Shao, Sarah Billington, and Matthew J. Bandelt
An Eco-Friendly UHPC for Structural Application:
Tensile Mechanical Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1056
Amin Abrishambaf, Mário Pimentel, and Sandra Nunes
xx Contents
Characterization of Ultra High Performance Fiber Reinforced

Concrete (UHPFRC) Tensile Behaviour . . . . . . . . . . . . . . . . . . . . . . . . . 1068
Nicola Generosi, Jacopo Donnini, and Valeria Corinaldesi
Slip-Hardening Bond: A Key to the Success of Ultra High
Performance FRC Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1079
Antoine E. Naaman
Testing of Thin UHPFRC Cantilever Stairs
with Bolted Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1090
Ioan Sosa, Camelia Negrutiu, Bogdan Heghes, and Adel Todor
Studying of Processing-Structure-Properties Relation of Strain
Hardening Cementitious Composites (SHCC) . . . . . . . . . . . . . . . . . . . . 1100
Zhenghao Li, Jiajia Zhou, Cong Lu, and Christopher K. Y. Leung
The Effect of Fiber Content on the Post-cracking Tensile Stiffness
Capacity of R-UHPFRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1112
M. Khorami, Juan Navarro-Gregori, and Pedro Serna
Controlling Strength and Ductility of Strain-Hardening Cementitious
Composites by Nano-Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1124
Ousmane A. Hisseine and Arezki T. Hamou
Comprehensive Characterization of UHPFRC Mixes for Seismic
and Durability Rehabilitation of Bridge Piers . . . . . . . . . . . . . . . . . . . . 1137
C. Sevigny-Vallières, P. Marchand, B. Terrade, N. Roy, F. Toutlemonde,
and A. Tagnit-Hamou
Evaluation of the Splitting Tensile Strength of Ultra-High
Performance Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1149
An Hoang Le
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1161

RILEM Publications
The following list is presenting the global offer of RILEM Publications, sorted by
series. Each publication is available in printed version and/or in online version.
RILEM Proceedings (PRO)
PRO 1: Durability of High Performance Concrete (ISBN: 2-912143-03-9; e-ISBN:

2-351580-12-5; e-ISBN: 2351580125); Ed. H. Sommer
PRO 2: Chloride Penetration into Concrete (ISBN: 2-912143-00-04; e-ISBN:
2912143454); Eds. L.-O. Nilsson and J.-P. Ollivier
PRO 3: Evaluation and Strengthening of Existing Masonry Structures (ISBN:
2-912143-02-0; e-ISBN: 2351580141); Eds. L. Binda and C. Modena
PRO 4: Concrete: From Material to Structure (ISBN: 2-912143-04-7; e-ISBN:
2351580206); Eds. J.-P. Bournazel and Y. Malier
PRO 5: The Role of Admixtures in High Performance Concrete (ISBN: 2-912143-
05-5; e-ISBN: 2351580214); Eds. J. G. Cabrera and R. Rivera-Villarreal
PRO 6: High Performance Fiber Reinforced Cement Composites - HPFRCC 3
(ISBN: 2-912143-06-3; e-ISBN: 2351580222); Eds. H. W. Reinhardt and
A. E. Naaman
PRO 7: 1st International RILEM Symposium on Self-Compacting Concrete
(ISBN: 2-912143-09-8; e-ISBN: 2912143721); Eds. Å. Skarendahl and
Ö. Petersson
PRO 8: International RILEM Symposium on Timber Engineering (ISBN:
2-912143-10-1; e-ISBN: 2351580230); Ed. L. Boström
xxi
xxii RILEM Publications
PRO 9: 2nd International RILEM Symposium on Adhesion between Polymers and

Concrete ISAP ’99 (ISBN: 2-912143-11-X; e-ISBN: 2351580249); Eds. Y. Ohama
and M. Puterman
PRO 10: 3rd International RILEM Symposium on Durability of Building and
Construction Sealants (ISBN: 2-912143-13-6; e-ISBN: 2351580257); Eds.
A. T. Wolf
PRO 11: 4th International RILEM Conference on Reflective Cracking
in Pavements (ISBN: 2-912143-14-4; e-ISBN: 2351580265); Eds.
A. O. Abd El Halim, D. A. Taylor and El H. H. Mohamed
PRO 12: International RILEM Workshop on Historic Mortars: Characteristics and
Tests (ISBN: 2-912143-15-2; e-ISBN: 2351580273); Eds. P. Bartos, C. Groot and
J. J. Hughes
PRO 13: 2nd International RILEM Symposium on Hydration and Setting (ISBN:
2-912143-16-0; e-ISBN: 2351580281); Ed. A. Nonat
PRO 14: Integrated Life-Cycle Design of Materials and Structures - ILCDES 2000
(ISBN: 951-758-408-3; e-ISBN: 235158029X); (ISSN: 0356-9403); Ed. S. Sarja
PRO 15: Fifth RILEM Symposium on Fibre-Reinforced Concretes (FRC) -
BEFIB’2000 (ISBN: 2-912143-18-7; e-ISBN: 291214373X); Eds. P. Rossi and
G. Chanvillard
PRO 16: Life Prediction and Management of Concrete Structures (ISBN:
2-912143-19-5; e-ISBN: 2351580303); Ed. D. Naus
PRO 17: Shrinkage of Concrete – Shrinkage 2000 (ISBN: 2-912143-20-9; e-ISBN:
2351580311); Eds. V. Baroghel-Bouny and P.-C. Aïtcin
PRO 18: Measurement and Interpretation of the On-Site Corrosion Rate (ISBN:
2-912143-21-7; e-ISBN: 235158032X); Eds. C. Andrade, C. Alonso, J. Fullea,
J. Polimon and J. Rodriguez
PRO 19: Testing and Modelling the Chloride Ingress into Concrete (ISBN:
2-912143-22-5; e-ISBN: 2351580338); Eds. C. Andrade and J. Kropp
PRO 20: 1st International RILEM Workshop on Microbial Impacts on Building
Materials (CD 02) (e-ISBN 978-2-35158-013-4); Ed. M. Ribas Silva
PRO 21: International RILEM Symposium on Connections between Steel and
Concrete (ISBN: 2-912143-25-X; e-ISBN: 2351580346); Ed. R. Eligehausen
PRO 22: International RILEM Symposium on Joints in Timber Structures (ISBN:
2-912143-28-4; e-ISBN: 2351580354); Eds. S. Aicher and H.-W. Reinhardt
PRO 23: International RILEM Conference on Early Age Cracking in Cementitious
Systems (ISBN: 2-912143-29-2; e-ISBN: 2351580362); Eds. K. Kovler and
A. Bentur
RILEM Publications xxiii
PRO 24: 2nd International RILEM Workshop on Frost Resistance of Concrete

(ISBN: 2-912143-30-6; e-ISBN: 2351580370); Eds. M. J. Setzer, R. Auberg and
H.-J. Keck
PRO 25: International RILEM Workshop on Frost Damage in Concrete (ISBN:
2-912143-31-4; e-ISBN: 2351580389); Eds. D. J. Janssen, M. J. Setzer and
M. B. Snyder
PRO 26: International RILEM Workshop on On-Site Control and Evaluation of
Masonry Structures (ISBN: 2-912143-34-9; e-ISBN: 2351580141); Eds. L. Binda
and R. C. de Vekey
PRO 27: International RILEM Symposium on Building Joint Sealants (CD03;
e-ISBN: 235158015X); Ed. A. T. Wolf
PRO 28: 6th International RILEM Symposium on Performance Testing and
Evaluation of Bituminous Materials - PTEBM’03 (ISBN: 2-912143-35-7; e-ISBN:
978-2-912143-77-8); Ed. M. N. Partl
PRO 29: 2nd International RILEM Workshop on Life Prediction and Ageing
Management of Concrete Structures (ISBN: 2-912143-36-5; e-ISBN: 2912143780);
Ed. D. J. Naus
PRO 30: 4th International RILEM Workshop on High Performance Fiber
Reinforced Cement Composites - HPFRCC 4 (ISBN: 2-912143-37-3; e-ISBN:
2912143799); Eds. A. E. Naaman and H. W. Reinhardt
PRO 31: International RILEM Workshop on Test and Design Methods for Steel
Fibre Reinforced Concrete: Background and Experiences (ISBN: 2-912143-38-1;
e-ISBN: 2351580168); Eds. B. Schnütgen and L. Vandewalle
PRO 32: International Conference on Advances in Concrete and Structures 2 vol.
(ISBN (set): 2-912143-41-1; e-ISBN: 2351580176); Eds. Ying-shu Yuan, Surendra
P. Shah and Heng-lin Lü
PRO 33: 3rd International Symposium on Self-Compacting Concrete (ISBN:
2-912143-42-X; e-ISBN: 2912143713); Eds. Ó. Wallevik and I. Níelsson
PRO 34: International RILEM Conference on Microbial Impact on Building
Materials (ISBN: 2-912143-43-8; e-ISBN: 2351580184); Ed. M. Ribas Silva
PRO 35: International RILEM TC 186-ISA on Internal Sulfate Attack
and Delayed Ettringite Formation (ISBN: 2-912143-44-6; e-ISBN: 2912143802);
Eds. K. Scrivener and J. Skalny
PRO 36: International RILEM Symposium on Concrete Science and Engineering –
A Tribute to Arnon Bentur (ISBN: 2-912143-46-2; e-ISBN: 2912143586);
Eds. K. Kovler, J. Marchand, S. Mindess and J. Weiss
PRO 37: 5th International RILEM Conference on Cracking in Pavements –
Mitigation, Risk Assessment and Prevention (ISBN: 2-912143-47-0; e-ISBN:
2912143764); Eds. C. Petit, I. Al-Qadi and A. Millien
xxiv RILEM Publications
PRO 38: 3rd International RILEM Workshop on Testing and Modelling the
Chloride Ingress into Concrete (ISBN: 2-912143-48-9; e-ISBN: 2912143578);
Eds. C. Andrade and J. Kropp
PRO 39: 6th International RILEM Symposium on Fibre-Reinforced Concretes -
BEFIB 2004 (ISBN: 2-912143-51-9; e-ISBN: 2912143748); Eds. M. Di Prisco,
R. Felicetti and G. A. Plizzari
PRO 40: International RILEM Conference on the Use of Recycled Materials in
Buildings and Structures (ISBN: 2-912143-52-7; e-ISBN: 2912143756);
Eds. E. Vázquez, Ch. F. Hendriks and G. M. T. Janssen
PRO 41: RILEM International Symposium on Environment-Conscious Materials
and Systems for Sustainable Development (ISBN: 2-912143-55-1; e-ISBN:
2912143640); Eds. N. Kashino and Y. Ohama
PRO 42: SCC’2005 - China: 1st International Symposium on Design, Performance
and Use of Self-Consolidating Concrete (ISBN: 2-912143-61-6; e-ISBN:
2912143624); Eds. Zhiwu Yu, Caijun Shi, Kamal Henri Khayat and Youjun Xie
PRO 43: International RILEM Workshop on Bonded Concrete Overlays (e-ISBN:
2-912143-83-7); Eds. J. L. Granju and J. Silfwerbrand
PRO 44: 2nd International RILEM Workshop on Microbial Impacts on Building
Materials (CD11) (e-ISBN: 2-912143-84-5); Ed. M. Ribas Silva
PRO 45: 2nd International Symposium on Nanotechnology in Construction, Bilbao
(ISBN: 2-912143-87-X; e-ISBN: 2912143888); Eds. Peter J. M. Bartos, Yolanda
de Miguel and Antonio Porro
PRO 46: ConcreteLife’06 - International RILEM-JCI Seminar on Concrete
Durability and Service Life Planning: Curing, Crack Control, Performance in Harsh
Environments (ISBN: 2-912143-89-6; e-ISBN: 291214390X); Ed. K. Kovler
PRO 47: International RILEM Workshop on Performance Based Evaluation and
Indicators for Concrete Durability (ISBN: 978-2-912143-95-2; e-ISBN:
9782912143969); Eds. V. Baroghel-Bouny, C. Andrade, R. Torrent and
K. Scrivener
PRO 48: 1st International RILEM Symposium on Advances in Concrete through
Science and Engineering (e-ISBN: 2-912143-92-6); Eds. J. Weiss, K. Kovler,
J. Marchand, and S. Mindess
PRO 49: International RILEM Workshop on High Performance Fiber Reinforced
Cementitious Composites in Structural Applications (ISBN: 2-912143-93-4;
e-ISBN: 2912143942); Eds. G. Fischer and V.C. Li
PRO 50: 1st International RILEM Symposium on Textile Reinforced Concrete
(ISBN: 2-912143-97-7; e-ISBN: 2351580087); Eds. Josef Hegger, Wolfgang
Brameshuber and Norbert Will
RILEM Publications xxv
PRO 51: 2nd International Symposium on Advances in Concrete through

Science and Engineering (ISBN: 2-35158-003-6; e-ISBN: 2-35158-002-8);
Eds. J. Marchand, B. Bissonnette, R. Gagné, M. Jolin and F. Paradis
PRO 52: Volume Changes of Hardening Concrete: Testing and Mitigation (ISBN:
2-35158-004-4; e-ISBN: 2-35158-005-2); Eds. O. M. Jensen, P. Lura and
K. Kovler
PRO 53: High Performance Fiber Reinforced Cement Composites - HPFRCC5
(ISBN: 978-2-35158-046-2; e-ISBN: 978-2-35158-089-9); Eds. H. W. Reinhardt
and A. E. Naaman
PRO 54: 5th International RILEM Symposium on Self-Compacting Concrete
(ISBN: 978-2-35158-047-9; e-ISBN: 978-2-35158-088-2); Eds. G. De Schutter and
V. Boel
PRO 55: International RILEM Symposium Photocatalysis, Environment and
Construction Materials (ISBN: 978-2-35158-056-1; e-ISBN: 978-2-35158-057-8);
Eds. P. Baglioni and L. Cassar
PRO 56: International RILEM Workshop on Integral Service Life Modelling of
Concrete Structures (ISBN 978-2-35158-058-5; e-ISBN: 978-2-35158-090-5);
Eds. R. M. Ferreira, J. Gulikers and C. Andrade
PRO 57: RILEM Workshop on Performance of cement-based materials in
aggressive aqueous environments (e-ISBN: 978-2-35158-059-2); Ed. N. De Belie
PRO 58: International RILEM Symposium on Concrete Modelling - CONMOD’08
(ISBN: 978-2-35158-060-8; e-ISBN: 978-2-35158-076-9); Eds. E. Schlangen
and G. De Schutter
PRO 59: International RILEM Conference on On Site Assessment of Concrete,
Masonry and Timber Structures - SACoMaTiS 2008 (ISBN set: 978-2-35158-061-
5; e-ISBN: 978-2-35158-075-2); Eds. L. Binda, M. di Prisco and R. Felicetti
PRO 60: Seventh RILEM International Symposium on Fibre Reinforced Concrete:
Design and Applications - BEFIB 2008 (ISBN: 978-2-35158-064-6; e-ISBN: 978-
2-35158-086-8); Ed. R. Gettu
PRO 61: 1st International Conference on Microstructure Related Durability of
Cementitious Composites 2 vol., (ISBN: 978-2-35158-065-3; e-ISBN: 978-2-
35158-084-4); Eds. W. Sun, K. van Breugel, C. Miao, G. Ye and H. Chen
PRO 62: NSF/ RILEM Workshop: In-situ Evaluation of Historic Wood and
Masonry Structures (e-ISBN: 978-2-35158-068-4); Eds. B. Kasal, R. Anthony and
M. Drdácký
PRO 63: Concrete in Aggressive Aqueous Environments: Performance, Testing
and Modelling, 2 vol., (ISBN: 978-2-35158-071-4; e-ISBN: 978-2-35158-082-0);
Eds. M. G. Alexander and A. Bertron
xxvi RILEM Publications
PRO 64: Long Term Performance of Cementitious Barriers and Reinforced

Concrete in Nuclear Power Plants and Waste Management - NUCPERF 2009
(ISBN: 978-2-35158-072-1; e-ISBN: 978-2-35158-087-5); Eds. V. L’Hostis,
R. Gens, C. Gallé
PRO 65: Design Performance and Use of Self-consolidating Concrete - SCC’2009
(ISBN: 978-2-35158-073-8; e-ISBN: 978-2-35158-093-6); Eds. C. Shi, Z. Yu,
K. H. Khayat and P. Yan
PRO 66: 2nd International RILEM Workshop on Concrete Durability and Service
Life Planning - ConcreteLife’09 (ISBN: 978-2-35158-074-5; ISBN: 978-2-35158-
074-5); Ed. K. Kovler
PRO 67: Repairs Mortars for Historic Masonry (e-ISBN: 978-2-35158-083-7);
Ed. C. Groot
PRO 68: Proceedings of the 3rd International RILEM Symposium on ‘Rheology of
Cement Suspensions such as Fresh Concrete (ISBN 978-2-35158-091-2; e-ISBN:
978-2-35158-092-9); Eds. O. H. Wallevik, S. Kubens and S. Oesterheld
PRO 69: 3rd International PhD Student Workshop on ‘Modelling the Durability of
Reinforced Concrete (ISBN: 978-2-35158-095-0); Eds. R. M. Ferreira, J. Gulikers
and C. Andrade
PRO 70: 2nd International Conference on ‘Service Life Design for Infrastructure’
(ISBN set: 978-2-35158-096-7, e-ISBN: 978-2-35158-097-4); Ed. K. van Breugel,
G. Ye and Y. Yuan
PRO 71: Advances in Civil Engineering Materials - The 50-year Teaching
Anniversary of Prof. Sun Wei’ (ISBN: 978-2-35158-098-1; e-ISBN: 978-2-35158-
099-8); Eds. C. Miao, G. Ye, and H. Chen
PRO 72: First International Conference on ‘Advances in Chemically-Activated
Materials – CAM’2010’ (2010), 264 pp, ISBN: 978-2-35158-101-8; e-ISBN: 978-
2-35158-115-5, Eds. Caijun Shi and Xiaodong Shen
PRO 73: 2nd International Conference on ‘Waste Engineering and Management -
ICWEM 2010’ (2010), 894 pp, ISBN: 978-2-35158-102-5; e-ISBN: 978-2-35158-
103-2, Eds. J. Zh. Xiao, Y. Zhang, M. S. Cheung and R. Chu
PRO 74: International RILEM Conference on ‘Use of Superabsorsorbent Polymers
and Other New Addditives in Concrete’ (2010) 374 pp., ISBN: 978-2-35158-104-9;
e-ISBN: 978-2-35158-105-6; Eds. O. M. Jensen, M. T. Hasholt, and S. Laustsen
PRO 75: International Conference on ‘Material Science - 2nd ICTRC - Textile
Reinforced Concrete - Theme 1’ (2010) 436 pp., ISBN: 978-2-35158-106-3;
e-ISBN: 978-2-35158-107-0; Ed. W. Brameshuber
PRO 76: International Conference on ‘Material Science - HetMat - Modelling of
Heterogeneous Materials - Theme 2’ (2010) 255 pp., ISBN: 978-2-35158-108-7;
e-ISBN: 978-2-35158-109-4; Ed. W. Brameshuber
RILEM Publications xxvii
PRO 77: International Conference on ‘Material Science - AdIPoC - Additions

Improving Properties of Concrete - Theme 3’ (2010) 459 pp., ISBN: 978-2-35158-
110-0; e-ISBN: 978-2-35158-111-7; Ed. W. Brameshuber
PRO 78: 2nd Historic Mortars Conference and RILEM TC 203-RHM Final
Workshop – HMC2010 (2010) 1416 pp., e-ISBN: 978-2-35158-112-4;
Eds J. Válek, C. Groot, and J. J. Hughes
PRO 79: International RILEM Conference on Advances in Construction Materials
Through Science and Engineering (2011) 213 pp., ISBN: 978-2-35158-116-2,
e-ISBN: 978-2-35158-117-9; Eds. Christopher Leung and K. T. Wan
PRO 80: 2nd International RILEM Conference on Concrete Spalling due to Fire
Exposure (2011) 453 pp., ISBN: 978-2-35158-118-6, e-ISBN: 978-2-35158-119-3;
Eds E.A.B. Koenders and F. Dehn
PRO 81: 2nd International RILEM Conference on Strain Hardening Cementitious
Composites (SHCC2-Rio) (2011) 451 pp., ISBN: 978-2-35158-120-9, e-ISBN:
978-2-35158-121-6; Eds R.D. Toledo Filho, F.A. Silva, E.A.B. Koenders and
E.M.R. Fairbairn
PRO 82: 2nd International RILEM Conference on Progress of Recycling in the
Built Environment (2011) 507 pp., e-ISBN: 978-2-35158-122-3; Eds V.M. John,
E. Vazquez, S.C. Angulo and C. Ulsen
PRO 83: 2nd International Conference on Microstructural-related Durability of
Cementitious Composites (2012) 250 pp., ISBN: 978-2-35158-129-2; e-ISBN:
978-2-35158-123-0; Eds G. Ye, K. van Breugel, W. Sun and C. Miao
PRO 84: CONSEC13 - Seventh International Conference on Concrete under
Severe Conditions – Environment and Loading (2013) 1930 pp., ISBN: 978-2-
35158-124-7; e-ISBN: 978-2- 35158-134-6; Eds Z.J. Li, W. Sun, C.W. Miao,
K. Sakai, O.E. Gjorv & N. Banthia
PRO 85: RILEM-JCI International Workshop on Crack Control of Mass Concrete
and Related issues concerning Early-Age of Concrete Structures – ConCrack 3 –
Control of Cracking in Concrete Structures 3 (2012) 237 pp., ISBN: 978-2-35158-
125-4; e-ISBN: 978-2-35158-126-1; Eds F. Toutlemonde and J.-M. Torrenti
PRO 86: International Symposium on Life Cycle Assessment and Construction
(2012) 414 pp., ISBN: 978-2-35158-127-8, e-ISBN: 978-2-35158-128-5; Eds
A. Ventura and C. de la Roche
PRO 87: UHPFRC 2013 – RILEM-fib-AFGC International Symposium on Ultra-
High Performance Fibre-Reinforced Concrete (2013), ISBN: 978-2-35158-130-8,
e-ISBN: 978-2-35158-131-5; Eds F. Toutlemonde
PRO 88: 8th RILEM International Symposium on Fibre Reinforced Concrete
(2012) 344 pp., ISBN: 978-2-35158-132-2, e-ISBN: 978-2-35158-133-9;
Eds Joaquim A.O. Barros
xxviii RILEM Publications
PRO 89: RILEM International workshop on performance-based specification and

control of concrete durability (2014) 678 pp, ISBN: 978-2-35158-135-3, e-ISBN:
978-2-35158-136-0; Eds. D. Bjegović, H. Beushausen and M. Serdar
PRO 90: 7th RILEM International Conference on Self-Compacting Concrete and
of the 1st RILEM International Conference on Rheology and Processing of
Construction Materials (2013) 396 pp, ISBN: 978-2-35158-137-7, e-ISBN: 978-2-
35158-138-4; Eds. Nicolas Roussel and Hela Bessaies-Bey
PRO 91: CONMOD 2014 - RILEM International Symposium on Concrete
Modelling (2014), ISBN: 978-2-35158-139-1; e-ISBN: 978-2-35158-140-7; Eds.
Kefei Li, Peiyu Yan and Rongwei Yang
PRO 92: CAM 2014 - 2nd International Conference on advances in chemically-
activated materials (2014) 392 pp., ISBN: 978-2-35158-141-4; e-ISBN: 978-2-
35158-142-1; Eds. Caijun Shi and Xiadong Shen
PRO 93: SCC 2014 - 3rd International Symposium on Design, Performance and
Use of Self-Consolidating Concrete (2014) 438 pp., ISBN: 978-2-35158-143-8;
e-ISBN: 978-2-35158-144-5; Eds. Caijun Shi, Zhihua Ou, Kamal H. Khayat
PRO 94 (online version): HPFRCC-7 - 7th RILEM conference on High perfor-
mance fiber reinforced cement composites (2015), e-ISBN: 978-2-35158-146-9;
Eds. H.W. Reinhardt, G.J. Parra-Montesinos, H. Garrecht
PRO 95: International RILEM Conference on Application of superabsorbent
polymers and other new admixtures in concrete construction (2014), ISBN: 978-2-
35158-147-6; e-ISBN: 978-2-35158-148-3; Eds. Viktor Mechtcherine, Christof
Schroefl
PRO 96 (online version): XIII DBMC: XIII International Conference on
Durability of Building Materials and Components(2015), e-ISBN: 978-2-35158-
149-0; Eds. M. Quattrone, V.M. John
PRO 97: SHCC3 – 3rd International RILEM Conference on Strain Hardening
Cementitious Composites (2014), ISBN: 978-2-35158-150-6; e-ISBN: 978-2-
35158-151-3; Eds. E. Schlangen, M.G. Sierra Beltran, M. Lukovic, G. Ye
PRO 98: FERRO-11 – 11th International Symposium on Ferrocement and 3rd
ICTRC - International Conference.on Textile Reinforced Concrete (2015), ISBN:
978-2-35158-152-0; e-ISBN: 978-2-35158-153-7; Ed. W. Brameshuber
PRO 99 (online version): ICBBM 2015 - 1st International Conference on
Bio-Based Building Materials (2015), e-ISBN: 978-2-35158-154-4; Eds.
S. Amziane, M. Sonebi
PRO 100: SCC16 - RILEM Self-Consolidating Concrete Conference (2016),
ISBN: 978-2-35158-156-8; e-ISBN: 978-2-35158-157-5; Ed. Kamal H. Kayat
RILEM Publications xxix
PRO 101 (online version): III Progress of Recycling in the Built Environment
(2015), e-ISBN: 978-2-35158-158-2; Eds I. Martins, C. Ulsen and S. C. Angulo
PRO 102 (online version): RILEM Conference on Microorganisms-Cementitious
Materials Interactions (2016), e-ISBN: 978-2-35158-160-5; Eds. Alexandra
Bertron, Henk Jonkers, Virginie Wiktor
PRO 103 (online version): ACESC’16 - Advances in Civil Engineering and
Sustainable Construction (2016), e-ISBN: 978-2-35158-161-2; Eds. T.Ch.
Madhavi, G. Prabhakar, Santhosh Ram and P.M. Rameshwaran
PRO 104 (online version): SSCS’2015 - Numerical Modeling - Strategies for
Sustainable Concrete Structures (2015), e-ISBN: 978-2-35158-162-9
PRO 105: 1st International Conference on UHPC Materials and Structures (2016),
ISBN: 978-2-35158-164-3, e-ISBN: 978-2-35158-165-0
PRO 106: AFGC-ACI-fib-RILEM International Conference on Ultra-High-
Performance Fibre-Reinforced Concrete – UHPFRC 2017 (2017), ISBN: 978-2-
35158-166-7, e-ISBN: 978-2-35158-167-4; Eds. François Toutlemonde & Jacques
Resplendino
PRO 107 (online version): XIV DBMC – 14th International Conference on
Durability of Building Materials and Components (2017), e-ISBN: 978-2-35158-
159-9; Eds. Geert De Schutter, Nele De Belie, Arnold Janssens,
Nathan Van Den Bossche
PRO 108: MSSCE 2016 - Innovation of Teaching in Materials and Structures
(2016), ISBN: 978-2-35158-178-0, e-ISBN: 978-2-35158-179-7; Ed. Per
Goltermann
PRO 109 (2 volumes): MSSCE 2016 - Service Life of Cement-Based Materials
and Structures (2016), ISBN Vol. 1: 978-2-35158-170-4, Vol. 2: 978-2-35158-171-
4, Set Vol. 1&2: 978-2-35158-172-8, e-ISBN : 978-2-35158-173-5; Eds. Miguel
Azenha, Ivan Gabrijel, Dirk Schlicke, Terje Kanstad and Ole Mejlhede Jensen
PRO 110: MSSCE 2016 - Historical Masonry (2016), ISBN: 978-2-35158-178-0,
e-ISBN: 978-2-35158-179-7; Eds. Inge Rörig-Dalgaard and Ioannis Ioannou
PRO 111: MSSCE 2016 - Electrochemistry in Civil Engineering (2016), ISBN:
978-2-35158-176-6, e-ISBN: 978-2-35158-177-3; Ed. Lisbeth M. Ottosen
PRO 112: MSSCE 2016 - Moisture in Materials and Structures (2016), ISBN: 978-
2-35158-178-0, e-ISBN: 978-2-35158-179-7; Eds. Kurt Kielsgaard Hansen,
Carsten Rode and Lars-Olof Nilsson
PRO 113: MSSCE 2016 - Concrete with Supplementary Cementitious
Materials (2016), ISBN: 978-2-35158-178-0, e-ISBN: 978-2-35158-179-7;
Eds. Ole Mejlhede Jensen, Konstantin Kovler and Nele De Belie
xxx RILEM Publications
PRO 114: MSSCE 2016 - Frost Action in Concrete (2016), ISBN: 978-2-35158-
182-7, e-ISBN: 978-2-35158-183-4; Eds. Marianne Tange Hasholt, Katja Fridh
and R. Doug Hooton
PRO 115: MSSCE 2016 - Fresh Concrete (2016), ISBN: 978-2-35158-184-1,
e-ISBN: 978-2-35158-185-8; Eds. Lars N. Thrane, Claus Pade, Oldrich Svec
and Nicolas Roussel
PRO 116: BEFIB 2016 – 9th RILEM International Symposium on Fiber
Reinforced Concrete (2016), ISBN: 978-2-35158-187-2, e-ISBN: 978-2-35158-
186-5; Eds. N. Banthia, M. di Prisco and S. Soleimani-Dashtaki
PRO 117: 3rd International RILEM Conference on Microstructure Related
Durability of Cementitious Composites (2016), ISBN: 978-2-35158-188-9,
e-ISBN: 978-2-35158-189-6; Eds. Changwen Miao, Wei Sun, Jiaping Liu,
Huisu Chen, Guang Ye and Klaas van Breugel
PRO 118 (4 volumes): International Conference on Advances in Construction
Materials and Systems (2017), ISBN Set: 978-2-35158-190-2, Vol. 1: 978-2-
35158-193-3, Vol. 2: 978-2-35158-194-0, Vol. 3: ISBN: 978-2-35158-195-7,
Vol. 4: ISBN: 978-2-35158-196-4, e-ISBN: 978-2-35158-191-9; Eds. Manu
Santhanam, Ravindra Gettu, Radhakrishna G. Pillai and Sunitha K. Nayar
PRO 119 (online version): ICBBM 2017 - Second International RILEM
Conference on Bio-based Building Materials, (2017), e-ISBN: 978-2-35158-192-6;
Eds. Sofiane Amziane, Mohammed Sonebi
PRO 120 (2 volumes): EAC-02 - 2nd International RILEM/COST Conference on
Early Age Cracking and Serviceability in Cement-based Materials and Structures,
(2017), Vol. 1: 978-2-35158-199-5, Vol. 2: 978-2-35158-200-8, Set: 978-2-35158-
197-1, e-ISBN: 978-2-35158-198-8; Eds. Stéphanie Staquet and Dimitrios Aggelis
PRO 121 (2 volumes): SynerCrete18: Interdisciplinary Approaches for Cement-
based Materials and Structural Concrete: Synergizing Expertise and Bridging
Scales of Space and Time, (2018), Set: 978-2-35158-202-2, Vol.1: 978-2-
35158-211-4, Vol.2: 978-2-35158-212-1, e-ISBN: 978-2-35158-203-9; Eds.
Miguel Azenha, Dirk Schlicke, Farid Benboudjema, Agnieszka Knoppik
PRO 122: SCC’2018 China - Fourth International Symposium on Design,
Performance and Use of Self-Consolidating Concrete, (2018), ISBN: 978-2-35158-
204-6, e-ISBN: 978-2-35158-205-3; Eds. C. Shi, Z. Zhang, K. H. Khayat
PRO 123: Final Conference of RILEM TC 253-MCI: Microorganisms-
Cementitious Materials Interactions (2018), Set: 978-2-35158-207-7, Vol.1: 978-2-
35158-209-1, Vol.2: 978-2-35158-210-7, e-ISBN: 978-2-35158-206-0;
Ed. Alexandra Bertron
PRO 124 (online version): Fourth International Conference Progress
of Recycling in the Built Environment (2018), e-ISBN: 978-2-35158-208-4;
Eds. Isabel M. Martins, Carina Ulsen, Yury Villagran
RILEM Publications xxxi
PRO 125 (online version): SLD4 - 4th International Conference on Service Life
Design for Infrastructures (2018), e-ISBN: 978-2-35158-213-8; Eds. Guang Ye,
Yong Yuan, Claudia Romero Rodriguez, Hongzhi Zhang, Branko Savija
PRO 126: Workshop on Concrete Modelling and Material Behaviour in honor of
Professor Klaas van Breugel (2018), ISBN: 978-2-35158-214-5, e-ISBN: 978-2-
35158-215-2; Ed. Guang Ye
PRO 127 (online version): CONMOD2018 - Symposium on Concrete Modelling
(2018), e-ISBN: 978-2-35158-216-9; Eds. Erik Schlangen, Geert de Schutter,
Branko Savija, Hongzhi Zhang, Claudia Romero Rodriguez
PRO 128: SMSS2019 - International Conference on Sustainable Materials,
Systems and Structures (2019), ISBN: 978-2-35158-217-6, e-ISBN: 978-2-35158-
218-3
PRO 129: 2nd International Conference on UHPC Materials and Structures
(UHPC2018-China), ISBN: 978-2-35158-219-0, e-ISBN: 978-2-35158-220-6
PRO 130: 5th Historic Mortars Conference (2019), ISBN: 978-2-35158-221-3,
e-ISBN: 978-2-35158-222-0; Eds. José Ignacio Álvarez, José María Fernández,
Íñigo Navarro, Adrián Durán, Rafael Sirera
PRO 131 (online version): 3rd International Conference on Bio-Based Building
Materials (ICBBM2019), e-ISBN: 978-2-35158-229-9; Eds. Mohammed Sonebi,
Sofiane Amziane, Jonathan Page
PRO 132: IRWRMC’18 - International RILEM Workshop on Rheological
Measurements of Cement-based Materials (2018), ISBN: 978-2-35158-230-5,
e-ISBN: 978-2-35158-231-2; Eds. Chafika Djelal, Yannick Vanhove
PRO 133 (online version): CO2STO2019 - International Workshop CO2 Storage
in Concrete (2019), e-ISBN: 978-2-35158-232-9; Eds. Assia Djerbi, Othman
Omikrine-Metalssi, Teddy Fen-Chong
RILEM Reports (REP)
Report 19: Considerations for Use in Managing the Aging of Nuclear Power Plant
Concrete Structures (ISBN: 2-912143-07-1); Ed. D. J. Naus
Report 20: Engineering and Transport Properties of the Interfacial Transition Zone
in Cementitious Composites (ISBN: 2-912143-08-X); Eds. M. G. Alexander,
G. Arliguie, G. Ballivy, A. Bentur and J. Marchand
Report 21: Durability of Building Sealants (ISBN: 2-912143-12-8); Ed. A. T. Wolf
Report 22: Sustainable Raw Materials - Construction and Demolition Waste
(ISBN: 2-912143-17-9); Eds. C. F. Hendriks and H. S. Pietersen
Report 23: Self-Compacting Concrete state-of-the-art report (ISBN: 2-912143-23-
3); Eds. Å. Skarendahl and Ö. Petersson
xxxii RILEM Publications
Report 24: Workability and Rheology of Fresh Concrete: Compendium of Tests

(ISBN: 2-912143-32-2); Eds. P. J. M. Bartos, M. Sonebi and A. K. Tamimi
Report 25: Early Age Cracking in Cementitious Systems (ISBN: 2-912143-33-0);
Ed. A. Bentur
Report 26: Towards Sustainable Roofing (Joint Committee CIB/RILEM) (CD 07)
(e-ISBN 978-2-912143-65-5); Eds. Thomas W. Hutchinson and Keith Roberts
Report 27: Condition Assessment of Roofs (Joint Committee CIB/RILEM) (CD
08) (e-ISBN 978-2-912143-66-2); Ed. CIB W 83/RILEM TC166-RMS
Report 28: Final report of RILEM TC 167-COM ‘Characterisation of Old Mortars
with Respect to Their Repair (ISBN: 978-2-912143-56-3); Eds. C. Groot, G. Ashall
and J. Hughes
Report 29: Pavement Performance Prediction and Evaluation (PPPE):
Interlaboratory Tests (e-ISBN: 2-912143-68-3); Eds. M. Partl and H. Piber
Report 30: Final Report of RILEM TC 198-URM ‘Use of Recycled Materials’
(ISBN: 2-912143-82-9; e-ISBN: 2-912143-69-1); Eds. Ch. F. Hendriks,
G. M. T. Janssen and E. Vázquez
Report 31: Final Report of RILEM TC 185-ATC ‘Advanced testing of cement-
based materials during setting and hardening’ (ISBN: 2-912143-81-0; e-ISBN:
2-912143-70-5); Eds. H. W. Reinhardt and C. U. Grosse
Report 32: Probabilistic Assessment of Existing Structures. A JCSS publication
(ISBN 2-912143-24-1); Ed. D. Diamantidis
Report 33: State-of-the-Art Report of RILEM Technical Committee TC 184-IFE
‘Industrial Floors’ (ISBN 2-35158-006-0); Ed. P. Seidler
Report 34: Report of RILEM Technical Committee TC 147-FMB ‘Fracture
mechanics applications to anchorage and bond’ Tension of Reinforced Concrete
Prisms – Round Robin Analysis and Tests on Bond (e-ISBN 2-912143-91-8);
Eds. L. Elfgren and K. Noghabai
Report 35: Final Report of RILEM Technical Committee TC 188-CSC ‘Casting of
Self Compacting Concrete’ (ISBN 2-35158-001-X; e-ISBN: 2-912143-98-5);
Eds. Å. Skarendahl and P. Billberg
Report 36: State-of-the-Art Report of RILEM Technical Committee TC 201-TRC
‘Textile Reinforced Concrete’ (ISBN 2-912143-99-3); Ed. W. Brameshuber
Report 37: State-of-the-Art Report of RILEM Technical Committee TC 192-ECM
‘Environment-conscious construction materials and systems’ (ISBN: 978-2-35158-
053-0); Eds. N. Kashino, D. Van Gemert and K. Imamoto
Report 38: State-of-the-Art Report of RILEM Technical Committee TC 205-DSC
‘Durability of Self-Compacting Concrete’ (ISBN: 978-2-35158-048-6);
Eds. G. De Schutter and K. Audenaert
RILEM Publications xxxiii
Report 39: Final Report of RILEM Technical Committee TC 187-SOC

‘Experimental determination of the stress-crack opening curve for concrete in
tension’ (ISBN 978-2-35158-049-3); Ed. J. Planas
Report 40: State-of-the-Art Report of RILEM Technical Committee TC 189-NEC
‘Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete
Cover’ (ISBN 978-2-35158-054-7); Eds. R. Torrent and L. Fernández Luco
Report 41: State-of-the-Art Report of RILEM Technical Committee TC 196-ICC
‘Internal Curing of Concrete’ (ISBN 978-2-35158-009-7); Eds. K. Kovler and
O. M. Jensen
Report 42: ‘Acoustic Emission and Related Non-destructive Evaluation
Techniques for Crack Detection and Damage Evaluation in Concrete’ - Final
Report of RILEM Technical Committee 212-ACD (e-ISBN: 978-2-35158-100-1);
Ed. M. Ohtsu
Report 45: Repair Mortars for Historic Masonry - State-of-the-Art Report of
RILEM Technical Committee TC 203-RHM (e-ISBN: 978-2-35158-163-6);
Eds. Paul Maurenbrecher and Caspar Groot
Report 46: Surface delamination of concrete industrial floors and other durability
related aspects guide - Report of RILEM Technical Committee TC 268-SIF
(e-ISBN: 978-2-35158-201-5); Ed. Valerie Pollet
Rheology and Early-Age Properties
Influence of Different Fibre Types
on the Rheology of Strain Hardening
Cementitious Composites
Hassan Baloch1(&), Steffen Grünewald1,2, Karel Lesage1,

and Stijn Matthys1
1
Ghent University – Ghent, Ghent, Belgium
Hassan.Baloch@ugent.be
2
Delft University of Technology – Delft, Delft, The Netherlands
Abstract. Strain-hardening cementitious composites (SHCC) have a high

tensile strength and display a remarkable strain-hardening behaviour. These
unique characteristics make them an interesting choice for improving the
strength and durability of new and existing structures. The tensile strain beha-
viour of SHCC is strongly influenced by its rheological properties as they
determine the hardened state behaviour such as fibre-bridging strength and
ultimately the degree of multiple cracking. The presence of fibres significantly
affects the rheological performance of SHCC.
This study aimed at investigating the relationship between rheological char-
acteristics of SHCC mortar before and after the addition of different fibres.
Polyvinyl alcohol (PVA), high modulus polyethylene (HDPE) and glass fibres
were added at three different contents in order to assess their effect on the
workability of SHCC. Flow tests along with rheological assessment were con-
ducted to evaluate the fresh state behaviour of SHCC. The addition of fibres
reduced the flowability of mix, especially at high dosages. A modified fibre
influence factor was developed to characterize different types of fibres and was
related to the viscosity and yield stress of the mix.
Keywords: Fibres Strain-hardening cementitious composites Rheology

Flowability
1 Introduction
Strain-hardening cementitious composites (SHCC) are a special class of material which

display an increase in stress beyond the formation of initial cracks under uniaxial
tensile loading [1]. This behaviour is usually achieved by the addition of discontinuous
fibres which bridge the cracks resulting in a reduced crack width and higher fracture
toughness [2]. In recent years extensive research has been carried out to improve the
mechanical properties of SHCCs resulting in ultra-high performance SHCC having
tensile strengths up to 15 MPa and a strains capacity of up to 8% [3, 4]. These
properties make SHCC an excellent choice not only as a construction material but also
as repair material for existing structures.
© RILEM 2021
P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 3–11, 2021.
https://doi.org/10.1007/978-3-030-58482-5_1
4 H. Baloch et al.
The addition of fibres comes with undesired effects on the flow and rheology of
concrete and should be taken into account when designing SHCC. Short synthetic
fibres are mostly used in SHCC to achieve micro-cracking behaviour. The type and
content of the fibres, the mixing regime and the binder content greatly influence the
rheological properties of the resulting mortar. It is important to analyse rheological
properties as they can cause variation in ductility and mechanical properties of SHCC.
Most work in this regard is done on rigid fibres which has shown that increasing the
fibre content decreases the workability of the fibre reinforced cementitious matrix
[5–7]. The influence of steel fibres is found to increase as the fibre content increases
This reduction in workability was observed to become non-linear at increasing fibre
content. At the end, these studies successfully demonstrated the correlation between
viscosity, yield stress and flow of steel fibre reinforced matrices. However, for plastic
and glass fibres these correlations are not established in detail. The objective of this
study was to develop correlations for different fibre types at varying fibre contents. This
study was carried out using PVA, HDPE and glass fibres at different fibre contents.
Flow and rheological correlations were established for all fibre types based on the
experimental results.
2 Materials and Methods
2.1 Materials
A suitable control mixture was designed with a target flow spread of 340 mm using
Haegermann’s flow test and fibre contents tested were 1.0, 1,5 and 2.0 vol.% of the
mixture keeping other parameters constant. Ordinary Portland Cement (CEM 1 52.5 N,
Holcim, Belgium) was used as binder. Very fine silica sand (M34, Sibelco, Belgium)
having an average particle size of 174 lm was used to increase the packing density of
the granular skeleton and to obtain a high strength matrix. A polycarboxylate-based
superplasticizer (Glenium 51, 35% con, BASF) was used to achieve the desired
workability. Three different types of fibres including PVA, HDPE and glass fibres were
tested having properties listed in Table 1.
Table 1. Properties of fibres used

Fibre type Aspect ratio Diameter (lm) Length (mm) Young Modulus (Gpa)
PVA 307 39 12 42.8
HDPE 600 20 12 80
Glass 857 14 12 72
2.2 Mix Formulations and Mixing Regime

A total of 10 formulations were investigated having a w/c ratio of 0.22. The amount of
cement, sand and superplasticizer (SP) were kept constant in order to better understand
the effect of fibres on the mixture characteristics. The mixture proportions along with
Influence of Different Fibre Types on the Rheology 5
constituent dosage per weight relative to the cement are listed in Table 2; the fibre
contents are represented as volume fraction of the mixture.
Table 2. Mixture formulations

a
Mix Cement Sand Water Fibre volume (%) SP
C 1 0.5 0.22 – 0.021
PE1 1 0.5 0.22 1.0 0.021
PE1.5 1 0.5 0.22 1.5 0.021
PE2 1 0.5 0.22 2.0 0.021
Glass1 1 0.5 0.22 1.0 0.021
Glass1.5 1 0.5 0.22 1.5 0.021
Glass2 1 0.5 0.22 2.0 0.021
PVA1 1 0.5 0.22 1.0 0.021
PVA1.5 1 0.5 0.22 1.5 0.021
PVA2 1 0.5 0.22 2.0 0.021
a
Fibre content is represented as volume fraction of mix,
while rest of the ingredients are weight proportion of
cement
A Hobart mixer was used to prepare the mortar mixes. A total of 6-min wet mixing
time was arranged to ensure a homogenous fibre dispersion throughout the matrix.
Cement and sand were first dry mixed in the Hobart mixer for 60 s at 145 rpm followed
by the addition of water and suerplasticizer (SP). The mixer was then operated at
145 rpm for 180 s followed by the fibre addition and 180 s of fast mixing at 285 rpm.
Fibres were added in 4 about equal batches to ensure an equal distribution of fibres.
2.3 Test Methods

2.3.1 Flow Tests
Flow tests of SHCCs mortars (without compacting action) were conducted using
Haegermann’s cone with a height of 60 mm and diameters of 70 and 100 mm at the top
and the bottom of the cone, respectively. The value of SP was adjusted to maximize the
flow in the control formulation without any apparent segregation. This SP value was
then kept constant for formulations containing fibres.
2.3.2 Rheological Investigation

The rheological parameters were measured using the rotational type MCR 52
Rheometer produced by Anton Paar. This rheometer consists of a fixed rheometer cup
having an internal diameter of 70 mm and a 39 mm rotating probe cylinder attached to
the motor. SHCC mortars were prepared using the specified mixing regime and were
poured in the rheometer cup. Excess material was carefully scraped off using a spatula
and the filling level of material was adjusted so that it won’t flow out after probe
insertion. The cup was then fixed in the machine and the rotating probe was inserted to
start the measurement.
6 H. Baloch et al.
The shear rate was increased from 1 1/s to 100 1/s with 100 measuring points in
between (2 s point duration). Assuming a plastic Bingham model, the average yield
stress value is calculated as the axis intercept of the linear regression line of the
measured points while the viscosity is calculated as the slope of the regression line.
A test resulted in the shear rate-shear stress curve, further interpolations were then
executed to determine the yield values and viscosities of the mixtures.
3 Results and Discussion
3.1 Flowability
3.1.1 Measurements of Flow Spread
Table 3 shows the results of flow spread tests carried out. The addition of different
fibres considerably reduced the flow spread of the specimens compared with the control
specimen. The increase in fibre content results in a non-proportional decrease in
flowability to the point where almost no flow is observed at a fibre content of 2 vol.%.
This effect is visually depicted in Fig. 1 for HDPE fibres.
Table 3. Flow spread test results.

Mix Spread (cm)
C 34.0
PE1 17.3
G1 18.5
PVA1 18.8
PE1.5 12.5
G1.5 11.5
PVA1.5 12.8
PE2 10.7
G2 10.3
PVA2 11.0
The flowability dependent on both type and content of fibres. Glass fibres had a
similar effect as the other fibre types at 1 vol.% fibre content, however they had a more
pronounced effect on the workability at higher fibre contents. This outcome can be
explained by the higher aspect ratio of glass fibres compared with the other fibre types.
3.1.2 Effect of the Fibre Factor

The fibre factor of fibres is calculated by multiplying the volume of the fibres with the
respective aspect ratio [Vf*L/D] [8]. This factor was applied to assess the flow mea-
surements and the results are displayed in Fig. 2a. It can be seen that the fibre factor
Fig. 1. Effect of HDPE fibre addition on the flow spread
does not corelate well when considering the different fibre types. This deviating effect
can be explained by an entirely different physical and chemical nature of the fibres. As
this study also was carried out with glass and plastic fibres, a more comprehensive
factor was needed to take into account the different nature of the studied fibres.
A generalized fibre influence factor was then developed to accurately predict the tests
results obtained with the studied fibres. Considering different fibre materials, the
modulus of elasticity was also included as dependent variable [9]. After incorporating
all governing parameters, the fibre influence factor is stated as:
F ¼ Vf La Db Ec ð1Þ
Where a, b,c are influence coefficients. Chu et al. [10] also used a similar type of fibre
influence indicator for FRC mixes without E parameter as only steel fibres were studied
and E was constant. After performing several regression analyses, the authors found
that a = 1, b = −0.5 and c = −0.5 results in the best correlation. The best fitting curve
is plotted in Fig. 2b. It can be seen that Vf LaDb Ec seems to be a more suitable factor
than the traditional Vf (L/D) which is overly simplistic in order to take into account
different fibre types.
40 40
y = 1E-05x2 - 0,0308x + 29,858 y = 0,3074x2 - 5,4424x + 34,026
35 R² = 0,7322 35 R² = 0,9742
30 30
Flow Spread (cm)
Flow Spread (cm)
25 PE
25
PE PVA
20 20
PVA
Glass
15
Glass 15
10
10
5
5
0
0 500 1000 1500 2000 0
0 2 4 6 8 10
Vf (L/D) Vf Lα Dβ Eγ
(a) (b)
Fig. 2. a. Traditional fibre factor b. Modified fibre influence factor

8 H. Baloch et al.
Another interesting observation is that the flow spread decreased at increasing fibre
dosage. With increase in fibres content the decrease in flow spread is less than pro-
portional which contradicts with other studies carried out on steel fibre types. This
trend can be explained by the very stiff consistency of mixes at 2 vol.% fibre dosage as
the flow spread is close to 100 mm, which is the base diameter of the flow cone itself.
3.2 Rheological Investigation

The shear stress was measured at increasing shear rate. The effect of different fibre
additions is shown in Fig. 3 for different fibre types. It can be seen that because of the
addition of the fibres the curves are not completely straight which is often the case with
plain mortars or concretes, which are typical Bingham materials. The yield value was
obtained as the y-axis intercept of each curve idealised as a straight line and the plastic
viscosity was determined by calculating the slope of the linear model.
3.2.1 Relation Between Yield Stress and Flow Spread

In order to study the effect of different shear-rates, the curve was assessed for five
maximum shear-rate values of 20, 40, 60, 80 or 100 1/s. These values were plotted
against the flow spread as shown in Fig. 4. A considerable increase in the yield stress is
noted at increasing shear rate. At a shear rate of 20 1/s an almost linear increase in yield
stress is noted at decreasing flow spread, however as the shear rate increased to 100 1/s,
a more than proportional rise in yield stress is observed. This can be explained by the
higher mechanical energy required during the initial seconds to get the mix flowing at
higher shear rates. This leads to a high torque registration, which is translated in a high
yield stress.
3.2.2 Relation Between Viscosity and Fibre Influence Factor

The addition of more fibres increases the internal friction among the fibres. In order to
characterize the relation between plastic viscosity and fibre content, the viscosity values
were calculated from rheological readings and plotted against both fibre factor and the
modified fibre influence factor at three different strain-rate ranges as shown in Fig. 5.
R2 values along with equations were plotted for each trendline which depicts a better
correlation of viscosity with the modified fibre influence factor. The plastic viscosity
increased at increasing fibre factor in both cases. With a maximum shear rate of 20 1/s
there is a steep increase in viscosity values at increasing fibre influence factor, however
as the strain-rate increases, the viscosity values drop indicating the breakdown of the
mixes. A more preferred fibre orientation during testing could be an explanation for the
decrease in viscosity, which first increases at increasing fibre dosage and is about
constant at a higher fibre influence factor.
1400
1400
1200
1200
1000 1000
Shear Stress (Pa)

Shear Stress (Pa)
800 800
600 600
PVA 1 PE1
400 400
PVA1.5 PE 1.5
200 200
PVA2 PE2
0 0
0 20 40 60 80 100 0 20 40 60 80 100
She ar rate 1/s She ar rate 1/s
(a) (b)
1400
1200
1000
Shear Stress (Pa)
800
600
Glass1
400
Glass1.5
200
Glass2
0
0 20 40 60 80 100
She ar rate 1/s
(c)
Fig. 3. a. Shear stress-shear rate curves for PVA fibres b. Shear stress-shear rate curves for PE
fibres c. Shear stress-shear rate curves for glass fibres
20 20 1/s 20
60 1/s 20
100 1/s
18 18
18
Slump Flow (mm)
Slump Flow (mm)
Slump Flow (mm)
16 16
16
14 14
14
12 12
12
10 10
10
y = 2E-05x2 - 0,0572x + 21,976 8 y = 0,0001x2 - 0,0982x + 27,109 8 y = 0,0001x2 - 0,1033x + 31,088

8
R² = 0,7474 R² = 0,8163 R² = 0,8152
6 6 6
0 100 200 300 400 0 100 200 300 400 0 100 200 300 400
Yield Stress (Pa) Yield Stress (Pa) Yield Stress (Pa)
Fig. 4. Relationship between yield value and flow spread at different shear rates
10 H. Baloch et al.
25 25
20 1/s 60 1/s 25
100 1/s
y = -6E-06x2 + 0,016x + 2,0343 y = -4E-06x2 + 0,0104x + 2,8298
20 R² = 0,3923 R² = 0,3096
20 20
Viscosity (pa.s)
Viscosity (pa.s)
Viscosity (pa.s)
15 15
15
10 10
10
y = 5E-07x2 + 0,0045x + 11,885 5 5

R² = 0,3237 200 700 1200 1700 200 700 1200 1700
5 Vf (L/D)
200 700 1200 1700 Vf (L/D)
Vf (L/D)
(a)
25 25
20 1/s 60 1/s 25
100 1/s
y = -0,4997x2 + 7,241x - 13,773 y = -0,358x2 + 5,1022x - 8,6393
20 20 R² = 0,7853 20 R² = 0,7282
Viscosity (pa.s)
Viscosity (pa.s)
Viscosity (pa.s)
15 15 15
10 10 10
y = -0,0005x2 + 1,8979x + 5,2359
R² = 0,6544
5 5 5
2 4 6 8 10 2 4 6 8 10 2 4 6 8 10
Vf L1 D-0.5 E-0.5 Vf L1 D-0.5 E-0.5
Vf L1 D-0.5 E-0.5
(b)
Fig. 5. a. Relationship between viscosity and fibre factor b. Relationship between viscosity and
modified fibre influence factor
4 Conclusions
This paper discusses the effect of different fibre types and contents on flow and rhe-
ological properties of SHCC. An experimental program of 10 SHCC mixes was exe-
cuted with 3 different fibre types with three contents per fibre type. Based on the
experimental results, the following conclusions can be drawn:
• The addition of fibres significantly reduced the flow spread of the mixes. This effect
is more pronounced at higher fibre dosages and should be accounted for in the mix
design.
• The yield stress exhibits an almost linear relation with flow spread when assessing
the flow curves at a lower shear rate of 20 1/s, however the yield stress increases
significantly as the shear-rate values increased.
• The effect of fibres on the flowability depends not only on the fibre content but on
the type of fibres as well. A modified fibre influence factor was developed to depict
fibre behaviour resulted in a better prediction ability compared with the traditional
fibre factor.
• The plastic viscosity increased steeply with increase in fibre influence factor at
lower shear rates. No significant increase in viscosity was observed at higher strain
rates.
Acknowledgements. The authors would like to acknowledge the financial support of the Higher
Education Commission, Pakistan [HRDI-UESTP (BATCH-V)]. The support of Owens-Corning
for free supply of glass fibres used in this research is also appreciated.
References
1. Kong, H.J., Bike, S.G., Li, V.C.: Constitutive rheological control to develop a self-
consolidating engineered cementitious composite reinforced with hydrophilic poly(vinyl
alcohol) fibers. Cem. Concr. Compos. 25(3), 331–341 (2003). https://doi.org/10.1016/
S0958-9465(02)00056-2
2. Li, V.C., Mishra, D.K., Wu, H.C.: Matrix design for pseudo-strain-hardening fibre
reinforced cementitious composites. Mater. Struct. 28, 586–595 (1995). https://doi.org/10.
1007/BF02473191
3. Ranade, R., Li Prof, V.C., Rushing, M.D., Roth, J., Heard, W.F.: Micromechanics of high-
strength, high-ductility concrete. ACI Mater. J. 110, 4–375 (2013). https://doi.org/10.14359/
51685784
4. Yu, K., Wang, Y., Yu, J., Xu, S.: A strain-hardening cementitious composites with the
tensile capacity up to 8%. Constr. Build. Mater. 137, 410–419 (2017). https://doi.org/10.
1016/j.conbuildmat.2017.01.060
5. Grünewald, S., Walraven, J.C.: Parameter-study on the influence of steel fibers and coarse
aggregate content on the fresh properties of self-compacting concrete. Cem. Concr. Res. 31
(12), 1793–1798 (2001). https://doi.org/10.1016/S0008-8846(01)00555-5
6. Stähli, P., van Mier, J.G.: Rheological properties and fracture processes of HFC’-in
proceeding BEFIB (2004)
7. Grunewald, S.: Performance-based design of self-compacting fibre reinforced concrete. Delft
University Press (2004)
8. Grunewald, S., Walraven, J.C.: Rheological study on the workability of fibre-reinforced
mortar. In: Ozawa, K., Ouchi, M., (Eds.), Proceedings of the second international
symposium on self-compacting concrete, pp. 127–136 (2001)
9. Martinie, L., Rossi, P., Roussel, N.: Rheology of fiber reinforced cementitious materials:
classification and prediction. Cem. Concr. Res. 40(2), 226–234 (2001). https://doi.org/10.
1016/j.cemconres.2009.08.032
10. Chu, S.H., Li, L.G., Kwan, A.K.H.: Fibre factors governing the fresh and hardened
properties of steel frc. Constr. Build. Mater. 186, 1228–1238 (2018). https://doi.org/10.1016/
j.conbuildmat.2018.08.047
Using Fiber Reinforced Concrete to Control
Early-Age Shrinkage in Replacement Concrete
Pavement
Nakin Suksawang1(&) and Daniel Yohannes2

1
Mechanical and Civil Engineering, Florida Institute of Technology,
Melbourne, USA
nsuksawang@fit.edu
2
Structural Division, SBA Communications, Boca Raton, USA
Abstract. Unlike ordinary concrete pavement, replacement concrete pavement

needs to be open to traffic within 24 h (sooner in some cases). Thus, high early-
strength concrete is used; however, it frequently cracks prematurely as a result
of high heat of hydration that leads the slab to develop plastic shrinkage. FRC is
known to provide good resistance to plastic shrinkage. This paper explores the
potential use of fiber- reinforced concrete (FRC) in concrete pavement
replacement particularly in controlling plastic shrinkage. Five different fiber
types, including steel, glass, basalt, nylon, and polyethylene fibers were inves-
tigated. Additionally, the effect of fiber length was also investigated for the
polyethylene fiber. The fibers were added at low dosage amounts of 0.1% and
0.3% by volume. A retrained shrinkage test was conducted to assess the
cracking potential of the concrete mixtures and the ability for each fiber type to
resist cracking. Results indicated that both polyethylene and nylon fibers pro-
vided the best resistance to early-age shrinkage. Short fibers (<1-in.) also had the
best performance in resisting early-age shrinkage, while long fibers (>1-in.)
provided additional post-cracking capacity. For replacement concrete pavement,
it is recommended that a short polyethylene fiber be used to eliminate uncon-
trolled cracking.
Keywords: High early-strength concrete Concrete pavement FRC

Polyethylene fibers Cracking Plastic shrinkage
1 Introduction
The Florida Department of Transportation (FDOT) Design Standards [1] require a full
depth replacement of concrete slab with severe distresses. The construction standards
and requirements of the replacement slab are provided in Section 353 of the Standard
Specifications for Road and Bridge Constructions [2]. Two of the most important
acceptance criteria for the replacement slab are the plastic property, specifically the 6-h
compressive strength of 2,200 psi (15.2 MPa), and the 24-h compressive strength of
3,000 psi (20.7 MPa). The 6-h compressive strength is also used as the determination
point to open the slab to traffic, and therefore, it is highly emphasized in the standard
specifications. In fact, if the replacement slab does not meet the plastic property
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_2
Using Fiber Reinforced Concrete 13
requirements and the engineer determines that this will severely impact the replacement
slab service life, the contractor would have to replace the slab at no cost to the FDOT.
Thus, to ensure the replacement slab meets the plastic property requirements, low
water-cement ratio concrete and concrete accelerators are used. As a consequence, the
heat of hydration increases causing larger early-age shrinkage, which could potentially
lead to cracking [3–7]. As a consequence, the FDOT also specified a limit on the
concrete temperature not to exceed 100 °F (38 °C) and requires the contractor to cure
the slab with curing compound and cover the surface with white burlap-polyethylene
curing blanket immediately after the slab hardens. Furthermore, if uncontrolled cracks
appear on the replacement slab during the life of the contract, the contractor will have
to replace the slab free of charge.
While Section 353 is comprehensive and provides many levels of protection to
FDOT, replacement slabs do crack. If cracks are discovered during the contract period,
then the contractor will be responsible for replacing the slab again at his expense.
However, some contractors may file a lawsuit against the FDOT to avoid their con-
tractual obligation. Regardless of the outcome, further delay to the roadway, in which
the slab is constructed, opening to traffic would have a direct impact on the traveling
public. Therefore, there is a need in a better solution that can prevent or at least
minimize the number of cracks for replacement slabs.
One proposed solution is to replace ordinary concrete used in the construction of
replacement slabs with fiber reinforced concrete (FRC). The use of FRC in slab-on-
grade is not new as the building industry had been benefitting from it for over 30 years
[8]. However, unlike slabs-on-grade in buildings, replacement slabs on roadways are
exposed to the outdoor environment and need to withstand heavy truck traffic.
FRC can enhance concrete in many different ways. Concrete fracture toughness and
ductility can be improved using steel and synthetic fibers at high-dosage amount (>1%
by volume) [9–13]. The glass fiber can improve fracture toughness but does not pro-
vide ductility enhancement. At low-dosage amount (<0.5%), synthetic fibers can be
used to provide concrete resistance to plastic shrinkage and have been used for many
years in building construction for the construction of slab-on-grade [13]. Despite all
these advantages, there are some drawbacks with the use of fiber as well. First, there is
no standard specification for its application in civil infrastructure. Second, exposed
steel fibers would corrode particularly in Florida’s environment. Third, the FRC needs
to be externally vibrated. Internal vibration would cause the fibers to cling together.
Fourth, during mixing the fibers tend to cling together forming clumps and balls instead
of being evenly distributed throughout the concrete matrix. These clumps and balls
cause void pockets in the concrete, which potentially lead the concrete to fail prema-
turely. This balling effect could be even more pronounced when used for concrete
pavement slab replacement because of the lean concrete mixture and low workability
needed for immediately opening the replacement slab to traffic. Furthermore, the
impact on basalt fiber on concrete cracking performance is not well established. For
these reasons, more research is needed to better understand the characteristic of FRC
with high-early strength.
14 N. Suksawang and D. Yohannes
2 Methodology
As a first step in exploring the potential used of FRC in concrete pavement slab
replacement, the mixture proportion, mixing procedure, workability, and method of
consolidation were investigated. The mixture proportion was based on FRC manu-
facturer recommendations and published articles on the use of FRC in preventing early-
age shrinkage. Based on this information, low-dosage (<0.5% by volume) FRC was
used and evaluated for its potential in eliminating early-age cracking. Two mixing
procedures were investigated: 1) mixing the fiber to the dried ingredients before adding
water to the mixture; and 2) mixing the fiber to plastic concrete after all ingredients,
including the water, have been mixed together. Both procedures provided good dis-
tribution of fiber in concrete but the latter procedure was selected because the ease of
preparation under field condition. As there is no active standard to evaluate the
workability of FRC, withdrawn standard, ASTM C995, was used. This standard
determines the workability in term of time of flow of FRC mixture through an inverted
slump cone under internal vibration. This standard was withdrawn because of the
inherent bias in the result generated by the variability in the operation of the vibrator.
Thus, to minimize the bias in the result, a designated operator was used in this study to
control a constant frequency and vibrator’s operation. The FRC mixture was consol-
idated using external vibration (i.e., vibrating table) that had been calibrated to operate
at a constant frequency to yield the best consolidation.
2.1 Mixture
Specimens were prepared with Type II Portland cement, silica sand and limestone and
various chemical admixtures, i.e., high range water reducers, air-entraining agent and
accelerator. All mixtures were proportioned to comply with the FDOT provisions for
minimum required workability and strength. It should be noted that no attempt was
made to optimize the mixture containing fibers in order to isolate the fiber’s effect on
concrete properties. Two dosages of fibers were used in this project. A dosage of 0.1%
by volume of fibers was used for mixtures M02–M08, while the last two mixtures, M09
and M10 had a fiber volume fraction of 0.3%. Mixture M01 was a control mixture with
no added fiber content. All mixtures had the same mixture proportion except for the
fiber type and dosage amount used. The general mixture proportion and fiber content
are given in Tables 1 and 2.
2.2 Laboratory Tests

The laboratory tests that were conducted were i) compressive strength, ii) modulus of
rupture, iii) flexural toughness, (iv) residual strength, and v) restrained shrinkage tests.
The compressive strength test was carried out in accordance with ASTM C39 using
4 8-in. (100 200 mm) cylindrical specimens at 6-h, 24-h, and 28-days. The
modulus of rupture, flexural toughness and residual strength tests were tested using
four-point flexural test in accordance with ASTM C78, C1018, and C1399, respec-
tively. Figure 1 illustrates the compressive strength and four-point flexural tests.
Table 1. Mixture proportion

Materials lb/yd3 kg/m3
Type II Portland cement 840 498
Water 318 189
Silica sand 1200 712
#57 limestone 1500 890
AEA 2.1 oz/yd3 0.08 L/m3
Type C accelerator 672 oz/yd3 26 L/m3
Type F HRWR 135 oz/yd3 5.2 L/m3
Table 2. Fiber type and volume fraction of the mixtures

Mix Fiber type Volume fraction (%)
M01 None 0
M02 Polypropylene 0.5ʺ multifilament 0.1
M03 Steel 0.5ʺ monofilament 0.1
M04 Glass 0.5ʺ monofilament 0.1
M05 Basalt 0.5ʺ monofilament 0.1
M06 Nylon 0.5ʺ monofilament 0.1
M07 Polypropylene 1.5ʺ macro synthetic 0.1
M08 Polypropylene ¾ʺ fibrillated 0.1
M09 Nylon 0.5ʺ monofilament 0.3
M10 Polypropylene ¾ʺ fibrillated 0.3
Fig. 1. Compressive strength and four-point flexural tests
The restrained shrinkage test was performed in accordance with the ASTM C1581
with slight modifications. A 6-in. (150 mm) concrete ring with inside diameter of 13-
in. (330 mm) and outer diameter of 18-in. (460 mm) was cast around a 0.5-in. (13 mm)
thick steel ring. The specimen was placed on a nonabsorbent base. Six hours after
casting, the molds were removed from the specimens and the top surface of the con-
crete was coated with paraffin wax to avoid evaporation. The specimen was left to dry
only from the sides at an ambient temperature of 73 ± 2 °F (23 ± 1 °C) and a relative
of humidity of 50 ± 5%. Two strain gages were placed at mid height on diametrically
opposite sides of the steel ring to measure the induced steel strain. A data-logger was
used to continuously collect the data from different specimens. The data was then
plotted to obtain the age at cracking. A sudden drop in strain of more than 30
microstrain indicates a crack. The specimens were inspected for visible cracks with the
aid of a digital microscope. A Dino-Lite AM-413TA digital microscope with the
capabilities to detect micro cracks and measure the width of cracks was used. The crack
width at the end of 28 days was then recorded for each specimen. Figure 2 illustrates
the testing detail of the restrained shrinkage test.
Fig. 2. Restrained shrinkage test setup
3 Results
Plastic properties, mechanical properties and cracking performance as well as any

observations made during mixing are reported below.
3.1 Plastic Properties

Two plastic properties were of concern in this project consisted of 1) the unit weight
and 2) the workability of the concrete. The workability of the concrete was tested using
the inverted slump cone method shown. The time of flow of the concrete through an
inverted cone was measured. The results of the unit weight and time of flow through an
inverted cone are given in Table 3 for the different mixtures. According to the test
results there was no notable reduction in workability due to the addition of fibers. There
was a slight reduction in the time of flow in the case of stiff fibers, i.e., steel, glass and
long (1.5-in.) polypropylene fibers. However, the reduction was insignificant with only
about 2 to 3 s increase. It should be noted that a 15 s time of flow is equivalent to a 1-
in. slump, which was the targeted slump for this particular control mixture used in this
study. Therefore, unlike FRC with high-dosage (>1%) fiber content that has lower
workability reported in the literatures, the workability of FRC with low-dosage
(<0.5%) fiber content is not affected. The mixture can further be optimized to increase
the workability of the concrete by using superplasticizer but the Research Team
decided to stick to the targeted slump of the control mixture obtained from FDOT.
Table 3. Plastic properties of different mixtures

Mix Time of flow (sec) Unit weight (lb/ft3)
M01 14 141.0
M02 14 141.0
M03 17 141.3
M04 15 141.1
M05 14 141.1
M06 13 141.1
M07 16 141.1
M08 15 141.1
M09 15 141.1
M10 14 141.1
3.2 Mechanical Properties

The mechanical properties that were investigated in this study were compressive
strength, modulus of rupture, flexure toughness, and residual strength.
3.2.1 Compressive Strength

The results of the compressive strength tests at three different specimen ages (6 h, 24 h
and 28 days) are illustrated Fig. 3. As can be seen from the test results, in general FRC had
higher compressive strength at early-age but then there was little effect on the com-
pressive strength when comparing to ordinary concrete. Higher effects could be achieved
at higher volume fractions, but at a low-dosage of 0.1% and 0.3% used in this project,
there was no significant increase in the compressive strength. One noticeable difference
was the 24-h compressive strength of FRC was in general higher (could be as much 1000
psi (7 MPa)) in comparison to ordinary concrete (control mixture). This could be
attributed to the fiber confining the concrete at early-age. The specimens with fiber
showed higher shutter resistance as compared to the control specimens. Unlike the fiber
reinforced concrete, the plain concrete specimens were shattered, while the former stayed
in one piece after been subjected to ultimate load. All mixtures had an initial setting time
of approximately 5 h. Another observation that was made was the mixture containing
nylon fiber had a strength reduction at 28 days, which could be attributed to fiber balling
encountered for these mixtures that prevented proper consolidation and fiber distribution.
12000 83 MPa
10000 69 MPa
Compressive Strength (psi)
8000 55 MPa
6000 41 MPa
4000 27 MPa
14 MPa
2000
0 6 hours
M01 M02 M03 M04 M05 M06 M07 M08 M09 M10 24 hours
Mixtures 28 days
Fig. 3. Compressive strength of mixtures at different ages
3.2.2 Modulus of Rupture

The modulus of rupture is a measure of flexural strength. The behavior of the FRC was
expected to be superior as compared to ordinary concrete. However, the fiber volume
fraction chosen was intended to improve the early age cracking resistance, not the
flexural strength. As a result, only a small increase in capacity was observed. It is also
observed that fibers with higher stiffness had the highest modulus of rupture. The
results of the modulus of rupture test are summerized in Table 4. As can be seen from
the test results, almost all mixtures had low modulus of rupture. This is typical of an
ordinary concrete which had low flexural strength. By introducing a fiber into the
mixture, the modulus of rupture was increased only to a small degree. To attain
significant improvements the fiber volume fraction should be much higher. In addition,
the tensile and bond strengths of the fiber should be higher. In this test fibers with
higher tensile strength performed better than those with lower tensile strengths. Steel
and 1.5-in. (38 mm) monofilament polypropylene fibers increased the modulus of
rupture by 21% and 31%, respectively. Although the percentages showed good
improvements, in reality it is not as high as the percentages. Two modes of failure were
identified in this test. For fibers with lower tensile strength, the mode was tensile failure
in concrete immediately followed by tensile failure of the fibers. For fibers with higher
tensile strength the mode was tensile failure of concrete followed by a delayed pull out
of the fibers. Unlike in the first mode of failure discussed above, in the second mode of
failure the specimen stayed intact after the onset of tensile cracking. Another factor was
the fiber length. The longer the fiber, the better the bond strength. An ideal fiber to
boost the modulus of rupture would be one that has higher tensile and bond strengths
and are long.
3.2.3 Residual Strength

The residual strength test was conducted to compare the post cracking strength of
different mixtures. The result of the residual strength test is summarized in Table 4. As
expected, the ordinary concrete mixture had no post-cracking residual strength. The
analysis of the test results clearly showed that most of the mixtures had no post-
cracking strength due to the lower fiber volume fraction used. Fibers with higher tensile
strength have showed some post cracking residual strength. For example, mixture
containing steel fiber (M03) showed the highest post-cracking residual strength.
Mixtures containing nylon and polypropylene fibers (M06–M10) also showed some
residual strength as well. The rest of the mixtures could not be reloaded after cracking
as the specimen shattered into two pieces. Hence the reloading load was taken as zero.
Table 4. Summary of modulus of rupture and residual strength of the mixtures

Mix Average modulus of rupture Average residual strength
M01 515 psi (3.55 MPa) 0
M02 520 psi (3.59 MPa) 0
M03 625 psi (4.31 MPa) 110 psi (0.76 MPa)
M04 555 psi (3.83 MPa) 0
M05 615 psi (4.24 MPa) 0
M06 575 psi (3.96 MPa) 45 psi (0.31 MPa)
M07 675 psi (4.65 MPa) 90 psi (0.62 MPa)
M08 515 psi (4.24 MPa) 50 psi (0.34 MPa)
M09 590 psi (4.07 MPa) 60 psi (0.41 MPa)
M10 620 psi (4.27 MPa) 70 psi (0.48 MPa)
3.2.4 Flexural Toughness

Flexural toughness is one means of measuring the post-crack strength of FRC. An
attempt was made to find the flexural toughness parameters of the different mixtures
used in this project. However, the flexural toughness parameters could not be computed
for most of the mixtures. This was because most of the specimens split in two at first
crack, hence it was not possible to measure further. The results of the flexural
toughness test are given in Table 5. As can be seen from Table 5, flexural toughness
indices were computed for mixtures M03 and M07 only, which stayed intact after the
first crack. Comparing the residual strength factor R5,10, M03 proves to be superior to
the rest of the mixtures.
3.3 Cracking Performance

The cracking performance was evaluated using the restrained shrinkage test. The
restrained shrinkage test was carried out to compare the relative potential of different
fiber reinforced mixtures in preventing and controlling early age shrinkage cracking.
The results of the shrinkage test are plotted in Fig. 4. Two parameters were of interest
in this test: 1) the age at cracking and 2) the crack width of a 28-day-old specimen. The
age at cracking was determined from the strain-age graphs. A full depth crack was
detected when there was a sudden drop in strain (more than 30 le) or a consistent drop
of strain. Some cracks were superficial which do not penetrate to full depth of the
specimen. Superficial cracks were noted by a localized drop in the strain-age graph. For
comparative reasons the age at cracking was taken at the onset of cracking in case
superficial cracks.
A closer look of the strain-age graphs for the first 8 mixtures (M01–M08) revealed
that all but M02 and M08 had developed full depth cracks within the 28 days period
after casting. This was clearly indicated by a sudden drop in strain. However, mixtures
M02 and M08 had no indications of a superficial crack, which did not develop to full
depth crack within the 28-day period. This result is further bolstered by the reduced
crack width for these two mixes.
Table 5. Flexural toughness test results

Mix First- First-crack First-crack First-crack I5 I10 R5,10
crack Deflection Strength Toughness
Load
M01 2720 lb 0.0008 in 510 psi 1.1 lb-in. – – –
(12.1 kN) (0.020 mm) (3.52 MPa) (124 N-mm)
M02 2800 lb 0.0009 in 525 psi 1.3 lb-in. – – –
(12.5 kN) (0.023 mm) (3.62 MPa) (147 N-mm)
M03 3281 lb 0.0010 in 615 psi 2.0 lb-in. 4.1 7.7 72
(14.6 kN) (0.025 mm) (4.24 MPa) (226 N-mm)
M04 2933 lb 0.0009 in 550 psi 1.3 lb-in. – – –
(13.0 kN) (0.023 mm) (3.79 MPa) (147 N-mm)
M05 3307 lb 0.0009 in 620 psi 1.5 lb-in. – – –
(14.7 kN) (0.023 mm) (4.27 MPa) (169 N-mm)
M06 3067 lb 0.0009 in 575 psi 1.4 lb-in. – – –
(13.6 kN) (0.023 mm) (3.96 MPa) (158 N-mm)
M07 3548 lb 0.0011 in 665 psi 2.3 lb-in. 4.2 6.7 50
(15.8 kN) (0.028 mm) (4.59 MPa) (260 N-mm)
M08 3253 lb 0.0010 in 610 psi 1.6 lb-in. – – –
(14.5 kN) (0.025 mm) (4.21 MPa) (181 N-mm)
M09 3120 lb 0.0011 in 585 psi 1.7 lb-in. – – –
(13.9 kN) (0.028 mm) (4.03 MPa) (192 N-mm)
M10 3280 lb 0.0012 in 615 psi 2 lb-in. – – –
(14.6 kN) (0.030 mm) (4.24 MPa) (226 N-mm)
The crack width for all mixtures was measured at 28 days after casting. The
measurement of the crack width also is in complete accord with the strain-age graphs.
Mixtures M02, M07 and M08 had registered the smallest crack width as compared with
the rest of the mixtures.
As far as age at cracking and crack width were concerned, polypropylene fibers
have shown the best results. Mixtures M02 and M08 showed a combined effect of
longer age at cracking and the smallest crack width without developing into a full depth
crack within a 28-day period. The age at cracking increased by 62% (from 13 days for
the ordinary concrete, M01, to 21 days for polypropylene FRC, M02 and M08). The
crack width had also been reduced by 84% (0.4 mm for M01 to 0.065 mm for M02 and
M08). The steel fiber reinforced concrete, M03, had also shown smaller crack width.
This is due to the fact that two major cracks developed, hence the total crack width can
safely be assumed to be twice as much as the measured value. In addition, the steel
fibers had shown severe corrosion even in laboratory condition. There were also
multiple cases where the crack width was found to be greater than the crack in the
control specimen. M03 and M07 had actually increased the crack width and lessened
the age at cracking as compared to the control specimen. This can be related to the
relatively higher flexural strength of the fibers used in these mixtures. M04 and M05
had highly scattered age at cracking (indicated by the standard deviation of the age at
cracking) with an average age of slightly higher than the control specimen, however,
with a wider crack. M06 had extended the age at cracking considerably, however, the
crack width remained the same as the control specimen.
After assessing the results of the first eight mixtures, two mixtures (M09 and M10)
were tested at a higher fiber dosage amount. The fiber volume fraction was increased to
0.3% to investigate the effect of increased fiber volume. In both mixtures, no full depth
crack was detected. Hence, it was required to increase the observation period to 40
days. Within this period both mixtures only developed superficial cracks, no full depth
35 0.065
Age at cracking (days)
30
25 0.4 0.065
20 0.45
0.065
15 0.1
10
0.4
0.2 0.45
5 0.6
0
0 1 2 3 4 5 6 7 8 9 10
Crack width (mm) Mix number
Note: The size of the bubbles represents the crack width
Fig. 4. Age at cracking and crack width for the different mixtures
crack or a sudden drop in strain was recorded. Although M09 had a better performance
in reducing the crack width and preventing a full depth crack, it developed the
superficial cracks at a much earlier time as compared to M10. M10 on the other hand
increased the age at cracking considerably as compared to the control mixture as well
as all the other mixtures. With M10 mixture, it is possible to reduce the crack width by
84% (from 0.4 mm for M01 to 0.065 mm for M10) and extend the age at cracking by
138% (from 13 days for M01 to 31 days for M10).
4 Conclusions
The following conclusions can be made from this project:

1. It can also be concluded that there will not be significant influence on workability
and unit weight due to low volume fraction fiber addition to concrete.
2. Adding a fiber to concrete can increase the early age compressive strength by up to
48%. Furthermore the age at cracking can be more than doubled by just adding
0.3% volume of polypropylene fibers and the crack width is reduced to a fourth.
Other fiber can also be used for intermediary effects. Steel fibers are found to
provide residual strength, however they are also prone to deterioration due to
corrosion. The steel fibers used were found to be rusted even under laboratory
conditions.
3. When viewed holistically the use of fiber reinforced concrete for pavement slab
replacement has an advantage of controlling early age cracking and increasing early
age strength. Moreover the choice of fiber should be given due consideration as
different types of fibers perform differently. A fiber with high tensile strength,
higher pull out strength and lower flexural strength will be the best candidate to
control early age shrinkage cracking. The ultimate choice of fiber type and volume
fraction would depend on the desired effect or property. As far as controlling early
age shrinkage cracking is concerned a mix with a combined effect of higher early
age compressive strength, longer age at cracking and smallest crack width would be
an ideal candidate.
Acknowledgements. This research was funded by the Research Center of the Florida Depart-
ment of Transportation (FDOT) under the direction of Dr. Darryll Dockstader. We are particu-
larly grateful to our Project Manager, Mr. Donald Vanwhervin, of the FDOT District 4/6
Materials Office, for his guidance and support throughout the project. Thanks to Titan America,
BASF, and Cemex for their donation of materials for this study.
References
1. Florida Department of Transportation (FDOT): Design Standards, Topic No. 625-010-003,
Roadway Design Office, Tallahassee, FL (2019)
2. Florida Department of Transportation (FDOT): Standard Specifications for Road and Bridge
Construction, Specifications and Estimates Office, Tallahassee, FL (2019)
3. Bentz, D.P., Peltz, M.A.: Reducing thermal and autogenous shrinkage contributions to early-
age cracking. ACI Mater. J. 105(4), 414–420 (2008)
4. Bentz, D.P., Peltz, M.A., Winpigler, J.: Early-age properties of cement-based materials. II:
influence of water-to-cement ratio. J. Mater. Civ. Eng. 21(9), 512–517 (2009)
5. Bernard, O., Brühwiler, E.: Influence of autogenous shrinkage on early age behaviour of
structural elements consisting of concretes of different ages. Mat. Struct. /Materiaux et
Constructions 35(253), 550–556 (2002). https://doi.org/10.1007/BF02483123
6. Byard, B., Schindler, A., Barnes, R., Rao, A.: Cracking tendency of bridge deck concrete.
Transp. Res. Rec. 2164, 122–131 (2010)
7. Nassif, H.H., Suksawang, N., Mohammed, M.: Effect of curing methods on early-age and
drying shrinkage of high-performance concrete. J. Transp. Res. Board 1834, 48–58 (2003)
8. ACI 360R-10: Guide to Design of Slabs-on-Ground, American Concrete Institute,
Farmington Hills, MI (2010)
9. Roque, R., Kim, N., Kim, B., Lopp, G.: Durability of Fiber-Reinforced Concrete in Florida
Environments, Final Report for FDOT, Contract No.: BD545-41 (2009)
10. Folliard, K., Sutfin, D., Turner, R., Whitney, D.P.: Fiber in Continuously Reinforced
Concrete Pavements, Final Report Submitted to the Texas Department of Transportation,
Report No. 0-4392-2 (2006)
11. Hasan, M.J., Afroz, M., Mahmud, H.M.I.: An Experimental investigation on mechanical
behavior of macro synthetic fiber reinforced concrete. Int. J. Civ. Environ. Eng. 11(3), 19–23
(2011)
12. Richardson, A.E., Coventry, K., Landless, S.: Synthetic and steel fibers in concrete with
regard to equal toughness. Struct. Surv. 28(5), 355–369 (2010)
13. Roesler, J.R., Altoubat, S.A., Lange, D.A., Rieder, K.-A., Ulreich, G.R.: Effect of synthetic
fibers on structural behavior of concrete slabs on ground. ACI Mater. J. 103(1), 3–10 (2006)
Early Age Shrinkage Crack Distribution
in Concrete Plates Reinforced with Different
Steel Fibre Types
Sébastien Wolf1(&), Simon Cleven2, and Oldrich Vlasák1

1
ArcelorMittal Fibres, Bissen, Luxembourg
sebastien.wolf@arcelormittal.com
2
Institute of Building Materials Research at RWTH Aachen University,
Aachen, Germany
Abstract. In structural concrete elements reinforced with steel fibres, the

bearing capacity is an important criterion to be verified and designed at ultimate
limit state following the existing standards or codes already in use to design and
build structures in Steel Fibre Reinforced Concrete (SFRC), as for example the
Model Code 2010, the Swedish standard SS813310:2014, the German guideline
“Stahlfaserbeton” by DAfStb, the future Eurocode 2, etc.…
This approach was obtained by defining post cracking constitutive curves
validated by beam tests following EN14651 to reach residual flexural tensile
strength in a given cross section, and it is basically defined today in these
numerous reference documents and validated by experience with only a few
points, like orientation of fibres, being topic of further research.
The Serviceability Limit State is however less well defined for steel fibres
only solutions, meaning without any additional rebars, and some existing
equation and theories can still be today under discussion.
In that context, a parametric study has been undertaken to analyse the crack
distribution and the crack opening in SFRC plates made of steel fibres only in order
to put in evidence the influence of the fibre type and the dosage on this crack
formation under shrinkage conditions at early age, meaning during the first 7 h.
The purpose of this paper is to address these results and to highlight the level
of performance of a given steel fibre type compared to another, and the influence
on the crack distribution and the crack opening that a given steel fibre type can
play. The paper gives evidences of this statement to be able to support design
methods at Serviceability Limit State.
Keywords: Steel Fibre Reinforced Concrete Shrinkage Cracks
1 Introduction
The following testing process consists in exposing rectangular concrete plates made of
steel fibre reinforced concrete using different steel fibre shapes to shrinkage by ambient
air exposure. The concrete mix and the concrete compressive strength remain the same
for all plates so that the only varying parameters in this test process are the steel fibre
shapes and dimensions.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_3
Early Age Shrinkage Crack Distribution 25
2 Experimental Program
2.1 Test Samples
Each test set consisted of two samples. The first one was a slab with dimensions
160 60 8 cm3, which was anchored on the perimeter of the formwork (see
Fig. 1–left). The dimensions of the slabs have been chosen in order to facilitate the
crack initiation counting on the size ratio between x and y axis close to 3. The thickness
of 8 cm is thick enough to allow a good fibre distribution and thin enough to generate
quickly the cracking pattern. The anchoring prevented the plate from shrinking freely.
The second sample, a smaller plate with dimensions 30 30 8 cm3, was used only
for the measurement of water evaporation. After pouring of the steel fibre reinforced
concrete in these two castings, a transparent plastic panel was put on the steel frame to
cover the both plates (see Fig. 1–right): objective is to create a rectangular wind tunnel
with a nearly laminar wind flow to activate the drying process and to generate the
conditions of the restrained shrinkage.
Fig. 1. Testing formwork (left) and sample in a wind tunnel during testing (right).
26 S. Wolf et al.
2.2 Concrete Mixture

The composition of concrete mixture is shown in Table 1. Volume of cement, water
and fly ash is constant. Volume of gravel is slightly changing with regards to different
steel fibres dosages. The concrete mixture is based on the “‘Prüfplan für die Zulas-
sungsprüfung von Polymerfasern zur Verwendung in Beton nach DIN EN 206-1 in
Verbindung mit DIN 1045-2 mit nachgewiesener Wirksamkeit” by Deutsches Institut
für Bautechnik (DIBt) [7]. Air content was measured on the plane concrete with 2.0%
and for the compostion of the fibre reinforced concretes assumed to be nearly constant
to ensure a comparable concrete mix. The concrete is composed to show a huge drying
shrinkage and therefor also a crack pattern with large crack width in the reference
concrete and to allow fibers a good possibility to stabilize the concrete and reduce crack
width. In best way cracks should be reduced to a number of zero, but the main aim is to
achieve crack width and a total crack area that are suitable for Serviceability Limit
State. The maximum grain size of 8 mm is used to ensure that fibres are long enough to
bridge cracks and are anchored in sufficient way.
Table 1. Concrete mixture.

CONCRETE MIX composition
in function of the steel fibres dosage
Parameter Unit Steel fibre dosage (kg/m3)
0 20 30 40 60
3
CEM I 32,5R (DIN EN 197-1) kg/m 360 360 360 360 360
Water 270 270 270 270 270
Gravel 1.197 1.190 1.186 1.183 1.176
Powdered limestone 368 368 368 368 368
Air content (assumed) % 2.0 2.1 2.2 2.2 2.1
Water cement ratio 0,75 0,75 0,75 0,75 0,75
Maximum grain size of gravel mm 8 8 8 8 8
2.3 Steel Fibres

Three different steel fibre shapes have been used: The Hooked End (named HE), the
undulated (named TABIX) and the Hooked Flat End (named HFE).
The HE steel fibres were tested in four different wire diameters (0.55 mm,
0.75 mm, 0.90 mm and 1.00 mm) and three different lengths (35 mm, 50 mm and
60 mm). The HFE were tested with a 0.90 mm wire diameter and a length of 60 mm.
The TABIX Steel Fibres were used with wire diameters of 0.90 mm and 1.00 mm with
a length of respectively 35 mm and 50 mm.
The wire quality was also tested, meaning the wire tensile strength that the steel fibres
are made of. For that purpose, three different wires with varying tensile strength were
used: 1150 MPa (HE Fibres), 1500 MPa (HE+ Fibres) and 1900 MPa (HE++ Fibres).
To be able to have a significant range of results, 4 different fibre dosages have been
analysed in the tests: 20, 30, 40 and 60 kg/m3. The highest dosage has only been used
for steel fibres types that are effectively used in real projects. A summary of the test
program is exposed in Table 2:
Table 2. Test program with all the different steel fibres and dosages tested
Fibre type Dosage
20 kg/m3 30 kg/m3 40 kg/m3 60 kg/m3
Reference sample Ref
(without fibres)
HE 1/50 HE 1/50-20 HE 1/50-30 HE 1/50-40 HE 1/50-60
HE 1/60 HE 1/60-20 HE 1/60-30 HE 1/60-40 HE 1/60-60
HE+ 1/60 HE+ 1/60-20 HE+ 1/60-30 HE+ 1/60-40 HE+ 1/60-60
HE 90/60 HE 90/60-20 HE 90/60-30 HE 90/60-40 –
HE++ 90/60 HE++ 90/60-20 HE++ 90/60-30 HE++ 90/60-40 HE++ 90/60-60
HE 75/50 HE 75/50-20 HE 75/50-30 HE 75/50-40 –
HE++ 75/50 HE++ 75/50-20 HE++ 75/50-30 HE++ 75/50-40 –
HE 55/35 HE 55/35-20 HE 55/35-30 HE 55/35-40 –
HE 75/35 HE 75/35-20 HE 75/35-30 HE 75/35-40 –
HFE 90/60 HFE 90/60-20 HFE 90/60-30 HFE 90/60-40 –
TABIX 1/50 TABIX 1/50-20 TABIX 1/50-30 TABIX 1/50-40 –
TABIX 90/35 TABIX 90/35-20 TABIX 90/35-30 TABIX 90/35-40 –
To highlight the nomenclature, as an example HE++ 90/60 means a Hooked End

shape with 1900 MPa wire tensile strength, a wire diameter of 0.90 mm and a length of
60 mm.
So, in total 42 samples were tested: Twelve types of steel fibres were selected for
the experimental program and two reference samples were made of plain concrete.
Table 3 shows the calculated number of fibres per m3 of concrete. The calculation is
based on the assumption of a circular cross section of the fibres and a straight form with
a density of 7850 kg/m3.
2.4 Sample Preparation and Storage Conditions

The air temperature of the testing room and of the storage room, the testing machine
and the constitutive material temperature was ranging between 18 °C and 22 °C. The
mixing of the concrete mix components, meaning the water, the cement, the aggregates
and the powdered limestone, which was used to achieve a high-water retention capacity
28 S. Wolf et al.
Table 3. Calculated number of fibres per m3 concrete of the different mixes

Fibre type Number of fibres in thousand per m3
concrete
20 kg/m3 30 kg/m3 40 kg/m3 60 kg/m3
Reference sample 0
(without fibres)
HE 1/50 65 97 130 195
HE 1/60 54 81 108 162
HE+ 1/60 54 81 108 162
HE 90/60 67 100 133 –
HE++ 90/60 67 100 133 200
HE 75/50 115 173 231 –
HE++ 75/50 115 173 231 –
HE 55/35 306 460 613 –
HE 75/35 165 247 330 –
HFE 90/60 67 100 133 –
TABIX 1/50 65 97 130 –
TABIX 90/35 114 172 229 –
to allow the use of a water cement ratio of 0.75 without extreme bleeding effects, was
done in a compulsory mixer the capacity of which is about 150 L. First the cement, the
limestone and aggregates were homogeneously mixed and only then the water was
added while the mixing process was ongoing in a time period of 30 s before stopping
shortly. Then a further mixing time of 5 min was started so that the steel fibres could be
added and then mixed finally with the whole amount of fibres for 1 more minute.
The fabrication of the concrete cubes of 15 15 15 cm3 foreseen to determi-
nate the concrete compressive strength was done following the DIN EN 12390-2:2009-
02. The specimens were stored till demoulding at 24 h in a full humidity room at a
temperature of 20 °C and a relative humidity grade of 95%. Till the 7th day, they were
stored in the water at a temperature of 20 °C and afterwards in an ambient air of 20 °C
and 65% relative humidity. The compressive strength was then obtained following the
DIN EN 12390-3:2009-07 at an age of 28 days. The average compressive strength
measured was 33.7 MPa, ranging from 31.8 to 35.3 MPa, meaning a concrete grade of
C25/30 with a good reproducibility. The reference samples without fibres were
reaching the same value.
Two main fresh concrete characteristics were measured: the slump and the air
content.
• The slump was done immediately after the end of the mixing time following
DIN EN 12350-5:2009-08.
• The density of fresh concrete and air content were also analysed at the same
moment following DIN EN 12350-7:2009-08 and DIN EN 12350-6:2011-03. The
average air content measured in steel fibre reinforced concrete was about 2.1%,
ranging from 1.5% to 2.5%, whereas in the reference sample without fibres it was
2.0%.
2.5 Testing Conditions

The slab was exposed to unfavourable conditions in a wind tunnel immediately after
concreting for 7 h, and afterwards the wind tunnel was removed to analyse the crack
behavior. Wind speed in the tunnel was 5 m/s, temperature 20 ° C and relative
humidity about 60%, as shown in Fig. 2. Within this test period the water evaporation
was measured continuously on the small concrete plate by recording the loss of mass
with an integrated load cell. Also wind speed, temperature and humidity were recorded
with a combined anemometer and climate measuring system.
Wind speed in m/s rel. humidity in % Water evaporation in % by mass

Temperature in °C 2,0
25 100 1,8
90
1,6
20 80
1,4
70
1,2
15 60
1,0
50
10 40 0,8
30 0,6
5 20 0,4
Wind speed Temperature rel. humidity 10 0,2
0 0 0,0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Concrete age in hours Concrete age in hours
Fig. 2. Temperature, relative humidity, air speed (left). Water evaporation (right).
Fig. 3. Testing sample with highlighted cracks (left). Map of cracks (right).
30 S. Wolf et al.
2.6 Samples Processing

After removing the wind tunnel, the position of the shrinkage induced cracks was
highlighted and accurately recorded (see Fig. 3). Next, the length and width of each
crack section were measured. The main monitored parameters were:
• number of cracks
• length of cracks (total, max, average)
• width of cracks (average, max)
• total crack area.
3 Test Results and Test Observations
The experiments resulted in a large data set. The following results have been selected
for the purposes of this article: Total number of cracks, average crack width, total
cracks area and total crack length. The water evaporation for example can give hints on
the water retention ability of the concrete but is not presented in this paper.
3.1 Total Number of Cracks

The first parameter selected is the total number of cracks in each sample (see Fig. 4.).
A small number of cracks can be observed in the plain concrete slab. Similar behaviour
can be seen also in slabs made from concrete reinforced by HE 1/50 fibres at dosages of
20 and 30 kg/m3. For the other samples selected, the number of cracks increases
significantly. The effect of the addition of the HE 55/35 fibre, which has the smallest
diameter and therefor the highest number of fibres at a specific dosage, is extremely
high since a dosage of 20 kg/m3 leads to 185 cracks, which is ten times as much as the
plain concrete sample. In general, there is a correlation between the amount of fibres
and the number of cracks visible due to the property of fibres to bridge a crack and
redistribute stress, resulting in new cracks with smaller crack widths, which means a
higher effectivity of short fibres as long as their length leads to good anchoring
mechanisms under consideration of the maximum grain size.
3.2 Average Crack Width

The next parameter that should be focused on is the average crack width (see Fig. 5.). It
can be seen that in the plain concrete slab the average crack width is greater than
0,8 mm. The influence of HE 1/50 fibres is again significant only from the 40 kg/m3
dosage, while the other monitored fibres show an average crack width below 0,3 mm in
all tested dosages, beginning with 20 kg/m3.
Fig. 4. Number of cracks in function of the fibres dosage: 0–20–30–40 kg/m3.
Fig. 5. Average crack width in function of the fibres dosage: 0–20–30–40 kg/m3.
Here the crack bridging effect and so the ability of stress redistribution can be very
well observed. A small number of cracks in Fig. 4 corresponds with a high average
crack width in Fig. 5. Especially high amounts of fibres lead to small average crack
width, which is visible by the comparison of the results of HE 1/50 to HE 55/35, where
at the same fibre content the amount of fibres of HE 55/35 is nearly 4.5 times higher
than of HE 1/50. This means stress in the concrete always leads to local exceeding of
the concrete tensile strength and therefor the risk of cracking. Fibres bridging this
potential crack can overtake the stress and in a certain amount lead to very small crack
widths which are acceptable for the Serviceability Limit State of a real concrete ele-
ment. Small cracks below a crack width of 0.3 mm fulfil the requirements according to
DIN EN 1992-1-1:2011-01 that shall prevent water penetration and lead to a high
32 S. Wolf et al.
durability for reinforced concrete with standard reinforcement – not post or pre-
tensioned. In such cases small cracks are only an optical damage in case of exposed
concrete, but no structural problem.
3.3 Total Crack Area

The effect of the addition of steel fibres on the total crack area (equal to crack length
multiplied by crack width) is also evident, but less dramatic than the other monitored
parameters (see Fig. 6.). The total crack area in the plain concrete sample is about
25 cm2. Significant effect of the addition of fibres can be observed for fibres with
smaller diameter (HE 75/50 and HE 55/35) at a dosage of 40 kg/m3, when the crack
area decreased to approx. half of the reference concrete. An explanation of this has
already been given in the chapter above, meaning the bridging effect of fibres leading to
a higher amount of smaller cracks. Looking on the results of the total crack area it can
be reasoned that fibres don’t hinder the concretes tendency to build cracks when being
stressed but only distribute cracks to a higher amount and smaller crack width. It could
be assumed that, taking into account cracks with a not detectable crack width below
0.05 mm, the total crack area only is influenced by the concrete and therefor is a
material parameter of the plain concrete and shows no influence to the addition of
fibres. Only because of a limitation in detectable crack width differences in total crack
area are visible. This leads to the suggestion that for purpose of characterisation of fibre
performance in Serviceability Limit State a lower limit of crack width should be set to
compare results of total crack area. Such a useable limit could be defined taking into
account the field of application based on the tolerable crack width according to [8],
which for example could be 0.3 mm for constructions with standard reinforcement –
not post or pre-tensioned.
Fig. 6. Total crack area in function of the fibres dosage: 0–20–30–40 kg/m3.
3.4 Total Crack Length

The last parameter that should be considered is the total crack length (cp. Figure 7). It is
obvious that a certain amount of fibres leads to an increase in total crack length. While the
plain concrete only shows cracks in a length of about 3000 mm, all of the fibre reinforced
concretes showed a larger total crack length. Similar to the results of number of cracks and
crack width the number of single fibres in the concrete seems to have the biggest effect on
the crack length, what is recognisable through the comparison of the results of HE 1/50
and HE 55/35. In lower fibre dosages of 20 and 30 kg/m3 the concretes reinforced with HE
55/35 significantly possess the largest total crack length whereas concretes with HE 1/50
show only a small increase compared to the plain concrete. With an amount of 40 kg/m3
of fibres an upper limit seems to be found where no more visible cracks can be detected
and so the crack length will not increase any more.
All in all, it can be concluded that the total crack length as well as the total crack
area is not a reliable parameter to evaluate the performance of a fibre reinforced
concrete in the Serviceability Limit State. However, the total crack length seems still to
be a better parameter than the total crack area.
Fig. 7. Total crack length in function of the fibres dosage: 0–20–30–40 kg/m3.
4 Discussion and Conclusions
Based on the experimental program and detailed evaluation of the results it can be
concluded that the addition of steel fibers significantly improves the properties of
concrete with respect to the formation of shrinkage induced cracks. In particular, the
following were observed:
• Number of shrinkage induced cracks is increasing by increasing amount of fibres.
• Width and length of shrinkage induced cracks is decreasing by increasing amount of
fibres.
34 S. Wolf et al.
• Total crack area is no useable parameter to characterise the performance of steel

fibres in concrete due to the effect of fibres to redistribute stress leading to the
formation of new cracks.
• Positive influence of steel fibres is increasing with increasing number of pieces in
unit volume of concrete, so the main parameter of a fibre reinforced concrete in the
Serviceability Limit State is not fibre dosage or any characteristic of the fibre itself
but number of fibres.
• Fibre lengths of 35 mm are enough to ensure a good bond between fibre and
concrete matrix with maximum grain size of 8 mm. In this case short fibres are
more effective than long steel fibres (50 or 60 mm) to increase the number of cracks
because of the higher amount of fibres, meaning to redistribute the cracks in the
concrete.
• Average crack width is dependent on the total number of cracks. An increased
number of fibres leads to a higher amount of small cracks.
• Steel Fibres with High tensile strength wire with 1900 MPa compared to normal
tensile strength wire of 1150 MPa, from a dosage of 40 kg/m3, show an increased
performance considering each parameter.
• The undulated shape TABIX compared to the hooked end shape is effective to
multiply the cracks especially for low dosages. Comparing the TABIX 1/50 to the
HE 1/50 brings a real advantage with 20 and 30 kg/m3 in terms of crack length (for
example: a total of 10.6 m crack length for the TABIX at 20 kg/m3 compared to
4.2 m for the HE 1/50) and in terms of crack width (0.27 mm crack opening for the
TABIX compared to 0.52 mm for the HE for the same dosage of 20 kg/m3).
REFERENCES
1. EN14651:2005+A1: Test Method for metallic fibre concrete–Measuring the flexural tensile
strength (limit of proportionality LOP, residual)
2. fib Model Code for concrete structures (2010)
3. DIN EN 12390-2:2009-02/DIN09c: Testing hardened concrete - Part 2: Making and curing
specimens for strength tests, German Version EN 12390-2:2009
4. DIN EN 12390-3:2009-07/DIN09d: Testing hardened concrete - Part 3: Compressive
strength of test specimens; German version EN 12390-3:2019
5. DIN EN 12350-7:2009-08/DIN09b: Testing fresh concrete - Part 7: Air content - Pressure
methods; German version EN 12350-7:2019
6. DIN EN 12350-6:2011-03/DIN11: Testing fresh concrete - Part 6: Density; German version
EN 12350-6:2019
7. DIBt: Prüfplan für die Zulassungsprüfung von Polymerfasern zur Verwendung in Beton nach
DIN EN 206-1 in Verbindung mit DIN 1045-2 mit nachgewiesener Wirksamkeit. Deutsches
Institut für Bautechnik, Berlin, January 2013
8. DIN EN 1992-1-1:2011-01 Eurocode 2: Bemessung und Konstruktion von Stahlbeton- und
Spannbetontragwerken - Teil 1–1: Allgemeine Bemessungsregeln und Regeln für den
Hochbau; Deutsche Fassung EN 1992-1-1:2004+AC:2010
Technological Aspects
Influence of Synthetic Fibres on Seismic
Resistance of Reinforced Concrete Sections
E. Stefan Bernard(&)
TSE Pty Ltd., Sydney, Australia

s.bernard@tse.net.au
Abstract. To provide resilient load capacity during a seismic event, longitu-

dinal reinforcing bars within a reinforced concrete section need to be confined
against buckling in compression, and concrete within the section needs to be
restrained against loss from the section. The traditional approach to achieving
these goals has been to include numerous stirrups at close centers along all RC
members, but this approach is expensive and time-consuming to construct. An
alternative to the use of closely spaced stirrups is the inclusion of fibres within
the concrete. In this investigation, an experimental assessment has been
undertaken into beam-columns made with steel reinforcing bars and a moderate
dosage rate of various fibre types including macro- and micro-synthetic fibres,
steel fibres, and amorphous metallic fibres. The beam-columns were subjected to
reverse-cycle simulated seismic loading in accordance with ACI 374 [1], and
results were compared in terms of residual load carrying capacity at various
levels of deformation. The data indicate that stirrup spacing can safely be
increased without a reduction in seismic resistance when certain types of fibre
are added, but that fibres used at moderate dosage rates cannot entirely replace
stirrups.
Keywords: Fibres Concrete Seismic Stirrups
1 Introduction
Reinforced Concrete (RC) members such as beams and columns often include stirrups
to maintain post-crack shear capacity, and in earthquake-prone areas may include
additional stirrups that act as confining steel to maintain structural integrity during a
seismic event (e.g. NZS3101 [2], ACI 318 [3]). To achieve seismic resistance, the
spacing between stirrups is usually closer than is ordinarily prescribed for non-
seismically resistant structures. The close spacing between stirrups can lead to con-
gestion of reinforcement within members, and can also be expensive to construct due to
the labour-intensive nature of steel fixing. A possible alternative to closely-spaced
stirrups is the use of fibres to provide restraint and confinement during a seismic event,
although this is not yet recognized in codes. Previous experimental research [4–7] has
demonstrated that steel fibres are relatively effective in this application, but suffer the
problem that their stiffness can make them difficult to consolidate within an already
tight reinforcement cage, especially if they are relatively long (>40 mm).
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_4
38 E. S. Bernard
An alternative to steel fibres is the use of macro-synthetic fibres. This type of fibre
usually offers the advantage of being flexible so that consolidation is easier between
reinforcing bars. Several recent small-scale investigations [8–11] have indicated that
macro-synthetic fibres can produce meaningful improvements in seismic resistance if
used in combination with conventional steel reinforcing bars.
Micro-synthetic and amorphous metallic fibres have also been included in this
investigation, primarily because they are much smaller in size than either steel or
macro-synthetic fibres and therefore offer the advantage of inducing a larger ‘network
effect’ which might be effective for holding the cracked concrete together during
reverse-cycle loading. A total of 11 mixtures have been produced with a variety of
metallic and synthetic fibres included to examine their effectiveness in reducing the
need for confining stirrups compared to RC beam-columns made with plain concrete
and subject to reverse-cycle loading (Table 1).
Fig. 1. a) Load imposition and corresponding moment distribution on a beam-column subjected

to twin point loads spaced 500 mm apart, b) steel bars and stirrups in cross-section.
The intended application for the current work is cast-in-place and pre-cast concrete
members. Many previous investigations have focused on joint behaviour in these
applications because joints tend to suffer the most rapid breakdown during seismic
loading, but the current investigation has focused on the potential for increasing stirrup
spacing in members between joints. A relatively simple beam-column configuration
comprising four longitudinal bars with square stirrups at regular centres has been used
for all specimens (Fig. 1). Plain concrete controls were at first produced with stirrups at
50, 100, and 250 mm centres. The 50 mm stirrup spacing configuration was designed
in partial accordance with the New Zealand concrete structures code NZS 3101. The
subsequent FRC specimens were all produced with stirrups at 250 mm centres.
Influence of Synthetic Fibres 39
2 Laboratory Investigation
The current investigation consisted of several laboratory-based trials in which a series

of RC members were cast using a nominal 35 MPa concrete mix consisting of either
plain concrete or FRC dosed with various fibres (Table 1). The cementitious mix
design was the same for all the specimen sets and included a 10 mm maximum
aggregate size (Table 2). The mean UCS was equal to 47 MPa at the time of testing (91
days). The beams-columns all measured 2000 mm in length and 200 200 mm in
cross-section with one 12 mm longitudinal reinforcing bar in each corner. With the
pivots at each end included, each beam-column measured 2524 mm total length
(Figs. 1 and 2). The 12 mm N-grade deformed longitudinal steel bars had a nominal
tensile yield strength of 500 MPa and exhibited normal levels of post-yield ductility
(AS4671). The cover to the bars was nominally 20 mm from each face. Four replicates
were produced and tested for each reinforcement configuration for a total of 52
specimens in the investigation. The beam-columns included stirrups consisting of
6 mm plain round bar bent into a fully enclosed square with 40 mm long end hooks.
Seven ASTM C1609/C1609M beams were also cast and tested for each FRC mixture.
Fig. 2. Test apparatus for simulated seismic loading of beam-columns in flexure.
The experimental apparatus developed for this investigation is shown in Fig. 2.

Load was imposed on each beam-column specimen using two servo-hydraulic actua-
tors and deformation was measured using an assortment of displacement transducers
(LVDTs). A constant axial compressive stress of 4 MPa was applied using a 250 kN
MTS servo-controlled actuator operating in load-control. The lateral load was imposed
using a 100 kN MTS servo-controlled actuator operating in stroke-control. All the
equipment was mounted on a 5500 2650 mm slotted steel work plate measuring
40 E. S. Bernard
150 mm thick on a 600 mm deep steel waffle-plate. The reinforced concrete beam-
columns were bolted into place between two large end pivots in the longitudinal
direction while axial tension rods spanned the distance between the pivots in order to
prevent slip between the mounting blocks and the slotted steel base.
The fibres used in the investigation included the following. The FM650S fibers
listed in Table 1 were approximately 58 mm long and 2 mm wide, and were made of
straight tape-based polypropylene of 613 MPa tensile strength. The BC54 fibers were
54 mm long, 2 mm wide, embossed, and made of work-hardened polypropylene of
640 MPa tensile strength. The MQ58 macro-synthetic fibers were made of embossed
polypropylene, 58 mm in length, nominally 100 lm in diameter, and 700 MPa tensile
strength. The RC65/60 3D Dramix cold-drawn hooked-end steel fibres were 60 mm in
length and about 1100 MPa tensile strength. The BC48 macro-synthetic fibres were
approximately 48 mm long and 2 mm wide, and made of embossed polypropylene of
640 MPa tensile strength. The RECS15 micro-synthetic fibres were 12 mm in length,
40 lm diameter, and made of PVA of 1600 MPa tensile strength. The PA fibres were
made of polyamide and were 225 lm diameter and 18 mm long. The Fibraflex
FF/20L6 fibres were amorphous metallic whiskers of 20 mm length and 24 lm
thickness and approximately 1400 MPa tensile strength. The mechanical properties of
each of the FRC mixes are presented in Table 1.
Table 1. Specimen sets, reinforcement details, and mean FRC post-crack performance
summary.
Set Reinforcement Dosage Spacing f600 f150 UCS
(kg/m3) (mm) (MPa) (MPa) (MPa)
1 Plain – 50 – – 55
2 Plain – 100 – – 55
3 Plain – 250 – – 55
4 BC54 macro-synthetic PP 6 250 3.02 2.95 42
5 MQ58 macro-synthetic PP 6 250 2.00 2.75 44
6 MQ58 macro-synthetic PP 9 250 2.95 3.79 46
7 FM650S macro-synthetic PP 6 250 2.38 2.61 40
8 BC48 macro-synthetic PP 10 250 4.33 3.92 35
9 PA synthetic polyamide 9 250 0.51 0.22 54
10 RECS15 micro-synthetic PVA 4 250 0.47 0.11 42
11 Fibraflex FF/20L6 AMF 20 250 1.41 0.28 44
12 Dramix RC 65/60BG steel fibre 20 250 2.96 3.52 50
13 Dramix RC 65/60BG steel fibre 40 250 5.56 6.48 47
2.1 Simulated Seismic Testing

For each of the beam-columns, a moment was induced by imposing a saw-tooth pattern
of steadily increasing lateral deformation at the lateral load points to assess the ability
of the specimen to resist progressive degradation under reverse cycle simulated seismic
deformations [1]. Some preliminary quasi-static bending tests established the reference
lateral deflection Q (at which yielding of bars commenced) to be equal to 8 mm. At the
start of each cyclic test, a maximum lateral deflection of 4 mm (0.5Q relative to the axis
of the beam-column) was therefore imposed at the load point(s) through two cycles in
accordance with ACI 374. The rate of lateral deflection was 20 mm/min and load was
imposed through displacement-control while recording the resistance to deformation.
A second pair of cycles, to 8 mm lateral deflection (Q), was thereafter imposed at a rate
of 40 mm/min. Subsequent pairs of cycles were imposed to a maximum deflection of
2Q, 3Q through to 13Q (104 mm maximum lateral deflection relative to the centerline).
This was equivalent to a peak drift of 2 104/1012 = 20.5%. The rate of displacement
was progressively increased as Q was increased so that the duration of each cycle
remained constant at 48 s.
Table 2. Mix design for 32 MPa concrete mix used in all trials.
Ingredient Quantity (kg/m3)
10 mm CRG 1000
Coarse sand 460
Fine sand 250
GP cement 280
Fly ash 100
LR Water reducer 1330 mL/m3
3 Results for Laterally Loaded Specimens

3.1 Plain Concrete Bar-Reinforced Beam-Columns
The result for each test consisted of a load-deformation plot of the type shown in
Fig. 3a. Comparable plots are shown for various FRC mixes in Figs. 3b and 4. These
graphs show the load resistance of the beam-column (expressed as moment) plotted as
a function of lateral deformation (expressed as drift, in percent) for specimens with a
stirrup spacing of 50, 100, and 250 mm. Note that the axial force was held constant at
160 kN throughout each test and the resulting P-D effect induced an additional moment
in the beam-column that was proportional to the lateral deformation at the center D.
The P-D moment was calculated as the axial force multiplied by the eccentricity of the
axial load from the plastic centroid of the deformed section. For simplicity, the plastic
centroid has been taken as the geometric center of the section despite the presence of
cracks at larger lateral deformations. This additional moment has been included in all
the test results.
42 E. S. Bernard
The moment-drift curves have been reduced to a set of peak drift and moment
values that occurred at the maximum drift imposed during each cycle of deformation.
The moments have also been normalized relative to the average peak moment sustained
in each test and plotted against the corresponding drift in each cycle for each set of
nominally identical specimens (Fig. 5). A non-linear curve fit to each set of data was
then performed and is shown with each set of test data. These curves show that for a
stirrup spacing of 50 mm, the rate of fall in moment resistance was moderate following
the peak. Minor variations in moment resistance occurred before the peak as a result of
differences in the perpendicularity of the sides of the specimens. After the peak, there
also existed minor differences in moment resistance out to about 6% drift, but beyond
this a major difference arose due to progressive break-up of the concrete in the beam-
columns, especially as spacing between stirrups increased.
Fig. 3. Moment resistance plotted as a function of drift for RC section with a) plain concrete and
50, 100 or 250 mm stirrup spacing, and b) 6 kg/m3 FM650S and 250 mm stirrup spacing.
Fig. 4. Moment resistance plotted as a function of drift for RC section with plain concrete and
100 or 250 mm stirrup spacing, compared to a) 9 kg/m3 MQ58, or b) 40 kg/m3 Dramix
RC65/60.
The plain RC specimens with a 100 mm stirrup spacing suffered fall-outs of large
pieces of concrete during the test that led to a loss of confinement and buckling of the
bars as drift increased. For the specimens with 250 mm stirrup spacing, the break-up of
the concrete was quite rapid beyond 7% drift. The principal reason why the members
with 50 mm stirrup spacing performed better was that the longitudinal bars were more
effectively restrained against lateral buckling, particularly in the range 8–14% drift.
All the specimen sets showed a general increase in performance as the stirrup
spacing was reduced. However, failure through longitudinal buckling nevertheless
occurred in even the most heavily reinforced members, as represented by the abrupt
termination in each moment-drift curve.
3.2 Beam-Columns with Macro-Synthetic Fibres

Just as for the plain concrete specimens, the specimens containing 6 kg/m3 of macro-
synthetic fibres exhibited a peak in moment capacity between 6 and 8 per cent drift
followed by a fall in resistance that appeared quite similar for each fibre type (Fig. 5b).
All tests ended with longitudinal compression failure of the beam-columns caused by
loss of surface concrete and buckling of bars at the central hinge. The fibres were well
distributed through the concrete and acted to hold spalled concrete in place around and
between the longitudinal bars, thereby helping to enhance the capacity of the stirrups.
Fig. 5. Normalized moment versus drift for a) plain concrete specimens with stirrup spacing of
50, 100, and 250 mm, and b) MSFRC with 6 kg/m3 of fibre and 250 mm stirrup spacing.
The addition of 6 kg/m3 of macro-synthetic fibre to each mix had a minimal effect
on performance out to around 10 per cent drift, but thereafter assisted in holding the
concrete fragments together as the concrete steadily broke up under reverse-cycle
loading. As spalling of surface concrete outside of the main longitudinal bars com-
menced, the presence of the fibres acted to constrain fall-outs and confine the rubble-
like concrete fragments inside the reinforcement cage. This was most obvious at large
deformations (Fig. 5b). The longer and more slender fibres (MQ58 and FM650S)
44 E. S. Bernard
appeared to be more effective in this role, and several of the beam-columns reinforced
with these fibres continued to support load up to about 20% drift. After this, they failed
in compression as the longitudinal bars buckled. It is notable that while the presence of
the fibres assisted the FRC specimens (which all had stirrups spaced at 250 mm) to
support greater load than the plain concrete RC members beyond about 11–12% drift,
the residual load resistance in the 16–18% drift range was only about 20% of the peak
load resistance.
Fig. 6. Normalized moment versus drift for all FRC specimens with a) 9–10 kg/m3 macro-
synthetic FRC, or b) 4 kg/m3 micro-synthetic PVA FRC, compared to mean plain concrete
performance curve-fits for stirrups spacing of 50 to 250 mm.
The relatively small improvement in performance exhibited by the beam-columns

with 6 kg/m3 of macro-synthetic fibres compared to the plain concrete RC members
suggested that a higher dosage rate was required. Three additional sets of specimens
with 9–10 kg/m3 of synthetic fibres were therefore trialled (Table 1). These sets
exhibited an improvement in residual load resistance beyond 16 per cent drift compared
to the specimens with 6 kg/m3 of fibre, but the reduction in workability of the concrete
associated with the increased fibre dosage meant that a considerable effort was required
to achieve this improvement.
The majority of macro-synthetic fibres that emerged from the concrete as it broke
up appeared to be in pristine condition with none of the usual signs of work damage
associated with pull-out. This was because the concrete appeared to break up around
the fibres, with cracks occurring in every direction, such that the majority of fibres were
subject to minimal tensile stress during the reverse-cycle loading process. This sug-
gested that a fine denier fibre was required to hold the concrete together as it broke up,
because this would increase the likelihood that fibres would intersect every crack as it
widened. A mixture containing 4 kg/m3 of RECS15 fibres and 250 mm stirrup spacing
was therefore produced and tested (use of a higher dosage rate resulted in an
unworkable concrete). However, the results for this mixture were quite disappointing
given that all the specimens collapsed before 16 per cent drift had been attained
(Fig. 6b). The overall improvement in performance was marginal compared to a plain
concrete RC member with 250 mm stirrup spacing.
3.3 Beam-Columns with Metallic Fibres

The reverse-cycle loading tests of the beam-columns containing macro- and micro-
synthetic fibres did not produce results sufficient to warrant replacement of the stirrups
with fibres. However, they did suggest that the spacing between stirrups could be
increased if a minimum amount of suitable fibre was included in a mixture. Given that
the majority of previous research into the seismic performance of FRC has been
conducted using steel fibres, three mixtures were trialled using hooked-end steel fibres
and amorphous metallic fibres to examine how these perform in comparison to syn-
thetic FRC mixtures in reverse-cycle loading.
A mixture was produced with 20 kg/m3 of Fibraflex FF/20L6 amorphous metallic
fibres, and two mixtures incorporating 20 and 40 kg/m3 of Dramix RC65/60 BG steel
fibres were also produced and tested. All specimens incorporated steel stirrups at
250 mm centres. They were all subject to the same curing and testing regime as the
macro-synthetic FRC beam-columns, and the results are presented in Fig. 7a. The
immediate conclusion from this is that 20 kg/m3 of Dramix RC65/60 fibres resulted in
a minimal improvement in performance, and was equivalent to 4 kg/m3 of RECS15
fibre in its effect on resilience during reverse-cycle loading. In contrast, 20 kg/m3 of
FF/20L6 and 40 kg/m3 of Dramix RC65/60 BG fibre both resulted in a level of resi-
lience in response to reverse-cycle loading that was comparable to that achieved with
9–10 kg/m3 of macro-synthetic fibre. The relatively good performance of the FF/20L6
fibre, which measured only 20 mm in length but had a very high fibre count, suggests
that a small highly dispersed fibre exhibiting high stiffness and a large fibre count was
more effective in holding shattered concrete together during reverse-cycle loading than
larger and more sparsely distributed fibres, even if the latter exhibit far superior flexural
performance in standard beam tests (Table 1).
Fig. 7. Normalized moment versus drift for all FRC specimens with a) 20–40 kg/m3 metallic
FRC, compared to mean plain concrete performance curve-fits for stirrups spacing of 50 to
250 mm, and b) comparison of curve-fitted trends for the four best performing FRC sets relative
to plain concrete sets.
46 E. S. Bernard
4 Discussion
A comparison of mean normalized moment resistance as a function of drift for the four
best performing FRC mixes is shown in Fig. 7b together with the plain concrete sets.
The data indicate that any of these four FRC mixes, which included stirrups at 250 mm
centres, is capable of matching the performance of plain concrete beam-columns
incorporating stirrups at 100 mm centres. It therefore appears that both synthetic and
metallic fibres (at moderate dosages) can be used to increase the spacing between
stirrups, but not by more than 100%. None of the FRC mixtures achieved the per-
formance of the plain concrete with stirrups at 50 mm centres in the 8–16% drift range
(Fig. 7a), primarily because the fibres did not match the stirrups in their capacity to
restrain the longitudinal bars from deforming laterally when compressed. However, at
very large drifts, the fibres exceeded the capacity of the stirrups, apparently because the
shattered concrete was held together more competently by the fibres than by the stir-
rups. It was notable that few large pieces of broken concrete fell out of any of the
macro-synthetic FRC specimens, especially for the MQ58 and FM650S fibre rein-
forced mixes, but numerous such pieces fell out of the plain concrete members during
testing. The hooked-end steel fibres were less effective in this regard. A comparison of
normalised moment resistance at 16% drift with residual flexural strength measured in
the ASTM C1609/C1609M beams indicated that performance in reverse-cycle testing
was largely unrelated to the toughness of the FRC (Fig. 8).
The results of this investigation are difficult to compare to previous investigations
of the capacity of macro-synthetic fibres to replace or supplement conventional con-
fining stirrups for seismic resistance because most of the previous investigations [8–10]
included an insufficient number of specimens to estimate the systematic contribution of
macro-synthetic fibres. The only conclusion that could be reached in the earlier studies
was that fibres were effective in increasing seismic resistance provided that conven-
tional stirrups were retained.
More numerous studies have been performed on the contribution of steel fibres to
the seismic resistance of Reinforced Concrete members, primarily beam-column joints.
Many of the early studies simply demonstrated that steel fibres could increase seismic
resistance through their capacity to enhance confinement [12], but later studies
addressed the issue of potentially reducing the amount of confining steel (by increasing
the spacing between stirrups) through the inclusion of steel fibres [5, 6, 13, 14]. In
general, these studies showed that stirrup spacing could be increased while maintaining
seismic resistance, but very high levels of steel fibre reinforcement (typically 1–2% by
volume) were required to achieve this. Moreover, almost all the studies focused on joint
behavior in order to reduce bar congestion, not the members between joints.
Fig. 8. Relation between residual flexural strength of FRC and normalized moment
resistance/peak measured at 16% drift.
Almost all previous investigations into the seismic resistance of RC/FRC hybrid
members have included dosage rates of steel or macro-synthetic fibres that are
uneconomic and require serious re-design of concrete mixes to compensate for the
reductions in workability associated with very high volume fractions of fibre (usually
well over 1.0% v/v). The additional cost associated with the need for mixes that can be
consolidated with 1.0+% v/v of fibre, plus the cost of the fibres themselves, is likely to
be more than the savings possible through reductions in confining steel. For RC/FRC
hybrid designs to be economically viable, the dosage rate of fibre required to double the
spacing between stirrups must remain below about 1.0% (v/v) to be viable.
5 Conclusions
A series of Reinforced Concrete (RC) beam-columns were produced and tested to

examine the possibility of increasing the spacing between confinement stirrups in a
member subject to seismic loading by including synthetic fibres in the concrete mixture.
The performance of beam-columns produced with various synthetic fibres was com-
pared with beam-columns manufactured using plain concrete with stirrups at a spacing
ranging from 50 to 250 mm. These represented beam-column of low to high seismic
resistance designed in partial accordance with the New Zealand concrete structures code
NZS 3101. Simulated seismic loading was imposed in the form of a uniform bending
moment at the center of the beam-column in accordance with ACI 374.
It was found that macro-synthetic fibres dosed at 9–10 kg/m3 could supplement the
seismic resistance of a RC member such that the maximum spacing between stirrups
could be doubled. Moreover, it appeared that long slender fibres with a low denier were
more effective in this regard than stocky fibres. In addition, hooked-end steel fibres
48 E. S. Bernard
dosed at 40 kg/m3, and amorphous metallic fibres dosed at 20 kg/m3, could achieve a
similar effect. Thus, the reverse-cycle seismic flexural resistance of a member with a
stirrup spacing of 250 mm and 9–10 kg/m3 of macro-synthetic fibres is equivalent to a
plain concrete member with a stirrup spacing of 100 mm. The addition of macro-
synthetic fibres was most effective for the purpose of increasing residual flexural
capacity at very large drifts.
References
1. ACI 374 .2R13: Guide for Testing Reinforced Concrete Structural Elements under Slowly
Applied Simulated Seismic Loads, American Concrete Institute, Farmington Hills, MI
(2013)
2. NZS 3101 Code of Practice for the Design of Concrete Structures. Standards Association of
New Zealand, Wellington, p. 256 (2006)
3. ACI 318: Building Code Requirements for Structural Concrete and Commentary, American
Concrete Institute, Farmington Hills, MI (2011)
4. Jiuru, T., Chaobin, H., Kaijian, Y., Yongcheng, Y.: Seismic behavior and shear strength of
framed joints using steel fibre reinforced concrete. ASCE J. Struct. Eng. 118(2), 341–358
(1992)
5. Filiatrault, A., Pineau, S., Houde, J.: Seismic behavior of steel fibre-reinforced concrete
interior beam-column connection. ACI Struct. J. 92(5), 543–552 (1995)
6. Bayasi, Z., Gebman, M.: Reduction of lateral reinforcement in seismic beam-column
connection via application of steel fibres. ACI Struct. J. 99(6), 772–780 (2002)
7. Billington, S., Yoon, J.: Cyclic response of unbonded posttensioned precast columns with
ductile fibre-reinforced concrete. J. Bridge Eng. 9(4), 353–363 (2004)
8. Ma, H., et al.: The effect of macro-synthetic fibre on the seismic performance of reinforced
concrete columns carried out by experiment study. Adv. Mater. Res. 479–481, 1220–1224
(2012)
9. Carnovale, D.J.: Behaviour and Analysis of Steel and Macro-Synthetic Fibre Reinforced
Concrete Subjected to Reversed Cyclic Loading: A Pilot Investigation, Master of Applied
Science thesis, Department of Civil Engineering, University of Toronto (2012)
10. Osorio, L.I., Paultre, P., Eid, R., Proulx, J.: Seismic behavior of synthetic fibre-reinforced
circular columns. ACI Struct. J. 111(1), 189–200 (2014)
11. Bernard, E.S.: Enhancement of seismic resistance of reinforced concrete members using
embossed macro-synthetic fibers. In: World Tunnelling Congress 2016, San Francisco, April
25–28 (2016)
12. Jindal, R., Sharma, V.: Behavior of steel fibre reinforced concrete knee-type beam-column
connections. In: Shah, S.P., Batson, G.B. (eds.) Fibre Reinforced Concrete Properties and
Applications, SP-105, American Concrete Institute, Farmington Hills, MI, pp. 475–491
(1987)
13. Gefken, P., Ramey, M.: Increased hoop spacing in type 2 seismic joints using fibre
reinforced concrete. ACI Struct. J. 86(2), 168–172 (1989)
14. Tavallali, H., Lepage, A., Rautenberg, J.M., Pujol, S.: Concrete beams reinforced with high
strength steel subjected to displacement reversals. ACI Struct. J. 111(5), 1037–1048 (2014)
Development and Mechanical Characterization
of Dry Fiber-reinforced Concrete
for Prefabricated Prestressed Beams
Kamyar Bagherinejad Shahrbijari1, Suman Saha1(&),

Joaquim A. O. Barros1, Isabel B. Valente1, Salvador Dias1,
and João Leite2
1
ISISE, Department of Civil Engineering at the University of Minho,
Guimarães, Portugal
sumansaha.civil@gmail.com
2
Império Company, Braga, Portugal
Abstract. In this study, compression, bending and shear properties of the dry
fibre reinforced concrete (FRC) were experimentally characterized at its early
age. Considering the application of the developed FRC in prefabricated pre-
stressed beams, three-point notched beam bending tests were executed to
analyse the post-cracking flexural behaviour, and direct shear tests were per-
formed to evaluate its behaviour shear failure. Furthermore, the consistency of
the proposed material on the design has been investigated. Eurocode Standard
Specifications for concrete buildings is considered for the purpose of loading
and analysing of a prefabricated prestressed FRC beam, which was designed by
a cross section layered model for deriving its moment-curvature response. Using
the properties of developed dry FRC, the section of the prestressed beam has
been suggested.
Keywords: Prefabricated prestressed beam Dry fibre reinforced concrete

Flexural behaviour Shear behaviour Early age mechanical properties
1 Introduction
Prestressed concrete slab system is widely used for residential and office buildings,
hotels, hospitals and parking garages in Europe. The system has proven its structural
efficiency and economy. The use of prestressing allows for material savings due to
reduced total slab thickness and self-weight, and also reduces the deflection of the slab
under service loads. The present research is intended to develop a prestressed-
prefabricated beam capable of assuring slabs of 12 m span length. This requires
developments not only at the level of the prestressed-prefabricated beam, but also the
type of lightweight block, the geometry of the slab and the properties of the concrete
cast in place. For attaining this span length, the shear resistance of both the prestressed-
prefabricated beam and of the slab is a critical issue, since no stirrups are intended to be
used for economic reasons and lack of efficiency of these reinforcements in shallow
concrete elements [1]. For assuring the required shear resistance, short fibres are
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_5
50 K. B. Shahrbijari et al.
adopted in this project, since available research on the use of steel fibre reinforced
concrete (SFRC) demonstrates the high potential of this reinforcement for eliminating
totally (mainly in shallow elements, such is the actual case) steel stirrups without the
occurrence of shear failure [2].
FRC is becoming very popular in many construction applications, such as slabs on
grade, industrial floors, tunnel linings, precast and prestressed concrete products [3].
Incorporation of fibres in concrete mixtures leads, fundamentally, to the improvement
of the post-cracking tensile capacity of the hardened concrete offering resistance to
crack propagation [4–6], which has benefits in terms of durability [5], impact [7] and
energy dissipation capacity [4].
Metallic, synthetic, mineral and vegetable fibres are the most commonly used for
the reinforcement of cement-based materials. Since the major part of the available
research on the use of fibres for the shear reinforcement evidences that steel fibres are
still more efficient, mainly with hooked-ends [8], in the present work this type of fibres
are also used. Properties of steel fibre reinforced concrete are usually affected by the
behaviour of the matrix concrete, steel fibre geometry, aspect ratio, volume fraction,
distribution and orientation, the bonding strength at the fibre-matrix interface, etc. [9].
In the actual technology, the prestressed-prefabricated beams are made by very dry
concrete for assuring the desired geometry of these elements without any type of
moulds, while cement content is relatively low, which constitute serious obstacles on
assuring adequate fibre distribution and orientation.
This work aims to produce the most economic dry SFRC for pre-fabricated pre-
stressed beams, by minimizing the cement content and assuring the required
mechanical properties at early age and long term. The SFRC was also developed for
designing a new generation of prestressed-prefabricated beams capable of constituting
the main structural element of lightweight prefabricated slabs of 12 m span without
using any type of conventional shear reinforcement. In parallel, analytical and
numerical research is being carried out for optimizing the prestressed-prefabricated
beam and slab in order to accomplish their design loading cases/combinations for
serviceability and ultimate limit state conditions. Models for the evaluation of the
moment-curvature of FRC beams and slabs reinforced with passive and/or prestressed
bars are available [10, 11], but not of them integrates materials of different concrete
strength class in a different age of maturation (FRC of the beams and concrete cast in
place). The numerical strategy that is being developed aims to simulate these
particularities.
2 Experimental Programme
2.1 Materials
Cement type I 52.5R, fine sand (0/2), coarse sand (0/4), coarse aggregate (4/8) and
superplasticizer were used in the experimental program for the development of the dry
steel fibre reinforced concrete for the prestressed-prefabricated beams. All the
Development and Mechanical Characterization of Dry Fiber-reinforced 51
aggregates (fine and coarse both) were used in completely dry conditions, having been
kept at 100°C–110°C for 24 h, followed by a period of cooling down of 24 h before its
use in the production of concrete mix. Hooked end steel fibres having a length of
30 mm and an aspect ratio of 80 were used in this study. According to the supplier, the
tensile strength of steel fibres is higher than 1300 MPa.
2.2 Compressive Strength

In order to determine the compressive strength at early age, cubic specimens of
100 mm edge were cast with the produced SFRC mixes and were tested at 3 days
according to the recommendations of EN 12390-3: 2001.
2.3 Flexural Behaviour

To determine the post-cracking behaviour of the developed SFRC, three-point notched
beam bending tests (3PNBBT) were conducted according to the guidelines proposed by
RILEM-TC162-TDF (2002b) and EN-14651 (2007). From this test is directly obtained
the load - crack mouth opening displacement (CMOD) relationships, from which the
limit of proportionality, fL and the flexural residual strength parameters for CMOD =
0.5, 1.5, 2.5 and 3.5 mm are obtained (indicated by fR1, fR2, fR3,and fR4, respectively).
2.4 Shear Behaviour

To assess the shear behaviour of the developed SFRC, the FIP test proposed in [12]
was adopted, whose setup is represented in Fig. 1, since it is possible to obtain the
applied force versus crack opening and crack sliding. This information can be later
used in order to obtain the local crack shear stress versus crack sliding by inverse
analysis [13], in order to characterize the fracture mode II in advanced numerical
modelling based on the nonlinear fracture mechanics [14]. The specimens adopted for
these shear tests (150 mm 150 mm 290 mm dimensions) are obtained by
extracting the intact SFRC concrete of the lateral parts of the 3PNBBT after have been
tested. Each shear specimen has a notch at its mid-span section in the front and rear
faces, of 5 mm thickness and 25 mm depth along the full cross-section. Two LVDTs
(LVDT3 and LVDT4) were installed for measuring the crack opening in the rear face at
the position of 50 mm and 100 mm from the top of the beam specimen. One LVDT
(LVDT2) was installed at the bottom in the front face of the specimen for recording the
crack opening, while LVDT (LVDT1) was used for measuring the slip in this face. The
vertical LVDT1 has a measuring stroke of 10 mm, while the remaining ones have a
stroke of 5 mm, all of them with an accuracy of 1 µm. The applied force was measured
with a load cell of 300 kN capacity. The tests were controlled by an LVDT attached to
the piston of the actuator, at a displacement rate 2 µm/s.
Fig. 1. (a) Schematic diagram of Shear test (dimensions in mm) (b) Front side of the Specimen
during test (c) Rear side of the Specimen during test

To obtain the best cement-concrete matrix, which leads to achieving higher com-
pressive strength at the early stages, different trial mixes were considered with the
variation of water-cement (W/C) ratio followed by the cement content. Exceeding
30 MPa of compressive strength in cubic specimens of 100 mm edge at 3 days is the
target, due to requirements for removing, transporting and stocking the prestressed-
prefabricated beams, as imposed by the company associated to this applied research
project. The first set of compositions executed had the aim of obtaining the
water/cement (W/C) ratio capable of assuring adequate fibre distribution into the
concrete. This process was also guided by the compressive strength obtained on
compressive tests with SFRC cubic specimens from each composition, having been
concluded that a W/C = 0.5 attained the objectives. This is a relatively high W/C,
which is justified by the total dry state of the aggregates, but the final SFRC has the
aimed dry nature required by the pre-fabrication technology of the prestressed beams.
In the second phase of the SFRC development, the W/C = 0.5 was preserved, but the
cement content was increased up to exceed 30 MPa of compressive strength at 3 days.
For 300 kg/m3 of cement, 40 MPa at 3 days was obtained. By critical analysis of
existing database on SFRC beams tested in shear loading configuration [12], where
steel fibres were used to avoid shear failure in beams without steel stirrups, and taking
into consideration economic requirements and restrictions imposed the relatively low
dimensions of the cross-section of the prestressed-prefabricated beam, 60 kg/m3 of the
smallest hooked-ends steel fibres found in the Portuguese market were selected. By
executing compositions with 0, 30, 45 and 60 kg/m3 of these fibres, it was verified a
small decrease of the compressive strength with the increase of the fibre content,
indicating that the composition has sufficient paste for the aimed content of fibres. By
adopting the final composition indicated in Table 1, compressive strength of 40 MPa
was obtained at 3 days.
Table 1. Materials considered for one cubic metre concrete

Cement W/C Fine sand Coarse Coarse Admixture Fiber
(kg) (kg) Sand (kg) aggregate (kg) (l) (kg)
300 0.5 194.29 874.29 971.43 1.6 60

Figure 2a represents the relationship between load and CMOD for the SFRC reinforced
with 30, 45 and 60 kg/m3 of fibres at 3 days. A pronounced deflection hardening stage
is only obtained for the SFRC with 60 kg/m3, and Fig. 2b shows the small dispersion
on the flexural response of the specimens of this SFRC, caused by the relatively high
number of fibres per m3 of concrete in consequence of the high fibre aspect ratio (80).
The residual flexural strength (fRi) obtained in these tests are presented in Table 2.
Fig. 2. (a) Influence of the fibre content on the load-CMOD response and (b) Flexural stress vs
CMOD response for the specimens reinforced with 60 kg/m3 of fibres
Table 2. Values of residual flexural tensile strength at different CMOD levels

MIX Specimen no Residual Flexural Tensile Strength
(MPa) at CMOD (mm)
0.5 1.5 2.5 3.5 4.5
SF30 2.32 2.64 2.59 2.38 2.15
SF45 4.77 5.71 5.41 4.83 4.15
SF60 I 5.33 7.09 6.93 6.15 4.69
II 6.16 6.55 5.90 4.76 4.08
III 6.46 6.88 6.23 5.41 4.71
Average (SF60) 5.98 6.84 6.35 5.44 4.49
Std. Dev. (SF60) 0.59 0.27 0.53 0.70 0.36
3.3 Shear Behaviour

Figure 3a and 3b and Fig. 3c and 3d represent the relationship between the load, the
average crack opening and the slip for the specimens tested at 7 days and 28 days,
respectively, while the corresponding values are presented in Table 3. It is verified that
the maximum shear capacity at 7 and 28 days are similar, indicating that this FRC
attains its full maturation at a relatively early age. By considering the average shear
capacity of all tested specimens, average shear stress of about 13.49 MPa was obtained.
Fig. 3. Load-slip response and load-average crack opening response for the specimens during
the shear test at: (a-b) 7 days, and (c-d) at 28 days.
Table 3. Maximum load observed during the shear test

Specimen no. Maximum load (kN) Specimen no. Maximum load (kN)
Age of testing = 7 Days Age of testing = 28 Days
1 186.58 4 215.55
2 221.50 5 188.93
3 214.54 6 187.17
Average 207.54 Average 197.22
Std. Dev. 34.19 Std. Dev. 23.36
4 Design of FRC Prestrestressed Beam
With the consideration of the properties of the developed SFRC (reinforced with
60 kg/m3 of fibres), a moment-curvature diagram was obtained for designing a pre-
stressed beam for residential building applications. Consider a simply supported pre-
stressed FRC beam from a beam and block slab floor system as shown in Fig. 4.
(Figure 5). In the present stage of the research activities, a plain concrete (C20/25) is
considered for the concrete cast in place, as well as for the block, since this component
is still in development (a block of structural performance is being developed). Taking
the maximum possible span length for the block, the slab’s module has a width of
800 mm.
For loading conditions, according to Eurocode 1 [15], in residential buildings, the
permanent loads and live loads on floors are 1.9 kN/m2 and 2 kN/m2, respectively. This
permanent load does not include the self-weight of the slab’s module. The self-weight
of the slab’s module will be calculated based on the section size and the density of
materials. The minimum required dimensions of the reverse T prestressed beam to
carry the design loads are achieved by determining the moment-curvature diagram of
the slab’s section.
Fig. 4. Simply supported composite prestressed SFRC beam.
Fig. 5. Part of a composite slab section.
The prestressing force in a prestressed concrete member continuously decreases

with time. Prestress losses can be classified as two types: instantaneous loss or elastic
shortening loss and long-term prestress losses including creep and shrinkage of con-
crete and steel relaxation [16, 17]. The total losses of prestressing for the SFRC
prestressed beam is considered equal to 20% [18]. For simplification of the design
process, the position of prestressing steel during the design is considered to be at the
intersection of the web and the flange of the beam (the optimum centroid of this
reinforcement is still being investigated numerically).
4.1 Analysis of the Beam

Eurocode Standard Specifications [19] for concrete buildings are considered for the
purpose of loading and analysing of an FRC prefabricated prestressed beam. The
analysis of the beam is performed by accomplishing the service limit state (SLS) in
terms of deflection, and the ultimate limit state (ULS) in terms of flexural and shear
capacity. Eurocode gives partial safety factors for the loads and materials [15]. The slab
system was subjected to a load combination of (1.35 Total Permanent loads + 1.5
Live load). Also, for the design situation, the partial safety factors (cM) for concrete
(cc= 1.50) and prestressing steel (cs= 1.15) are considered.
4.2 Flexural Design

The selection of a prestressed concrete section is a common problem in a design
process. This is because the depth of the section will be not the only variable to be
changed, but also the web thickness, flanges width, and thickness. The type of cross-
sections used for the precast prestressed concrete SFRC beam is an inverted T section,
as shown in Fig. 5. The design is performed to satisfy the ULS and SLS design criteria
at the stage corresponding to the production of the precast prestressed concrete SFRC
beam (moment of releasing the prestress while the beam is in the prefabrication line, its
lifting, transporting and stockage), and at the slab’s composite section level.
To evaluate the flexural capacity of the SFRC beam at its production stage and of
the slab’s module, the moment-curvature response of the cross-section is obtained by
DOCROS software [20]. DOCROS can analyse sections of irregular shape and size,
composed of different types of materials, and the section can be subjected to an axial
force and variable curvature. DOCROS has a wide database of constitutive laws for the
simulation of monotonic and cyclic behaviour of cement-based materials, polymer-
based materials, reinforced with steel and/or fibre-reinforced polymer bars [21].
The cross sectional area of the steel reinforcement in the SFRC beam section is
taken equal to the balanced steel reinforcement ratio [22], which assures simultane-
ously the yield stress of this reinforcement and the concrete crushing strain (0.0035).
The constitutive laws adopted for the concrete and the prestressed steel wires are
shown in Fig. 6a and 6b, respectively, whose defining values adopted in the simula-
tions are included in Table 4.
Fig. 6. Stress-strain diagrams adopted in DOCROS for the simulation of a) concrete/SFRC, and
b) prestressed steel wires.
Table 4. Properties of the parameters defining the constitutive laws of the concrete/SFRC and
steel wires.
SFRC Compression (28 Days)
fcc(MPa) ecc eccr Ec(MPa)
40 0.0026 0.0032 39000
Tension
fct1(MPa) fct2(MPa) fct3(MPa) fct4(MPa) ect1 ect2 ect3 ect4 ect5
5.98 6.84 6.35 5.44 0.001 0.0015 0.002 0.003 0.0035
Steel Es (MPa) esy esh esu fsy fsu fsh Esh
wires (MPa) (MPa) (MPa) (MPa)
205000 0.0068 0.008 0.01 1262 1388 1152 102500
For the slab loading condition according to Sect. 4.1 of this study, the maximum
bending moment for the factored loads (1.35 Permanent load + 1.35 Self Weight +
1.5 Live load) is 139.124 kN.m, which is shown as a dashed line in Fig. 8. In this
study, the flexural capacity of the slab’s section is obtained from the moment-curvature
diagram. For this purpose, a window application was developed to do an incremental
procedure to generate the prestressed-prefabricated SFRC beam cross-section geome-
try, and to analyse the generated section by DOCROS and compare the moment-
curvature results with the maximum bending moment on the beam. Figure 7 shows the
flowchart of the developed numerical strategy. In Fig. 8 the solid line shows the
DOCROS moment-curvature diagram of the designed composite slab with SFRC
prefabricated beam. According to the DOCROS moment-curvature diagram and
maximum moment, the ultimate moment capacity of the designed section is adequate.
Fig. 7. Flowchart of the developed computational strategy for the optimum design of SFRC
beam and slab’s module.
Figure 9 shows the dimensions of the designed composite slab section including
prefabricated prestressed SFRC beam and cast-in-place concrete. The cross-section
area of the prefabricated prestressed SFRC beam is 37500 mm2. The total self-weight
of the composite section for the SFRC beam and cast-in-place concrete is 0.8 kN/m and
1.6 kN/m respectively. The steel ratio (= Asp/Ac) in the designed SFRC beam section is
equal to 1% and the total cross-sectional area of prestressed steel is 375.15 mm2.
Fig. 8. Moment-curvature diagram of the designed section.
Fig. 9. Designed composite FRC section (dimensions in mm)
4.3 Shear Design

For determining the shear capacity of FRC beam, the fib Model Code 2010 [23]
recommendations were used. This approach based on the concept of residual flexural
strength for FRC. By this approach, the shear resistance is obtained from [24]:
VRd ¼ VRd;F þ VRd;s ð1Þ
where VRd,F and VRd,s are the components of shear carried by FRC and steel stirrups,
respectively. The fibres component is given by:
( 1=3 )
0:18 fFtuk
VRd;F ¼ k 100qeq 1 þ 7:5 fck þ 0:15rcp bw deq ð2aÞ
cc fctk
where
Ap
qeq ¼ ð2bÞ
bw dp
is the flexural reinforcement ratio of a FRC beam with prestressed reinforcements,

where Ap and dp are steel cross-sectional area and internal arm, and bw the width of the
web of the section, while
dl Al þ dp Ap
deq ¼ ð2cÞ
Al þ Ap
is the equivalent internal arm of the flexural reinforcement.

In Eq. (2a) fFtuk is the post-cracking residual tensile strength obtained from either a
direct tensile test of by inverse analysis on prism bending test data; fct is the design
tensile strength of the FRC; cc is a partial safety factor (cc = 1.5), rcp= Nsd/Ac< 0.2fck/
cc is the average stress acting on the concrete cross-section, Ac, for an axial force, Nsd,
due to loading or prestressing actions (Nsd> 0 for compression); and k is a factor that
takes into account the size effect and given by:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
k ¼ 1þ 200=deq 2:0 deq inmm ð3Þ
The characteristic post-cracking residual tensile strength (fFtuk) of the FRC for shear
is determined at a crack opening displacement (COD) of wu= 1.5 mm and is given by:
fFtuk ¼ 0:45fR1k 0:6ð0:65fR1k 0:5fR3k Þ 0 ð4Þ
where fR1k and fR3k are flexural strengths determined in accordance with MC2010. The
fRi values of the SFRC beam are available in Table 4. In the present stage of the
optimization of the SFRC, relatively few number of tests were carried out for the
evaluation of the fRi, the average values of these parameters were adopted for deter-
mining fFtu, by avoiding a too detrimental conjugated effect of this low number of
specimens and relatively large CoV on the fRi,k values. The proposed section has no
stirrups, hence the equation for determining the contribution of the transverse bar
reinforcement (VRd,s) is not provided here. The design shear resistance cannot be
greater than the crushing capacity of concrete in the web:
fck coth þ cota

VRd;max ¼ kc bw :z: ð5Þ
cc 1 þ cot2 h
where z = 0.9 deq is the effective shear depth, h is the inclination of the CDC, kc = ke
ηfc, ke = 0.55 and:
gfc ¼ ð30=fc Þ1=3 1:0 ð6Þ

According to the above-mentioned equations, the design shear capacity of the FRC
beam is 251 KN which is much higher than the maximum shear force (52.5 KN).
4.4 Deflection Control

For the deflection calculation of the slab’s module, the DefDOCROS software was
used. The DefDocros program uses the M-v relationship data derived from DOCROS
and predicts the force versus deflection (F-u) response of simply supported beams by
using the displacement method described elsewhere [25]. According to the Eurocode-2
[19], the maximum deflection should not exceed L/250, where L is the length of the
beam. Hence, the deflection limit for the slab’s module with 12 meters span length is
48 mm. The total long-term deflection of the studied beam is 36 mm where the creep
factor is 1.73 and we have a 6.72 mm upward deflection due to the prestress force.
Hence, the deflection of the beam is in the safe zone.
5 Conclusions
From the observations of this experimental work, it can be concluded that dry SFRC
can be produced with a minimum cement content of 300 kg and the steel fibre content
of 60 kg per cubic metre with 40 MPa compressive strength at 3 days. At 3 days, the
average fR1 and fR3 of SFRC containing fibres 60 kg/m3 were found to be 5.98 MPa
and 6.38 MPa respectively and average shear strength of 13.49 MPa was attained by
the produced SFRC. Analytical results show that using prefabricated prestress SFRC
beam in slab’s module in a residential building the height of slab will be 350 mm and
the shear resistance in the section is less than the maximum shear force so the beam can
be cast without stirrups which are economically advantageous. Further, experimental
investigations on the prototype precast long-span prestressed beams will be conducted
for having a better understanding of the post-cracking behaviour against the flexural
and shear. A design methodology was developed capable of optimize the prestressed
prefabricated SFRC beam and the slab’s module by attending the SLS (deflection) and
ULS (flexural and shear capacity) design criteria.
Acknowledgements. The authors acknowledge the support provided by FEDER funds through
the Operational Programme for Competitiveness and Internationalization (POCI) within the
scope of the project n. 33883, SlabImp- Prefabricated lightweight and multifunctional large span
slabs. The first two and the last Authors would like to acknowledge the grant provided by this
project.
References
1. Park, H., Kang, S., Choi, K.: Analytical model for shear strength of ordinary and prestressed
concrete beams. Eng. Struct. 46, 94–103 (2013)
2. Adebar, P., Mindess, S., StPierre, D., Olund, B.: Shear tests of fiber concrete beams without
stirrups. ACI Struct. J. 94(1), 68–76 (1997)
3. Mobasher, B., Yiming, Y., Soranakom, C.: Analytical solutions for flexural design of hybrid
steel fiber reinforced concrete beams. Eng. Struct. 100, 164–177 (2015)
4. Frazão, C., Camões, A., Barros, J., Gonçalves, D.: Durability of steel fiber reinforced self-
compacting concrete. Constr. Build. Mater. 80, 155–166 (2015)
5. Mo, K.H., Yap, K.K.Q., Alengaram, U.J., Jumaat, M.Z.: The effect of steel fibres on the
enhancement of flexural and compressive toughness and fracture characteristics of oil palm
shell concrete. Constr. Build. Mater. 55, 20–28 (2014)
6. Atiş, C.D., Karahan, O.: Properties of steel fiber reinforced fly ash concrete. Constr. Build.
Mater. 23(1), 392–399 (2009)
7. Mohammadhosseini, H., Tahir, M., Sam, A.R.M.: The feasibility of improving impact
resistance and strength properties of sustainable concrete composites by adding waste
metalized plastic fibres. Constr. Build. Mater. 169, 223–236 (2018)
8. Thomas, J., Ramaswamy, A.: Mechanical properties of steel fiber reinforced concrete.
J. Mat. Civ. Eng. 19, 385–392 (2007)
9. Mohammadi, Y., Singh, S.P., Kaushik, S.K.: Properties of steel fibrous concrete containing
mixed fibres in fresh and hardened state. Constr. Build. Mater. 22(5), 956–965 (2008)
10. Padmarajaiah, S.K., Ramaswamy, A.: Flexural strength predictions of steel fiber reinforced
high-strength concrete in fully/partially prestressed beam specimens. Cem. Concr. Compos.
26(4), 275–290 (2004)
11. Barros, J.A.O., Taheri, M., Salehian, H., Mendes, P.J.D.: A design model for fibre reinforced
concrete beams pre-stressed with steel and FRP bars. Compos. Struct. 94(8), 2494–2512
(2012)
12. Khanlou, A., MacRae, G., Scott, A., Hicks, S., Clifton, G., et al.: Shear performance of steel
fibre-reinforced concrete. In: Australasian Structural Engineering Conference 2012: The
Past, Present and Future of Structural Engineering, p. 400. Engineers Australia (2012)
13. Zamanzadeh, Z., Lourenço, L., Barros, J.: Recycled steel fibre reinforced concrete failing in
bending and in shear. Constr. Build. Mater. 85, 195–207 (2015)
14. Brocks, W., Cornec, A., Scheider, I.: Computational aspects of nonlinear fracture mechanics.
GKSS Forschungszentrum Geesthacht GMBH-Publications-GKSS, no. 30 (2003)
15. European Committee for Standardisation (CEN), ‘Eurocode 1: Basis of Design and Actions
on Structures. Part 1: Basis of Design’, CEN, Brussels, ENV 1991–1 (1994)
16. Bymaster, J.C., Dang, C.N., Floyd, R.W., Hale, W.M.: Prestress losses in pretensioned
concrete beams cast with lightweight self-consolidating concrete. Structures, vol. 2, pp. 50–
57 (2015)
17. AASHTO, LRFD. Specifications for Highway Bridges (1997)
18. Robitaille, S., Bartlett, F.M., Youssef, M.A.: Evaluating Prestress Losses During Pre-
tensioning. St. Johns Canadian Society for Civil Engineering, Canada (2009)
19. European Committee for Standardisation (CEN), ‘Eurocode 2: Design of Concrete
Structures. Part 1: General Rules and Rules for Buildings’, European Prestandard, CEN,
Brussels, ENV 1992–1-1 (1991)
20. Basto, C.A.A., Barros, J.A.O.: Numeric simulation of sections submitted to bending.
Technical report 08–DEC/E–46, p. 73 (2008)
21. Barros, J.A.O., Taheri, M., Salehian, H., Mendes, P.J.D.: A design model for fibre reinforced
concrete beams pre-stressed with steel and FRP bars. Compos. Struct. 94(8), 2494–2512
(2012)
22. Soranakom, C., Mobasher, B.: Flexural design of fiber-reinforced concrete. ACI Mater.
J. 106(5), 461–469 (2009)
23. Model Code 2010. fib Model Code for Concrete Structures 2010. International Federation
for Structural Concrete (fib). Ernst & Sohn, Berlin, Germany (2013)
24. Barros, J.A.O., Foster, S.J.: An integrated approach for predicting the shear capacity of fibre
reinforced concrete beams. Eng. Struct. 174, 346–357 (2018)
25. Varma, R.K.: Numerical models for the simulation of the cyclic behaviour of RC structures
incorporating new advanced materials (2013)
Simulation of Fibre Orientation
in Self-compacting Concrete: Case Studies
Thomas Bauwens1,2, Steffen Grünewald1,3(&), and Geert De Schutter1

1
Ghent University, Ghent, Belgium
Steffen.Grunewald@ugent.be
2
Dirk Bauwens nv, Evergem, Belgium
3
Delft University of Technology, Delft, The Netherlands
Abstract. Recent developments in concrete technology with high potential

include ultra high performance concrete and self-compacting fibre reinforced
concrete, which have a flowable consistency and can transport relatively high
fibre dosages. Flowability is achieved by adopted mix design and both the mix
design and flow affect the distribution and orientation of the fibres, which affect
the post-cracking behaviour and accordingly the structural performance. With
new materials also come new manufacturing and design approaches. The pre-
diction of fibre orientation with computational fluid dynamics (CFD) simulations
can be an important instrument to predict, understand and influence fibre ori-
entation. With better understanding the mix design and casting process can be
optimized.
This paper reports about a study executed to determine the applicability of the
software package Autodesk Moldflow for fluid dynamics simulations of flow-
able fibre concrete. After a discussion of relevant literature, two reference cases
address stretching and shearing flow conditions in a qualitative and quantitative
way. Concrete was modelled as an incompressible Bingham material with
addition of a fibre orientation model that was developed by Folgar and Tucker.
A third case, a square panel, was used as a reference and structural element for
flow simulations. Parameters varied were among others rotary diffusion, wall-
slip and duration of casting.
Keywords: Self-compacting concrete Fibres Fibre orientation Flow

simulation CFD
1 Introduction
Flowable Concrete
The use of flowable concrete (FC) facilitates the construction with concrete and very
specific solutions can be realised such as remote casting, casting in very congested
areas and the production of concrete surfaces with architectural appearance. FC is a
cluster of types of concrete, which distinguishes itself from traditional concrete through
rheological characteristics obtained by tailor-made mix design and component selec-
tion. The flow behaviour of concrete can be modelled by rheological laws with a
relation between the shear stress s and the shear rate c_ . The simplest model that
captures the nature of the flow of concrete is the Bingham model [1], which is shown
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_6
Simulation of Fibre Orientation in Self-compacting Concrete 65
by Eq. (1). This relatively simple law can be defined by two fundamental parameters:
the yield stress sy and the plastic viscosity µp. Fibres, especially at higher dosages, have
a significant effect on both rheological characteristics.
s ¼ sy þ lp c_ ð1Þ
Fibre Reinforced Concrete

Fibres have been added to concrete to transfer forces across cracks occurring during
hardening or in service. Accordingly, the most important contribution of fibres is
situated in the post-cracking tensile behaviour of concrete. The contribution of the
fibres depends on their location and orientation with regard to the crack surface. The
fibre orientation in flowable concrete types depends on the boundary conditions of the
flow; the influence of fibre orientation on the structural performance is caused by the
probability that a fibre will cross a crack. Fibre orientation can be experimentally
determined for example by manual counting, image analysis, computed tomography,
magnetic methods and electric conductivity. The knowledge of how fibres orient and
how this information can be implemented for design and execution is key for the
optimisation of their application. Potentially, fibres can be used more efficiently with
adequate material selection, mix design and casting operation.
Fluid Dynamics Simulations with Fibres

The flow of concrete is a rather complex problem to model in a simulation [2]. A first
method to model concrete is computational fluid dynamics (CFD), which assumes that
the concrete is a single-phase liquid [3, 4]. The problem of the flow of concrete is
solved by a set of flow equations, that are defined as partial differential equations
(PDE). These PDE are converted into a set of algebraic equations which are solved at
discrete points in space and time. The second method is the discrete element method
(DEM) [5]. This method will reduce the concrete as if it was a set of interacting
particles. The interaction in itself is modelled as normal and tangential forces that act
on each other using springs, dashpots and slip elements to model the flow. This method
primarily focusses on the discrete character of concrete. As always, there is a research
field that tries to combine the best of both worlds. More advanced simulations will
require more assumptions and come at higher computational costs, but more phe-
nomena can be modelled and explained.
All fibre orientation phenomena can be explained by two main drivers for fibre
orientation: shearing and stretching flows. In an extensional flow mode, the shear forces
will apply a torque on the fibre to align in the direction of the extensioning flow. In a
flow dominated by shear stresses, the torque exerted on the fibre reaches a minimum
when the fibre is parallel to the flow direction (e.g. in the presence of a wall). Both
phenomena are illustrated by Figs. 1a and 1b. In order to express the orientation
distribution of fibres, first of all, the orientation of a single fibre is defined. The fibres
are assumed to be rigid cylinders, uniform in length and diameter; the orientation of a
single fibre can be described as a unit vector p. The theory of fibre orientation is based
on the works of Einstein [6] and Jeffery [7], who described respectively the movement
of a sphere and ellipsoid in a fluid. The movement of a single fibre in a fluid can be
assumed similar to the movement of an ellipsoid in a fluid. Therefore, the differential
66 T. Bauwens et al.
equation of Jeffery is applicable to the modelling of fibre orientation. Folgar and Tucker
[8] suggested a diffusion coefficient to take into account the interaction between fibres
and they proposed the relation of Dr ¼ CI c_ , in which CI is the fibre-fibre interaction
coefficient depending on the fibre geometry and volume fraction which can be obtained
through experiments. c_ is the effective shear rate. From a computational point of view,
the modelling of every single fibre is not feasible, since this approach requires a lot of
storage and computational effort. Therefore, it was necessary to capture the nature of
the fibre orientation with a simplification. This simplification was investigated and
implemented in the fibre orientation model by replacing the fibre orientation distri-
bution by a second order tensor.
a) Stretching flow: fibres align in the

stretching direction
b) Shearing flow: fibres align in the

shearing direction
Fig. 1. Principal mechanisms for fibre orientation a, top) stretching flow and b, right) shearing
flow.
The use of fibre orientation modelling has been applied to the study of the orien-
tation of fibres in concrete a few times in the past decade [9, 10]. With new design
approaches that allow taking into account fibre orientation, this knowledge can also be
used to optimize material and structural performance and the casting process. In this
study, the commercial software Autodesk Moldflow was applied as a tool, that still
needed verification through appropriate experimental and numerical studies. Three
cases were selected for first validations and to determine the need for future research.
2 Simulation Model
2.1 Software Applied

Autodesk Moldflow is a software type of the CFD-class that was designed for simu-
lating and optimizing injection molding processes, which has some similarities with the
casting of concrete and therefore, with an adaptation of the software, the flow of
concrete can be modelled. These adaptations can be done by User Defined Functions
for rheological and constitutive material laws. The biggest advantage of Moldflow over
other softwares is the already implemented fibre orientation model of Folgar-Tucker.
The latter was the reason why this software package was selected to perform
simulations.
2.2 Material Characteristics

To make Moldflow suitable for the modelling of concrete, adaptations were made with
regard to the constitutive and rheological properties of the material [11]. Since concrete
was not yet modelled within Moldflow Insight, some adaptations were developed on
the pressure, volume, velocity-relationship and rheological laws of the model. Mold-
flow models the fluid default with a 2-domain Tait pvT-model and Cross-WLF vis-
cosity model. Concrete was modelled as an incompressible fluid with a Bingham
plastic rheological law. This was done by assembling a dynamic link library (.dll-file)
in Visual Studio and using Moldflow Application Programming Interface (API), which
allows the user to insert the desired plastic viscosity and yield stress in the model.
Three parameters are necessary to characterize the implemented Bingham model: the
plastic viscosity, the yield stress and the shear rate cut-off. The first two parameters are
necessary to facilitate the standard Bingham model. The third parameter is necessary to
reduce the computation time by limiting the viscosity for low shear rates. This sim-
plification does not influence the final solution since fibre orientation happens at high
shear rates and the cut-off is chosen low enough to not influence the flow. Potentially,
computation time could be saved by increasing this criterion. For the following three
cases, a plastic viscosity of 105 Pas, a yield stress of 37 Pa and a shear rate cut-off of
0.0001 s−1 were used. The first two parameters are the same as measured by [12]; the
addition of steel fibres typically increases both characteristics compared to a reference
SCC without fibres. The dosage and the type of fibres were taken into account by the
rotary diffusion coefficient CI. A CI. of 0 means that no fibre-fibre interactions were
taken into account and therefore the simulations were run with the differential equation
that was derived by Jeffery.
3 Case 1: Radial Flow
The fibre orientation model was verified by simulating flow fields which induce
shearing and stretching flows (Sects. 3 and 4). First a radial flow field was simulated in
which the fibres were expected to align perpendicular to the flow direction. The model
consists of a cylinder with a radius of 0.50 m and a height of 0.30 m. The results of the
fibre orientation at the end of the simulation are shown in Fig. 2 and they confirm
qualitatively the principle of flow-induced fibre orientation by stretching. At the wall of
the panel mould the fibres are mainly oriented perpendicular to the flow direction, as
was also observed e.g. in experimental studies executed with panel casting [13] and on
tunnel segments [14]. In the centre, the fibre orientation is relatively random. A second
observation is that fibre orientation occurs very fast, such observation can be also made
after the execution of the slump flow test, after which the preferred orientation of fibres
at the border of the flow spread is visible. The transition from green to yellow/red
colour in the simulations mainly is observed in the last third of the panel radius
(orientation number larger than 0.70). The maximum fibre orientation obtained by
panel testing was 0.866 [13], as this was an average value the results of the simulations
are realistic. Orientation numbers of up to about 0.90 were found in tunnel segments
[14] in a similar flow condition.
Fig. 2. Final average fibre orientation after radial flow, fibre orientation tensors in X-Y plane.
4 Case 2: Flow Through a Pipe
4.1 Simulation Without Wall Slip

The shearing flow is obtained by simulating the flow in a cylindrical pipe with a
diameter of 300 mm and a length of 10 m (time to fill the pipe: 140 s). As illustrated in
Fig. 3, the fibres quickly align parallel to the flow direction close to the walls as
expected (location 2 m from inlet). This confirms in a qualitative manner that the fibre
orientation model is correct for shearing flow fields. Martinie and Roussel [15] carried
out flow simulations on the shear flow between two parallel plates; with a fibre rein-
forced SCC having a yield strength of 50 Pa orientation numbers of up to 0.966 were
obtained close to the walls.
4.2 Simulation of Wall Slip

Furthermore, the influence of a wall slip model on the fibre orientation was investigated
with Moldflow. This was done by variation of wall slip model parameters for the pipe
flow simulation. From a macroscopic point of view, the physical phenomenon of wall
slip can be modelled as a slip velocity. In this approach the velocity at the boundary is
not taken equal to zero, as is assumed in classic no-slip boundary conditions of
Newtonian fluid mechanics. When wall slip occurs, the shearing of the fluid will reduce
Fig. 3. Fibre orientation within the pipe.
Fig. 4. Flow simulations a, left) velocity profile in the pipe, with wall slip and without wall slip
and b, right) fibre orientation within the pipe, with wall slip.
and the influenced zone of fibre orientation due to shearing will be smaller. As a result,
there is only a noticeable velocity gradient at the wall of the pipe and therefore the fibre
orientation will occur slower compared to a model without wall slip. Fibre located in
the centre will orientate only very slowly. The effect on the velocity profile is shown in
Fig. 4a. Wall slip was observed with a flow velocity of about 8 cm/s. In the middle of
the pipe the flow occurs as a plug flow. In the model without wall slip, a parabolic
velocity profile is obtained. The fibre orientation was investigated at the same location
(2 m away from the inlet) as shown by Fig. 3 and the orientation profile is shown in
Fig. 4b. The results show that the influence of the wall slip on the fibre orientation is
not negligible. It can be noticed that the fibre orientation is lower at the walls of the
pipe compared to the model without wall slip and that the fibre orientation remains
random in the middle of the pipe. Both phenomena can be explained by the difference
in velocity profiles. Since shearing is equal to the gradient of the velocity profile, the
shear rate will be less at the walls of the pipe and zero in the middle of the pipe. As
fibre orientation is driven by shearing, it is logical that the fibre orientation is relatively
lower at the walls and no fibre orientation takes place in the centre of the pipe. It can be
concluded that the influence of the wall slip on the fibre orientation is not negligible
and therefore further analysis is required on the subject of fibre orientation and the
influence of wall slip.
5 Case 3: Casting of a Slab
5.1 Simulation Set-up

5.1.1 Material Characteristics
In order to verify the accuracy of the model, experimental data from literature was
chosen as a reference case. Žirgulis [12] discusses in his PhD-thesis experimental
results as well as numerical simulations of the casting of fibre reinforced concrete. The
experimental work was done at the Norwegian University of Science and Technology
in Trondheim and the numerical simulations were carried out at the Technical
University of Denmark in Lyngby. The experimental work consisted of casting square
slabs with a width of 1.2 m and a thickness of 150 mm. The pouring was executed at a
distance of 200 mm from the sides through a hose with a diameter of 150 mm from a
height of 200 mm. The concrete contained 0.5 vol.% steel fibres. The fibres had
hooked ends and were produced of cold-drawn wire, having a length of 60 mm, an
aspect ratio of 80, and a minimum tensile strength of 1050 N/mm2. The rheological
properties were determined 20–30 min after mixing with the 4C rheometer (plastic
viscosity: 105 Pas; yield stress: 37 Pa). The density of the fresh concrete was
2486 kg/m3. The surfaces of the formwork were smooth.
The following five general conclusions were drawn by Žirgulis: 1) The fibres align
at the walls due to shearing, 2) The fibres align perpendicular to the flow in the middle
of the plate due to stretching, 3) The stretching effect is more present in the top half of
the plate than the bottom half and 4) The fibres shear over the bottom at the left under
corner and 5) At the inlet the fibres are randomly oriented. Based on all comparisons
between results of CT scans and flow simulations, it was concluded in [12] that the
stretching and shearing were underestimated in the numerical simulation, but qualita-
tively, the results were accurate. The conclusions of Žirgulis were qualitatively
assessed and confirmed in this study.
5.1.2 Geometry, Mesh and Process Settings

The geometry of the slab was then reduced to the simplest form which was a simple
box with a cylindrical inlet condition with a diameter of 150 mm. The final mesh
contained 1,377,439 3D tetrahedral elements. These were generated by first creating a
grid on the planes of the model and then meshing over the depth of the elements. A bias
factor of 2 and a global mesh factor of 0.9 were used as local mesh refinements to solve
the viscous effects. The viscosity function and the pvt-relation were installed by user-
defined API’s. The modelled fluid was an incompressible Bingham material with
properties reported in Sect. 2.2. The inlet was automatically controlled in such a way
that the full slab would be filled in 300 s. The mould was underfilled (concrete kept a
slope during casting) and hence the simulation was aborted when the slab had a
minimum thickness of 150 mm. This was after approximately 120 s.
5.2 Simulation Results

5.2.1 Model Without Rotary Diffusion
The slab model was simulated with Moldflow to show its capabilities. No rotary
diffusion was included in the first simulation since it was used as a benchmark.
Practically this means that fibre interactions are neglected, the stretching patterns are
less visible. Some rotary diffusion is necessary to show stretching behaviour. This can
be explained by the Jeffery orbitals, the orbitals that are perpendicular to the flow are
less stable and therefore during simulation the numerical solution can occasionally fail.
When using rotary diffusion, this disturbance is averaged over the neighbouring cells.
Figure 5a shows the fibre orientation tensors at the bottom of the slab. The range of
orientation numbers is indicated in Fig. 5 from dark red (0.90) to dark blue (0.333). At
the casting location, shearing flow is most present and hence the fibres orientate parallel
to the flow direction, which is radial. In the corner opposite to the inlet, the fibre
orientation is even more pronounced due to the converging flow field. Figure 5b shows
the fibre orientation tensors at the top of the slab. At the top of the slab, the shear rate is
smaller and hence the stretching flow is more present. The fibres aligned more per-
pendicularly to the flow direction compared to the bottom of the mould. The fibre
orientation simulation seems to be accurate, also when compared with the outcomes of
simulations described by [12].
Fig. 5. Overview of fibre orientation in the slab a, left) at the bottom and b) at the top of the slab,
fibre orientation tensors in X-Y directions
The conclusions of Žirgulis are confirmed qualitatively with the Moldflow simu-
lations (Fig. 5): At the inlet where the concrete was cast fibres were randomly oriented.
Fibres preferably oriented along walls and perpendicular to the flow direction in the
middle of the plate. In the top half of the plate stretching was more pronounced
compared to the bottom half.
5.2.2 Influence of Rotary Diffusion

Figure 6 shows the effect of the rotary diffusion on the local fibre orientation at the side
of the plate; the fibre-fibre interaction coefficient CI was varied. As discussed, rotary
diffusion takes into account fibre-fibre interactions and increases the stability of the
solution. The position of the simulation location for Fig. 6 is at half the length of the
mould (side of the mould); values in the x-axis are distances from the bottom of the
mould. The fibres aligned along the wall at this location. In general, the principal value
of the fibre orientation is reduced at increasing interaction. It can be concluded that the
rotary diffusion reduces the fibre orientation component Tyy and increases the fibre
orientation in the other directions.
Fig. 6. Principal values of fibre orientation tensor for various CI-values (wall closest to inlet,
half the length, close to the wall).
5.2.3 Influence of Wall Slip

The influence of wall slip on the principal value of the fibre orientation tensor in the
centre of the slab is shown by Fig. 7. The fibre orientation at the bottom of the slab is
not greatly influenced, the fibre orientation at the top is. The fibre orientation at the top
shows more a stretching fibre orientation, which resembles the experimental results
better compared to the model without wall slip.
5.2.4 Influence of Duration of Casting

Figure 8, the influence of a longer duration of casting (about doubled to 220 s) on the
principal value of the fibre orientation tensor in the centre of the slab is shown. This
reduces the flow rate, and hence less shearing will be present. It can be noticed that, at
the bottom, the fibre orientation is lower when the flow velocity decreases, and fibre
orientation is higher at the top of the slab. Both phenomena can be explained by the
lower shearing flow condition during the casting of the slab.
Fig. 7. Principal values of fibre orientation tensor with and without wall slip, in the centre of the
plate.
Fig. 8. Principal value of fibre orientation tensor for longer casting duration, in the centre of the
plate.
6 Conclusions
This paper discusses the application of Autodesk Moldflow fluid dynamics software for
the simulation of fibre orientation in specific cases with self-compacting fibre rein-
forced concrete. Case studies were executed to simulate flow fields which induce
shearing and stretching flows; cases studied were a round panel, a pipe and a slab.
Affecting parameters on the outcome of the simulation were assessed. Based on the
study the following conclusions can be drawn:
• The qualitative accuracy of the fibre orientation model was demonstrated through
case studies, where the fibre orientation tensors aligned respectively parallel and
perpendicular to the flow direction.
• The addition of rotary diffusion in the model smears out the fibre orientation tensor.
This reduces the principal values of the fibre orientation tensors and increases the
stability of the solution.
• With wall slip the gradient in the velocity profile decreases and therefore, the
shearing of the flow is reduced.
• A lower flow rate reduced the velocity and shear rate in the concrete. This reduced
the principal value of fibre orientation at the walls.
• A more detailed analysis of the quantitative accuracy is required for further
implementation.
References
1. Wallevik, O.H.: Rheology – a scientific approach to develop self-compacting concrete. In:
3rd International Symposium on SCC, Rilem, Reykjavik, pp. 23–31 (2003)
2. TC 222-SCF, Simulation of fresh concrete flow, State-of-the-art report of the RILEM
Technical Committee 222-SCF, Eds.: Roussel, N., Gram, A., ISBN: 978–94-017-8883-0
(2014)
3. Thrane, L.N.: Form Filling with Self-compacting Concrete. PhD-thesis, DTU Lyngby (2007)
4. Gram, A.: Modelling Bingham Suspensional Flow - Influence of Viscosity and Particle
Properties Applicable to Cementitious Materials. PhD-thesis, KTH Royal Institute of
Technology, Stockholm, ISSN 1103–4270 (2015)
5. Ferrara, L., Shyshko, S., Mechtcherine, V.: Predicting the flow-induced fiber dispersion and
orientation in self-consolidating concrete by distinct element method. BEFIB 2012, ISBN:
978–2-35158-132-2, pp. 213-214 (2012)
6. Einstein, A.: Eine neue Bestimmung der Moleküldimensionen. PhD-thesis, Universität
Zürich. https://doi.org/10.3929/ethz-a-000565688 (1905)
7. Jeffery, G.B.: The motion of ellipsoidal particles immersed in a viscous fluid. Proc. R. Soc.
Lond. Ser. A 102(715), 161–179 (1922)
8. Folgar, F., Tucker, C.L.: Orientation behavior of fibers in concentrated suspensions. J. Reinf.
Plast. Compos. 3(2), 98–119 (1984)
9. Martinie, L.: Comportement rheologique et mise en œuvre des materiaux cimentaires fibres.
PhD-thesis, 2010PEST1077 (2010)
10. Svec, O., Skocek, J., Stang, H., Olesen, J.F., Poulsen, P.N.: Flow simulation of fiber
reinforced self-compacting concrete using Lattice Boltzmann method. Dissemination (2011)
11. Bauwens, T.: Fibre Orientation in Self-compacting Concrete: Effect of Mix Design and
Casting Process. Master-thesis, Ghent University, Ghent (2018)
12. Zirgulis, G.: Fibre Orientation in Steel-Fibre-Reinforced Concrete: Quantification methods
and influence of formwork surface and reinforcement bars in structural elements. PhD-thesis,
NTNU Trondheim (2015)
13. Abrishambaf, A., Barros, J.A.O., Cunha, V.M.C.F.: Relation between fibre distribution and
post-cracking behaviour in steel fibre reinforced self-compacting concrete panels. Cem.
Concr. Res. 51, 57–66 (2013)
14. Grünewald, S.: Performance-based design of self-compacting fibre reinforced concrete. PhD-
thesis, Delft University of Technology (2004)
15. Martinie, L., Roussel, N.: Fiber-reinforced cementitious materials: from intrinsic properties
to fiber alignment. In: Khayat, K.H., Feys, D. (eds), Design, Production and Placement of
Self Consolidating Concrete, Dordrecht, RILEM Bookseries, vol. 1, pp. 407–415 (2015)
Mix Design and Properties of Self-compacting
Fibrous Concrete
Rafael R. Polvere(&), Ana R. L. Pires,

Sidiclei Formagini, and Andrés B. Cheung
FAENG/UFMS, Federal University of Mato Grosso do Sul,

Campo Grande, Brazil
rafaelpolvere@gmail.com
Abstract. The development of self-compacting fiber reinforced concrete

(SCFRC) marks an important milestone of the Brazilian building industry,
because it combines the benefits of high fluidity in the fresh state, better per-
formance on tensile strength and the control of the cracks. To an efficient
performance, it is necessary a good granular mixture proportioning. Thus, the
objective of this research is the evaluation of the influence on fiber contents in
self-compacting concrete of 40 MPa and its properties in fresh and hardened
state. The difference of the concretes was the volume levels of the steel fibers in
each mixture proportioning: SCC0F (no addition), SCC0.5F (0,5% in volume)
and SCC1F (1,0% in volume). The concrete mix design was based in the
compressible particle model. The results show that the insertion of steel fibers
interferes diminishing the workability and fluidity of the mixtures and improve
the tensile strength.
Keywords: Self-compacting fiber reinforced concrete Steel fibers Mixtures

design, compressible packing model
1 Introduction
Self-compacting concrete (SCC) is characterized by its self-compacting capability

without the need for additional internal or external vibration. Your development in
1988 enabled the execution of concrete structures without the need for vibration due to
its flow ability and self-consolidating properties [1]. It was an important milestone in
improving the efficiency of the construction industry, because it is an easy and fast to
apply and requires a reduced number of people for application. The requirements
properties of SCC in the fresh state are mainly resumed to the filling ability, the passing
capacity, and the resistance to segregation.
As concrete has low tensile strength, the introduction of fibers improves its
behavior when requested to tensile and flexural bending. Fibers, when added in
appropriate volumes, can overcome this deficiency by attributing ductile behavior.
Steel fibers proved to have the potential to increase the post cracking energy absorption
capacity of cement-based materials, enhancing the ductile properties of concrete
structures behavior [2].
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_7
76 R. R. Polvere et al.
This way steel fiber reinforced self-compacting concrete (SCFRC) combines the
benefits of SCC in the fresh state reducing cracks and showing an improved perfor-
mance in the hardened state compared to conventional concrete. The composition and
production of SCFRC are more complicated than SCC, because it has an ideal volume
fraction of steel fiber to ensure higher workability and better mechanical performance.
Due to the packing effect of the fiber-aggregate solid skeleton, the self-compacting
performance could not be achieved once the volume fraction of steel fiber in concrete
exceeds the limited value even if the concrete mixture is a homogeneous and stable
suspension [3]. The limitation, usually less than 2%, is mainly affected by the prop-
erties of raw material [4] and granular mix proportion [5].
Grunewald [6] proposed a performance-based mix design method for SCFRC. The
method used the Compressible Packing Model (CPM) [7] and assumes that the steel
fiber is the equivalent packing diameter [8]. The packing density of the aggregates and
also the fibers in the mix of SCFRC determine the amount of cement paste that is
required to fill the interstices of the granular skeleton.
When reinforced in modulus and amount with fibers, concretes minimize this
fragile behavior. According to Gois [9], fibers act as connecting bridges, transferring
stresses from side to side of the matrix, minimizing stresses at the ends of the cracks.
Naaman [10] proposed that the main advantage in adding fibers to concrete is the
ability to modify the behavior of the material from brittle to ductile when concrete is
evaluated the traction, because the fibers, when crossing the cracks, create bridges that
make it difficult to increase their opening. Already Melo Filho [11] observed significant
increases in tensile strength after the appearance of the first crack when evaluating the
direct tensile tests was performed.
Although the mechanical benefits of fibers added to concrete are known, few
studies exist on the joint action of the hybrid effort of steel fibers on concrete. In this
paper, the objective is the comparison and analysis between SCC and SCFRC, which
have two different levels of steel fiber incorporation. To achieve the overall goal it’s
necessary to fulfill the specific objective, which is the study of the determination of the
dosage and characterization of the properties of steel fiber reinforced self-compacting
concretes, in the fresh and hardened state.
Three concrete mixes were designed: one reference mix of SCC without fibers
(SCC0F) and two SCFRC with steel fibers volume insertion of 0.5% (SCC0.5F) and
1.0% (SCC1F). All concrete were produced with the same dry granular mixture pro-
portioning (except the steel fiber contents) in an inclined shaft concrete mixer, being a
production of 122 L of concrete for each mix. This amount was sufficient for the fresh
state characterization tests, molding of specimens to determine the properties in the
hardened state.
Mix Design and Properties of Self-compacting Fibrous Concrete 77
2.1 Materials
The cementitious materials used to produce the concretes were Brazilian Portland
cement type CP II E– 32 (addition varying from 5 to 32% of blast furnace slag) and
silica fume (SF). Aggregates were used: coarse aggregate originating from the region’s
basaltic rock crushing (G12.5), natural very fine sand (NS) and gravel sand (GS). As
the natural sand is very fine crushing sand was used for grain size correction. It was
chosen to work with a superplasticizer additive that provided fluidity, avoiding seg-
regation and giving concrete more workability and water from the supply network. The
steel fibers inserted in the last two mixes had hooks at the ends and a ratio (l/d) of 50,
30 mm of length and 0.60 of diameter.
The materials were experimentally characterized according to the input parameters
required by the packing density model [12] used for the granular mixture propor-
tioning. The particle size distributions of aggregates (by mechanical sieving) [13] are
illustrated in Fig. 1. The maximum aggregate size was limited to 12.5 mm, so that there
was no discontinuity in the granular skeleton, since that the natural sand was very thin,
as well as reducing the risk of segregation during the application of self-compacting
concrete. The other physical properties of aggregates required for the preparation of
self-compacting concrete mixtures are presented in Table 1.
0.0
Accumulated Fraction
20.0
40.0 G.12,5%
(%)
GS
60.0 NS
80.0
100.0
0.1 1
Diameter (mm) 10
Fig. 1. Aggregate grading curve.
2.2 Concrete Mixture Proportioning

The theoretical mixer design of self-compacting concretes was performed by the
computer model [12] based on the theoretical formulation using the compressible
packing density model [7]. From various simulations of different granular composi-
tions, appropriate selection of materials was carried out to produce the maximum
packing density of the granular mixture, ensuring that all desired physical properties
both in the fresh state and after hardening were achieved. The challenge of mixture was
to associate the use of very fine natural sand with the high density coarse aggregate. For
high levels of fines required for a good consistency of the self-compacting concrete, the
continuity of the granular skeleton is guaranteed by the mortar, with the large aggregate
Table 1. Aggregate properties.

Property Test Methodology Unity Material
G 12.5 SN CS
3
Unit Mass [14] g/cm 1.68 1.59 1.90
Specific Mass [15] g/cm3 2.88 2.63 2.85
Water Absorption % 0.71 – –
Shape Index [16] – 2.03 – –
Organic Impurities [17] ppm – 300 –
Powdery Material [18] % 1.89 1.40 11.0
Characteristic Maximum Dimension [13] mm 12.5 0.60 4.75
Fineness Module [13] – 6.06 1.17 3.14
being dispersed in smaller quantities in the mortar, which facilitates the segregation of
the concrete. The three mix designs are presented in Table 2.
Table 2. Mixes of the concrete produced to 1 m3.

Material Unity Consumption by m3 of
concrete
SCC0F SCC0.5F SCC1F
Cement kg/m3 390 390 390
Silica Fume kg/m3 38 38 38
Natural Fine Sand kg/m3 661 661 661
Crushing Fine Sand kg/m3 301 301 301
Coarse aggregate kg/m3 966 966 966
Steel Fibers kg/m –
3
40 81
Water l/m3 190 198 198
Superplasticizer l/m3 4.80 4.90 4.50
Mortar Content % 69.7 69.7 69.7
Expected Compressive Strength, 28 days MPa 40 40 40
Ratio w/c 0.49 0.51 0.51
2.3 Testing Method

Tests were performed shortly after concrete production to evaluate their behavior in the
fresh state. Fresh state tests evaluate and frame self-compacting characteristics and
properties according to the requirements of the technical standard of Brazilian NBR
15823 [19]. The tests were performed to determine the spreading, flow time and visual
stability index, determining the passing ability, and viscosity determination. Figure 2
shows parts of the tests performed in the fresh state of the concrete: Abrams cone
spreading; flow through L box; and flow through V funnel.
(a) Abrams cone. (b) L box. (c) Funnel V.
Fig. 2. Fresh concrete tests.
After completion of the tests in the fresh state, 20 cylindrical and 3 prismatic
specimens were molded in each mixture to characterize the hardened state concrete. All
cylindrical specimens were 10 cm in diameter and 20 cm high and the prismatic
specimens were 15 15 50 cm3. The samples were kept in a humid chamber for
curing until the tests were performed. To determine the mechanical properties of
concrete, compression tests were performed [20], tensile strength [21], splitting test
[22], modulus of elasticity [23], and determination of the absorption, voids index and
specific mass of concrete [24].
3 Presentation and Analysis of Results
3.1 Fresh Concrete Properties

The behavior of the concrete in the fresh state was different in the mixtures produced,
since the same composition of the materials was considered and only the fiber content
was varied. Even with the mixture of very fine natural sand with the high density
aggregate it was possible to produce SCC with satisfactory behavior in the fresh state.
In general, they had good spread through the Abrams cone as shown in Fig. 3 and
experimental data available in Table 3. SCC0F was homogeneous, cohesive and
without evidence of segregation and exudation. SCC0.5F indicated evidence of exu-
dation as showed at the edges of the spread concrete, not being significant to affect its
stability. In SCC1F the spread showed that fresh concrete was less homogeneous, with
a small accumulation of coarse aggregate in the central region. It is believed that for the
granular mixture used, the 1% fiber percentage has affected the mobility of the
aggregates contributing to such behavior and appearance, being a critical point for this
fibers percentage and the mixture proportioning not ware adjusted for the fiber
contents.
The spreading values of the Abrams cone varied within the necessary range to be
classified as SCC according to NBR 15823 [19]. In respect to the apparent plastic
a. SCC0F. b. SCC0.5F. c. SCC1F.
Fig. 3. Spreading and aspects of concretes.
Table 3. Properties and framing of concrete in the fresh state.

Test SCC0F SCC0.5F SCC1F
Result Group Result Group Result Group
Visual Inspection Highly stable IEV 0 Stable IEV 1 Stable IEV 2
Spreading (slump flow) 750 mm SF2 720 mm SF2 670 mm SF2
Apparent viscosity t500 <2s VS1 2 s VS1 7 s VS2
Passing skill - box L (3 bars) 0.82 PL2 1.0 PL2 Blocked –
Viscosity - funnel V t10s 6s VF1 6 s VF1 11 s VF2
viscosity t500, the concrete SCC0F and SCC0.5F presented spreading times of less than
2 s, while SCC1F took longer to reach this mark.
It was observed that the higher the fiber content, the less concrete was spread.
Several hypotheses were suggested for this behavior: the very fine sand has a high
specific area, which requires a large amount of water to wet the grains, increasing the
internal friction and reducing the mobility of the concrete as the fiber content in the
mixture increased [25]; the difference in density between the matrix and the coarse
aggregate may also have contributed to this, because the coarse aggregate is denser and
is immersed in a mixture where the continuity of the granular skeleton is governed by
the matrix, which may have favored the beginning of segregation with the increase in
fiber contents; and the unadjusted mixture proportioning for the fiber contents.
The experimental results of the “L” box test using three bars as concrete flow
restriction were satisfactory for concrete SCC0F and SCC0.5F. The 1% fiber content in
SCC1F promoted concrete blocking, indicating that its flow was not appropriate for the
number of bars adopted. This indicates that SCC1F cannot be recommended for use in
places where concrete application is restricted such as high reinforcement rates.
The same confined flow time for SCC0F and SCC0.5F concrete was observed over
the apparent plastic viscosity determined by the V funnel. For SCC1F, the confined
flow time of the concrete increased slightly indicating that the 1% fiber content made
this mobility more difficult. This flow capacity, which depends on the relationship
between the maximum aggregate size and the size of the bottom funnel opening, as
well as the viscosity and stability of the material, was reduced with the increase in fiber
contents.
3.2 Hardened Concrete Properties

The average values of the mechanical characterization tests of hardened concretes and
their respective coefficients of variation (CV) are presented in Table 4. Making a
statistical analysis with the three concretes produced, we are obtained the resistance fcm
of 42.9 MPa with a variation coefficient of the 5.3%. In the case of the obtained results,
there was no significant difference between all mixtures because the mixture propor-
tioning not ware adjusted for the fiber contents. The fibers may only have acted
similarly to a concrete confinement system, restricting the development of longitudinal
cracks.
Table 4. Properties of concrete in the hardened state.

Proprieties SCC0F SCC0.5F SCC1F
Average Resistance to Simple fcm CV fcm CV fcm CV
Compression, fcm (MPa) (%) (MPa) (%) (MPa) (%)
fcm - 3 days 17.9 0.4 15.9 0.2 14.4 0.6
fcm - 7 days 25.8 3.3 26.7 0.8 26.3 3.4
fcm - 14 days 35.7 4.0 35.4 1.6 31.6 3.3
fcm - 28 days 41.8 1.5 45.6 4.8 41.5 4.0
Static Modulus of Elasticity to Em CV Em CV Em CV
Compression, Em (GPa) (%) (GPa) (%) (GPa) (%)
Em - 28 days 26.7 3.4 30.5 3.2 29.4 3.2
Splitting tensile strength, fct,sp fct,sp CV fct,sp CV fct,sp CV
(MPa) (%) (MPa) (%) (MPa) (%)
fct,sp - 28 days 3.7 1.5 5.7 4.5 5.1 9.7
Figure 4 shows the evolution curves of compressive strengths of concrete as a

function of test ages. These curves indicate a close behavior between them, with
significant strength gains between the ages of 14 and 28 days. Faced with this behavior,
a growth potential of this resistance is noticed for ages greater than 28 days, because
the Portland cement used is composed of the addition of blast furnace ground slag.
The modulus of elasticity is the most important characteristic for deformation and
stiffness control, then, when verifying the results, it is considered that there is a slight
50
40
Compressive Strength (MPa)
30
20 SCC0F
SCC0.5F
10 SCC1F
0
0 7 14 21 28
Age (days)
Fig. 4. Evolution of compressive strengths of concretes up to 28 days.
increase with the insertion of the fibers. This difference is an indication that fibers can
help with the rigidity and integrity of the material. In the SCC1F mixture it is noted that
the modulus of elasticity had a decrease when compared to the SCC0.5F. This dis-
continuity may have affected this property by the fact that the inclusion of many fibers
causes a disturbing effect on the concrete packing, distancing the grains from the
aggregates.
When the effects of the influence of steel fiber on the tensile strength in splitting test
are evaluated, it is observed that they provide a significant average increase in relation
to the SCC0F as showed in Table 5. This increase of 54.8% for SCC0.5F and 37.1%
for SCC1F in comparison to the SCC0F is explained by the limitation of the concrete as
to the tensile strength. The effects of the analysis of the results of the flexural tensile
strength test prove that there is an increase in strength from the insertion of steel fibers.
This increase in stress and deflection varies with the growth in fiber contents in the
mixture.
Table 5. Flexural tensile strength results (fct,f) at 28 days (analyzing the stress of first crack
(rcr), first crack deflection (dcr), maximum strain (ru) and maximum deflection(du)).
Mixture First Crack Peak (maximum)
rcr (MPa) CV (%) dcr (mm) CV (%) ru (MPa) CV (%) du (mm) CV (%)
SCC0F 3.7 9.9 1.47 5.1 3.7 9.9 1.47 5.1
SCC0.5F 3.9 8.0 1.41 3.14 3.9 8.0 1.41 3.14
SCC1F 4.1 9.9 1.76 8.64 4.9 10.6 2.4 6.0
Regardless of the ultimate loads achieved in tensile testing, steel fibers promoted
changes in concrete tensile behavior. The Fig. 5 shows the flexural tensile strength
versus deflection of specimens. The SCC0F concrete was classified as fragile material,
because it broke in the first crack (Fig. 5a). In SCC0.5F, the rupture was also related to
the SCC0F, but the material changed from fragile to apparently semi-fragile behavior,
since it is notorious the analysis of a short period where the fibers held for a while effort
(Fig. 5b). The softening branch curve was not obtained completely because the test was
finished early not showing the ductile behavior. In SCC1F the ductile behavior was full
observed in the development of load versus deflection curves of concrete (Fig. 5c). The
fibers held the load for a period of time after the maximum breaking peak, demon-
strating that they have important deflection characteristics, as they help and increase the
concrete to combat the tensile effects. When analyzing concretes with fibers, the curves
show that at the beginning of the test, the prisms exhibit a linear elastic behavior, that
is, the moment when the concrete alone can withstand the tensile effects. After the
appearance of the first microcracks, the behavior becomes nonlinear before reaching
the maximum rupture, known as the hardening process. After the peak, there is non-
linearity in the concrete curve, but with elastic recovery parallel to the initial stiffness
due to the loss of fiber adhesion in the concrete matrix. After this phase, the mixtures
showed a progressive tension softening due to the pullout of the fibers and concrete
fracture. It is also noteworthy that the tension of the first crack was higher in concrete
with greater fibers insertion. The similar comportment was also observed by [26].
The average values of the physical index tests are presented in Table 6. It is
observed that the concretes with fibers addition showed a slight tendency of decrease of
voids and absorption, when compared to the SCC0F. The specific mass results only
corroborated what was noticeable during the execution of the tests, that is, the work-
ability of self-compacting concrete is reduced by incorporating steel fibers.
3.3 Indications for Use

The three mixtures produced have different characteristics in fresh and hardened state
that will possibility distinct application. SCC0F can be used for most current appli-
cations because was presented the best results for the fresh state as ability to flow
without losing its uniformity good workability and free flow fill without the need for
vibration or compaction. Pillars, walls, beams, slabs, slender elements and foundation
elements can be recommended without problems. The SCC0.5F is most useful for
structures that require a short horizontal spreading distance on structural elements with
maximum reinforcement spacing of 80 mm and where vibrators are difficult to use and
to prevent shrinkage cracks. The SCC1F is suitable for structures with low steel armor
reinforcement, as slabs, floors, hard concrete floors, when concrete needs to be placed
for a short distances.
Flexural Tensile (MPa)

4
3
Prism 1
2 Prism 2
1 Prism 3
0
0 0.5 1 1.5 2
Displacement (mm)
a. SCC0F.
5
3
Prism 1
2
Prism 2
1 Prism 3
0
0 0.5 1 1.5 2
Displacement (mm)
b. SCC0.5F.
6
5 Prism 1
4 Prism 2
Prism 3
3
2
1
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Displacement (mm)
c. SCC1F.
Fig. 5. Flexural tensile strength versus deflection of concrete.

Table 6. Physical properties of concrete specimens.

Physical indexes SCC0F SCC0.5F SCC1F
Amount CV Amount CV Amount CV
(%) (%) (%)
Void Indices (%) 7.5 2.89 5.96 5.86 7.13 2.64
Absorption (%) 3.27 2.94 2.47 6.21 3.01 2.25
Dry Sample Specific Mass 2.30 0.15 2.41 0.37 2.37 0.48
(g/cm2)
Saturated Sample Specific 2.37 0.15 2.47 0.22 2.44 0.53
Mass (g/cm2)
Real Specific Mass (g/cm2) 2.48 0.24 2.56 0.05 2.55 0.66
4 Conclusions
The three concrete products were considered self-compacting in at least one test. The
incorporation of metallic fibers in the concrete was satisfactory, emphasizing that it
interferes with the properties in the fresh and hardened state.
• There was a decrease in the spread and flow of the concrete as well as disturbances
in the packing of aggregates as the fiber contents increased.
• Better performance in tensile strength as the fiber contents increased.
• The increase in fiber contents up to 1% improved ductile behavior.
It is possible to produce self-compacting fiber reinforced concrete with the content
up to 1% of steel fibers using fine natural sand, even without adjusting the proportion of
the granular materials of the reference mixture and the use of very fine natural sand
with the high density coarse aggregate.
References
1. Okamura, H., Ouchi, M.: Self-compacting concrete. J. Adv. Concr. Technol. 1(1), 5–15
(2003)
2. Barros, J.A.O., Figueiras, J.A.: Experimental behaviour of fiber concrete slabs on soil.
J. Mech. Cohesive-frictional Mater. 1988(3), 277–290 (1998)
classification and prediction. Cem. Concr. Res. 40, 226–234 (2010)
4. Ghazi, F.K., Rand, S.A.J.: New method for proportioning self-consolidating concrete based
on compressive strength requirements. ACI Mater. J. 107, 490–497 (2010)
5. Vijaykumar, H., Shamu, S.: A critical study on the influence of steel fiber on performance of
fresh and hard self-compacting concrete. J. Struct. Eng. (India) 42, 237–245 (2015)
6. Grunewald, S.: Performance based desing of self-compacting steel fober reinforced concrete.
Ph.D Thesis, Delft University of Technology, Holland, The Netherlands (2004)
7. Larrard, F.D.: Concrete Mixture Proportioning. A Scientific Approach. Modern Concrete
Technology Series, vol. 9, pp. 77–169. Spon Press, London (1999)
8. Yu, A.B., Standish, N., McLean, A.: Porosity calculation of binary mixtures of non-spherical
particles. J. Am. Ceram. 76, 2813–2816 (1993)
9. Góis, F.A.P.: Avaliação experimental do comportamento do concreto fluido reforçado com
fibras de aço: influência do fator de forma e da fração volumétrica das fibras nas
propriedades mecânicas do concreto. 2010 Dissertação de mestrado, p. 156. Universidade
Federal de Alagoas, Maceió (2010)
10. Naaman, A.E.: Sifcon: tailored properties for structural performance. In: High Performance
Fiber Reinforced Cement –Composites – HPFRCC, Department of Civil Engineering,
University of Michigan, Ann Arbor, USA, pp. 18–36 (1992)
11. Melo Filho, J.A.: Desenvolvimento e Caracterização de Laminados Cimentícios Reforçados
com Fibras Longas de Sisal. Dissertação de Mestrado. PEC/COPPE/UFRJ, Rio de Janeiro,
p. 144 (2005)
12. Formagini, S.: Dosagem científica e caracterização mecânica de concretos de altíssimo
desempenho. Tese de doutorado, COPPE/Universidade Federal do Rio de Janeiro, Rio de
Janeiro, p. 284 (2005)
13. ABNT NBR NM 248:203. Agregados – Determinação da composição granulométrica. RJ
14. ABNT NBR NM 45:2006. Agregados – Determinação da massa unitária e volume de vazios.
RJ
15. ABNT NBR NM 52:2009 and 53:2009. Agregado miúdo e graúdo – Determinação da massa
específica e massa específica aparente. Rio de Janeiro
16. ABNT NBR 7809:2006. Agregado graúdo – Determinação do índice de forma pelo método
do paquímetro. Rio de Janeiro
17. ABNT NBR NM 49:2001. Agregado miúdo – Determinação de impurezas orgânicas. RJ
18. ABNT NBR NM 46:2003. Agregado miúdo – Determinação do material fino que passa
através da peneira 75 um, por lavagem. Rio de Janeiro
19. ABNT NBR 15823:2017. Concreto Autoadensável. Rio de Janeiro
20. ABNT NBR 5739:2018. Concreto – Ensaio de compressão de corpos de prova cilíndricos.
RJ
21. ABNT NBR 12142:2010. Concreto – Determinação da resistência à tração na flexão de
corpos de prova prismáticos. Rio de Janeiro
22. ABNT NBR 7222:2011. Concreto – Determinação da resistência à tração por compressão
diametral de corpos de prova cilíndricos. Rio de Janeiro
23. ABNT NBR 8522:2017. Concreto – Determinação dos módulos estáticos de elasticidade e
de deformação à compressão. Rio de Janeiro
24. ABNT NBR 9778:2005. Argamassa e concreto endurecidos – Determinação da absorção de
água, índices de vazios e massa específica. Rio de Janeiro
25. Figueiredo, A.D.: Concreto com fibras de aço. Boletim Técnico da Escola Politécnica da
USP. Departamento de Engenharia de Construção Civil e Urbana. BT/PCC/260. São Paulo
(2000)
26. Cotterell, B., Mai, Y.W.: Fracture Mechanics of Cementious Materials.
London/Glasgow/Weinheim/New York/Melbourne/Madras. Academic & Professional
(1996)
Aligned Interlayer Fibre Reinforcement
for Digital Fabrication with Concrete
Lukas Gebhard(&), Jaime Mata-Falcón, Tomislav Markić,

and Walter Kaufmann
Institute of Structural Engineering, ETH Zurich, Zurich, Switzerland

gebhard@ibk.baug.ethz.ch
Abstract. This paper presents a novel concept of fibre reinforcement placement

for digital fabrication with concrete, particularly suitable for 3D concrete
printing, which aims at overcoming the limitations of adding the fibres to the
concrete mix. In this process, the fibres are placed in-between layers, which
allows aligning the fibres and grading their content according to the structural
needs. The mechanical performance of this concept is investigated through a
series of four-point bending tests in which the influence of different fibre con-
tents and distributions, fibre types, sample geometries and time intervals
between consecutive printing layers is studied. The crack kinematics were
recorded using digital image correlation. Based on the crack kinematics, a
refined inverse analysis is proposed to predict the direct tension behaviour. The
results show that a deformation hardening structural behaviour might be reached
at a relatively low fibre content (0.7 vol%) when the fibres are aligned. The
performance is further increased when grading the fibre distribution over the
height but keeping the overall fibre content constant. With an increasing number
of fibres in-between the layers, however, a delamination failure caused by a
concentration of anchorage stresses is observed, which limits the peak load and
leads to severe softening behaviour.
Keywords: Fibre reinforced concrete Fibre alignment Digital fabrication

Digital image correlation Inverse analysis
1 Introduction
Digital fabrication with concrete (DFC) brings many new possibilities to the conceptual
design, dimensioning, detailing, and production of concrete structures, with great
potential to lead to a major impact on the construction market. However, in spite of
huge efforts of industry and academia in the recent years – in particular in the field of
3D concrete printing (3DCP) – most digital technologies still encounter difficulties in
satisfying structural integrity requirements. One reason for this unsatisfactory situation
is that most of these novel technologies do not allow adding reinforcement during the
production [1]. Reinforcement is essential to ensure structural safety and ductile
behaviour of concrete structures. Fibre reinforcement is one of the most promising
reinforcing strategies for DFC due to its adaptability to complex geometries [2].
However, the addition of fibres to the concrete mix affects (i) concrete rheology (a key
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_8
88 L. Gebhard et al.
aspect of DFC), which in turn challenges the development of robust processes; and
(ii) concrete pumpability (usually only very short, expensive fibres can be used). First
steps in the application of fibre reinforced concrete (FRC) in digital fabrication have
already been made [3–5] showing promising results, but only short fibres have been
used so far.
This study proposes a novel reinforcement concept for layered concrete processes,
where the fibres are placed in-between consecutive layers (Fig. 1). This approach
overcomes the processing constraints mentioned above by allowing the use of large and
inexpensive fibres and the alignment and grading of the fibres according to the
structural requirements. As a side effect, since the fibres can be positioned respecting a
concrete cover, durability and aesthetical concerns of conventional steel FRC, where
part of the fibres are at the surface, are eliminated.
Fig. 1. Schematic sketch of the reinforcement concept: (a) concrete layers printed by a nozzle
with continuous placement of fibres; (b) resultant layered concrete with aligned fibres in-between
the layers.
While the mechanical behaviour of FRC has been studied for decades [6], its
application is still limited to secondary structural elements such as tunnel linings or
foundation slabs. The combination of fibre reinforcement and digital fabrication could
lead to a win-win situation, in which the structural performance of digitally fabricated
concrete elements is enhanced, and the application of fibre reinforcement is extended.
In this paper, the structural behaviour of layered concrete elements reinforced with
aligned interlayer fibres is studied by means of an experimental campaign consisting of
four-point bending tests. The mechanical performance in direct tension is estimated
based on a new inverse analysis based on detailed results of crack kinematics measured
with digital image correlation.
2 Experimental Campaign
2.1 Specimens, Materials and Casting

The properties of the steel fibres used in this project are given in Table 1 and Fig. 2.
The 3D and 5D fibres were added as interlayer fibre reinforcement. The OL fibres were
mixed in the concrete matrix of some samples as in conventional FRC.
Table 2 shows the composition of the mortar used for this study. This mix was
developed for digital fabrication processes. The mechanical properties of the mortar
Aligned Interlayer Fibre Reinforcement for Digital Fabrication with Concrete 89
was characterised at 14 days. The average compressive strength measured for standard
cylindrical samples was 79.9 MPa (nine samples); the average tensile strength tested by
double punch tests also on cylinders (D = 150 mm; H = 150 mm) was 3.2 MPa
(14 samples).
Fig. 2. Various types of fibres used in this study: (a) 3D fibres; (b) 5D fibres; (c) OL fibres.
Two different sample geometries were produced (Fig. 4). The smaller geometry has
a length (L) of 300 mm, a width (W) of 100 mm and is 40 mm high (H) and contains
three layers of fibres (four layers of mortar, each 10 mm thick). The samples produced
with this geometry are referred to as small geometry series (SG series). The larger
geometry was adapted from the standard three-point bending test in the fib Model Code
2010 [10], measures 800 mm 150 mm 150 mm (L W H) and contains nine
layers of fibres (ten layers of mortar, each 15 mm thick). The samples produced with
this geometry are referred to as larger geometry series (LG series).
The samples were cast layer by layer. After each mortar layer, the fibres were
placed manually on top of the mortar before the next layer of mortar was cast. Two
fibre contents (0.7 and 1.4% by volume) were investigated. The total amount of fibres
was calculated with respect to the entire sample volume, and afterwards, the weight of
the fibres was divided by the number of layers to get the fibre amount per layer. For
some samples, 0.2% of OL fibres were added to the mortar to reinforce the matrix
itself.
In the SG series, the fibres were placed in different configurations (Fig. 3). The
ordered (O) configuration refers to the fibres being placed perfectly ordered on top of
the mortar layer. Hooked (H) refers to the end hooks of the fibres being pushed into the
mortar matrix during placing, and diagonal (D) refers to a diagonal staggering of the
end hooks position. In the aligned (A) configuration, the fibres were dropped on top of
the layer mainly aligned in the direction of occurring tensile stresses, and random in-
plane (R2D) refers to the fibres being randomly dropped on top of the mortar layer with
no particular alignment. For the LG series, only the aligned configuration was used. For
one of the LG samples, the fibres were graded over the height of the sample with the
lower three layers containing a fibre content equal to 1.4%, the inner three layers an
equivalent of 0.7% and the top three layers containing no fibres. This grading results in
an overall fibre content of 0.7% and is referred to as graded alignment (GA).
Table 1. Properties of the steel fibres [7–9].

Name End Diameter Length lf Slenderness Tensile strength E-Modulus
hook df [mm] [mm] lf/df [−] fft,u [MPa] Ef [GPa]
3D fibres 3D 0.9 60 65 1180 200
5D fibres 5D 0.9 60 65 2300 200
OL fibres − 0.15 6 40 3000 200
Table 2. Composition of 10 L of mortar.

Aggregates Cement CEM Microslilica Water Superplasticizer Sika
(0–4 mm) [kg] Normo5 [kg] [kg] [kg] VC3082 [kg]
14.20 5.91 0.51 2.37 0.09
For the time passing between the deposition of consecutive layers, the term ‘in-
terlayer time’ is used. This interlayer time was varied between 3, 6 and 10 min. After
production, the samples were stored in a climate chamber at 95% humidity and 20 °C
until testing. Table 3 shows an overview of all tested samples.
Fig. 3. Fibre configurations in top and side view of a layer.
2.2 Test Setup and Protocol

The test setup is shown in Fig. 4 for the two tested geometries. The four-point bending
configuration allows the formation of multiple cracks without any influence of shear in
the constant bending moment zone. To ensure a bending failure, an external shear
reinforcement was provided outside the region with a constant bending moment in all
LG samples. Loading was applied using displacement control at a rate of 0.005–
0.01 mm/s for SG and 0.02 mm/s for LG. The surface of the samples was tracked using
a 3-dimensional digital image correlation system based on the commercial software
VIC-3D [11], which provided full-field measurement of the displacements and allowed
measuring precisely the crack kinematics [12, 13].
Table 3. Overview of the test specimens (SG: small geometry; LG: larger geometry; tIL: time
between the production of layers; qfm: concentration of OL-fibres in the mortar; Age: age at
testing).
Name Size Fibres Config qf [%] qfm [%] tIL [min] Age ½d
SG-3D-O-0.7-10 SG 3D O 0.7 − 10 7
SG-3D-OH-0.7-10 SG 3D OH 0.7 − 10 9
SG-3D-A-0.7-10 SG 3D A 0.7 − 10 9
SG-3D-A-1.4-10 SG 3D A 1.4 − 10 9
SG-3D-R2D-0.7-10 SG 3D R2D 0.7 − 10 9
SG-5D-O-0.7-10 SG 5D O 0.7 − 10 7
SG-5D-OH-0.7-10 SG 5D OH 0.7 − 10 7
SG-5D-A-0.7-10 SG 5D A 0.7 − 10 7
SG-5D-A-1.4-10 SG 5D A 1.4 − 10 7
SG-5D-R2D-0.7-10 SG 5D R2D 0.7 − 10 7
SG-3D-OHD-0.7-10 SG 3D OHD 0.7 − 10 7
SG-3D-A-0.7-3 SG 3D A 0.7 − 3 7
SG-3D-A-1.4-3 SG 3D A 1.4 − 3 7
SG-3D-A-0.7-3-FC SG 3D A 0.7 0.2 3 7
SG-3D-A-1.4-3-FC SG 3D A 1.4 0.2 3 7
SG-5D-OHD-0.7-10 SG 5D OHD 0.7 − 10 7
SG-5D-A-0.7-3 SG 5D A 0.7 − 3 7
SG-5D-A-1.4-3 SG 5D A 1.4 − 3 7
SG-5D-A-0.7-3-FC SG 5D A 0.7 0.2 3 7
SG-5D-A-1.4-3-FC SG 5D A 1.4 0.2 3 7
LG-3D-A-0.7-6 LG 3D A 0.7 − 6 25
LG-3D-A-1.4-6 LG 3D A 1.4 − 6 25
LG-3D-GA-0.7-6 LG 3D GA 0.7 − 6 25
LG-5D-A-0.7-6 LG 5D A 0.7 − 6 21
LG-5D-GA-0.7-6 LG 5D GA 0.7 − 6 21
Fig. 4. Test setup: (a) SG series; (b) LG series (dimensions in [mm]).

3 Results
3.1 Load-Deformation Behaviour of Small Geometry Series
Figure 5 shows the load-deformation behaviour of the SG series. All samples, except
for 3D fibres with the R2D configuration (grey), exhibit a deformation hardening
behaviour after the formation of the first bending cracks at around 0.2 kNm. Ordering
the fibres on the mortar layer increases the strength (light green). Embedding the fibre
hooks in the matrix (purple) and ordering the hooks diagonally (dark green) further
increases the performance. The aligned configuration (yellow) shows a similar beha-
viour as the ordered samples with a slightly decreased cracked stiffness. Doubling the
fibre amount (yellow dashed) leads to a stiffer response after cracking as well as an
increase of the maximum load. While the increase in stiffness seems to be similar to the
increase in fibre concentration, the increase in peak load is not proportional to the
increase in the fibre dosage.
The observed structural behaviour suggests that the location of the fibre anchorage
plays an important role, and the main load transfer in the fibre is concentrated in the
area of the anchorage. The sudden softening after the peak load observed for all ordered
samples with 5D fibres hints that the anchorage stresses are higher for the 5D fibres
than for the 3D fibres, which show overall a more ductile softening. Doubling the fibre
amount does not double the maximum capacity. This indicates that the failure is not
only caused by fibre pull out but also by delamination of the layers. This delamination
could also be visually observed during the testing of the samples.
Fig. 5. Load-deformation behaviour of different fibre configurations for SG series with 10 min
interlayer time: (a) samples with 3D fibres (the sample SG-3D-O-0.7–10 (light green) was
measured with a different test setup, and only the maximum value is displayed); (b) samples with
5D fibres.
The influence of the interlayer time, as well as the presence of OL fibres in the
mortar matrix (OL fibres content of 0.2%), is shown in Fig. 6 for the aligned fibre
configuration. No significative influences are observed for a fibre content of 0.7%
(Fig. 6a-b), while for the double fibre content both, the reduction of the interlayer time
and the addition of OL fibres in the mix, enhance the structural performance slightly
(Fig. 6c-d). This indicates that delamination and hence, the strength of the matrix
govern the failure for high fibre contents.
Fig. 6. Influence of different times between layers and presence of OL fibres in the matrix on SG
series with aligned fibre configuration: (a) 3D fibres at 0.7%; (b) 5D fibres at 0.7%; (d) 3D fibres
at 1.4%; (d) 5D fibres at 1.4%.
3.2 Load-Deformation Behaviour of Larger Geometry Series

The results of the LG series (Fig. 7) are qualitatively consistent with the SG series. For
3D and 5D fibres, the behaviour at a fibre dosage of 0.7% is similar (yellow). After
cracking of the mortar matrix, the tension is transmitted to the fibres, and the load can
be further increased up to a peak after which softening occurs. Doubling the fibre
content again increases the peak load and cracked stiffness. While the stiffness is
increased proportionally to the increase of the fibre content, the strength is not doubled
(orange). The softening behaviour at large fibre contents is more severe than in the SG
series. Grading the samples over the height increases the cracked stiffness as well as the
peak load (blue). However, for the 5D fibres, strong softening right after the peak load
can be observed. The analysis of the cracking in the first layer shows that this softening
originates from delamination of the lowest layer.
Fig. 7. Series with larger geometry: (a) 3D fibres; (b) 5D fibres.

3.3 Example of Crack Kinematics

An example of the results of crack kinematics is given in Fig. 8 for sample SG-3D-A-
1.4-10. This test presented a total of three bending cracks inside the constant bending
region, and its behaviour is representative of the behaviour in the tests showing
deformation hardening. At the cracking moment (around 0.2 kNm) the cracks originate
and then open quasi-linearly with increasing bending moments. This quasi-linear
behaviour then transitions into a non-linear response up to the peak load as the failure
starts to localise in the weakest crack (Crack 1 in this case). After reaching the peak
load, this crack governs the response: the load decreases and the entire additional
deformation localises in that single crack while the other crack openings remain
roughly constant. The distribution of the crack opening over the specimen height is
linear, and the position of the neutral axis remains almost constant (Fig. 8b).
Fig. 8. Example of crack kinematics for bending cracks of specimen SG-3D-A-1.4–10: (a) crack
opening at the height of the first layer of fibres; (b) crack opening over the height for several
measuring stages (dark grey: early in the experiment, light grey: late in the experiment).
4 Inverse Analysis
FRC is usually characterised by its behaviour in direct tension. Direct tension tests are,
however, quite difficult to perform. Hence, it is common practice to estimate the direct
tension behaviour by inverse analysis of bending experiments. Most inverse analysis
procedures, such as [10] and [14], are developed for three-point bending tests. Pro-
cedures for four-point bending experiments on strain-hardening materials are also
possible, but require using detailed instrumentation [15]. In this section, it is studied
how to perform such inverse analysis using refined crack-kinematic results extracted
from DIC measurements.
The inverse analysis takes the crack kinematics of the localising crack measured by
DIC, discussed in Sect. 3.3 (Fig. 9a), and derives the direct tension behaviour
(Fig. 9b). To this end, characteristic crack openings and slopes are read from the
experimental curve. By satisfying equilibrium at the cross-section, the direct tension
behaviour can be calculated. The resulting direct tension behaviour is then applied to
predict the bending behaviour of the sample (Fig. 9c).
Fig. 9. Schematic display of the inverse analysis.
For this analysis, the input parameters need to be visually determined from the
experimental curve of the bending moment-crack opening at the lowest fibre, Fig. 9a.
The parameters are the crack opening right after the cracking moment uA and the crack
opening where the quasi-linear hardening behaviour transitions to a non-linear beha-
viour uH. Next, the quasi-linear hardening slope between uA and uH and the quasi-linear
softening slope after the peak load need to be calculated from the experimental curve.
The direct tension behaviour of the fibres is idealised by means of the linear fibre
stress-crack opening phases represented in Fig. 9b: (1) linear activation phase up to
cracking (uA), i.e., even though the fibres are not active before cracking, a linear
activation phase is assumed; (2) linear hardening phase up to the crack opening uH; and
(3) linear softening at crack openings exceeding uH.
Based on the idealised direct tension behaviour and to simplify the procedure,
several phases are also assumed for the input bending moment-crack opening data
(Fig. 9a). In the first phase, all fibres are in the linear activation phase (1), until the
crack opening right after cracking (uA) is reached. Then all fibres are assumed to be in
the linear hardening phase (2). When the maximum crack opening in the lowest fibres
reaches a value of uH, the lower fibres start softening, while upper fibres are still
hardening (2 + 3). After the peak bending moment, all fibres experience linear soft-
ening behaviour (3). In the absence of refined measurements of the crack kinematics
and based on the observations of Sect. 3.3, the concrete behaviour should be consid-
ered as linear elastic, the variation of the crack opening over the height should be
assumed to be linear, and the compression zone height can be estimated as 12% of the
specimen’s depth.
The measured average slope of these phases in the experimental curve allows to
directly calculate the slope of the corresponding phase in the fibre stress-crack opening
diagram using simple equilibrium conditions at the cross-section (Fig. 10). For a more
detailed description of the inverse analysis, see [16].
In Fig. 11, the various steps of the inverse analysis and the back-calculation can be
seen for one of the samples. The behaviour in direct tension assumes that the mortar
cracks at around 3.5 MPa (*fct), then hardens up to uH after which it transitions into a
linear softening response. The assumptions can be verified by applying the direct tension
behaviour (Fig. 11b) to the three fibre layers of the used sample and calculating the
response of the section. The results (Fig. 11c) fit the experimental curve used for the
Fig. 10. Cross-section analysis and assumptions for the inverse analysis illustrated for aligned
fibre reinforcement with three interlayers.
Fig. 11. Steps of the inverse analysis for SG-3D-A-1.4-10 with uA = 0.03 mm and
uH = 0.19 mm: (a) Localising crack with the linear hardening and softening slope; (b) direct
tension behaviour; (c) back-calculation compared to experimentally observed bending cracks.
inverse analysis, as well as the other cracks in the specimen (up to the peak load, since
only one crack enters the softening range). This proves that the idealisation is appropriate.
Once the direct tension behaviour is known, the behaviour of the FRC in any
geometry in bending can be predicted. The direct tension behaviour obtained from
inverse analysis of the SG series with aligned fibre configuration was used to predict
the behaviour of the LG series, as shown in Fig. 12 for specimen LG-3D-A-1.4-6. Two
predictions were conducted based on the results of the SG samples with 10 and 3 min
of interlayer time, showing a very good agreement with the experimental results.
Fig. 12. Comparison of experimental results and behaviour predicted based on inverse analysis
for specimen LG-3D-A-1.4-6.
5 Conclusions
• A novel reinforcement strategy for layered digital fabrication with concrete is
presented. The strategy consists in placing fibres in-between concrete layers, which
could overcome pumpability and rheological constraints of conventional FRC.
Fibre performance can be significantly improved when the fibres are aligned in-
between the layers. Moreover, the amount of fibres can be adjusted for each layer,
which allows grading the fibre concentration to the structural needs. As a side effect,
durability and aesthetic concerns due to fibre corrosion in conventional FRC can be
avoided.
• Failure is caused by the delamination of layers, by fibre pull-out or by a combi-
nation of these two. Several strategies have been studied to improve the bond of the
fibres (addition of small fibres in the mortar and the reduction of the interlayer
time), but no significant improvement could be observed. Hence, to fully activate
high fibre contents and 5D fibres, the bond of the adjoining layers needs to be
enhanced (e.g. prestressing perpendicularly the layers).
• Severe softening, especially at the larger geometry, indicates that this reinforcement
strategy is not yet suitable for carrying the main structural loads and fulfilling
ductility requirements. The combination with other reinforcements (e.g. conven-
tional reinforcement or post-tensioning) could result in a promising reinforcement
strategy for digital fabrication, but the automation of the process is challenging.
• The proposed inverse analysis shows the possibility to derive the direct tension
behaviour of FRC by measuring the crack kinematics of the crack, where damage
localises. The results allow the prediction of the structural behaviour at larger
geometries.
Acknowledgements. The authors would like to thank the support of the Swiss National Science
Foundation, which partially funded this work within the National Centre for Competence in
Research in Digital Fabrication in Architecture (project number 51NF40_141853).
References
1. Asprone, D., Menna, C., Bos, F.P., Salet, T.A.M., et al.: Rethinking reinforcement for digital
fabrication with concrete. Cem. Concr. Res. 112, 111–121 (2018)
2. Wangler, T., Roussel, N., Bos, F.P., Salet, T.A.M., et al.: Digital Concrete: A Review. Cem.
Concr. Res. 123, 105780 (2019)
3. Hambach, M., Volkmer, D.: Properties of 3D-printed fiber-reinforced Portland cement paste.
Cem. Concr. Compos. 79, 62–70 (2017)
4. Panda, B., Paul, S.C., Tan, M.J.: Anisotropic mechanical performance of 3D printed fiber
reinforced sustainable construction material. Mater. Lett. 209, 146–149 (2017)
5. Bos, F.P., Bosco, E., Salet, T.A.M.: Ductility of 3D printed concrete reinforced with short
straight steel fibers. Virtual Phys. Prototyp. 14, 160–174 (2019)
6. Zollo, R.F.: Fiber-reinforced concrete: an overview after 30 years of development. Cem.
Concr. Compos. 19, 107–122 (1997)
7. Bekaert, Dramix® 3D technical documents. Bekaert n.d
8. Bekaert, Dramix® 5D technical documents. Bekaert n.d

9. Bekaert, Dramix®OL CE documents. Bekaert n.d
10. fib Model Code 2010, fib Model Code for Concrete Structures 2010 (2013)
11. Correlated Solutions, VIC 3D Software, Reference Manual v7 (2014)
12. Haefliger, S., Mata-Falcón, J., Kaufmann, W. : Application of distributed optical measure-
ments to structural concrete experiments, In : SMAR 2017 Proceedings, p. 159. ETH Zurich
(2017)
13. Mata-Falcón, J., Haefliger, S., Lee, M., Galkovski, T., et al. : Combined application of
distributed fibre optical and digital image correlation measurements to structural concrete
experiments. Eng. Struct. J. (2020, submitted)
14. Amin, A., Foster, S.J., Muttoni, A.: Derivation of the r-w relationship for SFRC from prism
bending tests. Struct. Concr. 16, 93–105 (2015)
15. López, J.Á., Serna, P., Navarro-Gregori, J., Camacho, E.: An inverse analysis method based
on deflection to curvature transformation to determine the tensile properties of UHPFRC.
Mater. Struct. 48, 3703–3718 (2015)
16. Gebhard, L., Mata-Falcón, J., Markić, T., Kaufmann, W.: New opportunities for fiber
reinforced concrete in digital fabrication. Master’s Thesis, ETH Zurich, Switzerland (2018)
Mixture Proportioning of Steel Fibre
Reinforced Self-compacting Concrete Based
on the Compressible Packaging Method:
Comparison with ACI 237R-07 and RILEM
TC 174-SCC Recommendations
M. G. Cardoso(&), R. M. Lameiras, T. T Oliveira, F. B. Santana,

and V. M. S. Capuzzo
Department of Civil and Environmental Engineering (ENC), University of

Brasília (UnB), Brasília, Brazil
matheus-ssdo@hotmail.com
Abstract. This work evaluates the production of self-compacting concrete

reinforced with steel fiber (SFRSCC). The mix designs concretes were obtained
with the compressible packaging method (CPM), the concrete properties found
were compared to parameters established for self-compacting concrete in ACI
237R-07 and RILEM TC 174-SCC. Using BetonLab Pro 3 software, 39 con-
crete mix design were produced. In this process, three classes of compressive
strength were tested 20 MPa, 30 MPa and 40 MPa, each one were combined
with 3 different types of steel fibers, in contents of 0.0%, 0.5%, 0.75%, 1.0% and
1.5%. Fourteen types of concrete were produced from those mixes, the concretes
values of compressive strength where 20 MPa and 40 MPa, with three different
types of steel fibers, in the contents of 0.5% and 1.0%. In addition, a reference
concrete was produced for each compressive strength class, without fibers, a
total of 14 mixtures. All the produced concrete presented the rheological
characteristics of a self-compacting concrete, showing the effectiveness of CPM.
When comparing all the compositions obtained and the limits cited by ACI and
RILEM, the 20 MPa and 30 MPa concretes presented most of the values within
the limits proposed by the standards, unlike the 40 MPa concrete compositions
that in many parameters did not follow the proportions described in the concrete
self-compacting standards. The results showed that, for the materials and con-
ditions used in this research, the values proposed by the standards would serve
as a reference for the experimental dosage of SFRSCC only for concretes with
compressive strength range of 20 MPa and 30 MPa. In addition, a reference
concrete was produced for each compressive strength class, without fibers, a
total of 14 mixtures. All the produced concrete presented the rheological
characteristics of a self-compacting concrete, showing the effectiveness of CPM.
When comparing all the compositions obtained and the limits cited by ACI and
RILEM, the 20 MPa and 30 MPa concretes presented most of the values within
the limits proposed by the standards, unlike the 40 MPa concrete compositions
that in many parameters did not follow the proportions described in the concrete
self-compacting standards. The results showed that, for the materials and con-
ditions used in this research, the values proposed by the standards would serve
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_9
100 M. G. Cardoso et al.
as a reference for the experimental dosage of SFRSCC only for concretes with
compressive strength range of 20 MPa and 30 MPa.
Keywords: Steel fiber reinforced Self-compacting concrete Compressible

packing model Fiber reinforced concrete Mixture proportioning
1 Introduction
The Steel Fibre Reinforced Self-Compacting Concrete (SFRSCC) is an innovation that

combines the benefits of Fiber Reinforced Concrete such as increased toughness, and
the ability to withstand post-cracking residual shrinkage stresses. Also, it has the
characteristics of Self-compacting Concrete, such as workability, ability to fill and
transpose restrictions without segregating, eliminating the use of vibration. As for the
application, research already shows that SFRSCC can be used in tunnel segments,
precast roof elements, beams, sandwich panels, sheet piles, floors and slabs [1, 2, 3, 4,
5, 6, 7, 8].
There are several ways to define the SFRSCC composition, from empirical
methods, such as those proposed by Ferrara and Sah [9], Okamura Method, EFNARC
Method, Gomes Method, Tutikian Method [20, 4], to more precise methods, with a
more solid scientific approach, such as the Compressible Packaging Method
(CPM) proposed by De Larrard and used by several researchers, among themby
Grunewald, Silva and Marangon [1, 2, 19, 21].
The CPM is based on a theory that solves the question of packaging dry mixtures in
all components used in concrete dosage. It relates the constituent materials with the
desired characteristics of the concrete, through equations. This method also has a great
advantage over other methods, which is the possibility of being programmable. Using
this possibility, Betonlab Pro 3 was developed, a computational tool developed based
on the Compressible Packaging Model (CPM). This software is capable of optimizing
granular mixtures to achieve the desired concrete properties, such as strength and
workability [22].
In addition, there are also parameters for self-compacting concretes proposed by
ACI 237R-07: 2007 and RILEM TC 174-SCC: 2000 [11, 12] wherein some indicators
are established regarding the constituent materials to obtain the dosage of self-
compacting concrete; such as the amount of fines, consumption of fine and coarse
aggregates, and volume of paste. However, these values proposed by these standards
were not defined for self-compacting concrete reinforced with fibers.
The main objective of this research is to verify if the limits proposed by ACI 237R-
07: 2007 and RILEM TC 174-SCC: 2000 [11, 12] are also acceptable for obtaining
Steel Fibre Reinforced Self-Compacting Concrete by an experimental mixture pro-
portioning methodology.
Mixture Proportioning of Steel Fibre Reinforced Self-compacting Concrete 101
In this work, the CPM was used to define the composition of different SFRSCC with
fiber contents of 0.5% and 1.0% and the obtained dosages were validated experi-
mentally. Once validated, different compositions defined for SFRSCC with the CPM
were obtained with the limits proposed by the ACI 237R-07 and RILEM TC 174-SCC
[11, 12] for simple self-compacting concretes.
BetonLab Pro 3 software was used to obtain the mixtrure proportioning of con-
cretes divided into three compressive strength classes 20 MPa, 30 MPa and 40 MPa.
Wherein 3 types of steel fiber were adopted, varying the contents by 0.5%, 0.75%,
1.0% and 1.5%, in volume. Thus, 12 SFRSCC were dosed for each resistance, plus a
reference concrete for each resistance, totalling 39 concrete mixes. The characteristics
of the fibers used in the study are described in . The fibers were considered in the
mixture as a spherical grain calculating an equivalent diameter, according to the
methodology proposed by Yu et al. [22] and also adopted by Grunewald [1].
Table 1. Physical characteristics of the steel fibers used in the research.

Fiber type Diameter (mm) Length (mm) Aspect ratio
ST-33/44 0.75 33 44
ST-33/60 0.55 33 60
ST-50/67 0.75 50 67
The materials used in this research were: Portland Cement composed with car-
bonate material CP II 32 F, average river sand with a maximum diameter of 2.36 mm,
coarse aggregate with a maximum diameter of 9.5 mm, limestone filler and porlicar-
boxylate superplasticizer. The dosage of SFRSCC adopting the CPM requires a
complete characterization of the materials, in which the minimum parameters to be
inserted in the Betonlab Pro 3 software, and the standards used to obtain them are
described in Table 2. The packing obtained described in Table 2 followed the proce-
dure performed by De Larrard [10], adapted by Silva [19]. The superplasticizer used in
the research was a Master Glenium 51 supplied by BASF. This superplasticizer con-
sists of a third generation additive, made from modified polycarboxylate ether.
For the dosage of concretes using the software Betonlab Pro 3, in addition to the
characterization of the materials used, the parameters described in Table 3 were
adopted, which are reference values established by De Larrard [10] that a concrete must
respect to be self-compacting.
When dosing using the Betonlab Pro 3 software, the following procedure was used:
first, the granular skeleton was optimized in order to achieve the proportion of small
and coarse aggregates that presented the greatest possible compactness. Later, with the
parameters described in Table 3 fixed in the software, the goal characteristics were
obtained by minimizing cement consumption.
Table 2. Parameters used for SFRSCC dosing adopting CPM.
Materials Data Results

Density (g/cm³), [23] 2.6
Absorption (%), [23] 2,00
Sand
Particle size distribution,[16]
Packing, [2] 0.632
Density (g/cm³), [14] 2.75
Absorption (%), [14] 0.5
Coarse agreagte
Particle size distribution, [16]
Packing, [2] 0.608
Density (g/cm³), [17] 2.97
Particle size distribution, a laser
Cement
Packing, [2] 0.523
Compressive strength of Portland cement 33.28
Density (g/cm³), [17] 2.7
Limestone Filer Particle size distribution, laser
Packing, [2] 0.523
Table 3. Parameters for a concrete to be self-compacting by De Larrard [10].

Parameters Limits
Yield Stress (MPa) <400
Plastic Viscosity (Pas) <200
Influence of fines - k0f >3.3
Influence of coarse k0gg aggregates <1.4
In order to validate the dosages and check if the concretes were really self-compact
concretes, they were produced and tests were carried out according to NBR 15823:
2017 [12]. Slump flow test, T500, L box and V Funnel tests were performed.
After the validation of the obtained compositions, they were compared with the
limits of ACI 237R-07: 2007 and RILEM TC 174-SCC: 2000 [11, 12], for self-
compacting concretes, in order to verify if these indicators can be used as parameters to
obtain experimentally the necessary mix compositions of SFRSCC.

3.1 SFRSCC Validation
For the validation of the adopted dosage method, some mix concrete were selected to
be produced. The results are described in Table 4 for the self- compacting concretes of
20 MPa strength class and in Table 5 for the self-compacting concretes of 40 MPa
strength class.
Table 4. Tests to check the self-compacting properties for the SFRSCC of 20 MPa strength
class.
Fiber type ST-33/60 ST-50/67 ST-33/44 REF20
Fiber content (%) 0.50 1.00 0.50 1.00 0.50 1.00 0.00
V funnel (s) 9.16 10.14 7.23 6.94 9.29 10.15 8.86
T 500 (s) 2.81 4.06 2.17 2.86 2.18 3.19 2.41
Slump flow test (mm) 625 630 625 635 600 605 635
L box (H2 / H1) 0.89 0.93 0.89 0.91 0.85 0.85 0.90
For SFRSCC with characteristic compressive strength equal to 20 MPa, as can be

seen in Table 4, all concretes reached characteristics to be considered self-compacting,
achieving high fluidity and good spreading, without segregations being observed.
The worst results observed were obtained for the ST-33/44 fiber, in which less
fluidity and less spreading was observed. The behavior of the SFRSCC produced with
the ST-33/60 fiber and with the ST-50/67 fiber was very similar, and the observed
small variations may be due only to the adjustment of superplasticizer.
Table 5. Tests to check the self-compacting properties for the SFRSCC of 40 MPa strength
class.
Fiber type ST-33/60 ST-50/67 ST-33/44 REF40
Fiber content (%) 0.50 1.00 0.50 1.00 0.50 1.00 0.00
V funnel (s) 10.86 13.18 9.96 8.84 6.94 8.82 9.09
T 500 (s) 3.56 4.30 3.86 2.97 3.84 3.85 2.74
Slump flow test (mm) 675 665 645 625 620 618 650
L box (H2 / H1) 0.89 0.88 0.89 0.82 0.92 0.87 0.82
In Table 5 it is shown that, for all concrete compositions, among the SFRSCC of
40 MPa strength class, the increase in fiber content resulted in a decrease in the fluidity
of the concrete. As a consequence, less spreading in concrete with a fiber content equal
to 1% was observed, compared to concrete with a fiber content equal to 0.5%. This
shows that the increase of fibers generated a disturbance in the particles and impaired
the rheological behavior of the concrete.
3.2 Comparison with ACI 237R-07 and RILEM TC 174-SCC

Parameters
It was observed that, for all the concretes of the 20 MPa strength class, the absolute
volume of coarse aggregate were all within the range cited by the ACI 237-07.
However, when compared with the RILEM TC174, some mixtures were outside the
limits recommended by the standard. The same behavior was observed for concretes
with compressive strength equal to 30 MPa and 40 MPa. The results are shown in
Fig. 1.
Fig. 1. Coarse aggregate volume in SFRSCC the compressive strength classes equal to (a),
20 MPa (b), 30 MPa (c) 40 MPa.
The paste fraction for the concrete of the compressive strength classes equal to
20 MPa and 30 MPa were within the limits established by ACI 237R-07 and
RILEM TC 174-SCC. For the mixtures concrete of strength class equal to 40 MPa, the
paste fraction was higher than the limit indicated by ACI 237R-07 AND RILEM TC
174-SCC in all concrete mixes, shown in Fig. 2.
Fig. 2. Paste volume in SFRSCC the compressive strength classes equal to (a), 20 MPa (b),
30 MPa (c) 40 MPa.
The mortar fraction was within the range established by ACI 237R-07 (2007) in
concretes of all classes of compressive strengths for all types and contents of fibers, as
shown in the Fig. 3. The concrete mixtures of 40 MPa, even with the high consumption
of cement, presented the fraction of mortar within the limits predicted by the ACI. This
can be explained because this work didn’t used any other type of supplementary
cementitious materials, leading to a higher consumption of cement to attain the required
strength class that, consequently, led to a reduction in the consumption of limestone
filer.
Fig. 3. Mortar fraction in SFRSCC the compressive strength classes equal to (a), 20 MPa (b),
30 MPa (c) 40 MPa.
The cement consumption was much higher than the maximum indicated, according
to the parameters in the concrete of the compressive strength classes equal to 40 MPa
shown in Fig. 4. On the other hand, the consumption of cement in concrete mixtures of
compressive strength equal to 20 MPa and 30 MPa were within the ranges provided by
the ACI. ACI 237R-07 (2007) for self-compacting concretes.
The water-cement ratio, shown in Fig. 5, was the one predicted by the ACI in the
concretes of all strength classes, for all types of fibers used. It is noteworthy that water
consumption was higher than expected in all analyzed concrete mixtures.
As expected, the ST-33/60 fiber, having the smallest size and smallest diameter,
was the one that least increased the consumption of cement and less afftected the
compactness when compared to ST-50/67 and ST-33/44 fibers. The increase in fiber
content resulted in a higher consumption of cement in concretes for all types of fibers.
Fig. 4. Cement consumption in SFRSCC the compressive strength classes equal to (a), 20 MPa
(b), 30 MPa (c) 40 MPa.
It was observed that the parameters provided by ACI and RILEM provided good
references for the experimental dosage of the SFRSCC mixture composition for steel
fiber reinforced self-compacting concretes with a compressive strength class of 20 MPa
and 30 MPa. This shows that they can be used for self-compacting concrete dosages
adopting the materials adopted in the research, for these strength classes. On the other
hand, the SFRSCC of the 40 MPa strength class presented some parameters analyzed
outside the ranges provided by the standards, showing that they do not provide a
reliable reference for the dosages of this SFRSCC.
Fig. 5. Water /cement ratio in SFRSCC the compressive strength classes equal to (a), 20 MPa
(b), 30 MPa (c) 40 MPa.
4 Conclusions
• All SFRSCC mix compositions obtained using the Compressible Packaging
Method experimentally tested in this research presented self-compacting
characteristics;
• The mix compositions obtained for the SFRSCC of 20 MPa and 30 MPa strength
classes fit the parameters cited for by ACI 237R-07 [11] for SCC without fibers;
• The SFRSCCs of 20 MPa and 30 MPa strength class fit almost all parameters
proposed by RILEM TC 174-SCC: 2000 [12], except that they had a higher water
consumption than that specified by the standard;
• The results showed that the parameters of RILEM TC 174-SCC: 2000 and ACI
237R-07 [11, 12] are a good reference to obtain the mixture proportioning of
SFRSCC of 20 MPa and 30 MPa strength classes, however it could not be used to
obtain the mix composition of SFRSCC of 40 MPa strength class.
References
1. Grünewald, S.: Performance-Based Design of Self-Compacting Fibre Reinforced Concrete.
Delft University of Technology, TU Delft (2004)
2. Grünewald, S., Walraven, J.C.: Parameter-study on the influence of steel fibers and coarse
aggregate content on the fresh properties of self-compacting concrete. Cem. Concr. Res. 31
(12), 1793–1798 (2001). https://www.sciencedirect.com/science/article/pii/
S0008884601005555. ISSN 0008-8846
3. Ferrara, L., Meda, A.: Relationships between fibre distribution, workability and the
mechanical properties of SFRC applied to precast roof elements. Mater. Struct. 39(4), 411–
420 (2006). ISSN 1359-5997
4. Barros, A.R., Gomes, P.C.C., Barboza, A.D.S.R.: Avaliação do comportamento de vigas de
concreto auto-adensável reforçado com fibras de aço (2009)
5. Lameiras, R., Barros, J., Valente, I.B., Azenha, M.: Development of sandwich panels
combining fibre reinforced concrete layers and fibre reinforced polymer connectors. Part I:
conception and pull-out tests. Compos. Struct. 105, 446–459 (2013). ISSN 0263-8223
6. Hedebratt, J., Silfwerbrand, J.: Full-scale test of a pile supported steel fibre concrete slab.
Mater. Struct. 47(4), 647–666 (2013). https://doi.org/10.1617/s11527-013-0086-5
7. Salehian, H., Barros, J.A.: Assessment of the performance of steel fibre reinforced self-
compacting concrete in elevated slabs. Cem. Concr. Compos. 55, 268–280 (2015). ISSN
0958-9465
8. Ahmad, H., et al.: Steel fibre reinforced self-compacting concrete (SFRSC) performance in
slab application: a review. In: AIP Conference Proceedings, p. 030024. IP Publishing (2016)
9. Ferrara, L., Park, Y.-D., Shah, S.P.: A method for mix-design of fiber-reinforced self-
compacting concrete. Cem. Concr. Res. 37(6), 957–971 (2007). ISSN 0008-8846
10. De Larrard, F.: Concrete Mixture Proportioning: A Scientific Approach. CRC Press, Boca
Raton (1999). ISBN 1482272059
11. ACI 237R-07 C. Concrete, Self-Consolidating. American Concrete Institute: Farmington
Hills, MI, USA, p. 30 (2007)
12. Rilem, T.C.: 162 TDF Test and Design Methods for Steel Fiber Reinforced concrete. Design
of Steel Fibre Reinforced Concrete using the S-W method: Principles and Applications
Mater. Struct. 35 262–278 (2000)
13. ASSOCIAÇÃO BRASILEIRA DE NORMAS TÉCNICAS. NBR 15823: Concreto
autoadensável. Rio de Janeiro (2017)
14. Associação Brasileira De Normas Técnicas. NBR NM 53: Agregado graúdo—Determinação
de massa específica, massa específica aparente e absorção de água. Rio de Janeiro (2009)
15. Associação Brasileira De Normas Técnicas. NBR 15823: Concreto autoadensável. Rio de
Janeiro (2017)
16. Associação Brasileira De Normas Técnicas: NBR NM 248. Agregados - Determinação da
composição granulométrica, Rio de Janeiro (2003)
17. Associação Brasileira De Normas Técnicas: NBR NM 23. Cimento Portland e outros
materiais em pó- Determinação da Massa Específica, Rio de Janeiro (2003)
18. SILVA, A. D. Dosagem de concreto pelos métodos de empacotamento compressível e
Aitcin-Faury modificado. : Dissertação de Mestrado. Universidade Federal do Rio de
Janeiro, Rio de Janeiro (2004)
19. EFNARC S. Guidelines for Self-Compacting Concrete. vol. 32, p. 34. Association House,
London (2002)
20. Marangon, E.: Desenvolvimento e caracterização de concretos auto-adensáveis reforçados

com fibras de aço. Dissertação de Mestrado em Ciências em Engenharia Civil, Rio de
Janeiro (2006)
21. Grabois, T.M.: Desenvolvimento e caracterização experimental de concretos leves
autoadensáveis reforçados com fibras de sisal e aço – Rio de Janeiro: UFRJ/COPPE (2012)
22. Yu, A.B., Standish, N., McLean, A.: Porosity calculation of binary mixtures of non-spherical
particles. J. Am. Ceram. Soc. 76(11), 2813–2816 (1993)
23. Associação Brasileira De Normas Técnicas. NBR NM 52: Agregado miúdo–Determinação
da massa específica e massa específica aparente. Rio de Janeiro (2009)
24. ASSOCIAÇÃO BRASILEIRA DE NORMAS TÉCNICAS. NBR 7215: Determinação da
resistência à compressão. Rio de Janeiro (1996). 8p.
Evaluation the Yield and Ultimate Strain
of FRC in Compression
Salam Wtaife1, Nakin Suksawang2(&), and Ahmed Alsabbagh3

1
University of Misan, Amarah, Iraq
2
Mechanical and Civil Engineering Department, College of Engineering
and Science, Florida Institute of Technology, Melbourne, USA
nsuksawang@fit.edu
3
Department of Civil Engineering, College of Engineering,
University of Babylon, Hillah, Iraq
Abstract. The design of reinforced concrete structures depends greatly on its

compressive behaviours. For normal concrete, most codes assume an ultimate
strain of concrete of 0.003. However, for concrete containing discrete fibres,
fibre reinforced concrete (FRC), the same assumption could not be made
because of its strain softening and hardening behaviours. This research used 250
point-database from the previous 24 research papers that evaluated the yield and
ultimate strain in compression for different types of fibre and volume fraction for
both FRC and fibre reinforced cement. In this research, a volume fraction of
fibre did not affect yield strength. The significant impact on yield strength was a
compressive strength. For fc0 < 69 MPa, the yield strength is 0.85fc0 . For fc0
69 MPa, the yield strength is roughly fc0 because the shape of the ascending
branch of the stress-strain relationship becomes more linear and steeper, and the
slope of the descending part also becomes steeper. The general shape of the
stress-strain relationship becomes more likely to be a triangle. The ultimate
strain at the extreme concrete compression fibre is taken to be 0.0035 for volume
fraction less than 1%, and 0.005 for volume fraction more than or equal 1%.
Keywords: Yield and ultimate strain Fibre reinforced concrete Compressive

strength
1 Introduction
Fibre reinforced concrete (FRC) has become increasingly popular as more and more
owners rely on it to improve the service life of their concrete structures. FRC provides
many advantages over normal concrete, particularly in minimizing early-age cracking,
increasing toughness, and enhancing post-crack load carrying capacity by improving
the energy absorption [1–4]. Overall, fibre enhances durability, fatigue life, and impact
resistance of concrete [5–8]. Compared with unreinforced concrete, FRC provides great
advantages to resist blast load, which help saving people’s lives and improving the
durability of concrete buildings [9]. Additionally, FRC also provides significant
improvements in creep characteristic of the concrete structure [10].
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_10
112 S. Wtaife et al.
One of the most important design criteria of FRC is its compressive behaviours.
Currently, there is no consensus of what the ultimate compressive strain should be.
Various codes, including AASHTO [11], ACI 318 [12], ACI 544 [13], Canadian
Highway [14], Eurocode 2 (EC2 2004), and others [15–19] provide different upper
limits of concrete compressive strength and ultimate strain of concrete compression.
Table 1 summarizes all the yield compressive strain equations design codes Chote and
Mobasher (2009) used a set of closed-form equations for the flexural design of FRC
that were based on simplified tensile and compressive constitutive response. The design
procedures of this model depend on the same concept of theoretical derivations in
RILEM TC 162-TDF.16 [22, 23] and ACI 318 (2014) with an ultimate compressive
pffiffiffiffi
strain ecu of 0.0035. The yield compressive strain is 0:00018 fc0 , so fc0 is the ultimate
uniaxial cylinder compressive strength.
Table 1. Database for Stress Block Models.

Reference eu
AASHTO (2012) and ACI (2014) 0.0030
NZS 3101 (1995) 0.0030
CSA A23.3 (2001) 0.0035
0
CEB-FIB Model Code (1990) 0.004 − 0.002
fc
100
0 4
Eurocode2 (2004) 90f c 0
0.0026 − 0.035 100 50 f c 90
ACI 544 0.035
2 Database of Experimental Data
For more accuracy in evaluating the impact of discrete fibres on comprehensive stress-
strain carver’s database were collected using experimental results obtained from var-
ious literature listed in Tables 3. Based on the database gathered using experimental
results for compressive stress-strain curves obtained from the various literature were
listed in Table 2. This Table mentions some details as Type, aspect ratio, the volume of
fraction and compressive strength of Fibre. A total of 20 kinds of research consisting of
over 250 data points using steel, PVA, polypropylene, and basalt fibres were collected
for both FRC and FRCC mixtures. The length and volume fractions of fibre ranged
from 0.20 in (5 mm) to 2.36 in (60 mm) and 0.1% to as high as 3.0%, respectively. The
concrete compressive strengths were between 2,900 psi (20 MPa) and 17,400 psi
(120 MPa).
Table 2. Database of experimental data.
Reference Type of fibre Sp Vf, % Strength (MPa)
Steel PVA Polypropylene Basalt <1 1 fc < 50 50 fc < 100 fc 100
1 Altun et al. * 80 * *
2007
2 Mansur et al. * 50 and 60 * * * *
1997
3 Yang 2011 * 35, 60 and 82 * * *
* 533, 800, 1000, * *
and 1267
5 Yoo et al. 2011 * 60 * * * *
6 Leeet al. 2015 * 45, 65 and 80 * * * *
7 Ou 2011 * 50, 60, 70 and * * *
100
8 Ezeldin et al. * 60, 75 and 100 * * *
1992
11 Fanella and * 47, 83, and 100 * *
Antoine 1985 * 100 and 250 * * *
12 Oliveira et al. * 64 * * *
2010
13 Poon et al. * 60 * *
2004 * 60 and 360 * *
14 Bhargava et al. * 60 * * * *
2006
15 Neves and * 55 and 80 * * * *
Almeida 2005
(continued)
Evaluation the Yield and Ultimate Strain of FRC in Compression
113
Table 2. (continued)
114
Reference Type of fibre Sp Vf, % Strength (MPa)

Steel PVA Polypropylene Basalt <1 1 fc < 50 50 fc < 100 fc 100
16 Hsu and * 60 * * *
ChengTzu 1994
17 Ayub et al. * 1389 * *
2015
S. Wtaife et al.
18 Bencardino * 40 * * *
et al. 2007
19 Dhakal et al. * 47, 64, 83 and * * *
2005 100
20 Altun et al. * 80 * *
2007
21 Ünal et al. 2007 * 60 and 80 * *
22 Srikar et al. * 35 * *
2016
23 Şahmaran et al. * 308 * *
2011
Evaluation the Yield and Ultimate Strain of FRC in Compression 115
3 Theoretical Analysis
To determine the stress block in the compression zone, it should be obtained from
stress-strain curves in compression. From 250-points database from previous research
[23–42] that explained the stress-strain curves in compression for different types of
fibre and volume fraction for both FRC and FRCC, it was easy to determine the stress
block through determining the parameters study. These parameters depended on yield
strain, ultimate strain, and elastic modulus. This paper used the Suksawang Equa-
tion [43] to determine the elastic modules for FRC and FRCC.
3.1 Yield Strain

The yield point represents the end of the elastic range of behaviour, which is repre-
sented by the elastic modulus. The elastic modulus is the slope of stress-strain rela-
tionship that it is the triangle shape. In ACI 544 the yield compressive strength for FRC
is adopted as:
rcy ¼ 0:85fc0 ð1Þ
Therefore,
0
0:85fc
ecy ¼ ð2Þ
Ec
0:85fc0
ecy ¼ pffiffiffiffi ð3Þ
4700 fc0
Then:
qffiffiffiffi
ecy ¼ 0:00018 fc0 ðMPaÞ
qffiffiffiffi ð4Þ
ecy ¼ 0:000015 fc0 ðpsiÞ
The new equation depends on the concept of elastic behaviour that the yield strain
is equal to the slope of the yield strength- elastic modules of the stress-strain curve. The
equation below determines the yield strain for FRC by determining factor a to account
for the influence of additional fibres in concrete.
a: fc0
ecy ¼ ð5Þ
Ec
This research adopted the Suksawang equation [43] for elastic modulus of FRC and
FRCC, shown below:
pffiffiffiffi
Ec ¼ 4700 kVf fc0 ðMPaÞ
pffiffiffiffi ð6Þ
Ec ¼ 57000 kVf fc0 ðpsiÞ
Where:
kVf ¼ 1 for C=S [ 1

Vf ð7Þ
kVf ¼ 1 þ 0:7
2 for C=S 1
3.2 Ultimate Strain

An elastic-perfectly plastic model defines the compression model. Since the com-
pressive behaviour of FRC will not vary significantly from that of normal concrete or
concrete without fibres, the same concept can be used for FRC in compression. Some
data indicate 0.003 may be conservative. Williamson (1973) and Pearlman (1979)
suggested the ultimate strain for steel fibre concrete should be 0.0033, while Swamy
et al. (1979) suggested a value of 0.0035.
According to the RILEM model16 [22, 23], the ultimate compressive strain ecu is
limited to 0.0035, which is the lower bound value of typical SFRC [21]. The value of
ultimate strain is taken as 0.0035, instead of 0.003, which recommended in ACI 318-
14, based on the results of compressive strain corresponding to the maximum com-
pressive strength. The Canadian standard (CSA) and European standard (EC2) also
suggest the value of ultimate strain as 0.0035. It has been reported that fibre-RC can
sustain a strain value of 0.005–0.006 at failure, and the corresponding failure stress
improves to 0.9fc0 from 0.85fc0 , where fc0 is the cylinder strength [48]. Accordingly,
the maximum design flexural stress of fibre-RC can be taken as 0.5fc0 by applying a
partial safety factor of 1.5, and the corresponding ultimate strain can be taken as 0.004
[48]. Based on the experimental work and the other researcher’s database, this study
will check a volume fraction of fibre affected by the ultimate strain. The value for the
ultimate strain was chosen as a lower bound on the test data that will be given the safer
design for FRC.
3.3 Statistical Indicators

In order to further evaluate the deviation between experimental data points and pre-
diction curves, the statistical indicators are employed. The statistical indicators of the
model and data are coefficients of variation (C.O.V), the mean square error (M.S.E),
and the mean deviation (S.D) to indicate systematic overestimation or underestimation
of a given model. The 45° line represent a perfect correlation between the calculated
and measured for the model. Data points above this line represent non-conservative
deviations of the model equation, while the data points below this line represent
conservative deviations. To better understand the variability between the calculated and
measured results, a coefficient of variation (C.O.V) is used and also illustrated in the
Figures for each equation. The C.O.V was computed by dividing the standard deviation
(r) by the mean (l).
To further examine the accuracy of these equations, other statistical indicators were
also used in the evaluation, including the mean square error and mean deviation. Both
of these statistical indicators provide a better mean to assess the quality and dispersion
of the equation in predicting the model.
The proposal methods and study parameters’ relationships are also validated and
confirmed for FRC and FRCC by using statistical and parametric analyses.
4.1 Yield Strain

A new equation for yield strain was developed by modifying the ACI 544 and RILEM
equations to account for the influence of additional fibres in concrete.
0
rcy ¼ a:fc ð8Þ
rcy
ecy ¼
Ec
ð9Þ
a:f 0
ecy ¼ c
Ec

0:85 for fc0 \69 MPað10000 psiÞ
a¼
1 for fc0 69 MPað10000 psiÞ
Substituting Eqs. (6) into (9) for C/S 1, Eq. 9 can be express as
pffiffiffiffi0
ecy ¼ 4700a kV f c . . .ðN=mm2 Þ
f pffiffiffiffi ð10Þ
ecy ¼ 57000a kV fc0 . . . Ib=in2
f
For C/S > 1,

pffiffiffiffi
ecy ¼ 0:00018 fc0 ðMPaÞ
pffiffiffiffi ð11Þ
ecy ¼ 0:000015 fc0 ðpsiÞ
In this research, a volume fraction of fibre did not affect yield strength. The sig-
nificant impact on yield strength was compressive strength. For normal strength
fc0 69 Mpa, the yield strength is 0.85 fc0 . For high strength fc0 69 Mpa, the yield
strength is roughly fc0 because the shape of the ascending branch of the stress-strain
relationship becomes more linear and steeper, and the slope of the descending part also
becomes steeper. The general shape of the stress-strain relationship becomes more
likely to be a triangle. Figures 1 and 2 illustrate the comparison between the calculated
and measured yield strain using the proposed equation. It is determined that the C.O.V.
of the proposed equation is 18.81%, while the C.O.V. of the ACI 544 is 27.90%. ACI
544 provides the best correlation between the calculated and measured yield strain with
mean square error and mean deviation of 20.10 and 0.89, respectively. Thus, the new
equation provides a good prediction of the yield strain with mean square error and
mean deviation of 18.48% and 0.99, respectively.
Fig. 1. Comparison of calculated and mea- Fig. 2. Comparison of calculated and mea-
surement of yield strain for ACI 544 surement of yield strain for proposal model.
4.2 Ultimate Strain

Based on the other researcher’s database, a volume fraction of fibre affected the ulti-
mate strain. There is no significant trend of an ultimate compressive strain of FRC and
FRCC as compressive strength increases. The ultimate strain at the extreme concrete
compression fibre is taken to be 0.0035 for volume fraction less than 1%, and 0.005 for
volume fraction more than or equal 1%. The value for ultimate strain was chosen as a
lower bound on the test data, as indicated by the line in Figs. 3 and 4. It will be given
the safer design for FRC.
Fig. 3. Ultimate compression strain for the volume fraction of fibre less than 1%.
Fig. 4. Ultimate compression strain for the volume fraction of fibre more than 1%.
5 Conclusion
Experimental results of this research, combined with other researches in the literature,
were analytically evaluated using statistical analysis. These analyses were used to
evaluate the validity of the provisions of codes and researches. Based on the experi-
ment in this dissertation and 250- stress-strain curves of the database for other studies
investigations compression strain and investigations carried out on the FRC and FRCC
for three groups of fibres (steel, PVA, basalt and synthetic fibre with volume fraction
ranging from 0 to 2%.
1. A new equation for yield strain was developed by modifying the ACI 544 and
Rilem equations to account for the influence of additional fibres in concrete. It was
observed that the new equation depends on the compressive strength.
2. The adopted the Suksawang equation for elastic modulus of FRC and FRCC that
depend on the types of mix and volume fraction of fibre provides better prediction
of the measured yield strain as compared to ACI 544 equations.
3. There is no significant trend of an ultimate compressive strain of FRC and FRCC as
compressive strength increases. Based on the other researchers’ database, a volume
fraction of fiber affected the ultimate strain. The ultimate strain at the extreme
concrete compression fiber is taken to be 0.0035 for volume fraction of less than
1%, and 0.005 for volume fraction more than or equal to 1%.
References
1. Grzybowski, M., Shah, S.P.: Shrinkage cracking of fiber reinforced concrete. Mater. J. 87(2),
138–148 (1990)
2. Sadiqul Islam, G.M., Gupta, S.D.: Evaluating plastic shrinkage and permeability of
polypropylene fiber reinforced concrete. Int. J. Sustain. Built Environ. (2016)
3. Saje, D., Bandelj, B., Šušteršič, J., Lopatič, J., Saje, F.: Shrinkage of polypropylene fiber-
reinforced high-performance concrete. J. Mater. Civ. Eng. 23(7), 941–952 (2010)
4. Giaccio, G., Bossio, M.E., Torrijos, M.C., Zerbino, R.: Contribution of fiber reinforcement in
concrete affected by alkali–silica reaction. Cem. Concr. Res. 67, 310–317 (2015)
5. Löfgren, I.: Fibre-reinforced Concrete for Industrial Construction-a fracture mechanics
approach to material testing and structural analysis. Chalmers University of Technology
(2005)
6. Martínez, D.M.: Bending with r-e and r-w approach (2006)
7. Jansson, A.: Fibres in reinforced concrete structures-analysis, experiments, and design
(2008)
8. Jiang, L.: Strain-hardening behavior of fiber reinforced concrete. In: Masters Abstracts
International, vol. 47, no. 03 (2003)
9. Luccioni, B., Isla, F., Codina, R., Ambrosini, D., Zerbino, R., Giaccio, G., Torrijos, M.C.:
Effect of steel fibers on static and blast response of high strength concrete. Int. J. Impact Eng.
107, 23–37 (2017)
10. Zerbino, R., Monetti, D.H., Giaccio, G.: Creep behaviour of cracked steel and macro-
synthetic fibre reinforced concrete. Mater. Struct. 49(8), 3397–3410 (2015)
11. AASHTO LRFD Bridge Design Specifications, Third Edition including 2005 and 2006
Interim Revisions, American Association of State Highway and Transportation Officials,
Washington DC (2004)
12. ACI Committee 318. Building code requirements for structural concrete (ACI 318–14) and
commentary. American Concrete Institute, Farmington Hill, MI (2014)
13. ACI Committee 544. 544.4R-18: Guide to Design with Fiber-Reinforced Concrete.
American Concrete Institute, Farmington Hill, MI (2018)
14. Canadian Highway Bridge Design Code, CSA S6 2001. Canadian Standards Association,
Rexdale, Ontario (2001)
15. EC2. BS EN 1992-1-2: Eurocode 2: Design of Concrete Structures–Part 1-2: General Rules:
Structural Fire Design. (2004)
16. Standard, Norwegian. Design of Concrete Structures (NS 3473). Norwegian Council for
Building Standardization, Oslo, Norway (1992)
17. Standard, NewZealand. New Zealand Concrete Structures Standards (NZS 3101) (1995)
18. Standard, Concrete Structures. Part 1-The Design of Concrete Structures. Standards New
Zealand, Wellington, New Zealand (2006)
19. Code, CEB-FIP Model. “Model code for concrete structures.” Bulletin D’Information 117-E
(1990)
20. AS3600-06. Australian Standard for Concrete Structures, Australia (2006)
21. Chote, S., Mobasher, B.: Flexural design of fiber-reinforced concrete. ACI Mater. J. 106(5),
461 (2009)
22. Vandewalle, L., Nemegeer, D., Balazs, L., di Prisco, M.: Final recommendations of
RILEM TC 162-TDF: test and design methods for steel fibre reinforced concrete sigma-
epsilon design method. Mater. Struct. 36(262), 560–567 (2003)
23. Vandewalle, L.: RILEM TC162-TDF: Test and Design Methods for Steel Fibre Reinforced
Concrete: sigma-epsilon Design Methods, “Technical Recommendation”. Mater. Struct. 33,
75–81 (2000)
24. Altun, F., Haktanir, T., Ari, K.: Effects of steel fiber addition on mechanical properties of
concrete and RC beams. Constr. Build. Mater. 21(3), 654–661 (2007)
25. Mansur, M.A., Chin, M.S., Wee, T.H.: Flexural behavior of high-strength concrete beams.
Struct. J. 94(6), 663–674 (1997)
26. Yang, K.-H.: Tests on concrete reinforced with hybrid or monolithic steel and polyvinyl
alcohol fibers. ACI Mater. J. 108(6) (2011)
27. Yoo, D.-Y., Yoon, Y.-S., Banthia, N.: Predicting the post-cracking behavior of normal and
high-strength steel-fiber-reinforced concrete beams. Constr. Build. Mater. 93, 477–485
(2011)
28. Lee, S.-C., Joung-Hwan, O., Cho, J.-Y.: Compressive behavior of fiber-reinforced concrete
with end-hooked steel fibers. Materials 8(4), 1442–1458 (2015)
29. Ou, Y.-C., Tsai, M.-S., Liu, K.-Y., Chang, K.-C.: Compressive behavior of steel-fiber-
reinforced concrete with a high reinforcing index. J. Mater. Civ. Eng. 24(2), 207–215 (2011)
30. Ezeldin, A.S., Balaguru, P.N.: Normal-and high-strength fiber-reinforced concrete under
compression. J. Mater. Civ. Eng. 4(4), 415–429 (1992)
31. Fanella, D.A., Naaman, A.E.: Stress-strain properties of fiber reinforced mortar in
compression. J. Proc. 82(4), 475–483 (1985)
32. Júnior, O, de Álvaro, L., dos Santos Borges, V.E., Danin, A.R., Machado, D.V.R., de Lima
Araújo, D., Debs, E., Khalil, M., Rodrigues, P.F.: Stress-strain curves for steel fiber-
reinforced concrete in compression. Matéria (Rio de Janeiro) 15(2), 260–266 (2010)
33. Poon, C.S., Shui, Z.H., Lam, L.: Compressive behavior of fiber reinforced high-performance
concrete subjected to elevated temperatures. Cem. Concr. Res. 34(12), 2215–2222 (2004)
34. Bhargava, P., Sharma, U.K., Kaushik, S.K.: Compressive stress-strain behavior of small-
scale steel fibre reinforced high strength concrete cylinders. J. Adv. Concr. Technol. 4(1),
109–121 (2006)
35. Neves, R.D., De Almeida, J.C.O.F.: Compressive behaviour of steel fibre reinforced
concrete. Struct. Concr. 6(1), 1–8 (2005)
36. Hsu, L.S., Hsu, C.T.T.: Stress-strain behavior of steel-fiber high-strength concrete under
compression. Struct. J. 91(4), 448–457 (1994)
37. Ayub, T., Shafiq, N., Khan, S.U.: Compressive stress-strain behavior of HSFRC reinforced
with basalt fibers. J. Mater. Civ. Eng. 28(4), 06015014 (2015)
38. Bencardino, F., Rizzuti, L., Spadea, G.: Experimental tests v/s theoretical modeling for FRC
in compression. In: Proceedings of 6th International Conference on Fracture Mechanics of
Concrete and Concrete Structures—FraMCoS-6, Catania, Italy, vol. 3, pp. 1473–1480
(2007)
39. Dhakal, R.P., Wang, C., Mander, J.B.: Behavior of steel fibre reinforced concrete in
compression (2005)
40. Ünal, O., Demir, F., Uygunoğlu, T.: Fuzzy logic approach to predict stress-strain curves of
steel fiber-reinforced concrete in compression. Build. Environ. 42(10), 3589–3595 (2007)
41. Srikar, G., Anand, G., Suriya Prakash, S.: A study on residual compression behavior of
structural fiber reinforced concrete exposed to moderate temperature using digital image
correlation. Int. J. Concr. Struct. Mater. 10(1), 75–85 (2016)
42. Şahmaran, M., Özbay, E., Yücel, H.E., Lachemi, M., Li, V.C.: Effect of fly ash and PVA
fiber on microstructural damage and residual properties of engineered cementitious
composites exposed to high temperatures. J. Mater. Civ. Eng. 23(12), 1735–1745 (2011)
43. Suksawang, N., Wtaife, S., Alsabbagh, A.: Evaluation of elastic modulus of fiber-reinforced
concrete (2018)
44. Williamson, G.R.: Compression characteristics and structural beam design analysis of steel
fiber reinforced concrete. No. Cerl-TR-M-62. Construction Engineering Research Lab
(Army) Champaign, IL (1973)
45. Pearlman, S.L.: Flexural performance of reinforced steel fiber concrete beams. Ph.D. diss.,
MS thesis, Carnegie-Mellon University, Pittsburgh (1979)
46. Swamy, R.N., Al-Taan, S., Ali, S.A.R.: Steel fibers for controlling cracking and deflection.
Concr. Int. (1979)
47. Lim, T.-Y., Paramasivam, P., Lee, S.-L.: Behavior of reinforced steel-fiber-concrete beams
in flexure. J. Struct. Eng. 113(12), 2439–2458 (1987)
48. Singh, H.: Flexural modeling of steel fiber-reinforced concrete members: analytical
investigations. Pract. Period. Struct. Des. Constr. 20(4), 04014046 (2014)
Electromagnetic Shielding Characteristics
of High Performance Fiber Reinforced
Namkon Lee(&), Sungwook Kim, and Gijoon Park
Department of Infrastructure Safety Research, Korea Institute of Civil

Engineering and Building Technology, Goyang-si, South Korea
nklee@kict.re.kr
Abstract. This study aims to investigate the effect of multi-walled carbon

nanotubes (MWCNTs) and steel fibers on the electrical conductivity and elec-
tromagnetic shielding effectiveness (SE) of a high-performance, fiber-reinforced
cementitious composite (HPFRCC). The electrical conductivity of the 100 MPa
HPFRCC with 0.30% MWCNT was 0.093 S/cm and that of the 180 MPa
HPFRCC with 0.35% MWCNT and 2.0% steel fiber was 0.0173 S/cm. At 2.0%
steel fiber and 0.3% MWCNT contents, the electromagnetic SE values of the
HPFRCC were 45.8 dB (horizontal) and 42.1 dB (vertical), which are slightly
higher than 37.9 dB (horizontal) at 2.0% steel fiber content or 39.2 dB (hori-
zontal) at 0.3% MWCNT content. The incorporation of steel fibers by 3.0 vol. %
did not result in any electrical percolation path in the HPFRCC at the micro
level; thus, a high electrical conductivity could not be achieved. At the macro
level, the proper dispersion of the steel fibers into the HPFRCC helped reflect
and absorb the electromagnetic waves, increasing the electromagnetic SE. The
incorporation of steel fibers helped improve the electromagnetic SE regardless
of the formation of percolation paths, whereas the incorporation of MWCNTs
helped improve the electromagnetic SE only under the condition that percolation
paths were formed in the cement matrix.
Keywords: HPFRCC MWCNT Steel fiber Electrical conductivity

Shielding effectiveness
1 Introduction
A high-performance fiber-reinforced cementitious composite (HPFRCC) mainly con-

sists of cement, filler, aggregate, microsilica, and superplasticizer. Among the con-
stituent materials of HPFRCCs, steel fibers have a high tensile strength of 2,500 MPa
and a good electrical conductive property.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_11
124 N. Lee et al.
The popularity of electronic devices, especially wireless and communication sys-

tems, brings the problems of electromagnetic interference (EMI) pollution and infor-
mation security, which are also harmful to the personal health and confidential military
information [1]. In order to mitigate the electromagnetic interference pollution prob-
lems, it is essential to develop high frequency electromagnetic shielding materials
which act as a barrier to limit the penetration of electromagnetic wave by reflecting or
absorbing [2]. Steel fibers and CNTs are recommended for electromagnetic shielding as
they are conductive; their use as macrofibers and microfibers, respectively, is expected
to create a synergistic effect on the electrical conductivity of concrete by providing a
conductive network in the cement matrix, thus giving rise to a percolation path.
However, studies on the electrical conductivity and electromagnetic shielding effec-
tiveness of steel fiber/CNT-incorporated cementitious composites are lacking. In par-
ticular, there is no study on the electromagnetic shielding effectiveness of HPFRCCs.
Steel fibers are generally utilized to increase the ductility of HPFRCCs as well as to
provide a good electromagnetic shielding owing to its conductive property. CNTs can
be better dispersed in an HPFRCC matrix than in an OPC matrix as HPFRCCs have an
extremely low water-to-cement ratio and contain a high amount of silica fume, which
plays a role of ball bearing effect. Under an oven dry condition, a low water-to-cement
ratio can help improve the dispersion level of CNTs in cement paste.
This study aims to investigate the electromagnetic shielding characteristics of
HPFRCC. Steel fibers and MWCNTs are used as conductive materials to improve
shielding effectiveness (SE) of HPFRCC. The shielding effectiveness measurement of
the HPFRCC samples is conducted inside a dual shielding room system in accordance
with MIL-STD-188-125.
2 Experimental Method
2.1 Materials and Mix Proportions

The ordinary Portland cement (OPC, ASTM C150 Type I) used in this study was
provided by Sungshin Cement Corp., South Korea. The OPC cement had a Blaine
fineness of 3700 cm2/g and a specific gravity of 3.17. Silica fume was supplied by
Elkem Corp., South Korea. Silica powder (average grain size: 14 lm) and quartz sand
with a diameter in the range of 100–800 lm were employed for the fabrication of
HPFRCCs. Steel fibers (diameter: 0.2 mm, length: 19.5 mm) and MWCNTs (diameter:
6–9 nm, length: 50–200 lm, purity: min 98.5%) were used as conductive materials. In
addition, an MWCNT liquid solution prepared by dispersing MWCNTs in distilled
water by sonication was used as a conductive material Table 1.
Electromagnetic Shielding Characteristics 125
Table 1. Chemical composition of binder materials [3]

(wt%) XRF Rietveld analysis
OPC Fly ash Silica fume Compound OPC
SiO2 20.6 38.07 95.31 C3S 62.2
Al2O3 5.0 14.54 0.1 C2S 11.0
Fe2O3 3.4 5.42 0.35 C4AF 11.0
CaO 60.7 22.78 0.21 C3A 4.2
MgO 2.6 2.67 0.8 Calcite 5.1
SO3 2.38 5.45 0.55 Gypsum 0.7
K2O 0.98 5.83 – Anhydrite –
Na2O 0.15 0.92 0.19 Ye’elimite (C) –
TiO2 0.27 3.62 – Ye’elimite (O) –
P2O5 0.11 1.52 0.03 CT –
Others <0.25 1.19 – C2AS –
LOI 0.75 7.1 2.46 Hemihydrate –
Quartz –
Arcanite –
Others 2.2
Table 2 lists the mix proportions of the HPFRCC. The mix proportions are divided
into two categories: HPFRCCs with 100 MPa and 180 MPa of the compressive
strengths.
The water-to-binder (cement + fly ash + microsilica) ratio was 0.30 for the
100 MPa HPFRCC and the ratios were 0.20 and 0.30 for the 180 MPa HPFRCC. The
Table 2. Mix proportion of HPFRCC (mass ratio)

Mixture w/b OPC Fly Micro CNT Fine Filler SP Steel
ratio ash silica (%) aggregate fiber** (%)
100 MPa 0.3 1 0.2 0.1 – 1.2 0.2 0.015 2.0
100MPa_S1.0 0.3 1 0.2 0.1 – 1.2 0.2 0.015 1.0
100MPa_S2.0 0.3 1 0.2 0.1 – 1.2 0.2 0.015 2.0
100MPa_S3.0 0.3 1 0.2 0.1 – 1.2 0.2 0.015 3.0
100MPa_N0.3 0.3 1 0.2 0.1 0.30 1.2 0.2 0.015 2.0
100MPa_S2.0_N0.3 0.3 1 0.2 0.1 0.30 1.2 0.2 0.015 2.0
180 MPa 0.2 1 – 0.25 – 1.2 0.2 0.015 2.0
180 MPa_S2.0 0.2 1 – 0.25 – 1.2 0.2 0.015 2.0
180MPa_N0.35* 0.3 1 – 0.25 0.35 1.2 0.2 0.015 2.0
180MPa_S2.0_N0.35* 0.3 1 – 0.25 0.35 1.2 0.2 0.015 2.0
*
The w/b ratio of the 180MPa samples with the addition of MWCNT is 0.30, and that of the 180 MPa
samples without MWCNT is 0.20.
**
It indicates a volume fraction of steel fibers in HPFRCC.
126 N. Lee et al.
amount of superplasticizer agent (SP) was adjusted to meet the mini slump flow
requirement of 180 mm (measured after 25 times hits); however, some samples could
not meet this requirement because of the high amount of MWCNTs added.
Steel fibers were added to the HPFRCC at 0, 1.0, 2.0, and 3.0 vol. %, respectively.
The MWCNTs were added to the HPFRCC at 0.30 wt% and 0.35 wt% by weight of
cement. The 100 MPa and 180 MPa HPFRCCs were manufactured in accordance with
the study conducted by Lee et al. (2018) [3].
2.2 Testing Methods

An unconfined compressive strength test was conducted using a 300 kN universal
testing machine in accordance with ASTM C39. The AC impedance of the
60 60 160 mm specimen embedded with two copper electrodes (20 60
0.5 mm) was measured at an interval of 30 mm. LCR meters (Keysight Technologies,
model: E4980A) were employed to measure the AC (Alternative Current)-impedance.
The frequency was swept from 1 MHz down to 1 Hz using a logarithmic point spacing
of 50 points. To minimize the conduction effect of the pore solution, the tests were
performed after the sample was dried at 60 °C for three days. To confirm the true bulk
resistance of the HPFRCC specimens, four-point DC measurements were also
conducted.
The shielding effectiveness measurement of the HPFRCC samples was conducted
in accordance with the military standard MIL-STD-188-125. Figure 1 shows the sys-
tem diagram used for measuring the shielding effectiveness of the HPFRCC samples.
Fig. 1. Configuration of shielding effectiveness measurement system
3 Results

Table 3 shows the compressive strength results of the 100 and 180 MPa HPFRCCs. As
the steel fibre amount increases, the compressive strength is slightly increased. The
addition of MWCNT in the HPFRCC led to the reduction in the compressive strength
from 106.5 to 99.9 MPa. When both steel fiber and MWCNT are added to the
100 MPa HPFRCC, the compressive strength (106.7 MPa) was not reduced, compared
to that (106.5 MPa) of plain 100 MPa HPFRCC.
For the 180 MPa samples, the addition of 2.0% steel fibres resulted in the increase
of compressive strength from 174.0 to 191.6 MPa. The compressive strength of
180 MPa samples with addition of MWCNT was significantly lower than that of plain
180 MPa sample due to the increased w/b ratio (0.2 ! 0.3). In case of the 180 MPa
samples with MWCNT, the w/b ratio was adjusted to be 0.3 to improve the reduced
flowability of HPFRCC resulting from the addition of MWCNT.
Table 3. Compressive strength results [4]

Sample Compressive strength (MPa)
100 MPa 106.5
100MPa_S1.0 106.6
100MPa_S2.0 106.4
100MPa_S3.0 107.8
100MPa_N0.3 99.9
100MPa_S2.0_N0.3 106.7
180 MPa 174.0
180MPa_S2.0 191.6
180MPa_N0.35 141.7
180MPa_S2.0_N0.35 141.2
3.2 Electrical Conductivity

Table 4 lists the electrical conductivity of the 100 and 180 MPa HPFRCCs. The
electrical conductivity (r) can be calculated as follows:
1 L
r¼ ð1Þ
R A
where, R is the resistance (X); A is the area of contact between the material and the
electrodes (cm2); L is the distance between the electrodes (cm); and r is the electrical
conductivity (S/cm).
With the increase in the steel fiber content from 0% to 3%, the electrical con-
ductivity of the HPFRCC increases from 1.32 10−4 to 8.75 10−4 S/cm. However,
the extent of increment in the electrical conductivity by the addition of steel fiber is not
significant, compared to that in the electrical conductivity by addition of MWCNT. At
the addition of 0.3% MWCNT, the electrical conductivity of the 100 MPa sample was
932.4 10−4 S/cm. At the addition of 0.35% of MWCNT, that of the 180 MPa sample
was 172.7 10−4 S/cm. The influence of MWCNT on the electrical conductivity of
HPFRCC is higher than that of steel fiber.
128 N. Lee et al.
3.3 Electromagnetic Wave Shielding Effectiveness

Table 4 gives a summary of the SE (Shielding effectiveness) results measured at a
frequency of 1 GHz. The SE of the 100 MPa HPFRCC without steel fibers is 6.5 dB at
the horizontal antenna and 9.1 dB at the vertical antenna. As the steel fiber content is
increased from 0 to 1.0%, the SE increases. However, the SE does not increase further
when the steel fiber content is increased from 1.0 to 3.0%. This indicates that the steel
fibers in the HPFRCC are well dispersed at 1.0 vol.%. When the fiber content is more
than 2.0%, the clumping or aggregation of the steel fibers may occur in the HPFRCC,
making it difficult to further increase the SE of the HPFRCC.
Table 4. Electrical conductivity and SE results [4]

Sample Electrical Shielding effectiveness (SE)
conductivity (S/cm) Frequency Horizontal Vertical
(GHz) (dB) (dB)
100 MPa 0.000132 1.0 6.5 9.06
100MPa_S1.0 0.000669 1.0 31.86 47.33
100MPa_S2.0 0.000858 1.0 34.65 37.9
100MPa_S3.0 0.000875 1.0 41.25 35.59
100MPa_N0.3 0.09324 1.0 34.72 39.19
100MPa_S2.0_N0.3 0.1333 1.0 45.82 42.11
180 MPa N/A 1.0 1.42 2.72
180MPa_S2.0 0.0000184 1.0 40.65 43.91
180MPa_N0.35 N/A 1.0 28.37 21.87
180MPa_S2.0_N0.35 0.017271 1.0 37.47 51.72
The addition of 0.3% MWCNTs led to a significant increase in the SE of the

HPFRCC without the steel fibers. The use of both MWCNTs and steel fibers did not
increase the SE value as high as expected. At 2.0% steel fiber and 0.3% MWCNT
contents, the SE values of the HPFRCC are 45.8 dB (horizontal) and 42.1 dB (verti-
cal), which are slightly higher than 37.9 dB (horizontal) at 2.0% steel fiber content and
39.2 dB (horizontal) at 0.3% MWCNT content. The synergy effect with the use of both
steel fibers and MWCNTs was not clear in the HPFRCC.
Table 4 shows that there is no direct relationship between the electrical conduc-
tivity and the SE results of the HPFRCC incorporating steel fiber. In terms of micro
level, the incorporation of steel fibers did not result in an electrical percolation path in
the HPFRCC; therefore, the electrical conductivity could not be increased. In terms of
macro level, the proper dispersion of the steel fibers into the HPFRCC helped reflect
and absorb the electromagnetic waves, thus increasing the electromagnetic SE. The
electrical conductivity of the HPFRCC with 2.0% steel fiber content was 0.00086 S/cm
whereas its SE values were 45.8 dB (horizontal) and 42.1 dB (vertical) at 1 GHz.
Despite the electrical conductivity of the HPFRCC being as low as 10−5 S/cm, the SE
value was as high as 40 dB.
The formation of the electrical percolation path with the addition of MWCNTs had
a positive effect on the electromagnetic SE. The electrical conductivity of the HPFRCC
with 0.3% MWCNT content was 0.093 S/cm, which was high enough to form a
percolation path, and the SE values were 34.7 dB (horizontal) and 39.2 dB (vertical) at
1 GHz.
4 Conclusions
This study presented the experimental results and discussions pertaining to the elec-
trical conductivity and electromagnetic shielding effectiveness of HPFRCCs. To
improve the electrical conductivity of HPFRCCs, MWCNTs and steel fibers were
added to the HPFRCC. The following conclusions can be drawn from the results
presented in this paper.
• The electrical conductivity of the 100 MPa HPFRCC with 0.30% MWCNT was
0.093 S/cm and that of the 180 MPa HPFRCC with 0.35% MWCNT and 2.0% steel
fiber was 0.0173 S/m. To achieve a high electrical conductivity, adding MWCNTs
was more beneficial than adding steel fibers.
• At 2.0% steel fiber and 0.3% MWCNT contents, the SE values of the HPFRCC
were found to be 45.8 dB (horizontal) and 42.1 dB (vertical), which were slightly
higher than those (37.9 dB (horizontal)) at 2.0% steel fiber content and that
(39.2 dB (horizontal)) at 0.3% MWCNT content.
• The incorporation of steel fibers did not result in any electrical percolation path in
the HPFRCC at the micro level; therefore, a high electrical conductivity could not
be achieved. At the macro level, the proper dispersion of the steel fibers into the
HPFRCC helped reflect and absorb the electromagnetic waves, consequently
increasing the electromagnetic SE. Although there was no electrical percolation
path in the HPFRCC, a high electromagnetic SE could be achieved.
• The incorporation of steel fibers at the macro level can help improve the electro-
magnetic SE regardless of the formation of electrical percolation paths whereas the
incorporation of MWCNTs at the micro level can help improve the electromagnetic
SE only when electrical percolation paths are formed in the cement matrix.
Acknowledgements. This research was supported by a grant (20SCIP-B146646-03) from

Construction Technology Research Project funded by the Ministry of Land, Infrastructure and
Transport of Korea government.
References
1. Wansom, S., Kidner, N.J., Woo, L.Y., Mason, T.O.: AC-impedance response of multi-walled
carbon nanotube/cement composites. Cement Concr. Compos. 28(6), 509–519 (2006)
2. Singh, A.P., Gupta, B.K., Mishra, M.: Multiwalled carbon nanotube/cement composites with
exceptional electromagnetic interference shielding properties. Carbon 56, 86–96 (2013)
130 N. Lee et al.
3. Lee, N.K., Koh, K.T., Kim, M.O., Ryu, G.S.: Uncovering the role of micro silica in hydration
of ultra-high performance concrete (UHPC). Cem. Concr. Res. 104, 68–79 (2018)
4. Lee, N., Kim, S., Park, G.: The effects of multi-walled carbon nanotubes and steel fibers on
the AC impedance and electromagnetic shielding effectiveness of high-performance. Fiber-
Reinforced Cement. Compos. Mater. 12(21), 3591 (2019)
Mechanical Properties
The Manufacture of Fiber Cement Blocks
Using Chemical and Thermomechanical Pulps
and Rice Husk Ash
Javad Torkaman(&)
Associate Professor, Faculty of Natural Resources,

University of Guilan, Rasht, Iran
torkaman@guilan.ac.ir
Abstract. In this work, We made fiber cement blocks by using the raw
materials of Cement, Rice husk ash, Fine aggregate, Fibers of Chemical pulps
(CP) and Thermomechanical pulps (TMP). The replacement amount were zero
(control), 5, 10, 15, 20 and 25%. In total 36 cement blocks with dimensions of
15 * 15 * 15 cm were made. The properties of blocks which were measured
include compressive strength, water absorption, density of before and after
soaking. The data Statistically analysed by Spss software. Statistical analysis
showed that the type of fibers had significant effect on both mechanical and
physical properties at confidence level of 0.05. Based on the findings of this
work, The CP fibers had better effects on the compressive strength of specimens
than The TMP fibers approximately twofold. Increasing the replacement level of
TMP fibers tends to reduce the compressive strength due to the low binding
ability. The water absorption and density values of specimens contain fibers
were lower than control. The fibers cause lighter weight, resistance to cracking
and a degree of flexibility.
Keywords: Chemical pulp Ash Block Compressive strength Water

absorption
1 Introduction
Since many years ago, natural fibers have been used as reinforced inorganic materials
and after finding hazardous effects of asbestos fibers on human health. New fibers with
both economic and environmental benefits are being considered for applications in the
most of industries [1]. The aims of fibers applications are to achieve desired properties
and to reduce the cost of the products [2]. The use of lingocellulosic fibers have both
advantages (Low density, Low cost, Low energy consumption, non-hazardous) and
disadvantages (High moisture absorption and the low compatibility between fibre and
cement) [1, 3]. In thermo mechanical pulping (TMP), Pressurized steam is applied
before and during refining to raise the wood temperature to soften the lignin. The bonds
between fibers to break gradually and fiber bundles, single fibers and fiber fragments to
be released. But in the chemical pulping (CP) processes, the fibers are liberated from
the wood matrix as the lignin is removed by dissolving it into the cooking chemical
solution at a high temperature. However there are many differences both in quality and
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_12
134 J. Torkaman
the cost between them. In general, chemical pulps are not only longer but they are
much more flexible and have a nearly pure cellulose surface that forms strong bonds.
The CP yield is low (40–70%) whereas TMP yield is high (90–98%). Also there has
been a strong desire to use of recycled paper and agricultural residues such as rice husk
as ash. Rice husk during milling process obtained from the outer covering of rice grains
[4, 5]. When rice husk is burned between 500 and 700 °C amorphous silica is soluble
and reactive in an alkaline solution [6, 7]. Two types of ashes (white and black) are
produced during burning process [6]. The rice husk ash is a pozzolanic material which
can to replace cement by up to 30% [4, 7]. Addition of fibers to concrete blocks imparts
a number of attributes include resistance to cracking and a lighter weight [6] to reduce
the cost of the final product [8]. The type of fibers that can be used as reinforcing agent
in cement composites must be technically and economically acceptable [1, 2]. Moslemi
(2008) has reported that due to a particular fibre’s specific gravity, tensile strength and
cost per unit weight Kraft pulp fibers are favour fibers in place of asbestos [2]. We
demonstrated before 25 wt% replacement of wood fibre waste (WFW) and rice husk
ash (RHA) have good physic-mechanical properties [3]. In this work has focused the
effects of the type and amount of pulps on the concrete blocks properties.
2 Experimental Procedures
2.1 Materials
The fibrous raw materials for this study were thermomechanical pulp (TMP) and
chemical pulp (CP) which were collected from fiberboard and paper companies called
Royal and Latif. The ash of rice husk (RHA) was used as cement replacement and
plays a pozzolanic role. Also calcium chloride (Ca cl2) was used as cement setting
accelerator. Ordinary Portland Cement (OPC) a product of Khazar Cement Co. Iran
that was employed as binder agent. Graded River Sand (GRS) was used as fine
aggregate.
2.2 Mixing and Fabrication of Blocks

Mixture proportion of the raw materials are summarized in Table 1 for using each pulp
seven different types of mixtures were prepared in the laboratory trails.
The amount of water was calculated for each block using the following Eq. (1) [9].
Water amountðmlÞ ¼ 0:25 cement weightðgrÞ þ 2:7 fiber weightðgrÞ ð1Þ
According to flow chart (Fig. 1) raw materials were placed in a concrete mixer and
mixed for 3 min and then the dilute aqueous solution of Cacl2 and water were added.
In order to obtain more homogeneous mixes, the paste was mixed for another 2 min.
Consequently, the blended mortars were immediately fed into the steel moulds
(150 150 150 mm3). The cast moulds were vibrated for 1 min to achieve adequate
compaction afterward, the cast specimens were covered with plastic to prevent water
The Manufacture of Fiber Cement Blocks 135
Table 1. Mixture proportion of raw materials
-Portland Cement
-Graded river sand(GRS)
-Rice husk ash
Compressive Strength
Water absorption
Mixing &Casting Curing
Bulk density after and
before Soaking
TMP&CP
(Variable Parameters)
Fig. 1. Flow chart of concrete black fabrication and characterization
loss. After 24 h the blocks were decamped and conditioned for 28 days at 25 ± 1 °C
and 65 ± 5%RH to allow the cement to cure and gain strength.
2.3 Tests Methods

The series of tests were carried out according to ASTM.C67 [10] to determine the
compressive strength, water absorption and bulk density of the block samples.
2.3.1 Compressive Strength

The composite specimens were prepared in accordance with ASTM C109 [11]. Each
compressive strength value reported is the average of three samples. The dry com-
pression strength was determined using an Instron Universal Testing Machine (Model
4486) with a loading speed of 10 mm/min.
2.3.2 Water Absorption

Water absorption was carried out using ASTM.C642 [12]. The cube specimens for
water absorption were completely submerged horizontally under distilled water
maintained at 25 °C for 24 h. After soaking, the samples were drained on paper towels
for 10 min to remove excess water. The water absorption was calculated from the
increase in weight of the specimen during submersion. At least four specimens of every
treatment were tested to obtain a reliable average and standard deviations.
136 J. Torkaman
2.3.3 Bulk Density

Specimens were tested following ASTM C642 for bulk density. The densities of the
composites were determined by measuring the mass and volume of each sample. The
air-dried samples were oven-dried up to 103 ± 2 °C until they reached constant
weights. Then, the samples were cooled in a desiccators containing calcium chloride
and weighed in an analytic balance with ±0.001 g sensitivity. The mass of each
sample was obtained by calculating the arithmetic mean of the mass of all of the test
samples. Afterward, the dimensions of the specimens were measured using a digital
caliper with ±0.001 mm sensitivity and the volumes were determined by the stereo
metric method. The density (D) was then calculated using the following Eq. (2).
Mo
D¼ ð2Þ
Vo
Where Mo is the oven dry weight (g) and VO is the dry volume (cm3) of the sample.
2.3.4 Data Analysis

Measured data on mechanical and physical properties of the composites were analysed
with analysis of variance (ANOVA) procedure using Spss software (version13).
Duncan’s Multiple Range tests were used to compare the difference among the mean
values for the groups properties at the level of 0.05.
The results of the mechanical and physical testes, with statistical analysis, are shown in
Tables 2 and 3. All blocks made with different amount of the CP fibers had the highest
values of the compressive strength than TMP fibers reinforcement. The unsatisfactory
results in compressive behaviour were obtained in TMP fiber-cement blocks. Mostly
depends on the formation of fiber-matrix, matrix-matrix and fiber-fiber bonds. The
bonding can be affected by quantity and quality of fibers in given volume of materials [1,
13, 14]. In addition, pulping process can influence the mechanical and physical prop-
erties of the fiber-cement blocks. The TMP fibers creates a lack of homogenous mixture
in the blocks than the CP fibers. There is not significant difference between the com-
pressive strength of control and CP fiber blocks (P-value = 0.526, a = 0.05). The CP
fibers are shown to be superior to the TMP fibers in increasing the compressive strength.
The compressive strength of blocks content 5% of CP fibers were showed more strength
(7.15 MPa) than control specimens (6.5 MPa). Statistical analysis showed that the
compressive strength and Physical properties in terms of water absorption and density of
the samples were influenced by the type of fibers (Fig. 2 and 3).
The experiment was arranged in a completely randomized design. Average values
were compared by Tukey test at 5%. According to Analysis of Variance (Table 3)
between three groups of the control, CP and TMP fibers there were significant dif-
ference at level of 0.05 on the mechanical and physical properties.
Table 2. Mechanical and physical properties of concrete blocks
Table 3. Analysis of variance on some mechanical and physical properties
All Samples made with CP and TMP fibers had the lowest values of water
absorption among the other specimens (Fig. 2). Because two types of fibers to absorb
water and to reach to fiber saturation point (FSP) in mixing process. Consequently after
Soaking don’t tend extra water. In relation to the density measurements before and after
138 J. Torkaman
Compressive Strength
8 6.48 6.37
6
(Mpa)
4 3.18
2
0
Control CP TMP
Type of Fibers
(a)
10 8.66
Water Absorption(%)
4
2.23
2 1.15
0
Control CP TMP
Type of Fiber
(b)
Fig. 2. Effect of Type of fibers on the (a) Compressive Strength and (b) Water Absorption
2.5 before soaking

2.23
1.9 after soaking

2 1.83
1.925
Density(gr/cm3)
1.84 1.78
1.5
0.5
0
Control CP TMP
Type of fibers
Fig. 3. Effect of Type of fibers on the bulk density before and after Soaking
soaking by replacing fibers the density of fiber-cement blocks were decreased gradu-
ally, similar results were reported by Torkaman et al. (2014). By replacing 15% CP
fibers to made of fiber-cement blocks the best properties of the compressive strength
and water absorption were obtained.
4 Conclusion
Based on the results of this research the following conclusion can be drown:
1. The CP fibers had best results in compare of control samples and TMP fibers on the
mechanical and physical properties of the blocks.
2. There wasn’t significant difference between different levels of both types of fibers
on the compressive strength.
3. The optimum condition was obtained the CP fiber contents were 15% by weight.
References
1. Ashori, A., Tabarsa, T., Valizadeh, I.: Fiber reinforced cement boards made from recycled
newsprint paper. Mater. Sci. Eng. A528, 7801–7804 (2011)
2. Moslemi, A.: Technology and market considerations for fiber cement composites. In: 11th
International Inorganic-Bonded Fiber Conference. November 5–7, Madrid-Spain (2008)
3. Torkaman, J., Ashori, A., Sader Momtazi, A.: Using wood fiber waste, rice husk ash and
Limestone power waste as cement replacement materials for lightweight concrete blocks.
Construc. Build. Mater. 50, 432–436 (2014)
4. Ramezanianpour, A.A., Mahdi Khani, M., Ahmadibeni, G.H.: The effect of rice husk ash on
mechanical properties and durability of sustainable concretes. Int. J. Civ. Eng. 7(2), 83–91
(2009)
5. Ganesan, K., Rajagopal, K., Thanagavel, K.: Rice husk ash blended cement: assessment of
optimal level of replacement for strength and permeability properties of concrete. Construc.
Build. Mater. 22, 1675–1683 (2008)
6. Hamzeh, Y., Ziabari, K.P., Torkaman, J., Ashori, A., Jafari, M.: Study on the effects of white
rice husk ash and fibrous materials addition on some properties of fiber cement composites.
J. Environ. Manage. 117, 263–267 (2013)
7. Ghofrani, M., Mokaram, K.N., Ashori, A., Torkaman, J.: Fiber-cement composite using rice
husk fiber and rice husk ash: mechanical and physical properties. J. Compos. Mater. 6, 1–6
(2014)
8. Chatveera, B., Lertwattanaruk, P.: Durabillity of conventional concretes containing block
rice husk ash. J. Environ. Manage. 92, 59–66 (2011)
9. Pereira, C., Jorge, F.C., Ferreira, J.M., Irle, M.: Characterizing the setting of cement when
mixed with cork, blue gum, or maritime pine, grown in Portugal I: temperature profiles and
compatibility indices. J. Wood Sci. 52, 311–317 (2006)
10. ASTM C67. Standard test methods for sampling and testing brick and structural clay tile.
Annual book of ASTM standards, Philadelphia, PA (2003)
11. ASTM C109. Standard test methods for Compressive strength of hydraulic cement mortars
(using 2-in or [50 mm] cube specimens]. Annual book of ASTM standards, Philadelphia, PA
(2012)
12. ASTM C642. Standard test method for density, absorption, and voids in hardened concrete.
Annual book of ASTM Standards. Philadelphia, PA (2006)
13. Enayati, A.A., Hooshmand, N.H., Doosthoseini, K., Latibari, J.A., Rahimi, S.: Evaluation of
the properties of wood sawdust-cement perforated Blocks. Iran. J. Wood Paper Sci. Res.
27(2), 294–305 (2012)
14. Fernandez, E.C., Taja-on,V.P.: The use and processing of rice straw in the manufacture of
cement-bonded fiber board. In: Forest Products and Paper Science pp. 49–54 (2000)
Post-Fire Flexural Tensile Strength of Macro
Synthetic Fibre Reinforced Concrete
Olivia Mirza1(&), Brendan Kirkland1, Kurt Bogart1,

and Todd Clarke2
1
Western Sydney University,
Locked Bag 1797, Penrith 2751, NSW Sydney, Australia
o.mirza@westernsydney.edu.au
2
BarChip Australia Pty Ltd, Sydney, NSW, Australia
Abstract. Tensile cracks in plain concrete occur and propagate when the ten-
sile stress exceeds the bond strength of the cement. These cracks are generally
bridged by coarse aggregates however the inclusion of fibres in concrete pro-
vides significant additional crack bridging. The utilisation of fibres in concrete
has been extensively researched however most studies have focussed on steel
fibre or newer types of fibre including glass, basalt and carbon. Current Aus-
tralian and international design standards do not address the loss of strength for
macro synthetic fibre reinforced concrete after exposure to elevated tempera-
tures. This paper is an experimental study of the post-fire performance of macro
synthetic fibre reinforced concrete. The flexural performance of the concrete is
determined by performing 4-point bending tests on prisms after being heated in
a furnace. Specimens were exposed to temperatures up to 800 °C. It was
observed that the addition of macro synthetic fibre in concrete has a negligible
effect on the overall peak cracking load of the specimens. The significant
benefits of fibres are seen after the onset of cracking as the fibres hold the crack
together which allows the specimen to withstand greater deflections and achieve
a high residual strength. After exposure, the specimens displayed post-cracking
flexural strengths between approximately 2–3 MPa for temperatures 600 °C.
The specimens with 0.8% fibre displayed much higher strengths than the
specimens with 0.4% fibre for temperatures 400 °C. However, for strengths
after exposure to 600 °C and 800 °C, far less variance was observed.
Keywords: Flexural tensile strength Post-fire Macro synthetic fibre Fibre

reinforced concrete
1 Introduction
Tensile cracks in plain concrete occur and propagate when the tensile stress exceeds the
bond strength of the cement. These cracks are generally bridged by coarse aggregates
however the inclusion of fibres in concrete provides significant additional crack
bridging [1]. Fibres are used in a range of applications for reinforcing concrete
including pavements, roads, shotcrete, railway transoms and tunnel linings. The pop-
ularity of fibre reinforced concrete (FRC) stems from the considerable improvement in
the concrete properties that can be achieved in comparison to traditional reinforced
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_13
Post-Fire Flexural Tensile Strength 141
concrete. The most common types of fibre used in the Australian construction industry
are steel and plastic (macro synthetic). The tensile strength of steel fibre is approxi-
mately 1300 MPa while the macro synthetic fibre varies from 250–800 MPa depending
on the type of synthetic polymer [2, 3]. Steel’s superior elastic modulus results in
higher flexural and direct tensile strengths compared to other FRCs [2–4].
Tensile failure of FRC occurs when the stress exceeds the bond strength between
the fibre and the cement or if it exceeds the tensile strength of the fibre. In the case of
exceeding the bond strength, a gradual pull out of the fibres occurs. This is also
possible where thin fibres are bent during casting or if relatively short fibres are used.
Where the bond strength or pull out are not observed, the large tensile strength of the
fibres governs the ultimate failure.
Although corrosion of uncracked steel fibre reinforced concrete is considered
insignificant and limited only to the surface [5–7], corrosion of cracked concrete is a
concern where the highly corrosive nature of steel fibres can gradually reduce the
tensile strength [7]. Additionally, the deterioration of the tensile strength is increased as
the crack width increases [8]. This creates a demand for the utilisation of macro
synthetic fibre reinforced concrete (MSFRC) [9, 10] where there is no concern for
corrosion [11, 12].
The utilisation of fibres in concrete has been extensively researched however most
studies have focussed on steel fibre or newer types of fibre including glass, basalt and
carbon [13]. Current Australian and international design standards do not address the
loss of strength for MSFRC after exposure to elevated temperatures. Additionally, a
limited amount of studies have been identified in open literature investigating the
impact of fire on other MSFRC properties [14–17]. This necessitates the need for the
current study. This paper presents the results of an experimental study of the post-fire
performance of macro synthetic fibre reinforced concrete.
2 Methodology
2.1 Materials and Specimen Preparation

Three batches of concrete were cast with the mix proportion provided in Table 1 based
on a general mix for precast applications. The fibres utilised in this study are made from
a polypropylene-based material with characteristics provided in Table 2 and presented
in Fig. 1. The fibres were introduced at three ratios in order to compare the effect of
fibre quantity on the post-cracking behaviour. Each of the 3 required batches was
0.13 m3 with the batch quantities shown in Table 3.
The specimens were cast into rectangular prisms measuring 100 100 340 mm.
This size was selected in order to best fit them into the small furnace whilst still
complying with AS 1012.11-2014 [19]. For each fibre dosage, cylinders of size Ø 100
200 mm were also cast to assess the compressive strength. Each specimen was
named using the fibre % and target surface temperature. E.g. “0.4-200-a” designates
0.4% fibre, 200 °C target surface temperature and specimen a of the two duplicate
specimens.
142 O. Mirza et al.
Table 1. Mix proportion

Component Dosage (kg/m3)
Cement (30% fly ash) 550
Coarse aggregate 1100
(50% 20 mm, 50% 10 mm)
Nepean paving sand (coarse) 472
Newcastle sand (fine) 82
Water 188
Water-cement ratio * 0.35
Table 2. Properties of the macro synthetic fibres [18]

Fibre Primary component Length Diameter Tensile Elastic Specify
type (mm) (mm) strength modulus gravity
(MPa) (GPa) (kg/m3)
BC48 Virgin polypropylene 48 0.7 640 12 890–910
Table 3. Batch quantities

Batch number Fibre % Fibre (g) Dosage (kg/m3) HRWR (mL)
1 0 0 0 0
2 0.4 490 3.77 127
3 0.8 970 7.46 253
Fig. 1. Barchip BC48 macro synthetic fibre reinforcement
2.2 Experimental Tests

The specimens were wrapped in an insulating blanket with only one side exposed as
shown in Fig. 2. This arrangement ensures they are exposed on only one surface in
order represent a precast panel for tunnel lining. The specimens were then placed into
the furnace and exposed to temperatures of 200 °C, 400 °C, 600 °C or 800 °C. These
temperatures were measured by the thermocouple tied to the specimen surface as
shown in Fig. 2. The rate of increase in temperature was not set to any standard curve
but rather controlled by the power of the small furnace. Furnace operation was also
required to be controlled by the air temperature in the furnace, not the desired surface
temperature. This required the operation to be manual and the furnace was turned off by
the operator when the measured temperature on the specimen surface reached the target
temperature. After exposure to the elevated temperatures, the specimens were allowed
to cool to the testing laboratory room temperature of *23–25 °C.
The flexural tensile strength of the specimens was determined by performing a
modulus of rupture test as presented in Fig. 3. The test setup adopted for the flexural
tests is a modified arrangement based on the methods detailed in AS 1012.11-2014 [19]
and ASTM C1609M-10 [20]. This ensures the accuracy and reliability of the experi-
mental results herein presented. In addition to the requirements of the standards, the
displacement of the specimens at the midspan was measured with a linear poten-
tiometer as shown in Fig. 3 in order to view the load-deflection behaviour. The
cylinders were tested to determine the compressive strength using the method detailed
in AS 1012.9-2014 [21].
The flexural strength of the macro synthetic fibre reinforced concrete specimens
was evaluated from the experimental data through Eq. (1):
PLð1000Þ
fcf ¼ ð1Þ
BD2
where fcf is the modulus of rupture (MPa), P is the applied force (kN), L is the span
length (mm) and B and D are the average width and average depth of the specimen
(mm) at the section of failure respectively.
Fig. 2. Furnace setup

144 O. Mirza et al.
Fig. 3. Flexural test setup. AS 1012.11 [19] (left) and actual (right)
3.1 Concrete Properties

Testing began 31 days after casting the specimens firstly with the ultimate compressive
strength (fc). The results from these compressive tests are shown in Table 4. One of the
specimens tested for batch 3 had a slightly lower strength which reduced the average
for that batch. The variation in strength for all other specimens is considered negligible
(± < 2%) which is consistent with previous research that concluded the compressive
strength is not influenced by the inclusion of fibres at these ratios.
Table 4. Concrete compressive strength (31 days)

Batch number Fibre % Compressive strength Average strength
(MPa) (MPa)
1 0 47.9 47.3
48.0
2 0.4 48.6 48.2
49.0
3 0.8 47.3 44.7
41.6*
3.2 Furnace Temperature Results

For each target temperature, the actual surface exposure temperature was recorded.
A summary of all the peak temperatures is presented in Table 5. The temperature
recorded for all specimens showed a similar curve to that presented in Fig. 4. The
temperature for each specimen peaked at approximately the target temp before being
switched off and allowed to cool. Some variance between the target temperature and the
measured temperature are noted due to operator error when switching off the furnace.
Table 5. Measured peak surface temperatures

Specimen Peak surface Specimen Peak surface Specimen Peak surface
temp (oC) temp (oC) temp (oC)
0-200-a 201 0.4-200-a 217 0.8-200-a 215
0-200-b 153 0.4-200-b 212 0.8-200-b 307
0-400-a 409 0.4-400-a 419 0.8-400-a 405
0-400-b 412 0.4-400-b 425 0.8-400-b 407
0-600-a 603 0.4-600-a 604 0.8-600-a 605
0-600-b 604 0.4-600-b 604 0.8-600-b 603
0-800-a 800 0.4-800-a 806 0.8-800-a 802
0-800-b 801 0.4-800-b 801 0.8-800-b 802
Fig. 4. Surface exposure temperature results for the 400 °C and 800 °C specimens
3.3 Flexural Test Results

The pre-cracking flexural strengths for all 32 specimens are presented in Fig. 5. The
results show that the fibre percentage has negligible influence on the pre-cracking
flexural strength. However, the important benefit of incorporating macro synthetic
fibres into concrete is for improved post-cracking characteristics. Figure 6 presents the
load-deflection results from the flexural strength tests for the six ambient temperature
specimens. The specimens with 0.4% fibre showed ductile post-cracking behaviour and
the specimens 0.8% showed similar behaviour with a much greater residual strength.
146 O. Mirza et al.
Fig. 5. Pre-cracking flexural strength (cracking stress)
Fig. 6. Flexural test results for ambient temperature specimen
By recording the load at a deflection of 2 mm (1/150 deflection), the post-cracking

strengths of the specimens can be compared. Figure 7 presents the post-cracking flexural
strength of all specimens where the inclusion of fibres is seen to significantly increase the
post-cracking performance of the concrete. The specimens without fibres show negligible
post cracking strengths as expected. The specimens with 0.4% fibre show post-cracking
strengths between approximately 2–3 MPa for temperatures 600 °C. The specimens
with 0.8% fibre display much higher post-cracking strengths than the specimens with
0.4% fibre for temperatures 400 °C. However, for strengths after exposure to 600 °C
and 800 °C, far less variance was observed.
Fig. 7. Post-cracking flexural strength (residual stress)
It can also be observed that there is a much greater variance between the some of
the duplicate a-b specimens for the post-cracking strengths in Fig. 7 compared to the
more consistent pre-cracking strengths in Fig. 5. The 0.4–25 and the 0.8–25 specimens
had variances of 2.2 MPa (67%) and 1.2 MPa (54%) respectively. This is attributed to
the variances in the distribution of fibres near the crack location. The post-cracking
flexural strength of the specimen is noted to be highly dependent on the location of the
fibres. This impact is more significant for the small specimens required by the current
study as some specimen had just a few fibres bridging the crack.
3.4 Design Model

Based on the DBV-Guide to Good Practice for steel fibre reinforced concrete [22], a
design model can be proposed for MSFRC. The model proposes that the flexural tensile
strength of MSFRC for surface exposure temperatures less than 150 °C is equal to the
strength at ambient temperature. After 150 °C, the strength is decreased linearly to zero
at 700 °C. This is presented as the ratio fct(T)/fct(25) in Fig. 8 where fct is the flexural
tensile strength, T is the surface exposure temperature and fct(25) is the strength at
ambient temperature.
Utilising the lower of the strengths measured for the ambient temperature speci-
mens, the proposed design models for 0.4% and 0.8% fibre can be applied and pre-
sented in Fig. 9. The test results exceed the model for 95% of the tested specimens and
the model is conservative at temperatures 600 °C.
It is noted that the DBV-Guide to Good Practice [22] assumes that the temperature
is consistent through the depth of the section; an assumption not valid in this study. The
temperature spread through the depth of the specimen in this study was not measured
but was likely limited to close to the surface due to the short exposure duration.
However, the guide still serves as the base for proposing a simplistic design model for
the post-cracking residual strength of MSFRC after short-term exposure to elevated
temperatures.
148 O. Mirza et al.
Fig. 8. Proposed design model
a) 0.4% fibre b) 0.8% fibre
Fig. 9. Comparison of post-cracking strengths and proposed design model
4 Conclusions
In this paper, the effect of macro synthetic fibre reinforcement on the post-cracking
performance of concrete exposed to elevated temperatures has been explored. The
following conclusions have been drawn from this study:
• The addition of fibres did not noticeably influence the ultimate compressive strength
of the concrete or the pre-cracking flexural strength of the specimens. The speci-
mens with no fibres showed a negligible post-cracking flexural strength. This is
consistent with preliminary expectations and conclusions from previous research.
• In general, the specimens with 0.8% fibre exhibited better performance than the
0.4% fibre specimens in terms of residual (i.e. post-cracking) flexural strength and
ductility.
• The results presented in this paper are utilised to propose a design model based on
the short term exposure to elevated temperatures on only one surface. It is
acknowledged that the temperature inside the specimens is not uniform and the
mechanical properties vary based on distance to the exposure surface. Further
research utilising longer exposure times and measuring the temperature spread
throughout the depth of the section is recommended to further validate the model.
The fibre condition at different depths of the specimens will be investigated in
subsequent research.
• The post-cracking flexural strength of the specimen was quite variable and was
noted to be highly dependent on the placement of the fibres. Variances in the
distribution of fibres near the crack location significantly impacted the strength.
More confidence in the results could be achieved by conducting further study using
a larger number of specimens and specimens with a greater width to reduce the
impact of variable fibre distribution.
Acknowledgements. The authors gratefully acknowledge Western Sydney University’s School

of Computing, Engineering and Mathematics and the Institute for Infrastructure Engineering lab
staff for their support to the author’s work described herein.
References
1. Li, V., Stang, H., Krenchel, H.: Micromechanics of crack bridging in fibre-reinforced
concrete. Mater. Struct. 26(8), 486–494 (1993)
2. Al-Masoodi, A.H.H., Kawan, A., Kasmuri, M., Hamid, R., Khan, M.N.N.: Static and
dynamic properties of concrete with different types and shapes of fibrous reinforcement.
Constr. Build. Mater. 104, 247–262 (2016)
3. Yao, W., Li, J., Wu, K.: Mechanical properties of hybrid fiber-reinforced concrete at low
fiber volume fraction. Cem. Concr. Res. 33, 27–30 (2003)
4. Xu, Z., Hao, H., Li, H.N.: Experimental study of dynamic compressive properties of fibre
reinforced concrete material with different fibres. Mater. Des. 33, 42–55 (2012)
5. Balouch, S.U., Forth, J.P., Granju, J.L.: Surface corrosion of steel fibre reinforced concrete.
Cem. Concr. Res. 40(3), 410–414 (2010)
6. Berrocal, C.G., Lundgren, K., Löfgren, I.: Corrosion of steel bars embedded in fibre
reinforced concrete under chloride attack: state of the art. Cem. Concr. Res. 80(C), 69–85
(2016)
7. Marcos-Meson, V., Michel, A., Solgaard, A., Fischer, G., Edvardsen, C., Skovhus, T.:
Corrosion resistance of steel fibre reinforced concrete-a literature review. Cem. Concr. Res.
103, 1 (2018)
8. Anandan, S, Manoharan, S.V., Sengottian, T.: Corrosion effects on the strength properties of
steel fibre reinforced concrete containing slag and corrosion inhibitor. (Research Article)
(Report) (2014)
9. Berrocal, C.G., Löfgren, I., Lundgren, K., Tang, L.: Corrosion initiation in cracked fibre
reinforced concrete: Influence of crack width, fibre type and loading conditions. Corros. Sci.
98(C), 128–139 (2015)
10. Hwang, J., Jung, M.S., Kim, M., Ann, K.Y.: Corrosion risk of steel fibre in concrete. Constr.
Build. Mater. 101, 239–245 (2015)
150 O. Mirza et al.
11. Blunt, J., Jen, G., Ostertag, C.P.: Enhancing corrosion resistance of reinforced concrete
structures with hybrid fiber reinforced concrete. Corros. Sci. 92, 182–191 (2015)
12. Lee, S.-J., Won, J.-P.: Resistibility of structural nano-synthetic fibre-reinforced cementitious
composites in various chemical and physical environments. Compos. Struct. 179, 495–501
(2017)
13. Simões, T., Costa, H., Dias-Da-Costa, D., Júlio, E.: Influence of fibres on the mechanical
behaviour of fibre reinforced concrete matrixes. Constr. Build. Mater. 137, 548–556 (2017)
14. Rambo, D.A.S., Blanco, A., de Figueiredo, A.D., dos Santos, E.R.F., Toledo, R.D., Gomes,
O.F.M.: Study of temperature effect on macro-synthetic fiber reinforced concretes by means
of Barcelona tests: an approach focused on tunnels assessment. Constr. Build. Mater. 268,
443–453 (2018)
15. Choumanidis, D., Badogiannis, E., Nomikos, P., Sofianos, A.: ‘Barcelona test for the
evaluation of the mechanical properties of single and hybrid FRC, exposed to elevated
temperature. Constr. Build. Mater. 138, 296–305 (2017)
16. Sukontasukkul, P., Pomchiengpin, W., Songpiriyakij, S.: ‘Post-crack (or post-peak) flexural
response and toughness of fiber reinforced concrete after exposure to high temperature.
17. Serafini, R., Dantas, S.R.A., Salvador, R.P., Agra, R.R., Rambo, D.A.S., Berto, A.F., de
Figueiredo, A.D.: Influence of fire on temperature gradient and physical-mechanical
properties of macro-synthetic fiber reinforced concrete for tunnel linings. Constr. Build.
Mater. 214, 254–268 (2019)
18. BarChip. 20 Nov 2019. https://barchip.com/product/
19. Standards Australia, Methods of testing concrete–Method 11: Determination of the modulus
of rupture, AS 1012.11–2014, SAI Global (2014)
20. ASTM 2010, ASTM C1609/C1609M-10, Standard Test Method for Flexural Performance of
Fibre-Reinforced Concrete (Using Beam With Third Point Loading), ASTM International.
http://www.astm.org/cgi-bin/resolver.cgi?C1609C1609M-19
21. Standards Australia, Methods of testing concrete-Method 9: Compressive strength tests-
Concrete, mortar and grout specimens, AS 1012.9–2014, SAI Global (2014)
22. GSCCT (ed.).: DBV-Guide to Good Practice, Steel Fibre Concrete, German Society for
Concrete and Construction Technology, Kurfurstenstrabe 129 10785 Berlin Germany (2001)
Experimental Investigation on the Cyclic
Behaviour of Steel Fibre Reinforced Concrete
Under Bending
Maure De Smedt(&), Rutger Vrijdaghs, Els Verstrynge,

Kristof De Wilder, and Lucie Vandewalle
Department of Civil Engineering, KU Leuven, Leuven, Belgium

maure.desmedt@kuleuven.be
Abstract. This paper presents the results of an experimental program investi-

gating the monotonic and cyclic flexural behaviour of steel fibre reinforced
concrete (SFRC). Three-point bending tests are assessed with acoustic emission
(AE) monitoring to detect concrete micro-cracking and damage propagation
during the different stages of cyclic loading. LVDTs over the prism’s height
investigate the evolution of the neutral axis position. Three loading patterns are
applied, namely monotonic, progressive cyclic and variable cyclic. 3D and 5D
hooked end steel fibres are both used in a content of 20 kg/m3 and 40 kg/m3,
representing a softening and a hardening behaviour respectively. By combining
traditional and advanced measurement methods, the developed test setup allows
a better understanding of the cyclic behaviour of SFRC. The monotonic
envelope curves agree with the cyclic stress-CMOD curves. The neutral axis
position appears to be independent of the SFRC classification. Fatigue damage
development occurs for multiple load cycles at high load limits. Furthermore,
AE monitoring proves to be complementary to mechanical testing for evaluating
damage accumulation in SFRC under bending.
Keywords: Steel fibre reinforced concrete Cyclic behaviour Damage

assessment Acoustic emission
1 Introduction
While steel fibres have shown a positive influence on the concrete’s fatigue behaviour
[1–3], research has been mainly focussed on the seismic behaviour of elements with
both conventional reinforcement and fibres. Moreover, there is still a lack in cyclic
building design codes [1, 2, 4]. Three-point bending tests (3PBT) are widely performed
to study the post-cracking behaviour of SFRC [5]. However, solely measuring load and
displacement rapidly becomes restricted to obtain a more profound damage assessment.
Therefore, advanced measurement methods are required.
Acoustic emission (AE) monitoring is such a method to detect and localise early
stages of micro-cracking, by recording the elastic waves generated by a strain energy
release within the material [6, 7]. This non-destructive technique benefits from the
continuous measurement, although interpretation must be done carefully to relate AE
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_14
152 M. De Smedt et al.
data to possible sources [6, 8, 9]. Previous research proved the adequacy of AE
monitoring on SFRC elements in monotonic bending tests [7, 10–13].
While SFRC, cyclic loading and AE monitoring have been studied before, the
combination of all three together is a novelty of this research. The aim is to compare
monotonic and cyclic bending behaviour and to discuss the influence of fibre type and
content. AE monitoring assesses the damage development by combined AE and
mechanical parameters evolution. This research is relevant for applications in which
cyclic loading is an important design parameter, such as bridges and infrastructure.
2 Experimental Research and Methods
Three types of tests were performed, namely (1) monotonic, (2) progressive cyclic and
(3) variable cyclic 3PBTs. Four groups of specimens were investigated, with varying
fibre type and dosage. For each group, three specimens were tested in each type of test,
leading to nine specimens per group.
2.1 Materials and Test Specimens

The concrete composition was identical for all specimens. It belonged to concrete class
C40/50 and had a W/C-ratio of 0.5. The mixture existed of 350 kg/m3 CEM I 52.5 R
HES, 835 kg/m3 sand 0/4, 1099 kg/m3 gravel 4/14, 175 kg/m3 water and 1 kg/m3
(0.3%) superplastizer Glenium 51. The measured mean cube compressive strength
equalled 59 MPa, with a standard deviation of 3.5 MPa (tested on 28 days according to
European Standard EN 12390-3 [14]).
Two steel fibre types were used in this research, namely Dramix 3D-80/60-BG
(single end-hook) and 5D-65/60-BG (double end-hook). Table 1 shows the fibre
properties as provided by the manufacturer. Both fibre types were used in two dosages,
namely 20 kg/m3 (0.25 V%) or 40 kg/m3 (0.5 V%). The fibre dosages were chosen in
order to investigate both softening (0.25 V%) and hardening (0.5 V%) behaviour under
bending loads, as determined in [15]. The combination of fibre types and dosages
resulted in four tested SFRC groups, namely 3D20, 3D40, 5D20 and 5D40.
The tested SFRC prisms (150 150 660 mm3, according to EN 14561 [5]) had a
notch of 5 mm wide and 25 mm high in the middle of the lower side of the prism in
order to initiate the crack location. Each specimen is indicated by the fibre type and
dosage and a number varying from 1 to 9 (e.g. 3D206).
Table 1. Properties of the steel fibres.

Steel fibre Length Diameter Tensile strength Elastic modulus
[mm] [mm] [MPa] [GPa]
3D-80/60-BG 60 0.75 1125 200
5D-65/60-BG 60 0.92 2300 200
Experimental Investigation on the Cyclic Behaviour of Steel Fibre 153
2.2 Test Setup and Loading Patterns

Each SFRC prism was subjected to a 3PBT, according to EN 14651, with a hydraulic
press (Dartec, max. capacity of 5 MN). The test setup is shown in Fig. 1. The testing
machine measured the load and the displacement of the midspan during the test.
A linear variable differential transformer (LVDT) registered the crack mouth opening
displacement (CMOD) at the bottom of the notch. Furthermore, five LVDTs along the
height of the prism measured the deformation profile of the cracked section during
loading. By assuming a linear fit through these measurements, the position of the
neutral axis can be determined [16].
Three loading patterns were performed on each group of specimens. The first test
type was a monotonic CMOD-controlled 3PBT. The CMOD increased with
0.05 mm/min until 0.1 mm, followed by a rate of 0.2 mm/min until 4 mm. The second
test type was a progressive cyclic CMOD-controlled 3PBT. The monotonic loading
pattern was extended with 12 progressive loading cycles at CMOD equal to 0.1, 0.2,
0.3, 0.4, 0.5, 1, 1.5, 2, 2.5, 3, 3.5 and 4 mm. The unloading and reloading rates
equalled 0.2 mm/min and a minimum load of 1 kN was kept. The third test type was a
variable cyclic 3PBT, where multiple load cycles with variable frequency and load
limits were applied. By performing the three different loading patterns, cyclic effects
can be distinguished from general monotonic behaviour and both progressive and
fatigue damage development can be investigated.
Fig. 1. Overview of the test setup.
2.3 Acoustic Emission (AE) Monitoring and Analysis

The AE technique has been applied to study the internal damage development. Four
piezoelectric 150 kHz resonance sensors were attached to one side of the specimen’s
surface (refer to Fig. 1). Each sensor was connected to a 4-channel Vallen AMSY-5
acquisition system, using a preamplifier with 34 dB gain. In the acquisition system, a
frequency filter of 50–850 kHz and an amplitude threshold of 38.2 dB were set to
avoid background noise. During post-processing, an additional amplitude filter was set
on 50 dB. From pencil lead breaks, the maximal localisation uncertainty was expected
as 20 mm.
The AE signals are investigated by a parameter-based approach. The AE activity

presents the amount of recorded localised AE events. Based on the first hit of each
event, the amplitude (dB), the frequency (kHz) and the source localisation of the signal
are studied. The AE activity rate is an indicator for the degree of damage development
within the specimen.
2.4 Damage Evolution

For subsequent loading cycles, the damage evolution can be characterised by the
stiffness response [13, 17, 18]. It is calculated as the slope of the unloading branch, as
indicated on Fig. 2. The damage D for each loading cycle is expressed as Eq. (1), with
i the cycle number, K0 the undamaged slope and Ki the slope of the ith cycle. D varies
from 0 (undamaged material) to 1 (failed material).
Di ¼ 1 K i =K 0 ð1Þ
In case that multiple load cycles with different upper and lower limits are applied, a
decrease of damage can be seen while increasing the load limits. This apparent
recovery of damage or regain of stiffness is caused by the non-linearity of the unloading
path, as indicated by the example of Ka and Kb on Fig. 2. Higher load limits results in a
steeper slope and therefore smaller damage D. To correct for this non-linearity, the
slope could be calculated between the overlapping load limits of all loading parts.
Fig. 2. Schematic presentation of the stiffness K and the influence of the non-linear unloading
path on the damage parameter.
Firstly, monotonic and progressive cyclic testing are compared on both mechanical
characteristics and AE properties, and the fibre influence is discussed. Secondly, a
variable cyclic loading test is elaborated.
3.1 Comparison Between Monotonic and Progressive Cyclic Behaviour

3.1.1 Load-Displacement Behaviour
Figure 3 shows the average stress-CMOD curves for both monotonic and progressive
cyclic loading. In general, the cyclic curves agree with the monotonic envelope curves
[19, 20]. The load cycles do not significantly influence the post-peak behaviour, except
for 5D40. However, this difference is attributed to the experimental scatter, as all
prisms were made from the same concrete batch on the same day. The cyclic scatter
and the lower boundary of the monotonic results are therefore shown as well.
As confirmed in literature, fibres mainly influence the post-peak behaviour, not the
maximal stress. The flexural tensile strength varies between 4.0 and 4.2 MPa, inde-
pendent of fibre dosage or loading pattern. Double end-hooked fibres lead to an
enhanced post-peak behaviour, as the residual strength class (according to Model Code
2010 [21]) varies from 1b to 1c for 3D20 to 5D20, and 1.5b to 1.5e for 3D40 to 5D40.
A higher fibre content increases the post-peak strength as well.
Fig. 3. Average monotonic and progressive cyclic load-displacement behaviour.

3.1.2 Position of the Neutral Axis

Figure 4 presents the average neutral axis evolution during monotonic and cyclic tests,
with y = 0 mm corresponding to the bottom of the prism. Amongst all SFRC groups
with different residual strength classes, there appears to be no significant influence on
the neutral axis position. For all tests, there is a small compressive zone, as the neutral
axis lies close to the top of the prisms. At a CMOD of 0.5 mm, the average position
equals 137.8 mm (COV of 2.7%) for all monotonic tests and 140.1 mm (COV of
1.1%) for all cyclic tests. After a CMOD of 0.5 mm, 90% of the section remains in
tension.
Fig. 4. Average the neutral axis position in monotonic and progressive cyclic tests.
The neutral axis shifts upwards in case of cyclic loading. A small difference in
position occurs between a CMOD of 0.08 and 0.3 mm. Applying load cycles during
matrix cracking and stress transfer to the fibres leads to unrecoverable internal changes
and a slightly adapted stress profile in the cracked section. Furthermore, unloading
decreases the compressive zone, reloading increases the compressive zone again until
its height before unloading.
3.1.3 AE Activity
Figure 5 provides the cumulative AE events. Double end-hooked fibres or an increased
fibre content lead to more AE activity, due to the additional matrix damage during
straightening of the fibres. The total amount of AE events is comparable for both
loading patterns. During unloading, almost no AE events are recorded. For 5D fibres,
an unloading-reloading cycle is only a small part of the total AE events. For 3D fibres,
a larger increase in AE events starting at the first cycles is seen. The highest increase in
AE events occurs right after reaching the peak stress, around a CMOD of 0.08 mm, due
to concrete matrix cracking. During transfer of the stresses to the fibres, less AE events
are recorded. A second rise of AE events occurs at a CMOD of 0.15 mm, when the
fibres are being straightened. Thereafter, AE activity gradually develops with growing
CMOD, when fibres are subjected to frictional sliding.
Fig. 5. Cumulative AE events in monotonic and progressive cyclic tests.
3.1.4 Damage Evolution

For the progressive cyclic loading tests, the averaged damage evolution (Eq. (1)) for all
SFRC groups is presented in Fig. 6. The coefficients of variation for each loading cycle
range from 0.2 to 10.8%. The largest increase of damage occurs during the peak phase,
Fig. 6. Average damage evolution during progressive cyclic loading.
namely at load cycles with a CMOD smaller than 0.5 mm. Thereafter, each cycle
gradually increases the damage. An increase of fibre dosage leads to a less rapid
damage increase. Double end-hooked fibres have a faster damage evolution than single
end-hooked fibres. More concrete matrix damage while straightening is caused by the
double end-hook and the larger diameter and tensile strength, as also indicated by the
higher AE activity. However, this damage is only induced at higher stress levels
compared to the 3D fibres. SFRC groups 3D20 and 5D40 show an almost identical
evolution.
3.2 Example of Variable Cyclic Loading

As an example of the variable cyclic loading tests, specimen 3D406 is elaborated in
Fig. 7(a), (b). After reaching the peak stress around a CMOD of 0.08 mm (based on the
monotonic behaviour), four consecutive parts of multiple load cycles at 0.04 Hz are
applied, starting at a CMOD of 0.1 mm. The load limits are expressed as a percentage
of the flexural tensile strength of 4.1 MPa. Since the specimen demonstrates a hard-
ening behaviour, the load limits are gradually increased, leading to:
– Part (1): 50 cycles between 90 and 44%
The increased post-peak strength can be attributed to the scatter on fibre dispersion
within the specimens. Figure 7(c) shows the cumulative AE activity. The total amount
of AE events at the end of the test (CMOD = 4 mm) is 10702, which is a 41% increase
compared to the average monotonic and progressive cyclic behaviour. Increasing the
number of load cycles increases the amount of AE events.
During part (1) and (2), the increase of CMOD is limited, as well as the increase of
AE activity. Only 8% of AE activity during cyclic loading is detected in these parts,
although representing 120 load cycles. The load limits are too low to induce detectable
damage. During part (4), CMOD and AE activity clearly increase with each loading
cycle, despite the constant load limits. 75% of the AE activity is detected in this part,
which represents only 39 load cycles. Both mechanical parameters and AE activity
display characteristics of fatigue damage.
Figure 7(d) presents the damage evolution during the multiple load cycles with
respect to time. It is observable that the increase of AE activity and of damage similarly
evolves during the loading parts (refer to Fig. 7(c) and (d)). These tests confirm that
mechanical testing and AE monitoring are complementary techniques for evaluating
damage accumulation, which opens interesting possibilities for situations where the
characteristics are difficult to measure, for example on-site.
Fig. 7. Variable cyclic loading on specimen 3D406: (a) stress-time curve, (b) stress-CMOD
curve, (c) cumulative AE events and (d) damage evolution.
Lastly, Fig. 8 presents the damage parameter in function of CMOD and compares
to the average damage evolution of the 3D40 group. Figure 8(a) does not correct for
the non-linearity of the unloading path (as discussed in Fig. 2), resulting in an apparent
recovery of damage for each transition to the next loading part. Figure 8(b) calculates
the damage between the overlapping load limits of 90 and 68%. In general, the evo-
lution corresponds to the average 3D40 damage curve. The damage increase with each
loading cycle is the most pronounced during part (4), again indicating the occurring
fatigue behaviour. Multiple load cycles at too low load limits do not lead to detectable
fatigue damage since AE activity, CMOD and damage only increase for load cycles at
high load limits.
Fig. 8. Damage evolution for 3D406 during the load cycles: (a) based on the whole unloading
part, and (b) based on the overlapping unloading part.
4 Conclusions
This paper experimentally investigated monotonic and cyclic behaviour of SFRC in

3PBT assessed by AE monitoring. The main research conclusions are summarised as
follows.
• The progressive cyclic behaviour is in agreement with the monotonic envelope
stress-strain curves. However, the results are sensitive to the experimental scatter of
SFRC testing. Despite different SFRC post-cracking behaviour, the neutral axis
position appears to be largely independent on the residual strength classification.
Applying load cycles slightly changes the deformation profile, the neutral axis
increases and decreases during unloading and reloading.
• AE events are maximised at matrix cracking and reaching the maximal stress.
Thereafter, a second increase occurs when straightening the fibres. Higher fibre
dosages or double end-hooks increase the total event amount. During an unloading
branch, almost no AE events are recorded. Applying multiple load cycles increases
the event amount as well.
• An increase of fibre dosage leads to less rapid damage increase and double end-
hooked fibres have a faster damage evolution than single end-hooked fibres. Cor-
rection for the non-linear unloading path is needed to avoid an apparent regain of
stiffness while increasing the load limits.
• A relation between damage evolution and AE activity was found for multiple cyclic
loading tests. All parameters indicate the occurrence of detectable fatigue damage
development for load cycles at high load limits. The AE measurements confirm the
mechanical characteristics.
Acknowledgements. The authors gratefully acknowledge the financial support of Research

Foundation Flanders (FWO, PhD-grant no. 1S32717N), and the supply of steel fibres used in this
study by Bekaert nv.
References
1. Lee, M., Barr, B.: An overview of the fatigue behaviour of plain and fibre reinforced
concrete. Cem. Concr. Compos. 26(4), 299–305 (2004)
2. Parvez, A., Foster, S.: Fatigue behaviour of steel-fiber-reinforced concrete beams. J. Struct.
Eng. 141(4), 04014117 (2014)
3. Wilkes, W.: Fatigue concrete vs. steel. PCI J. 34(4), 76–79 (1989)
4. di Prisco, M., Plizzari, G., Vandewalle, L.: ‘Fibre reinforced concrete: new design
perspectives. Mater. Struct. 42, 1261–1281 (2009)
5. CEN: EN 14651: Test method for metallic fiber concrete – measuring the flexural tensile
strength (2014)
6. Grosse, C., Ohtsu, M.: Acoustic Emission testing. Springer (2008)
7. Behnia, A., Chai, H., Shiotani, T.: Advanced structural health monitoring of concrete
structures with the aid of acoustic emission. Constr. Build. Mater. 65, 282–302 (2014)
8. Noorsuhada, M.: An overview on fatigue damage assessment of reinforced concrete
structures with the aid of acoustic emission technique. Constr. Build. Mater. 112, 424–439
(2016)
9. Wevers, M.: Listening to the sound of materials: acoustic emission for the analysis of
material behaviour. NDT&E Int. 30(2), 99–106 (1997)
10. Aggelis, D., et al.: Acoustic emission characterization of the fracture process in fibre
reinforced concrete. Constr. Build. Mater. 25, 4126–4131 (2011)
11. Soulioti, D., et al.: Acoustic emission behavior of steel fibre reinforced concrete under
bending. Constr. Build. Mater. 23, 3532–3536 (2009)
12. Behnia, A., et al.: Integrated non-destructive assessment of concrete structures under flexure
by acoustic emission and travel time tomography. Constr. Build. Mater. 67(part B) 202–215
(2014)
13. Nguyen-Tat, T., et al.: Characterization of damage in concrete beams under bending with
acoustic emission technique (AET). Constr. Build. Mater. 187, 487–500 (2018)
14. CEN: NBN EN 12390-3: Testing hardened concrete - Part 3: Compressive strength of test
specimens (2009)
15. De Wilder, K., et al.: Experimental analysis of the mechanical behaviour of shear-deficient
pretensioned steel-fibre reinforced concrete beams. In: Banthia, N., di Prisco, M. (eds.)
Proceedings of the 9th RILEM International Symposium in Fiber Reinforced Concrete
(BEFIB), Canada, pp. 1248–1262 (2016)
16. Vrijdaghs, R., Van Itterbeeck, P., De Smedt, M., Vandewalle, L.: Experimental study into
the location of the neutral axis in fiber reinforced concrete prisms. Struct. Concr. (2020)
17. Gebuhr, G., et al.: Deterioration development of steel fibre reinforced high performance
concrete in low-cycle fatigue. In: Zingoni, A. (ed.) Proceedings of the 7th International
Conference on Structural Engineering, Mechanics and Computation (SEMC), South Africa,
pp. 1444–1449 (2019)
18. Boulekbache, A., Hamrat, M., Chemrouk, M., Amziane, S.: Flexural behaviour of steel fibre-
reinforced concrete under cyclic loading. Constr. Build. Mater. 126, 253–262 (2016)
19. Chanvillard, G., Pimienta, P., Pineaud, A., Rivillon, P.: Fatigue flexural behavior of pre-
cracked specimens of Ductal® UHPFRC. In: di Prisco, M., Felicetti, R., Plizzari, G.A. (eds.)
Proceedings of the 6th RILEM International Symposium in Fiber Reinforced Concrete
(BEFIB), Italy (2004)
20. Ulfkjaer, J.P.: Low cycle fatigue of ultra-high performance steel fibre reinforced concrete. In:
Middendorf, B., Fehling, E., Wetzel, A. (eds.) Proceedings of the 5th International
Symposium on Ultra-High Performance Concrete and High Performance Construction
Materials (HIPerMat), Germany, pp. 113–114 (2020)
21. Fédération Internationale du Béton (fib): fib Model Code for Concrete Structures 2010,
(Wilhelm Ernst und Sohn Verlag für Architektur, Berlin, Germany (2013)
Effect of Test Setups on the Shear Transfer
Capacity Across Cracks in FRC
Alejandro Giraldo Soto(&) and Walter Kaufmann
Chair of Concrete Structures and Bridge Design,

Institute of Structural Engineering, ETH Zürich, Zürich, Switzerland
giraldo@ibk.baug.ethz.ch
Abstract. The shear transfer capacity of concrete across cracks is highly rel-
evant in situations where the principal concrete stress directions are not aligned
with the cracks. Fibres are effective in controlling the crack opening, thereby
enhancing aggregate interlock and hence, the ability of transferring shear
stresses across cracks. Compared to plain concrete, higher stresses can therefore
be transferred across cracks in fibre reinforced concrete (FRC). However, the
shear transfer across cracks in FRC has received much less attention over the
past decades than the residual tensile stress transfer across orthogonally opening
cracks, and no generally accepted model for this behaviour is available today.
The shear transfer capacity and the calibration of most existing models are
obtained from experimental tests which presume failures occurring in “pure
shear”, being the most popular used test setups: (i) Z-type push-off specimen;
(ii) modified JSCE-G 553; and (iii) FIP shear test method (asymmetrical four-
point bending). However, significant differenced have been observed in the
experimental resulting shear strength, depending on the test setup. These dif-
ferences have not been evaluated systematically until now. In order to address
this issue, the authors carried out an experimental campaign on specimens made
from identical SFRC mixes with varying fibre dosage, testing each mix in all
three mentioned setups. The paper presents the results of this experimental
campaign.
Keywords: Steel fibre reinforced concrete Shear transfer Interface shear

Experimental testing Push-off test JSCE shear test FIP shear test Cracks
Crack kinematics Aggregate interlock Shear test setup
1 Introduction
Contrary to the residual tensile stresses in orthogonally opening cracks in fibre rein-
forced concrete (FRC), the transfer of shear stresses across cracks in FRC has received
only limited attention over the past decades, and no generally accepted model is
available for the case of skew crack opening in FRC. However, this is highly relevant
in situations where the principal concrete stress directions are not aligned with the
cracks, such as in orthotopically reinforced panels and webs of girders. In such ele-
ments, the principal concrete compressive stress direction rotates towards the stronger
reinforcement after the onset of yielding, causing substantial deviations from the
direction of initial cracks. Furthermore, cracks deviating from the principal
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_15
164 A. Giraldo Soto and W. Kaufmann
compressive stress direction may be caused by restrained shrinkage or previous, non-

proportional loading.
Fibres are effective in controlling the crack opening, thereby enhancing aggregate
interlock and hence, the ability of transferring shear stresses across cracks. Compared
to plain concrete, higher stresses can therefore be transferred across cracks in fibre
reinforced concrete.
Different experimental setups have been used in the past to examine the shear
transfer across cracks in FRC [1–11]. While apparently simple to conduct, such
experiments are demanding since large forces are required, and very small displace-
ments of the irregular crack faces need to be controlled and measured. The state of
stress in the interface is hardly ever uniform, even in complex experimental setups, due
to the irregularity of the crack faces and the difficulty of generating shear without
concomitant bending moments. In addition, the fibre distribution and orientation in the
interface are also subject to variations. Hence, test results are subject to considerable
scatter. Furthermore, a recent study aimed at establishing a mechanical model for the
shear transfer across cracks in FRC, relating the shear and normal stresses to the crack
kinematics, i.e. crack opening and slip [12], identified substantial differences in the
results obtained from different experimental setups used for determining the “shear
strength” of FRC.
In order to address this issue, the authors carried out an experimental campaign on
specimens made from identical SFRC mixes, tested in three different setups adopted in
previous experimental studies investigating the “shear strength” of FRC, illustrated in
Fig. 1: (a) Z-type push-off tests as known from classic investigations of aggregate
interlock in plain and reinforced concrete [13–18]; (b) modified JSCE-G 553 shear tests
[19]; and (c) FIP shear tests [20]. More details on the test setups and specimen
geometry are given in Sect. 2.
In order to study the test setups effect on the shear transfer capacity across cracks in
SFRC, three different, existing experimental test setups (Z-type push-off, modified
JSCE-G 553 and FIP shear tests), used to determine the shear resistance in plain
concrete and fibre reinforced concrete, are carried out. For this purpose, the specimens
of each type of test are made with the same concrete mixture using different amounts of
steel fibres: 0 kg/m3, 40 kg/m3, 60 kg/m3 and 80 kg/m3 (0%, 0.5%, 0.75%, 1%).
Therefore, a total of 12 shear tests, 4 tests for each test setup (3 4), were conducted.
Table 1 identifies the tests and specimens.
Effect of Test Setups on the Shear Transfer Capacity Across Cracks 165
Table 1. Tests and specimen identification.

Specimen ID qf [%] [kg/m3]
Z-Type JSCE FIP
Z-0 JSCE-0 FIP-0 0.00 0
Z-40 JSCE-40 FIP-40 0.50 40
Z-60 JSCE-60 FIP-60 0.75 60
Z-80 JSCE-80 FIP-80 1.00 80
2.1 Experimental Setups and Specimen Geometry

The experimental shear test setups and the geometry of the specimens are shown in
Fig. 1.
The Z-Type push-off specimens (Fig. 1a) consist of two reinforced L-shaped blocks
that are 300 mm wide, 474 mm high and 150 mm deep. In order to avoid cracking
outside the intended failure plane, a 15 mm deep and 2.5 mm wide notch was sawed on
the front and rear of the specimens. The shear plane has a theoretical area of
120 120 mm (14’400 mm2). The support and the applied load F are vertically
aligned in the specimen axis, having a width of 60 mm. The average shear stress s in
the shear plane is calculated as s ¼ F=Ac , where Ac is the area of the shear plane.
540
270 270
F 240 30 220 50
150 150 B = 150 mm
B = 150 mm
B = 250 mm
B = 250 mm B = 250 mm
F
150
2φ12
220 40 205 50
27 150 150 150 25
notch
297
15 120 15
3φ8 (stirrups)
250
B = 150 mm
474
notch
B = 150 mm 25 50 205 40 220
65
27
B = 250 mm
297
B = 250 mm
147,5 2,5 150 2,5 147,5 B = 250 mm
150
notch
150
B = 150 mm
B = 150 mm
B = 150 mm B = 150 mm Load cell
30
60 132,5 155 132,5 50 220 30 240

300 15 450 15
(a) Z-Type push-off (b) Modified JSCE-G 553 shear test (c) FIP shear test
Fig. 1. Shear test setups and geometry of the specimens.
The modified JSCE-G 553 (Fig. 1b) is a modified version of the standardized shear
test JSCE-G 553-2010 [19], specified by the Japanese Society of Civil Engineering
(JSCE), where notches in the two shear planes are included. The specimen consists of a
square prism of 450 mm length, with a square cross-section of 150 mm side length.
This test is composed of two shear planes, and to avoid cracking outside these intended
failure shear planes, a 15 mm deep and 2.5 mm wide notches were sawed all around
the specimen. The theoretical total area of both shear planes is twice 120 120 mm
(2 14’400 = 28’800 mm2). The specimen is supported by two relatively small
bearing surfaces of 15 mm width, located just outside of the notches. The load F is
applied symmetrically on the specimen, acting over a width of 15 mm located on the
inner side near the shear planes. The average shear stress s in the shear plane is
calculated as s ¼ F=Ac , where Ac is the sum of the area of both shear planes. Note that
due to the geometric configuration of the test – which can be seen as a four-point-
bending test with short shear span – the shear stress s is always accompanied by a
normal stress r 0.1 s (if loads and reactions act centric on the loading plates), and
furthermore, the relevant sections are subject to global bending, which accumulates
compression stresses in the upper part.
The FIP standard shear test (Fig. 1c) is performed according to the FIP report
published in 1978 [20]. The FIP specimen consists of a square prism of 540 mm
length, with a square cross-section of 250 mm side length. In order to avoid cracking
outside the intended failure shear plane, a 15 mm deep and 2.5 mm wide notch was
sawed all around the specimen, which leads to a theoretical shear area of
220 220 mm (48’400 mm2). The specimen is loaded in asymmetrical four-point
bending by means of a sophisticated arrangement of steel elements used to introduce
the load F (the load axis coincides with the shear plane in the middle of the specimen)
and to support the sample. The shear force V acting in the shear plane is equal to
V ¼ F ð220 30Þ=ð220 þ 30Þ ¼ 0:76 F, therefore the average shear stress s in the
shear plane is calculated as s ¼ 0:76 F=Ac , where Ac is the area of the shear plane.
2.2 Equipment and Tests Procedure

The tests are carried out using a Servohydraulic Universal Testing Machine with a
maximum capacity of 1600 kN. The crack and the relative displacements of the shear
planes were measured by two different optical measuring systems. Digital Image
Correlation (DIC) was used in the front face and the laser scanning, NDI system, in the
rear face. In addition, a load cell with 200 kN capacity was used for the FIP test setup in
order to measure and control the reaction on one of the two bearings (Fig. 1c), espe-
cially in the post-cracking state where certain eccentricities may be caused by dis-
placements of the specimen.
2.2.1 Digital Image Correlation (DIC)

The DIC measurement system (Fig. 2) used consisted of two Prosilica GT600 29MPx
cameras (with Rodagon lenses, f = 80 mm). The distance between the cameras and the
test specimens varied between 1.6 m and 2 m and the frequency of image capture was
2 Hz. The dot size of the speckle pattern used on the samples was 0.66 mm (Fig. 2).
Fig. 2. Overview of the DIC measurement system and speckle pattern with dots of 0.66 mm.
2.2.2 Laser Scanning, NDI System

An optical measurement system NDI was used to measure the displacements on the
rear face of the specimens. Figure 3 shows the optical sensors attached to the speci-
mens; all of them were tracked with a frequency of 50 Hz. The sensors around the
notch can capture the initiation and propagation of the crack and the other sensors serve
as reference points and allow information about rotational movements to be obtained.
11 13
11 1 2 13
12 14 15
3 4
1 2 5 6
22 3 4
5 6
7 8
7 8
16
9 10
21 1 2 11 12 23 12 9 10 14
3 4 13 14
21 20 17
5 6 15 16
7 8 17 18
22 9 10 19 20 24
19 18
25 26 25 26 25 26
(a) Z-Type push-off (b) Modified JSCE-G 553 shear test (c) FIP shear test
Fig. 3. Optical sensors on the rear face of the specimens.
2.2.3 Test Procedure

In order to take into account the softening behaviour of the SFRC, the load on the
specimens is applied in displacement control, and two load phases were considered.
The first load phase was running with a displacement rate of 0.06 mm/min until a
displacement of 1 mm after the peak (the peak is detected by a load drop of 3 kN) was
achieved. Once the first load phase is concluded, continuously the second load phase is
started with a displacement rate of 0.5 mm/min until the relative displacement reaches
15 mm (see Table 2 and Fig. 4).
Table 2. Displacement ratio of the applied load phases.

Phase 1 Phase 2
Displacement rate [mm/min] 0.06 0.50
F Fmax
-3 kN
1 mm 15 mm
w
phase 1 phase 2
Fig. 4. Load application criteria for phase 1 and 2.
2.3 Materials and Concrete Mix

All specimens were made from identical SFRC mixes with four different fibre dosages:
0 kg/m3, 40 kg/m3, 60 kg/m3 and 80 kg/m3 (0%, 0.5%, 0.75%, 1%), using a concrete
mix (Table 3) suited to obtain a normal strength concrete C30/37. Industry-standard
Dramix 3D 80/50BG steel fibres were used, whose geometry and material properties
are shown in Table 4.
Table 3. Mixture composition and properties of the concrete.

Mixture constituent Concrete properties1)
Cement 42.5 384 [kg/m ]3
qf [%] Days2) fc [MPa] Ec [GPa]
Water 204 [kg/m3] 0.00 28 37.1 –
Sand 939 [kg/m3] 0.00 72 43.9 30.2
Aggregate 4–8 mm 816 [kg/m3] 0.50 72 – –
Additive 0 [kg/m3] 0.75 72 42.4 30.4
W/C 53 [%] 1.00 72 42.4 30.2
Fibres 80/50BG 0-40-60-80 [kg/m3] 1)
Standard concrete cylinder test
2)
Shear tests: 72 days
Table 4. Steel fibres geometry and material properties.

Geometry Material properties
Fibre family 3D [-] Nom. tensile strength 1’270 [N/mm2]
Length 50 [mm] Modul of elasticity 200’000 [N/mm2]
Diameter 0.62 [mm] Strain at ultimate strength 0.8 [%]
Aspect ratio 80 [-]
2.4 Experimental Results

The maximum average shear stress capacity smax for each specimen, considering the
effective shear plane areas (Ac,eff), are shown in Table 5. For a proper interpretation of
the results, it is essential to know the effective (real) fibre volume fraction qf ;eff in the
shear plane, which may differ from the theoretical fibre volume fraction (0%, 0.5%,
0.75% and 1%), due to the non-homogeneous 3D distribution and orientation of fibres
produced by the phenomenon known as wall effect or by obstruction of reinforcements
concentrated in certain areas. In order to address this issue, the effective fibre volu-
metric fractions qf ;eff in the shear plane of the specimens have been determined by
counting the number of fibres in the crack plane Nf . However, it is not entirely accurate
to visually quantify the effective volumetric fraction in the shear plane, due to (i) the
difficulty in counting fibres (some of them are broken and others are slipped), and
(ii) the theoretical assumption which refers to a homogeneous 3D fibre distribution and
orientation. The total number of fibres counted in the shear plane are divided by the
theoretical 3D fibre orientation factor Kf ð3Þ ¼ 0:5 [21]. The effective fibre volume
fractions qf ;eff in the shear plane of the specimens and the shear test results versus
theoretical and effective fibre volume fraction plots are shown in Table 5 and Fig. 5,
respectively. As shown Table 5, the effective volumetric fibre contents qf ;eff in the
shear plane vary from the theoretical values qf by −8% to +12% in the Z-type test and
by −33% to −16% in the JSCE and FIP tests.
Regarding Specimen Z-40 (qf ¼ 0:5%), its maximum average shear stress is closer
to the shear resistance of the plain concrete specimen Z-0 (difference less than 10%),
which indicates that the fibres have not had a significant influence on the maximum
shear resistance. This issue can be attributed to the non-homogeneous distribution of
the fibres in the shear plane, which were concentrated mainly in the edges, thus pro-
ducing in the central area, where the highest shear stresses should be transferred (see
Fig. 6a and Fig. 6b), a zone practically free of fibres (see Fig. 6c).
Table 5. Maximum average shear stresses capacity (fc,test day = 43 N/mm2).

qf ;eff
Specimen ID qf [%] Nf [-] qf,eff [%] qf
Fmax [kN] Ac,eff [mm2] smax [N/mm2]
Z-0 0.00 0.00 0.00 – 79 13’800 5.7
Z-40 0.50 135 0.56 1.12 92 14’760 6.2
Z-60 0.75 163 0.69 0.92 131 14’520 9.0
Z-80 1.00 242 1.03 1.03 148 14’400 10.3
JSCE-0 0.00 0.00 0.00 – 182 27’848 6.5
JSCE-40 0.50 96 0.42 0.84 220 27’848 7.9
JSCE-60 0.75 122 0.54 0.72 246 27’966 8.8
JSCE-80 1.00 189 0.82 0.82 275 28’321 9.7
FIP-0 0.00 0.00 0.00 – 269 49’506 4.1
FIP-40 0.50 332 0.42 0.84 367 48’730 5.7
FIP-60 0.75 440 0.53 0.71 427 48’840 6.6
FIP-80 1.00 533 0.67 0.67 457 48’841 7.1
Figure 5 shows that the shear stress capacity increases approximately linearly with
the steel fibre content, and the trendlines of the test results per setup (Fig. 5b) are nearly
parallel. From the latter, it can be concluded that the test configuration had a similar
influence on the shear capacity, independently of the fibre volume (note that Z-40
specimen test result (Z-type push-off) is excluded from the trendline for the reasons
outlined above).
12 12
10 10
8 8
τ [N/mm2 ]
τ [N/mm2 ]
6 6
excluded from excluded from
4 trendline 4 trendline
2 2
0 0
0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2
ρ f [%] ρ f,eff [%]
K Z-type I JSCE J FIP
(a) (b)
Fig. 5. (a) Shear test results – theoretical fibre volume fraction. (b) Shear test results – effective
fibre volume fraction.
Considering the Z-type tests results as reference (Fig. 5b) – as explained below, this
test setups should theoretically be the most reliable to determine the shear transfer
capacity –, the experimental shear capacities obtained by JSCE and FIP tests setups are
around 6% higher and 20% lower, respectively. These differences can be attributed to
the test setups. The FIP test setup (asymmetrical four-point bending test) produces
tensile stresses in the upper and lower part of the specimen (according to a strut-and-tie
model and confirmed by the DIC measurements, see Fig. 7a and Fig. 7b). These tensile
stresses cause a premature failure of the specimen in comparison with the other two test
setups; in fact, the plain concrete specimen did not fail in the expected shear plane
despite the notch that had been sawed all around the specimen (Fig. 7c and Fig. 7d).
Fig. 6. Z-Type specimens. (a) Minimum principal strain of Z-60 test (DIC capture). (b) Failure
of specimen Z-60 (DIC capture). (c) Fibre distribution in share plane of Z-40 after test.
The higher nominal shear stresses obtained in the JSCE tests are attributed to the
beneficial compressive stresses produced by the arching effect, which is expected due
to the test setup and confirmed by the DIC measurements (Fig. 8a). The narrow bearing
surfaces, subject to high vertical stresses, can resist high horizontal reactions resulting
from the test setup (the specimen tends to extend at the bottom), which avoids hori-
zontal tensile stresses at the bottom of the specimen necessary for equilibrium with the
compression in the upper part. On the other hand, due to the high stresses concen-
trations, the contact surfaces are prone to fail and consequently, negatively affect the
maximum load capacity. In fact, Specimens JSCE-40 and JSCE-80 failed by local
indentation of the contact surfaces, see below (Fig. 8b and Fig. 8d).
Fig. 7. DIC captures of FIP tests. (a) Minimum principal strain of FIP-80 test. (b) Maximum
principal strain of FIP-80 test. (c) und (d) Failure of the plain concrete specimen of FIP-0 test.
Fig. 8. DIC captures of JSCE tests. (a) Arching effect. (b) Failure of specimen JSCE-40.
(c) Failure of specimen JSCE-60. (d) Failure of specimen JSCE-80.
No particularities that could significantly have affected the shear transfer were
observed in the Z-type push-off tests (see Fig. 6a and Fig. 6b). Therefore, the authors
consider that the results obtained with the Z-Type test setup give the most reliable
results among the three test setups studied. The shear stress - crack width and shear
stress - slip curves for all tests are shown in Fig. 9. The crack kinematics was calculated
as the average of the relative displacements along the vertical crack, at both sides of the
specimen (front and rear), which were obtained by the DIC measurement system on the
front side and by the NDI measurement system on the rear side.
Z-Type Z-Type Z-Type Z-Type

12 12 1.2 1.2
Z-40 Z-40 Z-40 Z-40
10 Z-60
10 Z-60
1 Z-60 1 Z-60
Z-80 Z-80 Z-80 Z-80
8 8 0.8 0.8
[MPa]
[MPa]
/ max
/ max
6 max
= 5.7 MPa 6 max
= 5.7 MPa 0.6 0.6
(plain concrete) (plain concrete)
4 4 0.4 0.4
2 2 0.2 0.2
0 0 0 0
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4
[mm] [mm] n
[mm] t
[mm]
n t
(a.1) (b.1) (c.1) (d.1)
JSCE JSCE JSCE JSCE

12 12 1.2 1.2
JSCE-40 JSCE-40 JSCE-40 JSCE-40
10 JSCE-60
10 JSCE-60
1 JSCE-60 1 JSCE-60
JSCE-80 JSCE-80 JSCE-80 JSCE-80
8 8 0.8 0.8
= 6.5 MPa = 6.5
[MPa]
[MPa]
/ max
/ max
max max
6 (plain concrete) 6 (plain concrete) 0.6 0.6
4 4 0.4 0.4
2 2 0.2 0.2
0 0 0 0
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4
[mm] [mm] n
[mm] t
[mm]
n t
(a.2) (b.2) (c.2) (d.2)
FIP FIP FIP FIP

12 12 1.2 1.2
FIP-40 FIP-40 FIP-40 FIP-40
10 FIP-60
10 FIP-60
1 FIP-60 1
FIP-60
FIP-80 FIP-80 FIP-80
8 8 0.8 0.8 FIP-80
[MPa]
[MPa]
/ max
/ max
6 6 0.6 0.6
max
= 4.1 MPa max
= 4.1 MPa
4 (plain concrete) 4 (plain concrete) 0.4 0.4
2 2 0.2 0.2
0 0 0 0
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4
[mm] [mm] n
[mm] t
[mm]
n t
(a.3) (b.3) (c.3) (d.3)
Fig. 9. (a) Shear stress s – crack width dn. (b) Shear stress s – crack slip dt. (c) Shear stress to
max. shear stress ratio s/smax – crack width dn. (d) Shear stress to max. shear stress ratio s/smax–
crack slip dt.
Comparing the curves for the different fibre volume fractions, it can be seen that the
curves of the Z-Type (Fig. 9c.1 and Fig. 9d.1) and FIP (Fig. 9c.3 and Fig. 9d.3) test
setups are fairly congruent; particularly the FIP tests exhibit quite similar load-average
crack kinematics curves. It is also observed that at equal ratios s=smax , larger average
crack widths and slips are observed in the FIP tests, and that at the same average crack
opening, higher ratios s=smax are observed in these tests (e.g. for dn ¼ 1 mm, the ratios
s=smax for Z-60, JSCE-60 and FIP-60 are around 0.82, 0.58 and 0.97, respectively).
This can be explained by the tensile stress fields produced by the test setup (discussed
above), tending to increase the crack opening in the shear plane. Regarding the JSCE
test, since the arching effect (discussed above) causes compressive stresses in the shear
plane, the maximum shear capacity smax should be higher, and the crack opening and
slip in the shear plane tend to be less than the other setups. However, the shear failures
in these tests occurred in combination with local concrete failures at the supports (see
Fig. 8b and Fig. 8d). The failures at the supports were dominant in the JSCE-40 and
JSCE-80 tests, while a shear failure was observed for JSCE-60 test (see Fig. 8c). This
explains why average crack width and slip curves of the specimen JSCE-60 have a
smoother softening behaviour than the JSCE-40 and JSCE-80 specimens (Fig. 9a.2 to
Fig. 9d.2), which exhibit a very pronounced softening after the maximum load is
reached (just when the concrete supports failed, as confirmed by the DIC images).
However, it cannot be concluded how much influence this combined failure had on the
maximum shear stress capacity because both the JSCE-40 and JSCE-80 specimens and
the JSCE-60 fit the trendline (Fig. 5b).
3 Conclusions
From the considerations of the previous paragraphs the following main conclusion can
be drawn:
• The shear capacity depends on the stress state in the shear plane, which is not
homogeneous in reality, and accompanied by normal stresses depending on the test
setup which also influence the crack opening. The test setups therefore clearly bias
the measured shear transfer capacity across the crack. In this respect, no particu-
larities were observed in the Z-type test setup that could significantly affect the
shear transfer capacity. The authors therefore consider this test setup as the most
reliable of the three tests studied to determine the shear transfer capacity.
• In the modified JSCE-G 553 tests, the shear transfer capacity is clearly affected by
an arching effect (Fig. 8a), which compresses the shear plane and reduces the crack
opening. Consequently, this setup tends to overestimate the shear capacity. The
experimental shear capacities observed in the JSCE test setup carried out by the
authors were only approximately 6% higher than those obtained by the Z-type tests
due to local failures at the bearing surfaces. However, due to the arching effect and
the kinematics of the specimen after cracking (Fig. 8c), further modifications on the
JSCE-G 553 (or previous version JSCE-SF 6 [22]) shear test to avoid such failures,
e.g. by providing additional supports at the ends of the specimen [8], favour the
formation of the arc effect by increasing the normal compressive stresses and,
accordingly, lead to a much more pronounced overestimation of the shear transfer
capacity. This explains the much higher shear transfer capacities observed using
such setups [8], as observed by [12].
• The FIP test setup causes tensile stresses in the upper and lower part of the spec-
imen (Fig. 7a and Fig. 7b), which cause a premature failure of the specimen and
thus lower shear transfer capacities compared to the other two shear test setups.
Therefore, the FIP shear test setup underestimated the shear transfer capacity by
approx. 20% compared to the Z-type test setup.
• The authors are not aware of any specific study in the literature regarding the
influence of the stress state in the shear plane, produced by the test setups, on the
shear transfer capacity. Neglecting this clearly significant influence of the test setup
partly explains the wide variation in the predictions of the shear transfer capacity
obtained from empirical models, which were calibrated using different test setups.
The comparison of the experimental results with theoretical models underlying this
conclusion will be published in a separate paper.
• The results show that the shear transfer capacity increases approximately linearly
with the steel fibre volume fraction, and the trendlines of the test setup results
(Fig. 5b) tend to be parallel to each other.
• Due to the non-homogeneous 3D distribution and orientation of fibres produced by

the phenomenon known as wall effect or by obstruction of reinforcements con-
centrated in certain areas, the effective fibre volumetric fractions qf ;eff in the shear
plane differ from the theoretical fibre volume fractions qf , where in some cases the
differences are significant, with up to 33% deviation in the tests.
Acknowledgements. The support of Patrick Studer and Yanik Pfister in carrying out the
experimental campaign and post-processing the results, as part of their MSc Project at ETH
Zurich, is greatly acknowledged.
References
1. Jayaprakash, J., Abdul Samad, A.A., Abbasvoch, A.: Experimental investigation on shear
capacity of reinforced concrete precracked push-off specimens with externally bonded bi-
directional carbon fibre reinforced polymer fabrics. Mod. Appl. Sci. 3 (2009). https://doi.org/
10.5539/mas.v3n7p86
2. Khaloo, A.R., Kim, N.: Influence of concrete and fiber characteristics on behavior of steel
fiber reinforced concrete under direct shear. ACI Mater. J. 94(6), 592–601 (1997)
3. Boulekbache, B., Hamrat, M., Chemrouk, M., Amziane, S.: Influence of yield stress and
compressive strength on direct shear behaviour of steel fibre-reinforced concrete. Constr.
Build. Mater. 27(1), 6–14 (2012). https://doi.org/10.1016/j.conbuildmat.2011.07.015
4. Mostafazadeh, M., Abolmaali, A.: Shear behavior of synthetic fiber reinforced concrete.
Adv. Civ. Eng. Mater. 5, 371–386 (2016). https://doi.org/10.1520/ACEM20160005
5. Khanlou, A., MacRae, G.A., Scott, A.N., Hicks, S.J., Clifton, G.C.: Shear performance of
steel fibre-reinforced concrete. In: Presented at the Steel Innovations Conference 2013,
Christchurch, New Zealand, pp. 21–22 (2013)
6. Mirsayah, A.A., Banthia, N.: Shear strength of steel fiber-reinforced concrete. ACI Mater.
J. 99(5), September 2002. https://doi.org/10.14359/12326
7. Mansur, M.A., Vinayagam, T., Tan, K.-H.: Shear transfer across a crack in reinforced high-
strength concrete. J. Mater. Civ. Eng. 20(4), 294–302 (2008). https://doi.org/10.1061/
(ASCE)0899-1561(2008)20:4(294)
8. Soetens, T., Matthys, S.: Shear-stress transfer across a crack in steel fibre-reinforced
concrete. Cem. Concr. Compos. 82, 1–13 (2017). https://doi.org/10.1016/j.cemconcomp.
2017.05.010
9. Appa Rao, G., Sreenivasa Rao, A.: Toughness indices of steel fiber reinforced concrete
under mode II loading. Mater. Struct. 42(9), 1173 (2009). https://doi.org/10.1617/s11527-
009-9543-6
10. Cuenca, E., Serna, P.: Shear behavior of self-compacting concrete and fiber-reinforced
concrete push-off specimens. In: Design, Production and Placement of Self-Consolidating
Concrete, pp. 429–438. Springer, Dordrecht (2010)
11. Li, B., Maekawa, K., Okamura, H.: Contact density model for stress transfer across cracks in
concrete. J. Fac. Eng. 40(1), 9–52 (1989)
12. Kaufmann, W., Amin, A., Beck, A., Lee, M.: Shear transfer across cracks in steel fibre
reinforced concrete. Eng. Struct. 186, 508–524 (2019). https://doi.org/10.1016/j.engstruct.
2019.02.027
13. Valle, M., Buyukozturk, O.: Behavior of fiber reinforced high-strength concrete under direct
shear. ACI Mater. J. 90(2) (1993). https://doi.org/10.14359/4006
14. Van De Loock, L.: Influence of steel fibres on the shear transfer in cracks. In: Proceedings of
International Symposium Fibre Reinforced Concrete, Madras, India, pp. 1101–1112 (1987).
https://www.scopus.com/record/display.uri?eid=2-s2.0-85062080024&origin=inward
15. Kaufmann, W.: Strength and deformations of structural concrete subjected to in-plane shear
and normal forces (1998)
16. Bazant, Z.P., Gambarova, P.: Rough cracks in reinforced concrete. ASCE J. Struct. Div. 106
(4), 819–842 (1980)
17. Walraven, J.C.: Fundamental analysis of aggregate interlock. ASCE J. Struct. Div. 107(11),
2245–2270 (1981)
18. Li, B., Maekawa, K.: Contact density model for cracks in concrete. IABSE 54, 51–62
(1987). https://doi.org/10.5169/seals-41916
19. Japan Society of Civil Engineers (JSCE): JSCE-G 553 Test method for shear strength of
steel fiber reinforced concrete (2010)
20. Fédération Internationale de la Précontrainte: Technical reaport: Shear at the interface of
precast and in situ concrete (1978)
21. Pfyl, T.: Tragverhalten von Stahlfaserbeton, Swiss Federal Institute of Technology in Zurich
(2003)
22. Japan Society of Civil Engineers (JSCE): JSCE-SF 6 Method of test for shear strength of
steel fiber reinforced concrete (1990)
Bearable Local Stress of High-Strength SFRC
Sven Plückelmann1(&), Rolf Breitenbücher1, Mario Smarslik2,

and Peter Mark2
1
Institute of Building Materials, Ruhr University Bochum, Bochum, Germany
sven.plueckelmann@rub.de
2
Institute of Concrete Structures, Ruhr University Bochum, Bochum, Germany
Abstract. In the case of partial-area loading, compressive forces are trans-

mitted into concrete members only over a limited area. For plain concretes and
conventionally reinforced concretes, numerous investigations have already been
carried out analyzing the load-bearing behavior under partial-area loading. Due
to the tendency towards higher concrete strengths and the increasingly wide-
spread use of steel fibers in recent years, it becomes also necessary to investigate
the performance of high-strength steel fiber reinforced concrete (SFRC) under
partial-area loading. This paper describes experimental tests on high-strength
steel fiber reinforced concrete under partial-area loading with spatial and plane
load distribution. Different area ratios and concretes with different fiber types
and contents as well as fiber cocktails were considered. On the basis of the test
results, a calculation approach is proposed for the determination of the bearable
ultimate local stress. It is shown that by referring to the flexural tensile strength,
instead of the compressive strength, as in the case of common calculation
approaches, a more precise approximation of the ultimate local stresses for high-
strength steel fiber reinforced concrete is possible.
Keywords: Steel fiber reinforced concrete High-strength concrete Partial-

area loading Claculation approach
1 Introduction
If compressive forces are transmitted into concrete members over a limited area, this is
referred to as partial-area loading. This load situation occurs, for example, in the case of
anchorages in post-tensioned concrete members, bearing over piers in bridge structures,
stanchions over concrete footings or tunnel lining segments due to the concentrated
jacking forces of the tunnel boring machine during the construction stage [1–4].
The load bearing capacity of concrete members subjected to partial-area loading
can be a multiple of the load derived from the uniaxial compressive strength, since the
surrounding unloaded concrete causes a confinement effect. However, due to the
accompanying load distribution, not only concentrated compressive stresses but also
transverse tensile stresses are generated. If these transverse tensile stresses exceed the
concrete’s tensile strength and are not absorbed by appropriate reinforcement they lead
to splitting and thus premature failure of the concrete member.
The first experiments focusing on partial-area loading were carried out on stone
blocks by Bauschinger [5] in the 1870’s. Since then, a large number of further
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_16
Bearable Local Stress of High-Strength SFRC 177
experimental and theoretical studies have been conducted that deal with the load
bearing behaviour under partial-area loading (e.g. [6–14]). In most cases, the focus of
the studies lay on unreinforced or conventionally reinforced, normal-strength concretes.
However, in a few studies high-strength [15–17] and ultra-high-strength concretes [18]
as well as steel fiber reinforced concretes (SFRC) [12–18] were also considered. On the
basis of the test results, various approaches for the calculation of the bearable ultimate
local stress were developed and corresponding design approaches were formulated,
some of which are included in the acutal standards and guidelines (e.g. Eurocode 2
[21], Modelcode 2010 [22]).
Most of the known approaches for the calculation of the ultimate local stress are
based on the dependency of the ratio between the total cross-sectional area and the
loaded area, which was clearly ascertained by experimental tests. This area ratio is
usually included as a square or cubic root factor in the calculation of the ultimate local
stresses. The second factor used in most approaches is the compressive strength.
However, while the reference to the compressive strength shows a good agreement
with the factual ultimate local stresses for unreinforced, normal-strength concretes, this
leads to an overestimation of the load-bearing capacity for high-strength concretes.
This is due to the under-proportional increase of concrete tensile strength with
increasing compressive strength in connection with the aforementioned splitting phe-
nomenon. Taking this into account, the calculation formulas for high-strength and
ultra-high-strength concretes in [16, 18] are modified by a reduction factor.
In investigations of SFRC under partial-area loading [18–20] it could be clearly
proven that steel fiber reinforcement with already low fiber content leads to an increase
in the load-bearing capacity. In [18] Klotz proposes a calculation formula for ultra-
high-strength SFRC, which also refers to the compressive strength. Further calculation
approaches for SFRC under partial-area loading are currently not known.
Since the tensile strength has a strong influence on the load-bearing capacity under
partial-area loading its consideration in a calculation approach could provide a better
agreement with the factually bearable ultimate local stresses, especially for high-
strength concretes. In this, the non-proportionality between compressive and tensile
strength would be directly considered. Furthermore, in the case of SFRC, the load-
increasing effect of steel fibers could also be taken into account.
2 Experimental Investigations
2.1 Materials
Both the plain concrete (reference) and all SFRCs that were investigated within the
study were produced using the same base mix (Table 1). In total, seven different
SFRCs with three different fiber types were included in the experimental study. The
properties of the used steel fibers are listed in Table 2. The fiber content was varied
between 40 and 120 kg/m3. Two of the SFRC mixtures were produced with a fiber
cocktail. Table 3 gives an overview of the SFRCs included in the test program. In
comparison to the reference concrete, the dosage of superplasticiser was adjusted in
such a way that a plastic to soft consistency was achieved for all mixes, irrespective of
178 S. Plückelmann et al.
the fiber content. For further details of the mix designs and relevant properties of the
fresh concretes, see [19].
Table 1. Components of the base-mix.

Component Unit Base-mix
3
Cement (CEM I 52.5R) [kg/m ] 330
Fly ash 90
Aggregate (sand, gravel) 1850
w/c-ratio [-] 0.45
Superplasticiser (PCE) [M.-%] 0.4
Table 2. Properties of steel fibers

Fiber type Shape Length Diameter Aspect ratio Tensile
(designation) L [mm] D [mm] L/D [-] Strength ft
[MPa]
3D 80/60 BG straight, 60 0.75 80 1.225
(L) hooked end
3D 65/35 BG straight, 35 0.55 65 1.345
(M) hooked end
FM 0.19/13 straight 13 0.19 68 2.000
(S)
Table 3. Fiber content of SFRCs

SFRC designation L40 L60 L80 M60 S60 L40S40 L60S60
L [kg/m3] 40 60 80 – – 40 60
M [kg/m3] – – – 60 – – –
S [kg/m3] – – – – 60 40 60
2.2 Test Program and Test Setup

The mechanical properties of the hardened concretes were characterized by assessing
the compressive strength according to EN 12390-3 [23], the splitting tensile strength
according to EN 12390-6 [24] as well as the Young’s modulus according to EN 12390-
13 [25]. In addition, flexural tensile tests (four-point bending tests) were performed on
six beams for each concrete according to the German guideline for steel fiber reinforced
concrete (DAfStb Richtline “Stahlfaserbeton” [26]).
In order to investigate partial-area loading with both spatial and plane load distri-
bution, test specimens with two different dimensions (150 150 300 mm3 and
300 150 300 mm3) were produced (see Fig. 1, left). The casting direction was
parallel to the load direction or perpendicular to the load introduction surface of

the specimens, respectively. At the age of 28 days, the specimens were tested under
centric partial-area loading. For specimens with a square cross-sectional area (Ac1 =
150 150 mm2), the loaded area Ac0 was also square (spatial load distribution). For
specimens with a cross-sectional area of Ac1 = 300 150 mm2, the length of the
loaded area was equal to the length of the specimens (plane load distribution). The area
ratio d = Ac1/Ac0 was varied between 2.25 and 16. In Table 4 the test program is
summarized. For each of the listed combinations, 3 to 4 specimens were tested.
Fig. 1. Dimensions of the specimens for partial-area loading tests with spatial and plane load
distribution (left) and photograph of the test setup (right).
The partial-area loading tests were performed using a servo-hydraulic testing

machine with a maximum load capacity of 5 MN. The load was transmitted onto the
upper surface of the specimens through steel bearing plates with dimensions corre-
sponding to the area ratios listed in Table 4. The load was applied with a constant loading
rate of 0.5 mm/min. The deformations of the specimens in longitudinal and transverse
direction to the load axis were measured by LVDTs. Figure 1 (right) shows the test setup
exemplary for the spatial case of partial-area loading. For further details see [19].
Table 4. Test program.

Series designation Area ratio d = Ac1/Ac0 [-]
2.25 4 9 16
REF s/p s/p s/p s
L40 s s/p s –
L60 – s/p s/p s
L80 – s/p s –
M60 – s s –
S60 – s/p s –
L40S40 P s/p s/p –
L60S60 – s s s
s: partial-area loading with spatial load distribution
p: partial-area loading with plane load distribution
3 Results
3.1 Properties of Hardened Concretes
The mechanical properties (compressive strength, splitting strength, ultimate flexural
strength, Young’s modulus) of the hardened concretes are listed in Table 5. The
average compressive strength of the plain concrete (reference) is 84.5 MPa. In general,
it can be seen that the compressive strengths of the SFRCs are slightly higher than
those of the plain concrete. Regarding the concretes L40 to L60 (fiber type L), a
moderate increase in the compressive strength can be observed with increasing fiber
content. The splitting strength is also enhanced with increasing fiber content. However,
as expected, the influence of steel fiber reinforcement on the splitting tensile strength is
much stronger than on the compressive strength. For example, the increase in com-
pressive strength for concrete L80 (80 kg/m3 L-fibers) is only 12% compared to the
plain concrete, while the splitting tensile strength was increased by 85%. Obviously,
the effects of fiber reinforcement on the Young’s modulus are negligible.
Table 5. Mechanical properties of hardened concretes

Designation Compressive Splitting tensile Ultimate flexural Young’s
strength strength strength modulus
fc,cube [MPa] fct [MPa] f fcflm;ult Ecm [GPa]
REF 85.5 4.0 7.7 36.9
L40 85.5 5.3 8.7 35.9
L60 87.4 6.7 9.8 35.4
L80 94.5 7.4 11.6 36.6
M60 81.4 6.0 8.9 35.6
S60 88.5 5.1 8.5 36.7
L40S40 91.6 7.3 10.9 –
L60S60 94.4 7.8 12.9 –
Figure 2 shows the averaged stress-deflection curves of the flexural bending tests
for the plain concrete and each SFRC. The plain concrete (REF) is characterized by
failure immediately after the first crack has occurred. In contrast, in the cases of SFRCs,
tensile stresses are still absorbed by the crack-bridging fibers even after the first peak
stress has been exceeded and the first crack occurred. This ductile post cracking
behaviour is typical for SFRC and is reflected by the curves shown in Fig. 2. The
stress-deflection curve of the concrete reinforced with 60 kg/m3 of microfibers (S60) is
characterized by a continuous decline after the first peak stress is exceeded. In this case,
the first peak stress is equal to the ultimate flexural tensile strength. For all the other
SFRCs investigated, however, a further stress increase with growing deflection can be
detected. Depending on the type of fiber reinforcement, the stresses in the post cracking
range predominantly rise above the first peak stress. After reaching the ultimate flexural
strength, a decline of stresses can be observed with increasing deflection. The ultimate
flexural strengths are listed in Table 5. A more detailed evaluation of the flexural
bending test including first peak and residual strengths as well as a classification of the
SFRCs according to [26] can be found in [27].
Fig. 2. Averaged stress-deflection curves of beams according to [26]
3.2 Bearable Local Stress

This paper focusses on the ultimate local stresses rmax, i.e. the ultimate load divided by
the loaded area Ac0 (rmax, = Fult/Ac0). These are presented as mean values in Table 6
for partial-area loading tests with both spatial and plane load distribution. The results
show on the one hand the influence of the area ratio and on the other hand the influence
of the fiber reinforcement. Information about the fracture behavior in the form of load-
displacement graphs and fracture patterns are given in [19].
Table 6. Ultimate local stresses rmax [MPa] for spatial and plane load distribution
Area ratio d = Ac1/Ac0
Designation Spatial load Plane load
distribution distribution
2.25 4 9 16 2.25 4 9 16
REF 81 104 152 214 69 83 109 –
L40 97 132 194 – – 137 – –
L60 113 153 214 298 118 154 193 –
L80 121 171 254 – – 168 – –
M60 107 137 196 – – – – –
S60 – 141 194 – – 123 – –
L40S40 – 162 236 – 128 173 222 –
L60S60 – 184 266 379 – – – –
The bearable ultimate local stress depends decisively on the area ratio due to the
fact that the confinement effect is enhanced by the surrounding unloaded concrete with
increasing area ratio. As expected, a significant increase in the ultimate local stress can
therefore be overserved with increasing area ratio for both the plain concrete and
SFRCs (see Table 6). This correlation is illustrated exemplarily for the plain concrete
(REF) and L60 in Fig. 3 and is valid both for spatial and plane load distribution.
However, this effect is less pronounced for the plane case due to the incomplete
confinement.
Fig. 3. Ultimate local stress for spatial and plane load distribution (REF, L60)
Table 6 as well as Fig. 3 also show that the steel fiber reinforcement leads to a
considerable increase in the ultimate local stress. Due to the dominant splitting tensile
failure of the specimens (see [19]), this is mainly attributed to the higher tensile bearing
capacity caused by the fiber reinforcement. However, the reinforcement effectiveness,
i.e. the difference of the ultimate local stresses between the SFRC and plain concrete,
tends to improve as the area ratio increases. For example, the ultimate local stress in the
case of partial-area loading with spatial load distribution and an area ratio of 9 could be
increased by 40% compared to the plain concrete by adding 60 kg/m3 L-fibers (L60).
In the case of plane load distribution, the effectiveness of the steel fiber reinforcement
was even more pronounced. At an area ratio of 9, the increase in the ultimate local
stress for the concrete L60 was 77%. This correlation can also be observed for the other
concretes. The improvement of the bearable ultimate local stress with increasing area
ratio as well as due to the steel fiber reinforcement confirms the finding of earlier
investigations in [7, 9, 18, 20].
4 Calculation Approach
For concretes for which at least three different area ratios were tested, Fig. 4 shows the
percentage deviation of the ultimate local stress as a function of the area ratio with
reference to the ultimate local stress at an area ratio of 4. The values show good
agreement with the dotted reference curves, which represent the percentage deviation
between the square root (Fig. 4, left) or cubic root (Fig. 4, right) of the variable area
ratio and the reference area ratio of 4. For both plain concrete and SFRCs, this proves a
clear proportionality to the square root of the area ratio for partial-area loading with
spatial load distribution (cf. Fig. 4, left) and to the cubic root of the area ratio for
partial-area loading with plane load distribution (cf. Fig. 4, rigth). As in the case of
already known approaches, this correlation is included as a factor in the proposed
calculation formula.
In compliance with investigations in [18, 20] it was shown in Sect. 3.2 that the
ultimate local stress is significantly increased due to the steel fiber reinforcement. In
contrast, the compressive strength is only slightly influenced by the steel fibers (cf.
Table 5). For example, in comparison to the plain concrete, the ultimate local stress of
the concrete L80 (80 kg/m3 L-fibers) at an area ratio of 4 was increased by 65% (spatial
load distribution) or 102% (plane load distribution), while the compressive strength of
the same concrete was only increased by 12%. Therefore, a calculation approach for
(high-strength) SFRC based on a proportionality between the ultimate local stresses
and the compressive strength is not appropriate.
Fig. 4. Correlation between ultimate local stresses and area ratio for spatial load distribution
(left) and plane load distribution (right)
The results suggest that a calculation formula that refers to the tensile strength
could better reflect the actual load-bearing behaviour of high-strength SFRC. However,
since the determination of the centric tensile strength is comparatively complex and
very sensitive to eccentricities and other imponderables, it is advisable to resort to a
substitute tensile strength such as the flexural strength. This does not lead to any
additional testing effort, since the classification of SFRC usually requires the execution
of flexural bending tests. Alternatively, the use of the splitting tensile strength would
also be conceivable.
In accordance with common approaches, the following equations were chosen as
the basis for the calculation of the ultimate local stress:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rmax ¼ ai fi Ac1 =Ac0 ðfor spatial load distributionÞ ð1Þ
p ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rmax ¼ ai fi 3 Ac1 =Ac0 ðfor plane load distributionÞ ð2Þ
where ai are empirical factors and fi suitable strength values.

The factors ai were determined individually using the respective concrete strength
(compressive, ultimate flexural and splitting tensile strength) for the investigated
SFRCs according to Eq. (1) and (2), respectively, so that the sum of the deviation
between the ultimate local stress determined by the tests (Table 6) and by calculation is
minimal. The results for ai are listed in Table 7.
Table 7. Factors ai for the calculation of ultimate local stress with spatial and plane load
distribution
Designation Spatial load distribution: ai acc. Plane load distribution ai acc. to
to Eq. (1) with fi Eq. (2) with fi
fi = fc,cube fi = fc,ct fi = f fcflm;ult fi = fc,cube fi = fc,ct fi = f fcflm;ult
L40 0.76 12.28 7.48 1.01 16.37 9.97
L60 0.85 11.06 7.55 1.07 13.94 9.52
L80 0.90 11.35 7.24 1.12 14.28 9.10
M60 0.83 11.28 7.60 – – –
S60 0.76 13.14 7.87 0.88 15.21 9.11
L40S40 0.87 10.90 7.28 – – –
L60S60 0.95 11.81 7.40 1.15 14.39 9.60
Mean [-] 0.9 11.7 7.5 1.0 14.8 9.5
VarC. [%] 8.2 4.9 2.9 10.2 6.6 3.9
The lower the scatter between the single values of ai, the better is the agreement of
the experimentally determined and calculated ultimate local stress. On the basis of the
variation coefficients (VarC.) listed in Table 7, it can thus be shown which of the
concrete strengths is best suited for the precise calculation of the ultimate local stress,
when using the approach according to Eq. (1) and (2). With a variation coefficient for ai
of 2.9% (spatial load distribution) or 3.9% (plane load distribution), the use of the
ultimate flexural strength leads to the best agreement with the test results. For the
splitting tensile strength, the variation coefficient for ai is 4.9% and 6.6%, respectively.
Hence, the reference to the splitting tensile strength is also to be assessed as advan-
tageous in comparison to the compressive strength. With a variation coefficient of 8.2%
(spatial load distribution) and 10.2% (plane load distribution), the compressive strength
leads to the largest deviation between the calculation approach and the test results.
On the basis of these results, the ultimate flexural strength (f fcflm;ult ) is selected to be
included in the calculation formulas for the determination of the ultimate local stress.
The Equations are as follows:
f
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rmax ¼ 7:5 fcflm;ult Ac1 =Ac0 ðfor spatial load distributionÞ ð3Þ
f
p ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rmax ¼ 9:5 fcflm;ult 3
Ac1 =Ac0 ðfor plane load distributionÞ ð4Þ
Figure 5 shows the good agreement between the proposed calculation formulas
with respect to the ultimate flexural strength according to Eq. (3) and (4) and the test
results (Table 7). For comparison, Fig. 6 shows that the approach according to Eq. (1)
and (2), using the compressive strength and the respective factors ai according to
Table 7, leads to considerable deviations between the calculation approach and the test
results.
Fig. 5. Comparison between test results and calculation proposal with respect to the ultimate
flexural strength (left: spatial load distribution, right: plane load distribution)
Fig. 6. Comparison between test results and calculation proposal with respect to the ultimate
compressive strength (left: spatial load distribution, right: plane load distribution)
5 Summary and Conclusions
The paper describes experimental tests on high-strength steel fiber reinforced concrete
(SFRC) under partial-area loading with spatial and plane load distribution. On this
basis, a calculation approach is proposed for the determination of the bearable ultimate
local stress. Different area ratios and concretes with different fiber types and contents as
well as fiber cocktails were considered. The results of the investigations presented
allow to draw the main concluding remarks listed in the following:
• As expected, the ultimate local stress is significantly enhanced with increasing ratio
between the total cross-sectional area and the loaded area. This could be proven by
the tests for both the high-strength plain concrete and SFRC.
• In the case of partial-area loading with spatial load distribution, a proportionality
between the ultimate local stress and the square root of the area ratio could clearly
be ascertained. Due to the incomplete confinement in the case of partial-area
loading with plane load distribution, the increase in the ultimate local stress was
comparatively lower. In this case, a proportionality to the cubic root of the area ratio
was found.
• In comparison to the plain concrete, it could be shown that the ultimate local stress
is significantly increased by the steel fiber reinforcement. Thereby, a higher effec-
tiveness of the fiber reinforcement is observed with increasing area ratio. Further-
more, it was found that the reinforcement effectiveness was somewhat more
pronounced in the case of plane load distribution than in the case of spatial load
distribution.
• A calculation approach is proposed for the determination of the bearable ultimate
local stress. It is shown that by referring to the flexural tensile strength, instead of
the compressive strength as in the case of common calculation approaches, a more
precise approximation of the ultimate local stresses for high-strength steel fiber
reinforced concrete is possible.
Acknowledgements. The authors would like to acknowledge the financial support of the
German Research Foundation (DFG) for this work within the subproject B1 of the Collaborative
Research Center SFB 837 “Interaction Modeling in Mechanized Tunnelling”.
References
1. Gall, V.E., Marwan, A., Smarslik, M., Obel, M., Mark, P., Meschke, G.: A holistic approach
for the investigation of lining response to mechanized tunneling induced construction
loadings. Undergr. Space 3(1), 45–60 (2018)
2. Gödde, L., Strack, M., Mark, P.: M-N-Interaktionsdiagramme für stahlfaserverstärkte
Stahlbetonquerschnitte – Anwendung am Beispiel von Tübbingen. Beton- und Stahlbeton-
bau 105(5), 318–323. Ernst & Sohn, Berlin (2010)
3. Putke, T., Bohun, R., Mark, P.: Experimental analyses of an optimized shear load transfer in
the circumferential joints of concrete segmental linings. Struct. Concr. 16(4), 572–582
(2015)
4. Zhao, C., Alimardani Lavasan, A., Barciaga, T., Kämper, C., Mark, P., Schanz, T.:
Prediction of tunnel lining forces and deformations using analytical and numerical solutions.
Tunn. Undergr. Space Technol. 64, 164–176 (2017)
5. Bauschinger, J.: Versuche mit Quadern aus Naturstein. Mittleilungen aus dem Mechanisch
Technischen Laboratorien der Technischen Hochschule München 6, 1–20 (1876)
6. Ahmed, T., Burley, E., Rigden, S.: Bearing capacity of plain and reinforced concrete loaded
over a limited area. ACI Struct. J. 95, 330–342 (1998)
7. Hawkings, N.M.: The bearing strength of concrete loaded through rigid plates. Mag. Concr.
Res. 20(62), 31–40 (1986)
8. Lieberum, K.H.: Das Tragverhalten von Beton bei extremer Teilflächenbelastung. Ph.D.
thesis, Technische Hochschule Darmstadt (1987)
9. Niyogi, S.K.: The bearing strength of concrete – geometric variations. J. Struct. Div. 99(7),
1471–1490 (1973)
10. Schmidt-Thrö, G., Tabka, B., Smarslik, M., Scheufler, W., Fischer, O., Mark, P.:
Experimente zur Teilflächenbelastung mit vorwiegend ebener Lastausbreitung. Beton- und
Stahlbetonbau 113(2), 115–126. Ernst & Sohn, Berlin (2018)
11. Spieth, H.P.: Das Verhalten von Beton unter hoher örtlicher Pressung und Teilbelastung
unter besonderer Berücksichtigung von Spannbetontragwerken. Ph.D. thesis, Technische
Hochschule Stuttgart (1959)
12. Wichers, M.: Bemessung von bewehrten Betonbauteilen bei Teilflächenbelastung unter
Berücksichtigung der Rissbildung. Ph.D. thesis, Technische Hochschule Braunschweig
(2013)
13. Wurm, P, Daschner, F.: DAfStb Heft 286: Versuche über Teilflächenbelastung von
Normalbeton. Ernst & Sohn, Berlin (1977)
14. Wurm, P, Daschner, F.: DAfStb Heft 344: Teilflächenbelastung von Normalbeton an
bewehrten Scheiben. Ernst & Sohn, Berlin (1983)
15. Rheinhardt, H.-W., Koch, R.: Untersuchungen zum Einfluss der Bauteilgröße und
Betonzusammensetzung auf die Tragfähigkeit von hochfestem Beton unter Teilflächenbe-
lastung. Abschlussbericht, Universität Stuttgart (1997)
16. Rheinhardt, H.-W., Koch, R.: Hochfester Beton unter Teilflächenbelastung. Beton- und
Stahlbetonbau 113(2), 182–188. Ernst & Sohn, Berlin (2018)
17. Schön, A., Rheinhardt, H.-W.: Untersuchungen zur Teilflächenbelastung von hochfestem
Beton. Abschlussbericht, Universität Stuttgart (1994)
18. Klotz, S.: Ultrahochfester Beton unter Teilflächenbelastung. Ph.D. thesis, Universität
Leipzig (2018)
19. Song, F.: Steel Fiber Reinforced Concrete under Concentrated Load. Ph.D. thesis, Ruhr-
Universität Bochum (2017)
20. Yazdani, N., Spainhour, L, Haroon, S.: Application of Fiber Reinforced Concrete in the End
of Precast Pre-stressed Bridge Girders. Technical report of the Florida Department of
Transportation, Florida A & M University – Florida State University (2002)
21. EN 1992-1-1: Eurocode 2: Design of concrete structures – Part 1-1: General rules and rules
for buildings (2004)
22. Walraven, J., et al.: Model Code 2010. International Federation for structural concrete (fib)
23. EN 12390-3: Testing hardened concrete – Part 3: Compressive strength of test specimens
(2017)
24. EN 12390-6: Testing hardened concrete – Part 6: Tensile splitting strength of test specimens
(2010)
25. EN 12390-13: Testing hardened concrete – Part 13: Determination of secant modulus of
elasticity in compression (2014)
26. Deutscher Ausschuss für Stahlbeton, DAfStb-Richtlinie Stahlfaserbeton (2012)
27. Plückelmann, S., Breitenbücher, R., Smarslik, M., Mark, P.: Aufnehmbare Teilflächenspan-
nung von hochfestem Stahlfaserbeton. Beton- und Stahlbetonbau 114(9), 653–662. Ernst &
Sohn, Berlin (2019)
Impact Response of Different Classes of Fibre
Reinforced Concrete
Juan C. Vivas1, Raúl L. Zerbino1(&), María C. Torrijos1,

and Graciela M. Giaccio2
1
Faculty of Engineering UNLP, CONICET. LEMIT-CIC, La Plata, Argentina
zerbino@ing.unlp.edu.ar
2
Faculty of Engineering UNLP, LEMIT-CIC, CIC Researcher,
La Plata, Argentina
Abstract. The use of fibre reinforced concrete in structural elements exposed to

impacts or different types of extreme loading represents one of the main fields of
application of this high-performance material. Nevertheless, there is not a
general consensus about a test for impact characterization of fibre concretes and,
specifically a procedure to evaluate the contribution of fibres after cracking. It is
well known that fibres control the evolution of cracks, improving the durability
of concrete elements. Nowadays there are many structural fibres available; one
of the greatest advantages to enhance the use of different fibres is the intro-
duction of FRC classes in the fib Model Code 2010. However, there are not
references about the relationship between the residual capacity measured in
static tests (i.e. EN 14651) and the impact response. A drop weight impact test
method is proposed to evaluate the contribution of different fibres considering
both the cracking resistance and the behaviour in cracked state. Results of FRC
belonging to different classes, incorporating different contents of steel, glass and
polymer macrofibres are presented and compared. The effect of the residual
capacity measured on standard bending tests on the impact resistance is
discussed.
Keywords: Crack control FRC class Impact tests Glass macrofibres

Polymer macrofibres Steel fibres
1 Introduction
The benefits of using Fibre Reinforced Concrete (FRC) in structural elements exposed
to explosions, vibrations, ballistic shocks and different types of impact [1–7] have been
confirmed. Although numerous methods have been proposed [8–12] there is no a
general consensus about a method to characterize the impact response of FRC.
Considering that the main contribution of fibres takes places after the initial
cracking of concrete matrix, it is interesting to develop a procedure for evaluating the
impact resistance in FRC at the cracked state. As nowadays many different types of
fibres are available their performance in the presence of impact loads becomes of great
interest.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_17
190 J. C. Vivas et al.
Using a single drop-weight impact instrumented test Banthia et al. [13] compared
the flexure impact resistance of plain concrete, steel and polypropylene FRC con-
cluding that 0.5% in volume of polypropylene fibres increases around 50% the fracture
energy and 1.5% of steel fibres triplicated it. Bindiganavile et al. [14, 15] developed
fibre-matrix bond impact tests and bending impact tests on FRC prepared with poly-
meric and steel fibres; it was found that at high rates both fibres increased bond
strength, but at very high velocities steel fibres presented lower bond and energy
absorption capacity than the polymeric fibres; similar trends were observed in bending
impact tests. Testing steel, polypropylene and cellulose FRC by using the impact tests
proposed by the ACI committee 544 [16] it was observed that while the first crack
strength increased when cellulose fibres were incorporated, the use of polypropylene
fibres improved the post-crack behaviour.
As it is known, there can be quite different post-cracking behaviours in FRC. In
addition to the material with which the fibres are made (steel, polymer, etc.), many
factors as geometry, stiffness, strength and interface bond as well as the concrete matrix
characteristics can affect the composite response. In this way, there is a general
agreement that fibre performance must be valued by the FRC residual capacity.
Considering the huge alternatives of fibres for concrete reinforcement, the fib Model
Code 2010 [17] established FRC classes based on the residual stresses determined
using the EN14651 standard bending test [18].
In this paper a simple drop-weight test was implemented in order to explore the
relationship between the static residual capacity and the impact resistance. Different
classes of FRC incorporating different contents of steel, glass and polymer macrofibres
as well as the corresponding base concrete without fibres were evaluated.
2.1 Impact Test

With the aim of evaluating the impact resistance, both considering first crack and the
post-cracking behaviour, a repeated drop-weight test was implemented. In preliminary
experiences several alternatives for loads application and specimen geometry were
explored, as well as the variability of results and the minimum number of specimens
required and different parameters for impact resistance characterization [19].
Considering that one of the motivations of the research was to study the relation-
ship between the residual capacity measured in static standard tests and the impact
resistance, prismatic specimens of 150 150 300 mm were selected for the tests,
which may be the halves resulting from the bending characterization test (EN 14651),
with a 25-mm depth notch at the centre of the tensile face. As was confirmed in a
previous study, using this notch depth no shear failure occurs. The use of prisms of the
same section for static and impact tests avoids differences in fibre orientation due to the
mould geometry.
The impact test procedure and instrumentation is relatively simple and consists on
repeated drops of a projectile (with a mass m) on the top of the specimen from a certain
height (h). After each impact, Crack Opening Displacement (COD) was measured at
Impact Response of Different Classes of FRC 191
120 mm below the top face of the prism using a Dino-Lite Premier digital micro-
scope® AM4113T, 1.3 Megapixels, which with the provided software, makes possible
to measure elements with a precision of 0.0001 mm.
The impact testing device was done by adapting an equipment originally designed
for drop-weight tests on steel specimens [20, 21]. The system has two vertical steel rails
to guide the projectile; when it achieves the desired height, it is released and impacts
the specimen at the middle span. Figure 1 shows a general view of the machine and
some setup details; the projectile adopted for this study has 5 kg-mass, 150 mm width
and semi-circular section linear Tup (striking end of the projectile).
Fig. 1. Impact testing device (values in mm).
The prism rests on cylinders that allow the sample to rotate at the plane of impact;
each support has a metallic bar that fixes the prism and prevents lifting during and after
each impact. One of the supports is fixed to the base while the other is left free
(properly lubricated), so that it is allowed to move horizontally.
The impact test has two phases. In Phase 1, with the objective of determining the
cracking resistance, impact loads are applied on a sound specimen; the drop height is
progressively increased from an initial height (ho) of 100 mm with height increments
(Dho) of 50 mm; only one drop is applied for each height level and the process finishes
when a visible crack (*20 lm) is detected. With the aim of evaluating the contribution
of fibres in cracked state, Phase 2 starts using ho equal to 100 mm and Dho 100 mm, but
in this case three impacts are applied in each height level. The end of the test occurs
when the COD is greater than 3 mm. The heights gradual increase allows to evaluate
concretes with different impact resistances; and the post cracking impact repetitions in
each height level, prevents sudden leaps of energy, thus allowing a crack gradual
opening and a greater sensitivity to perceive the effect of each type of fibre. This is
useful specially when evaluating very low impact toughness concretes (or even in
simple concrete), otherwise it is not possible to measure crack growth after the
appearance of the first crack.
The energy of each impact is calculated as m.g.h (being g the acceleration of
gravity) and corresponds to the potential energy introduced to the system before
starting the drop. The cumulated energy is the sum of energies of all impacts received
by the specimen up to a certain point. The energy cumulated up to the first visible crack
appearance, named cracking energy (EC) and the initial visible crack opening (CODC),
are the impact test results from the Phase 1, while the post cracking energy (EP), which
is the cumulated energy between the first crack and the end of test (COD > 3 mm) and
the COD rate during post-cracking (VC), calculated from the cumulated energy between
crack openings of 0.5 mm and 2.5 mm (VC = 2 mm/(ECOD0.5mm - ECOD2.5mm),
expressed in mm/J), are obtained for the Phase 2. Finally, the total energy ET is also
calculated as sum of EC+EP.
2.2 Concretes
A reference concrete (R) without fibres and six FRC incorporating steel (S), polymer (P)
or glass (G) macrofibres were evaluated. The FRC were identified by a letter that corre-
sponds to the type of fibre and a number indicating the fibre content (in kg/m3). All
mixtures were prepared using similar materials proportions using ordinary Portland
cement, natural siliceous sand and 19 mm maximum size granite crushed stone. The
water/cement ratio was 0.40. In Table 1 the used fibres characteristics are presented.
Table 1. Characteristics of fibres used.

Designation S G P
Type Steel Glass Polymer
Shape hooked-end flat embossed
Length (mm) 50 36 58
Diameter (mm) 1.00 0.54 0.67
Tensile strength (MPa) >1100 >1700 >640
Elastic modulus (GPa) 210 72 6.8
Three prisms of 150 150 600 mm and three cylinders of 100 200 mm
were cast for each concrete mixture. All specimens were consolidated by external
vibration, cured in a moist room for 28 days and then remained in laboratory indoor, to
minimize the variation of concrete mechanical properties during the testing period.
Firstly, both EN14651 bending [18] and ASTM C-39 [22] compression tests were
carried out (at three months), and after that impact tests were performed.
Figure 2 shows the average stress-crack mouth opening displacement (CMOD) curves
of each concrete. It can be seen that FRC with a wide range of static residual capacity
were studied.
Fig. 2. Mean curves from bending tests performed in accordance to EN 14651.
Individual impact curves of concretes P5 and P10 are given in Fig. 3 as example. It
can be seen that the curves are consistent within the same FRC with an acceptable
variability in the post-cracking stage, which increases as the fibre content increases. In
preliminary studies it was found that for an error lower than 15% and reliabilities of 90
or 95% the minimum number of specimens statistically required is 5 and 7 respectively.
Fig. 3. Individual impact curves for P5 (left) P10 (right).
Figure 4 (left) shows impact curves representative of each concrete, a higher slope
implies a lower COD growth rate (VC). Each symbol corresponds to one drop. When
comparing the impact post-cracking response some differences in the shape of the
curves can be seen. Concretes with steel fibres show a continuous progressive growth
and the VC clearly decreases as fibre dosage increases. In G12 the cracking energy is
lightly higher but later the crack control capacity decreases, resulting in lower energy
compared with concretes incorporating other fibres. Polymer FRC show a particular
response, for COD lower than 1 mm the shape of the impact curve is similar to the
other FRC, between 1 and 2 mm a greater energy absorption capacity appears and,
finally, once it is exceeded COD > 2 mm, the VC increases again. The different
responses not only depend on the material of the filament, they are significantly
influenced by both the bond mechanism and the amount of fibres.
Fig. 4. Representative impact curves (left). Variation of toughness with CMOD during standard
bending tests (right).
For comparison Fig. 4 (right) represents the evolution of static toughness (calcu-
lated as the area below the load-CMOD curve) evaluated from static bending tests. As
expected, the values of energy are lower than those measured during impact tests;
impact loads are applied during a very short time and also part of the energy can be
dissipated by other mechanisms as vibrations or friction among others; it is well known
that the cracks propagation is time dependent. Although in general terms for each FRC
the increments in toughness are qualitatively consistent between static and impact tests,
it can be seen that for great crack openings polymer FRC behaves better during impact
tests than what could have been predicted based on static tests. While S25 or S50
achieved greater toughness in static bending than P5 or P10 respectively, the contrary
occurs when the impact curves are compared. Even though, this situation could vary
between fibres made of a same type of material, these results demonstrate that differ-
ential behaviours can appear. This highlights the necessity of a test for impact resis-
tance evaluation.
Table 2 presents the mean values of compressive strength (fc) as well as the first crack,
or limit of proportionality (fL) and the residual stresses fR1 and fR3, corresponding to
CMOD of 0.5 and 2.5 mm respectively, obtained in bending tests. In addition, the mean
values of the cracking and post-cracking energies (EC, EP), the total energy (ET), the
initial visible crack opening (CODC) and the COD growth rate (VC) from impact tests are
given. There are no significant differences in EC between all concretes, indicating that it
mainly depends on matrix strength. It was also confirmed that fibre incorporation reduces
the initial crack size even in the case of low toughness FRC. In the case of steel fibres
although more energy was applied, the initial opening was smaller in S50 than in S25; in
concretes incorporating polymer or glass macro fibres CODC was greater for the higher
dosages of fibres as greater impact energy was required to cause the initial crack as seen
when comparing the corresponding EC values.
Table 2. Summary of results from static and impact tests.

Concrete Static characterization Impact test
fc fL fR1 fR3 fR3/fR1 FRC class EC EP ET CODC VC
(MPa) (J) (lm) (mm/J)
R 44.2 4.04 – – – – 103 22 125 751 0.191
S25 44.5 4.75 3.27 2.93 0.89 3b 104 215 319 116 0.018
S50 44.8 4.42 5.36 4.87 0.91 5c 111 670 780 69 0.007
P5 47.3 4.38 1.77 1.94 1.10 2d 95 589 684 96 0.005
P10 46.3 4.21 2.67 3.67 1.38 2.5e 102 1183 1285 119 0.002
G6 47.1 4.51 1.81 0.90 0.50 2a 89 180 269 91 0.025
G12 46.6 4.83 2.88 1.67 0.58 3a 114 221 334 111 0.019
Analysing the results of VC it appears that both in steel and polymeric FRC the
COD growth rate decreases as the volume of fibres increases. The post cracking energy
EP is clearly associated with the COD growth VC as shown in Fig. 5.
Fig. 5. Variation of post-cracking energy (EP) with COD growth rate (VC).
Figure 6 plots the relationship between the results of impact tests and the residual
stresses fR1 and fR3. Contrary to the case of EC, which is practically independent of FRC
toughness, it can be seen that as the residual stresses increase the total cumulated energy
increases (Fig. 6 left). At the same time, although VC decreases as FRC residual capacity
increases, as expected, its values depend on the fibre type, and polymer FRC showed VC
results clearly lower than those obtained with other types of fibres (Fig. 6 right).
Fig. 6. Variation of ET (left) and VC (right) with the residual stresses fR1 (empty symbols) and
fR3 (filled symbols) calculated from standard bending tests.
Figure 7 shows the influence of the residual stresses fR3/fR1 ratio on impact test
results: the measured energies and the COD growth rate. The dashed vertical lines
indicate the limits for the fR3/fR1 ratio (a, b, c, d, e) used to characterise the shape of the
FRC post-cracking branch (hardening/softening) [17]. It should be noted that the FRC
studied cover all of the post-peak responses established in this Code. It can be seen that
as the residual stress ratio increases the cracking energy is not modified while the post-
cracking energy (and consequently the total energy) clearly increases (Fig. 7 left).
When the values of VC are plotted against fR3/fR1 ratio (Fig. 7 right), the rate strongly
decreases (R2 = 0.88). In both cases the results follow similar tendencies beyond the
type of fibre used. This would appear as an additional advantage of the classification
system adopted for FRC in the fib Model Code.
Fig. 7. Influence of fR3/fR1 ratio on impact test parameters. Left: effects on cracking, post-
cracking and total energies. Right: effects on VC.
4 Conclusions
A drop-weight test on notched prisms was used to evaluate the contribution of different
fibres on the impact resistance. Impact resistance was characterised in terms of cracking
(EC), post-cracking (EP) and total (ET) cumulated energy and the COD growth rate
(VC). Diverse classes of FRC obtained by incorporating different contents of steel, glass
and polymer macrofibres were analysed. Main conclusions for the studied FRC are
summarized as follows.
• A repeated drop-weight test was implemented which easily enables to evaluate both
the resistance to the first crack and the behaviour in cracked state of fibre concretes.
Impact test results among specimens of a same FRC show reasonable variability.
• The energy of cracking was not significantly affected by type and content of fibres.
Nevertheless, fibre incorporation reduced the initial crack size even in the case of
low toughness FRC.
• Increase in concrete impact toughness, expressed as total cumulated energy during
impact test, were mainly observed after matrix cracking, and for steel and polymeric
FRC.
• When comparing the post-cracking response some differences in the shape of the
impact curves were found as a function of the material type of the fibres. Poly-
meric FRC were particularly efficient at large crack openings.
• A consistent relationship between the residual stresses measured in the static
bending test and the post-cracking impact test results for the same fibre type was
observed. As the residual stresses increase the total cumulated energy increases and
VC decreases.
• The results of VC show an inverse tendency with the residual stresses ratio fR3/fR1
beyond the type of fibre incorporated in concrete, which would appear as an
additional advantage of the classification system adopted for FRC in the fib Model
Code.
At the present, these findings correspond to the studied FRC and further studies are
necessary for generalization of the conclusions. Experimental tests exploring the effects
of concrete strength and other varieties of steel, polymer and glass macrofibres are in
progress.
Acknowledgements. The authors thank the collaboration of Eng. Francisco Hours and Pablo
Bossio on the experimental works, funding from LEMIT-CIC and the projects CONICET
PIP112-201501-00861 and UNLP 11/I188.
References
1. Drdlová, M., Buchar, J., Krátký, J., Řídký, R.: Blast resistance characteristics of concrete
with different types of fibre reinforcement. Struct. Concr. 16, 508–517 (2015)
Experimental and numerical analysis of blast response of high strength fiber reinforced
concrete slabs. Eng. Struct. 175, 113–122 (2018)
Effect of steel fibers on static and blast response of high strength concrete. Int. J. Impact Eng.
107, 23–37 (2017)
4. Almusallam, T.H., Siddiqui, N.A., Iqbal, R.A., Abbas, H.: Response of hybrid-fiber
reinforced concrete slabs to hard projectile impact. Int. J. Impact Eng. 58, 17–30 (2013)
5. Yahaghi, J., Muda, Z.C., Beddu, S.B.: Impact resistance of oil palm shells concrete
reinforced with polypropylene fibre. Constr. Build. Mater. 123, 394–403 (2016)
6. Zhu, X.C., Zhu, H., Li, H.R.: Drop-weight impact test on U-shape concrete specimens with
statistical and regression analyses. Materials (Basel) 8, 5877–5890 (2015)
7. Hrynyk, T.D., Vecchio, F.J.: Behavior of steel fiber-reinforced concrete slabs under impact
load. ACI Struct. J. 111, 1213–1224 (2014)
8. ACI Committee 544: Measurement of Properties of Fiber Reinforced Concrete 544.2R-89.
In: ACI 544.2R (1999)
9. Banthia, N., Mindess, S., Trottier, J.: Impact resistance of steel fiber reinforced concrete.
ACI Mater. J. 93, 472–479 (1996)
10. Mohee, F.M.: The effects of strain rate on concrete strength under dynamic impact load.
J. Bangladesh Electron. Soc. 16, 83–90 (2016)
11. Radomski, W.: Application of the rotating impact machine for testing fibre-reinforced
concrete. Int. J. Cem. Compos. Light. Concr. 3, 3–12 (1981)
12. Zhang, X.X., Ruiz, G., Yu, R.C.: A new drop weight impact machine for studying the
fracture behaviour of structural concrete. WIT Trans. Built Environ. 98, 251–259 (2008)
13. Banthia, N.P., Mindess, S., Bentur, A.: Impact behaviour of concrete beams. Mater. Struct.
20, 293–302 (1987)
14. Bindiganavile, V., Banthia, N.: Polymer and steel fiber-reinforced cementitious composites
under impact loading, part 1: bond-slip response. ACI Mater. J. 98, 10–16 (1998)
15. Bindiganavile, V., Banthia, N.: Polymer and steel fiber-reinforced cementitious composites
under impact loading, part 2: flexural thoughness. ACI Mater. J. 98, 17–24 (1998)
16. Rahmani, T., Kiani, B., Shekarchi, M., Safari, A.: Statistical and experimental analysis on
the behavior of fiber reinforced concretes subjected to drop weight test. Constr. Build. Mater.
37, 360–369 (2012)
17. International Federation for Structural Concrete (fib): Model Code, vol. 1 (2012)
18. Technical Committee CEN/TC 229: EN 14651:2005 Test Method for Metallic Fibered
Concrete - Measuring the Flexural Tensile Strength (Limit of Proportionality (LOP),
Residual) Méthode (2005)
19. Vivas, J., Zerbino, R.L.: Estudio de la resistencia al impacto de hormigones reforzados con
fibras. In: 19 Congress International Metallurgy and Materials, CONAMET-SAM, Valdivia,
Chile, pp. 140–141 (2019)
20. American Society for Testing and Materials: ASTM E436 – 03 Standard Test Method for
Drop-Weight Tear Tests of Ferritic Steels. In: ASTM B. Stand. 91 (1997)
21. American Society for Testing and Materials: ASTM E208-17(2018) Standard Test Method
for Conducting Drop-Weight Test to Determine Nil-Ductility Transition Temperature of
Ferritic Steels. In: ASTM B. Stand. 06 (2000)
22. American Society for Testing and Materials: ASTM C 39 M:2003, Standard Test Method
for Compressive Strength of Cylindrical Concrete Specimens. In: ASTM B. Stand. 03 (2003)
An Experimental Study on the Fatigue Failure
Mechanisms of Pre–damaged Steel Fibre
Reinforced Concrete at a Single Fibre Level
Humaira Fataar1, Riaan Combrinck1, and William P. Boshoff2(&)

1
Unit for Construction Materials,
Stellenbosch University, Stellenbosch, South Africa
2
University of Pretoria, Pretoria, South Africa
billy.boshoff@up.ac.za
Abstract. In fibre reinforced concrete (FRC), energy is dissipated in the wake

of the crack tip through the actions of fibre bridging and fibre pull out. This is
the main mechanism which inhibits crack growth, thus increasing the load
carrying capacity of FRC by providing post–cracking ductility. Furthermore, the
same mechanism is present when FRC undergoes fatigue loading. Typical
applications for FRC which undergo significant fatigue loading during their
service life include paving applications such as bridge decks, highways and
industrial floors. The continuous exposure to cyclic loading results in a decrease
in apparent stiffness of the material, which may lead to fatigue failure [1, 2].
Fatigue failures are almost always unexpected, and can have a catastrophic
outcome [3, 4]. Thus, the fatigue characteristics become vital performance and
design parameters [1]. In this paper, the mechanisms of fatigue failure of pre-
damaged hooked-end steel fibre reinforced concrete (SFRC) are investigated at a
single fibre level. An initial pre-damage was applied to the fibres before the
cyclic loading commenced. The pre–pull out ranged from 0.6 mm to 2.5 mm,
and the cyclic loading was applied at 70% and 85% of the maximum static pull
out capacity of the fibre embedded in the concrete.
Keywords: FRC Fatigue Single fibre Steel fibres Cyclic loading
1 Introduction
As concrete technology progresses, concrete’s compressive strength is constantly being

increased. However, an increase in compressive strength results in a more brittle
concrete which is prone to sudden failures [5]. Since concrete is known for its weak
tensile strength, steel is added in the form of reinforcing bars or fibres to contribute to
the tensile capacity of the composite [6]. The use of discontinuous fibres in cementi-
tious materials have been known to suppress crack development and slow down crack
propagation, improve the composite durability as well as its fatigue life [7–13]. The
fibres’ bridging activity is only activated at crack initiation, and only then does it
contribute to the load carrying capacity. The load gets transferred from the concrete
matrix to the fibre by shear deformation at the fibre-matrix interface [7, 14, 15].
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_18
200 H. Fataar et al.
Fatigue is defined as the process of continuous, permanent internal structural

changes to a material which is subjected to cyclic loading. In concrete, the internal
structural changes are associated with the progressive growth of internal micro cracks [1,
16, 17]. A continuous exposure to cyclic loading results in decreased apparent stiffness,
which may lead to fatigue failure [1, 2]. The fatigue loading can be divided into two
categories, low-cycle loading and high-cycle loading. The low-cycle loading consists of
fewer load cycles which are applied at high stress levels. In contrast, the high-cycle
fatigue loading consists of a large number of load cycles at low stress levels [1].
Generally, there are three main failure mechanisms for fibre reinforced concrete
(FRC). These include fibre pullout, fibre failure, and matrix failure [14, 18]. Fibre
pullout is the most common type of failure mechanism. It occurs when the bond
mechanisms, which includes adhesion, friction, and mechanical interlock, have failed
[19]. Fibre failure occurs when the fibre ruptures, and matrix failure occurs when the
failure strain of the matrix has been reached [14].
The purpose of this research is to use single fibre pullout tests in order to under-
stand the failure mechanisms involved with pre-damaged fatigue SFRC.
2 Experimental Method
2.1 Material Properties and Concrete Mix

The materials used for the concrete mix, excluding the steel fibres, were locally
sourced. The concrete mix design is displayed in Table 1. The cement used was a
CEM II 52.5 N, supplied by Pretoria Portland Cement. The coarse aggregate used in
the concrete was a 13 mm crushed Greywacke stone, and a natural pit sand (locally
called Malmesbury sand) was used as a fine aggregate. Municipal water was used,
along with Chryso Optima 206, which was a high range water reducing superplasti-
ciser. The steel fibres were supplied by BEKAERT and their DRAMIX 3D–65/60-BG
hooked-end fibres were used in this study. Table 2 provides the properties of the steel
fibres.
Table 1. Concrete mix composition

Material kg/m3
Cement 450
Sand 880
Stone 820
Water 194
Superplasticiser (1.4% by weight of cement) 6.3
An Experimental Study on the Fatigue Failure Mechanisms 201
Table 2. Fibre properties (DRAMIX 3D-65/60-BG) [20]

Tensile strength 1 160 MPa
Modulus of elasticity 200 GPa
Length (l) 60 mm
Diameter (d) 0.9 mm
2.2 Specimen Preparation

The specimen preparation was done in two states – fresh state and the hardened state.
During the fresh state, the concrete constituents were mixed together in a pan mixer.
The concrete mix was designed to be used for steel fibre reinforced beams as well as
the single fibre pull out tests. The concrete was then sieved on a shaker table through a
4.75 mm sieve to collect the mortar alone.
The test specimens were created by dividing 100 mm cube moulds into quarters.
The moulds were divided using a wooden cross-shaped partition placed in the centre of
the mould as displayed in Fig. 1. This resulted in four specimens, each measuring
40 40 100 mm3. The fresh mortar was cast into the moulds and was vibrated for
30 s on a shaker table. No fibres were added to the concrete mix, only a single fibre was
embedded in each specimen after the mortar was cast. The single fibre was placed in
the centre of the specimen to a fixed embedment length of l/2 (30 mm), and was
vibrated once more to close any voids that could have developed when inserting the
fibre. All the specimens were then placed in an undisturbed part of laboratory and was
demoulded after 24 h. Once demoulded the fibre on each specimen was coated with a
wax to protect it from corrosion while in the curing tank. All the specimens were placed
in a curing tank to cure for an additional 26 days, completely submerged in water at
24 ± 1 °C.
Fig. 1. Specimen mould
One day prior to testing, the specimens were removed from the curing tanks and
allowed to dry, after which the wax was removed from the fibre with a clean cloth. The
exposed hooked-end of the fibre was carefully removed using pliers to allow the fibre
to be clamped onto the test setup. Once the specimens were dry, it was glued onto steel
plates with a two-part epoxy provided by Sika called Sikadur–30. The epoxy took 24 h
to set and the specimens were tested between 28 and 40 days.
2.3 Test Setup

The tests were conducted on a 50 kN servo–controlled hydraulic actuator, which was
fixed onto a steel frame. An external 5 kN load cell was added to the test setup to
ensure a better resolution on the load control of the test, since the maximum pullout
loads were expected to be below 5 kN based on similar pullout tests conducted at
Stellenbosch University [21]. Additionally, an external linear variable differential
transducer (LVDT) was incorporated in the test setup to verify the crosshead dis-
placement readings provided by the actuator. Once the specimens were ready to be
tested, the steel plate with a single specimen on it was bolted down onto a box section
which was fixed to two channel sections of the steel frame. The crosshead of the
actuator was then lowered into position until the exposed fibre was completely inserted
into the clamp. The fibre clamp was then tightened by two grub screws which held the
fibre in place. The test setup can be seen in Fig. 2.
Fig. 2. Test setup
2.4 Loading Regime

Before the fatigue cyclic loading commenced, static pullout tests were conducted. This
was done in order to obtain the average maximum pullout load. The static tests were
position-controlled and it measured the load required to completely pull a fibre

embedded at l/2 out of the mortar. Once the static tests were completed, the cyclic
fatigue pullout tests could commence. The fatigue loading was applied by a sinus load
application as displayed in Fig. 3. The maximum load (Amax) was taken as a percentage
of the average maximum static pullout load (70% or 85%). The minimum load (Amin)
was specified as a small positive value (generally 20 N) in order to maintain cyclic
tensile loads. The fatigue tests were loaded up to 2 million cycles at a stable frequency
of 6 Hz.
Load (N)
Amax
Mean
Amin
0 Time (s)
Fig. 3. Sinus loading
The cyclic tests were performed in three steps: (1) pre-damage; (2) load/unload; and
(3) cyclic load. The pre-damage step consisted of an initial pullout of the fibre to a
certain predetermined degree. This step was position-controlled and the pullout was set
by the user. The second step was the load/unload step, which required the pullout load
attained in the first step to either increase or decrease in order to reach the specified
mean value of the sinus load. The third step was the fatigue loading, which cycled from
the mean value to the maximum and minimum amplitude until the fibre completely
pulled out, or reached the 2 million cycle limit set. Both the second and third steps were
load-controlled.
3.1 Static Tests

Before any fatigue tests could be performed, it was necessary to conduct static pullout
tests. This was done in order to find the average maximum load required to completely
pull out the hooked-end fibre. A total of six specimens were tested, and the results are
plotted in Fig. 4. The test results followed the same trend for all the specimens. The
load increased rapidly before it reached a small peak at a displacement of around
0.1 mm, followed by a relaxation in the load. The load then increased further until it
reached a global peak at a displacement of 0.9 mm, then it started to decrease slightly.
It was then followed by another smaller increase in load to a less notable peak. The load
then begins to decrease steadily until the fibre has completely pulled out. The first peak
was possibly due to the initial fibre debonding, but further investigation is required.
The global peak was then associated with further fibre debonding and the initiation of
the hook straightening. The last peak was due to the fibre straightening out completely
and eventually pulling out. Similar trends can be found in literature [15, 22, 23].
Fig. 4. Static pullout tests
3.2 Fatigue Tests

The fatigue tests were conducted on specimens at load levels of 70% and 85%, at a pre-
damage of 0.6 mm, 1.2 mm, 1.8 mm, and 2.5 mm. The tests were conducted until
failure, or until the 2 million cycle limit was reached. Unless otherwise specified, three
specimens were tested for each load level percentage and pre-damage. Figure 5
illustrates the average number of cycles sustained for each test group. A higher load
level coupled with a higher pre-damage resulted in early failure, as can be seen with the
85% load level and 2.5 mm and 1.8 mm pre-damage. The 1.2 mm and 0.6 mm pre-
damage specimens could still withstand the 85% load level to a certain extent. How-
ever, only a single specimen did not fail. On average, the 70% load level tests were able
to withstand a large number of load cycles before any type of failure occurred. Only
one test batch – 70% at 1.2 mm pre–damage – sustained the full 2 million cycles
without any failure.
The results in Table 3 indicates the failure mechanisms for each test. Two types of
failure was experienced – fibre pullout and fibre rupture. The pullout failure mechanism
was mostly experienced by specimens with the higher pre-damage, since the fibre has
already debonded and the hook begins to deform [23, 24]. At the lower pre-damage, the
likelihood of fibre rupture increases, due to the fact that the fibre hook is still intact. It
can also be noted that the 85% load level resulted in more pullout failures. The 70%
70
Load level (%)
0.6 mm
1.2 mm
1.8 mm
85 2.5 mm
1 100 10 000 1 000 000

Average cycles
Fig. 5. Average cycles endured by each test group
load level had an equal number of fibre rupture and fibre pullout failures, but having no
failure dominate the test results.
Table 3. Failure mechanism summary

Load level Pre-damage (mm) Failure mechanisms
70% 2.5 2 pullout; 1 no failure
1.8 1 rupture; 2 no failure
1.2 4 no failure
0.6 1 rupture; 2 no failure
85% 2.5 3 pullout
1.8 2 pullout; 1 rupture
1.2 1 pullout; 2 rupture
0.6 1 pullout; 1 rupture; 1 no failure
3.3 Computed Tomography (CT) Scans

The CT scans provided a non-destructive method of assessing the internal structure of
the specimens both before and after the fatigue testing. Initially, CT scans were taken of
specimens with a pre–damage and no fatigue testing.
The specimens in Fig. 6 were damaged to 0.6 mm to determine the actual measured
degree of fibre displacement and compare it with the applied pre-damage. It was found
that the actual measured displacement of the tip of the fibre was slightly less that the
applied pre-damage. Additionally, there was the same trend in delamination observed
at the curved portion of the fibre for specimens A, B and C. This was caused by the
fibre debonding from the surrounding matrix as the pre-damage load was applied.
Fig. 6. CT scans for three specimens at 0.6 mm pre-damage
CT scans were also performed on two failed specimens after the fatigue loading as
displayed in Fig. 7. Specimen D and Specimen E were pre-damaged to 1.2 mm before
the fatigue loading of 85% load level was applied. Specimen E was found to have
ruptured at the start of the hooked-end which was embedded in the matrix, whereas
Specimen E ruptured at the surface of the specimen. Upon further investigation at the
surface of Specimen E displayed in Fig. 8, it was found that there was a network of
micro cracks. This indicated that the stress concentration at that point was high, thus
resulting in crack propagation and ultimately, fibre rupture at the surface.
Fig. 7. Post-test CT scan of ruptured fibres

Fig. 8. Side view of Specimen E
4 Conclusions
Fatigue tests at a single fibre level were conducted on pre-damaged specimens. Using
hooked-end steel fibres embedded at l/2 into a mortar, the failure mechanisms were
found for specimens tested at load levels of 70% and 85% of the average maximum
static load. The pre-damage varied for each set of tests between 0.6 mm to 2.5 mm.
The following conclusions can be made:
• The main failure mechanisms were found to be fibre pullout and fibre failure in the
form of fibre rupture.
• At the largest pre-damage of 2.5 mm, the dominating failure mechanism was fibre
pullout, due to the fact that the fibre hook anchorage has already been deformed and
there is not much resistance left.
• The lower pre-damage experienced a mixture of fibre pullout as well as fibre rupture
occurring.
• Overall, the 85% load level experienced the most failures, with only one of twelve
specimens not failing. In contrast, the 70% load level experienced four specimen
failures out of thirteen specimens.
References
1. Lee, M.K., Barr, B.I.G.: An overview of the fatigue behaviour of plain and fibre reinforced
concrete. Cement Concr. Compos. 26(4), 299–305 (2004)
2. Kolluru, S.V., O’Neil, E.F., Popovics, J.S., Shah, S.P.: Crack propagation in flexural fatigue
of concrete. J. Eng. Mech. 126(9), 891–898 (2000)
3. Bathias, C., Pineau, A.: Fatigue of Materials and Structures. ISTE , Wiley, London,
Hoboken (2010)
4. Anderson, T.L.: Fracture Mechanics: Fundamentals and Applications. CRC Press, Boca
Raton (2017)
5. Boulekbache, B., Hamrat, M., Chemrouk, M., Amziane, S.: Flexural behaviour of steel fibre-
reinforced concrete under cyclic loading. Constr. Build. Mater. 126, 253–262 (2016)
6. Hsu, T.T.C.: Fatigue of plain concrete. Am. Concr. Inst. 78(4), 292–305 (1981)
7. Namur, G.G., Alwan, J.M., Najm, H.S.: Fiber pullout and bond slip I: analytical study.
J. Struct. Eng. 117(9), 2769–2790 (1992)
8. Banthia, N., Trottier, J.F.: Deformed steel fiber—cementitious matrix bond under impact.
Cem. Concr. Res. 21(1), 158–168 (1991)
9. di Prisco, M., Plizzari, G., Vandewalle, L.: Fibre reinforced concrete: new design
perspectives. Mater. Struct. 42(9), 1261–1281 (2009)
10. Buratti, N., Mazzotti, C., Savoia, M.: Post-cracking behaviour of steel and macro-synthetic
fibre-reinforced concretes. Constr. Build. Mater. 25(5), 2713–2722 (2011)
11. American Concrete Institute: Report on fiber reinforced concrete (2002)
12. Zhang, J., Stang, H., Li, V.C.: Crack bridging model for fibre reinforced concrete under
fatigue tension. Int. J. Fatigue 23(8), 655–670 (2001)
13. Yoo, D.Y., Kim, S., Park, G.J., Park, J.J., Kim, S.W.: Effects of fiber shape, aspect ratio, and
volume fraction on flexural behavior of ultra-high-performance fiber-reinforced cement
composites. Compos. Struct. 174, 375–388 (2017)
14. Beaudoin, J.J.: Handbook of Fiber-Reinforced Concrete : Principles Properties, Develop-
ments and Applications. Noyes Publications, Park Ridge, N.J. (1990)
15. Ghoddousi, P., Ahmadi, R., Sharifi, M.: Fiber pullout model for aligned hooked-end steel
fiber. Can. J. Civ. Eng. 37(9), 1179–1188 (2010)
16. Parvez, A., Foster, S.J.: Fatigue of steel-fibre-reinforced concrete prestressed railway
sleepers. Eng. Struct. 141, 241–250 (2017)
17. Keerthana, K., Chandra Kishen, J.M.: An experimental and analytical study on fatigue
damage in concrete under variable amplitude loading. Int. J. Fatigue 111, 278–288 (2018)
18. Naaman, A.E.: Fiber reinforced concrete under dynamic loading. In: Fiber Reinforced
Concrete International Symposium, vol. 81, pp. 169–186 (1984)
19. Naaman, A.E., Najm, H.: Bond-slip mechanisms of steel fibers in concrete. ACI Mater. J. 88
(2), 135–145 (1991)
20. Dramix® steel fiber concrete reinforcement - Bekaert.com. https://www.bekaert.com/en/
products/construction/concrete-reinforcement/dramix-steel-fiber-concrete-reinforcement.
Accessed 21 Feb 2018
21. Nieuwoudt, P.D.: Time-dependent Behaviour of Cracked Steel Fibre Reinforced Concrete,
March 2016
22. Nieuwoudt, P.D., Boshoff, W.P.: Time-dependent pull-out behaviour of hooked-end steel
fibres in concrete. Cement Concr. Compos. 79, 133–147 (2017)
23. Marković, I.: High-performance Hybrid-fibre Concrete : Development and Utilisation. Delft
University, the Netherlands (2006)
24. Cunha, V.M.C.F., Barros, J.A.O., Sena-Cruz, J.: Pullout behaviour of hooked-end steel
fibres in self-compacting concrete (2007)
Development of an HPFRC for Use
in Flat Slabs
Julia Blazy, Sandra Nunes(&), Carlos Sousa, and Mário Pimentel
CONSTRUCT-LABEST, Department of Civil Engineering,

Faculty of Engineering, University of Porto, Porto, Portugal
snunes@fe.up.pt
Abstract. Fibre-reinforced cementitious materials represent one of the most

significant developments in the field of concrete technology of the last decades.
The improved performance of this new class of materials (in terms of worka-
bility, compressive strength, flexural/tensile behaviour and/or durability) allows
rethinking several of the existing structural solutions. This paper describes
research on high-performance fibre reinforced concrete (HPFRC) to be used at
the slab-column connection zones of flat slabs, in order to improve its punching
shear resistance. Design of Experiments (DoE) approach was used to design
HPFRC paste and aggregate particle phases. As such, a central composite design
was carried out to mathematically model the influence of mixture parameters
and their coupled effects on deformability, viscosity and compressive strength.
After that, a numerical optimization technique was applied to the derived models
to select the best mixture, which simultaneously, maximizes aggregates content
and allows achieving a compressive strength of 90–120 MPa, while maintaining
self-compactability (SF1 + VS2), incorporating 1% steel fibres content.
Keywords: High-Performance Fibre Reinforced Concrete (HPFRC) Mix-

design Self-compacting Flexural behaviour Steel fibres
1 Introduction
During the last three decades, concrete has emerged from a rather simple mass con-
struction material towards a high-performance material, which can be tailored for
specific applications and according to user specifications. The rapid evolvement in
terms of enhanced workability (self-compacting ability) and compressive strength was
made possible by the development of superplasticizers and mix-design methods. At
increasing strength, concrete becomes more brittle, which can be compensated for by
the incorporation of fibres. The fibres can increase the ductility in compression as the
fibres transfer stress during cracking. Figure 1 maps the most common fibre-reinforced
cementitious composites with regard to compressive strength and fibre content, namely,
engineered cementitious composites (ECC), ultra-high fibre performance fibre rein-
forced composites (UHPFRC), high-performance fibre reinforced concrete (HPFRC)
and traditional fibre reinforced concrete (FRC).
The potential for an increase in compressive strength due to an increase in fibre
content is less significant compared to the tensile strength [1, 12]. Fibres can improve
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_19
210 J. Blazy et al.
Fig. 1. Most common range of different fibre-reinforced cementitious composites, concerning

compressive strength and fibre content
the tensile and flexural behaviour of FRC; its effect depends on the content, fibre type,
fibre orientation and distribution, and matrix strength [1]. The use of steel fibre
hybridisation has also been found to enhance the pre- and post- cracking response of
FRC [2]. The performance of FRC in tension usually improves at increasing fibre
content; but beyond a threshold value, the performance might even decrease due to the
loss of workability and entangling of interacting fibres during pull-out (lower efficiency
of the fibres). The effect of coarser aggregates in this context is also crucial concerning
the performance in both the fresh and the hardened states. The addition of fibres in
concrete affects its characteristics in the fresh state due to the larger surface area of
fibres, which requires more paste to envelop and lubricate the fibres [12]. In particular,
stiff fibres decrease the packing density of the aggregates; this effect is more pro-
nounced the larger the aggregates are, compared to the fibre length [12]. In most cases,
the FRC mix-design is a trade-off between workability requirements, desired post-
cracking behaviour and economical aspects. Since fibres increase the material costs,
optimum fibre content has to be determined.
The current study deals with the materials selection, formulation and optimization
of an HPFRC mixture to be used in a localised area at the slab-column connection
zones of flat slabs (the rest of the slab being cast with normal-strength concrete), in
order to improve its punching shear resistance. The strategy followed in this study was
to optimize the two main parts of the composite separately. Firstly, optimise the paste
phase and the aggregate particle phase to achieve a target compressive strength of 90–
120 MPa and self-compacting ability. A fixed fibre content and a single fibre type were
used at this stage. Secondly, combining the optimized concrete with a hybrid fibres
mixture (new hooked-ends macro steel fibres + straight micro steel fibres) and assess
its influence on flexural behaviour, in order to find the best combination. The current
paper reports only on the first part of the study.
Some literature studies on HPFRC with compressive strengths close to the target
range of 90–120 MPa are reported in Table 1. Regarding the performance in the fresh
Development of an HPFRC for Use in Flat Slabs 211
state, these include conventional vibrated concretes (slump: 17.5 and 13.0 cm [6]) as
well as self-compacting concretes (slump flow: 71–78 cm [2]; 76 cm [7]; 65–75 cm
[11]). Compressive strength and fibre content data from these studies are also plotted in
Fig. 1, identifying the zone where most HPFRC mixtures can be found. It can be
observed that in most HPFRCs the fibre content ranges from 0.5% to 1.5%.
Table 1. Characteristics of several HPFRC materials from different studies.
Ref. Binder w/b dmax Vf lf/df; geometry fcm,28d

(mm) (MPa)
[1] AST M T ype I 0.25 10 0.5%; 1%; 1.5% 40/0.62; 1HE 88; 90; 93
cement (w/c)
0.5%; 1%; 1.5% 50/0.62; 1HE 87; 91; 93
0.5%; 1%; 1.5% 60/0.75; 1HE 87; 92; 94
[2] CEM I 52.5R 0.27 10 a 0.75%+0% 60/0.9; 1HE+13/0.20; S 88
(w/c)
0.562%+0.188% 84
0.375%+ 0.375% 83
0.188%+0.562% 88
0%+0.75% 98
1.0%+0% 60/0.9; 2HE+13/0.20; S 89
0.75%+0.25% 89
0.5%+0.5% 87
0.25%+0.75% 91
0%+1.0% 98
[3] PC+SF+FA 0.25 13 1.25% 30/0.50; 1HE 94
[4] Cement CPN50 0.30 10 0.76% -; - 95
(w/c)
[5] PC 42.5 T ype I+ 0.24 12.5 b 0.2% 28/0.40; 1HE 102
SF (w/c)
0.4% 100
0.6% 102
1.0% 103
[6] OPC+SF 0.295 10 c 1% 13/0.18; S 103
2% 114
[7] CEM+mS+LP 0.24 10 b 0.5% 30/0.55; C 100
[8] OPC+zSF+FA 0.195 10 d 0.50% 30/0.50; 1HE 93
1% 108
[9] PC T ype II+LP 0.40 12.7 1.6% 36/0.70; 1HE 91
(w/c)
[10] CEM II/B-M (S- 0.20 11 e 1%+0.5% 13/0.20+30/0.60 109
V) 42.5 N+ SF
0.18 8e 3% (total) 13/0.20+30/0.60+50/0.1 122
0.18 11 e 0.5%+0.5%+0.5% 13/0.20+30/0.60+50/0.1 95
[11] CEM I 42.5R 0.32 8c 0.7%+0.3% 6/- + 30/- 130-140
HSR LA+SF (w/c)
[12] CEM I 52.5R+ 0.20 8c 2% 13/0.20; S 164
mS+LP (w/p)
(O)PC: (ordinary) Portland cement; (z)SF: (zirconium) silica fume; mS: micro silica; FA: fly ash; LP: limestone
powder; w/b: water to binder weight ratio; w/c: water to cement weight ratio; w/p: water to powder weight ratio;
dmax: maximum diameter of coarse aggregate; Vf: fibre content; lf fibre length; df:fibre diameter fcm,28d: mean
compressive strength at 28 days; 1HE: simple hooked-end; 2HE: double hooked-end; S: straight; C: crimped ends;
nature of coarse aggregate: ( a) flint, ( b ) limestone, ( c) basalt, ( d ) granite, ( e) diabase
212 J. Blazy et al.
2.1 Design of Experiments Strategy
The experiments were planned according to a Central Composite Design (CCD), which
consists of three distinct sets of experimental runs: factorial (Fi), central (Ci) and axial
(CCi) [13]. This approach helps to avoid the trial and error method and as a result,
reduces the time necessary to develop the mixture of desired properties by collecting
and statistically analysing relevant data. Additionally, it allows identifying design
variables that have the most significant influence on the response variables, as well as
the interactions between them. This design of experiments (DoE) technique involves
the following steps: (1) choosing the mix-design variables and response variables;
(2) selecting design variables’ ranges; (3) conducting experiments and collecting data;
(4) statistical analysis of data and fitting the empirical model, by means of regression
analysis; (5) optimizing mixture proportions. In this work, the research program was
first conducted at the mortar level and then on the concrete level to minimize at the
beginning the number of design variables and therefore to reduce the number of
experiments required to create the empirical model. After choosing the optimum paste
composition, resulting from the study at the mortar level, the research on the concrete
level was conducted.
Regarding the mortar stage, four key design variables were chosen: water to
powder volume ratio (Vw/Vp); water to cement weight ratio (w/c); superplasticizer to
powder weight ratio (Sp/p) and silica fume to cement weight ratio (SF/c) and four
response variables: slump flow diameter (Dflow), V-funnel flow time (Tfunnel),
compressive strength after 28 (fc,28d) and 98 days (fc,98d). It must be noted that sand
to mortar volume ratio (Vs/Vm) and fibre content (Vf) were set fixed and equal to 0.475
and 1.5%, respectively. For concrete level, two key design variables were selected:
solid volume (Vg/Vg,lim), sand to mortar volume ratio (Vs/Vm) and four response
variables: time necessary to reach a 500 mm diameter in the slump flow test (t500),
slump flow diameter (Dflow), compressive strength after 7 (fc,7d) and 98 days (fc,98d).
The fibre content (Vf) was fixed and equals 1.0%. The main reason to incorporate the
steel fibres at this stage of the study was to avoid brittle failures in compression and
reduce the variability of test results. More information about the formulation of the
mixtures can be found elsewhere [13]. The effect of design variables was established on
five levels: - a, -1, 0, +1, + a. Table 2 characterizes the experimental plans used for
both the mortar and concrete studies.
After collecting the experimental data, the commercial software Design-Expert was
used to analyse the results for each response variable namely, to examine the summary
data plots, to fit 2nd order polynomial models using regression analysis and ANOVA, to
validate the models by examining the residuals for trends, autocorrelation and outliers,
as well as to interpret the obtained model graphically.
Table 2. Experimental plans characterization
M aterials Type of plan Design variables [-α; +α] Response

variables
CEM I 42.5R 24Fi+8CCi+6Ci Vw/Vp [0.450; 0.650] Dflow
Limestone powder α= 2.0 w/c [0.250; 0.350] Tfunnel
Mortar level
Silica fume nc = 6 Sp/p [1.5%; 2.5%] fc,28d

Superplasticizer nf= 16 SF/c [7.5%; 17.5%] fc,98d
standard sand na = 8
Fibres (9/0.175;S)
Tap water
CEM I 42.5R 22Fi+4CCi+4Ci Vg/Vg,lim [0.440; 0.510] t 500
Limestone powder α= 1.414 Vs/Vm [0.365; 0.435] Dflow
Concrete level
Silica fume nc = 4 fc,7d

Superplasticizer nf = 4 fc,98d
Sand na = 4
Coarse aggregate
Fibres (13/0.2;S)
Tap water
2.2 Constituent Materials

In the study at mortar level, the following materials were used: Type I Portland cement
(CEM I 42.5R); limestone powder Betocarb® HP-OU (LP); silica fume Elkem 940-U
(SF); standard sand; superplasticizer Sika® ViscoCrete® 20HE (Sp); straight steel
fibres (lf = 9 mm and df = 0.175 mm, fy = 2100 MPa) and tap water. The specific
gravity of cement, LP and SF were 3.07, 2.68 and 2.40 g/cm3, respectively. The
standard sand is siliceous natural sand with particles of diameter from 0.08 mm to
2 mm and round shape. Furthermore, specific gravity equals 2.63 g/cm3 and water
absorption 0.3% by mass. Regarding superplasticizers, the supplier provided the fol-
lowing information: specific gravity of 1.08 g/cm3 and solid content of 40%.
Concerning the study at the concrete level, the same type of cement, LP, SF, Sp and
tap water were used. However, standard sand was replaced by real aggregates: natural
sand with a specific gravity of 2.58 g/cm3 and water absorption of 0.7%; and a crushed
coarse aggregate (dmax = 8 mm; amphibolite rock) whose specific gravity equals
3.02 g/cm3 and water absorption is 0.6%. All concrete mixtures incorporated 1% of
steel fibres (lf = 13 mm and df = 0.2 mm, fy = 2500 MPa), by volume.
214 J. Blazy et al.
2.3 Mixing Sequence and Testing Methods

Mortar mixes were tested in 1.40L batches at the low speed. The mixing procedure is
presented in Table 3. To characterize the fresh state properties of mortar mixes (vis-
cosity and flowability) the V- funnel and slump flow tests were carried out, respec-
tively, according to Okamura & Ouchi proposal [14]. Firstly, the V-funnel was filled
and the time that mortar needed to pass through the funnel opening was measured.
Secondly, the slump flow test was conducted. At the beginning, the cone was filled and
lifted with a constant speed. Then, after the flow of the mixture stopped, two diameters
in perpendicular directions were measured. The final result (Dflow) is an average of
four diameters because the test was repeated two times. Finally, four prisms of
dimensions 40 40 160 mm3 were cast to evaluate the compressive strength after
28 and 98 days. All the samples were covered with a plastic sheet to avoid drying and
water evaporation. After 24 h, specimens were demoulded and placed underwater in a
climatic chamber at 20 ± 2 °C, until the testing date. In the meantime, after 7 days, all
the prism samples were cut into cubes of dimensions 40 40 40 mm3. As a result,
12 samples for each mixture were obtained: 6 for testing the compressive strength after
28 days and 6 after 98 days. Afterwards, the average of these measures was calculated:
fc,28d and fc,98d.
Table 3. Mortar mixing sequence
Task Duration
Add sand + cement + limestone powder + silica fume + 80% of total 2.5min+ * +2.5min
mixing water
Add 75% of superplasticizer + 20% of total mixing water 2.5min + *
Add 25% of superplasticizer 1.5min
Add fibers 2.0min+ * +1.0min
* Stop to scrape off the material adhering to the mixing walls
Concrete mixes were tested in 25L batches in an open pan mixer. Steps of mixing
procedure are presented in Table 4. The slump flow test was carried out according to
EN 12350-8. When the cone was withdrawn upwards, the time from commencing the
upward movement of the cone to when the concrete had flowed to a diameter of
500 mm was measured (t500). Then, after the flow of the mixture stopped, two diam-
eters in perpendicular directions were measured and their average was calculated
(Dflow). Finally, six cubes of dimensions 150 150 150 mm3 were cast to eval-
uate the compressive strength after 7 and 98 days. All the samples were covered with a
plastic sheet to avoid drying and water evaporation. After 24 h, the specimens were
demoulded and placed underwater and kept at 20 ± 2 °C until the testing age. Three
samples were used to obtain compressive strength after 7 days and three after 98 days.
Afterwards, the average of these measures was calculated: fc,7d and fc,98d.
Table 4. Concrete mixing sequence
Task Duration
Add sand + 25% of total mixing water + coarse aggregate 2.5min
Waiting for absorption 2.5min
Add cement + limestone powder + silica fume + 75% of total mixing 5.0min + *
water + superplasticizer
Add fibers 3.0min
* Stop to scrape off the material adhering to the mixing walls
3 Experimental Campaign at Mortar Level

3.1 Collected Experimental Data
In this section, the results from tests carried out on mortar level are presented and
discussed. Figure 2.a shows the range of results obtained in the fresh state. In general,
the mortars exhibited very good flowability (Dflow 293 mm) and relatively high
flow times, which indicates a low risk of segregation. In Fig. 2.b it can be seen that
almost all mixtures experienced an increase of compressive strength from 28 to 98
days, ranging from 1.0% to 8.1%. For one mixture, the compressive strength decreased
by about 1.1%. This can be a result of local defects of the tested samples.
Fig. 2. Mortar test results: a) Tfunnel .vs. Dflow b) fc,98d.vs. fc,28d
3.2 Statistical Analysis of Data, Fitted Models and Optimisation

Table 5 presents the results of the fitted models, including the residual error term (e),
along with the correlation coefficients. An analysis of variance showed that these
models are significant when describing the effect of Vw/Vp, w/c, Sp/p and SF/c on the
216 J. Blazy et al.
modelled responses. Since R2 and R2adj values are significantly high, a large proportion
of the variability of response variables is explained by the obtained regression models.
The estimates of the model coefficients presented in Table 5 indicate the relative
significance of the various design variables on each response. Higher values indicate
greater influence of the variable in the response and, on the other hand, a negative value
reflects a response decrease to an increase in the design variable. Results in Table 5
clearly show Vw/Vp and w/c are the most influencing variables on the fresh state
properties, while the compressive strength is influenced mainly by w/c and SF/c.
Based on the regression models presented in Table 5, the numerical optimization
technique was used to determine the range of mortar mixture parameters that would
lead to adequate self-compacting concrete (SCC) mixtures, with Tfunnel in between 12
and 14 s, a target compressive strength of 150 MPa and minimum cost of paste. No
constraint was added concerning Dflow because all mixtures exhibited very good
deformability. The selected optimised solution is presented in Table 6, in terms of the
actual values. The optimised mortar mixture was also prepared and tested in the lab.
The experimental results obtained compare well with the predicted values, which
confirms the accuracy of the obtained fitted models.
Table 5. Fitted numerical models at mortar level (coded values of parameters)
Dflow Tfunnel fc,28d fc,98d

(mm) (s) (M Pa) (M Pa)
Independent 322.931 9.993 147.534 155.585
Vw/Vp +11.439 -3.340 -1.172 -1.594
w/c +7.915 -1.278 -5.486 -4.994
Sp/p NS -0.367 NS +0.558
SF/c -3.019 +0.115 +3.728 +3.715
Vw/Vp×w/c NS NS NS +1.144
w/c×Sp/p NS +0.236 NS -1.246
w/c×SF/c NS NS NS +1.225
Sp/p ×SF/c NS -0.251 NS NS
(Vw/Vp) 2 -1.903 +0.764 NS NS
(w/c)2 NS +0.367 NS NS
(Sp/p)2 NS +0.264 NS -1.582
ε, std. dev.* 4.193 0.371 1.898 1.795
R2 / R2adj 0.910 / 0.895 0.988 / 0.983 0.912 / 0.902 0.924/0.895
(NS) non-significant terms; (*) error term is a random and normally distributed
variable with mean zero
Table 6. Optimised mortar mixture and corresponding predicted and measured test results
Design variables Dflow Tfunnel fc,28d

(actual values) (mm) (s) (M Pa)
Vw/Vp=0.536 M easured 315.5 15.2 156.3
w/c=0.273 Predicted 314.0 14.0 150.0
Sp/p=1.59% 95% Prediction Interval- Low 304.0 12.8 145.6
SF/c=0.100 95% Prediction Interval-High 324.0 15.2 154.4
Predicted/M easured 0.99 0.92 0.96
4 Experimental Campaign at Concrete Level

4.1 Collected Experimental Data
At the concrete level, concrete mixtures composition was obtained by considering the
paste mixture proportions defined in Table 6 and replacing reference sand by real
aggregates (fine and coarse aggregates described above). Tests on concrete are then
necessary to optimize the aggregate skeleton and aggregates content. The DoE tech-
nique can also be applied at this stage to optimize concrete mixtures, considering as
independent variables, only these variables related to the aggregates- Vg/Vg,lim and
Vs/Vm- as detailed in Sect. 2.2.
The results from the tests carried at the concrete level are presented in Fig. 3. These
results show that the designed experimental plan covers a wide range of Dflow (Fig. 3.
a), including mixes that can be classified from SF1 to SF3 consistency classes. For all
mixtures t500 is higher than 2 s, corresponding to VS2 class, according to EN 206-9.
The relation between compressive strength after 7 days and slump flow diameter is
plotted in Fig. 3.b. The range of fc,7d results obtained is very narrow (about 8 MPa),
meaning that the changes introduced in the aggregate skeleton do not affect signifi-
cantly the compressive strength. In Fig. 4 the slump flow of two extreme mixtures is
presented. In the case of F4 mixture (the least fluid) slight aggregate segregation in the
middle could be observed. On the other hand, with F1 mixture (the most fluid) sepa-
ration of paste from aggregates took place at the edges. Also, the reduced workability
of F4 led to a reduction in compressive strength (see Fig. 3.b), probably due to
increased air content in the specimens.
4.2 Statistical Analysis of Data, Fitted Models and Optimisation

Table 7 presents the results of the fitted models for Dflow and t500 responses, including
the residual error term (e), along with the correlation coefficients. An analysis of
variance showed that these models are significant when describing the effect of Vg/Vg,
lim and Vs/Vm on the modelled responses. Since R2 and R2adj values are significantly
high, a large proportion of the variability of response variables is explained by the
obtained regression models. As could be expected, both design variables - Vg/Vg,lim
and Vs/Vm- influence significantly the analysed fresh state properties of SCC; Vs/Vm
being the most influencing parameter. An increase of Vs/Vm or Vg/Vg,lim decreases
the flowability and decreases the viscosity of the mixture (lower t500).
218 J. Blazy et al.
Fig. 3. Concrete test results: a) t500 .vs. Dflow; b) fc,7d .vs. Dflow
Fig. 4. Spread flow area of tested concrete mixtures: a) F4; b) F1
Again, the fitted numerical models and numerical optimization technique were used
to determine the range of design variables that would lead to an SCC belonging to
SF1 + VS2 consistency classes, while minimizing the volume of paste (or maximizing
the total aggregates content). The best solution found corresponds to: Vg/Vg,lim =
0.453 and Vs/Vm = 0.430. The estimated response values of this mixture are:
Dflow = 600 mm; t500 = 29.8 s; fc,7d 104 MPa (the average compressive strength
of tested mixtures). In the next stage of this study, the benefits of hybridization
(considering fibres with lengths of 13, 35 and 60 mm) on flexural strength will be
assessed, in order to achieve the highest performance while keeping a relatively low
fibre content.
Table 7. Fitted numerical models at concrete level (coded values of parameters)
Dflow (1/t 5000.5)

(mm) (s)
Independent 636.181 0.2315
Vg/Vg,lim -57.454 -0.0634
Vs/Vm -71.329 -0.0858
ε, std. dev.* 17.650 0.026
R2 / R2adj 0.951 / 0.941 0.927 / 0.911
(*) error term is a random and normally distributed variable with mean zero
5 Conclusions
Most relevant conclusions of the current study are the following:

• The paste and aggregate particle phases and fibres mixture of HPFRC can be
adjusted separately using DoE approach, allowing for mix-design optimisation.
• Fitted numerical models revealed that the workability and compressive strength of
HPFRC mortars are mainly determined by the following pairs of design variables:
(Vw/Vp, w/c) and (w/c, SF/C), respectively
• Total aggregates content was maximized to achieve an HPFRC exhibiting self-
compacting ability (SF1 + VS2) while incorporating 1% straight steel fibres (lf =
13 mm; df = 0.2 mm).
Acknowledgements. This work was financially supported by: UID/ECI/04708/2019- CON-

STRUCT - Instituto de I&D em Estruturas e Construções and the project PTDC/ECI-
EST/30511/2017 funded by national funds through the FCT/MCTES (PIDDAC). Collaboration
and materials supply by EUROMODAL, Secil, Omya Comital, Sika, Krampharex and Dramix
and is gratefully acknowledged.
References
1. Abbass, W., Khan, M.I., Mourad, S.: Evaluation of mechanical properties of steel fiber
reinforced concrete with different strengths of concrete. Constr. Build. Mater. 168, 556–569
(2018)
2. Okeh, C.A.O., Begg, D.W., Barnett, S.J., Nanos, N.: Behaviour of hybrid steel fibre
reinforced self compacting concrete using innovative hooked-end steel fibres under tensile
stress. Constr. Build. Mater. 202, 753–761 (2019)
3. Jang, S.J., Jeong, G.Y., Yun, H.D.: Use of steel fibers as transverse reinforcement in
diagonally reinforced coupling beams with normal- and high-strength concrete. Constr.
Build. Mater. 187, 1020–1030 (2018)
4. Ruano, G., Isla, F., Pedraza, R.I., Sfer, D., Luccioni, B.: Shear retrofitting of reinforced
concrete beams with steel fiber reinforced concrete. Constr. Build. Mater. 54, 646–658
(2014)
220 J. Blazy et al.
5. Mousavi, S.M., Ranjbar, M.M., Madandoust, R.: Combined effects of steel fibers and water
to cementitious materials ratio on the fracture behavior and brittleness of high strength
concrete. Eng. Fract. Mech. 216(June), 106517 (2019)
6. Gholampour, A., Ozbakkaloglu, T.: Fiber-reinforced concrete containing ultra high-strength
micro steel fibers under active confinement. Constr. Build. Mater. 187, 299–306 (2018)
7. Deeb, R., Karihaloo, B.L., Kulasegaram, S.: Reorientation of short steel fibres during the
flow of self-compacting concrete mix and determination of the fibre orientation factor. Cem.
Concr. Res. 56, 112–120 (2014)
8. Yoo, D.Y., Shin, H.O.: Bond performance of steel rebar embedded in 80–180 MPa ultra-
high-strength concrete. Cem. Concr. Compos. 93(February), 206–217 (2018)
9. Kazemi, M.T., Golsorkhtabar, H., Beygi, M.H.A., Gholamitabar, M.: Fracture properties of
steel fiber reinforced high strength concrete using work of fracture and size effect methods.
10. Skazlic, M., Serdar, M., Bjegovic, D.: Influence of test specimens geometry on compressive
performance concrete. In: Second International Symposium on Ultra High Performance
Concrete (2008)
11. Piérard, J., Dooms, B., Cauberg, N.: Durability evaluation of different types of UHPC. In:
RILEM-fib-AFGC Int. Symp. Ultra-High Perform. Fibre-Reinforced Concr. UHPFRC 2013
–, no. 1, pp. 275–284 (2013)
12. Li, P.P., Yu, Q.L., Brouwers, H.J.H.: Effect of coarse basalt aggregates on the properties of
Ultra-high Performance Concrete (UHPC). Constr. Build. Mater. 170, 649–659 (2018)
13. Nunes, S.: Performance-based design of self-compacting concrete (SCC): a contribution to
enchance SCC mixtures robustness, p. 357 (2008)
14. Okamura, H., Ouchi, M.: Self-compacting concrete. Adv. Concr. Technol. 1(1), 5–15 (2003)
Influence of the Steel Fibres on the Tension
and Shear Resistance of Anchoring
with Anchor Channels and Channel Bolts Cast
in Concrete
Mazen Ayoubi1(&), Christoph Mahrenholtz2, and Wilhelm Nell3

1
Frankfurt University of Applied Sciences (FRA-UAS), Frankfurt, Germany
mazen.ayoubi@fb1.fra-uas.de
2
Jordahl GmbH, Berlin, Germany
3
KrampeHarex GmbH & Co. KG, Hamm, Germany
Abstract. The current design method for anchor channels with channel bolts is
based on test results for fasteners installed in conventional concrete. The field of
application for steel fibre concrete have been growth over the last years and
recently steel fibre reinforced concrete became popular e.g. for the production of
prefabricated tunnel elements. The existing design rules for fasteners including
anchor channels with channel bolts do not cover steel fibre reinforced concrete.
To study the load-displacement behaviour in tension and shear, exploratory tests
have been carried out on anchor channel-channel bolt-systems cast in plain and
steel fibre reinforced concrete. The test results demonstrate a superior perfor-
mance of channel bolts installed in anchor channels which were cast-in steel
fibre reinforced concrete if compared with systems cast in plain reinforced
concrete. The results of the experimental investigations will be explained und
discussed in this article.
Keywords: Anchor channel Channel bolt Anchorage Steel fibre reinforced

concrete (SFRC) Concrete Capacity Ductility
1 Introduction
The addition of steel fibres to the concrete mix increases the tensile strain capacity and
ductility as well as improves the structural properties of concrete in its hardened state
such as the tensile strength, impact strength, toughness, fatigue strength and the ability
to resist cracking and spalling. Due to better mechanical and physical properties
became the usage of steel fibre reinforced concrete (SFRC) over the last decades a
better alternative to the conventional reinforcing concrete and being widely used as a
construction material. Due to addition of steel fibres, the overbridging of cracks (Fig. 1)
is ensured and so the cracking resistance of concrete increases [1]. This beneficial effect
leads to an increase in the concrete resistance under quasi-static, cyclic and impact
loading and furthermore, to the increase of the spalling resistance [1–3].
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_20
222 M. Ayoubi et al.
Fig. 1. Crack-bridging mechanism, a) in normal concrete and b) in SFRC from [2, 3].
Various factors and mechanisms influence the tensile behaviour of SFRC, including
the pullout behaviour of individual fibres, the random distribution of fibres, and the
effects of finite member dimensions [4]. Especially in members with end-hooked fibres,
the tensile behaviour of mechanical anchorage is essential in addition to the frictional
bond behaviour between fibres and concrete matrix [4]. The use of medium-high steel
fibres contents significantly improves the post-peak behaviour in tensile for flexure, by
extending the softening branch and reducing the negative slope. Steel fibres give to the
concrete a sizable post-peak residual strength. This lead to an improved fracture energy
and toughness of SFRC materials with high fibre content compared to ordinary
concrete [4].
Anchoring in Concrete Using Anchor Channels with Channel Bolts
Anchor channels combined with channel bolts allow the reliable connection of steel
components to the reinforced concrete structure. To this end, T-shaped channel bolts
are locked into C-shaped anchor channels (Fig. 2a) that have been cast into the rein-
forced concrete. Conventional anchor channels allow the transfer of tension loads
(N) and shear loads perpendicular to the channel ðV? Þ. Serrated anchor channels and
matching serrated channel bolts have recently been developed to also enable load
transfer in the direction of the channel ðVk Þ, thus making load transfer in all directions
possible (Fig. 2b). Simulated seismic load tests showed that the load bearing behaviour
of the serrated connection is very robust. This is because adjacent teeth are activated
when the teeth in the contact area between the head of the channel bolt and the lips of
the anchor channel start to fail. This allows the qualification of the serrated channels
according to the highest seismic requirements [5, 6].
Influence of the Steel Fibres on the Tension and Shear Resistance 223
Anchor
Anchor channel
V|
Channel
Serration
V||
Channel bolt N
a) b)
Fig. 2. a) Components of anchor channel and channel bolt; b) Serrated systems allow load
transfer in all directions.
Fig. 3. Installation sequence a) attaching of anchor channel to formwork, b) pouring of

concrete, c) removing of filler, d) twisting-in of channel bolt.
For installation, the anchor channel is hot glued or nailed to the formwork (Fig. 3a).
Anchor channels are generally furnished with filler material to prevent concrete slurry
leaking into the profile during concreting. After the concrete is set and the formwork is
stripped off, the filler is removed (Fig. 3b and c). Channel bolts are then inserted and
twisted in the slot of the anchor channel to allow fastening of components at any point
along its length (Fig. 3d).
Connections using anchor channels with channel bolts have several advantages,
making anchor channels with channel bolts suitable for the connection of any kind of
component in concrete, e.g. elevator guide rails, curtain wall brackets, and particularly
technical equipment in tunnels [7–9]:
– Quick and easy installation of the anchor channel during concreting
– Compensation of building tolerances by adjusting the position of channel bolt along
the length of the anchor channel
– Robust load transfer due to mechanical interlock (bolt-channel, anchor-concrete)
– Unlike for anchor plates no on-site welding is required, thus no weld quality issues,
or fire risks
– Unlike for post-installed fasteners no on-site drilling required thus no hassle with
positioning or cutting of reinforcing bars, and no exposition of the workforce to
harmful silica dust
– Later positional adjustment, or replacement of attached components is made easy at
any time
The design of anchor channels with channel bolts was just recently codified as the
European standard EN 1992–4 (2018) [10] which requires a qualification of the system
according to the European assessment guideline EAD 330008-02-0601 (2016) [11].
SFRC in Precast Segment for Tunnel Lining
Due to the advantages of SFRC in performance, durability and in terms of cost
reductions compare to traditionally reinforced concrete the application and the use of
SFRC in precast tunnel lining design is a growing trend [12, 13] (s. Figure 4). Anchor
channels with channel bolts (aka T-bolts) allow an easy and reliable connection of any
kind of component for such structures, e.g. trays, railings, lights, sprinklers and
equipment. While SFRC makes the post-installation of fasteners difficult, it is an ideal
substrate for anchor channels-channel bolt-systems.
Fig. 4. Construction of the tunnel lining segments, from [14].
Anchoring in SFRC
The proliferation of SFRC also means that anchoring of structural and non-structural
components has to be carried out in this substrate. Not many tests on fasteners used for
the anchoring in SFRC have been conducted yet.
Walter, E. and Ammann, W. [15] studied the behaviour of post-installed undercut
fasteners, adhesive fasteners, and wedge fasteners under tension and concluded that a
statistically significant increase of the capacities cannot be inferred. Also Klug, Y.,
Holschemacher, K. et al. [16] found it difficult to predict the increase in tension capacity
of post-installed fasteners and reasoned that this is due to the inhomogeneous distri-
bution and orientation of the fibres. Kurz, C., Thiele, C. et al. [17] reported that fasteners
post-installed in SFRC develop tension capacities which are at least equivalent to the
capacity if post-installed in regular concrete. To the knowledge of the authors, no shear

tests on post-installed fasteners have ever been carried out yet. All studies mentioned
that drilling in SFRC is difficult because the steel fibres may cause jamming and
increased wear of the drilling tools. This challenge is exacerbated because high per-
formance concrete i.e. high strength concrete is typically used for SFRC elements. In
these regards, cast-in fasteners are more suitable for the anchoring in SFRC, e.g. anchor
channels with channel bolts. Nilforoush, R., Nilsson, M. et al. [18] carried out tension
tests on cast-in headed fasteners in plain and steel fibre reinforced normal- and high-
strength concrete with compressive strengths up to 80 MPa. It was concluded that the
concrete capacity (design) method (Fuchs, W., Eligehausen, R. et al. [19]) considerably
underestimates the tensile breakout capacity of headed fasteners in fibre reinforced
concrete and that fibres facilitate a pronounced ductile deformation at ultimate load and
prevent a brittle post-peak behaviour potentially associated with high-strength concrete.
Some of the first published shear tests on cast-in fasteners have recently been conducted
by Lee, J.-H., Cho, B.-S. et al. [20]. The tested cast-in headed fasteners showed a
pronounced correlation of ultimate shear capacity and fibre content. However, neither
tension nor shear tests on channel bolts installed in anchor channels which were cast in
SFRC have been carried out to date. To investigate the performance of channel bolts-
anchor channels-systems in SFRC, an extensive test program was launched.
2 Experimental Investigations
2.1 Materials
For all tests, the same concrete mixture with a tested compressive strength of about
95 N/mm2 was used which is representative for applications where SFRC is used. No
reinforcing bars were installed since anchor channel-channel bolt-systems are generally
cast in unreinforced concrete members to study their load-displacement behaviour, e.g.
for product qualification. The steel fibres provided from the producer KrampeHarex were
made of circular, non–alloy steel wire with end hooks, diameter 0.75 mm, length 60 mm
and a nominal steel yield strength of 1900 N/mm2. The fibre mass content was 40 kg/m3,
equal to about 0.5% by weight. 320 mm long anchor channels JTA W 53/34 from the
producer Jordahl were cast into the concrete members. These anchor channels were made
of hot–rolled profiles with two anchors at a distance of s = 250 mm. For embedment
depths smaller than the standard depth of hef = 155 mm with round headed anchors
riveted to the channel, I–shaped anchors made of cut I–beams were welded to the channel.
Shutter and concrete works were carried out in a precast yard of the company Max Bögl.
The installation of the T–bolts JB M16 prior to testing completed the test specimens. The
grade 8.8 channel bolts had a nominal steel yield strength of 640 N/mm2.
2.2 Test Program

The 8 shear and 8 tension test series presented in this paper (Table 1) are part of a
larger test program. The number of test repeats within each series was typically 3 for
shear and 5 for tension tests in fibre reinforced concrete and 2 for shear and 3 for
tension reference tests in plain concrete. The edge distance c1 and the embedment depth
hef varied for the shear and tension test series, respectively. In addition to the 48 tests
on anchor channel-channel bolt-systems, tests to determine the performance class of the
fibre reinforced concrete were carried out that are not discussed in this paper.
2.3 Test Setup

The tests were carried out at the Jordahl Test Laboratory (three point bending tests were
carried out at the Frankfurt University of Applied Sciences and concrete tests were
carried out at the Max Bögl Laboratory). Two different unconfined test setups were
used where the support has sufficient distance to the anchoring in order to allow the
development of a full concrete breakout (Fig. 5 and 6). For shear loading, the test
specimens were placed on a strong floor and tied down to counteract the vertical uplift
forces deriving from eccentricity effects. A support accommodated the actuator for
shear loading and provided horizontal bearings at a distance of 5c1 + s. A PTFE sheet
was placed on top of the anchor channel and surrounding concrete before the loading
fixture was connected with a balance beam to ensure equal loading of the two channel
bolts installed above the anchors. For tension loading, a support with the mounted
actuator for tension loading formed with the test specimen a self–equilibrium system.
The contact area had a distance of at least 2hef from the centre of anchoring. The
loading fixture was connected to a channel bolt installed above an anchor.
The anchor channel-channel bolt system was monotonically loaded to failure at a
constant rate within 2 to 3 min. Load cells and displacement transducers recorded load
Fu and displacement d at a rate of 5 Hz.
Table 1. Test Program, [6]

Series* Repeats Load direction Concrete type Concrete strength Edge Embedment depth
fc,test distance hef
[MPa] c1° [mm]
[mm]
S-p-50-155 2 Shear Plain 96.4 50 155
S-f-50-155 3 Shear Fibre 98.6 50 155
S-p-100-155 2 Shear Plain 96.4 100 155
S-f-100-155 3 Shear Fibre 98.6 100 155
S-p-150-155 2 Shear Plain 96.4 150 155
S-f-150-155 3 Shear Fibre 98.6 150 155
S-p-200-155 2 Shear Plain 96.4 200 155
S-f-200-155 3 Shear Fibre 98.6 200 155
T-p-∞-69 2 Tension Plain 94.1 ∞ 69
T-f-∞-69 5 Tension Fibre 98.2 ∞ 69
T-p-∞-95 2 Tension Plain 94.1 ∞ 95
T-f-∞-95 5 Tension Fibre 98.2 ∞ 95
T-p-∞-120 1 Tension Plain 94.1 ∞ 120
T-f-∞-120 5 Tension Fibre 98.2 ∞ 120
T-p-∞-155 3 Tension Plain 92.5 ∞ 155
T-f-∞-155 5 Tension Fibre 92.8 ∞ 155
* Code: Shear or Tension-plain or fibre-c1-hef; ° ∞ equals to any distance larger than 2hef
Fig. 5. Test setup for shear loading (left) and tension loading (right), [6].
Fig. 6. Pictures of the test setup for shear loading (left) and tension loading (right).
2.4 Test Results

The coefficients of variation (cv) were reasonable despite the small number of test
repeats per series (Table 2): The cv of the ultimate load Fu was typically well below
15% which is the threshold commonly accepted for concrete related failure modes in
fastener qualification testing. The high cv of one series (S-f-100-155) can be attributed
to a test with a bias towards an outlier. The cv of the displacement at 50% of the
ultimate load d(0.5Fu,m) was always below 40% which is the maximum accepted in the
context of fastener qualification.
Table 2. Test results, [6].

Series Ultimate load Coeff. of var. Displacement Coeff. of var. Displacement Failure mode*
Fu,m cv(Fu) d(0.5Fu,m)m cv(s(0.5Fu,m)) d(Fu,m)m
[mm] [%] [mm] [%] [mm]
S-p-50-155 47.6 1.6 0.71 13.46 3.00 2C
S-f-50-155 81.0 0.8 1.23 27.8 4.66 3C
S-p-100- 71.8 11.2 1.42 23.3 2.67 3C
155
S-f-100-155 172.0 21.3 1.94 15.6 5.43 3C
S-p-150- 105.0 4.1 1.58 16.6 3.23 3C
155
S-f-150-155 212.6 3.6 1.77 10.7 5.21 3C
S-p-200- 144.4 2.8 1.83 17.9 2.84 3C
155
S-f-200-155 269.4 2.0 2.71 12.4 7.98 3 Sb
T-p-∞-69 89.6 8.0 0.66 18.4 5.61 2C
T-f-∞-69 110.3 2.9 1.31 17.8 9.11 1 C, 1 Sl, 3 Sb
T-p-∞-95 99.6 15.0 0.90 5.5 8.29 2C
T-f-∞-95 106.0 3.6 1.23 9.4 9.13 1 C, 2 Sa, 2 Sl
T-p-∞-120 105.2 – 1.06 – 11.21 Sl
T-f-∞-120 109.2 2.4 1.42 10.2 11.94 4 Sl, 1 Sb
T-p-∞-155 93.5 4.3 0.95 37.4 20.70 1 Sl, 2 Sb
T-f-∞-155 91.3 7.1 0.93 37.7 22.44 4 Sl, 1 Sb
* Sb: steel failure bolt; Sl: steel failure lip; Sa: steel failure anchor; C concrete cone or edge breakout
Overall, no clear trend of the coefficients of variation with regard to the concrete type
(plain or fibre) could be inferred (s. last paragraph). By trend, the recorded displacements
d(0.5Fu,m) and d(Fu,m) confirmed that fibres consistently support a more ductile beha-
viour also of anchor channels with channel bolts. More prominent, the fibres significantly
influenced the ultimate load Fu and the failure modes of the tested anchor channel-
channel bolt-systems: Subjected to shear load, only the systems cast in fibre reinforced
concrete with the largest tested edge distance c1 = 200 mm failed in steel due to shearing
off the bolt, otherwise concrete edge breakout was decisive. Under tension load, only the
systems cast in plain concrete with the embedment depth hef 95 mm failed consis-
tently by concrete cone breakout, otherwise steel failure of bolt, lip or anchor occurred
(rupture or bending). Clearly, if steel failure is the controlling failure mode, fibres have no
effect. If failure occurs due to concrete breakout, the fibres increased the capacity sig-
nificantly by the factor of about 1.8 for shear and 1.4 for tension. Moreover, the fibres
allow the shift of the transition from concrete breakout to steel failure (s. Fig. 7, 8 and 9).
The dashed lines (Fig. 9a and 9b) represent the mentioned transmission to steel failure
depending on the test results (s. Table 2, failure mode).
Fig. 7. Pictures of the failure under tension loading.
Fig. 8. Pictures of the failure under shear loading.
To compare the performance of anchor channel-channel bolt-systems cast in plain

and fibre reinforced concrete further, the influence of the edge distance c1 and
embedment depth hef is illustrated by means of typical curves recorded during the shear
and tension tests (Fig. 10): The fibres cause a substantial increase of the shear capacity
where the failure mode remains concrete edge breakout if tested with an edge distance
Fig. 9. Capacities of anchor channels-channel bolt-systems cast in plain and fibre reinforced
concrete tested a) in shear and b) in tension for different edge distances and embedment depths,
respectively, [6].
Fig. 10. Load-displacement curves of anchor channels-channel bolt-systems cast in plain and
fibre reinforced concrete tested in shear with a) small and b) large edge distance and tested in
tension with c) small and d) large embedment depth, [6].
of c1 = 50 mm (Fig. 10a) but changes to steel failure if tested with an edge distance of
c1 = 200 mm (Fig. 10b). In this case, the displacement at ultimate load is roughly
tripled. The fibres also cause a change from concrete cone breakout to steel failure,
accompanied by a distinct increase in tension capacity and displacement, if customized
systems with an embedment depth of hef = 69 mm are tested (Fig. 10c). In contrast, no
significant influence of the fibres can be determined if the standard system with an
embedment depth of hef = 155 mm is tested since for this configuration steel failure is
controlling already in case the anchor channel-channel bolt-system is cast into concrete
without fibres (Fig. 10d). The examples demonstrate that the fibres increase the duc-
tility of concrete breakouts and may change it to steel failure modes.
3 Summary and Conclusion
Steel Fibre reinforced concrete (SFRC) gains importance for the construction of
structural members, e.g. tunnel segments. The drilling in SFRC for the post-installation
of fasteners is challenging, not least because of the high concrete strengths prevalent for
SFRC. For this and other reasons, anchor channels with channel bolts are often a
favourable solution to connect any component to structural elements made of SFRC.
However, no study on the performance of anchor channel-channel bolt-systems has
been published to date.
To this end, a research program was launched to compare the performance of
channel bolts installed in anchor channels cast in plain and fibre reinforced concrete.
The results of 48 shear and tension tests presented in this paper demonstrate that anchor
channel-channel bolt-systems cast in fibre reinforced concrete sustain higher ultimate
loads and develop larger corresponding displacements if compared with identical
systems cast in plain concrete. The increase in capacity and ductility may lead to a
positive shift from rather brittle concrete breakout to more ductile steel failure modes.
The views expressed in this paper are the views of the authors only and do not
necessarily reflect the views of Jordahl, Max Bögl and KrampeHarex.
Acknowledgements. Anchor Channels, Concrete test members and steel fibres were kindly
provided by the companies Jordahl, Max Bögl and KrampeHarex, which are greatly appreciated.
References
1. Bokor, B., Tóth, M., Sharma, A.: Influence of steel fiber content on the load-bearing capacity
of anchorages in concrete (2017)
2. Mechtcherine, V.: Rissbeherrschung durch Faserbewehrung, Beherrschung von Rissen in
Beton. In: 7. Symposium Baustoffe und Bauwerkserhaltung Karlsruher Institut für
Technologie, S. 83–94. KIT Scientific Publishing (23. März 2010)
3. Døssland, Ä.L.: Fibre Reinforcement in load carrying concrete structures, Norwegian
University of Science and Technology, Thesis for the degree of philosophiae doctor,
February 2008
4. Lee, S.-C., Cho, J.-Y., Vecchio, F.J.: Diverse embedment model for steel fiber-reinforced
concrete in tension: model verification. ACI Mater. J. 107, 526–535 (2011)
5. Mahrenholtz, C., Lambton, J., Julier, F.: Suitability of anchor channels with channel bolts for
use in nuclear power plants. In: Proceedings of the 24th International Conference on
Structural Mechanics in Reactor Technology (SMiRT 24), Proceeding Div VI, Busan (2017)
6. Mahrenholtz, C., Ayoubi, M., Müller, S., Bachschmid, S.: Performance of anchor channels
with channel bolts cast in fibre reinforced concrete (FRC). IOP Conf. Ser. Mater. Sci. Eng.
615, 728–735 (2019)
7. Gage, C.: Are we paying enough attention to elevator shaft connections? Elevator World,
October Edition (2014)
8. Gage, C.: Looking behind the façade at curtain wall connections. Chinese Curtain Wall
Magazin (2014b)
9. Gottschalk, B., Mahrenholtz, C.: Befestigung von Ankerschienen mit Installationskonen
(Fastening anchor channels with installation cones). BFT International Betonwerk +
Fertigteil-Technik, 2017, Heft 6 (2017)
10. EN 1992–4, Eurocode 2: Design of concrete structures – Part 4: Design of fastenings for use
in concrete. European Committee for Standardization (CEN); EN 1992-4 (2018)
11. EAD 330008-02-0601, Anchor channels. European Assessment Document, OJEU 2016/C
248/06, European Organization for Technical Assessment (EOTA) (2016)
12. Axhimusa. R.: Investigation of steel fiber reinforced concrete (SFRC) elements with regard
to their economic viability and market growth prospects. Master Thesis at the Frankfurt
University of Applied Sciences, Supervisor: Prof. Dr.-Ing. M. Ayoubi, 2019
13. Rivaz, B.: Steel fiber reinforced concrete (SFRC): the use of SFRC in precast segment for
tunnel lining, p. 66 (2009)
14. Carmona, S., Molins, C., Aguado, A., Mora, F.: Distribution of fibers in SFRC segments for
tunnel linings. Tunn. Undergr. Space Technol. 51, 238–249 (2016)
15. Walter, E., Ammann, W.: Fastening technology in fibre reinforced concrete. In: Proceedings
of the 4th RILEM International Symposium on Fibre Reinforced Concrete, Sheffield (1992)
16. Klug, Y., Holschemacher, K., Wittmann, F.: Tragverhalten von Befestigungselementen in
Stahlfaserbeton (Load carrying behaviour of fastening elements in steel fibre reinforced
concrete). Innovationen im Bauwesen, Beiträge aus Praxis und Wissenschaft (2002)
17. Kurz, C., Thiele, C., Schnell, J., Reuter, M., Vitt, G.: Tragverhalten von Dübeln in
Stahlfaserbeton (Load carrying behaviour of fasteners in fibre reinforced concrete).
Bautechnik 89. Heft 8, 545–552 (2012)
18. Nilforoush, R., Nilsson, M., Elfgren, L.: Experimental evaluation of tensile behaviour of
single cast-in-place anchor bolts in plain and steel fibre-reinforced normal- and high-strength
concrete. Eng. Struct. 147, 195–206 (2017)
19. Fuchs, W., Eligehausen, R., Breen, J.: Concrete Capacity Design (CCD) Approach for
fastening to concrete. ACI Struct. J. 92(6), 73–94 (1995)
20. Lee, J.-H., Cho, B.-S., Kim, J.-B., Lee, K.-J., Jung, C.-Y.: Shear capacity of cast-in headed
anchors in steel fiber-reinforced concrete. Eng. Struct. 171, 421–432 (2018)
Fiber Reinforced Concrete After Elevated
Temperatures: Techniques of Characterization
Ronney Rodrigues Agra1,2(&), Ramoel Serafini1,2,

and Antonio Domingues de Figueiredo1
1
Department of Civil Construction Engineering, Polytechnic School of
University of São Paulo, São Paulo, Brazil
ronney.agra@usp.br
2
Institute of Technological Research, São Paulo, Brazil
Abstract. The mechanical properties of fiber reinforced concrete (FRC) are

negatively affected when subjected to elevated temperatures. The main concern
is regarding its post-crack tensile strength, which can be severely impaired at
temperatures above 300 °C. In this composite, the mechanical characterization is
constantly performed by means of bending tests of prismatic specimens, as
recommended by EN 14651. However, due to limiting aspects, alternative
methodologies have been used for the characterization of FRC, among which
are the DEWS (Double Edge Wedge Splitting) and the Double Punch tests. In
this context, the present study compares the methodologies for evaluating the
mechanical behavior of FRC after elevated temperatures, discussing and
emphasizing its advantages and limitations. The Double Punch test does not
show satisfactory response as a consequence of the degradation suffered by the
sample and the puncture interaction induced by the test. On the other hand, the
indirect tensile DEWS test shows that it is capable of characterizing the FRC
even after exposure to elevated temperatures. Although the post-crack response
of the composite varies according to the method adopted, the post-crack tensile
strength in the service limit state (SLS) and ultimate limit state (ULS) are
considerably reduced when compared with the ambient temperature.
Keywords: Fiber reinforced concrete Test method Post-crack tensile

strength Elevated temperatures
1 Introduction
The crescent use of fiber reinforced concrete (FRC) in the construction industry can be
attributed to economical and technical aspects that benefits the production of precast
elements, slabs, and tunnel linings. A significant contribution has been made by the
publication of the fib Model Code 2010 [1], which presented parameters for the use of
FRC for structural purposes. This led to the further increase in the use of this tech-
nology for the construction of buildings and the production of segments for tunnels
built with Tunnel Boring Machine (TBM) technology [2].
The mechanical properties of FRC are negatively affected when subjected to ele-
vated temperatures [3] and the post-crack behavior varies according the type of fiber,
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_21
234 R. R. Agra et al.
amount of fiber used, the duration of temperature exposure, and maximum temperature
achieved. In this sense, the main concern is regarding the post-crack tensile strength,
which may be severely reduced for temperatures above 300 °C [4].
The research carried out by Dehn and Hermann [5] reveals the absent literature and
normative bases that allow the FRC subject to high temperatures to be adequately
treated for design checks, which can put structural safety at risk. This indicates the need
for further studies with regard to requirements, test methods and procedures for the
design of FRC structures when subjected to fire, especially with regard to the char-
acterization of the composite’s behavior and obtaining constitutive equations to be
applied in behavior prediction models.
Although the current studies on the post-crack behavior of the FRC exposed to fire
concentrate mainly on the assessment of the mechanical properties for a specific target
temperature [6–8], the research carried out by Serafini et al. [9] is highlighted for using
the unifacial heating fire methodology. The fire simulator operated by the burning of
methane to achieve a similar behavior to the hydrocarbon fire curve. In this study, the
authors evaluated the effect of fire on the mechanical properties of the macro-synthetic
fiber reinforced concrete (MSFRC) for tunneling applications. The results can be used
as a basis for the development of design-oriented models that allow to evaluate the
structural condition of the FRC after the fire exposure.
The high strength concrete (HSC) is more sensitive to the occurrence of spalling
due to the low permeability and porosity of the cementitious matrix [4]. It is widely
accepted that the use of polymeric micro-synthetic fibers promotes a sensitive reduction
in the process of explosive spalling [10] of concrete, and that the use of steel fibers is
beneficial to the fire resistance of the composite. In the absence of spalling, the veri-
fication of the residual mechanical properties of the structure and a detailed inspection
of the elements affected by fire is mandatory to evaluate the degree of safety of the
structure and to aid the decision-making process. In most of the fire events, the
structures do not collapse and the post-fire capacity must be evaluated, since rebuilding
a structure or reinforcing damaged elements has a considerable monetary impact.
Therefore, the knowledge regarding the residual properties of the cementitious matrix
and the reinforcing mechanism is key to decide which procedures must be employed
[11]. This analysis can be conducted by means of destructive and non-destructive
techniques.
In order to determine the FRC mechanical properties from a structural perspective,
the post-crack tensile strength of the composite must be determined [1]. The charac-
terization of the post-crack tensile strength of FRC is usually conducted based on
standard bending tests in prismatic specimens. This test has already been used to
evaluate macro-synthetic fiber reinforced concretes exposed to elevated temperatures
[9]. However, the experimental challenges related to bending tests are considerable due
to the size of the specimens and the difficulties with handling the specimens after
exposure to high temperatures. More than that, the flexural tensile strength values tend
to be overestimated in bending tests and questions can arise from the suitability of
using converting factors to reach a more reasonable direct tensile strength value,
equivalent to a mode I fracture type. Therefore, emerging methodologies have been
used in substitution to the traditional bending tests to evaluate FRCs, among them the
Double Punch test [12, 13] and the Double Edge Wedge Splitting (DEWS) test [14].
Fiber Reinforced Concrete After Elevated Temperatures 235
In this sense, the current study aims to present the techniques employed to deter-
mine the post-crack tensile properties of FRC after exposure to elevated temperatures,
discussing and highlighting the advantages and limitations of each method. This study
contributes to the limited literature regarding the comparison between tests capable of
determining the post-crack behavior of FRC.
2 The Effect of Temperature on the Mechanical Behavior

of FRC
A temperature gradient is induced in the FRC when it is subjected to elevated tem-

peratures, which results in different layers of dehydration of the cement paste and
micro-cracks. This culminates that the mechanical properties of the composite are
affected as a function of the distance to the heated surface. In this sense, any evaluation
of the global mechanical properties of FRC elements exposed to fire would results in
the average mechanical response of the affected layers. However, there are a very
limited number of studies that study the temperature gradient in FRC and its impact on
the mechanical properties of the composite [9].
The immediate consequence of exposing concrete to heat is a change in the
characteristics of the composite, such as the reduction of its strength and stiffness, as
well as the generation of additional deformations linked to the stress level during the
first stages of heating. In addition, as the temperature distribution in the section is not
uniform, these thermal deformations are also not uniform in the section, which gen-
erates effects related to spalling, thermal expansion and lateral deflections [11].
The parameters of compressive and tensile strength, stiffness and energy absorption
capacity of the FRC are largely affected by the increase of temperature. Yermak et al.
[4] in their studies noted that the mechanical strength of concrete with fibers
(0.75 kg/m3 of PP microfibers and 60 kg/m3 of steel fibers) is greater than that of
concrete without fibers at room temperature. There was an approximate difference of
15 MPa in the compressive strength and 3,5 MPa in the tensile strength, which
demonstrates that the steel fibers could act in this gain.
Some studies [4, 7, 8] show that in concretes with the addition of up to 70 kg/m3 of
steel fibers there are reductions in compressive strength that reach 40% after heating to
400 °C and 70% after heating to 600 °C, when compared to room temperature. After
exposure to high temperatures (above 300 °C) a high decrease of the compressive
strength was noted because of the thermal strain mismatch between aggregate (which
expand) and paste (which shrink) inducing tangential and radial cracks at the paste-
aggregate interface [4, 15]. However, the use of steel fibers reduced the rate of
degradation of compressive strength, although at 900 °C the presence of steel fibers
was negligible for the residual strength of concrete.
In addition, the decrease in the values of modulus of elasticity is more pronounced
when compared to that of compressive strength, especially between the temperatures of
150 and 450 °C [4, 6, 9, 16]. Reductions are directly associated to the mismatch
between aggregates and cement paste in terms of thermal expansion, which results in
the formation of extensive cracks in the interfacial transition zone.
The elastic properties of the composite are also negatively affected by the degra-
dation of synthetic fibers, the dehydration of hydrated products present in the concrete
matrix (especially portlandite and C-S-H), the breakage of bonds in the microstructure,
the increase in aggregate porosity, the reduction in the specific surface area of the
hydrates, and the coarsening of the cement paste pore structure. These processes result
in the increase in the porosity of the cement paste and the increase in capillarity pore
size [16]. Additionally, Yermak et al. [4] in their studies noted that the addition of steel
fibers did not bring a meaningful contribution to the reduction in the elastic modulus
with the temperature increase.
The post-cracking behavior of FRC at high temperatures changes according to the
type and fiber content, the exposure temperature and the duration of the exposure.
The SFRC shows lower losses in the post-peak response for temperatures above 400 °
C, while the MSFRC shows greater reductions in post-crack tensile strength and
toughness due to fiber degradation [17].
Yermak et al. [4] found that the post-crack tensile strength of the FRC shows a
minor decrease tendency for temperatures up to 300 °C. After that temperature mark,
severe damage is caused in the composite. When compared to plain concrete speci-
mens, the use of steel fibers has shown an increase of *10% in the post-crack tensile
strength for target temperatures of 300 and 600 °C. The steel fibers contributed to
improve the post-peak behavior of concretes for temperatures up to 750 °C, while at
around 900 °C the steel fibers did not show any capability to enhance the post-crack
tensile strength of the composite. The reduction at 900 °C occurred due to the oxidation
and corrosion of the steel fibers, which start to happen at 500 and 700 °C, respectively.
They also noted that the shape of the load-displacement curve changes with temper-
ature, since the peak load increases and the post-crack response is reduced at 600 °C.
The mechanical properties of steel fibers and the fiber-matrix interaction are factors
that considerably influence the post-cracking behavior of FRC exposed to high tem-
peratures. The fiber-matrix interaction begins to be affected for temperature of 500 °C
and above in terms of pullout load capacity [18].
3 Post-heat Evaluation of FRC by Means of Non-destructive

Tests
The evaluation of the capacity of concrete structures exposed to fire is a very difficult
task, since traditional destructive or non-destructive techniques are not suitable for the
inspection of a heterogeneous material. Possible approaches to this problem usually
involve inspection of the average response of the concrete element or a point-to-point
analysis of small samples extracted from different depths. The combination of several
techniques is constantly applied in this analysis, which allows investigating and
measuring physical, chemical and mineralogical changes, as well as the higher tem-
perature reached in the structural element [19].
In addition, these techniques must be applied together with mechanical tests, in
order to obtain a complete evaluation of the concrete after exposure to temperature.
A summary of these non-destructive techniques usually employed in the FRC is shown
in Table 1. The purpose of these tests is to determine the quality and integrity of the
materials, without compromising the load capacity of these elements. Therefore, the
methodology adopted must not harm the future use of the structural element, even if
commit actions considered invasive [20].
Table 1. Main non-destructive tests applied to concrete after exposure to elevated temperatures.
Analysis of concrete cover Point-to-point analysis of Special techniques
small specimens
Schmidt Hammer Differential Thermal Ultrasonic Pulse Velocity
(Sclerometer) Analysis (DTA) (UPV)
Internal fracture test Thermogravimetric Analysis Tomography and Resonance
(TGA)
Penetration resistance Scanning Electron Modal Analysis of Surface
(Windsor Probe) Microscopy (SEM) Waves (MASW)
CAPO test X-Ray Diffraction Digital Image Correlation
(DIC)
4 Post-crack Tensile Strength of FRC Exposed to Elevated

Temperatures
The existence of adequate methods to assess the mechanical behavior of the FRC
increases the reliability of the application of this material. The adopted method must be
compatible with existing design models, which is very important for the successful
implementation of the composite [21]. In order to characterize the mechanical per-
formance of the FRC, it is necessary to consider the post-peak behavior of the load-
deformation curve, with the determination of the post-crack tensile strength of the
composite and the increase in toughness due to the addition of fibers [1].
Experimental verifications from the structural point of view of the FRC are sug-
gested with the purpose of knowing its behavior under high temperature conditions,
since there is still no adequate approach to fire design for the composite. In addition,
technical parameters and numerical models to assess the behavior of the FRC after fire
events are very limited in the current literature [5]. For this reason, it has been usual to
carry out real scale (or at least representative scale) tests to verify the safety of structure
made with FRC under fire loading. Thus, mesoscale tests are required to obtain the
post-heat mechanical response of the composite.
The mechanical characterization of the FRC is traditionally carried out by means of
bending tests. However, after exposure to high temperatures, the thermal gradients
induced also vary significantly depending on the dimensions of the specimen and the
structural element [19]. The test recommended by EN 14651 [22] has already been
used (Fig. 1) to assess the post-crack flexural tensile strength of the macro-synthetic
fiber reinforced concrete (MSFRC) subjected to a real fire simulation [9, 23].
These papers contribute to the limited amount of literature data regarding the effect
of real fire scenarios on the properties of MSFRC. In the study of Serafini et al. [9]
prismatic specimens were subjected to single face heating under hydrocarbon fire curve
Fig. 1. Fire exposed specimen after flexural tensile strength test [23 adapted]
and flexural results demonstrated that the composite has no significant flexural resis-
tance after fire exposure (Fig. 2) due to reinforcement melting (170 °C) and ignition
(400–500 °C) in the flexural-tensile region of the specimen.
Fig. 2. Load-CMOD curves for MSFRC before and after fire exposure [9 adapted]
Reductions in matrix tensile strength values were associated with dehydration of

hydrates in cement paste, specific surface area of hydrates reduction, increase in the
total pore volume, and changes in cement paste pore distribution [24]. Post-crack
flexural tensile strength is greatly affected by fire in both service and ultimate limit
states. These changes are related to microstructural changes in the cement paste,
physical changes of the composite due to fire exposure, and the deterioration of fiber
reinforcement which results in a specimen with little to no post-crack flexural tensile
strength. Dynamic elastic and rigidity modulus values decrease by over 86% after fire
exposure due to the influence of cracks in the sample, physicochemical changes in the
matrix, and reinforcement fiber degradation.
However, experimental difficulties to perform the bending test are considerable due
to the relatively large dimensions of the prismatic specimens and the difficulties in the
process of handling the prisms after being subjected to high temperatures. In addition,
there is an overestimation of the flexural tensile strength values when compared to the
results obtained by direct tensile strength tests. Moreover, doubts arise in the scientific
community about the relevance of using conversion factors for flexural strength values
to obtain the equivalent direct tensile strength in room temperature. Therefore, new
methodologies have emerged with the proposition to replace the traditional bending
tests in the task of evaluating FRC. A few emerging tests can be cited, such as the
Double Punch test (DPT) and Double Edge Wedge Splitting (DEWS) test [12–14].
The Double Punch Test (DPT) has a number of advantages, such as its ease of
procedure, simple specimen preparation and lower variability in the measured results
than those resulting from bending tests using beam specimens [25]. Particularly, the
properties of FRC can be determined by using relatively small cylindrical specimens
which can be molded, cut from standardized cylinders, or cores drilled from an existing
structure. For application of the simplified methodology [26], accessible in most quality
control laboratories, a conventional compression test frame equipped with an axial
displacement control (not necessarily with a closed-loop control) is required. The DPT
also has correlations with the EN 14651 [22] test, and allows the detection of changes
in the characteristics of the FRC in face of the influences of parameters that determine
the mechanical performance of this composite. This make the test suitable for the
quality control of FRC in construction applications [27].
One of the obstacles to its use as a definitive technique for describing the FRC post-
cracking response is the difficulty in determining the constitutive stress-strain equa-
tions, mainly due to the unpredictability of the number of fracture planes [14]. How-
ever, the Double Punch test is one of the few tests presented in the literature that can be
performed on FRC specimens drilled from real structures that have been exposed to a
fire [6]. The DPT has also been applied in order to investigate the mechanical prop-
erties of FRC exposed to high temperatures, showing that the detrimental factors for the
tensile behavior of FRC are more the volume fraction and the aspect ratio of the fibers,
than their type [28]. However, one of the limitations of using the DPT for FRC after
exposure to high temperatures is that the deterioration of the concrete matrix results in
an increased frictional interaction in the region where the loads are applied, which may
influence the post-crack results [6].
Rambo et al. [6] investigated the mechanical properties of concrete samples rein-
forced with polymeric macrofibers exposed to temperatures up to 600 °C and the
applicability of the DPT to assess the post-crack tensile strength. The effect of lower
temperatures was not significant, however, as a result of the degradation suffered by the
sample surface, added to the puncture that results in the crushing of the deteriorated
matrix, the Double Punch test did not show a conclusive response after exposure to
high temperatures. This was mainly because the additional friction effect generated by
the punching interaction between the concrete matrix and the piston can be confused
with the bridge effect provided by the fibers. Nevertheless, the gradient of temperature
stablished within the specimens may preserve part of the material (i.e.: matrix and
fibers), and consequently, the composite post cracking performance.
The stress-strain curves obtained from the Double Punch tests demonstrated that the
MSFRC gradually loses tensile strength an energy consumption density with increasing
temperature and the post-cracking response varies significantly depending on the
temperature (Fig. 3). However, the effect of temperature on the mechanical behavior of
the MSFRC showed to be very similar to that known for conventional concrete with
relation to the loss of mechanical strength and elastic modulus.
Fig. 3. Stress-strain curves obtained from the results of the Double Punch test for the MSFRC
specimens submitted to different temperatures [6 adapted].
The MSFRC exposed up to 200 °C maintain similar values of the post-crack tensile
strength and overall ductility when compared to the MSFRC at room temperature.
However, from 400 °C upwards, the bearing capacity of the material is significantly
reduced and shown to be critical to the MSFRC mechanical performance.The DEWS
test constitutes an indirect tensile test by applying a compression load to a specimen
with wedges and notches on two opposite faces. This is a test that can be carried out on
small cubic samples, extracted from structures affected by fire or from larger speci-
mens. In addition, it can be executed in an open-loop control system, which makes its
application feasible in most laboratories. A type I fracture is obtained by means of the
DEWS test and, thus, the stress-strain relationship is obtained directly, which is fun-
damentally important for structural safety assessment based on the investigation of its
resistant capacity, in accordance with the fib Model Code [1].
Agra et al. [29] showed that the DEWS test is able to characterize the SFRC in
terms of post-crack tensile strength even in severe conditions, as in the case of samples
subjected to the action of fire. This was possible because the evaluation occurred
without prejudice to the values obtained as a response to the material, since no concrete
damage was found under the contact region. The authors cite that the tensile strength
values of the composite after exposure to fire were 71.1% lower than the values
obtained at room temperature. This is associated with dehydration of cement paste
products, loss of fiber reinforcing capacity and changes in pore distribution [24].
The values of post-crack tensile strength in SLS (COD = 0.5 mm) and ULS
(COD = 2.5 mm) were 64.1% and 59.8% lower than the values obtained at room
temperature (Fig. 4). These effects are related to physical-chemical changes in the
matrix, microstructural changes in the composite and degradation of the fibers used as
reinforcement.
Fig. 4. Average stress-COD curves for DEWS test before and after hydrocarbon fire exposure
[29 adapted].
Serafini et al. [30] also applied the DEWS test at SFRC exposed to 600 °C. The results
obtained also show that the tensile strength of the cementitious matrix and the post-crack
tensile strength in SLS and ULS are significantly affected. The composite tensile strength
values after temperature exposure of 600 °C were 82.5% lower than the value reached at
room temperature and the post-crack tensile strength values at SLS (COD = 0.5 mm)
reduced by 74.3% and in ULS (COD = 2.5 mm) by 72.2% when compared to room
temperature results. The authors also claim that no visible damage was caused by the
interaction between the roller and the SFRC for temperatures up to 600 °C.
In studies related to the DEWS and Double Punch tests the authors note that some
limitations – such as the presence of an unstable post-peak region – should be
observed, especially when the gap between the tensile strength of the matrix and the
post-crack tensile strength corresponding to the SLS is high. It can be observed when a
low fiber content is employed. The region of instability can also be noticed when the
open-loop machine is used at room temperature. Residual loads are underestimated
when open loop system is employed. Besides, the lack of rigidity of the equipment may
also cause instability.
It is possible to verify the region of instability more clearly in the results obtained
by means of the DEWS test, which reached values greater than 0.5 mm. However, the
effect of instability is mitigated after the exposure of the specimens to elevated
temperatures, since the gap between the tensile strength of the matrix and the tensile
strength in the SLS becomes smaller.
It is important to state that the service state conditions are certainly compromised
after a fire event and the main concern is regarding the ultimate state condition values,
according to the fib Model Code 2010 [1]. Furthermore, there is no direct influence of
the region of instability in determining the residual strength related to the ULS. Thus,
the evaluation occurs without damage to the values obtained as a response of the
material, which makes the DEWS test a viable methodology under these conditions.
5 Conclusions
This paper compares the methodologies for evaluating the mechanical behavior of FRC
after elevated temperatures, discussing and emphasizing its advantages and limitations.
From the present study, the following conclusions can be drawn:
• The parameters of compressive and tensile strength of the cement matrix, post-crack
tensile strength, stiffness and energy absorption capacity of the FRC are largely
affected by the increase of temperature.
• There are obstacles related to the use of flexural test setups for the determination of
the mechanical properties of FRC after elevated temperatures: the considerable
specimen volume, the difficulties in the process of handling the prisms, the over-
estimation of the flexural tensile strength values, and the relevance of using con-
version factors to estimate the tensile strength.
• The Double Punch test may not be adequate to assess the properties of SFRC after
temperature exposure. Results presented in the literature indicate that the puncture
caused by the test may result in crushing of the porous matrix and induced frictional
interaction.
• The DEWS test is a viable technique for the assessment of the post-crack tensile
properties of SFRC after temperature exposure. One of the main advantages of
DEWS test is that it generates a mode I fracture that simplify the determination of
constitutive equations from the results.
• The effect of post-cracking instability is mitigated after the exposure of the speci-
mens to elevated temperatures. Furthermore, there is no direct influence of the
region of instability in determining the residual strength related to the ULS, even
with the SLS compromised after a fire event.
Acknowledgements. The authors would like to thank the Institute for Technological Research
(IPT) and its foundation (FIPT) for their financial and institutional support though the New
Talents Program N.01/2017 and N.01/2018. The authors would also like to thank CAPES
(Coordenação de Aperfeiçoamento de Pessoal de Nível Superior). Antonio D. de Figueiredo
would like to acknowledge the financial support of the National Council for Scientific and
Technological Development - CNPq (Proc. Nº: 305055 / 2019-4).
References
1. Fédération Internationale du Béton – FIB. ‘Fib Model Code for Concrete Structures 2010’.
Switzerland (2013)
2. Liao, L., et al.: Design of FRC tunnel segments considering the ductility requirements of the
model code 2010. Tunn Undergr Sp Tech 47, 200–210 (2015)
3. Silva, C.L.: Projeto de estruturas de concreto em situação de incêndio conforme ABNT NBR
15200:2012 (Blucher 2012)
4. Yemark, N., et al.: Influence of steel and/or polypropylene fibers on the behavior of concrete
at high temperature: spalling, transfer and mechanical properties. Constr. Build. Mater. 132,
240–250 (2017)
5. Dehn, F., Herrmann, A.: ‘Concreto reforçado com fibras de aço em situação de incêndio –
requisitos normativos, pré-normativos e códigos-modelo’. Concreto & Construções, 87
(2017)
6. Rambo, D.A.S., et al.: Study of temperature effect on macro-synthetic fiber reinforced
concretes by means of Barcelona tests: an approach focused on tunnels assessment. Constr.
Build. Mater. 158, 443–453 (2018)
7. Tai, Y.S., et al.: Mechanical properties of steel fiber reinforced reactive powder concrete
following exposure to high temperature reaching 800 & #xB0;C. Nuclear Eng. Design 241,
2416–2424 (2011)
8. Poon, C.S., et al.: ‘Compressive behavior of fiber reinforced high-performance concrete
subjected to elevated temperatures’. Cement and Concrete Research, 34, 2215–2222 (2004)
9. Serafini, et al.: Influence of fire on temperature gradient and physical-mechanical properties
of macro-synthetic fiber reinforced concrete for tunnel linings. Constr. Build. Mater. 214,
254–268 (2019)
10. Kalifa, P.: High temperature behaviour of HPC with polypropylene fibres: from spalling to
microstructure. Cement Concrete Res. 31, 1487–1499 (2001)
11. Fédération Internationale Du Béton - fib Bulletin 46 ‘Fire design of concrete structures -
structural behaviour and assessment’ State-of-art report, Lausanne, Switzerland 2008
12. Choumanidis, D., et al.: Barcelona test for the evaluation of the mechanical properties of
single and hybrid FRC, exposed to elevated temperature. Constr. Build. Mater. 138, 296–
305 (2017)
13. UNE 83515: 2010 ‘Hormigones con fibras. Determinación de la resistencia a fisuración,
tenacidad y resistencia residual a tracción. Método Barcelona. The Spanish Association for
Standardisation’, Madrid 2010
14. di Prisco, M., Ferrara, L., Lamperti, M.G.L.: Double edge wedge splitting (DEWS): an
indirect tension test to identify post-cracking behaviour of fibre reinforced cementitious
composites. Mater. Struct. 46(11), 1893–1918 (2013). https://doi.org/10.1617/s11527-013-
0028-2
15. Xing, Z., et al.: Aggregate’s influence on thermophysical concrete properties at elevated
temperature. Constr. Build. Mater. 95, 18–28 (2015)
16. Zheng, W., Li, H., Wang, Y.: Compressive stress-strain relationship of steel fiber-reinforced
reactive powder concrete after exposure to elevated temperatures. Constr. Build. Mater. 35,
931–940 (2012)
17. Sukontasukkul, P., et al.: Post-crack (or post-peak) flexural response and toughness of fiber
reinforced concrete after exposure to high temperature. Constr. Build. Mater. 24, 1967–1974
(2010)
18. Abdallah, S., Fan, M., Cashell, K.A.: Pull-out behaviour of straight and hooked-end steel
fibres under elevated temperatures. Cem. Concr. Res. 95, 132–140 (2017)
19. Fernandes, B., Gil, A.M., Bolina, F.L., Tutikian, B.F.: Microstructure of concrete subjected
to elevated temperatures: physico-chemical changes and analysis techniques. Ibracon Struct.
Mater. J. 10, 838–863 (2017)
20. Helal, J., Sofi, M., Mendis, P.: Non-destructive testing of concrete: a review of methods.
Electr. J. Struct. Eng. 14, 97–105 (2015)
21. Monte, R., Toaldo, G.S., Figueiredo, A.D.: Avaliação da tenacidade de concretos reforçados
com fibras através de ensaios com sistema aberto. Revista Matéria 19, 132–149 (2014)
22. Standardization, E.C.F.: EN 14651: Test method for metallic fiber-reinforced concrete –
Measuring the flexural tensile strength (limit of proportionality (LOP), residual,
p. 15p. CEN, London (2007)
23. Serafini, R., et al.: Influence of fire exposure on the flexural behavior of macro-synthetic fiber
reinforced concrete. In: 9th international conference on concrete under severe conditions-
environment & loading, Porto Alegre 2019
24. Ma, Q., et al.: Mechanical properties of concrete at high temperature-a review. Constr. Build.
Mater. 93, 371–383 (2015)
25. Molins, C., Aguado, A., Saludes, S.: Double Punch Test to control the energy dissipation in
tension of FRC (Barcelona test). Mater. Struct. 42, 415–425 (2008)
26. Pujadas, P.: Caracterización y diseño del hormigón reforzado con fibras plásticas. Tesis
Doctoral - Universitat Politècnica de Catalunya, Barcelona (2013)
27. Silva, C.L.: ‘Proposta de metodologia alternativa para controle de qualidade da aplicação
estrutural do concreto projetado reforçado com fibras de aço’. Dissertação de Mestrado, São
Paulo 2017
28. Kim, J., Lee, G.P., Moon, D.Y.: Evaluation of mechanical properties of steel-fibre reinforced
concrete exposed to high temperatures by double-punch test. Constr. Build. Mater. 79, 182–
191 (2015)
29. Agra, R.R., et al.: ‘Avaliação dos efeitos do fogo na resistencia à tração residual do concreto
reforçado com fibras de aço por meio do ensaio DEWS (Double Edge Wedge Splitting). In:
5th Iberian-Latin-American Congress On Fire Safety, Porto 2019
30. Serafini, R., et al.: Double edge wedge splitting test to characterize the post-cracking design
parameters of fiber reinforced concrete subjected to high temperatures. J. Mater. Civ. Eng.
(forthcoming). https://doi.org/10.1061/(ASCE)MT.1943-5533.0003701
Influence of the Curing Temperatures
on the Mechanical Properties of Hemp Fibre-
Reinforced Alkali-Activated Mortars
Bojan Poletanovic(&), Gergely Nemeth, and Ildiko Merta
Institute of Material Technology, Building Physics, and Building Ecology,

Faculty of Civil Engineering, TU Wien, Vienna, Austria
bojan.poletanovic@tuwien.ac.at
Abstract. The aim of this research was to investigate how the curing tem-
perature influences the mechanical properties of fibre reinforced alkali activated
matrices. The mortar matrix contained fly ash as a binder, sodium water glass as
an activator and sand with a particle size between 0.4 and 0.8 mm. As rein-
forcement hemp fibres of 10 mm in length and of volume ratio 1% of the
composite were used. The prism mortars had dimension of 40 40 160 mm3
and were cured under three different temperatures: 20 °C, 60 °C and 80 °C.
The results showed that after adding the fibres specimens cured at room
temperature (20 °C) increased their compressive strength, flexural strength and
energy absorption for 4%, 57% and 711% respectively. At the 60 °C and 80 °C
curing temperatures the compressive strength of the specimens decreased for 8%
and 27% respectively, whereas the flexural strength decreased for 14% in case of
60 °C curing temperature but remained the same after the curing temperature of
80 °C. The energy consumption of the specimens increased for 331% and 153%
when specimens were cured at the 60 °C and 80 °C respectively.
Keywords: Hemp fibres Fly ash Activated materials Mortar Mechanical

properties
1 Introduction
The demand for cement is increasing constantly in the last decades and will expo-
nentially increase in the near future [1, 2]. The production of Portland Cement is one of
the main causes for concrete’s vast negative footprint on the environment. Approxi-
mately one ton of Portland cement production may release up to one ton of carbon
dioxide (CO2) in the atmosphere [3]. A possible way to reduce concrete’s footprint is to
reduce or completely replace the cement within it. A very effective way with main-
taining its physical and mechanical properties is to use pozzolanic materials, such as fly
ash, ground granulated blast furnace slag, metakaolin, etc. [4–7]. When pozzolanic
materials react with alkaline solutions they form a so-called alkali-activated material
(AAM) [4]. These materials not only reduce the CO2 release in comparison to the
traditional cement-based materials, but effectively re-use industrial by-products (such
as fly ash coming from thermal power plant coal combustion) that are so far mainly
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_22
246 B. Poletanovic et al.
landfilled. So far, AAMs are the most commonly researched alternative to cement-
based materials.
Similar to cement-based materials, AAMs have high compressive strengths, but
very low resistance to crack propagation under flexural stresses and low energy
absorption capacity. To improve these deficiencies the material should be reinforced
with short, randomly dispersed fibres. The traditional fibres used for this purpose are
steel, glass, synthetics, etc. [8–10]. Due to the growing environmental awareness
natural fibres such as hemp, flax, coir, sisal, cotton, etc. have been considered as an
environmentally friendly alternative to traditional fibres, since their mechanical char-
acteristics are in the same range [11, 12]. Additionally, natural fibres are worldwide
locally available, biodegradable, from renewable resources and need much less energy
for production than traditional ones.
Limited research work is published dealing with AAMs reinforced with natural
fibres [13–18] and the existing ones focus solely on matrices based on alkali-activated
pastes instead of mortars (with fine aggregates).
2 Materials and Experimental Methods
The mixtures used for the mortar contain fly ash as a solid prime material, water glass
as an alkali activator and sand. The fly ash is Micorsit20 produced by the NewChem
company. All particles in the fly ash are smaller than 20 lm. The chemical composition
of the fly ash is listed in Table 1. As the alkali activator water glass with the chemical
composition of 16.72% Na2O, 25.08% SiO2 and 58.2% H2O by mass was used. To
make a mortar, sand (with grain size 0.4-0.8 mm) produced by the Scherf GmbH &
Co KG company was used.
Table 1. Chemical composition of major oxides in fly ash.

Material Oxides (wt%)
SiO2 Al2O3 Fe2O3 CaO
Fly ash 52 25 7 5
As fibre reinforcement short, randomly dispersed primary bast hemp fibres (Can-
nabis sativa L) in the range of micro fibres (diameter from 8–60 lm) were used. The
morphology of a typical hemp fibre is presented in Fig. 1. The density of fibres was
1500 kg/m3 and according to the literature the tensile strength lies between 300–
1100 MPa, the Elastic modulus 23,5–90 GPa and the elongation at failure 1–3.5% [19].
The fibres consist of two cell walls and one inner part (the lumen). The primary wall’s
role is to protect the secondary wall which is responsible for the tensile strength of the
fibres [20]. Hemp fibres consist mainly of cellulose (74.4%), hemicellulose (17.9%),
lignin (3.7%), pectin (0.9%) and wax (0.8%) [20]. The length of the fibres was 10 mm
and the dosage 1vol.% of the total mixture. Generally, all plant-based fibres are
hydrophilic, therefore the hemp fibres were added to the mixture in a water saturated
dry surface condition.
Influence of the Curing Temperatures 247
Fig. 1. Morphology of hemp fibres
The fly ash to sand mass ratio was set to 1:3. The activator to fly ash ratio was
defined that the mass of Na2O from the activator is equal to 10% of the fly ash mass.
All mortar samples were prepared according to the European Standard EN 196-1 [21].
Six identical prism specimens were cast from each mortar group. The prism
specimens had dimensions of 40 40 160 mm3. After the mixture is poured into
the molds and left for one hour at laboratory conditions (20 °C temperature and 50%
relative humidity), the specimens were covered with a plastic foil and cured under
different conditions. Three different curing conditions were applied, i.e.: i) 28 days in a
climate chamber (20 °C temperature and 65% relative humidity); ii) 24 h in an oven
under the temperature of 60 °C and then additional 27 days in a climate chamber; iii)
24 h in an oven under the temperature of 80 °C and then additional 27 days in a climate
chamber. All specimens were tested at the age of 28 days.
Three-point bending test (3PBT) and compressive test were conducted and the
mortars flexural strength, energy absorption capacity (toughness) under flexure and
compressive strength were calculated. The tests were conducted on all specimen groups
according to ÖNORM EN 1015-11 [22]. 3PBT was carried out on a mechanical testing
machine Zwick/Roell Z250 with a load capacity of 200 kN, rigidity of 8 10−3
mm/kN under a deflection rate of 400 m/min at a room temperature of 21 °C and
relative humidity of 50%. The load was applied at the middle of the specimens with a
span length of 100 mm. The flexural strength was calculated from the maximal force
from the force-span deflection curve of the 3PBT and the energy absorption as the area
under the force-span deflection until the deflection of 6 mm, divided with a cross
section of the specimens (Fig. 2). Flexural strength and energy absorption are calcu-
lated for six specimens of each group and the mean value (with standard deviation) is
used as a representative.
The compressive tests were conducted on three halves (of three different) speci-
mens tested on 3PBT with applying the force on the area of 40 40 mm2 calculating
the mean value with standard deviation as a representative.
Fig. 2. Force-displacement curve from a fibre reinforced mortar group which is cured at 80 °C
The compressive strengths of non-reinforced and fibre reinforced mortar groups cured
under different conditions are presented in Fig. 3. Generally, it can be seen that when
cured at the higher temperature, the composites’ compressive strength increased. The
reason for that is the fact that higher temperatures speed up the chemical reactions. At
increased curing temperature the fly ash reacts faster with the water glass, forming
more minerals and increasing the compressive strength of the composite. However, the
faster chemical reactions and faster forming of a final structure of the mortar disturbs
the uniform development of adhesion of fibres and matrix which has in turn a negative
influence on the fibre reinforced mortars mechanical properties. With increasing curing
temperature the strength difference between plain and fibre reinforced mortars became
more pronounced.
Fig. 3. Compressive strength of the mortars
In case of cement-based mortars, the addition of fibres decreases the density of a

plain mortar and consequently also its compression strength [12]. When cured at higher
temperatures, fibre reinforced alkali-activated mortars resulted in lower compressive
strength than their corresponding plain mortar. The strength decreased for 8% and 27%
in case of 60 °C and 80 °C curing temperature respectively. But contrary, when cured
at room temperature (20 °C) the addition of fibres increased the plain mortar’s strength
for 4%. The reason could be that the lower temperature provides moderate chemical
reactions within the matrix, which resulted in the slower minerals formation and
consequently in better fibre-matrix bond. Generally, fibres in the matrix bridge cracks
and transfer the stresses which can increase the ultimate strength of the composite [14].
The flexural strength of mortars is presented in Fig. 4. The addition of short fibres
in composites generally does not have a significant influence on their flexural strength
since the flexural strength is mostly determined by the strength of the matrix itself. The
fibres’s main role is to carry stresses after the failure of the matrix [12]. However, at
lower curing temperature (20 °C) the flexural strength of the matrix increased for 57%
after the addition of the fibres. The reason for the increase could be similarly as in case
of the compressive strength, i.e. the moderate chemical reactions and mineral formation
within the matrix provide better adhesion between fibres and the matrix. The better
fibre-matrix bond contributes to better transfer of stresses by bridging the cracks. Under
60 °C curing temperature the flexural strength of mortars decreased for 14% when
reinforced with hemp fibres whereas at 80 °C the flexural strength remained almost
unchanged.
Fig. 4. Flexural strength of the mortars
The fibres main relevant contribution in a matrix is to increase the energy

absorption capacity of the composite. When the maximal flexural strength is reached in
a plain matrix a sharp drop occurs in the post peak region, indicating no deformation-
and energy absorption capacity of the material. On the other hand, when reinforced
with fibres the composite prolongs its deformation capacity and is able to additionally
carry a certain amount of stresses, increasing in such a way the overall energy
absorption capacity of the composite.
In the Fig. 5. the energy consumption of the non-reinforced and fibre reinforced
mortars cured at different temperatures are shown. It can be seen that the most sig-
nificant energy consumption improvement of the non-reinforced mortars is at room
curing temperature (20 °C). The energy consumption capacity of the mortar increased
for 711%. Under 60 °C and 80 °C curing temperature the energy consumption

increased for 331% and 153% respectively, when compared to their non-reinforced
counterparts. The highest energy consumption increase (at 20 °C) could be due to the
moderate matrix final forming, which results in an optimal attachment of fibres to the
matrix and have a proper bond. The too rapid growth of the minerals could result in
inappropriate fibres attachment to the matrix with a weaker fibre-matrix bond resulting
in lower energy consumption capacity of the composite.
Fig. 5. Energy absorption of the mortars
4 Conclusions
In the research plain and hemp fibre reinforced alkali activated mortars cured under
three different temperature were examined. The prism mortars with the dimensions of
40 40 160 mm3 were prepared using fly ash as solid prime material, water glass as
alkali activator, sand of particle size between 0.4 and 0.8 mm and short (10 mm),
randomly dispersed hemp fibres. Regarding curing regime, three different temperature
were used: 20 °C, 60 °C and 80 °C.
The results showed that with reinforcing mortars with hemp fibres:
• solely at curing temperature of 20 °C increased the composites compressive
strength (for 4%). At curing temperatures of 60 °C and 80 °C it decreased for 8%
and 27% respectively.
• the flexural strength of the mortars increased for 57% at the curing temperature of
20 °C and decreased for 14% at 60 °C. At 80 °C the flexural strength remained
almost unchanged.
• all mortars increased their energy consumption capacity. The most significant
increase (711%) was observed at the lowest curing temperature (20 °C), whereas at
60 °C and 80 °C a significantly lower increase of 331% and 153% respectively,
was observed.
Acknowledgements. The research is supported by the Stiftung Aktion Österreich-Ungarn in the

frame of a bilateral research cooperation project Nr. 101ou10 and this activity has received
funding from the European Institute of Innovation and Technology (EIT), a body of the European
Union, under the Horizon 2020, the EU Framework Program for Research and Innovation from
the project EIT RM RIS-ALiCE: Al-rich industrial residues for mineral binders in ESEE region,
no. 18258.
The authors greatly acknowledge the support of the NewChem company for providing the fly
ash for the research.
References
1. Cohen, B.: Urbanization, city growth, and the new united nations development agenda role
in achieving the least cost path to a sustainable. Cornerstone 3(2), 4–7 (2015)
2. Activity Report 2017 of The European Cement Association CEMBUREAU, (2017)
3. Bilodeau, A., Malhotra, V.M.: High-volume fly ash system: concrete solution for sustainable
development. Mater. J. 97, 41–48 (2000)
4. Pacheco-Torgal, F., Labrincha, J., Leonelli, C., Palomo, A., Chindaprasit, P.: Handbook of
Alkali-Activated Cements. Woodhead Publishing, Mortars and Concretes (2014)
5. Provis, J.L., Palomo, A., Shi, C.: Advances in understanding alkali-activated materials. Cem.
Concr. Res. 78, 110–125 (2015)
6. Provis, J.L.: Alkali-activated materials. Cem. Concr. Res. 114, 40–48 (2018)
7. Poletanovic, B., Kopecsko, K., Merta, I.: Durability of Hemp Fibre Reinforced Cementitious
Mortars by Means of Fibre Protection and Cement Substitution with Metakaolin. In:
International Conference on Interdisciplinary Approaches for Cement-based Materials and
Structural Concrete, Madeira Island, Funchal, Portugal; 24.10.2018–27.10.2018; in:
Cement-based Materials and Structural Concrete, p. 957–962 (2018)
8. Noushini, A., Hastings, M., Castel, A., Aslani, F.: Mechanical and flexural performance of
synthetic fibre reinforced geopolymer concrete. Constr. Build. Mater. 186(20), 454–475
(2018)
9. Tung, T., Tran, T.M., Hong Hao, P.: Experimental and analytical investigation on flexural
behaviour of ambient cured geopolymer concrete beams reinforced with steel fibers. Eng.
Struct. 200(1), 109707 (2019)
10. Sathanandam, T., Awoyera, P.O., Vijayan, V., Sathishkumar, K.: Low carbon building:
Experimental insight on the use of fly ash and glass fibre for making geopolymer concrete.
Sustain. Envir. Res. 27(3), 146–153 (2017)
11. Merta, I., Mladenovič, A,. Turk, J., Šajna, A., Pranjić, M.: Life cycle assessment of natural
fibre reinforced cementitious composites. In: 6th International Conference on Non-
Traditional Cement and Concrete, Brno, Czech Republic. 19.06.2017–22.06.2017
12. Merta, I., Šajna, A., Poletanović, B., Mladenović, A.: Influence of natural fibres on
mechanical properties and durability of cementitious mortars. In: CoMS - 1st International
Conference on Construction Materials for Sustainable Future, Zadar; 19.04.2017–
21.04.2017
13. Amalia, F., Akihaf, N., Nurfadilla, S.: Development of coconut trunk fiber geopolymer
hybrid composite for structural engineering materials. In: Materials Science and Engineering
(2017)
14. Alomayri, T., Low, I.M.: Synthesis and characterization of mechanical properties in cotton
fiber-reinforced geopolymer composites. J. Asian Ceram. Soc. 1, 30–34 (2013)
15. Alomayri, T., Shaikh, F.U.A., Low, I.M.: Characterisation of cotton fibre-reinforced
geopolymer composites. Compos. B 50, 1–6 (2013)
16. Chen, R., Ahmari, S., Zhang, L.: Utilization of sweet sorghum fiber to reinforce fly ash-
based geopolymer. J. Mater. Sci. 49(6), 2548–2558 (2013). https://doi.org/10.1007/s10853-
013-7950-0
17. Sá Ribeiro, R.A., Sá Ribeiro, M.G., Sankar, K., Kriven, W.M.: Geopolymer-bamboo
composite – A novel sustainable construction material. Constr. Build. Mater. 123, 501–507
(2016)
18. Assaedi, H., Shaikh, F.U.A., Low, I.M.: Characterizations of flax fabric reinforced nanoclay-
geopolymer composites. Compos. B 95, 412–422 (2016)
19. Yan, L., Kasal, B., Huang, L.: A review of recent research on the use of cellulosic fibres,
their fibre fabric reinforced cementitious, geo-polymer and polymer composites in civil
engineering. Compos. B 92, 94–132 (2016)
20. Merta, I., Poletanovic, B., Kopecsko, K.: Durability of natural fibres within cement-based
materials - review”, concrete structures. J. Hungarian Group fib (Federation International de
Beton) 18, 10–16 (2017)
21. ÖNORM EN 196–1, Methods of Testing Cement - Part 1: Determination of Strength, 2016
10 15
22. ÖNORM EN 1015–11: 2018 01 01, Methods of test for mortar for masonry - Part 11:
Determination of flexural and compressive strength of hardened mortar (2018)
Equivalence Between Flexural Toughness
and Energy Absorption Capacity of FRC
Sergio Carmona1(&) and Climent Molins2

1
Civil Engineering Department, Federico Santa María Technical University,
Valparaíso, Chile
sergio.carmona@usm.cl
2
Civil and Environmental Department, Barcelona Tech, Barcelona, Spain
Abstract. The increase of toughness and residual strengths are the great ben-
efits of incorporating fibres in concrete. However, there is still no agreement
regarding the most suitable experimental procedure for its determination and the
most accepted and used tests, the bending tests, are complex to execute, require
of sophisticated equipment with closed loop control and are unsuitable for the
systematic quality control in works. Due to the latter, some authors have pro-
posed equivalences between the different tests with the aim of simplifying the
control.
This paper presents the first stage of an experimental research in which a
linear correlation between the flexural toughness determined by mean of the
three-point bending test given in the European standard EN 14651 and the
energy absorption capacity determined by square panel test according to
EFNARC recommendation, was obtained, with differences less than 13%
between the experimental results and predicted values.
This achievement will allow to define a flexural equivalent resistance, based
on the energy absorption capacity for controlling fibre reinforced concrete
properties in the construction of tunnel linings.
Keywords: Toughness Energy absorption capacity Beam test Squared

panel test
1 Introduction
In recent years, fibres have begun to be widely used to reinforce concrete in the
construction of tunnel linings in different Latin American countries, such as Bolivia,
Chile, Colombia and Peru. However, a recurring problem, both for designers and
contractors, is the characterization and control of fibre reinforced concrete, because
most laboratories do not have the equipment and experience to perform the three-point
bending test (3PB) according to EN-14651 standard [1], which requires a closed loop
control (CLC) system for its execution.
In contrast, due to its greater simplicity and stability when concrete cracks, many
laboratories are able to perform the square panel test, as shown in Fig. 1, obtaining
satisfactory results. Nevertheless, this test does not allow determining the residual
strengths of fibre reinforced concrete, the parameters that, according to the MC 2010
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_23
254 S. Carmona and C. Molins
[2], characterize the post cracking resistance of FRC. Then arises the need to establish
an equivalence between both tests, which will allow to control fibre reinforced concrete
at works by mean of square panel test, and thus verify compliance with the project
specifications. In the particular case of Fig. 1, all data was recorded manually, dial
indicator measures were taken while testing and the load was applied by a manual jack.
Fig. 1 Square panel test performs in non-standard testing system in a laboratory in Bolivia.
In order to produce a robust relationship between the square panel and the 3PB
tests, the state-of-the-art on comparing test methods for the mechanical characterization
of fibre reinforced concrete was reviewed. Many authors establish correlations among
standard tests based on the crack width [3, 4. 5, 6]. Nevertheless, the crack width in the
square panel test is not measured as is in the 3PB test, and the panels exhibit many
cracks in the final state, at the end of the test.
Even though the failure mechanisms between both test might be different, this
should not pose a problem to obtain a correlation between both tests. In fact, several
codes and studies from the literature propose correlations between the results of test
methods with completely different cracking mechanisms [7].
Using this idea, a code-type expression has been proposed in which the energy
absorption capacity, determined by testing squared panels, and energy dissipated by
cylinders subjected to double punching test or Barcelona test are correlated [8] and [9].
With that correlation, a good fit between the absorbed energy at a deflection of 25 mm
in square panel with the dissipated energy at a total crack opening displacement of
6 mm.
Taking in account those results, in this research, it was decided to correlate the
energy measurement in both tests. However, in standardized 3PB test there is no
definition of toughness at all. Then, a new definition of a toughness parameter was
necessary for that. The aim of this paper is to present the results of the first stage of an
experimental research, design to develop a correlation between the energy absorption
capacity by the square panel test and flexural toughness obtained in 3PB bending test.
Equivalence Between Flexural Toughness and Energy Absorption 255
2 Square Panel Test
According to the EFNARC recommendations [10], this test is conducted on a square

panel of 600 mm side and 100 mm thick supported on its four edges, which is loaded
at the centre by a contact surface of 100 100 mm, as can be seen in Fig. 2. Load is
applied under deformation control at a rate of midspan deflection of 1.5 mm/min. This
test is similar to the square panel test defined in EN 14488-5 standard [11].
The load – deflection curve shall be continuously recorded until a displacement of
25 mm is achieved at the centre point of the panel. Using this response, the energy
absorption capacity, (d), can be calculated as:
25
Z
E25 ¼ PðdÞdd ð1Þ
0
where PðdÞ is the load as function of deflection.
Fig. 2 EFNARC square panel test setup [10].
3 Three-Point Bending Test (3PB)
As is well known, according to EN-14651 standard [1], the characterization of the

properties of the FRCs is carried out by testing a beam with a 25 mm deep central
notch, loaded at the midspan (3PB), as shown in Fig. 3. The test should be carried out
in a closed loop control system under crack mouth opening displacement (CMOD)
control, at a rate of 0.05 mm/min, and with a sufficiently high rigidity of the loading
system to avoid instability in the transition between the pre and post cracking stage.
During test, the load and CMOD must be recorded continuously at a rate of 5 Hz.
Using the P CMOD response obtained during test, the EN-14651 standard
defines the residual strengths, FR;j , as:
3Pi l
FR;j ¼ ð2Þ
2bh2sp
where l = 500 mm, b = 150 mm and hsp = 125 mm are the span, width and height of
the specimen, respectively, shown in Fig. 2, y Pi , with i = 1, 2, 3, 4, is load at
CMOD1 = 0.5 mm, CMOD2 = 1.5 mm, CMOD3 = 2.5 mm and CMOD4 = 3.5 mm,
respectively.
Based on the toughness definition given by ASTM C–1609 standard [12] and
considering the linear relationship between the CMOD and the deflection of the beam,
which is validated in the EN-14651 standard, in this research the flexural toughness has
been defined, T3:5 , as the area under the curve P CMOD up to CMOD4 = 3.5 mm,
calculated as:
3:5
Z
T3:5 ¼ PðCMODÞdCMOD ð3Þ
0
This value of flexural toughness will be used in this research to establish an

experimental correlation with the energy absorption capacity of the FRC.
Fig. 3. Test setup as EN – 14651 standard [1].
4 Experimental Research
All the experimental research was developed in the Laboratory at Federico Santa Maria
Technical University in Valparaiso, Chile.
4.1 Tested FRC

The concretes were prepared with a Chilean pozzolanic cement of Type IP [13] and
crushed river sand; the mix proportions are presented in Table 1. Considering that
Barchip 48 (BC – 48) is one of the most widely fibres used in Latin American project,
in this research eight different contents of those fibres were used to reinforce concretes.
The fibres features are given in Table 2 and are shown in Fig. 4.
Table 1. Features and properties of tested concretes.
Material FRS
(kg/m3) 1.5 2.0 2.5 3.0 4.0 6.0 8.0 12.0
Cement type IP 420
Sand 0/10 1955
Superplasticizer admixture 2.10
Water reducer admixture 2.10
Active admixture 2.94
Free water 156
Fiber content 1.5 2.0 2.5 3.0 4.0 6.0 8.0 12.0
Beams 10
Number of
Square panel 10
specimens
Cylinders 3
FRS properties
Compressive strength, (MPa) 45.3 42.1 39.9 43.0 46.7 44.5 43.7 41.2
Slump (mm) 255 260 245 250 235 210 220 180
(%) 0.16 0.22 0.27 0.33 0.44 0.66 0.88 1.32
The concretes were prepared at laboratory using a conventional paddle mixer of

200 litres’ capacity. For each FRC 600 600 100 mm square panels, standardized
beams and three compression cylindrical specimens were cast. Fibre content, volu-
metric substitution (Vf ), slump and compressive strength (fc ) are given in Table 1.
Table 2. Synthetic fibres properties (manufacturer’s data).

Designation lf df kf fst E Fibers/kg
(mm) (mm) lf/df (MPa) (GPa) (N°)
BC – 48 48 0.70* 68.6 640 12 59500
(*) Equivalent diameter determined with the manufac-
turer’s information.
Fig. 4. View of fibres Barchip – 48 used in this research.

4.2 Tests and Results

All the tests were conducted in a hydraulic closed-loop control system of 100 kN
capacity, under deformation control. In the panel tests, the deflection was measured
with a LVDT of 50 mm range, placed in the centre bottom face of the specimen.
The 3PB tests were conducted according to standard EN–14651 under CMOD
control, which was measured using a clip-gauge extensometer with a total range of
5.0 mm, at a rate of 0.05 mm/min until reaching a CMOD = 0.1 mm. Then, the speed
was increased to 0.2 mm/min until the end of the test. Before to the bending tests, a
25 mm deep notch was cut at midspan of each beam. In both tests, the load and the
deformation were recorded continuously by the testing system at a rate of 5 data/s.
The mean curves obtained with each tested concrete are shown in Fig. 5, in which
can be clearly seen the effect of the fibre content in the post peak response of FRC. As
can be observed in Fig. 5a, for concretes reinforced with low fibres content, i.e. less
and equal to 4.0 kg/m3, the curves exhibit two peaks related with flexural cracks. The
loads corresponding to the first and second cracks do not seem to depend on the amount
of reinforcing fibre. After the second peak, the curves show a softening behaviour, with
a decreasing slop as fibre content increases. However, in the panels reinforced with
medium and high amounts (FRS–6.0, FRS–8.0 and FRS–12.0), a flexural crack was
initially open, which caused the first peaks that can be observed in the P d curves.
However, due to the higher fibres content, the testing system had to increase the applied
load to maintain the established deformation rate, which in addition to the friction force
developed in the supported, gave rise to a punching failure, which is reflected in the
formation of cracks around the loaded section.
At the same time, in the P – CMOD curves show in Fig. 5b, it can be seen that the
first peak does not depend on fibre content, nevertheless the post peak behaviour varies
from a deep softening for concrete reinforced with low fibre content to strong hard-
ening for high fibre content.
(a) (b)
Fig. 5. Mean curves load–deformation obtained with (a) square panel tests, and (b) 3PB tests.
5 Experimental Correlation Proposed
With the results of panel and beam tests, the energy absorption capacity and flexural
toughness were calculated using the Eqs. (1) and (3), respectively. The diagrams
showing the energy absorption–deflection and flexural toughness–CMOD are shown in
Fig. 6.
(a) (b)
Fig. 6. (a) Curves energy absorption – deflection of panels and (b) flexural toughness obtained
with 3PB tests.
In order to obtain an experimental correlation based on standardized parameters that

show the FRC behaviour in advanced cracking states, the values of energy absorption
capacity of square panels at a deflection of 25 mm, denoted as E25 , and the flexural
toughness of beams at a CMOD = 3.5 mm, denoted as T3:5 , were used. The values of
E25 and T3:5 are summarized in Table 3, along with the coefficient of variation in
bracket. As can be seen, the results scatter trends to decrease when fibres content
increases and, in all cases, the scatter is larger in the 3PB tests.
Table 3. Mean results of conducted tests.

Concrete Fibre content E25 T3:5 Equation (4) Difference
(kg/m3) (J) (J) (J) (%)
FRS – 1.5 1.5 400 (18.6) 9.3 (19.0) 8.5 −5.3
FRS – 2.0 2.0 482 (15.4) 12.3 (21.9) 13.1 6.9
FRS – 2.5 2.5 553 (16.6) 18.3 (22.5) 17.1 −6.6
FRS – 3.0 3.0 637 (15.6) 19.4 (20.5) 21.8 12.7
FRS – 4.0 4.0 669 (14.8) 25.2 (17.0) 23.6 −6.1
FRS – 6.0 6.0 859 (13.3) 39.5 (16.0) 34.3 1.3
FRS – 8.0 8.0 1098 (9.5) 48.7 (15.1) 47.7 −2.1
FRS – 12.0 12.0 1744 (6.1) 83.8 (14.3) 83.9 0.3
The values of E25 and T3:5 are plotted in Fig. 7, where it can be seen there is a linear
relationship between both parameters, given by the expression:
T3:5 ðE25 Þ ¼ 0:0561 E25 13:92 ð4Þ
where T3:5 ðE25 Þ is the predicted flexural toughness as function of energy absorption
capacity, E25 . This expression fits well with the experimental data, with a correlation
coefficient r 2 ¼ 0:9972, and differences between experimental and predicted values
less than 12.7%, as can be seen in the Table 3.
Fig. 7. Linear correlation between E25 and T3:5 obtained for BC - 48.
The high value of the correlation coefficient ðr 2 Þ of this linear relationship shows
that there is a direct proportionality between the flexural toughness at a CMOD =
3.5 mm, equivalent to a beam deflection of 3.02 mm [1], and the energy absorption
capacity determined at a deflection d = 25 mm, for a wide range of fibre content, which
will allow establishing a correlation between equivalent strengths determined using the
flexural toughness and the energy absorbed by the square panel.
6 Conclusions
The conclusions of this research are the following:

• In recent years, the use of fibre reinforced shotcrete has replaced steel mesh rein-
forced shotcrete in tunnel linings construction in Latin America. However, the lack
of laboratories to characterize and control of fibre reinforced concrete properties is a
very common limitation.
• To overcome that limitation, this paper proposes to establish an empirical equiva-
lence between the square panel test and the notched beam test given in standard EN
14651, since the former is easier to perform with simple laboratory equipment.
Then, a flexural toughness has been defined as the area under the load-CMOD
curve, until a CMOD = 3.5 mm.
• To work out the correlation, an experimental research was carried out using con-
cretes reinforced with different synthetic fibre contents, obtaining for this particular
fibre, a linear relationship between the energy absorption capacity and the flexural
toughness, with differences less than 13% between the experimental results and
values predicted using the proposed linear correlation.
• This achievement will allow to define a flexural equivalent resistance, based on the
energy absorption capacity for controlling fibre reinforced concrete properties in the
construction of tunnel linings.
Acknowledgments. This research was supported by Fondecyt Project “Use of the Generalized
Barcelona Test for Characterization and Quality Control of Fibre Reinforced Shotcretes in
Underground Mining Works”, N°1150881.
References
1. European Committee for Standardization, EN 14651: Test method for metallic fibre
concrete–measuring the flexural tensile strength (Limit of Proportionality (LOP), Residual)
(2007)
2. CEB-FIP: Model code – first complete draft. FIB Bull. 55, 1–318 (2010)
3. Alberti, M.G., Enfedaque, A., Gálvez, J.C.: On the mechanical properties and fracture
behavior of polyolefinfiber-reinforced self-compacting concrete. Const. Build. Mater. 55,
274–288 (2014)
4. Conforti, A., Minelli, F., Plizzari, G.A., Tiberti, G.: Comparing test methods for the
mechanical characterization of fiber reinforced concrete. Struct. Concr. 19, 656–669 (2018).
https://doi.org/10.1002/suco.201700057
5. Carmona, S., Molins, C., Aguado, A.: Correlation between bending test and Barcelona tests
to determine FRC properties. Const. Build. Mater. 181, 673–686 (2018)
6. Carmona, S., Molins, C.: Use of BCN test for controlling tension capacity of fibre reinforced
shotcrete in mining works. Const. Build. Mater. 198, 399–410 (2019)
7. Galeote, E., Blanco, A., Cavalaro, S.H.P., De la Fuente, A.: Correlation between the
Barcelona test and the bending test in fibre reinforced concrete. Const. Build. Mater. 152,
529–538 (2017)
8. Carmona, S., Molins, C.: Application of BCN test for controlling fibre reinforced shotcrete
in tunnelling works in Chile. IOP Conf. Ser. Mater. Sci. Eng. 246, 012010 (2017)
9. Carmona, S., Molins, C., García, S.: Application of Barcelona test for controlling energy
absorption capacity of FRS in underground mining works. Const. Build. Mater. 246, 118458
(2020). https://doi.org/10.1016/j.conbuildmat.2020.118458
10. European Federation of national associations of specialist contractors and material suppliers
for the construction industry, European Specification for Sprayed Concrete (1996)
11. European Committee for Standardization: EN 14488-5:2006 Testing sprayed concrete.
determination of energy absorption capacity of fibre reinforced slab specimens (2006)
12. ASTM ‘C1609/C1609M-19: Standard test method for flexural performance of fibre-
reinforced concrete (Using Beam With Third-Point Loading)’, ASTM International, West
Conshohocken, PA (2019). www.astm.org
13. ASTM: C595/C595M-18, Standard specification for blended hydraulic cements. ASTM
International, West Conshohocken, PA (2018). www.astm.org
Alkali Resistant (AR) Glass Fibre Influence
on Glass Fibre Reinforced Concrete
(GRC) Flexural Properties
S. Guzlena(&) and G. Sakale
Institute of Technical Physics, Faculty of Material Science and Applied

Chemistry, Riga Technical University, P. Valdena 7, Riga, Latvia
sandra.guzlena@rtu.lv
Abstract. Glass fibre reinforced concrete (GRC) is lightweight material mostly

used for façade panels and decorative elements. GRC can be made using two
methods – spraying and premixing. Glass fibre in both cases has main influence
on material flexural properties and ductility. Historically ordinary E type glass
fibre has been used, but during concrete aging and alkaline medium fibres
become fragile (weight and diameter loses). New type, alkali resistant
(AR) glass fibres have been developed. In this research AR glass fibre amount
and length influence on GRC flexural properties is investigated. Fibre length
was changed from 6 mm till 41 mm for different samples and cut during
spraying process. Fibre amount was changed from 0–7%. Samples were anal-
ysed using SEM-EDX to evaluate AR glass fibre and concrete matrix bond.
GRC mechanical properties was evaluated using four-point bending tests and
characterised by level of proportionality (LOP) and modulus of rupture (MOR).
.
Keywords: GRC AR glass fibre Flexural strength
1 Introduction
Concrete is most widely used engineering material in construction all over the world
due to its strength, durability and low cost as compared to other construction materials.
The major drawback of concrete is its low tensile strength. Crack appearance due to
stress during loading reduces concrete strength, durability and makes concrete more
vulnerable to deleterious outside environment [1]. To increase tensile strength different
fibres, like organic, polymer, basalt, steel and glass fibres can be incorporated in matrix.
Glass fibre reinforced concrete (GRC) is a composite which consist of concrete and
glass fibres. Glass fibre has low water absorption, high modulus of elasticity and tensile
strength [2]. Fenu et al. have even concluded that basalt fibre reinforcement was shown
to be less performing than glass-fibre reinforcement when used under dynamic con-
ditions [3]. This material is used for non-structural parts of buildings like facade panels
for over than 30 years. GRC is known for its high tensile and impact strength which is
due to glass fibres. These properties allow make GRC panels 12 mm thin [4–6]. There
are two types of techniques used to produce GRC elements. First of them is premixing,
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_24
Alkali Resistant (AR) Glass Fibre Influence on Glass Fibre Reinforced Concrete (GRC) 263
when concrete is mixed with chopped glass fibres and then casted in formwork. Fibre
length usually is around 12 mm and it is added 3% of mortar mass. If the fibre length
and amount is increased workability of the mix is reduced. Barluenga et al. have noted
that using short fibres in premix the cracked area of concrete surface can be reduced. If
crack appears perpendicular to fibre it limits the size of the crack, but if the crack
appears parallel to fibre, reinforcement does not work and crack size grows freely [7].
For spray technique fibre length can be changed, but usually it is 25–40 mm and 5% by
mortar weight. During spraying process uniform well, mixed composite is achieved
comparing to premix method. This type of material is with high durability, impact and
tensile strength [5, 8].
Glass fibres used in material must be alkali resistant due to concretes high alka-
linity. In the past, ordinary glass fibres, like E glass fibres were used, but during the
time they showed low durability. Fibres lose their durability due to hydration reaction
in concrete. Portlandite (Ca(OH)2) is produced during hydration process and increases
alkalinity in cementitious matrix to high alkalinity (pH 12). Hydroxyl ions (OH-),
which is at high concentration in cementitious matrix due to alkaline pore solution,
brakes Si-O-Si bonds in glass fibres Eq. 1 [9, 10] and causes weight and diameter loses
making fibres fragile [6, 8, 9, 11].
in solution ð1Þ
To improve glass fibre chemical stability in concrete ZrO2 have been added. In the
EN 15422:2008 is mentioned AR glass fibre must contain minimum 16% of ZrO2, to
use it in concrete. But ZrO2 addition do not solve his problem completely, as researches
showed decrease in aged GRC flexural properties comparing to young GRC [12–14].
There exist different methods to reduce fibre embrittlement. Admixtures like fume,
metakaolin, nano silica can be added to a concrete mixture. This permits pozzolanic
reactions and transforms portlandite to C-S-H in this way decreasing alkalinity of the
concrete [2, 6, 9]. Use of low alkalinity cement, like sulphoaluminate cement could
solve the problem, because Portlandite (Ca(OH)2) is not produced during hydration
process, but pH value in pore solution still is high, around pH 10. To densify the
interface between fibers and cement matrix with polymers, like PVA, AC and others.
This technique reduces lime diffusion into fibres [9, 10].
Due to thickness and common application of GRC tensile test is used to evaluate
GRC properties. Level of proportionality (LOP) and modulus of rupture (MOR) is used
to describe sample tensile strength.
LOP value describes sample maximal linear elastic deformation, which means
every load that is put below LOP will not harm material, no cracks will be seen on
surface and material will return to its present state. Sample elastic region, below LOP,
can be described by Hookes law stress is proportional to the strain and the slope is
Young’s modulus. Basically, LOP value demonstrates matrix maximum flexural
strength. After load is increased over LOP value material exhibits plastic behaviour, it
deforms irreversibly. In this moment multiple cracks can appear on GRC surface, but
264 S. Guzlena and G. Sakale
material still have load bearing properties. Increasing load elongation increases of GRC
sample due to fibres connection in concrete matrix until crack is too large and fibres
pull out or brake at fracture point. MOR values describe samples ultimate strength
before failure [2, 8, 13]. LOP and MOR values are influenced by many factors as:
• fibre amount,
• fibre length,
• sprayed GRC or premix is used,
• compaction.
In this research impact of amount of glass fibres in composition, length of fibre on
sprayed GRC properties have been evaluated.
GRC samples were made using spraying method. Basic components of GRC was
cement CEM I 52,5R, fine quartz sand, superplasticizer and acrylic polymer. AR-glass
fibre roving was used. Fibre length from 6 mm till 41 mm was changed for different
samples and cut during spraying process. Fibre amount was changed from 0–7% by
weight of GRC material (concrete matrix) in the uncured, green state. Sample flexural
test was carried according to BS EN 1170 [15]. Samples were cut from test board in
size 275 50 mm and tested with 4-point bending test after 28 days. GRC samples
and AR glass fibres were investigated using Phenom ProX Desktop SEM.
3 Results
3.1 SEM – EDX Analysis

For all tests have been used AR glass fibre roving which was cut during manufacturing
process. In SEM image (Fig. 1) can be seen bundle of AR glass fibres. Fibre surface is
smooth, and the shape is regular. EDX analysis was done on glass fibre fracture zone,
marked in Fig. 1. In Table 1 is shown element analysis. Base oxides in fibre is Si, Na
and Zr. With EDX analysis is not possible to obtain precise amount of ZrO2 amount in
fibre, but this method can be used to indicate Zr in fibre and to know that this is AR
glass fibre.
In Fig. 2 calcium carbonate crystals [16] are grown on surface of AR glass fibre
which means that glass fibre surface has good bonding abilities with concrete matrix.
Fenu et al. have also noted that fibres in fracture are almost completely covered by a
thin surface layer formed by the products of reaction between the outer alkali resistant
glass of the fibre and the cementitious matrix [3].
In Fig. 3 is shown fibre bundles incorporated in GRC matrix. GRC performance in
strength depends on mortar bonding with fibre filaments, fibre bundles and concrete
matrix strength [8]. As can be concluded from the SEM image fibre bundles have good
incorporation in matrix, no air voids or cracks can be seen around interface.
Fig. 1. SEM image of AR-glass fibre used in GRC samples. EDX analysis was done at marked
location.
Table 1. AR-glass fibre element analysis

Element symbol Weight concentration, %
O 46.92
Si 25.32
Na 12.72
Zr 9.38
Ti 2.91
C 2.75
Fig. 2. AR-glass fibre in GRC sample

Fig. 3. AR-glass fibre bundles in GRC sample cross section

Concrete as itself is brittle material in tensile strength. Fibre addition increase material
elastoplastic behaviour. In this section GRC property dependence from fibre amount
and length are presented.
AR glass fibre addition to concrete matrix has significant impact on sample ultimate
load bearing capacity. As shown in Fig. 4, if 41 mm long fibres are used MOR values
are increased with increase of fibre amount in the concrete, meanwhile LOP values do
not change significantly. LOP value is more affected by the strength of concrete matrix
than fibre addition [2].
Concrete as material is nonelastic, brittle material, as shown in Fig. 4, LOP and
MOR values are almost the same if no AR glass fibres are added to concrete matrix. In
Fig. 5 is shown that test samples without fibres are with very low plasticity and brittle
fracture point. When the amount of fibres is increased from 0% to 5% and 7%, material
elasticity is increased. By increasing fibre amount from 5% to 7% increases the load-
bearing capacity of the material, but as the fibre length has not change the deflection of
material does not change.
Fibre length was changed from 6 mm to 41 mm, but amount during these exper-
iments was 5%. In Fig. 6 is shown LOP/MOR dependence from fibre length. Fibre
length do not considerably change LOP values. As we concluded before the LOP
values are more dependent on the matrix properties of the concrete. MOR values
increases by increasing fibre length from 6 mm to 41 mm.
As shown in Fig. 7 by adding longer fibres plastic behaviour of material increases.
H. Kasagani et al. have also noted that samples with longer fibres have higher defor-
mation capacity comparing to samples with shorter fibres [17]. Longer fibres have
higher chance to defom in the bundle structure improving post crack deformation and
35
30
25
Strenght, MPa
20
15
10
0
0% 5% 7%
LOP (7d) MOR (7d) LOP (28d) MOR (28d)
Fig. 4. Influence of fibre amount on LOP/MOR values
Fig. 5. Influence of fibre amount on deflection
still keep increment of force applied to GRC. Analysing collapse of samples can be
seen that using longer fibres deflection stays constant for higher ΔForce. The filaments
slowly pulls out of the bundle, not brittle when applied under load, crack size
(deflection) do not change.
35
30
25
Strenght, MPa
20
15
10
0
6 mm 12.5 mm 41 mm
LOP (7d) MOR (7d) LOP (28d) MOR (28d)
Fig. 6. Influence of fibre length on LOP/MOR values
Fig. 7. Influence of fibre length on deflection of material
4 Conclusions
Using AR-glass fibres is possible to change GRC flexural properties:

• by increasing fibre amount from 5% to 7% increase the load-bearing capacity of the
material, but as the fibre length of all samples have been the same (41 mm) the
deflection of material does not change.
• LOP value, which describes material elastic deformation, is dominated by concrete
matrix flexural strength, addition of fibres has inessential effect.
• by increasing fibre length from 10 mm till 41 mm, MOR values increases con-
siderably due to composite plastic behaviour.
Acknowledgements. Authors of the article are grateful to LTD Skonto Concrete Cladding for a
materials and opportunity to use spraying equipment.
References
1. Khaliq, W., Ehsan, M.B.: Crack healing in concrete using various bio influenced self-healing
techniques. Constr. Build. Mater. 102, 349–357 (2016)
2. Madhkhan, M., Katirai, R.: Effect of pozzolanic materials on mechanical properties and
aging of glass fiber reinforced concrete. Constr. Build. Mater. 225, 146–158 (2019)
3. Fenu, L., Forni, D., Cadoni, E.: Dynamic behaviour of cement mortars reinforced with glass
and basalt fibres. Compos. Part B Eng. 92, 142–150 (2016)
4. Branco, F.A., Ferreira, J., Brito, J.D.E., Santos, J.R.: Building structures with GRC
Fernando A. Branco, João Ferreira, Jorge De Brito and José R. Santos, no. April, pp. 1–11
(2001)
5. Ferreira, J.G., Branco, F.A.: GRC mechanical properties for structural applications, vol. 1,
no. 1, pp. 1–20
6. Correia, J.R., Ferreira, J., Branco, F.A.: A rehabilitation study of sandwich GRC facade
panels. Constr. Build. Mater. 20(8), 554–561 (2006)
7. Barluenga, G., Hernández-Olivares, F.: Cracking control of concretes modified with short
AR-glass fibers at early age. Experimental results on standard concrete and SCC. Cem.
Concr. Res. 37(12), 1624–1638 (2007)
8. Bartos, P.J.M.: Glassfibre reinforced concrete: a review. IOP Conf. Ser. Mater. Sci. Eng. 246
(1) (2017)
9. Arabi, N., Molez, L., Rangeard, D.: Durability of alkali-resistant glass fibers reinforced
cement composite: microstructural observations of degradation. Period. Polytech. Civ. Eng.
62(3), 1–7 (2018). SE-Research Article
10. Scheffler, C., et al.: Interphase modification of alkali-resistant glass fibres and carbon fibres
for textile reinforced concrete I: fibre properties and durability. Compos. Sci. Technol. 69(3–
4), 531–538 (2009)
11. Genovés, V., Gosálbez, J., Miralles, R., Bonilla, M., Payá, J.: Ultrasonic characterization of
GRC with high percentage of fly ash substitution. Ultrasonics 60, 88–95 (2015)
12. Moceikis, R., Kičaite, A., Keturakis, E.: Workability of glass reinforced concrete
(GRC) with granite and silica sand aggregates. IOP Conf. Ser. Mater. Sci. Eng. 251(1)
(2017)
13. Enfedaque, A., Cendón, D., Gálvez, F., Sánchez-Gálvez, V.: Analysis of glass fiber
reinforced cement (GRC) fracture surfaces. Constr. Build. Mater. 24(7), 1302–1308 (2010)
14. Holubova, B.: Corrosion of Glass Fibres in Ultra High Performance Concrete and Normal
Strength Concrete. Ceram. Silikaty 61(4), 1–9 (2017)
15. “EN 1170-5.pdf.”
16. Davies, R., et al.: Multi-scale cementitious self-healing systems and their application in
concrete structures. In: 9th International Concrete Conference 2016: Environment, Efficiency
and Economic Challenges for Concrete, 4–6 July 2016 (2016)
17. Kasagani, H., Rao, C.B.K.: Effect of graded fibers on stress strain behaviour of Glass Fiber
Reinforced Concrete in tension. Constr. Build. Mater. 183, 592–604 (2018)
Fiber Reinforced Concrete Crack Opening
Evaluation Using Digital Image Correlation
Techniques
Kaio Cézar da Silva Oliveira(&), Gabriela Silva Dias,

Isadora Queiroz Freire de Carvalho,
Wandersson Bruno Alcides de Morais Silva,
Danilo José Pereira Freitas,
Christiano Augusto Ferrario Várady Filho,
and Aline da Silva Ramos Barboza
Technology Center, Structure and Materials Laboratory,

Federal University of Alagoas - UFAL, Maceió, Brazil
kaio.oliveira@ctec.ufal.br
Abstract. The analysis of mechanical properties in fiber reinforced concrete

(FRC) elements is basically done through destructive tests since the results
obtained by these methods are already well established in normative codes. One
of them is the 3-point flexural test normalized by EN 14651 using a notched
beam to measure the crack width (CMOD). Within the context of the mechanical
properties evaluation that do not require the production or extraction of speci-
mens, and can be applied in fully functioning structures, the digital image
correlation (DIC) is a technique which has been proposed. This non-destructive
test analyzes a group of images correlating one with each other, evaluating the
changes that occurred during the load has been applied. It is a non-invasive test,
capable of results with acceptable precision and a considerable low cost, proving
to be a promising technique in the field of behavior analysis. Thus, this study
compares the cracking results obtained through the extensometry technique
(Linear Variable Differential Transformer – LVDT) and the digital image cor-
relation. The samples were made using steel fibers. The results obtained using
the DIC technique were validated by the data obtained through the LVDTs, and
the absolute error was considerably low.
Keywords: Concrete Correlation Fiber Image
1 Introduction
Concrete is one of the most used materials in civil construction, however, when sub-
mitted to tensile efforts, it shows a low resistance and deformation capacity. In view of
the problem presented, the addition of staple fibers (steel, glass, polypropylene and
natural) represents an improvement alternative for such behavior, since the main
contribution of the fibers occurs in the post-cracking stage. The fibers act as stress
transfer bridges between the cracks, redistributing the efforts in the cementitious matrix,
and ensuring that even after the cracking the composite presents resistive capacity,
restricting a sudden rupture [1].
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_25
Fiber Reinforced Concrete Crack Opening Evaluation 271
In order to expand the use of Fiber Reinforced Concrete (FRC) in elements for
structural purposes, a working group called Special Activity Group (SAG 5) of the
Federation Internationale du Beton (FIB), included in the fib Model Code 2010 (2013)
sections addressing material capacity and recommendations for sizing and behavior
analysis. Configuring itself as an important step for the SFRC, since it defined the
minimum performance requirements in view of the safety limit states Ultimate Limit
State (ELU) and Service Limit State (ELS). Although not a normative character, the fib
Model Code 2010 (2013) is a document that aims to subsidize future regulations on the
SFRC [2].
The fib Model Code 2010 (2013) establishes the test standardized by EN 14651
(2007) for the characterization of the FRC for structural purposes. It consists of a beam test
under three-point bending carried out in a closed system with displacement speed control,
and through it, it is possible to determine the post-crack strength, the SFRC toughness and
the crack opening induced by a notch at the lower end of the specimen [3].
The code states that alternative tests can be used to determine the same mechanical
performance parameters if the results found are correlated with the reference test. From
this context follows the proposal of the present work, which aims to use the correlation
of digital images.
The behavior assessment method using digital image correlation is suitable as a
Non-Destructive Test (NDT), since there is no need to impose any type of deformation
or destructive process on the structure, it is interesting to use the technique for structures
already in full operation, being a technique focused on the monitoring of structural
health [4–5].
2 Digital Image Correlation
DIC is a technique that allows studying qualitatively, and quantitatively, the

mechanical behaviour of materials when submitted to different types of requests. Ini-
tially proposed by Sutton et al. (1983), the technique is based on the comparison of
pairs of digital images, captured at different times of the loading application. Each of
the images is composed of pixels, and each pixel has a grayscale value, which varies
from 0 to 255, according to the amount of light reflected by the surface of the specimen
(Fig. 1).
Fig. 1. Grayscale intensity matrix and its digital image correlation.

272 K. C. da Silva Oliveira et al.
As there may be pixels with the same gray intensity value, a region or set of pixels,
called a subset, is defined and used for the correlation method. After defining the subset
in the image captured before the deformation, a computational tool is used to identify
the post-deformation image, for that same set of pixels. The result obtained for the
central part of the subset is the average of the displacements suffered by each of the
pixels that compose it (Fig. 2) [6–11].
Fig. 2. Displacement of a subset during specimen deformation.
A displacement vector is generated for each subset, and at the end of the analysis of
all defined subsets, a displacement field results. It is important to highlight that the
unique pattern of subsets is guaranteed when the region studied has no repetitions,
isotropic and has high contrast patterns. Obtaining significant results through the
method is directly related to the lighting conditions and standardization of the surface
captured by images [12].
3 Dic and Crack Opening Evaluation
One of the main parameters of the behaviour of steel fiber reinforced concrete (SFRC)
obtained through the test recommended by EN14651 (2007), is the opening of the
notch in the lower face of the beam, located in the middle of the span, called CMOD.
This parameter is recorded during the experiment using mechanical equipment such as
clip-gage or displacement transducer.
For the application of digital image correlation, the displacement values were
recorded in real-time through a library of the ITOM ® software, which was created and
has been updated by researchers from the Mechanical Research Group on Advanced
Structures and Materials from Federal University of Alagoas, to assist in the process of
acquiring the data generated by the LVDTs (Linear Variable Differential
Transformers), shown in Fig. 3, and to automate the capture of the images in a certain
time interval. Still, with the help of this library, image processing, crack opening
analysis, deformation and displacement fields are carried out.
Fig. 3. LVDT positioning details.
The LVDTs were connected to a system for the acquisition of analog signals, the
Spider 8®. The test apparatus is shown in Fig. 4. The choice of blue LEDs for lighting
the specimen was made from previous tests, which proved that this color allowed
greater contrast for the captured surface.
The camera used for the tests has a CMOS sensor with a resolution of 24.5 MP. To
capture the images, it was necessary to carry out a camera calibration process, with
which a conversion factor from pixels to millimetres is found. This conversion factor is
obtained with the aid of an ITOM® tool, which receives the distance between the
center of two pixels (pixel pitch) as input data. Then, about 10 (ten) photographs are
taken of a checkered mesh generated on the computer screen, as shown in Fig. 5,
without changing the focus or zoom of the camera, nor the positioning of the tripod,
and they are transferred to the test computer and loaded into ITOM® for a calibration
whose reference was pixel pitch. In this calibration, ITOM® determines the conversion
factor.
The image analysis process starts with reading at time t = 0 s, using ITOM®. For
this reading, the software uses a tool called linecut and generates a graph (Fig. 6), in
which the X-axis refers to the number of pixels contained in the linecut and the Y-axis
to the intensity of gray tones (luminosity data). To avoid measurement errors, the
linecut is standardized for all test images, aligned to the LVDTs.
Analyzing the graph generated on the Fig. 6, it is noticed that when the line cuts a
region with shades of gray that tend to black, the Y values fall dramatically. In turn,
when the linecut cuts pixels with tones closer to white, the Y values increase. With this,
when analyzing the graph of Fig. 6, the crack opening will be given by the number of
pixels between the red and green points called width, which indicates the beginning
Fig. 4. Test apparatus.
Fig. 5. Checkerboard used in camera calibration.

Fig. 6. Detail of crack opening obtained through ITOM.
and end of the crack, respectively. With the crack opening values (width), conversion
to millimeters is performed, multiplying the values obtained by the conversion factor
described above.
4 Results
To validate the image correlation process in obtaining the crack opening of beams
tested under three-point bending, a comparison of the values obtained by each method
was performed and then a percentage difference between the opening read by the
LVDT and the opening was calculated.
Starting with the analysis of the results, it was found that they were satisfactory
since the average percentage difference found between the values obtained through the
LVDT’s and the images was approximately 7.5%. It is worth mentioning that in
previous studies the percentage differences were around 35%.
The average values obtained demonstrate the method’s ability to obtain crack
opening values, with most of the values fluctuating around 5%. Tables 1, 2, 3 and 4
shows the comparative averages of the crack opening values obtained through the
LVDTs and the images. The general average of variation between the two methods is
also presented, which in this case was approximately 7.4%.
Table 1. Beam 1 and 2 results.

TIME BEAM 1 BEAM 2
(s) LVDT DIC Percent LVDT DIC Percent
(10−3 m) (10−3 m) Difference (%) (10−3 m) (10−3 m) Difference (%)
1300 1.7655 1.9439 −10.1048 2.6933 2.6567 1.3612
1400 2.1242 2.2679 −6.7635 3.0385 2.9807 1.9048
1500 2.4982 2.6567 −6.3439 3.3871 3.3047 2.434
1600 2.8654 2.9159 −1.7596 3.7542 3.6286 3.344
1700 3.2325 3.1103 3.7801 4.1119 3.823 7.0253
1800 3.5963 3.4342 4.5051 4.4636 4.147 7.0936
1900 3.9628 3.7582 5.1631 4.8221 4.5358 5.9369
2000 4.3145 3.8878 9.8887 5.1846 4.9246 5.0157
2100 4.6641 4.3414 6.9183 5.5285 5.1838 6.2362
2200 5.0219 4.6006 8.3892 5.8698 5.443 7.2724
2300 5.393 4.8598 9.8874 6.197 5.7021 7.9861
2400 5.7322 5.1838 9.5681 6.5228 5.9613 8.6073
2500 6.0772 5.3134 12.5686 6.8545 6.2853 8.3041
2600 6.4204 5.6373 12.1964 7.1893 6.4797 9.8696

BEAM 1 BEAM 2
Mean 7.7026 5.8851
Standard Deviation (SD) 3.2090 2.6861
Mean + SD 10.9116 8.5712
Mean − SD 4.4937 3.1990
Variation 7.7026 5.8851
When performing an analysis of the results obtained, a certain homogeneity was

found in the values found. Initially, it was observed that in the pre-cracking period, that
is, corresponding to the time interval that goes from 0 to 1300 s, the image analysis
demonstrated a certain limitation. This fact is easily explained due to the high level of
sensitivity of the LVDT, with small displacements being captured that are in the
thousandths of a millimetres, which justifies the distortion found in the comparison of
the results between the image and the LVDT.

TIME BEAM 3 BEAM 4
(s) LVDT DIC Percent LVDT DIC Percent
(10−3 m) (10−3 m) Difference (%) (10−3 m) (10−3 m) Difference (%)
1300 2.8159 2.6567 5.6535 1.6794 1.8791 −11.8890
1400 3.2284 3.0455 5.6664 2.0838 2.2031 −5.7249
1500 3.5939 3.3047 8.0473 2.4338 2.6567 −9.1586
1600 3.9712 3.6286 8.6257 2.7894 3.0455 −9.1786
1700 4.3371 3.8878 10.3596 3.1424 3.2399 −3.1027
1800 4.6828 4.2118 10.0571 3.5044 3.4990 0.1542
1900 5.0356 4.5358 9.9245 3.8468 3.8230 0.6177
2000 5.3773 4.7302 12.0348 4.2150 4.1470 1.6120
2100 5.7266 4.9894 12.8744 4.5769 4.7950 −4.7656
2200 6.0834 5.2486 13.7238 4.9182 4.8598 1.1870
2300 6.4259 5.5078 14.2880 5.2925 5.1838 2.0550
2400 6.7648 5.7021 15.7082 5.6654 5.3134 6.2137
2500 7.1004 6.0261 15.1300 6.0345 5.7021 5.5071
2600 7.4279 6.2853 15.3824 6.4108 6.0261 6.0008

BEAM 1 BEAM 2
Mean 11.2483 4.7976
Standard Deviation (SD) 3.4443 3.5726
Mean + SD 14.6926 8.3703
Mean - SD 7.8039 1.2250
Variation 11.2483 4.7976
5 Conclusions
The application of the digital image correlation (DIC) method proposed by Sutton et al.
(1983), the main focus of this work, was able to quantify and monitor the development
of the crack formation process in SFRC beams under loading and to quantify their
opening. Comparing the values obtained with the images, with values obtained uti-
lizing usual reference devices, such as displacement transducers, an average error of
around 7,4% was obtained, which represents a good approximation for using the
method in the context of Engineering.
The standardization of the experimental process was carried out to allow repro-
ducibility without the need to perform the tests of all parts at once. In this standard-
ization, we can mention the positioning of the LVDT’s in the image plane to avoid
reading errors during the correlation process; the positioning of the camera, being at a
fixed distance from the specimen; the use of the same test material (LVDT’s, Spider,
camera, LEDs, cables), to avoid any measurement errors.
The technology developed in the study is promising because it is a non-destructive

and non-invasive test, and can become suitable for monitoring structures already
implemented, detecting cracks throughout the useful life of the structural element.
References
1. Bentur, A., Mindess, S.: Fibre Reinforced Cementitious Composites. Crc Press, London
(2007)
2. FÉDÉRATION INTERNATIONALE DU BÉTON–FIB. Fib model Code for Concrete
Structures 2010. p. 402, Switzerland (2013)
3. EUROPEAN COMMITTEE FOR STANDARDIZATION. EN 14651: Test method for
metallic fiber-reinforced concrete – Measuring the flexural tensile strength (limit of
proportionality (LOP), residual), p. 15. CEN, London (2007)
4. Yoneyama, S., Ogawa, T., Kobayashi, Y.: Evaluating mixed-mode stress intensity factors
from full-field displacement obtained by optical methods. Eng. Fract. Mech. 74, 1399–1412
(2007)
5. Yoneyama, S., Murasawa, G.: Digital image correlation, in experimental mechanics.
Encyclopedia of Life Support Systems (EOLSS) (2009)
6. McComirck, N., Lord, J.: Digital image correlation. Mater Today, pp. 52–54 (2010)
7. McNeill, S.R., Peters, W.H., Sutton, M.A.: Estimation of stress intensity factor by digital
image correlation. Eng. Fract. Mech. 28, 101–112 (1987)
8. Chu, T.C., Ranson, W.F., Sutton, M.A.: Applications of digital-image-correlation techniques
to experimental mechanics. Exp. Mech. 25, 232–244 (1985)
9. Kuntz, M., Jolin, M., Bastien, J., Perez, F., Hild, F.: Digital image correlation analysis of
crack behavior in a reinforced concrete beam during a load test. Can. J. Civ. Eng. 33(11),
1418–1425 (2006)
10. Pan, B.: Digital image correlation for surface deformation measurement: historical
developments, recent advances and future goals. Meas. Sci. Technol. 29(8), 082001 (2018)
11. Peters, W.H., Ranson, W.F.: Digital imaging techniques in experimental stress analysis. Opt.
Eng. 21, 427–432 (1982)
12. Pan, B., Qian, K.M., Xie, H.M., Asundi, A.: Two-dimensional digital image correlation for
in-plane displacement and strain measurement: a review. Meas. Sci. Technol. 20, 062001
(2009)
Effect of Distribution and Orientation of Fibers
on the Post-cracking Behavior of Steel Fiber
Reinforced Self-compacting Concrete in Small
Thickness Elements
Néstor Fabián Acosta Medina1(&), Rodrigo de Melo Lameiras2,

Ana Carolina Parapinski dos Santos1, and Fábio Luiz Willrich3
1
Federal University of Latin America Integration, UNILA,
Foz do Iguaçu, Brazil
nestor.acosta10@gmail.com
2
University of Brasília, UnB, Brasília, Brazil
3
Laboratório de Tecnologia do Concreto, Itaipu Binacional,
Foz do Iguaçu, Brazil
Abstract. In this work, an experimental investigation focused on the distri-

bution and orientation of fibers on the post-cracking behavior of Steel Fiber
Reinforced Self-Compacting Concrete (SFRSCC) to cast structural small
thickness elements was assessed. To achieve this purpose, two SFRSCC panels
with 45 and 60 mm of thickness were cast from their center point. From each
panel, cylindrical specimens were extracted and notched either parallel or per-
pendicular to the SFRSCC flow direction. The post-cracking behavior was
determined by means of the Modified Splitting Tensile Test. The fiber distri-
bution was evaluated by counting the number of effective fibers crossing the
fractured surfaces. Moreover, the orientation of the fiber was verified using
X-ray method. Notched specimens loaded in the parallel direction of the
SFRSCC flux lines presented higher post-cracking strength when compared with
notched specimens loaded in the perpendicular direction. Likewise, it was also
determined that smaller thickness of the structural element, represents greater
residual stresses and energy absorption, in consequence of the wall effect.
keywords: Steel Fiber Reinforced Self-Compacting Concrete Post-cracking

behavior Thickness Fiber distribution
1 Introduction
Currently, Fiber Reinforced Concrete (FRC) is mainly applied to industrial floors

[1, 2], prefabricated elements and in tunnels [3, 4]. When used in large structural
elements, FRC can contribute cost reduction, since the labour required to place the
conventional reinforcement is reduced (because the density of the reinforcement is
lower when compared to conventional reinforced concrete) or even eliminated if the
steel fibers replace all conventional reinforcement [5].
The advantages associated with the addition of steel fibers in concrete mixtures as
mentioned above, can be enhanced by the use of a concrete with self-compacting
parameters [6], resulting in steel fiber reinforced self-compacting concrete (SFRSCC).
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_26
280 N. F. A. Medina et al.
SFRSCC can lead to a decrease or even eliminate the need for conventional
reinforcement used in a structure. In addition, SFRSCC is easy to pour, increasing
efficiency in the construction process [7]. Considering that, conventional concrete
presents a brittle behavior when pulled; the addition of fibers allows the modification of
this behavior, benefiting in two aspects. First, in the post-cracking resistance, since they
act as a means of transferring stresses and loads between the cracks; and the second in
the increase in the toughness of the material, providing energy absorption mechanisms,
which are related to the processes of detachment and pullout of fibers that make the
bridge through the cracks [8, 9].
Although there are several fibers of different natures, steel fibers are the most used
as a reinforcement element due to the excellent compatibility between steel and con-
crete, and because they have a high modulus of elasticity. After the crack appears, the
fibers allow the transfer of loads from the cementitious matrix to the fibers, in such a
way as to provide increases in the strength of the concrete [10].
As the fiber reinforcement mechanisms are activated mainly after the cracking of
the concrete matrix, the fibers have little influence on the behavior of the non-cracked
elements. Thus, the resistance to compression and traction of the FRC is related to the
strength of the matrix and not by the influence of fibers. In this context, as the residual
tensile strength (post-cracking) represents an important design parameter for FRC
structures, it effectively constitutes the mechanical property most influenced by fiber
reinforcement [11].
The post-cracking behavior of the FRC is influenced by the dispersion and ori-
entation of the fibers. The dispersion and orientation of the fibers in the hardened state
are the product of a series of steps that the steel fiber reinforced concrete (SFRC) passes
from the mixture to the hardened state. Namely, the main factors that indirectly
influence the dispersion and orientation of the fibers are: the properties of the fresh state
after mixing, the conditions of release in the form, the flow characteristics, the vibration
and the wall effect produced by the form [12, 13].
2 Experimental Study
2.1 Materials and Mixture Proportions
The constituent materials used in the composition of the SFRSCC were: Portland
cement CP V (C), water (W), superplasticizer of third generation (SP) based on
polycarboxylates (GLENIUM 51), limestone filler, fine river sand and coarse sand,
coarse aggregate and hooked end steel fibers (length, lf, of 50 mm; diameter, df, of
0.55 mm; aspect ratio, lf/df, of 67 and a yield stress greater than 1100 MPa). The
adopted mix proportions are shown in Table 1, where W/C is the water/cement ratio.
The content of steel fibers in all SFRSCC is kept constant and equal to 40 kg/m3.
The test program was conducted on samples whose mix design has followed the
recommendation presented in Melo [14] and Gomes [15]. Firstly, the proportions of the
constituent materials of the paste were defined, then the proportions of each aggregate
on the final skeleton were determined, and finally the paste and the solid skeleton were
mixed in different proportions (36, 37, 38 and 40% of paste) until self-compacting
Effect of Distribution and Orientation of Fibers on the Post-cracking Behavior 281
Table 1. Mix proportions of SFRSCC per m3.

Casting Cement Water W/C SP Limestone Fine Course Course
(kg) (kg) (−) (kg) filler (kg) sand sand (kg) aggregate
(kg) (kg)
A 399.95 148.69 0.4 4.79 199,97 471.85 459.46 900
requirements are assured in terms of spread ability and correct flow velocity. The final
mix proportions are shown in Table 1.
To evaluate the properties of SFRSCC in the fresh state, the slump flow test was
performed according to EFNARC recommendations [16].
2.2 Specimens
According to Barnett et al. [17], when casting a slab out from its centre, a better
mechanical behavior is assures in comparison with other casting methods. Therefore,
three SFRSCC panels were produced following the selected direction of casting. The
dimensions of the panels were 1500 x 1500 mm2, where the panels thickness varied
(45, 60 and 75 mm), according to the scheme represented in Fig. 1. The specimens that
were used in this work were the number 4, 5, 8 and 13 for the 45 mm thickness plate,
and the specimens 4, 5, 12 and 13 for the 60 mm thickness plate.
Fig. 1. Schematic representation of the core-extracting specimens (units in millimeters).
The test carried out on specimens extracted from these panels will therefore allow
to take conclusions about the flow induced orientation and dispersion of fibers on the
post-cracking behavior. The influence of fiber dispersion and orientation within the
panel was assessed by means of the Modified Splitting Tensile Test, improved by
Abrishambaf et al. [18] and Lameiras et al. [19]. In Fig. 1, the pale dash line represent
the supposed concrete flow direction.
No specimens were extracted from the vicinity of the edge of the formwork, in
order to reduce the interference of the wall effect from the lateral formwork of the
results. The drilling operation was performed when the panel were in their harden-
mature phase. The hatch cores were extracted from a distance of 300 mm and 600 mm
from its centre, all of them with 200 mm in diameter.
The extracted hatch cores were subjected to a loading coincident with the notched
plane performed on the specimens. In order to obtain information on the influence of
the plate thickness on the fiber orientation, the post-cracking behavior of the specimens
was assessed in two distinct directions. By considering h as the angle between the
notched plane and the direction of the concrete flow, the notch plane is designated
parallel for h = 0° or perpendicular for h = 90°.
According to Abrishambaf et al. [18], in order to localize the specimen’s fracture
surface along the notched plane, a 5 mm deep notches parallel to the loading direction
was executed (see Notch 2 in Fig. 2). Furthermore, two additionally notches in a
V-shaped groove with 45° inclination has been cut at the extremities, in both directions
of the notched pane, as shown in notch 1 of Fig. 2. This V-shaped groove configu-
ration, as reported by Di Prisco et al. [20], induced a stress field corresponding to an
almost pure fracture mode I in the notched plane at the loading. Else, Di Prisco et al.
[20] states that this configuration deviate the compressive stresses field from the not-
ched plane, creating, an uniaxial tensile stress field in the notched plane. Moreover,
following what was implemented by Di Prisco et al. [20], two more 5 mm deep straight
notches at the V groove vertices were executed. This force the crack to open at the
reduced section and move the crack tip away from the load application zones, where
high stress concentrations generally arise (see notch 3 in Fig. 2).
Fig. 2. Sequence of implementation of the notches made in the specimens (units in millimetres;
tf: total height; tr: notched height).
2.3 Test Setup and Procedure

2.3.1 Splitting Tensile Test
The test were carried out under displacement control, through a universal servo-
controlled electro-hydraulic testing machine, with a 1000 kN load cell and closed-loop
configuration. The load was applied by using two steel rollers of 20 mm diameter that
were accommodated into the V-shaped grooves. The test setup is depicted in Fig. 3. In
each specimen five linear variable differential transducers (LVDTs) of 5 mm stroke
were used to measure the crack opening displacement, as shown in Fig. 3(a) and (b).
Three LVDT were applied at the side of the specimen corresponding to bottom side
during casting (corresponding to the side in contact with the metallic formwork), and
the two remaining at the other side of specimen.
Fig. 3. Experimental setup of the specimen (a) detail of the bottom side during casting;
(b) upper side during casting and (c) positions of the LVDT’s (units in millimeters).
The test was conducted using the following displacement rates: 0.06 mm/min up to
the displacement of 2.0 mm; 0.12 mm/min up to 2.5 mm; 0.24 mm/min until the end
of the test. The exact position of each LVDT is schematically represented in Fig. 3(c).
2.3.2 Assessment of Fiber Distribution and Orientation

After testing, the fiber distribution in the specimens was determined. Each of the two
faces of the fracture surface of splitting test was divided in four equal regions (see
Fig. 4a). This stage made possible to quantify the fiber density in the fracture surface.
The fiber distribution was evaluated by counting the number of effective fibers crossing
the fractured surfaces (see Fig. 4b). A fiber was considered effective when it was
broken or when its visible length was, at least, twice the length of the hooked part of
the fiber. However, this procedure allows to draw conclusions about their orientation
and distribution in the structural element.
Furthermore, the great challenge to understand better the distribution and orienta-
tion of fibers within a specimen, is to correlate images using non-destructive methods
with the results obtained from the mechanical test performed. Then, the orientation of
the fibers was verified using X-ray method. In some specimens, a photographic rep-
resentation of the projection of the fibers was obtained.
Fig. 4. Assessment of the fiber density in the fractured zone of specimens. (a) Schematic
representation of the fiber counting regions; (b) overview of fiber counting for splitting test
specimens.

3.1 Splitting Test
The Fig. 5 and Fig. 6 depict the splitting tensile stress (r) versus crack width curves,
where r is determined from the following equation:
2P
r¼ ð1Þ
p tr h
4.0 4.0
Envelope Envelope
3.5 3.5
Average Average
3.0 L.I.95 3.0 L.I.95
L.S.95 L.S.95
2.5 2.5
Stress (MPa)
Stress (MPa)
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5
Crack width (mm) Crack width (mm)
(a) (b)
Fig. 5. Splitting tensile stress versus crack width relationship for the specimens obtained from
the panel with 45 mm thickness, and with loading direction (a) parallel, and (b) perpendicular, to
the SFRSCC flux lines.
4.0 4.0
Envelope Envelope
3.5 3.5
Average Average
3.0 L.I.95 3.0 L.I.95
L.S.95 L.S.95
2.5 2.5
Stress (MPa)
Stress (MPa)
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5
(a) (b)
Fig. 6. Splitting tensile stress versus crack width relationship for the specimens obtained from
the panel with 60 mm thickness, and with loading direction (a) parallel, and (b) perpendicular, to
the SFRSCC flux lines.
where P is the compressive load applied in the specimen, tr is the height of the
SFRSCC cylinder and h is the diameter of the remaining SFRSCC cylinder after the
notches are executed in the specimen, as shown in Fig. 2. The values for tr and h in the
equation were measured in each specimen. The crack width of the abscissa axes of
Fig. 5 and Fig. 6 corresponds to the average of the values measured in the five LVDTs
in the specimen. The average, envelope and Lower Bound (L.B.) and Upper Bound (U.
B.) characteristics curves corresponding to a confidence level equal to 95% are also
presented in Fig. 5 and Fig. 6.
The splitting tensile stress (r) versus crack width responses are almost linear up to
the stress at crack initiation in all specimens. Note that, the peak stress was equal to the
stress at crack initiation. Once the peak load was attained, a strain-softening branch is
verified on specimens. Despite the high dispersion of results observed in specimens,
this behavior is generally high in SFRSCC elements, even with specimens of the same
casting and executing under the same test condition. This is due to the high dependence
on post-cracking behavior in relation to the orientation and distribution of the fibers, as
already shown by Abrishambaf et al. [18].
Furthermore, it is quite evident that the stress at crack initiation, and mainly the
post-cracking tensile strength, were higher in specimens loaded in the direction of the
SFRSCC flux lines than in the orthogonal direction, due to the tendency of fibers to
orientate themselves orthogonally to the SFRSCC flux lines, as demonstrated by
Abrishambaf et al. [18], and verified by Lameiras et al. [19].
Other conclusions can also be drawn from residual stresses parameters (rw*) and
energy absorption (GFw*) during the fracture process, as presented in Table 2 and
Table 3. The results were presented separating the samples according to the casting,
thickness and loading direction according to the flow lines. It is noted that the value
attributed to the subscript “w*”, represents the crack width at which rw and GF are
evaluated.
Table 2. Maximum stress, residual stress and dissipated energy for the specimens and with load
applied parallel to the SFRSCC flux lines.
Specimen Direction rmax r0.3 r0.5 r1.0 r1.5 GF0.3 GF0.5 GF1.0 GF1.5
(MPa) (MPa) (MPa) (MPa) (MPa) (N/mm) (N/mm) (N/mm) (N/mm)
A4–45 mm 0° 2,72 1,49 1,56 1,71 1,41 0,41 0,65 1,28 1,88
A5–45 mm 2,63 2,04 1,95 1,68 1,68 0,54 0,87 1,63 2,33
Avg. 2,68 1,76 1,75 1,69 1,55 0,47 0,76 1,46 2,10
CoV 2,54% 22,12% 15,80% 1,16% 12,16% 19,36% 21,11% 16,99% 15,15%
A4–60 mm 0° 2,16 2,08 1,89 1,61 1,41 0,35 0,59 1,12 1,59
A5–60 mm 2,45 1,42 1,37 1,28 1,26 0,31 0,46 0,82 1,17
Avg. 2,68 1,76 1,75 1,69 1,55 0,47 0,76 1,46 2,10
CoV 2,54% 22,12% 15,80% 1,16% 12,16% 19,36% 21,11% 16,99% 15,15%
Table 3. Maximum stress, residual stress and dissipated energy for the specimens and with load
applied perpendicular to the SFRSCC flux lines.
Specimen Direction rmax r0.3 r0.5 r1.0 r1.5 GF0.3 GF0.5 GF1.0 GF1.5
(MPa) (MPa) (MPa) (MPa) (MPa) (N/mm) (N/mm) (N/mm) (N/mm)
A8–45 mm 90° 3,19 1,54 1,48 1,48 1,42 0,50 0,75 1,37 1,98
A13–45 mm 2,07 0,90 0,95 0,94 0,83 0,29 0,42 0,77 1,10
Avg. 2,63 1,22 1,22 1,21 1,12 0,40 0,59 1,07 1,54
CoV 30,29% 37,16% 30,85% 31,07% 36,79% 38,21% 39,44% 39,86% 40,18%
A12–60 mm 90° 1,99 1,69 1,56 1,04 0,97 0,31 0,52 0,91 1,22
A13–60 mm 2,30 0,82 0,77 0,76 0,83 0,33 0,42 0,64 0,89
Avg. 2,14 1,26 1,16 0,90 0,90 0,32 0,47 0,78 1,05
CoV 10,23% 49,20% 48,17% 21,62% 10,54% 2,91% 15,21% 24,14% 22,39%
From the results of Table 2 and Table 3, was verified the influence of the load
application in the parallel and perpendicular direction to the flow lines. When the load
was applied parallel to the flux lines, the specimens of the two thicknesses evaluated
presented higher residual tensile stresses and also higher energy absorptions. This when
compared to specimens loaded perpendicular to the flux lines.
Therefore, can be verified that the results obtained through the variation of thick-
ness of the SFRSCC specimens, the average residual tensile stresses and the average
absorbed energies increased, while decreasing thickness. However, it is presented that
the specimens with 45 mm thickness showed greater energy absorption, both parallel
and perpendicular direction to the flux lines.
3.2 Assessment of the Number of Effective Fibers

The preferential alignment of fibers are corroborated by the results of effective fiber
counting at the fractures surface of splitting test, as shown in Table 4 for the specimens
obtained from the plate with two different thickness. The values presented in Table 4
highlights the correlation between the post-cracking response of the specimens and the
effective fiber counting at the fracture plane. Hence, when the values of Table 4 are
examined separately by load direction, it is verified a greater number of effective fibers

in specimens loaded parallel to the SFRSCC flux lines, when compared with the
average number of fibers counted in the specimens with the load applied perpendicu-
larly to the SFRSCC flux lines. Thus, the values determined are consistent with the
highest post-cracking parameters obtained in specimens with the fracture plane parallel
to the flux lines, as shown by Abrishambaf et al. [18] and Di Prisco et al. [20].
Table 4. Fiber counting at fractured surface of specimens.

Thickness Loading direction Specimen Average number of effective fibers (fibers/cm2)
(mm) Top Bottom Total Per Top Bottom
thickness
45 Parallel to flow A4–45 1,06 1,45 1,25 1,49 1.33 1.41
A5–45 2,14 1,32 1,73
Perpendicular to A8–45 1,32 1,98 1,65 1,25
flow A13–45 0,81 0,89 0,85
60 Parallel to flow A4–60 1,94 1,83 1,88 1,42 1.18 1.55
A5–60 0,78 1,11 0,95
Perpendicular to A12–60 1,34 2,31 1,82 1,31
flow A13–60 0,65 0,95 0,80
Furthermore, it is verified the existence of segregation when compared the values in

Table 4 according to its position (top and bottom) in the plate. For the 45 mm thickness
plate, 5.4% more fibers per square centimeter can be observed when compared to the
top region. In the same way, for the 60 mm thickness plate, 31.5% more fibers were
found in the bottom side when compared to the top side.
3.3 Visualization of Fiber Distribution and Orientation

To view the fiber orientation, were selected for the paper only the specimen corre-
sponding to the position 5 (see Fig. 1), with load applied in the parallel direction of the
SFRSCC flow, and the specimen corresponding to the position 13 (see Fig. 1), with
load applied in the perpendicular direction of the SFRSCC flow.
In Fig. 7, can be observed a tendency of the fibers to align perpendicularly to
SFRSCC flux lines. The specimen with loading application in the parallel SFRSCC
flux lines presented a larger amount of fibers crossing the fracture plane. This verifi-
cation are in accordance with the values of maximum residual stresses (rmax) and
energy absorption (GF), obtained in Table 2 and Table 3. In addition, it was verified in
the same direction of the loading application, a greater horizontal distribution of the
fibers, matching fewer fibers crossing the fracture section and lower values of maxi-
mum residual stresses and energy absorption, according to the values obtained in
Table 3.
Load Flux Load Flux

application direction application direction
(a) (b)
Fig. 7. X-ray images of specimens extracted from the 45 mm thickness plate: (a) A5 specimen
and (b) A13 specimen.
4 Conclusions
• For the materials and casting condition used in the investigation, it was found that
fibers have a tendency to be oriented perpendicularly to the concrete flow, namely,
the radial flow; as already noted by some researchers.
• The specimens presented better behaviors in the post-cracking strength when loaded
in the direction parallel to the flux lines of the SFRSCC, when compared with
specimens loaded in the perpendicular direction. Likewise, it was also found that
smaller thickness of the structural element, represents greater residual stresses and
energy absorption, in consequence of the wall effect.
• The X-ray tests carried out on the specimens proposed in the study, allowed to
verify this preferential tendency of the fibers in relation to the concrete flow.
5 Acknowledgments
The authors wish to express their thanks for the financial support of Fundação Araucária.
The first author also acknowledge the CNPq and the Itaipu Binacional for the support.
References
1. Frazier, P.: Steel fibrous concrete for airport pavement applications. Technical Report S-74-
12. US Army Engineer Waterways Experiment Station, Vicksburg, MS (1974)
2. Suksawang, N., Mirmiran, A., Yohannes, D.: Use of f reinforced concrete for concrete
pavement slab replacement. Final Report, Florida Department of Transportation, Tallahas-
see, Florida (2014)
3. Burgers, R., Walraven, J., Plizzari, G.A., Tiberti, G.: Structural behavior of SFRC tunnel
segments during TBM operations. In: Barták, J., Hrdina, I., Romancov, G., Zlámal, J. (eds.)
Underground Space – the 4th Dimension of Metropolises, Prague, pp. 1461–1467. CRC
Press, Prague (2007)
4. Gettu, R., Barragán, B., Garcia, T., Gonzalo, R., Fernández, C., Oliver, R.: Steel fiber
reinforced concrete for the barcelona metro line 9 tunnel lining. In: International Rilem
Symposium on Fibre Reinforced Concretes, Proceedings, Varenna. RILEM Publications,
Varenna, 2004, pp. 141–156 (2004)
5. Soutsos, M., Le, T., Lampropoulos, A.P.: Flexural performance of fibre reinforced concrete
made with steel and synthetic fibres. Constr. Build. Mater. 36, 704–710 (2012)
6. Okamura, H.: Self-compacting high-performance concrete. Concr. Int. 19(7) (1997)
7. Tojal, T.L.: Contribution to the study of the adhesion of steel bars on self-compacting
concrete reinforced with steel fibers. Master Thesis (Master in Structure) – Graduate
Program in Civil Engineering, Federal University of Alagoas, Maceió (2011). 116 p.
8. Bentur, A., Mindess, S.: Fibre Reinforced Cementitious Composites, 2nd edn. Taylor &
Francis, New York (2007)
9. Rambo, D.A.S.: Self-compacting concrete reinforced with hybrid steel fibers: material and
structural aspects. Master Thesis (Master in Civil Engineering) - Graduate Program in Civil
Engineering, Federal University of Rio de Janeiro, Rio de Janeiro (2012)
10. Velasco, R.V.: Self-compacting concrete reinforced with high volumetric fractions of steel
fibers: rheological, physical, mechanical and thermal properties. PhD Thesis (PhD in Civil
Engineering) - Graduate Program in Civil Engineering, Federal University of Rio de Janeiro,
Rio de Janeiro (2008), 388 p.
11. Di Prisco, M., Plizzari, G., Vandewalle, L.: Fibre reinforced concrete: new design
12. Laranjeira, F.: Design-oriented constitutive model for steel fiber reinforced concrete. Thesis
(PhD) - Department dÉnginyeria de la Construcció, Universitat Politècnica de Catalunya,
Barcelona (2010). 318 p.
classification and prediction. Cem. Concr. Res. 40(2), 226–234 (2010)
14. Melo, K.A.: Contribution to the self-compacting concrete composition with the addition of
limestone filler. Master Thesis (Master in Civil Engineering) - Graduate Program in Civil
Engineering, Federal University of Santa Catarina, Florianópolis, (2005). 183 p.
15. Gomes, P.C.C.: Optimization and characterization of high-strength self-compacting
concrete. Thesis (Ph.D.) - Departament D’Enginyeria de la Construcció, Universitat
Politècnica de Catalunya, Barcelona (2002). 150 p.
16. EFNARC: The European Guidelines for Self-Compacting Concrete. United Kingdom (2005)
17. Barnett, S., Lataste, J.F., Parry, T., Millard, S.G., Soutsos, M.N.: Assessment of fibre
orientation in ultra high performance fibre reinforced concrete and its effect on flexural
strength. Mater. Struct. 43, 1009–1023 (2010)
18. Abrishambaf, A., Barros, J.A.O., Cunha, V.M.C.F.: Relation between fibre distribution and
post-cracking behaviour in steel fibre reinforced self-compacting concrete panels. Cem.
Concr. Res. 51, 57–66 (2013)
19. Lameiras, R.M., Barros, J.A.O., Azenha, M.: Influence of casting condition on the
anisotropy of the fracture properties of Steel Fibre Reinforced Self-Compacting Concrete
(SFRSCC). Cement Concr. Compos. 59, 60–76 (2015)
composites. Mater. Struct. 46(11), 1893–1918 (2013)
Ductility of the Four-Year-Old Steel Fibre
Reinforced Concrete
Jakob Šušteršič(&), Rok Ercegovič, David Polanec, and Andrej Zajc
IRMA Institute for Research in Materials and Applications, Ljubljana, Slovenia

jakob.sustersic@irma.si
Abstract. The paper deals with the results of an experimental investigation into
the ductility of Steel Fibre Reinforced Concrete (SFRC) with a steel fibre
content of between 0,5% and 2% by volume, and that of a comparable concrete
without fibres. These investigations are part of a large-scale research project on
the SFRC that lasted 4 years. At their age of 4 years high compressive strength
has been achieved, with average values in the range of approximately 60 to
115 MPa. These concretes can be divided into two groups: 1st group - concrete
with a maximum nominal grain size (Dmax) of 16 mm with compressive strength
of 90 to 115 MPa and 2nd group - concrete with Dmax = 4 and 8 mm with a
compressive strength of 60 to 80 MPa. The ductile behavior of SFRC was
evaluated by the ductility factor 1/B. 1/B is a parameter for evaluating the
behavior of FRC, which takes into account the entire surface under the load –
CMOD curve. In this way it is possible to evaluate the behavior of the FRC with
only one parameter. Wedge Spit Test (WST) method was used to obtain load –
CMOD curves. In 1st group, a large influence of the fiber aspect ratio (lf/df) on
the increase of 1/B and the lower impact of the aggregate type is visible. In 2nd
group, the influence of polymer, Dmax in addition to fibers on ductile behavior,
can be noticed.
Keywords: Steel Fibre Reinforced Concrete Ductility Ductility factor 1/B

Absorption energy Fiber aspect ratio Wedge split test
1 Introduction
As concrete ages, strength increases, but also the brittleness. By adding fibres, the
toughness and ductility of concrete can be increased. In research, we are most inter-
ested in how concrete behaves at older ages, when the hydration of cement and the
formation of cement stone are almost completely completed. Cement stone binds
individual aggregate particles and fibres into a monolithic structure (but heterogeneous
in composition). This increases the strength of the composite as well as the efficiency of
the added fibres, thereby increasing the toughness and ductility of the composite. Thus,
the ductility of SFRC at high strengths is investigated. Based on these properties, we
estimated which SFRC at high strengths still exhibit ductile behaviour. Concretes with
a low value of w/c ratio have a higher strength than the strength of normal concrete
(compressive strength > 60 MPa). The behaviour of such concrete during fracture can
be “explosive”, leading to rapid demolition of the structural element. The crack
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_27
Ductility of the Four-Year-Old Steel FRC 291
propagation in concrete can be very fast. When fibres are added in concrete, they
prevent rapid crack propagation and concentration into one devastating crack [8–10].
In practice, we tend to use fibres that will least affect the workability of concrete. This
means that we want to work with fibres with as little aspect ratio (lf/df) as possible. The
smaller the lf/df, the easier the fibres are to mix in the fresh concrete mass and the easier it is
to place the SFRC. Most commonly used are steel fibres with an aspect ratio lf/df = 50 to
75. In our study, we used steel fibres with lf/df = 66 (lf = 33 mm, df = 0.50 mm). The
other types of steel fibres used in our study are more effective in hardened concrete but less
used in practice. Their aspect ratio is lf/df = 87 (lf = 33 mm, df = 0,38 mm).
The added quantity of fibres has a major impact on the workability, and especially
on the SFRC price. From a technological point of view, steel fibres of up to 2% by
volume can be used in practice, which is approximately 160 kg/m3 of placed concrete.
SFRC with a steel fibre content of approximately 0,5 to 1,0% by volume (40 to
80 kg/m3) are mostly prepared in our construction practice [11].
Due to the practical interest in our study, we have prepared SFRC with the
aforementioned steel fibres in quantities of 0,5 to 2,0% by volume in various modi-
fications of the mix-proportion.
2 Experimental Details
2.1 Concrete Mix-Proportions

Seven subgroups of the mix-proportions listed in Table 1 are investigated.
Basic principles for choosing the concrete mix-proportions from Table 1:
• The values of water/binder (w/b) ratio are low, in order to obtain the highest quality
cement stone (its structure and adhesion with the aggregate particles, especially
fibres).
• The adhesion of cement stone and fibre is increased by the addition of SF or SBL
(polymer).
• Two characteristic values of the steel fibre aspect ratio were selected and used:
– lf/df = 66 - (lf = 33 mm, df = 0.5 mm),
– lf/df = 87 - (lf = 33 mm, df = 0.38 mm).
• The amount of fibres varies from 0,5 to 2,0% by volume, that is, from the most used
amount in practice (approximately 40 kg/m3 of placed concrete) to the amount of
fibre that can still be manipulated by conventional technological methods of
preparation and placement of concrete (160 kg/m3 of placed concrete). For some
subgroups, concrete mixtures without fibres have also been prepared to evaluate the
effect of added fibres on individual properties and behaviour under different loads.
• Superplasticizer (SPL) improves the workability of concrete, which reduces the
amount of water required, thereby reducing the value of w/b ratio. At a constant
value of w/b ratio, the use of superplasticizer also reduces the amount of cement,
which favourably affects the rheology of the concrete.
292 J. Šušteršič et al.
Table 1. Mix-proportions (per 1 m3 of concrete mixture).

Designation w/b CEM* SF** Steel Fibers SPL*** AE**** SBL+ AGG#
of concrete ratio l/d amount Dmax
– (kg) (% _ % by (% of (% of (% of (mm)
of volume CEM) CEM) CEM)
CEM)
I/0 0.35 320 9.4 – – 2 0.2 – 16
I/0.5 0.35 330 9.4 66 0.5 2 0.15 – 16
I/1 0.35 348 9.4 66 1.0 2 0.3 – 16
I/1.5 0.35 375 9.4 66 1.5 2 0.2 – 16
I/2 0.35 393 9.4 66 2.0 2 0.2 – 16
II/0 0.35 439 9.4 – – 2 0.07 – 16
II/1 0.35 457 9.4 66 1.0 2 0.15 – 16
II/1.5 0.35 466 9.4 66 1.5 2 0.15 – 16
III/0.5 0.35 330 9.4 87 0.5 2 0.15 – 16
III/1.5 0.35 375 9.4 87 1.5 2 0.2 – 16
IV/1 0.35 457 9.4 87 1.0 2 0.15 – 16
IV/1.5 0.35 466 9.4 87 1.5 2 0.15 – 16
V/1 0.30 580 – 66 1.0 3 – – 16
VI/0 0.42 650 – – – – – – 4
VI/0p 0.30 660 – – – – – 10 4
VI/0.5p 0.30 657 – 66 0.5 – – 10 4
VI/1 0.41 645 – 66 1.0 – – – 4
VI/1p 0.30 655 – 66 1.0 – – 10 4
VII/1 0.40 657 – 66 1.0 – – – 8
VIIp/1 0.41 665 – 66 1.0 – – 10 8
* ** *** ****
CEM - cement, SF - silica fume, SPL - superplasticiter, AE - air-entraining admixture,
+
SBL - solid particles of styrene-butadiene copolymer latex, #AGG – aggregate; aggregate types:
fine gravel aggregate + coarse eruptive aggregate in concrete from subgroups I and III; slag
aggregate in concrete from subgroups II and IV; gravel aggregate in concrete from subgroups V,
VI and VII.
• The use of air-entraining admixture (AE) introduces fine closed air bubbles into the
concrete which increase the resistance of the concrete to the freezing/thawing effects
of the concrete without or in the presence of de-icing salt. Introduced air bubbles
can influence the better distribution of added fibres in fresh concrete. The resistance
of concrete to the effects of freezing/thawing is also increased by reducing the value
of w/b ratio or by adding SBL, thereby closing open capillary pores and reducing
their number.
• The distribution of fibres in fresh concrete and their efficiency in hardened concrete
may also be affected by the largest aggregate grain (Dmax). At constant fibre length
(lf = 33 mm) the ratio lf/Dmax has the values: 2, 4 and 8.
• The use of high-density aggregate with high strength also enables the preparation of
high strength concrete. Of course, these concretes must have good quality cement
stone with low w/b ratio values.
• Slag aggregate, as a secondary raw material, can replace natural eruptive aggregate.
The pozzolanic properties of the slag can also be counted on, which increases the
adhesion of certain aggregate grains and cement stone, thereby increasing the
compactness of concrete, which in turn increases the strength and other resistance of
these concretes.
When designing the mix-proportions of concrete from subgroups I to IV, the
selected value of w/b ratio was kept constant. Other parameters of the mix-proportions
(especially the amount of cement) were chosen so as to obtain the best workability of
fresh concrete.
Concrete mix-proportions of subgroups VI and VII are designed for constant
workability of fresh concrete and constant ratio: cement mass/aggregate mass = 1/2.
Because of this, the values of w/b ratio have changed so that for polymer modified
concrete with SBL the obtained value of w/b ratio is less than w/b ratio of concrete
without SBL.
2.2 Characteristic Properties of the Materials Used for Concrete
• Steel fibres: Both types of fibres used (lf/df = 66 and 87) are of the same shape,
given in Fig. 1. End hooks allow better anchoring of fibres in concrete, while
adhesion with cement stone increases with increasing quality of cement paste (low
value of w/b ratio and use of admixtures). The fibres with lf/df = 66 are produced
from wire with an average maximum tensile strength Rm = 841 N/mm2, and the
fibres with lf/df = 87 from wire with Rm 2000 N/mm2.
Fig. 1. Characteristic shape of the steel fibres used.
• Cement: Portland cement CEM I 42,5 was used to prepare the concrete from
subgroups I to V, and Portland cement with slag PC II/A-S 42,5 for the concrete
from subgroups VI and VII.
• Silica fume: fine amorphous dust in the form of balls of average diameter 0,1 lm;
SiO2 content is 94,0%.
• Superplasticizer: modified polycarboxylate; density of SPL is 1,2 kg/dm3.
• Air-entraining admixture: it is made on the basis of vinyl resin; density of AE is

1,1 kg/dm3.
• Styrene-butadiene copolymer latex: with the following characteristics: solid
particles (dried at 105° C) = 46,0%, Cl content = 1.45%, pH value = 8,3,
density = 1,01 kg/dm3.
• Gravel aggregate: fraction of fine aggregate = 0-4 mm, fractions of coarse
aggregate = 4-8 and 8-16 mm.
• Eruptive aggregate: fractions of crushed amphibolite 4-8, 8-11 and 11-16 mm;
density slightly greater than 2900 kg/m3.
• Slag aggregate: Slag generated in the production of ferrochrome by the carboth-
ermic process. Its structure is similar to the porphyry structure of magmatic rocks.
The aggregate, which is obtained by crushing the slag, is one of the heavy aggre-
gates for concrete because it has a density greater than 3000 kg/m3. For the
preparation of concrete from subgroups II and IV, fractions of 0-4, 4-8 and 8-
16 mm of slag aggregate were used.
2.3 Program of Investigations

The main purpose of adding fibres to concrete is to increase its ductility, toughness and
resistance to crack propagation. Several types of tests were performed. This paper will
only provide some of the results of the wedge spit test (WST). WST is a test method to
perform stable fracture mechanics tests on concrete and concrete-like materials. It was
proposed by Brühwiler and Wittman [1], and Linsbauer and Tschegg [2]. The method
proposed by last authors [2] was used and discussed in previous investigations [3–5] in
order to obtain load – CMOD curves.
When a concrete element is loaded, small individual cracks begin to appear inside
the concrete at a certain load. Those cracks combine into a continual crack, which can
be seen on the surface of concrete element. At this point of the load - CMOD curve, the
slope of the curve increases significantly. Load and CMOD at this point are denoted as
first crack load, first crack CMOD, respectively. There are difficulties relating to precise
determination of the location of the first crack. ASTM C 1018 defines first crack as the
point on the load – deflection curve at which the form of the curve first becomes non-
linear. Determination of the point of first crack has been proposed [6, 7] as the point at
which the slope of the curve departs from linearity by more than 5% and lasts for an
interval of more than 0,01 mm.
At our Institute (IRMA), computer program has been developed, which works in
graphical form, for automatically drawing load-CMOD curves, for calculation of
parameters for evaluation of concrete behaviour, and for determination of the point of
first crack (FC) [4].
At the moment when the point FC is reached, the crack width begins to propagate
with further loading. From the point FC, the fracture zone of the concrete begins to
form. In the fracture zone, all further fracture processes proceed until the final sepa-
ration of the test specimen. Absorbed energy of fracture zone (WFZ) represents the
absorbed energy required to completely separate the test specimen. Absorbed energy is
expressed as the product of load and crack mouth opening displacement (CMOD).
When the load at first crack (FFC) and the absorbed energy of fracture zone (WFZ)
are determined from the load-CMOD diagram, the characteristic CMOD of fracture
zone (DCMOD) can be calculated according to Eq. (1):
WFZ
DCMOD ¼ ð1Þ
FFC
DCMOD is a property of a concrete, on the basis of which we can estimate the

ductility of a concrete with respect to CMOD at the first crack (CMODFC). Concrete is
ductile if DCMOD > CMODFC and vice versa concrete is brittle if DCMOD < CMODFC.
Ductility can be expressed by the dimensionless ductility factor 1/B, which could
be considered as a favourable parameter for the evaluation of the behaviour of fibre
reinforced concrete (FRC). The ductility factor 1/B is expressed by Eq. (2):
DCMOD
1=B ¼ ð2Þ
CMODFC
The ductility factor 1/B is higher and thus the ductility of the tested FRC, if the
characteristic crack mouth opening displacement of the fracture zone (DCMOD) is larger,
which means that the FRC must achieve as much large absorbed energy of the fracture
zone (WFZ) at the corresponding load at the first crack (FFC).
From prisms with dimensions of 10 10 40 cm, at the age of concrete 4 years,
test bodies were cut - cubes with a 10 cm edge with a notch in the middle to a depth of
5 cm and a necessary slot for wedge, which is pushed vertically during the experiment.
This widens the notch. The crack opening velocity was 0,04 mm/min.
The ductility of concrete at the age of 4 years was tested when high compressive
strengths were reached, with average values ranging from approximately 60 to over
115 MPa (Fig. 2). Thus, all the subgroups in Table 1 can be joined into two groups:
1st group: concrete with Dmax = 16 mm (concrete from subgroups I to V) with
compressive strength of 90 to 115 MPa and
2nd group: concretes with Dmax = 8 mm (concretes from subgroup VII) and con-
cretes with Dmax = 4 mm (concretes from subgroup VI) with compressive strengths of
60 to 80 MPa.
At such high strengths, the question first arises about the form of behaviour of these
concretes in loading, in our case with wedge splitting.
As an example, the characteristic load-CMOD diagrams of individual concrete are
visible from Fig. 3, in which concrete diagrams from subgroups I and III are given.
Figure 3 shows the influence of aspect ratio of steel fibre on the ductile behaviour
of SFRC. The visual evaluation of the ductile behaviour of concrete from the diagrams
in Fig. 3 may very well be replaced by a numerical evaluation using a ductility
factor1/B (Fig. 4 and 5).
Fig. 2. Average compressive strength of 4 years old concretes.
2,2 load
(kN)
2,0
1,8
1,6
1,4
1,2
1,0
0,8
III/1,5
0,6 I/2
I/1,5
0,4 III/0,5
I/0 I/1
0,2
I/0,5
0,0
0,0 0,5 1,0 1,5 2,0
CMOD (mm)
Fig. 3. Characteristic diagrams load - CMOD of concrete from subgroups I and III.
First, a relative comparison can be made from the dependence of the ductility factor
1/B on the amount of steel fibres in the individual concrete from both groups.
60
50
IV
40
ductility factor 1/B
30
III
20
II
10
I
0
0 0.5 1 1.5
steel fibres (% by volume)
Fig. 4. The dependence of the ductility factor 1/B on the amount of steel fibres in concretes
from subgroups I to IV.
30
VIIp
25
VII
20
15
VI VIp
10
0
0 0.5 1
steel fibres (% by volume)
Fig. 5. The dependence of the ductility factor 1/B on the amount of steel fibres in concretes
from subgroups VI and VII.
In 1st group, a large influence of the fibre aspect ratio (lf/df) on the increase of 1/B
and a smaller influence of the type of aggregate are seen (Fig. 4). The ductility factor
1/B of SFRC with 1,5% by volume of steel fibres with lf/df = 87 is greater than 1/B of
SFRC with 2,0% by volume of steel fibres with lf/df = 66. In concrete without fibres
and with slag, the total absorbed energy is higher than in concrete without fibres and
with an eruptive aggregate. However, in the case of concrete with slag, the absorbed
energy of the elastic region is higher, which is why the ductility factor 1/B for this
concrete is reduced compared to the concrete with eruptive aggregate. Due to the large
amount of absorbed energy of the elastic area of the concrete with the slag aggregate,
the absorbed energy of the fracture zone WFZ is relatively smaller but still higher than
the absorbed energy of the fracture zone of concrete with eruptive aggregate. By adding
fibres to the concrete with slag aggregate, especially fibres with lf/df = 87, the absorbed
energy of the fracture zone is greatly increased. Adhesion increases and thus the
efficiency of fibres. The behaviour of these SFRC is similar to the behaviour of high
performance SFRC.
Similar observations are valid for concretes from the second group. Only a polymer
(SBL) without fibres in concrete with Dmax = 4 mm increases the ductile behaviour, as
can be seen from the diagram in Fig. 5 and the increase in the ductility factor. The
addition of fibres increases the absorbed energy and 1/B, most of all in SFRC with 1%
by volume of steel fibres and higher in absolute values in SFRC with Dmax = 4 mm
than in SFRC with Dmax = 8 mm. But the ductility factor 1/B of a polymer-modified
SFRC with 1% by volume is less than 1/B of SFRC with 1% by volume without
polymer although the total absorbed energy of the first SFRC is greater than the second.
1/B of polymer modified SFRC with 1% by volume is reduced due to the increase in
the load at the first crack FFC and the expansion of the CMODFC. The added polymer
(SBL) affects the increase of the absorbed energy of SFRC, so that the behaviour of
SFRC VI/1p and SFRC VII/1p is similar to the behaviour of high performance SFRC.
The correlations of fibre quantity and 1/B are similar to the correlations of fibre
quantity and parameters DCMOD and WFZ with which 1/B is calculated.
60
50
40
30
20
10
0
IV/1.5
VII/1p
II/0
II/1
VI/0p
VI/1p
I/0
I/1
I/2
VII/1
I/0.5
I/1.5
III/0.5
III/1.5
IV/1
VI/0
VI/1
II/1.5
V/1
VI/0.5p
type of concrete
Fig. 6. Average values of ductility factor 1/B of the concrete from the 1st and 2nd group.
A comparison of the average values of 1/B of all concrete, given in Fig. 6, indicates
a significant increase in 1/B of SFRC with slag and 1,5% by volume of steel fibres with
lf/df = 87. An increase of 1/B of SFRC with Dmax = 4 and 8 mm and 1% by volume of
steel fibres without or with SBL is also observed. This indicates that by increasing ratio
of the fibre length and the largest grain diameter of the aggregate lf/Dmax, the ductility
factor 1/B also increases. Both types of fibres used are 32 mm long, so by reducing the
Dmax from 16 to 8 and 4 mm, the lf/Dmax ratio increases from 2 to 4 and 8.
4 Conclusions
Based on the results of four-year-old SFRCs tested by the WST method and having
high compressive strengths, the following main conclusions can be drawn:
• ductility factor 1/B of SFRC increases if the absorbed energy of the fracture zone
WFZ is increased sufficiently with respect to the increase in the absorbed energy of
the elastic area or area up to the first crack WFC;
• an increase in the ductility factor1/B is achieved by using increased quantities of
efficient steel fibres (fibres with a larger aspect ratio lf/df), increasing the adhesion of
the fibres and the matrix, and increasing the lf/Dmax ratio (ratio of the fibre length
and the largest grain diameter of the aggregate);
• behaviour of SFRC with slag aggregate and especially fibres with a larger aspect
ratio lf/df as well as behaviour of polymer modified SFRC are similar to the
behaviour of high performance SFRC.
References
1. Brühwiler, E., Wittman, F.H.: The Wedge splitting test, a new method of performing stable
fracture mechanics tests. In: Rossmanith, H.P. (ed.) Fracture and Damage of Concrete and
Rock, pp. 117 – 125. Pergamon Press (1988)
2. Linsbauer, H., Tschegg, E.K.: ‘Die Bestimmung der Bruchenegie an Würfelproben’
(Fracture Energy Determination of Concrete with Cube-Shaped Specimens). Zement und
Beton, 31(1), 38 – 40 (1986)
3. Šušteršič, J., Kolenc, M., Zajc, A., Riček, F., Zajc, P.M.: High-performance fibre reinforced
concrete for mine roadway support panels. In: Proceedings, Second CANMET/ACI
International Conference Gramado, RS, Brazil,. SP-186, pp. 101 – 112 (1999)
4. Šušteršič, J., Ukrainczyk, V., Zajc, A., Šajna, A.: Evaluation of Crack Opening Resistance -
8) of SFRC. In: Banthia, N., Sakai, K., Gjørv, O.E.: Concrete Under Severe Conditions. Vol.
2. Vancouver, pp. 1594 – 1601 (2001)
5. Šušteršič, J., Zajc, A., Leskovar, I., Dobnikar, V.: Improvement in the crack opening
resistance of FRC with low content of short fibres. In: Dhir, Ravindra, K. (ed.) Role of
concrete in sustainable development: proceedings of the International Symposium dedicated
to professor Surendra Shah, Northwestern University, USA held on 3–4 September 2003 at
the University of Dundee, Scotland, UK. London: ˝Thomas Telford˝, pp. 167-174 (2003)
6. Chen, L., Mindess, S., Morgan, D.R.: Toughness evaluation of steel fibre reinforced
concrete. In: Proceedings of 3rd Canadian Symposium on Cement and Concrete, Ottawa,
pp. 16 – 29 (1993)
7. Morgan, D.R., Mindess S., Chen L.: ‘Testing and Specifying Toughness for Fibre
Reinforced Concrete and Shotcrete’. Fibre Reinforced Concrete. Modern Developments.
Eds.: N. Banthia and S. Mindess. The University of British Columbia, Vancouver, pp. 29 –
50 (1995)
8. Banthia, N.: Advanced Fiber Reinforced Cement-based Composites for New Generation
Safe, Sustainable and Smart Infrastructure. 25th Slovenian Colloquium on Concrete:
Concrete with Improved Properties. Ljubljana, IRMA, pp. 3–8 (2018)
9. Plizzari, G.: Structural applications of Fibre Reinforced Concrete. In: 25th Slovenian
Colloquium on Concrete: Concrete with Improved Properties. Ljubljana, IRMA, pp. 9–18
(2018)
10. Silfwerbrand, J.: Fibre Concrete – Swedish Research & Experience. 25th Slovenian
Colloquium on Concrete: Concrete with Improved Properties. Ljubljana, IRMA, pp. 19–28
(2018)
11. Šušteršič, J., Zajc, A.: Review of Certain Applications of Fibre Reinforced Concrete in
Slovenia. Fibre Concrete 2015. Prague (2015)
Sensitivity of the Flexural Performance
of Glass and Synthetic FRC to Fibre Dosage
and Water/Cement Ratio
Razan H. Al Marahla and Emilio Garcia-Taengua(&)
School of Civil Engineering, University of Leeds, Leeds, UK

E.Garcia-Taengua@leeds.ac.uk
Abstract. A comparative analysis of the flexural performance of FRC mixes

with either glass or synthetic fibers is presented in this paper. The data used for
such analysis were obtained from an experimental programme which comprised
42 notched prismatic specimens, produced and tested to EN 14651 at the age of
28 days. Different fibre dosages up to 15 kg/m3 were considered in two series of
mixes with water/cement ratios of 0.26 and 0.39, which yielded average com-
pressive strength values of 65 MPa and 50 MPa respectively. A direct corre-
lation between fibre content and residual flexural strength was confirmed.
However, statistically significant differences were observed between the two
fibre types considered. FRC mixes with glass fibre contents up to 5 kg/m3 failed
immediately after the first crack and showed no residual flexural strength. In
general, specimens reinforced with synthetic fibres showed better levels of
residual flexural strength and toughness than their glass fibre counterparts. The
ratio between the residual flexural strength and the limit of proportionality
provides a good illustration of such observations: it was 61% on average for
FRC mixes with synthetic fibers at 10 kg/m3, whilst it was only 33% for FRC
mixes with the same amount of glass fibers.
Keywords: Synthetic fibres Glass fibres Residual flexural strength
1 Introduction
Fibre reinforced concrete (FRC) is well suited for a variety of structural as well as non-
structural applications, as concrete performance in the hardened state is generally
enhanced by the fibres role in bridging cracks, which results in improved crack control
and enhanced properties in the cracked state [1, 2]. For applications where the struc-
tural contribution of fibres is intended to be significant, steel fibres are generally pre-
ferred [3, 4]. Fibres made with other materials have lower tensile strength and elastic
modulus than steel fibres, and they are typically considered for applications where their
main contribution is concerned with restraining the plastic cracks, and shrinkage
cracking control. However, there is no clear-cut separation that restricts non-steel fibres
to non-structural applications, and this, together with current trends favouring materials
with lower carbon footprint, has motivated increasing interest in the mechanical per-
formance of FRC with synthetic or glass fibres.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_28
302 R. H. Al Marahla and E. Garcia-Taengua
Fibres partially counterbalance the brittleness which is intrinsic to concrete as a

material [5, 6], but their positive influence on flexural toughness varies greatly
depending on the type of fibre material, its dimensions, and the fibre dosage. A number
of studies [7–9] have investigated the flexural performance and toughness of FRC
made with different types of fibres, and there is clear consensus that the addition of
increasing fibre contents improves residual flexural strength and toughness.
To characterise, specify and design with FRC, the residual flexural strength
parameters or toughness as obtained from the bending test are used as reference, in
addition to the compressive strength at 28 days. In this study, the flexural response of
different FRC mixes with either glass or synthetic fibres was characterised by testing
notched prismatic specimens under three point bending test conditions, following the
standard EN 14651 [10]. Different fibre dosages were considered. Also, to evaluate the
influence of water-to-cement (w/c) ratio on flexural toughness, and how this modifies
the contribution of fibres to residual flexural strength, two reference concrete mix
designs were considered.
2.1 Variables Considered

The factors considered in this study were: w/c ratio, type of fibre, and fibre content.
Cement type CEM I 52.5 N was used, and the superplasticiser was Sika Viscocrete
25MP. Two different values were considered for the w/c ratio: 0.26 and 0.39, leading to
the definition of two reference mix designs, which were adjusted for a slump value of
120-150 mm so they could incorporate different fibre contents without further
adjustments. These reference mix designs are summarised in Table 1, and were
intended to be representative of a range of specified compressive strengths between 45-
60 MPa approximately.
Table 1. Reference mix designs (kg/m3)

w/c = 0.39 w/c = 0.26
Cement 440 510
Water 175 130
Fine aggregate 825 950
Coarse aggregate (10 mm) 637 580
Coarse aggregate (20 mm) 317 300
Superplasticiser 8 11
The two types of fibres considered in this study were: synthetic fibrillated macro-
fibres and alkali-resistant glass macro-fibres, and they are shown in Fig. 1. The syn-
thetic fibres were high-modulus 54-mm long polymeric fibres and had a tensile strength
of 600 MPa. The glass fibres, on the other hand, were 36 mm long and their tensile
Sensitivity of the Flexural Performance of Glass 303
strength was 1700 MPa. Synthetic fibres were used in dosages of 5, 7.5, and 10 kg/m3,
whilst the dosages considered for the glass fibres were 2.5, 5, and 15 kg/m3.
Fig. 1. Samples of the glass fibres (left) and synthetic fibres (right) used in this study.
2.2 Production of FRC Mixes

Taking as reference the two mix designs presented in Table 1, either glass or synthetic
fibres were incorporated in different dosages, leading to the combinations summarised
in Table 2.
Table 2. Summary of the FRC mixes considered in this study.

Fibre type w/c ratio Fibre content (kg/m3)
(None) 0.26 0.0
0.39 0.0
Glass 0.26 2.5
5.0
15.0
0.39 2.5
5.0
15.0
Synthetic 0.26 5.0
7.5
10.0
0.39 5.0
7.5
10.0
The same mixing sequence was followed in the production of all mixes. In
preparation before the mixing of every batch, the total amount of water to be added was
separated in two buckets: one containing 80% of the water, and the other containing the
mixture of the remaining 20% of the water and the required amount of superplasticiser.
First, cement and all aggregates were all poured into the mixer and dry-mixed for
2 min. After that, 80% of the water was added and mixed with the cement and
aggregates for 3 min. During this time, the fibres were poured gradually into the mixer.
Finally, the remaining 20% of the water with the superplasticiser predispersed in it was
added, and the mixing continued for 4 min. In all cases, a uniform distribution of the
fibres in the mix was observed.
2.3 Characterisation of FRC Mixes

For each of the combinations presented in Table 2, one batch of 70 litres was produced,
and this material was used to cast 3 cubes, 3 cylinders, and 3 prismatic specimens. All
specimens were tested at the age of 28 days. Prior to this, all specimens were kept in a
fog room with a controlled temperature of 20C and a relative humidity of 90%. The
cubes were 100-mm side and were used to determine the compressive strength, whilst
the 150 300 mm cylinders were used to determine the splitting tensile strength
following the standard EN 12390-6:2009 [11].
The prismatic specimens were produced and tested in flexure, according to the
standard EN 14651 [10], and were 150 mm side and 600 mm long. They were all
notched with a notch depth of 25 mm and 5 mm width using a wet sawing machine,
and cured for a minimum of 3 days after sawing according to EN 12390-2 [12] until the
age of testing. For the flexural strength test, a three point bending scheme was used, an
image of which is shown in Fig. 2. The tests were carried by imposing a constant rate
of 0.05 mm/min for the increasing CMOD until the CMOD reached 0.1 mm, after
which the rate was increased to 0.2 mm/min. CMOD values were monitored by means
of LVDTs placed right under the notch, and a K7500 service controller was used for
the data acquisition and control signal.
Fig. 2. Set up for tested specimen

3 Test Results and Discussion

3.1 Compressive and Splitting Tensile Strength
The results of the compressive strength and splitting tensile strength tests are presented
in Table 3. The average compressive strength was 67 MPa and 50.4 MPa for the
reference mixes without fibres and w/c ratios of 0.26 and 0.39 respectively.
Table 3. Compressive and splitting tensile strength results.

Fibre type w/c ratio Fibre content (kg/m3) Compr. strength Split. tensile
(MPa) strength (MPa)
Average Std. dev. Average Std. dev.
(None) 0.26 0.0 67.0 1.0 4.2 0.2
0.39 0.0 50.4 0.6 3.8 0.6
Glass 0.26 2.5 66.6 1.5 3.9 0.6
5.0 66.8 0.6 4.0 0.6
15.0 67.5 0.7 4.0 0.4
0.39 2.5 50.8 1.9 3.6 0.5
5.0 51.3 1.4 3.7 0.8
15.0 52.2 0.8 3.8 0.5
Synthetic 0.26 5.0 67.3 0.9 4.1 0.4
7.5 67.5 0.6 4.1 0.3
10.0 67.6 0.3 4.0 0.3
0.39 5.0 51.3 1.4 3.7 0.4
7.5 52.4 0.8 3.7 0.7
10.0 52.2 0.8 3.6 0.4
The addition of glass or synthetic fibres, at the dosages considered in this study, did
not introduce significant variations in terms of average compressive strength, and the
same can be said in relation to splitting tensile strength. Although fibres have been
reported to sometimes increase the compressive strength of concrete by up to 15%, ACI
544 [13], these results are in agreement with those other studies that reported no
significant improvements in compressive strength due to the presence of fibres [14, 15].
In terms of variability, low fibre contents were observed to increase the standard
deviation of compressive strength values with respect to the reference mixes without
fibres. However, increasing fibre contents were associated with decreasing standard
deviation values, for both types of fibres and w/c ratios.
3.2 Bending Test Results

For each of the FRC mixes as per the combinations listed in Table 2, three prismatic
specimens were tested under flexure to EN 14651 [10], and the corresponding load-
CMOD curves were obtained. From these curves, the equivalent stress corresponding
to the limit of proportionality, fL, and the residual strength parameters fR1, fR2, fR3, and
fR4 (corresponding to CMOD values of 0.5, 1.5, 2.5, and 3.5 mm, respectively) were
obtained. Their average values are given in Table 4.
Table 4. Bending test results: average values, expressed in MPa.

Fibre type Fibre content (kg/m3) w/c ratio fL fR1 fR2 fR3 fR4
Glass 2.5 0.26 5.68 0.00 0.00 0.00 0.00
0.39 5.69 0.00 0.00 0.00 0.00
5 0.26 6.96 0.00 0.00 0.00 0.00
0.39 6.31 0.00 0.00 0.00 0.00
15 0.26 5.41 2.49 2.00 2.00 1.93
0.39 5.67 2.61 1.74 1.61 1.57
Synthetic 5 0.26 7.45 2.66 2.56 2.53 2.42
0.39 6.24 2.12 2.08 2.14 2.18
7.5 0.26 7.48 3.10 2.84 2.79 2.75
0.39 6.79 2.53 2.50 2.44 2.39
10 0.26 6.68 3.17 3.10 3.09 3.08
0.39 7.43 2.78 2.60 2.61 2.58
All specimens produced with FRC mixes containing glass fibres at dosages of 2.5
and 5.0 kg/m3 failed without exhibiting any residual load-bearing capacity in flexure,
regardless of the w/c ratio. A softening response was observed in specimens where the
glass fibres dosage was 15 kg/m3. A graphical comparison of the fL values corre-
sponding to the FRC mixes with glass fibres, and the stress-CMOD curves corre-
sponding to specimens with a glass fibre dosage of 15 kg/m3 are shown in Fig. 3.
Fig. 3. Flexural test results for glass FRC mixes: limit of proportionality values (left) and stress-
CMOD curves for 15 kg/m3 fibre content.
On the other hand, specimens with synthetic fibres presented comparatively better
performance than their glass fibres counterparts, as residual flexural capacity was
developed for all fibre dosages and w/c ratios considered. Figure 4 shows the average
stress-CMOD curves for the different dosages of synthetic fibres and w/c ratios con-
sidered. Furthermore, comparing the results corresponding to those mixes where fibres
were added at the maximum dosages considered in this study (10 kg/m3 for synthetic
fibres, 15 kg/m3 for glass fibres), synthetic fibres did better than glass fibres in terms of
the level of residual flexural strength achieved. The incorporation of synthetic fibres at
a dosage of 10 kg/m3 led to fR1/fL ratios between 0.42 and 0.61, for w/c ratios of 0.39
and 0.26 respectively, as opposed to 0.26 and 0.33, corresponding to specimens with
15 kg/m3 of glass fibres.
Fig. 4. Flexural test results for glass FRC mixes with w/c = 0.39 (left) and w/c = 0.26 (right).
3.3 Residual Flexural Strength

In order to better quantify the differences in the flexural response of FRC introduced by
changes in the variables considered in this study, a regression analysis was done and
equations for the residual flexural strength parameters were obtained. In the regression
analysis, the interactions between fibre content, fibre type and w/c ratio were also
considered, with the purpose of determining whether these synergies were statistically
significant. The equations obtained are presented in Table 5, where Cf is the fibre
content in kg/m3. The R-squared values ranged between 0.91 and 0.97, which indicated
a very accurate fit with the experimental results.
Unsurprisingly, fibre content and the fibre type were found to have a statistically
significant effect on all residual flexural strength parameters. In addition to that, the
interaction between fibre type and fibre content was found to be statistically significant.
That is, the effect that a certain increase in the fibre content had on residual flexural
strength was modified depending on the type of fibre considered. Regarding the effect
of the w/c ratio, it was found that it had a statistically significant effect on fR1 and fR2.
However, the regression analysis showed that varying the w/c ratio did not introduce
statistically significant variations in fR3 and fR4.
Table 5. Regression equations obtained for the residual flexural strength parameters.
FRC with glass fibres FRC with synthetic fibres
Lnðf R1 Þ ¼ 3:845 þ 0:380Cf 3:31w=c Lnðf R1 Þ ¼ 1:733 þ 0:044Cf 3:31w=c
Lnðf R2 Þ ¼ 3:557 þ 0:314Cf 1:89w=c Lnðf R2 Þ ¼ 1:258 þ 0:041Cf 1:89w=c
Lnðf R3 Þ ¼ 4:097 þ 0:305Cf Lnðf R3 Þ ¼ 0:647 þ 0:040Cf
Lnðf R4 Þ ¼ 4:650 þ 0:335Cf Lnðf R4 Þ ¼ 0:628 þ 0:041Cf
The equations presented in Table 5 were useful in visualising the sensitivity of the
different residual flexural strength parameters to changes in w/c ratio, fibre type, and
fibre content.
Figure 5 shows the contour plots for fR1 values corresponding to mixes reinforced
with either glass fibres (left) or synthetic fibres (right), as a function of w/c ratio and the
fibre content. For the mixes with glass fibres, a theoretical minimum for the fibre
content could be identified in order for the material to present residual flexural capacity.
This minimum glass fibre content was in the range of 10 to 12 kg/m3 and was slightly
dependent on the w/c ratio.
Fig. 5. Contour plots for fR1 (in MPa) as a function of w/c ratio, type of fibre, and fibre content.
The effect of w/c ratio was much more marked in mixes with synthetic fibres. As
the contour plot in Fig. 5 (right) shows, reducing the w/c ratio could lead to
improvements in fR1 comparable to those which would be achieved by increasing the
amount of synthetic fibres in the mix. Contour plots for fR2 are not shown in this paper
because they were very similar to those obtained for fR1 and led to similar conclusions.
Figure 6 shows the trend followed by fR3 and fR4 values with respect to increasing
fibre contents, for the two types of fibres considered. As mentioned before, fR3 and fR4
were not sensitive to variations in the w/c ratio within the range considered in this
study. It can be seen that, for the mixes with synthetic fibres, the relationship between
fR3 or fR4 values and the fibre content was practically linear. Also, in all cases con-
sidered in this study there was almost no difference between fR3 and fR4 values.
Fig. 6. Regression lines for fR3 and fR4 as a function of fibre content.
3.4 Flexural Toughness

In order to obtain an indication of the flexural toughness from the bending test results,
the areas under the stress-CMOD curves up to a CMOD value of 3.5 mm were cal-
culated. The average toughness values obtained for each case considered in this study
are summarised in Table 6. Values for glass fibre contents lower than 15 kg/m3 are not
presented, as those cases exhibited a brittle failure.
Table 6. Toughness (mm.MPa)

Fibre type Fibre content (kg/m3) w/c = 0.39 w/c = 0.26
Glass 15.0 6.95 8.03
Synthetic 5.0 7.78 9.52
7.5 9.26 10.57
10.0 10.52 11.51
The toughness values corresponding to synthetic FRC mixes were analysed by

means of an analysis of variance, and this showed that the interaction between the fibre
content and the w/c ratio had a statistically significant effect. The response surface
shown in Fig. 7 was obtained, and the significance of such interaction is clearly
noticeable: the trend followed by toughness values with respect to the fibre content
changes depending on the w/c ratio, and vice versa.
Reducing the w/c ratio from 0.39 to 0.26 increases toughness values by 22% if a
fibre content of 5 kg/m3 is considered; however, this increase is 14% or 9.5% if the
fibre content is 7.5 kg/m3 or 10 kg/m3, respectively. In consequence, it was concluded
that mixes with higher synthetic fibre contents were less sensitive to changes in the w/c
ratio, in terms of their flexural toughness. Conversely, increasing the synthetic fibre
content led to increased toughness values, but the magnitude of such an increase was
found to depend on the w/c ratio. For instance, for a w/c ratio of 0.39, toughness values
were increased by 35% as a result of increasing the fibre content from 5 kg/m3 to
Fig. 7. Contour plot for toughness values (area under stress-CMOD curve, in mmMPa).
10 kg/m3; but this increase was 21% instead of 35% if the w/c ratio was 0.26 instead of
0.39.
4 Conclusions
Different FRC mixes were produced and tested in order to evaluate the sensitivity of
their mechanical properties to changes in the following parameters: w/c ratio (0.26 or
0.39), type of fibre (glass or synthetic macrofibres), and fibre content (up to 10 kg/m3 or
15 kg/m3 for synthetic or glass fibres, respectively). The following conclusions were
obtained:
• The addition of the glass or synthetic fibres considered in this study, irrespective of
their dosage, did not cause statistically significant changes to the average com-
pressive strength or splitting tensile strength. However, increasing the fibre content
was found to reduce the standard deviation of compressive strength values.
• FRC specimens with glass fibres in dosages of 2.5 kg/m3 and 5 kg/m3 presented
brittle failure in flexure, and only those with 15 kg/m3 of glass fibres showed
residual flexural capacity. Based on the results obtained, it was estimated that glass
fibres need to be added in contents of at least 10–12 kg/m3 in order to achieve
residual flexural capacity.
• No cases of brittle failure in bending were observed amongst the FRC specimens
with synthetic fibres. They all presented a softening post-peak behaviour, irre-
spective of the w/c ratio and fibre content.
• When glass and synthetic fibres are compared at the maximum contents considered
in this study, synthetic FRC specimens clearly outperformed their glass FRC
counterparts. The level of residual flexural strength, represented by the ratio fR1/fL,
varied between 0.42-0.61 for synthetic fibres, as opposed to 0.26-0.33 for glass
fibres.
• The residual flexural strength parameters fR1 and fR2 were found to be sensitive to
changes in w/c ratio, fibre type and fibre content. It was concluded that reducing the
w/c ratio could lead to improvements in fR1 and fR2 comparable to those achieved by
increasing the fibre content.
• Flexural toughness was evaluated through the area under the stress-CMOD curves
up to a crack opening of 3.5 mm. Increasing the synthetic fibre content was found to
decrease the sensitivity of this parameter to variations in the w/c ratio.
Acknowledgements. The authors wish to acknowledge the contribution of Oscrete Construction

Products, part of Christeyns UK Ltd, and Sika Ltd (UK), which very kindly provided some of the
materials used in this study, as well as the support and assistance provided by the technical staff
of the School of Civil Engineering, University of Leeds. The authors are also thankful to Al-
Zaytoonah University of Jordan for the financial support granted to Ms Al Marahla in under-
taking her PhD studies at the University of Leeds.
References
1. Biolzi, L., Cattaneo, S., Guerrini, G.L.: Fracture of plain and fiber-reinforced high strength
mortar slabs with EA and ESPI monitoring. Appl. Composite Mater. 7(1), 1–12 (2000)
2. Kazemi, M.T., Golsorkhtabar, H., Beygi, M.H.A., Gholamitabar, M.: Fracture properties of
steel fiber reinforced high strength concrete using work of fracture and size effect methods.
3. Soutsos, M.N., Le, T.T., Lampropoulos, A.P.: Flexural performance of fibre reinforced
concrete made with steel and synthetic fibres. Constr. Build. Mater. 36, 704–710 (2012)
4. Cho, B., Lee, J.H., Back, S.Y.: Comparative study on the flexural performance of concrete
reinforced with polypropylene and steel fibers. J. Korean Soc. Civil Eng. 34(6), 1677–1685
(2014)
5. Olivito, R.S., Zuccarello, F.A.: An experimental study on the tensile strength of steel fiber
reinforced concrete. Composites Part B: Eng. 41(3), 246–255 (2010)
6. Thomas, J., Ramaswamy, A.: Mechanical properties of steel fiber-reinforced concrete.
J. Mater. civil Eng. 19(5), 385–392 (2007)
7. Simoes, T., Costa, H., Dias-da-Costa, D., Júlio, E.N.B.S.: Influence of fibres on the
mechanical behaviour of fibre reinforced concrete matrixes. Constr. Build. Mater. 137, 548–
556 (2017)
8. Lee, J.H.: Influence of concrete strength combined with fiber content in the residual flexural
strengths of fiber reinforced concrete. Composite Struct. 168, 216–225 (2017)
10. BS EN 14651, Test Method for Metallic Fibre Concrete-Measuring the Flexural Tensile
Strength, British Standard Institute, UK, pp. 1–20 (2007)
11. BS EN 12390–6, Testing Hardened Concrete. Tensile Splitting Strength of Test Specimens.
British Standard Institution, London (2009)
12. BS EN 12390–2, Testing hardened concrete-Part 2: Making and curing specimens for
strength tests, European Committee for Standardization, Brussels (2009)
13. ACI committee, ACI 544.1 R-96, State-of-the-art report on fiber reinforced concrete-
Technical report, ACI Farmington Hills, Michigan (2003)
14. Mazaheripour, H., Ghanbarpour, S., Mirmoradi, S.H., Hosseinpour, I.: The effect of
polypropylene fibers on the properties of fresh and hardened lightweight self-compacting
concrete. Constr. Build. Mater. 25(1), 351–358 (2011)
15. Cifuentes, H., García, F., Maeso, O., Medina, F.: Influence of the properties of
polypropylene fibres on the fracture behaviour of low, normal and high-strength FRC.
Bond Between Steel Reinforcement Bars
and Fiber Reinforced Cement-Based
Composites
Margareth S. Magalhães(&), Paulo José B. Teixeira,

and Maria Elizabeth N. Tavares
Civil Engineering Postgraduate Program, State University of Rio de Janeiro,

UERJ, Rio de Janeiro , Brazil
margareth.magalhaes@uerj.br
Abstract. This paper deals with the bond between steel reinforcement and
strain hardening cement-based composites (SHCC). Pull-out tests were carried
out according to RILEM standard on specimens made with three mixtures,
characterized by different fiber content (1% and 2%) beyond the control plain
mortar. Ribbed steel reinforcement bars with 8 mm and 10 mm diameter were
used to observe the influence of steel bar diameter. Experimental bond-slip
relationships were analysed, and results show enhanced bond resistance when
fiber is used in the mixture. SHCC specimens (composite with fiber content
equal to 2%) presented the best bond performance in terms of bond strength and
stiffness retention capacity, as well as damage control ability.
Keywords: Strain hardening cement-based composites PVA fiber Bond

behavior Pull-out test Steel reinforcement
1 Introduction
Strain hardening cement-based composite (SHCC) is a kind of composite material that

display a ductile behavior when subjected to tensile loading, different to other fiber
reinforced cementitious composites and ordinary concrete. SHCC can resist the full
tensile load with strain capacity up to 5.0%, leading to a high energy absorption
capacity. During the increase of strain under tensile loading, a strain hardening effect is
found and many, closely spaced, micro cracks (less than 70 µm in width) are formed in
the material [1, 2]. The compressive behavior of SHCC is also more ductile compared
to ordinary concrete, but the real benefit is attained when SHCC is used to resist a force
or imposed tensile strain.
Lately, SHCC have been widely utilized in civil engineering due to the remarkable
mechanical properties, especially in flexural and shear dominated members, such as
coupling beams, low-rise walls and beam-column joints [3–8]. Findings suggested that
SHCC can effectively improve bond efficiency of reinforcement by reducing slippage of
column bars [5], enhance the shear strength, energy dissipation and damage tolerance of
members, and consequently improve the performance of reinforced SHCC structures.
The experiences of current applications of SHCC are of paramount importance.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_29
314 M. S. Magalhães et al.
However, review on the existing literature related to SHCC indicates that the research on
the bond behavior between steel reinforcement and SHCC is limited [8, 10–15].
Bond strength is one of the most important parameters in reinforced elements
design, both at the ultimate and serviceability limit state. Poor bond property weakens
the bearing capacity of a component and then leads to structural failure. Therefore,
effective bonding between steel bars and SHCC is crucial for the components to
function in a stable and collaborative manner. According to Fischer and Li [9], by using
SHCC, it is possible to reduce the slip of steel bars and improve interfacial bonding
between steel bar - SHCC.
Chao et al. [8] observed that the composite type plays a vital role in bond behavior
of steel bar - composites. According Toshiyuki and Hiroshi [10], by using SHCC is
possible to reduce the cover thickness and bar spacing of the main bars in concrete
members. Lee et al. [11] proposed a constitutive model applicable to the bond-slip
behavior between the ECC, a kind of SHCC, and steel bar and Li et al. [12] observed
that the bond strength decreased as the diameter and embedded length of steel bars
increased.
Other studies were conducted on bond between fiber (basalt or glass) reinforced
polymer bar and ECC [16, 17]. Wang et al. [16], based on the pull-out test of basalt fiber
reinforced polymer (BFRP) bar embedded in ECC, pointed out that adding PVA fiber
can enhance the damage tolerance of matrix and thus improve the bond behavior. Kim
and Lee [17] investigated the bond behavior between glass fiber reinforced polymer
(GFRP) bar and ECC. Test results showed that energy absorption capacity of specimens
was enhanced due to the addition of PVA fiber and the ductility of bond stress-slip
curves significantly increased as fiber volume content increased from 1% to 2%.
The purpose of the research described in this paper is to gain a better understanding
on the bond behavior between steel reinforcement and SHCC. Pull-out tests were
carried out on specimens made with three different mixtures, characterized by different
fiber content. Ribbed steel reinforcement bar with two different diameters were used to
observe the influence of steel bar diameter.
2.1 Materials and Composite Manufacturing
The raw materials used on the preparation of the SHCC specimens, named C02
mixture, were Portland cement CPII F-32, defined by the Brazilian standard [18],
composed of filler (in mass: 6–10%) with a 28 days-compressive strength of 32 MPa;
fly ash with a density of 2.26 g/cm3; silica sand with a maximum diameter of 0.30 mm;
a superplasticizer and water. The fiber used was a polyvinyl alcohol (PVA) fibre,
manufactured by kuraray company, Japan, with a diameter of 40 µm, density of
1.31 g/cm3, elastic modulus of 40 GPa and tensile strength of 1600 MPa [19].
To study the influence of the PVA fiber content on the bond behavior of steel bar-
composites, other two mixtures were produced: one mixture without fiber (CM) and
another mixture with 1% PVA fiber (C01). The mixtures proportion used in this study
are shown in Table 1.
Bond Between Steel Reinforcement Bars 315
Table 1. Mixture design (kg/m3) of composite.

CM C01 C02
Cement 505 505 505
Fly ash 605 605 605
Sand 404 404 404
Water 404 404 404
PVA fiber – 13 26
Superplasticizer/(cement + fly ash) (%) – 0.12 0.13
Fly ash/cement 1.20 1.20 1.20
Water/(cement + fly ash) 0.36 0.36 0.36
To produce the mixtures, all dry raw materials, except fibers, were mixed for 3 min
in a mechanical mixer with a 20 L capacity. Water and superplasticizer were added to
form the basic matrix. The mixture was stirred for another 8 min to allow appropriate
workability of the matrix. In the last step, when used, fibers were added manually to the
cementitious matrix and the mixture was stirred for another 5 min.
The formulation of the C02 mixture, given in Table 1, was developed by Magal-
hães and co-workers in earlier studies [2]. The composite exhibits a strain-hardening
behavior with average tensile strain capacity of approximately 3% and average crack
width around of 68 µm (see Fig. 1).
Fig. 1. Typical tensile stress – strain and average crack width - strain curves of SHCC [2].
2.2 Test Procedure

Compressive strength tests at 28 days of mixes were carried out on 3 cylindrical
specimens of 50 100 mm (diameter height).
Ten pull-out specimens per each mixture were casted with a central horizontal bar.
A ribbed steel rebar with a nominal yield tensile strength of 500 MPa was used. The
casting direction was perpendicular to the longitudinal axis of the bar. Cubes’ side
(100 mm) was 10 times the diameter of the bar (10 mm), and the embedded length
(50 mm) was 5 times the diameter of the bar, following RILEM guidelines [20].
A plastic sleeve was used to limit the non-adherent zone (50 mm), situated at the
loading face. The same procedure was used to all tests with the three mixes and
repeated for another test set with C02 mix and bar diameter equal to 8 mm.
Pull out tests were carried out at 28 days after casting: the test setup is shown in
Fig. 2. A universal testing machine with a capacity of 600 kN was used to perform the
tests, and displacement-control was applied to capture the post-peak behavior. The
lower surface of the cube was restrained by a stiff 15 mm steel plate, with a hole of
32 mm diameter in the center. Between the specimen and the plate, a thin layer of
rubber of 5 mm was placed, to ensure that a uniform contact was realized and to
minimize friction effects.
Fig. 2. Setup for pull-out tests.
3 Experimental Results and Discussion
Test results are given in Table 2 with regards to compressive strength (fc), ultimate
bond strength (su), corresponding to the peak bond stress, and the corresponding slip
(su), residual bond strength (sr) and the failure modes. The residual bond strength was
obtained at a slip of 7 mm. The bond stress between anchorage steel and composite
was calculated as follows:
F
s¼ ð1Þ
pdla
where s is the bond stress, F is the applied load, d is the bar diameter and la is the
anchorage length.
Two different types of failures, such as pull-out failure and slitting failure, were
observed in the tested specimens. Pull-out mode generally occurs in confined specimen
Table 2. Summary of results from compressive and pull-out test (standard deviation in
parenthesis).
Mix fc (MPa) su (MPa) Su (mm) sr (MPa) Failure mode
CM-10 39.83(0.58) 11.28 (0.43) 0.41 (0.11) Splitting
C01-10 38.99 (0.20 11.42 (0.68) 1.14 (0.61) 2.99 (0.65) pull-out
C02-10 36.27 (0.15) 11.86 (1.76) 1.30 (0.18) 3.06 (0.68) pull-out
C02-8 36.27 (0.15) 9.23 (1.72) 1.09 (0.11) 2.86 (0.51) pull-out
in which the bond failure is due to the pull-out of the bars. In this research, specimens
with 1% and 2% fiber showed appreciable damage tolerance and maintained their
integrity throughout the test due to the fiber bridging effect, which prevented internal
cracks from opening widely. A clean bar was pulled out in most specimens and no
cracks (visible to the naked eye) were observed on the surface of the specimens, as
shown in Figs. 3b, c, d, e. In contrast, the confinement was not sufficient to prevent
splitting of mortar cover and a longitudinal cracking failure occurred in all mortar
specimens without fibers (CM-10), which exhibited an obvious brittle feature (see
Fig. 3a).
(a) (b) (c) (d)
Fig. 3. Typical specimens after tests: (a) CM-10, (b) C01-10, (c) C02-10 and (d) C02-8.
The concrete-steel bars bond is considered to stem from three main factors:
chemical bond, friction force and the mechanical thread strengths, which is the mainly
responsible by the adhesion in ribbed steel bars. Figure 4 shows the bond stress versus
slip typical curves (on the left) and a zoom of the pre-peak zone (on the right) to clearly
highlight the ascending branch, for the specimens with 10 mm diameter bars. At the
initial loading portion, up to approximately 2–3 MPa, the curves showed a relatively
sharp slope for all specimens, indicating the load transfer by the chemical adhesion.
After, the ascending portion of the curves became nonlinear, up to the applied load
reached the maximum value. From Table 2, it can be observed that mortar mixture
(CM–10) displayed insignificant reduction on the peak stress as compared with the
composite specimens (C01–10 and C02–10) due to splitting of mortar cover. The
difference observed on the ultimate bond strength between mortar (CM–10) and
composites (C01–10 and C02–10) was up to –5.2%.
(a) (b)
Fig. 4. Bond stress-slip curves: (a) effect of fiber content and (b) a zoom of the pre-peak zone.
Comparing the slip values, corresponding to the ultimate bond stress, a significant
increase of this bond characteristic is obtained for the C01–10 (1.14 mm) and C02–10
(1.30 mm) mixes, if compared to the other mix without fiber, CM–10 (0.41), that are
instead characterized by lower value (0.41 mm). This result indicates that, the con-
finement effect in the mortar was enhanced due to fiber addition. The fiber has effec-
tively controlled the crack opening and widening and then improving the bond
behavior of mortar specimens that failed due to splitting cracks occurrence.
Regarding to the fiber content (mixes C01–10 and C02–10), less differences was
observed in the ultimate bond strength values of the specimens. This can be assigned to
the fact that pull-out strength is mainly governed by composite compressive strength,
which is very similar for the specimens (C01: 38.99 MPa and C02: 36.27 MPa), being
mostly affected by the similarity of the cementitious matrix and little effect of
increasing fiber content.
The contribution of PVA fibre subsequent to matrix cracking was more apparent in
the post-peak behavior, mainly in the SHCC specimen (C02–10 mix). After ultimate
bond strength, the bearing capacity of CM–10 specimen has a sudden loss. As contrast,
the curve of specimens with fibers declines slowly which exhibits better ductility
performance. With the further increase in slippage, the bond stress of C02–10 gradually
tends to be steadier than C01–10 due to higher fiber content. Indeed, the average
residual bond strength of C02–10 is slightly higher than the value of C01–10 (+2.4%).
The contribution of the fibers resulted in the improvement of the toughness com-
pared to the specimen with no fibers. The properties and content of fibers plays a very
important role in controlling the development of cracking. Therefore, the contribution
of fibers to interfacial bond property during the crack development and propagation can
be summarized as the totally dissipated bond energy and it can be evaluated by
determining and evaluating the toughness at different slips value. The toughness of the
composites was calculated from the area under the bond stress–slip curves up to slips
values of Su (Tu), 2 mm (T2), 4 mm (T4), 6 mm (T6) and 7 mm (T7) and the average
values are presented in Table 3. In this aspect, the PVA fibers with volume fraction of
2% in total volume (C02–10) showed a good result on the improvement of the
Table 3. Average toughness from pull-out test results (standard deviation in parenthesis).
Mix Toughness (KJ/m2)
Tu T2 T4 T6 T7
C01-10 12.70 (6.73) 20.73 (1.34) 38.77 (0.82) 49.66 (1.32) 53.04 (1.91)
C02-10 13.35 (2.11) 22.95 (3.89) 42.17 (5.97) 52.63 (7.63) 55.86 (8.18)
C02-8 7.88 (2.38) 14.76 (4.46) 25.40 (8.33) 30.61 (10.58) 32.76 (11.38)
toughness at any slip level and maintenance of the residual load (see Table 2) for
10 mm bar.
The Fig. 5 indicate that the basic shape of the bond stress–slip curve of SHCC
specimens is not affected by the bar diameter, however the bond stress of the C02–10
specimen was always higher than that of the specimen C02–8, at any given slip, due to
an increase of effective bonding area in bars with higher diameters, which leads to an
increase in bond strength. This finding is contrary to those obtained in other studies [12,
15]. The results, in Table 2, also showed that the slip at peak load and residual of C02
specimen decreased, respectively, 21.6% and 6.5% with a decrease of the diameter of
the steel bar. Furthermore, the toughness values, as seen in Table 3, were also reduced
up to approximately 42%.
Fig. 5. Bond stress-slip curves: (a) effect of fiber content, (b) effect of steel bar diameter.
4 Conclusions
The purpose of the research described in this paper was to gain a better understanding
on the bond behavior between steel reinforcement and composite. Most relevant
conclusions of the scientific research are:
• Two different types of failures were observed in the tested specimens: ductile failure
with steel bar pull-out in composite specimens and brittle failure in matrix
specimens.
• The increase in the content of PVA fiber don’t affect the peak stress of the com-
posites, however the slip values, corresponding to the ultimate bond stress, increase.
• The fiber has effectively controlled the crack opening and widening and then
improving the bond behavior of mortar specimens that failed due to splitting cracks
occurrence.
• Bond strength, residual bond strength and toughness increase as the diameter of
steel bars increase.
Acknowledgements. The authors acknowledge the Foundation for Research Support of the
State of Rio de Janeiro (FAPERJ) for the financial support and Pozo Fly for supplying the fly ash.
References
1. Li, V.C., Wu, C., Wang, S., Ogawa, A., Saito, T.: Interface tailoring for strain hardening
PVA-ECC. ACI Mater. J. 99(5), 463–472 (2002)
2. Magalhães, M.S., Toledo Filho, R.D., Fairbairn, E.M.R.: Influence of local raw materials on
the mechanical behaviour and fracture process of PVA-SHCC. Mater. Res. J. 17(1), 146–156
(2014)
3. Kunieda, M., Rokugo, K.: Recent progress of SHCC in Japan – required performance and
applications. J. Adv. Concr. Technol. 4(1), 19–33 (2006)
4. Lim, Y.M., Li, V.C.: Durable repair of aged infrastructures using trapping mechanism of
engineered cementitious composites. Cem. Concr. Compos. 19(4), 373–385 (1997)
5. Parra-Montesinos, G.J.: High-performance fiber-reinforced cement composites: an alterna-
tive for seismic design of structures. ACI Struct. J. 102(5), 668–675 (2005)
6. Deng, M.K., Zhang, Y.X.: Cyclic loading tests of RC columns strengthened with high
ductile fiber reinforced concrete jacket. Constr. Build. Mater. 153, 986–995 (2017)
7. Fischer, G., Li, V.C.: Deformation behavior of fiber-reinforced polymer reinforced
Engineered Cementitious Composite (ECC) flexural members under reversed cyclic loading
conditions. ACI Struct J. 100(1), 25–35 (2003)
8. Chao, S.H., Naaman, A.E., Parra-Montesinos, G.J.: Bond behavior of reinforcing bars in
tensile strain-hardening fiber-reinforced cement composites. ACI Struct. J. 106(6), 897–906
(2009)
9. Fischer, G., Li, V.C.: Effect of matrix ductility on deformation capacity behavior of steel-
reinforced ECC flexural members under reversed cyclic loading conditions. ACI Struct J. 99
(6), 781–790 (2002)
10. Toshiyuki, K., Hiroshi, H.: Bond-splitting strength of reinforced strain-hardening cement
composite elements with small bar spacing. ACI Struct J. 112(2), 189–198 (2015)
11. Lee, S.W., Kang, S.B., Tan, K.H., Yang, E.H.: Experimental and analytical investigation on
bond-slip behaviour of deformed bars embedded in engineered cementitious composites.
12. Li, X.L., Bao, Y., Xue, N., et al.: Bond strength of steel bars embedded in high performance
fiber-reinforced cementitious composite before and after exposure to elevated temperatures.
Fire Saf. J. 92, 98–106 (2017)
13. Bandelt, M.J., Frank, T.E., Lepech, M.D., et al.: Bond behavior and interface modeling of
reinforced high-performance fiber-reinforced cementitious composites. Cem. Concr. Com-
pos. 83, 188–201 (2017)
14. Deng, M., Pan, J., Sun, H.: Bond behavior of steel bar embedded in Engineered
Cementitious Composites under pullout load. Constr. Build. Mater. 168, 705–714 (2018)
15. Asano, K., Kanakubo, T.: Study on size effect in bond splitting behavior of ECC. High
Perform. Fiber Reinforced Cem. Compos. 6, 137–144 (2012)
16. Wang, H., Sun, X., Peng, G., et al.: Experimental study on bond behaviour between BFRP
bar and engineered cementitious composite. Constr. Build. Mater. 95(1), 448–456 (2015)
17. Kim, B., Lee, J.Y.: Polyvinyl Alcohol Engineered Cementitious Composite (PVAECC) for
the interfacial bond behaviour of Glass Fibre Reinforced Polymer Bars (GFRP). Polym.
Polym. Compos. 20(6), 545–558 (2012)
18. Brazilian Standard NBR 11578. Cimento Portland Composto. Associação Brasileira de
Normas Técnicas (ABNT) (1991)
19. Magalhães, M.S., Toledo Filho, R.D., Fairbairn, E.M.R., et al.: Durability under thermal
loads of polyvinyl alcohol fibers. Matéria (Rio J.) 18(4), 1587–1595 (2013)
20. RILEM/CEB/FIP. Recommendations on reinforcement steel for reinforced concrete.
Revised edition of: RC6 bond test for reinforcement steel: Pull-out test, CEB News 73,
Lausanne, Switzerland (1983)
An Experimental Study of the Influence
of Moderate Temperatures on the Behavior
of Macrosynthetic Fiber Reinforced Concrete
Marta Caballero-Jorna(&), Marta Roig-Flores, and Pedro Serna
Instituto de Ciencia y Tecnología del Hormigón,

Universitat Politècnica de València, Valencia, Spain
marcajor@upv.es
Abstract. This research shows an experimental campaign performed aiming to

understand the influence of moderate temperatures (5 °C, 20 °C and 50 °C) on
the mechanical behavior of macrosynthetic fiber reinforced concrete (MSFRC).
To do this, a 35 MPa concrete was selected. Two types of polypropylene fibers
were added in a relatively high content (10 kg/m3). Steel fiber was used in an
equivalent volume as a reference fiber.
Considering that there is no standardized protocol to analyze the temperature
effect on MSFRCs, a testing methodology based on UNE-EN 14651:2007 +
A1:2008 is proposed. This modified procedure includes a system of insulating
thermal covering which allows to maintain the temperature during the tests.
Additionally, this study compares the effect of temperature on the pre-cracked
and uncracked sections of MSFRCs. For this purpose, MSFRC and SFRC
specimens were stored in each conservation temperature. Some of these samples
were pre-cracked at the age of 28 days and then, they were reassigned in their
respective conservation temperature until testing at 60 days.
The results show that the analyzed temperatures have some influence on the
mechanical properties of MSFRCs in terms of their residual flexural strength but
to a limited extent, being suitable for environments with equivalent temperatures.
Keywords: Aging Fiber reinforced concrete Macrosynthetic fibers

Moderate temperature Residual flexural strength
1 Introduction
The application of fibers in concrete emerged in the sixties as a result of the constant
search of potential enhancements in this construction material related to its mechanical
performance and sustainability. Nowadays, there are numerous fiber types on the
market which can be used to reinforce concretes. Steel fibers are the most commonly
employed but in the last few decades, macrosynthetic fibers have appeared as an
alternative in several applications due to the enhancements they offer [1].
Despite the enhancement they can provide to concrete, there is still some uncer-
tainty regarding the durability of this type of fibers and its influence in the behavior of
concrete. One of the most critical conditions to which macrosynthetic fibers can be
subjected is temperature. The deterioration of their properties may occur since they are
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_30
An Experimental Study of the Influence of Moderate Temperatures 323
thermally dependent. Macrosynthetic fibers can be found dimensionally unstable or

more brittle; they can present a loss of bond or even be melted at moderate tempera-
tures. As a result, the mechanical behavior of macrosynthetic fiber reinforced concretes
(MSFRCs) may be affected.
An analysis of the literature has revealed that many of the studies are focused on the
behavior of polymeric fibers reinforced concretes at high temperature, mainly aiming to
assess fire conditions where it is well proven that micro synthetic fibers are suitable in
order to reduce spalling, since the fibers produce channels that reduce pore pressure
[2–4]. It is noteworthy that adding synthetic fibers into concrete to improve spalling
resistance is also recommended by Eurocode 2 [5]. However, only scarce information
can be found on MSFRCs behavior at no standard temperatures under serviceability
conditions, and the existing studies show discrepancies.
Richardson and Ovington [6] investigated the effect of temperature variation (be-
tween −20 °C and 60 °C) on concrete properties with steel and polymeric fibers. The
results showed that all beams exposed to a temperature of 60 °C experienced a flexural
strength reduction when comparing to ambient temperature. However, a lesser decrease
in peak flexural strength was observed when polypropylene fibers (12%) were com-
pared to steel fibers (49%) and to plain concrete (35%). For all beams tested at cold
temperature, flexural strength increased when compared to those performed at ambient
temperature. The plain beam showed an average 99% increase in flexural strength, the
polypropylene fiber beam showed an average a 150% increase in flexural strength and
the steel fiber beam showed an average a 79% increase in flexural strength. This
outcome concurs with Mirzazadeh et al. [7] who stated that concrete beams tested at
−25 °C demonstrated an increase in strength and ductility of 13% and 34% respec-
tively, compared to those tested at room temperature.
Strauss Rambo et al. [8] showed the influence of temperature on the mechanical
behavior of the MSFRCs were very similar to that known for plain concrete with
relation to the loss of mechanical strength and elastic modulus. The residual tensile
strength and elastic modulus of the macrosynthetic fibers were not affected by the
temperature up to 100 °C remaining resistant capacity may be enough to ensure safety
in service depending on the temperature reached.
In contrast, Buratti and Mazzotti [9] found that temperatures ranging from 20 °C to
40 °C reduced the short-term residual strength of some of the MSFRCs analyzed. They
concluded that temperature should be considered as factor when designing structures
with macrosynthetic reinforcement, given that the effect of temperature is more relevant
for MSFRCs than for SFRCs. This is backed up by [10], in which residual strength of
the MSFRC decreased (about 10%) for crack mouth opening displacement (CMOD)
higher than 1.5 mm since the temperature increased.
Additionally, looking at the experimental procedure to determinate the thermo-
mechanical behavior of MSFRCs, only a few investigations have examined mechanical
characterization at the targeted temperatures [11–13] since most of the investigations
have carried out residual tests at room temperature after heating/cooling.
The different conclusions reached by the aforementioned studies shows that the
temperature effect on the properties of MSFRCs is not completely clear yet. This
context has motivated the present research, which aims to evaluate the mechanical
behavior of macrosynthetic fibers reinforced concrete exposed to cold and hot moderate
324 M. Caballero-Jorna et al.
temperatures in order to analyze whether macrosynthetic fibers may be a viable

alternative in environments and applications where they have a currently limited use
and whenever MSFRC elements should be designed considering temperature-induced
mechanical degradation.

2.1 Concrete and Fiber Properties
This study assesses the effect of moderate temperatures on the structural properties of
macrosynthetic fiber reinforced concretes when compared steel fiber reinforced con-
crete control samples with focus on the behavior in cracked state. To this purpose, a
base concrete was selected, with a compression strength around 35 MPa and high
workability, common in the precast industry. The mix contains Ordinary Portland
Cement CEM II 42.5R, two coarse aggregates (4-6 mm and 6-10 mm), river sand (0–
4 mm) and limestone filler. A polycarboxylate-based superplasticizer (SikaViscoCrete-
5980) was employed to guarantee the workability of the mixtures, being its dosage
slightly adjusted for each batch. The effective w/b ratio of the mix was 0.55. The mix
designs are displayed in Table 1.
Three types of fibers were added into the base concrete: two polymeric fibers and
one steel fibers. Fiber contents used were equivalent in a volume to achieve near equal
flexural performance for both types of fibers, 10 kg/m3 for polymeric fibers and
30 kg/m3 for steel fibers. These quantities were selected due to their relatively frequent
use. The most relevant properties of the three types of fibers used are summarized in
Table 2.
Table 1. MSFRC and SFRC dosages used in the present study.

Dosage MFRC Dosage SFRC
(kg/m3) (kg/m3)
CEM II 42,5R 325 325
Gravel 6-10 mm 430 430
Gravel 4–6 mm 580 580
Sand 0–4 mm 835 835
Limestone Filler 80 80
Water 178.75 178.75
Fibers 10 30
Table 2. Properties of the fibers used in the present study.
2.2 Experimental Program

The experimental campaign was developed to study the effect of moderate temperatures
(5, 20 and 50 °C) on the mechanical behavior of MSFRCs, in particular their residual
flexural strengths.
Specimens were cast in eight concrete batches to produce prisms of 100 100
400 mm to perform three-point bending tests at target temperatures. Additionally, to
characterize the batches, slump test was performed and cylindrical samples of Ø
150 300 mm were casted for compression tests at 28 days.
To date, there is no standardized protocol to evaluate the influence of temperatures
on FRCs during testing. Thus, a methodology based on the three-point bending test
according to standard EN 14651 (2/3 scale) was implemented.
2.2.1 Preparation and Conservation Conditions of Samples

All concrete mixes were manufactured in a DIEM WERKE model DZ 180 V mixer by
the same sequence to ensure reproducibility of results. First, coarse aggregates and sand
were added and pre-mixed for two minutes. Then, cement and filler were added and
mixed for two minutes. Afterwards, water was added and then mixing for one minute.
Finally, following the gradual addition of the fibers, superplasticizer was added, and a
final additional mixing was applied for about five minutes.
After mixing, the slump was measured according to standard EN 12350-2:2009.
Then, the concrete was taken directly from the mixer and was poured into the molds
according to the UNE EN 12350-1.
On the one hand, 3 cylindrical specimens were cast for compression test for each
batch, testing in total 24 specimens. The guideline for determining the compression
strength in specimens of concrete in the hardened state was standard UNE-EN 12390-
3:2009.
On the other hand, prisms for three-point bending tests were cast and compacted
over a vibrating table, being in total 48 prisms. They were demolded after 24 h of
casting and kept at 20 ± 2 °C and 95% of relative humidity for 3 days. After that, they
were located to their respective condition of conservation for two months until testing
time.
In this study, three different conservation temperatures were investigated, repre-
senting possible situations during service life in a Spain’s Mediterranean climate:
• 5 °C, using an industrial fridge, this condition represents moderately cool weather.
• 20 °C and RH 95%, using a controlled temperature room, to represent warm and
humid weather.
• 50 °C and saturated, using a thermal bath in water at target temperature, to represent
situations such as in concrete heaters or industrial containers.
Different measurements of the temperature in each defined case were conducted in
order to verify the conservation conditions throughout the time. Figure 1 presents that
the temperatures of 5 °C, 20 °C and 50 °C were quite stable during the exposure
conditions.
All the mixes were studied in pre-cracked conditions. One mix with polymeric
fibers was studied in uncracked and cracked conditions. The objective of this process
was to evaluate if the exposure of the open crack to moderate temperatures affects the
residual flexural strength in MSFRCs. The specimens that were studied in cracked
conditions were taken out of their conservation conditions and pre-cracked up to
CMOD of 0.5 mm, at 28 days of conservation. Then, the specimens were returned to
their corresponding conservation conditions until final testing at 60 days. The rest of
the samples were tested without being subjected to the pre-cracked phase after two
months.
50
Temperature ( C)
40
30
20
10
0
0 7 14 21 28 35 42 49 56 63 70 77 84 91
Conservation (Days)
Fig. 1. Control of temperature during the conservation period for the three conditions studied.
2.2.2 Methodology of Flexural Test at Target Temperatures

This test methodology was designed to determine how the mechanical properties of
fiber reinforced concrete vary considering temperature, from 5 °C to 50 °C. To ensure
the temperature was constant during the test, as novelty in this procedure, a cover
system was used to insulate the specimens during testing when the specimens were
removed from their storage areas.
This cover system consisted on reusable cold and hot gel that provide and maintain
the required temperature, while thermal sacks were used for insulation. The blue gel is
not toxic and is 99% biodegradable, which makes the product environmental-friendly.
For the flexural tests, the adopted configuration was scaled with a factor 2/3 from
the standardized test according to UNE-EN 14651:2007 + A1:2008. Specimens were
notched at the center of the prisms and LVDTs were placed to register the displace-
ments. Figure 2 shows two pictures of the setup decided to maintain temperature inside
concrete specimen during the test. The process was run at a load velocity of 2 mm/min.
Fig. 2. System of insulating thermal covering in the sample during testing.
Finally, to calculate the residual flexural strength, Eq. (1) was employed:
3 Fj l
fR;j ¼ ð1Þ
2 bh2sp
where Fj is the axial load recorded during the test; l, the distance between supports
(330 mm); b, the width of the sample cross section (100 mm) and hsp, the distance
between top of the notch and top of cross section (85 mm).
The results of the characterization tests and residual strength are collected and ana-
lyzed, to determine the uniformity and quality of the different concrete mixes and to
show the temperature effect on MSFRCs and SFRCs, respectively.
3.1 Characterization Tests

Workability of fresh concrete was evaluated by slump test. Table 3 sums up thereby the
data related to the workability test for each batch. As is shown, the slump of concrete is
affected by the fiber type, its properties and its geometrical structure. This effect may be
verified in all cases, according to Bolat [14]. Moreover, it is of importance to
emphasize that this influence is quite significative due to the high fiber contents added
in the concrete mixes.
All in all, polypropylene fibers reduced more the workability of concrete than steel
fibers. The higher workability values were obtained for SFRC, without segregation,
while a clear loss of workability was experimented by mixes with P2 fibers. It could be
explained due to the aspect ratio of this kind of fiber and the number of fibers on the
cross section. However, this effect can be overcome using of effective method of
compaction (vibrating tables). In the case of P1 fibers, slump values obtained are also
lower than those obtained for S fibers, although this does not imply a reduction in the
workability. Table 3 also summarizes the mean results of the compression performed
tests for each batch and their coefficients of variation (COV) calculated from 3 spec-
imens per batch. All compressive strengths results were around 35 MPa, which were
initially defined. From a general point of view, a slight decrease in compressive
strength is observed for the SFRC batches. However, the results indicate that there is no
significant deviation between the results.
Table 3. Results of complementary tests.

Code Consistency class Compression
fcm (N/mm2) COV (%)
P1 N-05 S2 35.05 1.89
N-20
P1 N-50 S2 37.26 2.49
P1 P-05 S2 35.68 2.39
P-20
P1 P-50 S2 35.42 1.28
P2 P-05 S1 35.20 1.22
P-20
P2 P-50 S1 36.47 2.53
S P-05 S3 31.60 5.80
S P-20 S3 32.73 1.46
P-50
3.2 Three-Point Bending Tests

To reach the main purpose of this work, three-point bending tests were performed at
different temperatures. The results of the residual flexural strengths corresponding with
the limit of proportionality (fLm) and the residual flexural strength fR,jm corresponding
to CMOD = CMODj (j = 1, 3) where CMOD1 = 0.5 mm, CMOD3 = 2.5 mm, and
their corresponding coefficients of variation are collected in Table 4 for each type of
fiber and conservation condition. Herein, it can be observed that the variability of the
results is influential within this study, which may complicate to obtain strong con-
clusions. This variability can be produced by the small size of the specimens used, the
number of the samples and the scattering of the obtained outcomes between several
specimens from the same production for the three-point bending tests.
Table 4. Results of the three-point bending tests at moderate temperatures on the studied
prisms.
Code Temperature fLm COV fR,1m COV Non pre-cracked Pre-cracked
(°C) (MPa) (%) (MPa) (%) fR,3m COV fR,3m COV
(MPa) (%) (MPa) (%)
P1 05 3.98 10.87 2.03 11.46 3.00 12.35 3.52 6.53
20 4.36 11.31 1.42 24.67 2.18 13.94 3.14 18.36
50 4.24 10.28 1.83 30.28 4.04 38.84 2.23 13.57
P2 05 3.91 8.65 1.26 10.34 – – 1.86 10.64
20 4.36 6.85 1.79 38.69 – – 3.14 41.36
50 5.10 6.59 1.84 25.93 – – 2.33 29.93
S 05 4.43 6.73 2.31 4.74 – – 2.35 26.43
20 4.27 3.43 2.95 18.96 – – 3.51 26.94
50 4.36 2.64 2.38 17.20 – – 2.69 18.84
To visualize the comparation between these outcomes, the residual flexural

strengths are depicted in Figs. 3, 4 and 5. For a better understanding, a representative
color was defined for each temperature. The blue color represents the coolest tem-
perature (5 °C) in the bar charts. The temperature of 20 °C is identified by the purple
color. The hot color depicts 50 °C. In addition, the mean value of all the data for every
fiber type, without considering the influence of temperature, is also showed, which is
the green bar. Firstly, as can be seen in Fig. 3, similar values in terms of residual
flexural strengths at peak-load are shown for each type of fiber, between 4 and 5 MPa.
These values are thought to be mainly dependent of the quality of the concrete matrix.
Moreover, it is possible to observe that steel and macrosynthetic fibers behaved in a
different way. For S samples, the increase or descent of the temperature had no sig-
nificant effect on the fLm. For P1 and P2 specimens, the effect of temperature on fL was
negligible, although a slight rising trend may be observed for P2 specimens.
6 5°C 20°C 50°C AVG

5
4
fLm (MPa)
3
2
1
0
P1 P2 S
Fig. 3. Results of fLm for each type of fiber at different temperatures.

6 5°C 20°C 50°C AVG
5
fR,1m (MPa)
4
0
P1 P2 S
Fig. 4. Results of fR,1m for each type of fiber at different temperatures.
Figure 4 shows the residual flexural behaviour at CMOD1 mm for the tested fibres
at different target temperatures. P1 and P2 samples obtained lower strengths than those
obtained by S specimens. It is noteworthy that the coefficients of variation were higher
than in the limit of proportionality for all fibers (see Table 4), indicating a greater
variability within a same batch. S specimens presented a decrease on fR,1m for cold and
hot temperature compared with the reference temperature (20 °C). For P2 elements, it is
possible to affirm that a decrease in temperature may cause a reduction in fR,1m.
However, the obtained residual flexural strengths at CMOD1 at 20 °C and 50 °C were
similar. In the case of fR,1m for P1 samples does not seem to be affected by temperature
since all values presented for residual flexural strength at CMOD1 are very similar.
Figure 5 shows the temperature effect on the residual flexural strength fR,3 at
CMOD3 for all types of fibers and all temperatures. This parameter is the one that
suffered the highest variation levels. Since this test was performed at a different stage
after pre-cracking and storage in the indicated conditions, this variation may be pro-
duced by handling during the displacement of the prisms from the storage to the testing
press.
SFRC specimens (S) exhibited lower values of fR,3m for targeted temperatures of 5 °
C and 50 °C than for the standard temperature. As in the case of CMOD1, the cold and
hot temperature could influence the residual flexural strength of SFRCs more than the
standard temperature.
For P2 elements, the temperature can cause changes in the values of fR,3m, but they
are not relevant. The best results for this type of fiber were obtained at 20 °C. In
general, with this fiber, at CMOD3 the residual strengths displayed a reduction when a
variation (increase or decrease) of temperature occurs.
For the specimens reinforced with the macrosynthetic fiber P1 in the pre-cracked
condition (same as in P2 and S), the best behavior occurred at low temperatures, but the
temperature factor does not seem to influence on fR,3m.
So, from a general point of view, it is not possible to observe any trend within
results of fR,3m. for those specimens that had pre-cracking.
Regarding the effect of the pre-cracked condition, there is a negative influence on
the residual flexural strength for the group pre-cracked (P1-P) when temperature
increases, and on the contrary, this increase of temperature has a positive effect for the
uncracked group (P1-N). Hence, the interaction between the factors pre-cracking
condition and the temperature can be potentially significant in some cases and is worth
an in-depth study.
6 5°C 20°C 50°C AVG

5
fR,3m (MPa)
4
3
2
1
0
P1-N P1-P P2-P S-P
Fig. 5. Results of fR,3m for each type of fiber at different temperatures.
4 Conclusions
In this research, an experimental campaign in order to obtain a better understanding of

the influence of moderate temperatures on the mechanical behavior of macrosynthetic
fiber reinforced concretes was carried out. From the obtained results, the conclusions
drawn are:
• No clear trend was obtained for the flexural strength in specimens that were pre-
cracked, indicating that both MSFRC and SFRC specimens, despite their differences,
did not suffer noticeable degradation at the target temperatures of 5 °C – 50 °C after
60 days of exposure. Thus, MSFRCs have potential to be used at moderate
temperatures.
• Regarding the responses obtained using the two polypropylene fibers, P1 fibers
obtained more efficient flexural performance than P2 fibers at the tested
temperatures.
• The interaction of pre-cracking condition with moderate temperatures can be
influential in some cases but need a specific study to determine its influence.
• The proposed experimental procedure, based on the standard UNE-EN
14651:2007 + A1:2008 (2/3 scale), allows to maintain the temperature during the
tests. This procedure includes a system of insulating thermal covering which
manages to set up the influence of moderate temperature as a factor on the
mechanical behavior of reinforced concretes. However, the reduced scale is a factor
that may increase the dispersion.
Acknowledgements. The authors would like to express their gratitude to the Spanish Ministry
of Science, Innovation and Universities for funding received under the FPU Program
[FPU18/06145].
References
1. Yin, S., Tuladhar, R., Shi, F., Combe, M., Collister, T., Sivakugan, N.: Use of macro plastic
fibres in concrete: A review. Constr. Build. Mater. 93, 180–188 (2015). https://doi.org/10.
1016/j.conbuildmat.2015.05.105
2. Kalifa, P., Chéné, G., Gallé, C.: High-temperature behaviour of HPC with polypropylene
fibres. Cem. Concr. Res. 31, 1487–1499 (2001). https://doi.org/10.1016/s0008-8846(01)
00596-8
3. Sideris, K.K., Manita, P., Chaniotakis, E.: Performance of thermally damaged fibre
reinforced concretes. Constr. Build. Mater. 23, 1232–1239 (2009). https://doi.org/10.1016/j.
conbuildmat.2008.08.009
4. Han, C.G., Hwang, Y.S., Yang, S.H., Gowripalan, N.: Performance of spalling resistance of
high performance concrete with polypropylene fiber contents and lateral confinement. Cem.
Concr. Res. 35, 1747–1753 (2005). https://doi.org/10.1016/j.cemconres.2004.11.013
5. European Committee for Standarisation, Eurocode 2: Design of concrete structures - Part 1–
2: General rules - Sturctural fire design, AENOR, Madrid (2011)
6. Richardson, A., Ovington, R.: Temperature related steel and synthetic fibre concrete
performance. Constr. Build. Mater. 153, 616–621 (2017). https://doi.org/10.1016/j.
7. Mirzazadeh, M.M., Noel, M., Green, M.F.: Effect of low temperature on the shear-fatigue
performance of reinforced concrete beams, Proceedings. Annu. Conf. Can. Soc. Civ. Eng. 4,
2682–2691 (2016)
8. Rambo, D.A.S., Blanco, A., de Figueiredo, A.D., dos Santos, E.R.F., Toledo, R.D., Gomes,
O.d.F.M.: Study of temperature effect on macro-synthetic fiber reinforced concretes by
means of Barcelona tests: An approach focused on tunnels assessment. Constr. Build. Mater.
158, 443–453 (2018). https://doi.org/10.1016/j.conbuildmat.2017.10.046
9. Buratti, N., Mazzotti, C.: Experimental tests on the effect of temperature on the long-term
behaviour of macrosynthetic Fibre Reinforced Concretes. Constr. Build. Mater. 95, 133–142
(2015). https://doi.org/10.1016/j.conbuildmat.2015.07.073
10. Buratti, N., Mazzotti, C.: Temperature effect on the long term behaviour of macro-synthetic-
and-steel–fibre reinforced concrete. In: 8th RILEM Int. Symp. Fibre Reinf. Concr.
Challenges Oppor. (2012)
11. Castillo, A.J., Durrani, C.: Effect of transient high temperature on high strength concrete.
ACI Mater. J. 87, 47–53(1987)
12. Cifuentes, H., Ríos, J.D., Leiva, C., Medina, F.: Influencia del tiempo de exposición a altas
temperaturas en el comportamiento en fractura de hormigones autocompactantes reforzados
con fibras. An. Mecánica Fract, pp. 1–6 (2016)
13. Cheng, F.P., Kodur, V.K.R., Wang, T.C.: Stress-strain curves for high strength concrete at
elevated temperatures. J. Mater. Civ. Eng. 16, 84–90 (2004). https://doi.org/10.1061/(ASCE)
0899-1561(2004)16:1(84
14. Bolat, H., Şimşek, O., Çullu, M., Durmuş, G., Can, Ö.: The effects of macro synthetic fiber
reinforcement use on physical and mechanical properties of concrete. Compos. Part B Eng.
61, 191–198 (2014). https://doi.org/10.1016/j.compositesb.2014.01.043
Post-cracking Behaviour of Glass Fibre
Reinforced Concrete with Recycled Aggregates
Brecht Vandevyvere1, Lucie Vandewalle2, Els Verstrynge2,

and Jiabin Li1(&)
1
Research Group RecyCon, Department of Civil Engineering,
KU Leuven Campus Bruges, 8200 Brugge, Belgium
jiabin.li@kuleuven.be
2
Department of Civil Engineering, KU Leuven, 3001 Leuven, Belgium
Abstract. In the past two decades, numerous studies have demonstrated the
benefit of adding fibres in structural concrete. The addition of fibres in concrete
limits the crack opening after concrete cracking. For conventional concrete with
natural aggregates, the post-cracking behaviour is mainly determined by the type
and amount of the fibres. Due to the increased shortage of raw materials, the use
of recycled aggregates (RAs) in structural concrete has gained more and more
interests recently. Nevertheless, the influence of fibres on the post-cracking
behaviour of concrete with RAs has not well been understood. This paper
presents an extensive experimental study on the behaviour of concrete with RAs
reinforced with alkali resistant (AR) glass fibres. To better understand the effect
of the quality of the RAs, three types of RAs were used in the present work,
which include a high grade recycled concrete aggregate (RCA+), a normal
quality RCA (nRCA) as well as a mixed recycled aggregate (MA). In addition, a
natural aggregate (limestone) is also used as the reference aggregate. A total of
24 fibre reinforced notched beams are tested under three-point bending load
according to EN 14651. In all fibre reinforced concrete mixtures, the CEM-FIL
MinibarsTM was used with a fibre content of 0.75 V%. The test results indicate
that the quality of the RAs can have a significant influence on the post-cracking
behaviour of the beams, especially in the Crack Mouth Opening Displacement
(CMOD) range of 0.5–1.2 mm.
Keywords: Fibre reinforced concrete Recycled aggregates (RAs) Alkali

resistant (AR) glass fibres Three-point bending Post-cracking
1 Introduction
Recycling of construction and demolition waste into recycled aggregates (RAs) saves
land space, decreases the transportation cost and reduces the consumption of natural
resources as well as CO2 emissions. Therefore numerous research activities have been
carried out to investigate the mechanical and durability properties of concrete con-
taining RAs [1]. The test results generally indicated a decrease in the mechanical and
durability properties. This reduction is mainly caused by the increased porosity of RAs
in comparison to natural aggregates due to the attached mortar. The attached mortar
influences the interfacial transition zone (ITZ) and consequently the concrete
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_31
334 B. Vandevyvere et al.
performance [2]. In spite of this disadvantage, a proper mix design can still meet the
design requirements [3–5]. In structural concrete elements, fibres are commonly added
to improve the post-cracking behaviour of concrete. The most commonly used fibres
are steel fibres, polypropylene (PP) fibres and glass fibres [6]. Previous studies [7, 8]
showed that those fibres have a limited influence on the compressive strength and
modulus of elasticity of concrete. The biggest improvement is related to the flexural
strength, fatigue and abrasion resistance, deformation capability and toughness [6]. Due
to this beneficial effect, a number of research activities have been conducted in the last
decades on fibre reinforced concrete with natural aggregates. Those studies revealed
that the fibre amount and dimensions are very important to obtain a certain post-
cracking strength [9]. For concrete with RAs, another kind of behaviour occurs. Many
researchers [10, 11] reported that the strength of RAs is commonly lower than that of
natural aggregates. Therefore the fracture process and consequently the fibre activation
in concrete with RAs can be different from that of conventional concrete with natural
aggregates. To fully understand the fracture process in fibre reinforced concrete con-
taining RAs, the following research is carried out to get a better insight in the fracture
behaviour of fibre reinforced concrete with different types of RAs.
To investigate the post-cracking behaviour of glass fibre reinforced concrete with

different types of RAs, a total of 8 concrete mixtures are manufactured, which include
two conventional and six recycled concrete mixtures. The mix composition was
obtained based on the target strength class C40/50 and the slump class S4 (>16 cm) for
the plain natural concrete mixture. The mixture comprised ordinary Portland cement
CEM I 52.5R/HES (400 kg/m3), fine and coarse aggregates, water (W/C = 0.48) and
super-plasticizer (SP). In Table 2, the detailed concrete compositions for all the con-
crete mixtures are given. In the concrete mixtures, the used aggregates are sand
0–2 mm (fineness modulus: 2.25), limestone 2–6.3 mm and coarse natural and recy-
cled aggregates 6.3–14 mm. In total, three types of recycled aggregates are used: high
quality recycled concrete aggregates (RCAs+), normal recycled concrete aggregates
(nRCAs) and mixed aggregates (MAs). Those aggregates are produced by a local
recycling plant in Flanders, Belgium. In Table 1, the physical and mechanical prop-
erties of the aggregates are given. The physical properties, including the oven-dry
density (qrd), the saturated surface-dried (SSD) particle density (qssd), the apparent
particle density (qa) as well as the water absorption after 24 h (WA24) are obtained
according to the standard EN 1097-6. In addition, the resistance to fragmentation
(LA) of the aggregates is determined according EN 1097-2. In the fibre reinforced
concrete mixtures, RA glass fibres with a content of 0.75 V% are used. These fibres are
the CEM-FIL MinibarsTM produced by OWENS CORNING. The glass fibres have a
fibre length of 43 mm, diameter of 0.70 mm and a Young’s modulus of 42 GPa.
Post-cracking Behaviour of Glass Fibre Reinforced Concrete 335
Table 1. Physical and mechanical properties of the aggregates

qrd (kg/m3) qa (kg/m3) qssd (kg/m3) WA24 (%) LA (−)
Sand 0-2 2580 2600 2590 0.40 –
Limestone 2-6.3 2680 2710 2690 0.45 –
Limestone 6.3-14 2630 2680 2650 0.81 22
RCA+ 6.3-14 2380 2660 2500 4.42 24
nRCA 6.3-14 2230 2680 2650 6.06 34
MA 6.3-14 1890 2420 2190 11.52 46
Table 2. Concrete composition.

Concrete Aggregates (kg/m3) Fibre content
mixture + (V%)
Sand Limestone Limestone RCA nRCA MA
0-2 2-6.3 6.3-14 6.3-14 6.3-14 6.3-14
NAC – ref. 759 199 777 – – – 0
RAC+ – ref. 759 199 – 703 – – 0
nRAC – ref. 759 199 – – 659 – 0
MAC – ref. 759 199 – – – 560 0
NAC – 759 199 777 – – – 0.75
0.75 V%
RAC+ – 759 199 – 703 – – 0.75
0.75 V%
nRAC – 759 199 – – 659 – 0.75
0.75 V%
MAC – 759 199 – – – 560 0.75
0.75 V%
The strength properties of each concrete mixture are examined on 12 cubes

(150 150 150 mm3) and 6 cylinders (height: 300 mm – diameter: 150 mm).
These specimens are cured in water for 28 days until test. The compressive strength test
is performed according the NBN EN 12390-3:200 - Testing hardened concrete.
Compressive strength of test specimens. The splitting tensile strength and E-modulus
are determined according to the NBN EN 12390-6 Testing hardened concrete – Part 6:
Tensile splitting strength of test specimens and the NBN B 15-203 Concrete testing –
Statical modulus of elasticity with compression, respectively. The flexural behaviour of
the different concrete mixtures is examined on beam specimens (600 150 150
mm3). In total, 24 fibre reinforced concrete beams as well as 24 unreinforced beams
were casted and tested according to the standard EN 14651: 2005 + A1:2007 - Test
method for metallic fibre concrete: Measuring the flexural tensile strength (limit of
proportionality (LOP), residual). The test set-up is shown in Fig. 1.
Fig. 1. Load-CMOD curves of the concrete mixtures
During the displacement controlled three-point bending test, the concrete beam is
subjected to a concentrated force F at the mid-span. According to the standard, a notch
in the beam is required to locate the crack plane under the bending load. In the present
test, the notch in the concrete beams is made through a wet saw. During the bending
test, the applied force F is measured as a function of the crack mouth opening dis-
placement (CMOD). Each test is stopped as the CMOD reaches a value of 4.5 mm.
During the test, not only the CMOD is measured (LVDT CMOD), but also the
deflection (LVDT deflection) of the beams is measured. This is also indicated in Fig. 1.
From the load-CMOD curve, the limit of proportionality (LOP) and residual flexural
tensile strengths can be determined. The LOP represents the maximal tensile stress at
the top of the notch. By considering a linear stress distribution, the following equation
is valued:
3FL l
f fct;L ¼ ð1Þ
2bh2sp
where: f fct;L = the limit of proportionality (LOP) (MPa); FL = the load corresponding to
the LOP (N); l = span length (mm); b = width of the specimen (mm); hsp = distance
between the tip of the notch and the top of the specimen (mm).
The residual flexural tensile strengths corresponding to different CMODj values can
be computed according to Eq. (2).
3Fj l
f R;j ¼ ð2Þ
2bh2sp
where: fR,j = the residual flexural tensile strength corresponding with CMODj (MPa);
Fj = the load corresponding with CMODj (N).
The purpose of adding fibres is to increase the residual flexural tensile strength. The
fibres enhance the concrete toughness by their energy absorption capacity. In the
obtained P-CMOD curve, two parts are distinguished: a pre-crack part and a post-crack
Fig. 2. Load-CMOD curves of the concrete mixtures.
part. The pre-crack part occurs before the LOP, while the post-crack part occurs after the
LOP. Hereby, the area under the pre-cracking curve and the post-cracking curve rep-
resents the pre-cracking energy (Epr) and the post-cracking energy (Epost). Generally, it
is known that the fibres mainly influence the post-crack energy.

3.1 Workability
In Table 3, an overview of the workability of the different concrete mixtures is given.
For all the unreinforced concrete mixtures, the amount of superplasticizer (SP) was
determined to obtain a S4 slump class. It seems that for the unreinforced NAC, RAC+
and nRAC 0.77 g/l is required, while for the MAC-ref. mixture an amount of 0.33 g/l
is sufficient. This might be due to the fact that the additional added water (to com-
pensate the high water absorption of the MAs) is not yet completed absorbed. When
0.75 V% CEM-FIL MinibarsTM is added, the obtained slump class is reduced to S3.
Only for the MAC-0.75 V%, a S4 slump class can still be obtained. Generally, it seems
that the used amount of glass fibres has an important impact on the concrete
workability.
Table 3. Workability of the concrete mixtures.

NAC RAC+ nRAC MAC
ref. 0.75 V% ref. 0.75 V% ref. 0.75 V% ref. 0.75 V%
Amount of SP (g/l) 0.77 0.77 0.77 0.33
Slump (cm) 18 10 16 11 19 13 20 16
Slump class S4 S3 S4 S3 S4 S3 S4 S4
3.2 Strength and Flexural Behaviour

The mechanical properties of the reference and the
fibre reinforced concrete mixtures are summarized in
Table 4. Earlier research [7, 8] reported that the fibre
addition has only a small influence on the mechan-
ical properties of conventional concrete. For the
recycled concrete mixtures similar results are found.
By adding 0.75 V% glass fibres to the recycled
mixtures, a small increase of the compressive
strength is observed for the concrete mixtures with
100 V% RCAs+ and nRCAs. This improvement is
resulted from the bridging effect of the fibres when
micro-cracks occur. Only the fibre reinforced mix-
ture containing MAs shows a small decrease in
compressive strength. Despite this decrease, the
highest (splitting) tensile strength improvement is
observed for the MAC-mixture. The splitting
strength increases from 2.2 (MAC-ref.) to 3.3 MPa
(MAC-0.75 V%). In addition, it is observed that all
fibre reinforced concrete cubes do not collapse at the
maximum load. This indicates an enhanced ductile
behaviour of the concrete due to the glass fibres.
Despite the small increase of the mechanical
concrete properties, a significant improvement is Fig. 3. ANOVA-test
observed in the post-cracking behaviour of the fibre
reinforced concrete mixtures with 100 V% nRCAs and
MAs, compared to NAC-0.75 V%. In Fig. 2, the individual P-CMOD curve of the four
fibre reinforced concrete mixtures is shown. It seems that there is no real difference in
the flexural tensile strength (f fct;L ) for the fibre reinforced concrete mixtures with natural
aggregates, RCAs+ and nRCAs (see also Table 5). Only for the MAC-0.75 V% mix-
ture, the flexural tensile strength is 3.3 MPa. From the obtained P-CMOD curves, it is
clearly that a higher post-cracking behaviour is obtained when aggregates with a higher
LA-value are used. The statistical parametric mapping method (SPM) [12] in which an
one-way ANOVA test was carried out, shows that there is a significant difference in the
load in a CMOD-range of 0.5–1.2 mm (see Fig. 3) for the four fibre reinforced con-
crete mixtures. In this test, a significance level (a) of 0.05 is used.
Table 4. Mechanical concrete properties.

Cubical compressive Splitting tensile Modulus of elasticity
strength (MPa) [st. dev] strength (MPa) [st. dev] (GPa) [st. dev]
NAC-ref. 59.6 [2.7] 3.6 [0.2] 38.6 [2.3]
NAC-0.75% 59.3 [1.9] 3.9 [0.2] 36.6 [2.6]
RAC+-ref. 54.6 [1.0] 3.6 [0.2] 33.9 [1.3]
RAC+-0.75 V% 59.8 [2.3] 3.8 [0.4] 36.0 [1.5]
nRAC-ref. 52.4 [1.4] 2.9 [0.6] 33.4 [1.5]
nRAC-0.75 V% 55.9 [1.6] 3.5 [0.2] 35.9 [3.6]
MAC-ref. 45.2 [0.8] 2.2 [0.4] 32.9 [0.4]
MAC-0.75 V% 44.3 [1.0] 3.3 [0.3] 34.1 [0.5]
Table 5. (Residual) Flexural tensile strength according EN 14651:2005.

f fct;L (MPa) [st. dev] fR,j at prescribed CMODj values Energy (J)
(MPa) [st. dev]
0.5 mm 1.5 mm 2.5 mm 3.5 mm
NAC-0.75 V% 4.2 [0.6] 3.4 [0.4] 3.3 [0.3] 2.7 [0.3] 2.1 [0.2] 38.7 [3.5]
+
RAC -0.75 V% 4.1 [0.3] 3.2 [0.4] 3.3 [0.4] 2.8 [0.3] 2.2 [0.2] 39.9 [4.0]
nRAC-0.75 V% 4.2 [0.2] 3.8 [0.2] 3.8 [0.3] 3.0 [0.3] 2.3 [0.3] 43.3 [3.5]
MAC-0.75 V% 3.3 [0.3] 3.9 [0.3] 3.9 [0.3] 3.0 [0.1] 2.3 [0.0] 43.3 [2.2]
The higher P-CMOD curves of the nRAC-0.75 V% and MAC-0.75 V% mixture

definitely results in a higher amount of absorption energy. The total energy (Epre +
Epost) increases from 38.7 J to 43.3 J for the MAC-0.75 V% mixture. According to the
FIB Model Code 2010 [13], traditional rebars can partly be replaced if Eq. 3 and 4 are
satisfied:
f R1k
[ 0:4 ð3Þ
f Lk
f R3k
[ 0:5 ð4Þ
f R1k
in which fLk represents the characteristic flexural tensile strength at LOP, while fR1k and
fR3k are the characteristic residual strength at 0.5 mm CMOD and 2.5 mm CMOD,
respectively.
Due to the higher post-cracking energy in the range of 0.5–1.2 mm CMOD, a
higher residual flexural tensile strength is observed for Serviceability Limit State (SLS),
which corresponds to a CMOD-value of 0.5 mm. This increases from 3.4 to 3.9 MPa
for the fibre reinforced concrete mixture with MAs. At ULS (CMOD = 2.5 mm), only
a small difference in the obtained residual tensile strength is observed (see Table 5).
Therefore, it is important to notice that Eq. (4) becomes more critical for the recycled
concrete mixtures. By considering a lognormal distribution, the characteristic strength

values can be determined for all parameters (f fct;L and fRj), according to Eqs. (5)–(7).

f k ¼ exp f ln;m kn rln ð5Þ
1 Xn
f ln;m ¼ ln f i ð6Þ
n i
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 Xn 2
rln ¼ ln f i f ln;m ð7Þ
n1 i
In the above mentioned formulas, the kn-value depends on the number of samples and
the prior knowledge on the coefficient of variation. Those kn-values are summarized in
table D1 from the standard EN 1990. For five and six specimens the value is 2.33 and
2.18, respectively. In Table 6, the corresponding characteristic flexural tensile strengths
are given. It seems that although a higher residual tensile strength, Eq. 3 and Eq. 4 are
still satisfied. Therefore the traditional rebars can still (partly) be replaced when
recycled aggregates are used. This is even the case for the MAC-mixtures.
Table 6. Characteristic (residual) flexural tensile strength according EN 14651:2005.

f fct;Lk fR,jk at prescribed CMODj values Equation 3 Equation 4 FRC
(MPa) (MPa) class
0.5 mm 1.5 mm 2.5 mm 3.5 mm
NAC- 3.2 2.8 2.1 1.7 1.3 ✓ ✓ 3a
0.75 V%
RAC+- 3.4 2.4 2.5 2.1 1.7 ✓ ✓ 3b
0.75 V%
nRAC- 3.8 3.4 3.0 2.3 1.6 ✓ ✓ 3a
0.75 V%
MAC- 2.8 3.1 3.1 2.6 2.1 ✓ ✓ 2.5b
0.75 V%
The LA-values, shown in Table 1, indicate that the aggregate strength of the
nRCAs and MAs are significantly lower compared the natural and high quality recy-
cled aggregates. Due to this, it is expected that the crack propagation, and therefore the
fibre activation is also influenced by the aggregate strength. By considering the linear
asymptotic superposition assumption, Tada et al. [14] proposed formula (8) and (9) to
calculate the equivalent-elastic crack length:
6PSa
CMOD ¼ VI ðaÞ ð8Þ
D2 BE
Where: a = (a + H0)/(D + H0); P = the load at specific CMOD-value; S = specimen

loading span (500 mm); a = crack length (mm); D = beam depth (150 mm); B = beam
width (150 mm); E = modulus of elasticity (GPa); H0 = thickness of the clip gauge
holder (= 18 mm). For S/D = 4, the geometrical function VI(a) is given as follows:
0:66
VI ¼ 0:76 2:28a þ 3:87a22 :04 a3 þ ð9Þ
ð1 aÞ2
With the above mentioned equations and the obtained P-CMOD curves, the crack
length propagation for the different concrete mixtures can be determined. This is
visualized by the blue line in Fig. 4. In this figure, the crack length at LOP (PLOP) is
shown in red, while the crack length at the initial cracking load (Pcrack) and post-
cracking peak (PMOR) is visualized in green and yellow. The used P-CMOD values at
initial cracking and ultimate cracking are also given in Table 7. The initial cracking
point is defined as the point where the first cracking occurs. This is commonly reached
earlier than the flexural strength of the concrete matrix. The initial cracking point is
where non-linearity in the load-deflection curve becomes obvious or the inclination of
P-CMOD curve reduces significantly. In Fig. 4, the mean value of six beams is given,
while the colored zone indicates the range [l − r, l + r]. It seems that for the high
Fig. 4. Crack length at initial cracking, ultimate and post peak load.
Table 7. Load-CMOD values at initial and ultimate cracking point

Initial cracking point Ultimate cracking point
Load (kN) CMOD (mm) Load (kN) CMOD (mm)
NAC-0.75 V% 7.4 [1.6] 0.010 [0.002] 13.1 [1.8] 0.049 [0.028]
RAC+-0.75 V% 6.8 [2.6] 0.009 [0.002] 12.6 [1.0] 0.067 [0.051]
nRAC-0.75 V% 6.8 [0.6] 0.008 [0.001] 12.9 [0.8] 0.034 [0.007]
MAC-0.75 V% 4.1 [0.8] 0.007 [0.001] 10.3 [1.0] 0.081 [0.007]
quality recycled concrete mixtures a longer crack length (and scattering) is obtained at
the different load stages. Due to the fact that the crack length of the natural and recycled
mixtures does not differ much, no real difference is observed in the P-CMOD curves.
By using recycled mixed aggregates, a crack length of 78.8 mm is obtained at the LOP-
value. This is caused by the lower aggregate strength which results in a long crack
length. Therefore, many fibres are directly activated after concrete cracking and no
force-drop is observed in the P-CMOD curve (see Fig. 3). Also for the nRAC mixture,
a higher post-cracking strength is observed; although only a crack length of 49.4 mm is
noticed at LOP. This can be caused by the higher ductile behaviour of the nRAC-
0.75 V% mixture [15].
4 Conclusions
In this study, the mechanical behaviour of glass fibre reinforced concrete with different
types of RAs is studied. In addition to the tests on compressive strength, splitting
tensile strength and modulus of elasticity, three-point bending tests on notched beam
specimens are also carried out according to the EN 14651 to determine the post-
cracking behaviour of the beams. Within the scope of this study, the following con-
clusions can be drawn:
• The CEM-FIL minibarsTM AR glass fibres have a minor, but beneficial influence on
the mechanical properties of the recycled concrete mixtures. The most significant
improvement is observed for the splitting tensile test of the concrete with mixed
recycled aggregates. The splitting strength increases from 2.2 to 3.3 MPa when the
fibre amount increases from 0 to 0.75 V%.
• The three point bending tests on the fibre reinforced concrete specimens show that
there is a significant difference in the obtained P-CMOD curves, especially in the
CMOD-range between 0.5–1.2 mm.
• The fibre reinforced natural mixtures, as well as the fibre reinforced recycled
concrete mixtures show that they meet the FIB Model code requirements to (partly)
replace the traditional reinforcement.
• The crack length (at LOP) for the RAC+-0.75 V% and MAC-0.75 V% mixtures are
higher compared to the natural aggregate fibre reinforced mixture. For the fibre
reinforced MAC-mixture a crack length of 78.8 mm was obtained while the NAC-
0.75 V% mixture has a crack length of only 57.4 mm at LOP.
• Despite smaller crack length of the nRAC-0.75 V% (49.4 mm), a higher P-CMOD
curve is obtained. This can be caused by the higher ductility of those concrete
mixture.
Acknowledgements. This research work has been supported by a start-up funding at KU

Leuven – University of Leuven, Belgium (No. STG/16/011), and the Lvnong Chair in Con-
struction Waste Recycling at KU Leuven, Belgium. The financial supports are gratefully
acknowledged. The authors would also like to express their thanks to the technical staff Bernd
Salaets and technical manager Kurt Scherpereel at the laboratory of Department of Civil Engi-
neering, KU Leuven, as well as Owens Corning for providing the glass fibres, O.B.B.C nv for
delivering the RAs and Willy Naessens Group for supplying the natural aggregates.
References
1. Safiuddin, M., Alengaram, U.J., Rahman, M.M., Salam, M.A., Jumaat, M.Z.: Use of
recycled concrete aggregate in concrete: a review. J. Civ. Eng. Manag. 19, 796–810 (2013)
2. Otsuki, N., Miyazato, S., Yodsudjai, W.: Influence of recycled aggregate on interfacial
transition zone, strength, chloride penetration and carbonation of concrete. J. Mater. Civ.
Eng. 15, 443–451 (2003)
3. Levy, S.M., Helene, P.: Durability of recycled aggregates concrete: a safe way to sustainable
development. Cem. Concr. Res. 34, 1975–1980 (2004)
4. Malešev, M., Radonjanin, V., Marinković, S.: Recycled concrete as aggregate for structural
concrete production. Sustainability 2, 1204–1225 (2010)
5. Tu, T.-Y., Chen, Y.-Y., Hwang, C.-L.: Properties of HPC with recycled aggregates. Cem.
Concr. Res. 36, 943–950 (2006)
6. ACI Committee 544: ACI 544.1R-96: Report on Reinforced Concrete (Reapproved). ACI
Struct. J. 96, 1–66 (2002)
J. Mater. Civ. Eng. 19(5), 385–392 (2007)
8. Alhozaimy, A.M., Soroushian, P., Mirza, F.: Mechanical properties of polypropylene fiber
reinforced concrete and the effects of pozzolanic materials. Cem. Concr. Compos. 18, 85–92
(1996)
9. Frank, H.P.: Fibre Cements and Fibre Concretes, D. J. Hannant. Wiley-Interscience,
New York, 219 pp. (1978). J. Polym. Sci. Polym. Lett. Ed. 17, 464–465 (1979)
10. López-Gayarre, F., Serna, P., Domingo-Cabo, A., Serrano-López, M.A., López-Colina, C.:
Influence of recycled aggregate quality and proportioning criteria on recycled concrete
properties. Waste Manag. 29, 3022–3028 (2009)
11. Silva, R.V., De Brito, J., Dhir, R.K.: Properties and composition of recycled aggregates from
construction and demolition waste suitable for concrete production. Constr. Build. Mater. 65,
201–217 (2014)
12. Friston, K.J., Karl, J., Ashburner, J., Kiebel, S., Nichols, T., Penny, W.D.: Statistical
Parametric Mapping: The Analysis of Funtional Brain Images. Elsevier/Academic Press
(2007)
13. FIB 2011 Model Code 2010 FIB Model Code Concr. Struct. 114–148 (2010)
14. Tada, H., Paris, P.C., Irwin, G.R.: The Stress Analysis of Cracks Handbook (1985)
15. Zhang, J., Li, V.C.: Simulation of crack propagation in fiber-reinforced concrete by fracture
mechanics Cem. Concr. Res. 34, 333–339 (2004)
Long-Term Properties
A Computational Sectional Approach
for the Flexural Creep Behavior
of Cracked FRC
Rutger Vrijdaghs1(&), Marco di Prisco2, and Lucie Vandewalle1

1
Department of Civil Engineering, KU Leuven, Leuven, Belgium
rutger.vrijdaghs@kuleuven.be
2
Department of Structural Engineering, Politecnico di Milano, Milan, Italy
Abstract. This paper presents a computational model to calculate and predict

the flexural creep behavior in a cracked fiber reinforced concrete (FRC) section.
The proposed model is based on uniaxial creep data and consists of three steps.
In the first step, an inverse analysis algorithm is presented to model the
monotonic bending behavior of a notched FRC beam in accordance with EN
14651. A simplified and numerically optimized method is compared to exper-
imental data and a good agreement is found.
In a second step, the unloading behavior of the beam is taken into account.
Calibrated on experimental data, the model is able to accurately and precisely
predict the unloading behavior. Further validation comes from the location of
the neutral axis, and the deformation profile.
In a third step, the flexural creep behavior is predicted based on the results in
the second step. The creep data is supplied in uniaxial form, which allows
greater applicability across various FRC mixtures. The proposed approach is
able to take into account stress redistribution following fiber fracture. Further-
more, the time-dependent effects of the stress redistributions are also accounted
for. As such, the model is able to predict tertiary creep and structural failure
under sustained loading.
Keywords: Sectional analysis Creep of FRC Polymeric FRC Tensile and

bending creep
1 Introduction
Fiber reinforced concrete (FRC) is a composite material in which fibers are added to the
fresh concrete mix [1, 2]. These fibers are used to improve the properties of the
concrete in the fresh or hardened state. In structural applications, fibers can partially or
totally replace the traditional reinforcement. For these purposes, the fibers provide an
enhanced post-cracking tensile strength in the hardened state by bridging crack faces
[3]. Commercially available fibers can be made from a number of different materials:
steel, glass, synthetic and natural being the most common types [4].
Until recently, the research effort has been focusing on the properties of FRC in the
fresh and hardened state [5, 6]. It has been found that the inclusion of fibers decreases
plastic shrinkage cracking [7] and flowability [8] in the fresh state. In the hardened
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_32
348 R. Vrijdaghs et al.
state, fibers provide a post-cracking strength to the matrix [9–11]. However, investi-
gations on the long-term structural properties of FRC have only recently become a
research focal point. The recently published Model Code 2010 [12] acknowledges that
time-dependent effects may significantly alter the behavior of structural elements, both
in the serviceability limit state and in the ultimate limit state. In the former, creep may
cause unwanted and unacceptable deflections of FRC beams [13, 14], whilst in the
latter, the residual load-bearing capacity of elements may be significantly lower than
the design value based on the short-term strength assessment. However, the Model
Code (MC10) does not provide design rules to take long-term behavior into account. It
is clear that further research on the long-term structural properties of FRC is required.
One of these time-dependent phenomena is creep. While the creep of cracked FRC
has been under investigation for over two decades [15–18], much uncertainty still
remains on how to model the behavior. In this paper, the creep deformations of a
cracked, notched FRC beam are predicted based on uniaxial tensile creep data.
A sectional approach is presented, consisting of three steps, which gradually increase
the complexity and capabilities of the model. In the first step, a monotonic bending test
is simulated using uniaxial constitutive laws. Secondly, unloading of the sample is
taken into account, and in the third step, uniaxial creep data is used to predict time-
dependent flexural crack opening. The approach is validated with experimental data in
each step.
2 Experimental Program and Results
The experimental program focusses on a normal strength concrete (fcm,cube = 43 MPa)

with 1 V% of polypropylene fibers (diameter: 0.7 mm, length: 55 mm) and consists of
a series of short-term characterization and long-term creep tests. Three different tests
are performed: monotonic bending according to EN 14651 [19], unloading from var-
ious points in the bending tests and uniaxial tensile creep tests, refer to Table 1. Each
test setup and result is discussed in this section.
Table 1. Number of specimens per test type

Test type Number of specimens [-]
Monotonic bending acc. to EN 14651 12
Unloading-reloading in bending 6
Uniaxial tensile creep 14
2.1 Monotonic Bending Test

The monotonic bending test is performed according to EN 14651 [19]. The test setup is
shown in Fig. 1, and the reader is referred to the European Standard for all details
concerning the test speed, specimen geometry and strength calculations. The results are
shown in a stress-CMOD curve on the right, with indicating of the average,
A Computational Sectional Approach for the Flexural Creep Behavior 349
characteristic and experimental scatter values. According to MC10, the FRC is clas-
sified as 2d, and can be used in structural applications.
6
F
5
[MPa]
150
LVDT deflection mm 3
Stress
2 Exp. scatter
LVDT CMOD Exp. mean
500 mm f
1 R,i,m
f
650 mm R,i,k
0
0 0.5 1 1.5 2 2.5 3 3.5 4
CMOD [mm]
Fig. 1. (left) bending setup (right) bending results
2.2 Unloading-Reloading Bending Test

The unloading-reloading test uses the same specimens and test geometry as the
monotonic bending tests, but the imposed CMOD rate is – both for the unloading as
well as the reloading phase – set equal to 0.2 mm/min. Unloading cycles are imposed at
CMOD = 0.1, 0.2, 0.3, 0.4, and 0.5 mm. After the 5 cycles, a monotonic test speed is
imposed. Furthermore, the mid-span deflection is no longer measured but 5 LVDTs are
attached to the side of the FRC prism to measure the deformation profile over the
height. The test setup is shown in Fig. 2 (right), with the global and zoomed experi-
mental results shown in the (middle) and (right) respectively. It is noted that the
average monotonic response yields generally higher (post-peak) strengths, as it is
expected that the cycles induce additional damage in the specimen.
Global response Zoom on cyclic response

4.5 4.5
Cyclic scatter
Cyclic mean
Monotonic mean
3 3
[MPa]
[MPa]
Stress
Stress
1.5 1.5
Cyclic scatter
Cyclic mean
Monotonic mean
0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 0.3 0.4 0.5
CMOD [mm] CMOD [mm]
Fig. 2. (left) cyclic bending setup with LVDT (middle) global results (right) zoom on cycles
2.3 Uniaxial Tensile Creep

In order to assess the uniaxial tensile creep deformations, a finite element model was
built in which the fibers are modelled separately. The aim of the numerical model is to
describe the creep deformations for a time period of 50 years. Further details about the
construction and analysis of the model can be found in literature [20], but for the
purpose of this paper, it is sufficient to show the creep coefficient evolution in Fig. 3.
This figure represents the uniaxial tensile creep coefficient determined from 25 different
numerical analyses.
1.4
1.35
1.3
1.25
[-]
1.2
basic
1.15
1.1
1.05
1
0 3 6 9 12
Log time [s]

10
Fig. 3. Uniaxial tensile creep coefficient, determined from 25 FEM simulations
3 Sectional Analysis: Monotonic Behavior
The sectional analysis considers a discretization of the notched sections: the cracked
section is divided into 125 1 mm high elements, each assumed to operate in uniaxial
compression or tension. As such, the approach can be considered a fictitious crack
model, where the cracked section functions as a plastic hinge between two rigid bodies.
All deformations are smeared over a length equal to the height of the section, such that
the strain e can be related to the deformation d: e = d/125. In the sectional approach,
the deformation and stress profile, d(y) and r(y) respectively, are calculated such that
the horizontal and bending moment equilibria are fulfilled, given a certain constitutive
law in compression and tension.
3.1 Constitutive Laws in Compression and Tension

In compression, the stress-strain curve of FRC is described by a Thorenfeldt model,
which is based on the Young’s modulus and the compressive strength.
In tension, two different approaches are implemented and compared. Firstly, the
crack width-stress relation is based on the MC10 model, extended with the model
proposed in [21]. In this model, the constitutive law can be directly computed from the
compressive strength and the post-cracking parameters, measured in the EN 14651 test
(fR1 and fR3). Secondly, a multi-linear crack width-stress relation is proposed, and in a
numerical optimization scheme, the coordinates of each point of this curve are varied
such that the resulting bending response corresponds as good as possible with the
experimentally obtained result, as determined by the minimization of the sum of the
absolute error between the measured and calculated bending moment, refer to Eq. (1).

X X

error ¼ b h ðyj rj;i Þ Mexp;i ð1Þ
i
j

In Eq. (1), rj,i is the stress in the jth element corresponding at the ith CMOD value
in the bending test. The height on an element h and the width of the section b are
known, i.e. 1 mm and 150 mm respectively. In the optimization scheme, the location
of the neutral axis yNA is determined such that the horizontal force error DH is less than
2 N, in accordance with Eq. (2).
X
DHi ¼ b h rj;i ð2Þ
j
3.2 Results of the Monotonic Bending

The results of the sectional approach for the monotonic bending test are shown in
Fig. 4, with a comparison of the experimental result (as shown before) and the MC10
model as well as the numerical optimized model, for the global response (CMOD =
4 mm, left) and a zoomed in part up to CMOD = 0.5 mm (right). A very good
agreement is found between the experimental and calculated results, especially with the
optimized constitutive tensile law. This is expected as the latter has more degrees of
freedom, i.e. post-peak points in the stress-crack width curve, to achieve a better fit.
6 6
5 5
4 4
[MPa]
[MPa]
3 3
Stress
Stress
2 2
Scatter Scatter
Mean Mean
1 MC10/SC 1 MC10/SC
Numerical optimization Numerical optimization
0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 0.3 0.4 0.5
CMOD [mm] CMOD [mm]
Fig. 4. Comparison of the experimental and predicted monotonic bending behavior (left) global
(right) zoomed
4 Sectional Analysis: Unloading Behavior
The next step is the analysis of the unloading behavior. The aim is, as before, to
calculate the stress and deformation profile, rLR(y) and dLR(y) respectively, as a
function of the height y and of the applied load ratio LR, i.e. the load w.r.t. the
maximum bearable force. To ensure residual deformation after complete unloading, the
Euler-Bernoulli beam theory is abandoned and a bilinear deformation profile, rather
than a linear, is assumed. Additionally, the uniaxial constitutive laws are extended to
include unloading and damage laws.
4.1 Extension of the Constitutive Laws

The uniaxial compressive behavior is assumed linear-elastic, as the compressive stress
remains below 45% of the strength for CMOD < 0.5 mm. It is assumed in MC10 that
for these low stress levels, concrete remains elastic, such that no modifications are
needed for the compressive constitutive law.
In tension, especially in the cracked zone of the section, unloading incurs irre-
versible deformations, which needs to be taken into account. Therefore, the unloading
stiffness in uniaxial tension cannot be equal to the Young’s modulus E0. The unloading
stiffness E(e) depends on the scalar damage evolution D(e) in accordance with Eq. (3).
E ðeÞ ¼ E 0 ð1 D ðeÞÞ ð3Þ
The damage function is a monotonically increasing function, bound between 0 and

1, representing zero and complete loss of stiffness. In this approach, the damage
function is further limited to ensure that tensile elements remain in tension during the
unloading phase.
4.2 Implementation of the Inverse Analysis

The inverse analysis aims to find the bilinear deformation profile such that the hori-
zontal and bending equilibrium is fulfilled, whilst simultaneously finding the optimal
damage function. The input for the algorithm is every (CMOD, M)i point from the
experimental unloading branches and the constitutive laws in compression and tension.
The total error DEi is minimized per unloading point i, as defined in Eqs. (4–6).
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2
DEi ¼ DH i þ DM i ð4Þ
!
X
DH i ¼ bh rj;i =Fi ð5Þ
j
!
X
DM i ¼ bh ðyj rj;i Þ Mexp;i =Mi ð6Þ
j
4.3 Results of the Unloading Analysis

The error defined in Eqs. (4–6) is calculated for all points on the unloading branches in
Fig. 2, and is expressed as a percentage. The analysis shows that 99% of all points have
a relative horizontal and bending moment error within ± 1% and ± 5%, with an
average error of 0.0009% and 0.0073%, respectively, indicating a very good close
agreement with the experimental data. In Fig. 5, the damage evolution in terms of the
tensile deformation is shown. A clear and steep increase of the damage can be seen,
even at very low crack widths: at a crack width of 0.1 mm, over 90% of the stiffness is
already lost.
0.8
0.6
10%/90% quantile
Damage [-]
Min/max
Median
0.4
0.2
0
-0.4 -0.3 -0.2 -0.1 0 0.1
Deformation [mm]
Fig. 5. Evolution of the scalar damage parameter
As the unloading branches are input for the model, they cannot be used for the
validation. However, the 5 LVDTs on the side of the FRC prism can be used to validate
the calculated deformation profile, as well as the location of the neutral axis. In Fig. 6,
the calculated deformation profile (solid lines) at full load (LR = 100%, black) and half
load (LR = 50%, gray) is compared against the experimentally measured deformations
at various y heights (circle markers). It should be noted that the measured deformations
are not used in the calibration of the model, indicating that a very good agreement can
be reached. Finally, the unloading deformation profile is assumed bilinear, as to allow
for residual deformations, but at full load, the bilinear curve converges to a linear curve,
representing an Euler-Bernoulli beam. As such, the unloading sectional analysis can be
understood as an extension of the classical beam theory.
CMOD = 0.1 mm CMOD = 0.2 mm CMOD = 0.3 mm
150 150 150
100 100 100
50 50 50
0 0 0
-0.1 -0.05 0 0.05 -0.2 -0.1 0 -0.3 -0.2 -0.1 0
CMOD = 0.4 mm CMOD = 0.5 mm
150 150
LR = 100 % model
LR = 100 % exp.
LR = 50 % model
100 100
LR = 50 % exp.
50 50
0 0
-0.4 -0.2 0 -0.4 -0.2 0
Fig. 6. Comparison of the measured and predicted deformation profiles during unloading
5 Sectional Analysis: Creep Deformations
The final step is the extension of the approach to include creep effects. The start point
for the time-dependent analysis is the stress and deformation profile at a certain LR,
which was obtained in the previous step of the analysis and is kept constant in time.
The main assumption of the creep analysis is that the total deformation d(t, y) as a
function of time t and height y can be written as the linear superposition of instanta-
neous deformations Dd multiplied by their creep coefficient /, as shown in Eq. (7).
X
dðt; yÞ ¼ /k ðt; y; tk Þ Ddk ðtk ; yÞ ð7Þ
k
5.1 Extension of the Constitutive Laws

In Eq. (7), k refers to a certain stress redistribution at the moment tk incurring an
instantaneous deformation Ddk. The stress redistribution might be necessary due to two
reasons: (1) a certain element in the discretized section cannot take any further force
due to excessive straining (2) a time-varying load is applied. In this approach, the load
is assumed constant, so that (2) is not allowed to happen. On the other hand, the effect
of excessive straining is investigated by (artificially) limiting the uniaxial tensile
deformation capacity of the cracked concrete. Note that such an effect is not experi-
mentally or numerically encountered, and the analysis is purely from illustrative pur-
poses to highlight the capabilities of the model. Additionally, excessive compressive
straining is neglected.
Each instantaneous deformation Ddk can be related to an instantaneous stress
change in each element, depending on whether this element is in compression or
tension. After each stress redistribution, a new equilibrium is sought. Equations (8–9)
show the incremental constitutive laws in compression and tension.

Drj ti ; yj ¼ E0 Dej ti ; yj ð8Þ

Drj ti ; yj ¼ E0 1 D dj ti1 ; yj Dej ti ; yj ð9Þ
The compressed zone is assumed visco-elastic, while the previously defined scalar
damage function is included in the cracked tensile law. Note that this incremental
description of the constitutive laws does not take into account the change in defor-
mation due to creeping, and that the assumption of a bilinear deformation profile is
abandoned. This is a significant simplification in the model, and further work is
underway to include the interaction between creeping and continuous stress changes in
the cracked section.
Finally, the creep coefficients are determined for the compressive and tensile zone
separately. For the former, the evolution according to MC10 is adopted for undamaged
linear visco-elastic concrete. As shown previously, the creep evolution of the cracked
tensile zone is determined from a two-phased numerical analysis. Both creep coeffi-
cients are shown in Fig. 7.
1.4
4
1.35
3.5
1.3
[-]
3 1.25
[-]
Creep coefficient
1.2
2.5
basic
1.15
2
1.1
1.5
1.05
1 1
1 s. 1 m. 1 h. 1 d. 1 m. 1 y. 50 y. 0 3 6 9 12
Time Log time [s]

10
Fig. 7. Evolution of the creep coefficient in (left) compression (right) cracked tension
5.2 Results of the Creep Analysis

The sectional analysis takes thus input from the constitutive laws, an initial deformation
and stress profile and the creep coefficient evolution to determine flexural creep in a
cracked section from uniaxial creep descriptions.
1.4
Min/max
1.2 10%/90% quantile

Mean
1
CMOD [mm]
0.8
0.6
0.4
0.2
1 s. 1 m. 1 h. 1 d. 1 m. 1 y. 50 y.
Time
Fig. 8. CMOD evolution of a cracked section
The global behavior of a flexural creep test can be expressed as the CMOD evo-
lution as a function of time. As previously highlighted, the effect of excessive tensile
straining is investigated by limiting the allowable tensile strain in the cracked zone.
This value is (arbitrarily) chosen as the maximum calculated tensile strain in the
cracked section before unloading, but the model allows for any limit value to be
imposed (or none at all), based on experimental data or numerical analysis. The evo-
lution of the CMOD for a 50 year sustained load at LR = 50% is shown in Fig. 8 for
150 different simulations.
50 0.9
Min/max Min/max
0.8
40 10%/90% quantile 10%/90% quantile
Mean Mean
0.7
30
[MPa]
[MPa]
0.6
T
C
20
0.5
10
0.4
0 0.3
1 s. 1 m. 1 h. 1 d. 1 m. 1 y. 50 y. 1 s. 1 m. 1 h. 1 d. 1 m. 1 y. 50 y.
Time Time
Fig. 9. Evolution of the maximum (left) compressive and (right) tensile stress in the section
While most simulations indicate a relatively modest CMOD growth, the majority
exceeds the initial crack width of 0.5 mm after 50 years under simulated loading. Since
the model also accounts for stress redistributions when the limit strain is encountered,
the maximum compressive and tensile stress in the cracked zone can be evaluated as
well. The evolution of both is shown in Fig. 9.
The compressive stress evolution (left) can exceed the compressive strength of the
material for some simulations, similar to the tensile stress (right). The exceedance of
the strength values indicates structural failure. This is an important result of the model,
as it indicates that it is capable not only to predict the crack width growth (even under
various simplifications), but also to predict structural failure under sustained loading.
Further work is still ongoing on removing the various simplifications of the three-
stepped model and to extend its capabilities.
6 Conclusions
In this paper, the results of a sectional approach to model flexural creep based on
uniaxial creep data is presented. The sectional approach considers the cracked section
of the EN 14651 beam in a plastic hinge formulation and consists of three steps:
monotonic loading, unloading, and creep. The model is validated with experimental
data, both on a global scale (stress-CMOD curves) as well as on a local scale (de-
formation profiles). The main conclusions are summarized as follows:
• A fast algorithm is implemented to calculate and predict the monotonic bending
behavior based on simplified uniaxial constitutive laws. Furthermore, a uniaxial
tensile law optimization scheme is implemented to extend the MC10 model, which
further improves the agreement between the predicted and experimental result.
• The proposed sectional approach can also predict the unloading behavior of the
cracked sections, provided the uniaxial laws are extended to include unloading. This
is implemented through the use of a scalar damage function and a simplified bilinear
deformation profile under unloading. The proposed algorithm achieves low errors in
the equilibrium equations and the deformation profile is validated against measured
deformations along the height of the beam.
• Finally, the uniaxial creep behavior is included in the analysis, to predict flexural
creep from uniaxial creep data. In compression, the creep deformations are based on
the viscoelastic concrete behavior assumed in MC10. In tension, a two-phased
numerical model predicts the time-dependent crack opening input for the analysis.
The results indicate a slow CMOD increase under a 50 year sustained load. More
fundamentally, the model is also capable to predict structural failure.
Further works remains in removing as much as possible the various assumptions in
the different steps, but the proposed model highlights important structural aspects of
cracked FRC under sustained loading.
References
1. Balaguru, P., Shah, S.P.: Fiber-Reinforced Cement Composites. Mcgraw-Hill, Texas (1992)
2. Bentur, A., Mindess, S.: Fibre Reinforced Cementitious Composites, 449 p. Elsevier Science
Publishers Ltd, London (1990)
4. Daniel, J.I.: ACI Committee 544. State-of-the-Art Report on Fiber Reinforced Concrete.
American Concrete Institute (2002)
5. di Prisco, M., Felicetti, R., Plizzari, G.A. : Fibre-reinforced concrete. In: RILEM
Proceedings of the 6th RILEM Symposium (BEFIB 2004), PRO39, BEFIB 2004. RILEM
Publications S.A.R.L., Bagneux (2004)
6. Rossi, P., Chanvillard, G.: Fibre reinforced concrete. In: RILEM Proceedings of the 5th
RILEM Symposium (BEFIB 2000), PRO15, BEFIB 2000. RILEM Publications S.A.R.L.,
Bagneux, France
7. Grzybowski, M., Shah, S.P.: Shrinkage cracking of fiber reinforced concrete. ACI Mater.
J. 87(2), 138–148 (1990)
8. Pasini, F., et al.: Experimental study of the properties of flowable fiber reinforced concretes.
In: di Prisco, M., Felicetti, R., Plizzari, G.A. (eds.) 6th International RILEM Symposium on
Fibre Reinforced Concretes, pp. 279–288. RILEM Publications SARL (2004)
9. Barros, J., et al.: Post-cracking behaviour of steel fibre reinforced concrete. Mater. Struct. 38
(1), 47–56 (2005)
12. fédération internationale du béton (fib), Model Code 2010 First complete draft (2010)
13. Kurtz, S., Balaguru, P.: Postcrack creep of polymeric fiber-reinforced concrete in flexure.
Cement Concrete Res. 30(2), 183–190 (2000)
14. MacKay, J., Trottier, J.F.: Post-crack creep behavior of steel and synthetic FRC under
flexural loading. In: Shotcrete: More Engineering Developments, pp. 183–192. Taylor &
Francis (2004)
15. Zerbino, R.L., Barragan, B.E.: Long-term behavior of cracked steel fiber-reinforced concrete
beams under sustained loading. Mater. J. 109(2), 215–224 (2012)
16. Kiang Hwee, T., Mithun Kumar, S.: Ten-Year Study on Steel Fiber-Reinforced Concrete
Beams Under Sustained Loads. ACI Struct. J. 102(3), 472 (2005)
17. Arango, S., et al.: A comprehensive study on the effect of fibers and loading on the flexural
creep of SFRC. In: 8th RILEM International Symposium on Fibre Reinforced Concrete
(BEFIB 2012). RILEM Publications S.A.R.L., Guimarães, Portugal (2012)
18. Babafemi, A.J., Boshoff, W.P.: Tensile creep of macro-synthetic fibre reinforced concrete
(MSFRC) under uni-axial tensile loading. Cement Concrete Composites 55, 62–69 (2015)
19. European Committee for Standardization, EN 14651 Test method for metallic fibered
concrete - Measuring the flexural tensile strength (limit of proportionality (LOP), residual),
p. 17 (2005)
20. Vrijdaghs, R., di Prisco, M., Vandewalle, L.: A numerical model for the creep of fiber
reinforced concrete. In: Hordijk, D.A., Luković, M. (eds.) FIB Symposium 2017: High Tech
Concrete, pp. 366–373. Springer, Maastricht (2017)
21. di Prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in fib Model Code 2010:
principles, models and test validation. Struct. Concrete 14(4), 342–361 (2013)
Shrinkage of Steel-Fibre-Reinforced
Lightweight Concrete
Hasanain K. Al-Naimi and Ali A. Abbas(&)
University of East London, London, UK

a.abbas@uel.ac.uk
Abstract. Long-term behaviour of steel fibre reinforced concrete remains rather

unknown and to a large extent unquantified by equations and standards. This
paper studies experimentally the free drying shrinkage of steel fibre reinforced
lightweight concrete during the first 28 days using 100 100 500 mm
beams. The coarse lightweight material tested (LYTAG) is recycled and offers
an alternative to gravel and quarry resources which are subjected to depletion in
the future. Also, this material can lead to reduction in the mass of the structure
which results in economical designs. However, LYTAG aggregate can absorb
up to 15% of its own weight in water. This makes it susceptible to drying
shrinkage both at young age and long-term due to environmental diffusion.
Shrinkage can have a detrimental effect on the concrete by inducing cracks,
creating therefore weak zones in the concrete. It is thought that fibres can have a
favourable effect on the reduction of shrinkage due to their ability to bridge
cracks. This could be vital particularly in large concrete flat slabs, joints, beams
and even columns. This project uses modern hooked-end DRAMIX 3D and 5D
fibres with different dosages Vf and number of hooks and evaluates shrinkage
for concrete with different characteristic strengths fck.
KEYWORDS: Lytag Hooked-end fibres Early-age shrinkage Drying

shrinkage Shrinkage beams Environmental diffusion Cracks
1 Introduction
Shrinkage is the time-dependent change in volume of unrestrained concrete when

tensile stresses due to contraction exceeds that of the concrete itself, although creep can
play a factor in counter acting the latter due to stress relaxation in a restrained structure
(Hossain, 2003; Havlasek 2014). Shrinkage can take place due to either internal
reactions usually before concrete hardening, responsible for by autogenous, plastic and
chemical shrinkage, or due to their surrounding environment responsible for by thermal
and drying shrinkage that leads to water evaporation through concrete pores (Neville
2011). This leads to shrinkage cracking at the surface which can negatively affect the
strength and integrity of the concrete from a young age to long-term (Mindess and
Young 1981). Shrinkage can be affected by wind, humidity, temperature, cement
fineness, water content and curing (Shoya 1979; Mehta 1994). The lightweight
aggregate LYTAG used in this work is porous and capable of absorbing water of up to
15% of its weight. Besides, no pore size reducing materials such as silica fume are used
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_33
360 H. K. Al-Naimi and A. A. Abbas
for the concrete mixed. Hence, this makes it easy for the water to evaporate out of the
concrete and act as a catalyst for drying shrinkage unlike the denser high strength
concrete (Mehta and Monteiro 1993; Koenders 1997). The lower modulus of elasticity
of lightweight concrete can also facilitate the shrinkage of concrete (Kayali et al. 1835).
Therefore, lightweight concrete is more susceptible to shrinkage than normal weight
concrete (Neville 2011). This however has been proven wrong in some other studies
where lightweight concrete showed less shrinkage than normal weight concrete (Zhang
et al. 2005).
Lightweight concrete is noted by its brittle nature in tension due to the absence of
tension toughening aggregate interlock mechanism which causes micro- and macro-
cracking. The incorporation of steel fibres in the concrete mix has been shown and
proven to increase the tensile strength of the concrete by bridging the crack and
maintain tensile stress flow in the concrete (Gao et al. 1997; Campione and La Mendola
2004; Abbas et al. 2014; Di Prisco et al. 2013; Grabois et al. 2016; Mo et al. 2017).
Provided that shrinkage cracking only takes place if the contraction due to water
dissipation to environmental diffusion or cement hydration is higher than the concrete
tensile strength, steel fibre reinforcement can offer a control mechanism to shrinkage
cracking at all ages. Previous studies on lightweight concrete agree with the latter (Tan
et al. 1994; Zhang et al. 2001).
Since the current design codes such as the Eurocode 1992-1 use the drying
shrinkage to calculate the long term shrinkage strain due to drying shrinkage, while as
erroneous as it might be, autogenous shrinkage is assumed to be controlled in the
choice of material mixing and curing conditions, this work focuses on the measurement
of the free unrestrained drying shrinkage of SFRLC.
2.1 Experimental Programme

The experimental program included 3 mixes with 3 different fibre dosages: Vf = 0%,
1% and 2%. Each mix constituted of 3 specimens. The mixes also aimed to study the
effect of different W/C ratios and type of fibre (Table 1). The latter was either 3D fibres
regarded as the most commonly used in industry (Sadoon et al. 2018) or the 5D fibres
with the highest tensile strength and most comprehensive hooking system which
promises of being capable of primary reinforcement substitution.
Table 1. The specimens cast

Mix fck (MPa) Vf (%) Fibre type
1 30 0, 1, 2 3D
2 40 5D
3 30 5D
Shrinkage of Steel-Fibre-Reinforced Lightweight Concrete 361
It should be noted that the specimens were kept in an environmental chamber with
unrestrained supports.
2.1.1 Materials
Portland-Limestone cement (CEM 11) according to the specification supplied in EN
197-1 was used. Coarse aggregate Lytag, also known as Sintered Pulverised Fuel Ash
Lightweight Aggregate (LYTAG) was provided by LYTAG Ltd. The loose dry density
of LYTAG was calculated in the lab to be approximately 760 kg/m3 while the water
absorption was estimated to be around 15% per mass of LYTAG. For this reason the
aggregates were pre-soaked for 24 h before mixing. This was proven to help reduce
shrinkage. Natural river sand with a 4.75 mm maximum size was used as the fine
aggregate of the concrete. The sand had a water absorption coefficient of 0.09% and
specific gravity of 2.65 complying with BS EN 12620. The properties of the fibres used
are summarized in the table below. It should be noted that to prevent the possibility of
balling, fibres were collated from the manufacturer (Table 2).
Table 2. Properties of fibres

Fibre type ru (MPa) lf (mm) df (mm)
3D 65/60 1160 60 0.9
5D 65/60 2300 60 0.9
2.1.2 Mix Design

The mix designs used are summarized in Table 3 below. The mix design of the Lytag
concrete for the characteristic cylinder and cube compressive strengths used are
summarised below. These were directly adopted from Lytag (2011) manuals.
Table 3. Mix design used

(fck/fck, Cement Sand Loose bulk Lytag Effective water
cube) (kg/m3) (kg/m3) (kg/m3) (kg/m3)
LC30/33 370 592 668.8 175
LC40/44 480 485 668.8 175
Calculating the water content of Lytag was of high importance as Lytag aggregates
were found to absorb water of approximately 15% of their weight which is also
confirmed by Lytag manual. For this reason and as suggested by Lytag manual 5
(2011), excess water to saturate Lytag aggregates was added 30 min before mixing
(Fig. 1).
Fig. 1. Mixing process for plain and fibrous lightweight concrete

To measure the free axial early age and drying shrinkage, 100 100 500 mm beam
specimens designed according to the European standard UNI 6555 with steel inserts
fixed into both ends of the specimens were cast and the method of measuring shrinkage
was identical to UNI 6555. This relatively large specimen was preferred to other
smaller specimens such as the more conventional 75 75 280 mm specimen
according to BS ISO 1920-8:2009 in an attempt to avoid any possible favorable ori-
entation of fibres in the shrinkage direction i.e. orthogonal to crack plane for the
SFRLC beams.
The specimens are required to be housed in a measuring apparatus with a digital
displacement guage for the duration of the test with a constant relative humidity of
(50 ± 5)% and temperature of (22 ± 2) °C (Fig. 2). To measure shrinkage strain, simply
Fig. 2. Concrete specimen about to be housed in the measuring apparatus

the digital readings of the displacement are taken and divided by the total length of the
specimen. Shrinkage strains from repeated specimens are averaged. The drying shrinkage
is assumed to start from day 2 following early age shrinkage which is recorded 24 h after
concrete hardening.
Figure 3 below shows the recorded free drying shrinkage for the 2 different W/C ratios
chosen. Throughout the duration of the test, the mix with the higher W/C ratio
(fck = 30 MPa) seemed to shrink more than that with the lower W/C ratio (fck = 40
MPa). Both the readings reduced in gradient as the days of testing were increased. It
should be noted that while the shrinkage of the mix with the low W/C ratio relatively
plateaued at the end of the test, that of the mix with the high W/C ratio continued to
increase (75% more shrinkage).
Figure 4 below studies the effect of hook geometry on drying shrinkage for 2
fibrous beams with Vf = 1%. It can be seen that while the free shrinkage of the fibrous
beam reinforced with the stronger bonded 5D was slightly higher during the first 14
days, the free shrinkage of the fibrous beam reinforced with 3D fibres became higher
during the remaining days of the test.
Figure 5 below illustrates the effect of increasing the fibre volume fraction Vf on
shrinkage. It is clear that the higher the fibre dosage, the lower the shrinkage as the
tensile strength of fibres prevent cracking of the concrete beams. The axial shrinkage
strain of the beam reinforced with Vf = 1% was about 56% of that of the plain
lightweight concrete beam, while that reinforced with Vf = 2% was only 23%.
Drying shrinkage for different W/C ratios
140
120
Free shrinkage (με)
100
80
fck=30MPa, Vf=0%
60
fck=40MPa, Vf=0%
40
20
0
0 5 10 15 20 25 30
Days
Fig. 3. Effect of W/C ratio on shrinkage

Drying shrinkage for different fibre types
250
200
Free shrinkage (με)
150
fck=30MPa, Vf=1%, 5D
100
50
0
0 5 10 15 20 25 30
Days
Fig. 4. Effect of fibre type on shrinkage
Drying shrinkage for differnet Vf

160
140
120
Free Shrinkage (με)
100
80 fck=30MPa, Vf=0%, 5D
60 fck=30MPa, Vf=1%, 5D
40
20
0
0 5 10 15 20 25 30
Days
Fig. 5. Influence of increasing fibre dosage on the drying shrinkage of lightweight concrete
The column chart below (Fig. 6) displays the early age shrinkage due to drying
shrinkage and cement hydration process of the concrete. Overall, it is evident that the
early age shrinkage is responsible for the majority of the shrinkage at the end of the 28
day test which makes it impactful. This agrees with findings reported by Holt (2001).
For example, during the first day after hardening for the specimen with fck = 30 MPa
and Vf = 0%, shrinkage strain was recorded to be 742 microstrains which is 542%
higher than that recorded at the end of the testing at 28 day. This is due to the porous
nature of the lightweight concrete and the ability of the aggregates to absorb 15% of
their weight in water which makes water loss due to environmental diffusion more
likely. Another observation that can be made is with regards to fibre geometry. It
appears that the more complicated anchorage system of the 5D fibres created more air
voids in the concrete lattice making it more easily for the water to dissipate as com-
pared to the specimens reinforced with the less complicated hook-ended 3D fibres.
Similarly, the higher fibre fraction led to higher early age shrinkage. Hence, it can be
deduced that a better vibration process, smaller fibres, less extended hooks or high
humidity curing process is recommended for better shrinkage control.
Early age shrinkage (με)

800
700
600
500
400
300
200
100
Fig. 6. Early age shrinkage for SFRLC beams
4 Conclusions
• The methodology adopted in this work was successful at measuring the early age
and drying shrinkage for fibrous lightweight concrete beams.
• The higher the water cement ratio, the higher the drying shrinkage measured.
• The higher the fibre dosage, the lower the drying shrinkage recorded.
• The more extensive the hook anchorage system is the less the drying shrinkage.
However, it was seen that the extensive 5D hooks lead to a more pronounced
increase in early age shrinkage as compared to the more conventional 3D fibres.
• For the duration of the test and the material used, the early age shrinkage during the
first day after hardening appeared to be 300% to 500% higher than the drying
shrinkage by the end of the test.
• The porous nature of lightweight concrete and the ability of its aggregate to absorb
up to 15% of their weight in water make the concrete more susceptible to shrinkage.
• The addition of fibres creates more voids in the lightweight concrete, making it
likely to shrinkage especially during early age.
• Therefore, to control early shrinkage, a better compacting process, smaller fibres,
less extended hooks or 99% humidity curing is recommended during early age.
References
Hossain, A.B.: Assessing residual stress development and stress relaxation in restrained concrete
ring specimens. Ph. D Disseratation. Purdue University (2003)
Havlasek, P.: Czech Technical University in Prague (2014). http://mech.fsv.cvut.cz/wiki/images/
c/ca/PhD_thesis_Havlasek_2014.pdf. Accessed 8 July 2015
Neville, A.: Properties of Concrete, 5th edn. Pearson, Harlow (2011)
Mindess, S., Young, J.F.: Concrete, vol. 181. Patience-Hall, Inc., New York, (1981). 671 p.
Shoya, M.: II-5 Drying Shrinkage and Moisture Loss of Super Plasticizer Admixed Concrete of
Low Water Cement Ratio. Transactions of the Japan Concrete Institute (1979)
Mehta, P.K.: Concrete technology at the crossroads-Problems and opportunities. American
Concrete Institute, SP-144, pp. 1–30 (1994)
Mehta, P.K., Monteiro, J.M.: Concrete: Structure, Properties and Materials, 2nd ed. Prentice Hall,
Inc., New York (1993)
Koenders, E.A.B.: Simulation of Volume Changes in Hardened Cement-Based Materials. Delft
University Press, Delft (1997)
Kayali, O., Haque, M., Zhu, B.: Drying shrinkage of fibre-reinforced lightweight aggregate
concrete containing fly ash. Cem. Concr. 121 Res. 29(11), 1835–1840. http://www.
sciencedirect.com/science/article/pii/S0008884699001799. Accessed 15 Jan 2020
Zhang, M., Zakaria, M., Dang, L., Paramasivam: Shrinkage of high-strength lightweight
aggregate concrete exposed to dry environment. ACI 129 Mater. J. 102(2). http://www.
researchgate.net/publication/250613564_Shrinkage_of_highstrength_lightweight_aggregate_
concrete_exposed_to_dry_environment. Accessed 15 Jan 2020
Gao, J., Sun, W., Morino, K.: Mechanical properties of steel fiber-reinforced, high-strength,
lightweight concrete. Cement Concr. Compos. 19(4), 307–313 (1997)
Campione, G., La Mendola, L.: Behavior in compression of lightweight fiber reinforced concrete
confined with transverse steel reinforcement. Cement Concr. Compos. 26(6), 645–656 (2004)
Abbas, A., Syed Mohsin, S., Cotsovos, D.: Seismic response of steel fibre reinforced concrete
beam–column joints. Eng. Struct. 59, 261–283 (2014)
Di Prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in fib Model Code 2010:
principles, models and test validation. Struct. Concr. 14(4), 342–361 (2013)
Grabois, T., Cordeiro, G., Filho, R.: Fresh and hardened-state properties of self-compacting
lightweight concrete reinforced with steel fibers. Constr. Build. Mater. 104, 284–292 (2016)
Mo, K., Goh, S., Alengaram, U., Visintin, P., Jumaat, M.: Mechanical, toughness, bond and
durability-related properties of lightweight concrete reinforced with steel fibres. Mater.
Struct. 50(1) (2017)
Tan, K., Paramasivam, P., Tan, K.: Creep and shrinkage deflections of rc beams with steel fibers.
J. Mater. Civ. Eng. 6(4) (1994). http://ascelibrary.org/doi/abs/10.1061/(ASCE)0899-1561
(1994)6:4(474). Accessed 15 Jan 2020
Zhang, J., Li, V.: Influences of fibers on drying shrinkage of fiber-reinforced cementitious
composite. J. Eng. Mech. 127(1). http://ascelibrary.org/doi/abs/10.1061/%28ASCE%290733-
9399%282001%29127%3A1%2837%29?journalCode=jenmdt. Accessed 15 Jan 2020
Sadoon, A., Rees, D.W.A., Ghaffar, S.H., Fan, M.: Understanding the effects of hooked-end steel
fibre geometry on the uniaxial tensile behaviour of self-compacting concrete. Constr. Build.
Mater. 178, 484–494 (2018)
Lyag: Technical manual. Lytag ltd., London (2011)
Holt, E.: Early age autogenous shrinkage of concrete. Doctoral dissertation ed. University of
Washington, Seatle (2001)
Time Dependent Deflection of FRC Members
Under Sustained Axial and Flexural Loading
Murray Watts1, Ali Amin1(&), R. Ian Gilbert2, and Walter Kaufmann3

1
School of Civil Engineering, The University of Sydney, Sydney, Australia
ali.amin@sydney.edu.au
2
School of Civil and Environmental Engineering,
The University of New South Wales, Sydney, Australia
3
Institute of Structural Engineering (IBK), ETH Zürich, Zurich, Switzerland
Abstract. The inclusion of steel or polypropylene fibres into a concrete matrix

can considerably improve the serviceability performance of reinforced concrete
members. The benefits of including fibres in structural concrete have been
extensively studied, and as a result, provisions for strength, and short-term
serviceability conditions are contained in national codes of practice such as the
Australian Standards for Concrete Structures and Concrete Bridges. Provisions
relating to the long-term serviceability behaviour of fibre reinforced concrete
(FRC) are either not included or can be seen to provide limited guidance to
designers. This paper describes a method of analysis that can be applied to
predict the time-dependent behaviour of cracked fibre (steel or macro-synthetic)
reinforced concrete. The model is versatile and can handle a wide range of
geometries, material properties and loading conditions. The layered modelling
approach provides a high level of flexibility which allows for the consideration
of variable creep, shrinkage and fibre properties, as a function of time. Results
from the model have been compared to existing experimental data available in
the literature and have been shown to correlate well. In addition, a sample
analysis is presented to demonstrate the effects of residual tensile stress, tensile
creep and variable shrinkage gradients on a FRC flexural section.
Keywords: Fibre Concrete Combined axial/flexural loading Sustained

loading Serviceability Deformation Cracking Creep
1 Introduction
Satisfying the serviceability limit state is one of the primary objectives when designing
reinforced concrete structures. This is partly due to aesthetic concerns, since cracked or
excessively deformed of concrete may detract from architectural finishes and impair the
functionality of a structure. It is also important from a durability perspective, as wider
crack widths can allow the penetration of aggressive solutions that can cause damage to
internal reinforcing steels [1]. The inclusion of fibres into a concrete matrix has been
investigated and shown to significantly improve the serviceability performance of
reinforced concrete members [2, 3]. To date, with respect to in-service behaviour,
modelling and subsequent provisions included in National Standards have revolved
around short-term behaviour. Currently the Australian Standards for Concrete
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_34
Time Dependent Deflection of FRC Members 369
Structures and Concrete Bridges [4, 5] contain no provisions for predicting the long-
term serviceability performance of FRC. This can be partially attributed to the lack of
experimental data available due to the onerous nature of long-term testing but, also, the
research focus on the strength, and short-term behaviour found in the literature.
Failure to account for creep and shrinkage of the concrete and creep of the fibres, if
any, can lead to a serious under-estimation of both the extent and severity of cracking
and long-term deformations of FRC members [6]. The development of simple and
reliable methods to accurately predict long-term cracking and deformation is essential
if the full potential of FRC is to be realised in practice. The model presented herein
aims to solve this problem. The model is versatile as seen in its ability to handle a range
of inputs but remains simple enough for routine use. The model presented in this paper
is an extension of the modelling approaches found in Gilbert and Bernard [6], Gilbert
and Amin [7] and Watts et al. [8].
2 Assumed Constitutive Relationships
The first step towards the development of serviceability models for efficient analysis
and design of FRC members is to formulate accurate and reliable underpinning models
of material behaviour. These include the material constitutive behaviour of the FRC as
well as the long-term deformational characteristics of FRC in tension and compression.
2.1 Uncracked FRC

Under in-service conditions, maximum compressive stresses rarely exceed 50% of the
maximum compressive strength of the concrete. For the modelling undertaken in this
paper, it is assumed that the compressive concrete remains in this range and is modelled
as linear elastic. Specific compressive properties can be determined through experi-
mental testing, or through formulations given in [4, 5, 9].
If an initial stress, r0, is applied to a concrete element, an elastic strain, ece,
develops in the concrete. If the stress is sustained over a time interval (i.e. between tk-1
& tk), a creep strain, Decc, develops in the concrete as a result of the initial stress. Over
the same time interval, a shrinkage strain, Decs, will also develop. These conditions are
summarised in Eq. (1):
ece ¼ r0 =Ec ; Decc ¼ ucc ðtk ; tk1 Þ ece ; Decs ¼ Decs ðtk Þ Decs ðtk1 Þ ð1Þ
2.2 FRC in Tension

The most fundamental property when considering the design of structural concrete
containing fibres is its post cracking, or residual tensile strength [10–12]. This is
typically characterised by its stress verses crack opening displacement (COD) (r−w)
relationship. This relationship is primarily influenced by the geometry, material and
dosage of the fibres as well as the strength of the matrix in which the fibres are cast.
The recent revision of the Australian Standard [5] characterises the tensile response of
370 M. Watts et al.
FRC through the concrete cracking stress fct, the stress at a COD of 0.5 mm, f0.5, and
the stress at a COD of 1.5 mm COD, f1.5. From these defined points the entire short-
term behaviour of the fibre can be reasonably determined by interpolating and
extrapolating between these points provided the stress remains greater than 0 MPa. In
this paper, wf is defined as the crack width in which the fibres carry no stress:
1 ½mm
wf ¼ þ 0:5 ½mm ð2Þ
1 f1:5 =f0:5
The stress carried by the fibres, rw, at any COD may be determined as:
wf w
rw ¼ f0:5 ð3Þ
wf 0:5 ½mm
The r−w relationship is the most suitable method of modelling the post cracking
behaviour of FRC. However, it can be troublesome and often not compatible when
implementing into structural analysis and design procedures. For this reason, it is
beneficial to convert the r−w relationship to a r−e relationship. This is conveniently
achieved through the calculation of a characteristic length, lcs which is accounted for in
the fib Model Code 2010 [9] and facilitates the definition of longitudinal strain as a
function of crack width:
e ¼ w=lcs ð4Þ
For FRC elements with conventional longitudinal reinforcing steel, lcs may be
taken as:
lcs ¼ minðsr ; ðD dn ÞÞ ð5Þ
where sr is the distance between primary cracks, D is the total depth of the section and
dn is the depth to the neutral axis from the extreme compressive fibre. In this paper, the
crack spacing in FRC members is predicted based on the adaptation of the Tension
Chord Model proposed by Marti et al. [13] accounting for the presence of the fibres
[14, 15]:
ðfct f0:5 Þdb ðD dn Þb

sr ¼ ð6Þ
12fct Ast
where db is the diameter of the longitudinal reinforcement, D is the depth of the

member, b is the width of the member, dn is the neutral axis of the section and Ast is the
total area of tensile longitudinal reinforcement.
σ σ
fct fct
Ec=E t f0.5
1 f1.5
ε
0 ε cr 0 ε cr
w w
0 0 0.5 mm 1.5 mm wf
(a) (b)
Fig. 1. Stress vs strain/COD(w) relationship for softening FRC: (a) Actual Behaviour;
(b) Modelled Behaviour
2.3 Creep and Shrinkage of FRC Matrix

In this paper, uncracked concrete creep and shrinkage are modelled using the provi-
sions in fib MC2010 [9]. The creep coefficients are defined as a function of the concrete
strength, relative humidity and time at first loading. The tensile creep of cracked FRC is
modelled through a creep amplification factor, kfcc. This is a crude approximation of the
creep that occurs within the fibres, if any (i.e. which might be the case for
polypropylene fibres), as well as any loss of frictional bond between the fibres and
concrete matrix with time. For the purposes of modelling, we take the tensile creep as:
uc;av ¼ kfcc ucc ¼ kfcc ðecc =ece Þ ð7Þ
where kfcc = 1 for steel fibres and kfcc > 1 for polymer fibres. We note that this is an
area that requires further research, testing and modelling.
2.4 Conventional Reinforcing

When serviceability levels of loading are applied to FRC members, the stress in the
conventional reinforcing steels are typically well below the yield stress of the steel. In
this paper it is assumed that the reinforcing steels remain in the linear elastic range
through time and therefore the reinforcing strain, es, can be determined as:
es ðtÞ ¼ rs =Es ð8Þ
3 Layered Method of Analysis
The layered approach presented herein is a common approach to the solution of

structural problems. Implementing a discretized layered approach provides significant
flexibility in analysis and allows the designer to assign variable properties to each of the
layers in the analysis. The assignment of variable properties is well suited to the
analysis of FRC structures where the fibres response in tension contributes in cracked
372 M. Watts et al.
layers only. As part of the analysis, a conventionally reinforced FRC section sym-
metrical about its vertical axis is discretised into n layers of D/n thickness. This is
represented in Fig. 2 [8].
y
εr
ε0 σ0
σs1
ds1
1
As1
f ct
ds2
z
D
εi
As2 κ fi
Dn
i σs2
n
bi
(a) Section (b) Strains (c) Stresses

at crack at crack
Fig. 2. Layered Approach. (a) Typical Section. (b) Strain Resultant. (c) Layer Stress Resultant.
The sectional analysis is somewhat complicated by the softening behaviour of the

cracked concrete region in which the fibres are engaged and resist tension. For
implementation of the method described below, a simplification is made in the eval-
uation of the stiffness provided by cracked FRC layers. In this paper, we define the
effective modulus of FRC (refer to Fig. 3), Ec,eff(i), as:
(
Ec . . . for ec;i \ecr
Ec;effðiÞ ¼ fw;i ð9Þ
ec;i Ec ¼ efctcr . . . for ec;i ecr
Ec
fct 1
fi 1 Ec,eff(i)
ε
0 ε cr εi
Fig. 3. Effective Modulus of FRC
If the layer is uncracked, then the effective modulus is equal to the elastic modulus
of the concrete, Ec. However, if the layer is cracked then the effective modulus is
defined based on the fibre tensile response. The effective modulus of each cracked layer
is defined at the completion of the instantaneous analysis and is dependent on the extent
of cracking identified. It is assumed that the effective modulus remains constant for
each layer throughout each subsequent time interval. The concrete rigidities (cracked or
uncracked) for the cross-section are therefore defined as:
X
n X
n X
n
RA;c ¼ AcðiÞ Ec;effðiÞ ; RB;c ¼ AcðiÞ Ec;effðiÞ yi ; RI;c ¼ AcðiÞ Ec;effðiÞ y2i ð10Þ
i¼1 i¼1 i¼1
where Ac(i) is the cross-sectional area of the i-th concrete layer, and yi is the distance
between the centroid of the i-th layer and the chosen reference axis. Similarly, the steel
rigidities are defined as a function of the number of steel layers, ns:
X
ns X
ns X
ns
RA;s ¼ AsðiÞ Es ; RB;s ¼ AsðiÞ Es yi ; RI;s ¼ AsðiÞ Es y2i ð11Þ
i¼1 i¼1 i¼1
The corresponding instantaneous strain at the reference axis and the instantaneous
curvature produced by the applied actions can then be directly evaluated as:

RI RB
er ¼ Next þ Mext ð12aÞ
RA RI R2B RA RI R2B

RB RA
j¼ Next þ Mext ð12bÞ
RA RI R2B RA RI R2B
where RA, RB and RI are the overall cross-sectional rigidities and equal to:
RA ¼ RA;c þ RA;s ; RB ¼ RB;c þ RB;s ; RI ¼ RI;c þ RI;s ð13Þ
3.1 Instantaneous Cracked Sectional Analysis at First Loading (t0)

To determine the instantaneous response at time step t0 for a section subjected to an
applied moment and/or axial force, the section is first assumed to be uncracked. In the
uncracked state, Ec,eff(i) is constant for all layers and is equal to the intact concrete
elastic modulus, Ec (see Eq. (8)). Using the corresponding sectional rigidity charac-
teristics, Eq. (13), the strain at the reference axis, er and curvature of the section, j are
calculated using Eq. (12).
A search is then undertaken to identify whether the tensile stress of the concrete has
been exceeded in one or more concrete layers (ec,i > ecr = fct/Ec). If cracking is not
identified, the section is solved, and computation is terminated. If cracking is identified,
the cracked concrete tensile stress in each cracked layer is set equal to the residual
tension carried by the fibres based on the FRC material stress verses strain relationship
(as defined in Fig. 1b and Eq. (2) and (3)). The stress in each of the remaining
uncracked layers is equal to rc(i),0 = Ec,eff(i) ec,i with Ec,eff(i) equal to the intact
concrete modulus, Ec.
The stress in each of the steel layers is equal to rs(i) = Es es,i. Based on the
cracked residual tensile stresses and the current strain and curvature profile, Ec,eff(i) is
updated for each cracked concrete layer within the section (Eq. (9)). These yields
revised cross-sectional rigidities which in turn updates the strain (Eq. (12)) and stress
374 M. Watts et al.
profile of the section. This iterative procedure is repeated until axial and rotational
equilibrium is achieved on the cross-section:
X
n X
n
Nint;0 ¼ Ac;i rc;i þ As;i rs;i ¼ RA;c þ RA;s er;0 þ RI;c þ RI;s j0 ¼ Next ð14aÞ
i¼1 i¼1
X
n X
n
Mint;0 ¼ Ac;i rc;i yi þ As;i rs;i yi ¼ RB;c þ RB;s er;0 þ RB;c þ RB;s j0 ¼ Mext
i¼1 i¼1
ð14bÞ
3.2 Time Analysis During Each Time Interval (Tk-1 to Tk)

During each time interval, a relaxation procedure is adopted to determine the change of
strains and stresses in the section induced by creep and shrinkage. The strain state is
initially frozen; that is, the strain distribution at the start of a time interval is assumed
equal to that determined in the previous time interval. If the total strain in a concrete
layer is held constant, the creep and shrinkage components are fully restrained and
stresses develop due to the restrained strain. The restrained strain is equal in magnitude
to the creep and shrinkage strains but of opposite sign. The restrained creep strain in a
particular layer during the time interval tk-1 to tk is the creep strain that would develop
due to the initial stress in the layer at time t0 and the sum of creep strains caused by the
stress increments that have developed in each of the previous time intervals.
When transitioning from time interval tk-1 to tk, the layered approach presented
above involves the determination of a fictitious transitional force (Ntrans;tk ) and a fic-
titious transitional moment (Mtrans;tk ). These transitional actions may be interpreted as
the equivalent force and moment produced by the increments of creep and shrinkage in
a particular time interval. The creep and shrinkage components are calculated inde-
pendently and then added together to determine the total transitional strain in each
individual layer through the section in that time interval:
Detrans;i;tk ¼ Decreep;i;tk þ Deshrink;i;tk ð15Þ
The creep strain that develops in the interval tk-1 to tk in a particular layer is equal to
the summation of the increments of creep caused by the stress increments in each
previous time interval Each creep increment is the product of the stress increment and
the change in the associated creep coefficient versus time curve between tk-1 and tk
divided by Ec:
ri;t0 ut1 ;t0
Decreep;i;t1 ¼
Ec
1 h
kP i
ð16Þ
ri;t0 utk ;t0 utk1 ;t0 þ ri;tj ri;tj1 utk ;tj utk1 ;tj
j¼1
Decreep;i;tk [ 1 ¼
Ec
The shrinkage component of the transitional strain is determined by applying the

increment of shrinkage strain to each of the individual layers. The increment of
shrinkage strain is the additional shrinkage that occurs in each individual layer during
the time interval tk-1 to tk only.
Deshrink;i;tk ¼ esh;i;tk esh;i;tk1 ð17Þ
Once Decreep;i;tk and Deshrink;i;tk are defined, Ntrans;tk can be evaluated as the sum-
mation of the restrained change in strain due to creep and shrinkage multiplied by the
effective modulus and the area of each layer through the depth of the section:
X
n X
n
Ntrans;tk ¼ Ntrans;tk ;i ¼ Detrans;i;tk AcðiÞ Ec;effðiÞ ð18Þ
i¼1 i¼1
Similarly, the transitional moment is taken as the summation over all the layers of
the product of the axial force determined in Eq. (18) and the distance between the
centroid of each individual layer and the reference axis location:
X
n
Mtrans;tk ¼ Ntrans;tk ;i yi ð19Þ
i¼1
The cumulative effects of creep and shrinkage Incurred from t1 onwards are
determined through the summation of all transitional forces and moments from all
previous time step analyses undertaken combined with the externally applied loads.
This represents the cumulative (total) impacts of creep, shrinkage and externally
applied loads on a section:
X
k
Ncumulative;tk ¼ Next þ Ntrans;tj ð20aÞ
j¼1
X
k
Mcumulative;tk ¼ Mext þ Mtrans;tj ð20bÞ
j¼1
The corresponding strain profile at time tk is calculated by replacing Next and Mext
with Ncumulative;tk and Mcumulative;tk , respectively in Eq. (12). From the resultant strain
profile a revised stress profile is determined by taking the resolved strain in each layer
at time step tk, ec;i;tk and subtracting the summed transitional strains for that layer
obtained from the previous time interval analyses and then multiplying by Ec,eff(i):
!
X
k
rc;i;tk ¼ Ec;effðiÞ ec;i;tk Detrans;i;tj ð21Þ
j¼1
376 M. Watts et al.
The iterative solution presented above is necessary because the FRC exhibits strain
softening characteristics after cracking and the concrete rigidities (RA,c, RB,c, RI,c) are a
function of the effective modulus of each layer (Ec,eff,(i)), which in turn are dependent
on the strain (er,tk & jtk) and hence stress resultants of the FRC layers. A simplification
is made after the completion of the instantaneous analysis that assumes the concrete
section rigidities remain constant in time during each subsequent time interval.
4 Model Validation
Results obtained using the proposed model are plotted against deformation vs time and
crack width vs time data from, Vasanelli et al. [15], Gilbert and Nejadi [16] and Aslani
et al. [17] for beams ST1E, ST2E, ST3E, ST4E, B1a, B1b, SSCC-a. The output
curvatures at each time step were translated into member deflections through the
application of Mohr’s Analogy (the Conjugated Beam Method). The correlations to
existing experimental data sets are presented in Fig. 4.
0.20 15 25
Experiment Experiment
Model Model
B1a 20
0.15
10
15
B1b
δ [mm]
w [mm]
δ [mm]
Experiment
0.10
Model (kfcc = 1)
10 Model (kfcc = 5)
5 Model (kfcc = 10)
0.05 b = 250 mm; D = 250 mm; d = 205 mm; l = 2800 mm; b = 250 mm; D = 350 mm; d = 300 mm; l = 3500 mm; b = 400 mm; D = 161 mm; d = 130 mm; l = 3500 mm;
fc = 21.4 MPa; fct = 2.3 MPa; Ec = 27.7 GPa; fc = 25 MPa; fct = 2.8 MPa; Ec = 24.9 GPa; 5 fc = 38 MPa; fct = 3.1 MPa; Ec = 35.8 GPa;
Ast = 462 mm 2; ρf = 0.60%; lf = 30 mm; df = 0.60 mm; Ast = 400 mm 2; ρf = 0%; lf = N/A; df = N/A; Ast = 452 mm 2; ρf = 0.56%; lf = 65 mm; df = 0.85 mm;
f0.5 = 0.70 MPa; f1.5 = 0.61 MPa f0.5 = 0 MPa; f1.5 = 0 MPa f0.5 = 0.49 MPa; f1.5 = 0.47 MPa
0.00 0 0
0 100 200 300 400 500 600 700 0 100 200 300 400 0 50 100 150 200 250
t [Days] t [Days] t [Days]
(a) (b) (c)
Fig. 4. Model comparison: (a) Vasanelli et al. (2014): Series ST; (b) Gilbert and Nejadi (2004):
B1a/b; (c) Aslani et al. (2014): SSCC-a.
5 Sample Analysis
In this section, the results of a sample cross-sectional analysis are presented. The
geometry of the cross-section is D = 250 mm, b = 1000 mm, ds1 = 30 mm, ds2 =
220 mm and As1 = As2 = 452 mm2. The section is discretised into 50 layers of equal
thickness (i.e. D/n = 5 mm). The cross-sections are subjected to a sustained bending
moment Mext = 45kNm (applying tension to the soffit of the section) and axial com-
pression of Next = 90kN that are first applied at t = 28 days. This level of loading
induces cracking of the section at first loading, and the steel and concrete stresses are
typical of in-service conditions.
The creep coefficient and shrinkage strains are modelled using the fib MC2010 [9],
with ucc(t1000, t28) = 1.39 and esh(t1000) = −432 10−6. A variable tensile creep
amplification coefficient, kfcc is also considered in the sample analysis. kfcc is varied
from 1 to 10 with the results plotted in Fig. 5.
Two different cases are analysed in Fig. 5, the first one being an increase in the
residual tensile stress provided by the fibres. In this case the f0.5 stress is modelled as
0 MPa (plain concrete) (see Fig. 5a), 0.5 MPa (Fig. 5b) and 1 MPa (Fig. 5c) with f1.5/f0.5
equal to 0.8 in the latter two cases.
In Fig. 6, we have taken f0.5 = 1 MPa but have varied the shrinkage profile through
the depth of the section. A constant, increasing and decreasing shrinkage profile is
applied to the section through the layered model analysis–these distributions are shown
in Figs. 6a, 6b and 6c, respectively. This analysis is not specific to a particular FRC
however it does illustrate the flexibility of the model and the importance of under-
standing the influence of shrinkage profiles in the analysis of structures.
Fig. 5. Effect of varying residual stress on curvature in time (a) f0.5 = 0 (Plain Concrete);
(b) f0.5 = 0.5 MPa; (c) f0.5 = 1.0 MPa.
Fig. 6. Effect of varying shrinkage profile on curvature in time & corresponding strain profiles
(a) Constant shrinkage profile. (b) Increasing shrinkage. (c) Decreasing shrinkage.
378 M. Watts et al.
6 Conclusions
A method of analysis to predict the time dependent behaviour of FRC has been pre-
sented. The model can be used to predict member deflection and crack widths of
members subjected to sustained axial and flexural loads. The model can predict the
behaviour of either macro-synthetic or steel fibre reinforced concrete and has the
versatility to implement variable tensile creep properties in order to simulate the
behaviour of synthetic fibres. The model was validated against various results available
in the literature and shown to correlate well. The results reveal that inclusion of fibres in
the concrete reduces time-dependent deformations and can significantly reduce maxi-
mum crack widths when used in combination with conventional reinforcing bars.
A sample analysis was undertaken to provide an insight to the sensitivity of the outputs
as a function of the residual stress provided by the fibres along with any tensile creep of
the fibres. Additionally, the impacts of a variation in the shrinkage profile on an FRC
flexural section was investigated. The sample analyses demonstrate the flexibility of the
model but also the importance of understanding the impacts of residual stress, tensile
creep and shrinkage on the long-term serviceability performance of FRC structures.
Acknowledgements. This work was supported by an Australian Research Council Discovery

Grant (DP 200102114) awarded to the second and third Authors.
References
1. Chiaia, B., Fantilli, A.P., Vallini, P.: Evaluation of crack width in FRC structures and
application to tunnel linings. Mater. Struct. 42, 339–351 (2009)
2. Amin, A., Foster, S., Watts, M.: Modelling the tension stiffening effect in steel fiber
reinforced-reinforced concrete. Mag. Concr. Res. 68(7), 339–352 (2016)
3. Vrijdaghs, R., di Prisco, M., Vandewalle, L.: Uniaxial tensile creep of a cracked
polypropylene fiber reinforced concrete. Mater. Struct. 51(1), 1–12 (2018). https://doi.org/
10.1617/s11527-017-1132-5
4. AS5100.5.: Bridge Design Part 5: Concrete. Australian Standard, Standards Association of
Australia (2017)
5. AS3600.: Concrete Structures. Australian Standard, Standards Association of Australia
(2018)
6. Gilbert, R.I., Bernard, E.S.: Creep analysis of macro-synthetic fibre cross-sections in
combined bending and axial force. Concr. Aust. 41(1), 33–40 (2015)
7. Gilbert, R.I., Amin, A.A.: Time-dependent behaviour of fibre reinforced concrete. Concr.
Aust. 45(1), 31–38 (2019)
8. Watts, M.J., Amin, A., Gilbert, R.I., Kaufmann, W.: Behavior of fiber reinforced concrete
members under sustained axial/flexural load. Struct. Concr, pp. 1–17 (2019)
9. fib Model Code 2010. Fédération Internationale du Béton, p. 402 (2013)
perspectives. Mater. Struct. 42, 1261–1281 (2009)
11. Pfyl, T.: Tragverhalten von Stahlfaserbeton. PhD Dissertation, IBK-Report No. 279. Swiss
Federal Institute of Technology, Switzerland. (in German) (2003)
12. Amin, A., Foster, S., Muttoni, A.: Derivation of the r-w relationship for SFRC from prism
bending tests. Struct. Concr. 16(1), 93–105 (2015)
13. Marti, P., Alvarez, M., Kaufmann, W., Sigrist, V.: Tension chord model for structural
concrete. Struct. Eng. Int. 8(4), 287–298 (1998)
14. Amin, A., Foster, S., Watts, M.: Modelling the tension stiffening effect in steel fiber
reinforced-reinforced concrete. Mag. Concr. Res. 68(7), 339–352 (2016)
15. Vasanelli, E., Micelli, F., Aiello, M.A., Plizzari, G.: Crack width prediction of FRC beams in
short- and long-term bending condition. Mater. Struct. 47, 39–54 (2014)
16. Gilbert, R.I., Nejadi, S.: An experimental study of flexural cracking in reinforced concrete
members under sustained loads. UNICIV Report R-435. School of Civil and Environmental
Engineering, The University of New South Wales, Australia (2004)
17. Aslani, F., Nejadi, S., Samali, B.: Long-term flexural cracking control of reinforced self-
compacting concrete one way slabs with and without fibres. Comput. Concr. 14(4), 419–444
(2014)
Influence of the Residual Tensile Strength
on the Factor for Quasi-permanent Value
of a Variable Action w2
Darko Nakov1(&), Goran Markovski1, Toni Arangjelovski1,

and Peter Mark2
1
Faculty of Civil Engineering, University “Ss. Cyril and Methodius”,
Skopje, North Macedonia
nakov@gf.ukim.edu.mk
2
Faculty of Civil and Environmental Engineering, Ruhr-University Bochum,
Bochum, Germany
Abstract. Steel fibres are known to aid in deflection control, control the pro-
cess of cracking and mainly improve the toughness of structural elements and
the whole structure. Large experimental program was performed at the Faculty
of Civil Engineering-Skopje to find out how steel fibres and the residual tensile
strength affect the time-dependent deformation properties and deflections of
concrete. Specific realistic loading with permanent and repeated variable loads
in loading interval of 8 h per day was applied on full scale beams that were
monitored up to an age of concrete of 400 days. The beams were with cross
section dimensions 15/28 cm and total length of 300 cm, manufactured from
concrete class C30/37. They were reinforced with same percentage of longitu-
dinal and shear reinforcement, but with different amount of steel fibres (0,
30 kg/m3 and 60 kg/m3). Using the experimental results, detailed analysis of the
time-dependent deformation properties of concrete and their effect on the time-
dependent behaviour was done. A value for the factor for quasi-permanent value
of variable action w2 is proposed for each type of concrete. It was concluded that
the factor for quasi-permanent value of a variable action w2 depends linearly on
the residual tensile strength.
Keywords: Steel fibre reinforced concrete Residual tensile strength

Repeated variable loads Factor w2
1 Introduction
The effect of long-term actions is usually connected with the permanent load. However,
there are certain concrete structures such as: storage areas at warehouses, traffic areas at
parking garages and city bridges under severe traffic conditions, where the variable
loads are acting longer and are with significant magnitude. In these structures, the
variable actions could overcome serviceability limit states criteria of concrete
structures.
In the serviceability limit states design, effects due to creep and shrinkage of
concrete caused by variable load, are taken into consideration using quasi-permanent
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_35
Influence of the Residual Tensile Strength 381
combination of actions [1]. The level of quasi-permanent load is defined by the factor
for quasi-permanent value of variable action w2, Eq. (1):
X X
G þPþ
j 1 k;j
w Qk;i
i 1 2;i
ð1Þ
Where:
Gk;j – permanent actions,
P– prestressing force,
Qk;i – variable actions,
w2;i – factors for quasi-permanent values of variable actions
The National Annex of each country could also set the values for the factor w2. This
gives opportunity for research of effects of variable load and its replacement by a quasi-
permanent load.
Proposed values for the factor w2 for buildings in Eurocodes [1], are presented in
Table 1.
Table 1. Recommended values of factor w2 for buildings [5]

Action w2
Imposed loads in buildings, category (see EN1991-1-1)
Category A: domestic, residential areas 0.3
Category B: office areas 0.3
Category C: congregation areas 0.6
Category D: shopping areas 0.6
Category E: storage areas 0.8
Category F: traffic area, vehicle weight 30 kN 0.6
Category G: traffic area, 30 kN < vehicle weight 160 kN 0.3
Category H: roofs 0.0
The recommended value for the factor w2 for road bridges is 0 [1]. In DIN report
102 “Concrete bridges” based on Eurocodes, for city bridges, w2 = 0.2, while in the
National application document of Finland for EN 1992-2, w2 = 0.3. The recom-
mended value of the factor w2 for railway bridges is 0. But if deformation is considered
for persistent and transient design situations, the factor w2 should be taken equal to
1.00 for rail traffic actions [1].
Several studies have been conducted on long-term behavior of SFRC beams under
sustained loads [2–6], while studies which include the effect of variable repeated load are
uncommon [7]. Up to now, there is no research dealing with the factor w2 for steel fibre
reinforced concrete. Having in mind that the residual tensile strength is one of the main
characteristics which differ ordinary from fibre concretes, with this research an attempt
was made to find out the influence of the residual tensile strength to the factor w2.
382 D. Nakov et al.
The experiment was carried out at the “Ss. Cyril and Methodius” University, Faculty of
Civil Engineering in Skopje, Republic of North Macedonia. It involved testing of 24
full scale beams constructed from reinforced concrete and steel fibre reinforced con-
crete with additional reinforcement. The beams were with cross section dimensions
15/28 cm and total length of l = 300 cm, Fig. 1. Together with each series of beams,
control specimens were cast in order to test the compressive strength, flexural (and
residual) tensile strength [8], splitting tensile strength, elastic modulus and deforma-
tions due to creep and shrinkage. In addition to the tests on mechanical [9] and time-
dependent properties of concrete [10, 11], the used reinforcement was also tested.
Fig. 1. Geometry, reinforcement and loading scheme of full scale beams.
All 24 beams were manufactured with concrete class C30/37. According to the used
type of material, they were divided into three series:
– Series A, reinforced concrete (C30/37);
– Series B, SFRC with 30 kg/m3 steel fibres and additional reinforcement (C30/37
FL1.5/1.5);
– Series C, SFRC with 60 kg/m3 steel fibres and additional reinforcement (C30/37
FL2.5/2.0).
In each series, the plain reinforcement was kept the same. The longitudinal rein-
forcement was ribbed and of RA 400/500-2 quality, while the shear reinforcement was
smooth, with GA 240/360 quality. Reinforcement 2Ø10, 2Ø8 and Ø6/10/20 cm was
used as tension, compression and shear reinforcement, respectively. The used steel
fibres were hooked-end HE1/50, produced of cold-drawn wire, manufactured by
Arcelor Mittal, with a diameter of 1 mm, length of 50 mm and tensile strength of
1100 N/mm2.
The mixture proportions are presented in Table 2.
Regarding the loading history, the beams were divided into four groups, each with
6 beams. In this paper, only the results for group “4” are presented.
On the beams from group “4”, a long term permanent load with intensity “g” was
applied at the age of concrete of 40 days and was held for a year as a long term load.
On the fortieth day, variable repeated load “±q” was also applied in an interval of 8 h
+q and 16 h −q, for a year. This means that the beams were loaded additionally with
Table 2. Mixture proportions

Mixture proportions (kg/m3)
Cement CEM II/A-M 42.5 N 410
Water 215
Water/Cement ratio, w/c 0.524
Aggregate:
0–4 mm (river sand), 50% 875
4–8 mm (limestone), 20% 350
8–16 mm (limestone), 30% 525
Fibres:
C30/37 0
C30/37 FL 1.5/1.5 30
C30/37 FL 2.5/2.0 60
load “q” for 8 h every day, whereat the strains, deformations and crack widths were
measured. After 8 h, the beams were unloaded from load “q” and all measurements
were performed again. The long term loading scheme is presented in Fig. 2.
Fig. 2. Long term loading scheme.
The beams and control specimens were cured for 8 days and then they were
transported to the Laboratory at the Faculty of Civil Engineering – Skopje, where they
were kept under almost constant temperature with an average of 19.5 °C and constant
relative ambient humidity with an average of 60.2%, which was regulated with special
humidifiers and dehumidifiers.
In each step, the concrete strains in middle section of the beam through the
thickness as well as on the top of the beam, were measured by a mechanical deflection
meter, type Hugenberger, Switzerland, with a base of 250 mm. The mechanical
measurement of the deflections was done at 5 points through the length of the beam and
2 points over the supports by using deflection meters produced by Stopani, Italy. The
crack widths were also measured in each load step, in the region with constant moment,
by use of a crack microscope - product of Controls, Italy. The positions of the mea-
surement points are presented in Fig. 3.
The long term load, which consists of permanent sustained load “g” and repeated
variable load “q”, was applied by gravitation levers, which enabled an increase of the
384 D. Nakov et al.
Fig. 3. Positions of measurement points of full scale beams.
load for 13 times. The permanent load acts all the time, while the variable load was
applied and removed each day by secondary hand gravitation levers.
The bending moments are as follows: from self-weight of the beam, Msw = 1 kNm,
from permanent load “g”, Mg = 5.0 kNm, from variable load “q”, Mq = 3.1 kNm,
from self-weight, permanent and variable load (service) Msw+g+q = 9.1 kNm. The
bending crack moment was Mcr = 6.1 kNm, while the ultimate bending moment
Md = 15.6 kNm. The intensity of the load was chosen so that the Mcr is bigger than
Msw+g and smaller than Msw+g+q. The permanent load is 0.39 times the flexural
strength, while the service load is 0.58 times the flexural strength of the beam without
fibres.
3 Analytical Analyses
The analytical analyses of the results from the experimental research were performed in
two parts:
• Analytical analysis of time – dependent deformation properties,
• Analytical analysis of time – dependent deflections.
Data on the time – dependent deformation properties were later used to calculate
the time-dependent deflections using the Age-Adjusted Effective Modulus Method
(AAEMM). The quasi-permanent load procedure and the principle of superposition
were used to obtain the factor for the quasi-permanent value of the variable action w2 .
3.1 Analytical Analysis of Time – Dependent Deformation Properties

The analytical analysis of time – dependent deformation properties, drying shrinkage
and creep, were performed by the B3 model [12] and Fib Model Code 2010 [13]. In the
beginning, the analyses were done only for the time period considered in this research,
which was 400 days.
The B3 model offers the possibility of improvement of the model by its users and
updating of its predictions based on short-time measurements. The updating of the
drying shrinkage strain was done very efficiently by using the scaling parameter p6. The
experimental and analytical results for the drying shrinkage up to the age of 400 days
are presented in Fig. 4. It can be noticed that the Fib Model Code 2010 underestimates
the drying shrinkage strain for 29%, while the original B3 model underestimation is
11.5%. The obtained scaling parameter in the improved B3 model is p6 = 1.123. It can
be noticed that there is a very good agreement between the experimental results and the
improved B3 model. Having in mind the service life of designed structures, it is very
important to be able to predict the time – dependent deformation properties for their
serviceability period. Therefore, based on the results obtained for the age of up to 400
days, the analyses according to the previously mentioned models were extended to the
serviceability period of the structures of 100 years. The results are presented in loga-
rithmic scale in Fig. 5.
900
Drying shrinkage εds [10-6]μs
800
700
600
500 C30/37
400 C30/37 FL 1.5/1.5
300 C30/37 FL 2.5/2.0
FIB MC 2010
200
B3 model
100
B3 model IMP.
0
0 50 100 150 200 250 300 350 400
t [days]
Fig. 4. Experimental and analytical results for drying shrinkage up to 400 days.
Due to the differences in the creep strain between different types of concrete, an
analytical analysis of the creep strain was performed for each concrete type taken
separately, by the B3 model and Fib Model Code 2010. For the concrete type C30/37,
the experimental and analytical results are presented in Fig. 6. In addition to the
previously mentioned models, the improvement of the B3 model is also presented. On
the basis of linear regression analysis, the following values for adjustment of the creep
compliance in the B3 model were obtained: p1 = 1.143 and p2 = 1.122. These values
are valid only for the concrete type C30/37. Taking into consideration the coefficients
of variation of each model code, a good agreement was found in all cases. The Fib
Model Code 2010 overestimates the experimentally obtained creep coefficient at 400
days for 13%, while the B3 model underestimates it for 9.5%.
The experimental results for both steel fibre reinforced concretes, C30/37 FL
1.5/1.5 and C30/37 FL 2.5/2.0, show a very small difference in the final strain after 400
386 D. Nakov et al.
900
Drying shrinkage εds [10-6]μs
800
700
600
500 C30/37
400 C30/37 FL 1.5/1.5
C30/37 FL 2.5/2.0
300
FIB MC 2010
200
B3 model
100
B3 model IMP.
0
1 10 100 1000 10000 36500 100000
t [days]
Fig. 5. Experimental and analytical results for drying shrinkage up to 100 years.
800
700
600
Creep εcc [10-6]μs
500
400
300 C30/37
200 FIB MC2010

B3 model
100
B3 model IMP.
0
0 50 100 150 200 250 300 350 400
t [days]
Fig. 6. Experimental and analytical results for creep of C30/37 up to 400 days.
days when the original B3 model is used. Therefore, for these types of concrete,
modification of the flow compliance q4 in the B3 model is proposed. The modification
includes addition of an amount of fibers Gf to the amount of aggregate a, multiplied by
the ratio between the moduli of elasticity of the fibers and the aggregate Ef/Ea, Eq. (2).
0 10:7
E
a þ Gf f
q4 ¼ 20:3@ A
Ea
ð2Þ
c
The results for the concrete C30/37 FL 2.5/2.0 with 60 kg/m3 steel fibers are
presented in Fig. 7. Based on the results obtained for an age of up to 400 days, the
analyses of the creep strain according to the previously mentioned models were
extended to the serviceability life of the structures of 100 years. The results for the
concrete type C30/37 FL 2.5/2.0 are presented in logarithmic scale in Fig. 8.
800
700
600
500
400
300 C30/37 FL 2.5/2.0
200
B3 model
100
B3 model MOD.
0
0 50 100 150 200 250 300 350 400
t [days]
Fig. 7. Experimental and analytical results for creep of C30/37 FL 2.5/2.0 up to 400 days.
3.2 Analytical Analysis of Time – Dependent Deflections

There are many available analytical and numerical methods for obtaining the time –
dependent response of concrete structures. However, the big number of necessary input
data, complicated analysis and the big number of unknown parameters in the design
phase make these methods impractical for most of the engineers. The influence that
variable load has on the time – dependent response of concrete structures is another
aspect that, in certain structures, has a big effect on the total behaviour. With the
experimental program of this research, it was planned to use a simple quasi – per-
manent load procedure to obtain a factor by which only one part of the variable load
will be included in the time – dependent analysis.
This procedure is intended for calculation of the creep effects when serviceability
limit state design should be performed using the quasi – permanent combination of
actions. This combination of actions enables inclusion of the variable loads in the
calculation of creep effects. One part of the variable load is added to the permanent load
and is named as quasi – permanent load. The factor that defines the part of the variable
load is called factor w2 [1]. This factor depends on the category of the building and the
loading history. The loading history on different types of concrete has been the subject
of research at the Faculty of Civil Engineering – Skopje for almost 12 years. In this
research, the loading history has been chosen such that the variable load acts for 8 h
each day in the period of one year. Four approaches to determination of the effects of
variable load were considered (Fig. 9):
The simplest solution of the problem is taking into consideration approach 1,
because the initial and time-dependent deflection can be obtained with intensity of the
load as a sum of permanent and quasi-permanent load. On the basis of the results
obtained by the experimental research, an analytical solution was proposed in which
388 D. Nakov et al.
1000
800
600
400 C30/37 FL 2.5/2.0

FIB MC2010
200 B3 model
B3 model MOD.
0
10 100 1000 10000 36500 100000
t [days]
Fig. 8. Experimental and analytical results for creep of C30/37 FL 2.5/2.0 up to 100 years.
Fig. 9. Four considered approaches to determination of w2 .
the total deflection due to the permanent load “g” and variable load “q” obtained from
the experiments at,exp(g + q) was determined as a sum of the initial deflection ao(g +
w2 q) and long-term deflection at(g + w2 q) due to the permanent load “g” and variable
load represented as quasi permanent load “w2 q”, Eq. (3):
at;exp ðg þ qÞ ¼ ao ðg þ w2 qÞ þ at ðg þ w2 qÞ ð3Þ
Factor w2 was determined by the Age – Adjusted Effective Modulus Method

(AAEMM) using the quasi – permanent load procedure and the principle of super-
position. The CRACK computer program developed by Ghali A. and Elbadry M. from
the University of Calgary, Canada, was used for calculation of the time – dependent
deflections. Except the geometry, materials and bending moments, all other necessary
input data needed for the AAEMM were obtained by the previously presented analysis
of the time – dependent deformation properties, based on the experimentally obtained
results. Those data include values which vary each day, like the values for shrinkage,
creep coefficient, relaxation function and aging coefficient. For the concrete type
C30/37, the total experimentally observed deflection was obtained with intensity of
load g + 0.37q, which meant that factor w2 had a value of 0.37. In the case of the steel
fibre reinforced concrete types C30/37 FL 1.5/1.5 and C30/37 FL 2.5/2.0, the total
experimentally observed deflection was obtained with intensity of load g + 0.22q and
g + 0.18q, which meant that the factor w2 had a value of 0.22 and 0.18, respectively.
The results for C30/37 FL 1.5/1.5 are presented in Fig. 10, while all results are pre-
sented in Table 3.
Time-dependent deflection
Concrete type: C30/37 FL 1.5/1.5
5
Total deflection a(mm)
C30/37
4 FL 1.5/1.5
(EXP.)
3 C30/37
ψ2 =0.22 FL 1.5/1.5
(AAEMM)
2 g+0.22q
0
0 50 100 150 200 250 300 350 400
t (days)
Fig. 10. Experimental and analytical time-dependent deflection for concrete type C30/37 FL
1.5/1.5.
Table 3. Factor w2 as a function of the ratio of residual tensile strength to the concrete
compressive strength, af
Concrete type C30/37 C30/37 FL 1.5/1.5 C30/37 FL 2.5/2.0
af 0 0.043 0.063
w2 0.37 0.22 0.18
ψ2 (αf)
0.8
C30/37
0.6
ψ2
C30/37 Linear
C30/37 FL 1.5/1.5 (C30/37)
0.4
C30/37 FL 2.5/2.0
0.2
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
αf
Fig. 11. Factor w2 as a function of the ratio between the residual tensile and compressive
strength, af.
390 D. Nakov et al.
In the following figure (Fig. 11), the results are presented as a function of the ratio
between the residual tensile strength and the concrete compressive strength, af, which
is zero for ordinary concrete. A simple linear dependence between factor w2 and the
ratio af can be noticed.
4 Conclusions
• Analytical analysis of the drying shrinkage and creep, performed by the B3 model
and fib Model Code 2010, demonstrated agreement with the experimental results.
For concrete C30/37, an improvement of the B3 model was done on the basis on
linear regression analysis. New values of the coefficients p1 and p2 for adjustment of
the creep compliance were obtained, as well as scaling parameter for the drying
shrinkage p6. For steel fibre reinforced concrete types, a modification of the flow
compliance q4, which includes the amount of fibres is proposed.
• The repeated variable load has significant influence on the time-dependent beha-
viour of reinforced and steel fibre reinforced concrete beams.
• Using the experimental results and analytical analysis, the following factors for
quasi-permanent value of a variable action were obtained: for concrete type C30/37,
w2 = 0.37, while for concrete type C30/37 FL 1.5/1.5, w2 = 0.22 and for C30/37
FL 2.5/2.0, w2 = 0.18.
• The factor w2 depends linearly on the residual tensile strength.
References
1. European Standard EN1990+A1. Basis of structural design, Standardization Institute of R.
Macedonia, Skopje, R.Macedonia (2002)
2. Tan, K.H., Saha, M.K.: Ten-year study on Steel fiber-reinforced concrete beams under
sustained loads. ACI Struct. J. (2005)
3. Vasanelli, E., Micelli, F., Aiello, M.A., Plizzari, G.: Long term behaviour of fiber reinforced
concrete beams in bending. In: BEFIB2012 – Fibre Reinforced Concrete, Guimaraes (2012)
4. Casucci, D., Thiele, C., Schnell, J.: Behavior of cracked cross-section of fibre reinforced
UHPFRC under sustained load. In: Serna, P., et al. (ed.) Creep Behaviour in Cracked
Sections of Fibre Reinforced Concrete. RILEM Bookseries, vol. 14. Springer (2017)
5. Nishiwaki, T., Kwon, S., Otaki, H., Igarashi, G., Shaikh, F.U.A., Fantilli, P.: Experimental
study on time-dependent behaviour of cracked UHP-FRCC under sustained loads. In: Serna,
P., et al. (ed.) Creep Behaviour in Cracked Sections of Fibre Reinforced Concrete. RILEM
Bookseries, vol. 14. Springer (2017)
6. Candido, L., Micelli, F., Vasanelli, E., Aiello, M.A., Plizzari, G.: Durability of FRC beams
exposed for long-term under sustained service loading. In: Serna, P., et al. (ed.) Creep
Behaviour in Cracked Sections of Fibre Reinforced Concrete. RILEM Bookseries, vol. 14.
Springer (2017)
7. Nakov, D., Markovski, G., Arangjelovski, T., Mark, P.: Creeping effect of SFRC elements
under specific type of long term loading. In: Serna, P., et al. (ed.) Creep Behaviour in
Cracked Sections of Fibre Reinforced Concrete. RILEM Bookseries, vol. 14. Springer
(2017)
8. RILEM TC 162-TDF: Test and design methods for steel fiber reinforced concrete, bending
test, final recommendation. Mater. Struct. 35 (2002)
9. Nakov, D., Markovski, G., Arangjelovski, T.: Influence of steel fibre reinforcement on the
properties of concrete. In: 1st International Conference COMS, Zadar (2017)
10. Nakov, D., Markovski, G., Arangjelovski, T., Mark, P.: Experimental and analytical analysis
of creep of steel fibre reinforced concrete. Periodica Polytechnica Civil Engineering 62(1)
(2017)
11. Nakov, D., Markovski, G., Mark, P., Arangjelovski, T.: Analytical analysis of drying
shrinkage of SFRC based on experimental results. In: Conference Fibre Concrete 2015,
Prague (2015)
12. Bazant, Z.P., Baweja, S., Creep and shrinkage prediction model for analysis and design of
concrete structures: model B3. In: Al-Manaseer, A. (ed.) The Adam Neville Symposium:
Creep and Shrinkage-Structural Design Effects, SP-194, American Concrete Institute,
Farmington Hills, Michigan (2000)
13. Fib Model Code 2010, fib Bulletin 65, Final draft (2012)
Compressive and Tensile Creep and Shrinkage
of Synthetic FRC: Experimental Results
and Comparison to Codes
Razan H. Al Marahla and Emilio Garcia-Taengua(&)

E.Garcia-Taengua@leeds.ac.uk
Abstract. The mechanical properties of FRC mixes, namely compressive and

tensile strength, are generally improved with respect to their unreinforced
counterparts due to the contribution of fibres. This has implications in aspects
like crack propagation or the development of time-deferred strains such as those
resulting from creep and shrinkage, which in turn influence the long-term
deformation of structural elements. The effect of fibres on compressive creep has
been studied by various researchers, and codes and guidelines for the design of
concrete structures include provisions to predict creep under compression.
However, tensile creep of FRC has not attracted the same level of coverage
despite its relevance to the loss of tension stiffening. This paper presents the
results of an experimental study in concrete specimens reinforced with synthetic
fibres were subjected to constant loading in uniaxial tension and in compression
in order to evaluate the effect that increasing dosages of synthetic fibres have on
the resulting time-dependent strains. The evolution of shrinkage, creep strains
and the creep coefficient were analyzed in relation to the fibres dosage. The
experimental strain-time curves were compared to the theoretical curves from
the models adopted in the Eurocode 2, the Model Code, or by the ACI Com-
mittee 209.
Keywords: Synthetic fibres Strains Time-dependant
1 Introduction
Concrete exhibits volume changes with time. Due to the partial or total restraint of
concrete deformation, shrinkage induce stresses even in the absence of externally
applied loads, and the resulting tensile stresses can result in cracking [1, 2]. The long-
term behaviour of concrete structures is also influenced by creep deformations under
sustained stresses. Time-dependent strains due to shrinkage and creep affect the
mechanical capacity of reinforced concrete sections [3]. The creep and shrinkage
equations included in the Eurocode 2 [4], Model Code 2010 [5], or in ACI 209.2R [6]
are tools for estimating shrinkage and creep of concrete as a function of time.
The presence of fibres in concrete plays a significant role in controlling the prop-
agation and growth of microcracks, and increase the tensile strength of concrete [7] as
well as its toughness in the cracked state [8]. The behaviour of FRC under sustained
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_36
Compressive and Tensile Creep and Shrinkage 393
tensile and flexural loads, and how this behaviour is influenced by the combined effect
of creep and shrinkage, remain issues that attract considerable research interest.
The behaviour of concrete under sustained compressive loads has been studied by
different authors, who have investigated the effect of different variables such as the
compressive strength of concrete, the specimen shape, the stress level applied, the
duration of the sustained loading, or the temperature and relative humidity conditions
[9, 10]. An empirical expression to estimate concrete creep under compression was
developed by [11], based on experimental data obtained for FRC mixes with different
types and dosages of steel fibres, and the effect of fibres geometry and dosage was
studied in terms of bond of fibres to the cementitious matrix.
Different methodologies have been proposed for evaluating tensile creep. Domone
[12] evaluated the tensile creep of both sealed and immersed concrete specimens,
adopting the uniaxial tensile test introduced by Elvery and Haroun [13], in which
cylindrical (bobbin) specimens are tapered at both ends and a concentric load is
applied. Tensile creep reduces the stresses caused by restrained shrinkage to some
extent [14–16]. Swamy and Stavrides investigated the influence of fibres on restrained
shrinkage and subsequent cracking and concluded that the addition of fibres to normal
and lightweight concrete mixes can reduce shrinkage deformations up to 20% [17].
Various experimental and analytical investigations considered different fibre types and
contents and studied the effect that higher fibre dosages have on interfacial bond and to
what extent fibres contribute to crack control [18–20].
The effect of steel fibres on concrete time-dependent deformations has been
investigated widely and results show that steel fibres improve the tensile and flexural
capacity of FRC under sustained loads [21, 22]. This paper is concerned with the effect
that synthetic fibres have on creep and shrinkage under sustained tensile and com-
pressive loads. Different contents of synthetic fibres were added to the same reference
mix design and the sensitivity to this parameter was evaluated.
2.1 Materials and Reference Mix Design

The reference mix design considered in this study had a water-to-cement (w/c) ratio of
0.29 and an average compressive strength of 60 MPa at 28 days. It was adjusted for a
slump value of 120–150 mm so it could incorporate synthetic fibres at different
dosages without further adjustments. The mix proportions are given in Table 1. The
synthetic fibres were high-modulus 54 mm long polymeric fibres, with a tensile
strength of 600 MPa. A picture of these fibres is shown in Fig. 1.
2.2 Mixes Production and Testing Methodology

In total, three different mix designs were considered in this study, all based on the
reference mix design summarised in Table 1 and differing in the synthetic fibre content
only: 0, 5, and 10 kg/m3. Two identical batches of each of these mixes were produced,
Table 1. Mixture proportions kg/m3.

Specimen w/c = 0.29
Cement 510
Water 147
Fine aggregate 950
Coarse aggregate (20 mm) 300
Coarse aggregate (10 mm) 580
Superplasticiser 9
Fig. 1. Synthetic fibres used in this study.
and the same number of specimens were cast from each batch, with the objective of
generating a sufficient number of replicates for better accuracy of the results.
The same mixing sequence was followed in all cases. Prior to the mixing of every
batch, the total amount of water to be added was separated in two buckets: one
containing only 80% of the water, and the other containing the mixture of the required
amount of superplasticiser and the remaining 20% of the water. First, cement and all
aggregates were poured into the mixer and dry-mixed for 2 min. After that, 80% of the
water was added and mixed with the cement and aggregates for 3 min. During this
time, the fibres were poured gradually into the mixer. Finally, the remaining 20% of the
water with the superplasticiser predispersed in it was added, and the mixing continued
for 4 min. In all cases, a uniform distribution of the fibres in the mix was observed.
From each batch, specimens were cast and tested as follows:
• Characterisation: 3 cubic specimens (100 mm side) and 3 cylindrical specimens
(150 300 mm), to determine the compressive strength and splitting tensile
strength at 28 days.
• Creep and shrinkage under compression: 8 prismatic specimens with a
75 75 mm cross-sectional area and a length of 200 mm. Of these specimens, 4
were tested under sustained compressive load, and 4 were not loaded but instru-
mented to measure free shrinkage strains.
• Creep and shrinkage under tension: 3 cylindrical specimens with a diameter of
75 mm and a length of 365 mm. Throughout this paper, these are referred to as
‘bobbins’, to distinguish them from the 150 300 mm cylindrical specimens. Out
of these 3 bobbins, 2 were tested under sustained tensile load, and 1 was not loaded
but instrumented to measure free shrinkage.
Characterisation specimens were tested at the age of 28 days. Prior to this, they
were kept in a fog room with a controlled temperature of 20 °C and a relative humidity
of 90%.
Creep and shrinkage specimens were instrumented with DEMEC points, which
were glued and fixed to their sides and positioned 150 mm apart from each other, in
order to measure the average surface strain. A picture of the specimens and test
instrumentation is shown in Fig. 2. For the specimens tested under compression, a
sustained load corresponding to 25% of the average compressive strength measured at
28 days was applied. For the specimens tested in tension, the sustained tensile stress
applied was 1 MPa. Measurements were taken daily for 90 days. In both cases, the tests
were set up in a controlled room where the temperature and relative humidity were kept
at 21 ± 2 °C and a relative humidity of 50 ± 5% throughout the duration of the tests.
Fig. 2. Experimental setup for specimens in compression (left) and in tension (right).

3.1 Compressive Strength and Splitting Tensile Strength
For each of the batches, compressive strength and tensile splitting strength were
measured at 28 days. The average results as well as the corresponding standard
deviation values are shown in Table 2.
Table 2. Compressive and splitting tensile strength results.

Fibre content (kg/m3) Compressive strength Splitting tensile
(MPa) strength (MPa)
Average Std. deviation Average Std. deviation
0 60.9 0.8 3.9 0.1
0 60.4 0.6 3.9 0.4
5 61.1 0.7 4.4 0.2
5 61.6 0.7 4.3 0.2
10 60.4 0.6 4.6 0.5
10 59.8 1.4 4.6 0.4
Table 3. Strains measured in prismatic specimens under compression (in microstrain).

Fibre content Elastic strains Strains at 14 days Strains at 30 days Strains at 90 days
Loaded Unloaded Loaded Unloaded Loaded Unloaded
0 kg/m 3
−395 −807 −156 −968 −231 −1146 −325
−390 −811 −162 −966 −237 −1140 −326
−412 −789 −151 −962 −228 −1150 −319
−389 −819 −149 −958 −240 −1136 −312
5 kg/m3 −365 −728 −145 −858 −201 −1033 −288
−359 −730 −134 −864 −189 −1042 −286
−373 −724 −128 −851 −193 −1038 −280
−376 −731 −149 −863 −210 −1022 −277
10 kg/m3 −356 −655 −100 −786 −168 −881 −267
−353 −661 −92 −797 −178 −876 −256
−347 −647 −110 −814 −179 −890 −274
−354 −639 −118 −789 −163 −871 −265
Regarding the average compressive strength, the addition of synthetic fibres at 5 or

10 kg/m3 did not introduce statistically significant differences with respect to the
batches without fibres. The same can be said in relation to the standard deviation
values. In terms of the splitting tensile strength, an improvement was observed due to
the introduction of synthetic fibres. With respect to the batches without fibres, the
average splitting tensile strength increased by 11.5% when synthetic fibres were dosed
at 5 kg/m3, and by 18% when the fibres content was 10 kg/m3.
3.2 Creep Under Sustained Compression

For the analysis of creep and shrinkage under compression, strains were measured over
a period of 90 days on unloaded as well as loaded specimens. Measurements were
taken daily on a total of eight specimens per case. Table 3 shows the measurements
corresponding to 14, 30, and 90 days, as well as the elastic strains, which were
measured immediately after the compressive load was applied.
Individual values for the compressive creep strain at any age were determined as
the difference between the loaded and unloaded specimen strains, minus the corre-
sponding elastic strain. The average compressive creep strains for the three synthetic
fibre contents considered are shown in Fig. 3, together with the theoretical strains as
predicted by the equations in Model Code and Eurocode 2. The compressive creep
strains corresponding to the reference mix without fibres showed good agreement with
the theoretical values. In particular, strain values after 5 days were observed to be very
close to the predictions calculated according to the Model Code.
Fig. 3. Average compressive creep strains and comparison to codes.
The addition of fibres was correlated with a reduction in compressive creep strains,
consistently observed at all ages. Strain values corresponding to FRC mixes with
5 kg/m3 of synthetic fibres were between 10% and 15% lower than those observed in
mixes without fibres (strains were reduced in 15%, 11% and 10% at 14, 30, and 90
days, respectively). For the FRC mixes where the fibre content was 10 kg/m3, these
reductions were between 15% and 22% (strains were reduced in 22%, 16% and 15% at
14, 30, and 90 days, respectively).
The analysis of the results was also made in terms of the creep coefficient values,
and these are shown in Fig. 4, together with the theoretical values as predicted by the
equations in the Model Code, the Eurocode, and the ACI Committee 209 report. Again,
it was observed that the values corresponding the case without fibres showed very good
agreement with the theoretical values as per the Model Code.
At all ages, the presence of synthetic fibres was correlated with lower creep
coefficients than without fibres, and the reduction associated with a fibre content of
10 kg/m3 was higher than that observed for a fibre content of 5 kg/m3. If the com-
pressive creep coefficient at 90 days is considered, reductions of 5% and 8% were
observed for fibres contents of 5 kg/m3 and 10 kg/m3, respectively.
Fig. 4. Average compressive creep coefficient values and comparison to codes.

3.3 Shrinkage
Shrinkage strains were determined from the strains measure in unloaded conditions
(values in Table 3). The average shrinkage strains for the different synthetic fibre
contents considered, together with the theoretical values according to the Model Code,
Eurocode 2 and the ACI Committee 209 report, are shown in Fig. 5. Shrinkage strains
observed in specimens without fibres were well in agreement with the ACI 209 pre-
dictions up to approximately 50 days. For later ages, the model proposed by the
Eurocode 2 yielded better predictions. In any case, the experimental results seemed
conclusive in showing that the theoretical models tend to overestimate shrinkage strains.
Fig. 5. Average shrinkage strains and comparison to codes.
Similarly to what was observed in relation to creep under compression, the addition
of synthetic fibres led to generalised reductions in shrinkage strains, and these reduc-
tions were directly related to the fibre content. The addition of synthetic fibres at a
dosage of 10 kg/m3 decreased shrinkage strains in between 20% and 26% (the
reductions observed were 23%, 26%, and 20% at 14, 30, and 90 days, respectively). In
those cases where the fibre dosage was 5 kg/m3, shrinkage strains were between 8%
and 16% lower than the values corresponding to the specimens without fibres.
3.4 Creep Under Sustained Tension

To characterise the tensile creep, strains were measured for 90 days on unloaded as
well as loaded specimens. Table 4 summarises the measurements corresponding to 14,
30, and 90 days, as well as the elastic strains, which were measured immediately after
the tensile load was applied.
Table 4. Strains obtained from loaded and unloaded bobbins under tensile creep
Fibre content Elastic strains Strains at 14 days Strains at 30 days Strains at 90 days
Loaded Unloaded Loaded Unloaded Loaded Unloaded
0 kg/m 3
31.6 101 −244 145 −322 168 −381
27.9 99 139 155
5 kg/m3 26.2 78 −191 85 −241.5 126 −311
27.4 75 88 101
10 kg/m3 24.4 65 −172 85 −218 98 −282
23.8 66 79 94
Individual tensile creep values were determined as the difference between the
loaded and unloaded specimen strains, minus the corresponding elastic strain. The
average tensile creep strains observed for the three synthetic fibre contents considered
in this study are shown in Fig. 6.
Fig. 6. Tensile creep strain for samples with different fibre dosages.
Tensile creep strains were significantly reduced by the addition of synthetic fibres.
The differences observed between the specimens without fibres and those with 5 kg/m3
of fibres were significantly more pronounced than the differences observed when the
fibre content was increased from 5 kg/m3 to 10 kg/m3. The introduction of 5 kg/m3 of
synthetic fibres led to reductions in tensile creep strains between 20% and 11% with
respect to the unreinforced specimens (reductions were 20%, 16%, and 11% at 14, 30,
and 90 days, respectively). With 10 kg/m3 of synthetic fibres, reductions between 15%
and 24% were observed (24%, 22%, and 15%, at 14, 30, and 90 days, respectively).
These observations led to the conclusion that, in terms of tensile creep reduction,
the additional gains achieved by doubling the fibre content were less significant than
those achieved by the incorporation of fibres at the intermediate dosage considered in
this study. In consequence, dosing the fibres at the maximum dosage did not led to the
most advantageous control of tensile creep from a cost-benefit point of view.
4 Conclusions
In this study, a reference mix with w/c ratio of 0.29 and an average compressive
strength of 60 MPa at 28 days was considered as representative of most of the medium
to high specification mixes prevalent in real scale production. The effect of synthetic
fibres on compressive creep, tensile creep and shrinkage when dosed at 5 kg/m3 and
10 kg/m3 was analysed. To this end, daily measurements were taken on loaded and
unloaded specimens over a period of 90 days. Multiple replicates were produced and
tested to ensure accuracy of the final results and conclusions, and a comparison was
made to the predictive models proposed by Eurocode 2, Model Code, and the ACI
committee 209.
The following conclusions were drawn:
• In terms of compressive strength, no significant differences were observed between
the reference mix without fibres and the FRC mixes, irrespective of the fibre
content. However, the addition of 5 kg/m3 or 10 kg/m3 of synthetic fibres increased
the spitting tensile strength by 11.5% or 18%, respectively.
• The incorporation of synthetic fibres was associated with a general, consistent
decrease in compressive creep, shrinkage, and tensile creep strains. The magnitude
of these reductions was observed to increase with the fibre content.
• Compressive creep strains measured on FRC specimens with 5 kg/m3 of synthetic
fibres were up to 15% lower than those corresponding to the reference mix without
fibres. When the fibre content was 10 kg/m3, hihger reductions (up to 22%) were
observed. In terms of the creep coefficient at 90 days, it decreased by 5% and 8% for
fibre contents of 5 kg/m3 and 10 kg/m3, respectively.
• Tensile creep strains observed in FRC specimens with 5 kg/m3 of synthetic fibres
were up to 20% lower than the values obtained for the reference mix without fibres.
The additional gains achieved by doubling the fibre content were less significant:
reductions of up to 24% were observed when the fibre content was 10 kg/m3.
• The addition of 10 kg/m3 of synthetic fibres led to shrinkage strains being up to
26% lower than those corresponding to the reference specimens without fibres.
When the fibre content was 5 kg/m3, reductions of up to 16% were obtained.
• The models proposed by the Model Code, Eurocode 2 and the ACI Committee 209
for the prediction of creep and shrinkage strains showed good agreement with the
compressive creep strains observed in specimens without fibres, but they consis-
tently overestimated shrinkage strains.
Acknowledgements. The authors wish to acknowledge the contribution of Oscrete Construction

Products, part of Christeyns UK Ltd, and Sika Ltd (UK), which very kindly provided some of the
materials used in this study, as well as the support and assistance provided by the technical staff
of the School of Civil Engineering, University of Leeds. The authors are also thankful to Al-
Zaytoonah University of Jordan for the financial support granted to Ms Al Marahla in under-
taking her PhD studies at the University of Leeds.
References
1. Kristiawan, S.A.: Strength shrinkage and creep of concrete in tension and compression. Civ.
Eng. Dim. 8(2), 73–80 (2006)
2. Gilbert, R.I., Ranzi, G.: Time-Dependent Behaviour of Concrete Structures. CRC Press,
Boca Raton (2010)
3. Bazant, Z.P., Wittmann, F.H.: Creep and shrinkage in concrete structures, pp. 12–16 (1982)
4. BS EN 1992-1-1. Eurocode 2: Design of concrete structures: Part 1-1: General rules and
rules for buildings. European Committee for Standardization (2004)
5. FIB: fib model code for concrete structures 2010, Wiley, Germany (2013)
6. ACI Committee, ACI 209.2 R-08: Guide for Modeling and Calculating Shrinkage and Creep
in Hardened Concrete. American Concrete Institute Committee (2008)
7. Monteiro, P.: Concrete: Microstructure, Properties, and Materials. McGraw-Hill Publishing,
London (2006)
9. Gamble, B.R., Parrott, L.J.: Creep of concrete in compression during drying and wetting.
Mag. Concr. Res. 30(104), 129–138 (1978)
10. Brooks, J., Neville, A.M.: A comparison of creep, elasticity and strength of concrete in
tension and in compression. Mag. Concr. Res. 29(100), 131–141 (1977)
11. Mangat, P.S., Azari, M.M.: A theory for the creep of steel fibre reinforced cement matrices
under compression. J. Mater. Sci. 20(3), 1119–1133 (1985)
12. Domone, P.L.: Uniaxial tensile creep and failure of concrete. Mag. Concr. Res. 26(88), 144–
152 (1974)
13. Elvery, R.H., Haroun, W.: A direct tensile test for concrete under long-or short-term loading.
Mag. Concr. Res. 20(63), 111–116 (1968)
14. Igarashi, S.I., Bentur, A., Kovler, K.: Stresses and creep relaxation induced in restrained
autogenous shrinkage of high-strength pastes and concretes. Adv. Cem. Res. 11(4), 169–177
(1999)
15. Igarashi, S.I., Bentur, A., Kovler, K.: Autogenous shrinkage and induced restraining stresses
in high-strength concretes. Cem. Concr. Res. 30(11), 1701–1707 (2000)
16. Forth, J.P.: Predicting the tensile creep of concrete. Cem. Concr. Compos. 55, 70–80 (2015)
17. Swamy, R.N., Stavrides, H.: Influence of fiber reinforcement on restrained shrinkage and
cracking. J. Proc. 76(3), 443–460 (1979)
18. Chern, J.C., Young, C.H.: Factors influencing the drying shrinkage of steel fiber reinforced
concrete. Mater. J. 87(2), 123–139 (1990)
19. Mangat, P.S., Azari, M.M.: A theory for the free shrinkage of steel fibre reinforced cement
matrices. J. Mater. Sci. 19(7), 2183–2194 (1984)
20. Shah, S.P., Rangan, B.V.: Fiber reinforced concrete properties. J. Proc. 68(2), 126–137
(1971)
21. Llano-Torre, A., Arango, S.E., García-Taengua, E., Martí-Vargas, J.R., Serna, P.: Influence
of fibre reinforcement on the long-term behaviour of cracked concrete. In: Creep Behaviour
in Cracked Sections of Fibre Reinforced Concrete, pp. 195–209. Springer, Dordrecht, (2017)
22. García-Taengua, E., Arango, S., Martí-Vargas, J.R., Serna, P.: Flexural creep of steel fiber
reinforced concrete in the cracked state. Constr. Build. Mater. 65, 321–329 (2014)
Creep in FRC – From Material Properties
to Composite Behavior
Martin Hunger1(&), Jürgen Bokern1, Simon Cleven2,

and Rutger Vrijdaghs3
1
BASF Construction Solutions GmbH, Trostberg, Germany
martin.hunger@basf.com
2
Institute of Building Materials Research, RWTH Aachen University,
Aachen, Germany
3
Materials and Building Technology Section at KU Leuven, Leuven, Belgium
Abstract. Although being the subject of numerous studies during the last 5
decades, fiber reinforced concrete (FRC) only very recently started to penetrate
the construction market to a larger extent. This is mainly due to the lack of
appropriate standards in the past years and the recent availability of structural
codes such as fib Model Code 2010, the German DAfStb guideline
“Stahlfaserbeton”, the Italian code, or codes under development, such as the
Eurocode 2. However, FRC still struggles with some issues that potentially
hinder, at least at a first glance, a further penetration into structural applications.
Creep under tensile stress is one of these major obstacles. The creep response of
a structural member in tension results from the interaction of the creep of the
fiber material itself and the creep of the bond between fiber and surrounding
matrix. The latter at last leads to the fact that creep is also a subject of further
investigation for steel fiber reinforced concrete which shows that this is not an
exclusive problem for synthetic fibers, as publicized in common literature.
Unfortunately, commonly agreed test methods to investigate creep are missing.
In this regard more research is required in order to specify which load levels and
crack openings should be chosen to provide a realistic scenario for creep tests on
FRC in general.
Testing of creep in FRC is a very complex and time-consuming matter.
Therefore, the subject of this paper is an investigation of a possible correlation
of the creep properties of a single filament and the bond behavior of a single
fiber in mortar. The most important subject of this paper is whether it is possible
to derive the overall creep response of the composite material FRC from the
determined component-specific characteristics. Based on a thorough test plan,
two distinctly different performing polypropylene fibers (moderate to high-
performing) are tested and compared. The determination of short- and long-term
behavior of the filaments in this test plan is carried out in varying temperatures
in a range from −10 °C to 60 °C. Results indicate that the creep performances of
a filament and the bond between a filament and a mortar correlate with the
overall creep of the corresponding FRC. Moreover, a broad range of creep
performance is revealed which strongly suggests that the performance of syn-
thetic fibers and, in particular the creep performance, cannot be generalized.
Keywords: Creep Fiber reinforced concrete (FRC) Synthetic fibers

Tension Filament Bond Temperature
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_37
Creep in FRC – From Material Properties to Composite Behavior 403
1 Introduction
About 10 years ago, the state-of-the-art in terms of creep of synthetic fibers could be
summarized as follows: (i) neither a model to include creep in the design of FRC is
available nor threshold values in regard to tolerated creep deformations are defined,
(ii) a long-term load level of 50% of residual load in short-term test seems like a
maximum for well performing polypropylene (PP) fibers and (iii) long-term tests are
lacking as creep may occur after two years and even later [1].
Not much has changed in the general perception of synthetic fibers in literature
since then although results have become available, in particular during the last two
years, that would first put the above-mentioned three limitations into perspective and
even suggest synthetic fibers for structural applications [2, 3].
However, one of the most critical factors hindering a wider application or more
specifically a structural application of synthetic macro fibers is creep performance. This
paper aims to shed some light on the entire causal chain of creep – from material creep
(fiber level), over creep in the interface (bond creep) until creep performance of
structural members. An evaluation seems challenging as testing is not standardized and
some decisive test parameters often reported in literature (e.g. load levels, pre-cracking
width) seem at least arguable. This paper presents a potential testing plan for the creep
analysis of fibers for structural applications. Moreover, some indicative results are
presented to show how based on fiber creep and single-fiber pullout the creep per-
formance of FRC under sustained load can be modelled, a very valuable approach in
simplifying the time-consuming and cumbersome creep testing of FRC and a potential
way forward in design of FRC.
2 State-of-the-Art in FRC Creep
The long-term behavior of FRC under sustained load remains one item that still
requires intensive research. FRC (not necessarily with polymeric fiber reinforcement) is
in the meantime part of a number of international codes for structural concrete (e.g.
ACI 318, fib Model Code 2010 (MC 2010), the German DAfStb Guideline for steel
fiber-reinforced concrete or the Italian guideline (CNR-DT 204/2006). More codes
considering structural behavior of FRC are in development or about to include FRC,
such as Eurocode 2. In general, creep even for normal RC structures is only treated
with simplified models and design rules.
A good overview of how few creep related inputs are considered in MC2010 is
given in [3]. MC2010 highlights the need to take creep of cracked FRC into account
but does not provide any design guidelines. The German DAfStb-guideline for steel
fiber-reinforced concrete (November 2012) only mentions the possibility that defor-
mations can increase due to creep in the interface of steel fiber and matrix. However, no
testing or any form of proof is required.
Creep is a phenomenon to be considered in Serviceability Limit State (SLS) as
deformations have to be limited and crack width growth has to be controlled. In SLS,
quasi permanent loads in combination will only sum up to 30–40% of fR,1 (according to
EN 14651) of the FRC component which in turn is even less the 30–40% of the
404 M. Hunger et al.
ultimate tensile strength (UTS) of the fiber [3]. Hence, most of the creep testing done
with 60% load level or the like is too demanding and does not represent practical
application. This way a 40% load level can be accepted as a sound upper limit.
When it comes to creep in FRC, MC2010 does not consider any structural issues
other than creep of the fiber material itself. There is general consensus that the addition
of fibers, not exceeding a critical amount, does not change the creep behavior in
uncracked state in comparison to normal concrete, neither in compression nor in ten-
sion [4]. After cracking, however, time-dependent strains in FRC under tension or in
bending mainly stem from three sources: (i) creep in the compression zone, (ii) creep in
the interface between fiber and matrix and (iii) creep of the fiber material itself.
The authors believe that for a deterministic description of the phenomenon creep
(on composite level) the knowledge of single fiber creep and interfacial bond is
indispensable. So far, literature is very scarce when it comes to a systematic investi-
gation of creep on these three levels [5, 6]. However, for the technical approval of fibers
for structural purpose the ‘Deutsches Institut für Bautechnik’ prepared a guidance paper
(DIBt (2013)) that defines the necessary testing [7]. For structural purposes this also
includes a section on creep. Here, the above suggested split in fiber creep, interfacial
bond and creep on the composite level is followed. According to this test plan, next to
creep testing on concrete beams by means of 4–point bending, optionally testing of
fiber creep and interfacial bond and creep at temperatures of −10, 20 and 60 °C is
required. In order to define the performance envelope of typically used PP-fibers this
testing plan was used as guidance and two differently performing PP-fibers were
chosen to undergo this extensive test plan.
3 Applied Materials and Testing Setup
3.1 Applied Materials

A summary of the materials and test methods used in this research is given in the following
chapters. Two PP-fibers with different mechanical properties and shape were selected
(Table 1). For single fiber and fiber pull-out testing long filaments were used whereas for
the FRC bending creep tests short fibers in their nominal lengths were applied.
Table 1. Properties of the applied PP fibers*

Specimen Length Equivalent Shape Strength Modulus
[mm] diameter [mm] [N/mm2] [N/mm2]
Type A 48 0.85 embossed 363 11,000
Type B 54 0.81 crimped 408 15,000
*properties determined according to EN 14889-2, modulus shown here is tangent modulus at first
loading, i.e. the slope of the stress-strain curve at zero stress
To investigate the short- and long-term fiber-matrix interaction (pull-out) a mortar

defined in DIN EN 14649 (SIC test for glass fibers) is prescribed in the DIBt guideline.
It is composed, by weight, of three parts cement, one part CEN standard sand and 0.427
parts water. The composition of the FRC for bending tests is also specified in the DIBt
guideline. It contained 300 kg/m3 CEM I 32,5 R, 1.835 kg/m3 siliceous gravel,
180 kg/m3 water and 5.0 kg/m3 of either fiber Type A or B.
3.2 Single Fiber Testing

Single fiber tensile strength properties are determined according to DIN EN ISO 5079.
Both ends of the filament are embedded in a metal polymer resin to prevent any
slippage or edge pressure when clamping. The free length of filament is 150 mm and
the test machine is operated in position-controlled mode at a rate of 15 mm/min.
For the single fiber creep trials, apart from the free length of the filament, which is
reduced to 75 mm, filaments are prepared same as for the single fiber tensile tests. One
side of the filament cast in metal polymer resin is fixed at the top of the creep test setup
whereas the embedded filament end hanging down is loaded with steel plates and
allowed to freely move. The creep deformation is measured by an analogue dial gauge.
This setup is chosen to ensure an unhindered filament deformation and a centric
loading. Load levels of 40% and 70% of the fiber tensile strength are applied. All tests
described before are performed at temperatures of −10, 20, 40 and 60 °C.
3.3 Fiber Pull-Out and Pull-Out-Creep

For the fiber pull-out and pull-out-creep tests, mortar blocks of 30 mm side length are
cast, in which one side of the filament is centrally embedded in such way that the
filament is perpendicularly oriented to the mortar block surface. The thickness of the
mortar block is similar to the embedment depth of the fiber, which is 24 mm. The other
side of the filament again is cast in a steel polymer resin. Between mortar and resin a
free filament length of 150 mm is maintained to enable a stress-free vertical and centric
pull-out of the fiber. After demolding at the age of one day the specimens are stored for
6 days under water and afterwards for 7 days at 65% RH until testing. The temperature
is kept constant at 20 °C.
For pull-out testing the mortar block is placed on a notched steel plate where the
notch serves as lead-through for the fiber. Then the downwards oriented resin block is
clamped and the mortar block is pulled up with a rate of 25 mm/min until complete
pull-out. The test is performed at temperatures of −10, 20 and 40 °C for both fiber
types and additionally at 60 °C for fiber type B.
Pull-out-creep tests are conducted with the same test set-up as the fiber pull-out
tests, but only with fiber type B and a free fiber length of 75 mm. As load level 40% of
the respective pull-out load is defined. The creep deformation is again measured by an
analogue dial gauge, that is placed against the steel plate the mortar was positioned on.
The test is performed at temperatures of 20 °C and 40 °C for at least three months.
3.4 Fiber Reinforced Concrete Bending and Bending Creep

Finally, the composite creep behavior of the FRC with fiber type A and B is investi-
gated. Therefor six fiber-reinforced concrete beams per fiber type with dimension of
150 150 700 mm3 are cast, kept in the molds protected from evaporation for one
day and wrapped in plastic foil after demolding. At an age of 21 days the beams are
notched according to EN 14651 and wrapped again in plastic foil for further 7 days. In
order to determine the creep load the beams are first subjected to 3-point-bending test in
accordance with EN 14651 at an age of 28 days. The test is stopped when a CMOD of
0.3 mm is reached. The load applied at that point of time is documented and the beam
is unloaded. Room temperature while storing and testing is always kept at 20 °C.
Out of the series of six beams finally three beams are chosen which show a similar
behavior and load level in the short-term tests. Those beams are subsequently installed
in 3-point bending test frames where each beam could be loaded with its specific creep
load, which is defined as 50% of the beam’s residual strength at 0.3 mm crack width.
The midspan deflection of the beams is measured by two indicating calipers over a time
of one year at a controlled temperature of 20 °C and 65% RH.
4 Results
4.1 Single Fiber Tensile Strength
Tensile strength properties of both fiber types were investigated as a first step. Stress-
strain curves obtained at a temperature range from −10 °C to 60 °C are shown in
Fig. 1. Comparing the two fibers, type B shows at each temperature level an about 20–
30% higher ultimate tensile strength (UTS) than type A. Moreover, the mechanical
response to tensile stress is much stiffer for type B. The impact of temperature on
tensile fiber properties is different as well. While for fiber type A the UTS drops at
temperatures beyond 20 °C it remains in a comparable range up to 40 °C for fiber type
B. A similar conclusion can be derived concerning fiber stiffness. While for fiber type
A the response softens substantially with each temperature increment, the effect of
temperature on fiber type B becomes especially noticeable beyond 40 °C with a
reduction in strength and stiffness. Overall it can be concluded that increasing tem-
peratures have a negative impact on tensile strength of polymer fibers, but critical
temperatures at which a relevant decrease may occur seems to be fiber specific.
4.2 Single Fiber Creep

As shown in Fig. 2 where the creep deformation until 120 days of both fiber types is
presented, the difference in tensile strength properties transfers to their creep behavior.
For fiber type B, no unstable creep development occurs at load levels of 40% and
temperatures from −10 °C to 60 °C. Up to 40 °C it even seems that creep tends to
reach asymptotically a finite value. Instead, for fiber type A, an unstable creep
development may occur beyond 20 °C already. A more precise determination of the
critical temperature is unfortunately impossible, because a test at 40 °C is missing. At
the higher load level of 70%, fibers of both types show unstable creep development and
rupture. However, while for fiber type A rupture occurs at 20 °C already, it takes for
type B until 40 °C. Considering that fiber type B is – because of its higher UTS –
exposed to much higher absolute loads, this is in particular striking. In summary it is
observed that there are PP-fibers with an asymptotic stable performance in single fiber
creep even at high temperatures.
Tensile Stress in N/mm² Tensile Stress in N/mm²

500 500
-10 °C -10 °C
20 °C 20 °C
400 40 °C
400 40 °C
60 °C 60 °C
300 300
200 200
100 100
0 0
0 5 10 15 20 0 5 10 15 20
Strain in % Strain in %
(a) (b)
Fig. 1. Stress-strain curves for fiber type A (left) and B (right)
Creep strain in % Creep strain in %

0 0
-15 -15
-30 -30
-10 °C 40 %
-45 -45 -10 °C 70 %
-10 °C 40 % 20 °C 40 %
-10 °C 70 % 20 °C 70 %
-60 20 °C 40 % -60 40 °C 40%
20 °C 70 % 40 °C 70 %
60 °C 40 % 60 °C 40 %
-75 -75
0 30 60 90 120 0 30 60 90 120
Creep duration in d Creep duration in d
(a) (b)
Fig. 2. Single fiber tensile creep behavior of fiber type A (left) and B (right)
Finally, it can be concluded that quality of PP-fibers differs regarding creep, but
high-quality PP-fibers are capable to sustain relevant stress levels even at increased
temperatures.
4.3 Fiber Pull-Out

For characterizing the bond behavior of fibers and cementitious matrix, pull-out tests at
different temperatures are performed (Fig. 3). At a temperature of −10 °C both fibers
show the same behavior with a pull-out-load of about 150 N following a decrease in
bond with increasing pull-out-deformation. At higher temperatures of 20 °C to 60 °C
fiber type B shows the same behavior. In contrast, the bond of fiber type A is greatly
influenced by temperature. At 20 °C the pull-out-load is reduced by 50% and at 40 °C

by about 65%. For a temperature of 60 °C there is no result, however, considering the
impact on fiber tensile strength (Fig. 1) a significant further decrease is expected. In
conclusion it seems that fiber type B itself and eventually its shape (crimped) lead to a
much better bond at higher temperature and therefor allow a broader application scope
than for fiber type A.
Tensile Load in N Tensile Load in N

200 200
-10 °C -10 °C
175 175
20 °C 20 °C
150 40 °C
150 40 °C
125 125 60 °C
100 100
75 75
50 50
25 25
0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Pull-Out-deformation in mm Pull-Out-deformation in mm
(a) (b)
Fig. 3. Pull-Out behavior of fiber type A (left) and fiber type B (right)
4.4 Fiber Pull-Out Creep

The pull-out creep behavior is examined only on fiber type B at 40% load level and 2
temperature levels in order to investigate whether besides fiber creep also long-term
bond phenomena might influence the fiber pull-out creep (short-term bond is not
influenced by temperature for fiber type B). Results obtained at 20 °C and 40 °C are
shown in Fig. 4. At both temperature levels this fiber type first shows an initial steep
creep deformation which turns into an almost asymptotic behavior after about 60 days.
This is in line with fiber creep observations. A first estimate of the fiber creep defor-
mation at roughly similar loads indicates that in absolute terms, the fiber creep defor-
mation is in the same order of magnitude as the pull-out creep deformation. Therefore,
the effect of long-term bond phenomena appears limited and the pull-out creep seems to
be dominated by the fiber creep deformation. This confirms findings in [6].
4.5 Fiber Reinforced Concrete Bending Creep

Finally, the creep performance of the two fiber types is tested as composite material,
hence, in FRC. After pre-cracking of six beams for creep-test preparation, three of them
were fully tested in accordance with EN 14651 to characterize the short-term bending
behavior of the FRC (left side of Fig. 5). Fiber type A provides an average residual
Creep deformation in mm
0
20 °C 40 %
-1
40 °C 40 %
-2
-3
-4
-5
-6
0 30 60 90 120
Creep duration in d
Fig. 4. Pull-Out creep behavior of fiber type B
strength of fR,1 = 1.5 N/mm2 and fR,3 = 1.6 N/mm2. Type B fibers exceed this per-
formance especially at higher deformations and residual strength of fR,1 = 2.1 N/mm2
and fR,3 = 2.4 N/mm2 are achieved. These results perfectly fit to the short-term fiber
bond performance, where fiber type B showed much higher pull-out loads.
Residual tensile strength in N/mm² FRC bending deformation in mm

5,0 0,00
Type A precracking Type A
4,5 -0,05
Type A postcracking Type B
4,0 Type B precracking -0,10
3,5 Type B postcracking -0,15
3,0
-0,20
2,5
-0,25
2,0
-0,30
1,5
1,0 -0,35
0,5 -0,40
0,0 -0,45
0 1 2 3 4 5 0 50 100 150 200 250 300
CMOD in mm creep duration in d
Fig. 5. Results of 3-point-bending tests to determine the creep load (left) and creep behavior of
fiber reinforced concrete beams under bending load (right)
As a consequence, the absolute creep load of beams with fiber type B is about 15%
higher. Despite the higher applied load, fiber type B concrete beams show a smaller
creep deformation than beams with fiber type A (compare Fig. 5 right). This, according
to previous analysis, should predominately be related to the better creep performance of
fiber type B (Fig. 2). But, in contrast to fiber creep and fiber bond creep performance,
which both show an almost asymptotic trend after 60 days, for the beams unfortunately
an uncertainty remains in this respect and further investigations are required to create a
better understanding of potential phenomena behind and to build further confidence. If
the FRC creep deformation level for fiber type B should remain at this level or slightly
above, it would be acceptable for construction.
5 Modeling
The experimentally obtained results highlight the different behavior for both fiber types
on all scales, but time and space constraints limit the number of test samples as well as
the duration of the creep test to 1 or several years at most. However, most structures are
designed with at least a 50 years life span in mind. Finite element modelling (FEM) can
be used to verify the SLS and ULS requirements at these time scales. In this chapter, a
numerical approach to determine the crack width growth of cracked FRC under uni-
axial tension is presented. This approach is based on previous work [6] where cored
cylinders were notched at mid-height and subjected to sustained uniaxial tensile
loading. The load was applied after pre-cracking the specimen to 0.2 mm. The
numerical approach does not consider the pre-cracking procedure itself, rather it only
simulates the time-dependent crack widening upon and after load application. The
approach is validated for 2 types of polymeric FRC investigated before and is
implemented as such in this research project.
The numerical approach consists of three steps: (i) pre-processing and fiber gen-
eration, (ii) construction and analysis of a finite element model with discrete fibers and
(iii) post-processing and analysis.
5.1 Fiber Generation Algorithm

The fibers are modelled discretely in the numerical model, and their positions within a
cylindrical concrete core (diameter 100 mm, height 300 mm) are determined in a fiber
generation algorithm (FGA), written in MATLAB. The FGA requires the input of the
concrete volume, fiber geometry and fiber fraction and distributes straight fibers within
the volume without overlapping or intersecting, according to a user-defined distribu-
tion. Here, a random 3D distribution is assumed, and since the pre-cracking itself is not
modeled, the crack faces are assumed perfectly flat. The fibers are divided into bond-
slip and fully embedded fibers, whether the fiber crosses the crack face or not,
respectively. In Fig. 6, all fibers are shown in the core on the left and middle figure,
with the contributing fibers, i.e. those crossing the flat crack faces, shown in red. On the
(right), one half of the numerical model is also shown.
5.2 Finite Element Model: Material Models and Analysis Procedure

From the fiber distribution, a finite element model (FEM) with discrete fibers is built in
the numerical program DIANA. The material models are calibrated based on the
experimental results, discussed before. The model consists of 4 element types: (1) 3D
solid elements for concrete, (2) embedded reinforcement elements for the embedded
Fig. 6. Fiber distribution in the whole core (left and center), one half of the FEM model with
fibers crossing the crack (right)
fibers, (3) bond-slip reinforcement elements for the contributing fibers and (4) 3D beam
elements for the fiber part in the crack itself.
For the concrete (1), a linear-elastic material model is assumed as the tested
material is uniaxial strain softening and the initial pre-crack is explicitly modelled at the
start of the finite element analysis (FEA). The embedded fibers (2) adapt the stiffness of
the mother material in which they are embedded, based on the stiffness of the fibers, but
since they do not cross the crack, they take up only very limited forces. As such, the
embedded fibers are elastic with the elastic parameters determined in the short-term
tests and are assumed to be perfectly bonded with the concrete. The bond-slip fibers
(3) behave elastic as well, but in contrast with the embedded fibers, they can undergo
deformations relative to the mother elements. The pull-out displacement upon loading
is calibrated with the pull-out tests, where the material model requires the input of the
pull-out stiffness parameter, i.e. the initial slope in Fig. 3. The effect of the embedded
length and angle is effectively neglected by adopting an identical pull-out stiffness for
all fibers, which is verified for the pull-out data set presented here. The 3D beam
elements in the crack (4) can undergo creep deformations based on the single fiber
creep deformations presented in Fig. 2. In the FEM, fiber creep is implemented with a
time-dependent stiffness and based on analyses presented in [6] the pull-out creep can
be lumped in the single fiber creep for polymeric fiber types. As such, the model can
capture single fiber creep as well as (implicitly) pull-out creep.
The applied load is 50% of the residual strength of the specimen at the pre-crack
width of 0.2 mm as discussed before. The residual strength is numerically determined
as no uniaxial tensile tests are performed. The imposed load is 950 N and 1450 N for
fiber type A and B, respectively, and it is applied centrically and is kept constant for 50
years, divided into 125 logarithmically spaced time steps. A quasi-Newton iterative
scheme with a simultaneous force and displacement convergence criterion of 5% is
used.
5.3 Analysis and Discussion

The evolution of the crack width in mm and the corresponding creep coefficient evo-
lution, u, is shown in Fig. 7 on the left and right respectively. It is noted that the absolute
value of the crack width is greater for type B fibers, even though they outperform the
type A fiber, as shown in Fig. 1. It is attributed to the nearly 50% increase in load (950 N
to 1450 N), even if the load ratio remains constant at 50%. The superior performance of
the type B fiber is more clearly shown in the creep coefficient evolution.
0.4 1.8
Type A Type A
1.7
Type B Type B
0.35
1.6
1.5
Creep coefficient [-]

0.3
COD [mm]
1.4
0.25 1.3
1.2
0.2
1.1
0.15 1
1 s. 1 min. 1 h. 1 d. 1 m. 1 y. 10 y. 50 y. 1 s. 1 min. 1 h. 1 d. 1 m. 1 y. 10 y. 50 y.
Time
Time
Fig. 7. (left) Crack width evolution (right) creep coefficient evolution for both fiber types
While the load application is fundamentally different for the experimental results
shown at right in Fig. 5 (bending) and the numerical model (uniaxial tension), the ratio
of the creep coefficients of type A with respect to type B can be compared (Fig. 8).
While both methods yield different values for the coefficient, a similar conclusion can
be drawn, as both methods indicate that the type B fiber outperforms type A (ratio is
greater than 1, higher creep coefficients for type A). Additionally, for the time period
considered in the bending test, an increasing ratio in time can be observed (unsteady
course of experimental graph is a result of calculating the ratio of parameters only
slightly changing). In that sense, the numerical methodology and results are in line with
the experimental, bending results.
1.3
Uniaxial (FEM)
Bending (EXP)
1.25
1.2
ratio [-]
1.15
1.1
1.05
1
1 s. 1 min. 1 h. 1 d. 1 m. 1 y. 10 y.50 y.
Time
Fig. 8. Ratio of the creep coefficient for fiber type A with respect to type B in bending and
uniaxial tests
6 Conclusion
The work presented in this paper can be summarized as follows:

1. Standards and guidelines, due to limited understanding, are not yet considering
creep of FRC in a satisfying way.
2. A holistic approach considering creep from material, over interface to composite
level is suggested to better understand decisive phenomena.
3. Fiber creep and pull-out creep results indicate that a high performing PP-macrofiber
can sustain a load level of 40% related to maximum strength in short-term tests even
at a temperature of 40 °C. This would be appropriate for a lot of applications in
construction.
4. While for a temperature of 20 °C the superior performance is proven on composite
level (FRC), results at 40 °C still need to be verified in beam tests.
5. The importance of fiber creep for the creep of FRC with synthetic fibers is con-
firmed. Other factors, such as fiber shape (crimping vs. embossing) may have an
influence as well.
6. Finally, it seems feasible to simulate the crack opening due to creep on synthetic
FRC by considering test results in a model recently developed at KU Leuven [6].
Further proof is needed.
Investigations presented here shall be continued to understand creep of FRC up to
an extent that engineers are able to consider this factor in structural design calculations.
References
1. Kusterle, W.: Viscous material behavior of solids – creep of polymer fibre reinforced
concrete. In: 5th Central European Congress on Concrete Engineering, Baden, pp. 95–100
(2009)
2. Pujadas, P., Blanco, A., Cavalaro, S., de la Fuente, A., Aguado, A.: The need to consider
flexural post-cracking creep behavior of macro synthetic fiber reinforced concrete. Constr.
Build. Mater. 149, 790–800 (2017)
3. Plizzari, G., Serna, P.: Structural effects of FRC creep. Mater. Struct. 51(6), 1–11 (2018)
4. Balaguru, P., Ramakrishnan, V.: Properties of fibre reinforced concrete: workabilty, behaviour
under long-term loading and air-void characteristics. ACI Mater. J. 85(3), 189–196 (1988)
5. Babafemi, A.J.: Tensile creep of cracked macro synthetic fibre reinforced concrete. Ph.D.
thesis. Stellenbosch University, South Africa (2015)
6. Vrijdaghs, R.: Creep of synthetic fiber reinforced concrete - a multi-scale and two-phased
approach. Ph.D. thesis. KU Leuven, Belgium (2019)
7. DIBt: Prüfplan für die Zulassungsprüfung von Polymerfasern zur Verwendung in Beton nach
DIN EN 206-1 in Verbindung mit DIN 1045-2 mit nachgewiesener Wirksamkeit. Deutsches
Institut für Bautechnik, Berlin, January 2013
Durability
Morphology of Corrosion of Metallic Fibers
in Aggressive Media
Carmen Andrade1 and Miguel A. Sanjuán2(&)

1
International Center of Numerical Methods in Engineering, CIMNE-Madrid,
Madrid, Spain
2
Institute for Cement and Its Applications, IECA, Madrid, Spain
masanjuan@ieca.es
Abstract. In marine and industrial environments, the steel bars embedded in

concrete can corrode producing the cracking of the cover and impacting nega-
tively in the load-bearing capacity. Metallic fibres have shown a record of good
mechanical performance as reinforcing material for concrete. Those on carbon
steel can corrode in the aggressive environments and although publications have
reported a better impermeability in the fibre reinforced concretes, it still remains
the question of how the steel fibres corrode and, in this case,, whether they can
microcrack the surroundings. In present communication are presented results of
long-term performance of the fibres due to chlorides and carbonation attacks
during more than 20 years of exposure. The carbon fibres corrode but not
cracking seems to be produced in spite of the full conversion of the fibres into
oxides. The galvanized fibres corrode comparatively less, and the stainless-steel
ones are in perfect condition.
Keywords: Corrosion Metallic fibres Durability Aggressive environment

Concrete
1 Introduction
The introduction of fibres into the concrete is due to the historical disadvantage that
both cement mortar and concrete have with regard to their low tensile strength and high
brittleness. To counteract these characteristics, they are normally controlled by means
of steel reinforcement. As a measure to control the cracking, fibre reinforced concretes
were also used from the second half of the 20th century.
Steel fibre reinforced concrete (SFRC) is a composite material, combining a
Portland cement and steel fibres as a discontinuous reinforcement, which are randomly
distributed in the cement paste. The cement content in SFRC usually is ranged between
300 kg/m3 and 450 kg/m3. Lower amount of cement leads to a loss of workability.
SFRC needs from 35% to 45% by volume of cement paste while the plain concrete
only needs from 25% to 35%. SFRC is increasingly being used for the production of
slabs, tunnel linings, foundations, thin-shell structures and so on [1, 2].
From the durability viewpoint, corrosion of metallic fibers in aggressive environ-
ments is a risk due to the discontinuous nature of the metallic fibres in the SFRC. The
first effect is aesthetic as consequence of the surface corrosion [3]. Later, the corrosion
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_38
418 C. Andrade and M. A. Sanjuán
of the internal steel fibres may affect to the strength of the concrete structure [4, 5].
Given that, currently, the total replacement of steel bars reinforcement by SFRC is
controversial when the long-term durability of SFRC under severe environments is
addressed [6].
In present communication results are presented on the morphology of the corrosion
of fibers and whether the oxides have the expansive character as in normal
reinforcements.
2 Experimental
The main components found in the fibre reinforced concretes are the Portland cement,
both fine and coarse aggregates, water, superplasticizer and steel fibres. Therefore, the
only significant variation from a normal concrete is the addition of the fibers. Different
types of steel fibre reinforced concrete pipes have been studied for 30 years.
2.1 Materials
Portland cement CEM I 42.5 R, siliceous aggregates (sand and gravel), tap water and
superplastifier were used in the present research program. Three types of hooked steel
fibres were selected to produce the fibre reinforced concrete: carbon steel, stainless steel
and galvanized steel.
2.2 Aggressive Solutions

An aggressive solution of Na2S 9H2O (2.9 g/l) and NH4Cl (29 g/l) was prepared. In
addition, tap water was used as reference media.
2.3 Fibre Reinforced Concrete

Fibre reinforced concrete made of carbon steel, stainless steel and galvanized steel
cylindrical specimens (ø15 30 cm) were manufactured according to the mix design
shown in Table 1. Concrete slump test was performed to check the workability of
freshly made concrete, and the result was 5 cm. These specimens have an internal hole
of 5 cm of diameter in order to simulate a concrete pipe. Concrete was cured for 28
days under water and, later on, two series of the pipe specimens were subjected to the
aggressive solution (Na2S 9H2O (2.9 g/l) and NH4Cl (29 g/l)) and other two series of
the pipe specimens were kept in tap water. The concrete specimens’ layout is shown in
Fig. 1 (Table 2).
Table 1. Concrete mix design (kg/m3).

Constituent Cement Sand Gravel Water Water with superplastifier Steel fibres
Content 350 820 1110 190 154 35
Morphology of Corrosion of Metallic Fibers in Aggressive Media 419
SoluƟon
level
Fig. 1. Steel fibre reinforced concrete specimens’ layout.
Table 2. Steel fibre reinforced concrete codes.

Water/cement Witout Steel Steel fibres with Stainless Galvanized
ratio fibres fibres 3%wt Ca(NO2)2 steel fibres steel fibres
0.55 1 2 3 4 5
0.44 (with 6 7 8 9 10
superplastifier)
After the first months of testing, some of the specimens were held at the atmosphere
for 8–9 years (sheltered from rain) while other were kept in the aggressive solutions.
Some remaining after the atmospheric conservation where introduced in NaCl solution
for further years.
The analysis of the results was not systematic and is out of present work to give
rates for deterioration. However, having been submitted to aggressive solutions, some
of the fibres fully corrode and have allowed us to study which is the kind of damage
produced.
3 Results
Some aspects of the specimens will be presented in order to appreciate the oxides and
their impact in the concrete. The specimens immersed in tap water with and without
nitrites, presented corrosion at 4 months of exposure in the superficial fibres as shown
in Fig. 2 left. However, galvanized and stainless-steel fibres were found free of
corrosion.
In the specimens held in the aggressive solutions, those without fibres presented
cracks and expansion.
On the opposite the stainless-steel fibres were not corroding as can be appreciated
in Fig. 3.
Fig. 2. Aspect of the specimens after 4 months in tap water (left) and in the aggressive solutions
(right).
Fig. 3. Concrete with stainless steel fibres after 28 years of being in contact to aggressive
solutions, air and sodium chloride solution.
The aspect of the corroding fibres is shown in Fig. 4. In the left is shown that all the
internal fibres were corroding 4 months after of attack and in the right after 28 years.
Fig. 4. Aspect of the concrete with the fibres fully corroded.

Morphology of Corrosion of Metallic Fibers in Aggressive Media 421
Figure 5 shows aspects of the oxides and how they diffuse out of the fibre. In the
right can be detected that the concrete surrounding the fibre is not cracks in spite that in
some places the fibre has disappear completely.
Fig. 5. Aspect for the oxide’s diffusion from the fibres when corrosion is active. No cracking
around the fibres was identified.
4 Discussion
Due to the diverse environments and conditions where the specimens were held a
correct rate of degradation was not feasible, but however the work enables to deduce
well the morphology of the corrosion and that the concrete surrounding the fibres does
not seem to crack. That is, the fibres dissolve without producing visual damage in the
concrete, even at its surface. In the surface the fibres dissolve too until disappearing but
without visual cracking.
This performance can be explained in Fig. 6 where it has been tried to indicate that
the controlling factor for producing cracking due to corrosion oxides depends on the
ratio diameter of the bar/cover depth. This ratio expresses also the ratio of the surface
where the pressure is produced due to the expansive character of the oxides with
respect to the volume around the rod.
Fig. 6. Importance of the ratio cover depth/bar diameter for the cracking due to reinforcement
corrosion.
The oxides occupy a higher volume than the parent metal due to oxides have
oxygen and water in addition of the metallic atoms. The cracking of the cover due to
the pressure they impose depends on how large the cover thickness is. If it is large
enough the cracks do not reach the external surface.
In addition to this effect, we attribute to the cover thickness much larger in the case
of internal fibres, in the case of superficial fibres, very close to the surface but not
protruding, the oxides can however produce cracking which is mixed with the diffusing
oxide and the spalling is very superficial. The fibres protruding from the surface they
directly are dissolved and detached form the concrete mass due to the loss of bond.
The performance of the oxides detected enables to support the idea of a sacrificial
thickness in fibre reinforced concretes when in contact to aggressive solutions. The
corrosion is going to produce a concrete surface thickness where the fibres will dis-
appear and with this disappearance their reinforcing effect.
5 Conclusions
The specimens with stainless steel have been maintained for 28 years free of corrosion
in spite of the very aggressive solutions tested. Galvanized fibres presented a not so
good performance because some of them corrode and other not, Bare steel has corroded
in the aggressive media and in tap water. The morphology pf the corrosion shows
however that not microcracking is visually detected in the concrete surrounding the
bars which is indicative that the concrete in the attacked thickness is not to weaken and
become more permeable.
References
1. Serna, P., Arango, S., Ribeiro, T., Núñez, A.M., Garcia-Taengua, E.: Structural cast-in place
SFRC: technology, control criteria and recent applications in Spain. Mater. Struct. 42, 1233–
1246 (2009)
2. di Prisco, M., Plizzari, G.A.: Precast SFRC elements: from material properties to structural
applications. In: di Prisco, M., Felicetti, R., Plizzari, G.A. (eds.) 6th International RILEM
Symposium on Fibre-Reinforced Concretes, pp. 81–100. RILEM Publications SARL,
Varenna, Italy (2004)
3. Balouch, S.U., Forth, J.P., Granju, J.-L.: Surface corrosion of steel fibre reinforced concrete.
Cem. Concr. Res. 40, 410–414 (2010)
4. Mangat, P.S., Gurusamy, K.: Permissible crack widths in steel fibre reinforced marine
concrete. Mat. Struct. 20, 338–347 (1987)
5. Mangat, P.S., Gurusamy, K.: Corrosion resistance of steel fibres in concrete under marine
exposure. Cem. Concr. Res. 18, 44–54 (1988)
6. Hwang, J.P., Jung, M.S., Kim, M., Ann, K.Y.: Corrosion risk of steel fibre in concrete.
Effects of Fibres on the Flexural Behaviour
of Sound and Damaged RC Beams
Raúl L. Zerbino1(&), María C. Torrijos1, Graciela M. Giaccio2,

and Antonio Conforti3
1
CONICET, LEMIT-CIC, Faculty of Engineering, UNLP, La Plata, Argentina
zerbino@ing.unlp.edu.ar
2
LEMIT-CIC, CIC Researcher, Faculty of Engineering, UNLP, La Plata,
Argentina
3
Department of Civil, Environmental, Architectural Engineering and
Mathematics (DICATAM), University of Brescia, Brescia, Italy
Abstract. The incorporation of fibres in Reinforced Concrete (RC) beams

controls the width and evolution of cracks leading to positive effects on the
durability of the element. The study of damage processes in concrete and their
effects on the residual properties represents a key point related to the service life
of RC structures. The contribution of fibres on the bending behaviour of sound
and damaged RC beams was investigated. In order to use alkali silica reaction as
a damaging tool, RC beams with and without fibres and reactive aggregates
were subjected to service loading conditions during eight months in an envi-
ronment with high humidity. The evolution of deformations and the distribution
and propagation of cracks were recorded. As reference, similar RC beams
without reactive aggregates were also evaluated. After the treatment, all RC
beams were loaded up to failure. The free expansion, the compressive strength
and the bending residual capacity of plain and fibre concrete were measured on
companion specimens for material characterization. The effect of both fibres and
alkali silica reaction on bearing capacity and ductility of RC beams were
analysed. Results showed that alkali silica reaction damage provokes a signifi-
cant reduction of RC beam ductility, while the flexural strength is preserved.
Keywords: Alkali silica reaction Crack control Flexure Reinforced

concrete Steel fibres
1 Introduction
Alkali Silica Reaction (ASR) represents one of the processes of degradation of concrete
that generates most interest. Much progress has been made in terms of minerals and
causes that produce ASR as well as in the criteria and methods of evaluation and
prevention. Nevertheless, there is not much work on the development of ASR under
load [1–6] and even less under tensile loads. However, considering that the presence of
cracks substantially affects concrete permeability and water is essential for the ASR, it
is possible to infer that the kinetics of the reaction should be related to the evolution of
tensile cracking.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_39
424 R. L. Zerbino et al.
Fibre Reinforced Concrete (FRC) is a high-performance material with great pos-

sibilities for structural design as seen in the fib Model Code 2010 [7]. The incorporation
of fibres into concrete controls the propagation of fissures and increases residual
capacity and toughness. In addition to FRC traditional applications such as floor slabs
or tunnel lining, the combined use of FRC y Reinforced Concrete (RC) structural
elements improves shear resistance allowing the reduction of conventional steel bars
and improves steel-concrete bond, then, reductions in anchorage length are possible. In
addition, as fibers control crack width they give important benefits in terms of the
extension of service life. Many papers confirm that the use of steel fibers reduces cracks
width and spacing [8–10]. In a study on steel FRC elements loaded under uniaxial
tension, increases in toughness and reductions in crack spacing were found [11].
Although when combined with conventional reinforcements, fibers do not significantly
increase flexural strength and ductility in Ultimate Limit State [12] there are benefits in
Service Limit State referred to the control of cracks and deflections, as they enhance the
transfer of tensile stresses from the reinforcements to the concrete. The synergy
between this mechanism and the FRC residual tensile capacity reduces the width and
spacing of cracks. Finally, although the presence of fibres does not prevent the ASR, it
reduces the expansions; mainly in the case of steel ones [13]. As a part of a research
project on the advantages of using FRC for the extension of the service life and
durability of the structures, this paper shows the effects of fibres on the development of
ASR in RC beams subjected to long-term loads.
Four different types of concrete were used: two incorporating non-reactive aggregates
and two with potentially reactive coarse aggregates, and, in each case, one without
fibres and the other with steel fibres. Four RC beams, one for each concrete, were
subjected to service loads. Twin beams were also cast and they remain unloaded and
exposed to the same environmental conditions.
2.1 RC Beams
Beams of 150 150 900 mm including two 8 mm diameter steel bars as main
reinforcement (q = 0.53%) and 6 mm diameter stirrups (spaced 50 mm) were cast.
They were loaded using 4 PB configuration on a span of 840 mm; Fig. 1 shows the
geometry and load configuration adopted. Tensile and compression deformations were
continuously recorded by mechanical extensometers.
The study was carried out in a temperature controlled chamber (23 ± 2 °C) and, to
promote ASR all samples were covered with wet cotton cloths and then were isolated
by a plastic film and remain in that condition during all the creep test; in addition, water
was periodically injected (see Fig. 2). The beams were demoulded at 24 h and moist
cured (as described) for 7 days. Then, they were removed from the bags, instrumented,
wrapped again with humid cloths and plastic film and loaded by a lever system.
Unloaded reference RC beams also remain in the same conditions. The RC beams
Effects of Fibres on the Flexural Behaviour 425
5 stirrups 5 stirrups
150
Ø6@5cm Ø6@5cm
150 60
150
25
2Ø8
35 400
80 280 280 280

900
Fig. 1. Geometry and reinforcement details of the RC beams (in mm).
remain during 8 months under loading and then they were tested in bending up to
failure using the same loading configuration.
Fig. 2. Load RC beams (left) and reference unloaded ones (right).
2.2 Concretes Mixture Design and Properties

Two plain concrete (R, P) with similar mixture proportions, w/c ratio 0.42; 380 kg/m3
of ordinary portland cement (Na2Oeq 0.73%), natural siliceous sand (fineness modulus
2.07) and granitic crushed stone of 19 mm maximum size were done. To promote
ASR, 40% of the coarse aggregate was replaced by a very reactive quartzitic sandstone
in the first of them (R) and NaOH was added in the water to achieve a total alkali
content in concrete equal to 4 kg/m3. Two FRC, one reactive and the other non-reactive
(FR, FP), were prepared adding 40 kg/m3 of low carbon hooked ended steel fibres to
the same matrices (R, P). The slump was 170 ± 20 mm on concretes R and P and
90 ± 20 mm when fibres were incorporated.
In addition to RC beams, three 70 70 300 mm prisms to measure the evo-
lution of the length variations, six 100 200 mm cylinders to evaluate the com-
pressive strength and the modulus of elasticity (only in R and P concretes) and six
75 105 430 mm prisms to characterize the flexural response were cast.
All the specimens were compacted by external vibration and protected to prevent
water evaporation. Both RC beams and the rest of the specimens were demoulded after
24 h, covered with wet cotton cloths and stored in plastic bags. Throughout the study
the ambient temperature was maintained at 23 ± 2 °C.
At 7 days RC beams were loaded in the frames as already described. Simultane-

ously, the compressive strength and the modulus of elasticity and the flexural behaviour
of each concrete were evaluated on three cylinders and three notched prisms respec-
tively. Bending tests were performed following the general guidelines of EN 14651
[14, 15]. The rest of the cylinders and prisms were tested at the end of the experience
(near 9 months) in order to evaluate the impact of the ASR on the mechanical prop-
erties of the concrete.
3 Results
3.1 Plain and FRC Properties

Figure 3 shows the linear expansions measured on 70 70 300 mm prisms of each
concrete. It can be seen that after near 3 months R and FR show expansions greater than
0.04%, while in the reference mixtures (P and FP) the dimensional changes did not
exceed 0.01%. It can be seen that ASR expansions are smaller in FRC; although the
incorporation of fibres cannot inhibit ASR, it leads to some benefits such as a decreases in
expansions and, even more important, reductions in the width and length of cracks [16].
0.15
FR
0.12 R
FP
0.09
Expansion (%)
0.06
0.03
0.00
-0.03
0 60 120 180 240 300 360
Time (days)
Fig. 3. Evolution of expansions (FR: reactive FRC; R; reactive concrete; FP: non-reactive FRC;
P: non-reactive concrete).
Figure 4 shows typical stress-CMOD curves of each concrete at 7 days and 9

months. The effect of ASR damage on the mechanical behaviour of plain and FRC is
clearly appreciated when compared the post-peak responses.
The average results of compression and bending tests at both ages are given in
Table 1. As it can be seen, no great differences in the mechanical properties at 7 days
are present, indicating that ASR damage has not yet occurred at that age. On the
contrary, at later ages there is a significant decrease in compressive strength and
especially in the elastic modulus of concrete R.
Fig. 4. Stress – CMOD curves on notched prisms at 7 days (left) and 9 months (right) (FR:
reactive FRC; R; reactive concrete; FP: non-reactive FRC; P: non-reactive concrete).
Table 1. Mechanical properties of concretes (CV between brackets).

Concrete R FR P FP
Fibre content (kg/m3) Age 0 40 0 40
f’c (MPa) 7 days 42.5 (4) 44.5 (4)
9 months 32.2 (18) 47.8 (6)
E (GPa) 7 days 38.5 (4) 40.2 (4)
9 months 19.3 (20) 38.5 (3)
fL [MPa] 7 days 5.5 (3) 6.9 (3) 5.8 (6) 6.3 (4)
9 months 4.4 (12) 5.1 (1) 6.2 (9) 7.8 (11)
fR,1 [MPa] 7 days 5.8 (6) 5.1 (17)
9 months 4.1 (31) 7.6 (12)
fR,3 [MPa] 7 days 5.6 (4) 5.0 (12)
9 months 3.2 (28) 6.1 (8)
Regarding the bending behaviour, reactive concretes (R, FR) evidence the internal
damage in a lower initial slope (due to internal cracking) when compared to the one
observed in concretes P and FP, a reduction in the first peak load and, in the case of
plain concrete, an increase in the softening branch. In FR the first peak, the maximum
and the residual loading capacity decrease, however, it conserves a residual stress near
4 MPa. The comparison against P and FP is difficult since the crack opening before
testing in FR and R prisms is different than cero (i.e. samples are pre-cracked by ASR).
3.2 Instantaneous and Creep Behaviour of RC Beams Under Loading

Once placed in the frames, RC beams were loaded measuring tensile and compression
deformations in the middle third. Considering that the ultimate estimated load in
bending would be between 50 and 65 kN, 24 kN were applied enhancing the devel-
opment of flexural cracks as in SLS condition. In every case, when the load exceeded
18 kN (near tensile stress 4 MPa) a clear deviation from the linear behaviour was
observed during initial loading, indicating the presence of cracks (see Fig. 5).
Stress (MPa)
4
Compression Tensile
face face
2
0
-500 -250 0 250 500 750
Strain (10-6)
Fig. 5. Example of instantaneous deformations of a beam during the loading process.
Figure 6 shows the evolution of deferred deformations measured on the tensile face
of loaded and unloaded RC beams. It can be seen that the specimens incorporating
reactive aggregates (R, FR) show a clear increase in the deformations; no great dif-
ferences between specimens with and without fibres can be detected. Nevertheless, in
the case of beams which are not subject to long-term loading (and therefore are not pre-
cracked), slight expansions (<100 microstrains) product to a high humidity environ-
ment appeared in three cases while in concrete R the expansion growth from about 45
days indicating the presence of ASR.
0.15 0.15
Loaded beams FR
Unloaded beams
R
0.12 0.12
FP
Deformation (%)
Deformation (%)
P
0.09 0.09
FR
R
0.06 FP 0.06
P
0.03 0.03
0.00 0.00
0 30 60 90 120 150 180 210 240 0 30 60 90 120 150 180 210 240
Time (days) Time (days)
Fig. 6. Evolution of deferred deformations on loaded (left) and unloaded (right) RC beams (FR:
reactive FRC; R; reactive concrete; FP: non-reactive FRC; P: non-reactive concrete).
After 5 months under loading a survey of the RC beams surface was performed.
In R and FR gel points were observed together with several cracks of approximately
0.05 mm width in the R beam. The crack widths were much smaller in FR being
necessary to moisten the specimen surface to see the cracks. It must be noted that not
only the expected vertical cracks were found but also cracks with different orientations.
Only very small cracks (thickness 0.05 mm) were detected in concretes P and PF
which cannot be easily differentiate to those produced during loading (see Fig. 7).
Fig. 7. Appearance of the surface of loaded R and FR beams at the age of 5 months.
3.3 Residual Bending Behaviour of RC Beams

Figure 8 presents the load-deflection curves of RC beams after creep test, both that
loaded and those that remained unloaded. In addition, Table 2 shows some cracking
characteristics measured on the surface of damage beams (R, FR) at 9 months and the
main flexural test results evaluated both in terms of bearing capacity and ductility.
80 80
60 60
Load (kN)
Load (kN)
40 40
FR FR
20 R 20 R
FP FP
Loaded beams P Unloaded beams P
0 0
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Deflection (mm) Deflection (mm)
Fig. 8. Load-deflection curves of RC beams after creep test. Left: creep loaded beams. Right:
beams that remained unloaded (FR: reactive FRC; R; reactive concrete; FP: non-reactive FRC; P:
non-reactive concrete).
Regarding loaded beams, it can be observed that the typical flexural failure
expected for under reinforced beam section (i.e. rebar yield before the concrete in the
compression zone reached its maximum usable strain, CC) was observed only for RC
beam without alkali silica reaction. In fact, both FR and R showed a flexural collapse
characterized by rebar failure. This aspect can be also observed in Fig. 9, where the
final crack patterns of FR and FP beams are shown and compared. In addition, in case
of fiber incorporation, crack localization after rebar yielding was observed as expected
[17]. This underlines that the damage due to alkali silica reaction changed the mode of
failure of the beam, leading to a significant reduction of beam ductility, relative duc-
tility (RD) equal to 0.72 and 0.63 for R and FR beam, respectively. RD was evaluated
as (dmax/dy of FR)/(dmax/dy of FP) for concretes with fibers and (dmax/dy of R)/(dmax/dy
of P) for concretes without fibers. The reduction of ductility was even more pronounced
in the FRC beam as a consequence of the crack localization due to fibers. In terms of
Table 2. Survey of cracks before testing in concretes affected by ASR and the results of flexural
tests on RC beams.
Before testing Results of flexural tests
Concrete Crack Crack Type of flexural Pmax DPmax dy dmax dmax/ RD
density width failure dy
mm/mm2 mm kN % mm mm
Loaded FR 0.029 0.07 RF 63.6 16 1.86 11.7 6.29 0.63
R 0.023 0.20 RF 54.6 2.07 15.5 7.51 0.72
FP – – CC + CL 63.3 16 1.92 19.3 10.05
P – – CC 54.6 1.90 19.8 10.42
Unloaded FR 0.026 0.05 CC + CL 62.7 18 1.90 20.0 10.54 0.99
R 0.011 0.01 CC + CL 53.1 1.76 17.3 9.85 0.94
FP – – CC + CL 59.8 15 1.90 20.2 10.63
P – – CC 52.1 1.96 20.6 10.51
Type of flexural failure: RF: rebar failure due to crack localization; CC: concrete crushing
without crack localization; CC + CL: concrete crashing with crack localization.
Pmax: Maximum load.
DPmax: Increase in load due to fibre incorporation.
dy: Net mid-span deflection at rebar yielding.
dmax: Maximum net mid-span deflection.
RD: Relative ductility of damaged concrete (R/P; FR/FP).
bearing capacity, it can be observed that the damage due to alkali silica reaction does
not reduce the flexural bearing capacity. This is due to the fact that the main parameter
influencing the flexural bearing capacity of RC beams is the rebar amount and position,
while the concrete compressive strength (which is affected by alkali silica reaction,
Table 1) has a secondary role.
FR FP
Fig. 9. Final crack pattern of loaded beams after testing: FR (left) and FP (right).
Concerning unloaded beams, it can be underlined that all beams showed a flexural
failure characterized by concrete crushing after rebar yielding (no rebar failure was
observed). However, the presence of either fibres or alkali silica reaction damage led to
a premature crack localization, which is not generally observed in under reinforced
beams under flexure. Since this crack localization did not affect the beam ductility, the
overall ductility of the different beams resulted comparable. In fact, RD is equal to 0.99
and 0.94 for R and FR beam, respectively. This small reduction of beam ductility in the
unloaded specimens damaged by alkali silica reaction is probably due to the low
mechanical properties of concrete in the compression zone. Finally, the maximum
bearing capacity seems once again to be not affected by alkali silica reaction damage.
4 Conclusions
The influence of alkali silica reaction damage on the flexural behaviour of RC beam
with and without fibres was studied by means of an experimental program on small
beams subjected to service loads or not. Based on this experimental program, the
following conclusions might be drawn:
• With the selected materials and proportions, a concrete was achieved where the
damage caused by ASR was manifested for the times of the proposed research; after
a month the signs of reaction were already evident. From that period, the standard
prisms showed expansions that exceeded 0.04% and the ASR was also manifested
in the RC beams.
• The flexural behavior of RC beams damaged by alkali silica reaction is different
considering loaded or unloaded beams. In fact, alkali silica reaction effects are more
severe in service loading condition.
• Alkali silica reaction damage significantly affects the ductility of RC beams, due to
premature crack localization after rebar yielding. This effect is even more pro-
nounced in case of fibre reinforced concrete.
• The alkali silica reaction damage (corresponding to a reduction of 25% of concrete
compressive strength) does not clearly affect the flexural bearing capacity of RC
beams.
Acknowledgements. The authors specially thank the collaboration of Eng. Diego Monetti and
Pablo Bossio on the support of the experimental works and to Cementos Avellaneda SA and
Maccaferri SA for the provision of the cement and the steel fibres respectively. Funding from
LEMIT-CIC and from projects of National Scientific and Technical Research Council (CON-
ICET) PIP112-201501-00861 Advances in Fibre Reinforced Concretes, and UNLP 11/I188
Fibre reinforced concretes and other composites for the construction and repair of sustainable
infrastructure, is greatly appreciated.
References
1. Monette, L., Gardner, J., Grattan-Bellew, P.: Structural effects of the alkali-silica reaction on
non-loaded and loaded reinforced concrete beams. In: 11th ICAAR, Quebec, pp. 999–1008
(2000)
2. Jones, A.E., Clark, L.A.: A review of the institution of structural engineers report: structural
effects of alkali-silica reaction (1992). In: 10th ICAAR, Melbourne, pp. 394–401 (1996)
3. Takemura, K., Ichitsubo, M., Tazawa, E., Yonekura, A.: Mechanical performance of ASR
affected nearly full-scale reinforced concrete columns. In: 10th ICAAR, Melbourne,
pp. 410–417 (1996)
4. Koyanagi, W., Rokugo K., Uchida, Y., Iwase, H.: Deformation behavior of reinforced
concrete beams deteriorated by ASR. In: 10th ICAAR, Melbourne, pp. 458–465 (1996)
5. Multon, S., Toutlemonde, F.: Effect of applied stress in alkali-silica reaction-induced
expansions. Cem. Conc. Res. 36, 912–920 (2006)
6. Giaccio, G., Torrijos, M.C., Tobes, J.M., Batic, O.R., Zerbino, R.: Development of alkali-
silica reaction under compressive loading and its effects on concrete behavior. ACI Mat.
J. 106(3), 223–230 (2009)
7. Féd Int du Béton: fib Model Code for Concrete Structures 2010. Ernst & Sohn, Berlin (2013)
8. Mitchell, D., Abrishami, H.H.: Influence of steel fibres on tension stiffening. ACI Struc. J. 94
(6), 769–773 (1997)
9. Bischoff, P.H.: Tension stiffening and cracking of steel fibre reinforced concrete. J. Mat.
Civil Eng. ASCE 15(2), 174–182 (2003)
10. Vandewalle, L.: Cracking behaviour of concrete beams reinforced with a combination of
ordinary reinforcement and steel fibers. Mat. and Struc. 33, 164–170 (2000)
11. Tiberti, F., Minelli, G., Plizzari, G.: Cracking behavior in reinforced concrete members with
steel fibers: a comprehensive experimental study. Cem. Conc. Res. 68, 24–34 (2015)
12. Meda, A., Minelli, G., Plizzari, G.: Flexural behaviour of RC beams in fibre reinforced
concrete. Compos. B Eng. 43(8), 2930–2937 (2012)
13. Giaccio, G., Bossio, M.E., Torrijos, M.C., Zerbino, R.: Contribution of fiber reinforcement in
concrete affected by alkali–silica reaction. Cem. Conc. Res. 67, 310–317 (2015)
14. EN14651:2005: Test method for metallic fibered concrete - measuring the flexural tensile
strength (Limit of proportionality (LOP), Residual), Brussels (2005)
15. Giaccio, G., Tobes, J.M., Zerbino, R.: Use of small beams to obtain design parameters of
fibre reinforced concrete. Cem. Conc. Comp. 30, 297–306 (2008)
16. Giaccio, G., Torrijos, M.C., Milanesi, C., Zerbino, R.: Alkali–Silica reaction in plain and
fibre concretes in field conditions. Mater. Struct. 52(2), 1–15 (2019). https://doi.org/10.1617/
s11527-019-1332-2
17. Conforti, A., Zerbino, R., Plizzari, G.A.: Influence of steel, glass and polymer fibers on the
cracking behavior of reinforced concrete beams under flexure. Struc. Conc. 20, 133–143
(2019)
Fiber Reinforced Concrete Elements Exposed
to Accelerated Corrosion
Camelia Negrutiu(&), Ioan Sosa, Bogdan Heghes, Oana Gherman,

and Horia Constantinescu
Faculty of Civil Engineering, Technical University of Cluj-Napoca,

Cluj-Napoca, Romania
camelia.negrutiu@dst.utcluj.ro
Abstract. This study evaluates the structural performance of high performance

fibre reinforced concrete elements affected by corrosion and its subsequent
deteriorations.
With large scale applications already set in place and enhanced mechanical
and durability performance, high and ultra-high performance concrete is the
construction material of the present days. However, long term performance of
structural elements exposed to harsh environments is still linked to the cracking
pattern. Fibre reinforcement significantly improves the tensile strength of con-
crete, but serviceability cracks are nevertheless inherent. These cracks are the
path through which the aggressive media travels to the embedded reinforcement,
leading to the corrosion of the steel, the diminution of concrete cross-section,
loss of reinforcement area, increased deformations and crack widths. Therefore,
the present research is focused on the corrosion of the embedded traditional
reinforcement and the loss of section in high performance fibre reinforced
concrete elements with and without sacrificial zinc anodes. For this purpose,
several small scale beam elements are exposed to artificial accelerated corrosion
and their flexural behaviour assessed in comparison to non-exposed control
elements.
Keywords: High performance concrete Accelerated corrosion
1 Introduction
High and ultra-high performance fibre reinforced concrete has the main components of
an ordinary concrete: cement, aggregates, water, admixtures and additives. But in the
mix design, the amount and size of these constituents are what confers the concrete its
high density, thus its superior mechanical and durability properties. Specifically, a
cement content of more than 600 kg/m3, fine silica sand instead of coarse aggregates
and industrial waste such as silica fume. Furthermore, high packing density is ensured
by a very low water/binder ratio and the subsequent necessary superplasticizers and the
end result is a uniform, dense concrete mix with a minimum compressive strength of
150 MPa and tensile strength between 7–10 MPa [1, 2]. If steam and thermal curing is
applied, a further increase in mechanical properties is obtained [2]. Cracking in service
is characteristic for a structural reinforced concrete member and it is a significant factor
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_40
434 C. Negrutiu et al.
of durability performance, so a way to minimize crack width is by adding fibres of

different lengths and diameters. As a result, the flexural tensile strength will raise up to
15–30 MPa [2, 3]. Furthermore, water permeability, carbonation and chloride resis-
tance, as well as the ability to withstand freezing and thawing cycles are also well
above the durability performances of ordinary concrete [3, 4]. In fact, due to the very
small pore size, the freezing point in the ultra-high performance concrete matrix is
approximately −30 °C considering a combined exposure of freezing temperatures and
5.0 wt% NaCl solution [5], a temperature rather rare met in natural environments. Even
so, the rehydration rate of the matrix is able to repair the freeze-thaw damage by filling
pores smaller than 100 nm [5]. In addition from the obvious immediate performance,
utilizing ultra-high performance concrete may address our current sustainability con-
cerns in construction, prolonging the service life, improving the long term environ-
mental impact and reducing the equivalent annual cost [1].
With these remarkable properties, utilizing ultra-high performance concrete may
seem to lengthen the service life of a structural element indefinitely. In un-cracked
concrete, the electrical resistance of the ultra-high performance concrete matrix with
steel fibre addition of up to 3% showed an increasing density opposing the ingress of
the chloride ions, whereas the embedded steel bar became more corrosion resistant in
time [6].
However, even very small cracks within the matrix could provide the pathway of
the aggressive media to instil the corrosion in the embedded reinforcement. The
cracking pattern was observed on notched specimens with one main limited crack at the
notch and multiple internal cracks growing from the lugs of the rebar, with very narrow
widths at the concrete surface and wider regions at the reinforcement level [7]. The
fibres themselves do not create a preferential path for the aggressive media at the
interfacial zone around the fibres, but the overall conductivity of the fibres could
influence the initial corrosion current [7]. Moreover, some might be interconnected to
touch the reinforcing bar, and for natural corrosion, due to the early presence of oxygen
and aggressive agents like chlorides, the passivity of fibres would first break down,
becoming anodes, thus reducing the corrosion in the steel bar. In those fibres that were
not in contact with the steel reinforcing bar, the corrosion rate could be expected to be
minor [7].
Some of the effects of corrosion are increased crack widths and deformations, lower
bending and shear capacity and lower bond strength between the concrete and the
reinforcement. Although bond toughness initially increases at corrosion ratios below
10%, it gradually, but insignificantly declines up to 15% [8]. Similar results were
obtained by other authors with critical corrosion ratios of 12% to 13.5%, far superior to
those for plain concrete specimens [9].
Bending tests on low scale specimens showed that after accelerated corrosion, plain
concrete beams lost about 13% of their ultimate load-carrying capacity and 9% of their
deflection capacity, compared to functionally-graded concrete beams, where the bottom
part of the cross section is made of a ductile fibre reinforced cementitious composite,
lost about 11% and 6% respectively. Local pitting was also observed in the stirrups [7].
Still, an increase in the corrosion current and subsequently the deterioration attributed
to corrosion, although limited in high density fibre reinforced concrete, cannot effec-
tively be controlled. The formation of the cracks in the matrix will ultimately lead to the
Fiber Reinforced Concrete Elements Exposed 435
development of other active points on the reinforcement than the initial one and
therefore, the control of the anodic current response is regarded as being critical for
preventing increasing corrosion rates in reinforced concrete composites rather than the
durability superiority of a high performance concrete [10].
2 Scope
The purpose of this research is to assess the behaviour of high performance fibre
reinforced concrete small scale elements with traditional reinforcement in the tensile
area. In order to simulate the service life, the elements were pre-cracked and then
exposed to an accelerated corrosion through an external electrical current source.
Commercial zinc anodes were used to extend the service life and to reduce the effects
of the corrosion. After the exposure, a half-cell potential test was performed to identify
eventual corrosion spots. Furthermore, the elements were tested to failure and the load
and the crack width was recorded. The corroded elements were compared to control,
non-corroded elements. The mass loss was also recorded for the exposed elements.
3.1 Concrete Composition

The fibre reinforced concrete composition is presented in Table 1 where the con-
stituents are expressed by weight of cement. The fibre content is 5% by wt% concrete.
Furthermore, commercially available zinc anodes (Mapeshield I 10) were also used in
half of the exposed elements. They can be identified by the colour blue in Fig. 2(a)
attached to the tensile reinforcement.
Table 1. Concrete composition (wt% cement).

Steel Fibres WHS 6/0.75/S-26B160 (Baum Cas Fibers SRL) 0.436
Water + Superplasticizer (Sika ViscoCrete 1040) 0.266
SiO2 (Elkem Materials) 0.222
Cement CEM II/B-M(S-LL) 42.5R StructoPlus Holcim 1
Fine sand (local origin) 0.645
Extra fine sand (local origin) 0.362
Two local sand types were used to develop the high performance fibre reinforced
concrete, a fine sand with the maximum aggregate size of 2 mm and an extra fine sand
with the maximum aggregate size of 1 mm. The sand is acquired from local sources.
The sand sieve percentages are presented in Table 2.
Table 2. Silica sand size

Aggregate size (mm) Fine sand percentages % Extra fine sand percentages %
1–2 20.7 –
0.5–1 24.52 12.1
0.250–0.5 40 79.64
0.125–0.250 12 7.86
Powder 2.8 0.4
3.2 Element Reinforcement and Flexural Testing

The elements were tested in a 3 point bending test. The vertical deformations and the
crack width at the middle of the clear span were recorded using displacement trans-
ducers. A schematic of the elements is presented in Fig. 1 and a set-up bending test is
presented in Fig. 2.
2ÿ 4 - Smooth OB37
100
10
2ÿ 8 - Ribbed S500
500 100
550
2ÿ 4 - Smooth OB37 2ÿ 4 - Smooth OB37
2ÿ 8 - Ribbed S500
Fig. 1. Small scale elements in 3 point bending test. (dimensions in mm)
The type of the longitudinal tensile reinforcement was European steel S500 with a
characteristic yielding stress fyk = 500 MPa, whereas the compressed longitudinal and
the transversal reinforcement was mild OB37 of local Romanian production with
fyk = 255 MPa.
3.3 Accelerated Corrosion Test

Simulation of reinforcement corrosion due to environmental actions is a long-term
procedure. Therefore, a simplified, more aggressive action was set in place through a
process of electrolysis. The set-up for the accelerated corrosion is similar to other
research [11]. A schematic view and a picture of the corrosion equipment are presented
in Fig. 3(a) and (b). The circuit of the electric current generated by an external source
consisted of the embedded reinforcement, a stainless-steel plate covering 3 faces of the
element and the 5% NaCl solution. The intensity of the current was set to a 3 mA/cm2
and the elements were exposed for 7 days. Previous research reports that a moderate
corrosion rises up to 0.5 lA/cm2 which can produce a loss of section of 5.7 lm/year
Fig. 2. (a) Steel moulds for the experimental elements; (b) Bending test.
and an oxide layering up to 17.3 lm/year, whereas stronger corrosion is produced by

an intensity current of 1 lA/cm2, which corresponds to a maximum of 11.5 lm/year
and an oxide layering of maximum 34 lm/year [12].
(a) (b)
Source of electric
+ current
Stainless
steel plate
(CATHOD) Recipient
(fiber glass)
Reinforced
Reinforcement concrete
(ANOD) beam
Support
5%NaCl solution
Fig. 3. (a) Schematic view of the corrosion process; (b) Corrosion equipment.
3.4 Experimental Operational Sequence

As previously stated, the purpose of this research is to assess the behaviour of high
performance fibre reinforced concrete small scale elements with traditional reinforce-
ment in the tensile area after accelerated corrosion. Before casting, each tensile bar was
cleaned of rust, weighted and connected to electric cables which later on facilitated the
electric circuit of the accelerated corrosion. Zinc sacrificial anodes were also placed on
the tensile reinforcement. After casting and curing, the elements were pre-cracked and
then exposed to an external electrical current source. At the end, half-cell potential tests
were performed before the elements were tested to failure. The reinforcement was then
extracted from the elements and the mass loss recorded. Although multiple elements
were designed and cast, only 5 were presented in the present study.
A complete sequence of operations is presented as follows: cleaning of the rein-

forcement of rust and imperfections; recording mass for each rebar; binding the rebars,
connecting electric cables to the tensile reinforcement, inserting zinc anodes; casting
the concrete; demoulding the second day and applying thermal treatment for 5 days;
pre-cracking to a crack width of 0.1 mm; exposing the elements to accelerated cor-
rosion for 7 days; half-cell potential mapping; testing to failure; removing the rein-
forcement and recording new mass.
The elements exposed to accelerated corrosion are: 1 (with zinc anode); 4 (without
zinc anode). The control elements are: 13; 14; 15. The elements are part of a 15
elements batch, with the same characteristics as the ones in the present article. The rest
of the elements will be subsequently tested and therefore, the identification of the
specimens, although not very eloquent at this time, will be correlated with the next

The compression strength was tested on 3 cubes 50 50 50 mm, with vertical
deformations. The results are presented in Fig. 4. The resulted compression strength is
fcm = 146.44 MPa with a standard deviation s = 2.82 MPa.
Stress - Displacement
Compression
160
150
140
130
120
110
Stress (MPa)
100
90
80
70 Cube1
60
50 Cube2
40
30 Cube3
20
10
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5
Displacement (mm)
Fig. 4. Stress-displacement curve in compression.
The primary intention of the authors was to develop a high-performance concrete

composition with local aggregates, with the exact sizes as delivered by the suppliers,
made with the most common commercially available cement, and with an increased
workability that would not require any vibration at casting. With these necessities in
mind, a small decrease in compression strength was tolerated which placed the
experimental concrete under the limit of 150 MPa in compression of ultra-high per-
formance concrete.
The flexural tensile strength was tested on 3 prisms 40 40 160 mm, with
vertical deformations. The results are presented in Fig. 5. The resulted flexural tensile
strength is fct,fl = 12.34 MPa with a standard deviation s = 1.11 MPa. The flexural
tensile strength of the experimental concrete corresponds to the characteristics of ultra-

high performance concrete.
Force - Displacement
Flexural tensile test
4.25
4
3.75
3.5
3.25
3
2.75
Force[MPa]
2.5
2.25 Prism1
2 Prism2
1.75
1.5 Prism3
1.25
1
0.75
0.5
0.25
0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 2.75 3 3.25
Displacement [mm]
Fig. 5. Force-displacement curve in flexural tensile test.
3.6 Half-Cell Potential Mapping

The tests were performed with a half-cell potential kit with a Copper/Saturated Copper
Sulphate Electrode which gives the following range of measurements: for potential
level between −350 and −500 mV there is a 95% chance of steel corrosively active,
−200 to −350 there is a 50% chance and less negative than −200, 5%.
Figure 6 shows the mapping procedure, on points distanced at 50 mm, measured
on the bottom surface of the elements, for each bar.
100
0
550
Bottom 10
50 Bar #1
100
Bottom 1 2 3 4 5 6 7 8 9 10
Bar #2
Fig. 6. Half-cell measurement method and apparatus.
For both rebars of element 1, with zinc anode, there is a chance of corrosion of 95%
along their length.
For both rebars of element 4, without zinc anode, there is a chance of corrosion of
50% along their length, except one side which exhibits less chances of corrosion: 5%.
The value of the measurements can be seen in Fig. 7.
Element 1 Element 4
Acc, with Zn Anode Acc, without Zn Anode
0 0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
-100 -50
-200 -100
Potential (mV)
Potential (mV)
-300 -150
-400 -200
-500 -250
-600 -300
-700 -350
Bar #2 Bar #1 Bar #2 Bar #1
Fig. 7. Half-cell measurements.
Although the potential measurements showed a very high chance of corrosion of

the embedded reinforcement, after extraction, the steel bars presented only minor spots
of corrosion products, specifically at the contact with the electric cable, as seen in
Fig. 8. Moreover, only an insignificant mass loss was registered of 0.2% on bar #2 of
the first element. The zinc anode, however, was approximately half-way destroyed.
Fig. 8. Visual aspects of the extracted reinforcement.
It is an obvious disassociation between the results on the half-cell potential map-

ping and the unaltered surface of the tensile reinforcement extracted from the elements.
It seems that the steel fibres became a generalized anode surrounding and communi-
cating with the traditional reinforcement, protecting in fact the bars. An exception was
made for the element containing zinc, which normally sacrifices itself before the cor-
rosion in the steel starts. This supposition is supported by other authors [7].
3.7 Flexural Tests

Figure 9 presents the failure mode of the elements exposed to accelerated corrosion 1
and 4 and the non-exposed, control elements, 13, 14 and 15. All of them presented a
middle main crack which led to the failure of the elements, as predicted in a 3 point
bending test. Along the main crack, multiple cracks have formed, which should be a
sign of ductility, specifically, we should notice increased vertical and transversal
deformations before failure. However, in a reinforced element of reduced size such as
the experimental elements of this study, all deformations were rather small. Further-
more, the main crack width could not be recorded in elements 13 and 14, because it
passed out of the limits of the transducer. Still, the characteristic crack width of 0.1 mm
was recorded and the results presented in Fig. 10. It can be seen that a 0.1 mm was
reached at a bending moment of (0.64 – 0.84) of the maximum bending moment, which
supports the assertion on ductility limitations of low scale elements. The respective
values show no significant quantitative differentiation between the exposed elements
and the control elements.
Fig. 9. Failure mode (from bottom up: left - 1; 4 exposed; right – 13; 14; 15 control).
The maximum experimental bending moment was calculated for each element
based on the maximum load recorded during tests and is presented in Fig. 11. All the
elements failed in the same range. The maximum bending moment of element 1, with a
zinc anode, exposed to accelerated corrosion was with 5.68% smaller than the greatest
value of 8.077 kNm experienced by the control element 13. Element 4 displayed a
2.97% decrease, element 14 a 0.62% decrease, whereas element 15 a 13.97% decrease
compared to element 13. Although this is proof that the corrosion process did not affect
the tested elements, more specimens are in order for viable conclusions.
Furthermore, the design resistant bending moment calculated with the concrete
experimental compression strength and the yielding strength of the reinforcement is
MRd = 3.66 kNm. This value represents 48.04% for element 1, 46.70% for element 4,
45.31% for element 13, 45.60% for element 14 and 52.67% for element 15, showing an
important strength reserve for elements made with high performance fibre reinforced
concrete.
Load ratio at wcrack = 0.1 mm

1.00 0.86
0.80 0.67 0.71
0.64
Mexp / Mmax
0.60
0.40
0.20
0.00
0.00
1 4 13 14 15
Element
Fig. 10. Mexp / Mmax at wcrack = 0.1 mm (1 – exposed, with zinc anode; 4 – exposed, without
zinc anode; 13; 14; 15 – control)
Mmax,exp
8.5 8.077 8.027
7.837
Mmax,exp (kNm)
8 7.618
7.5
6.949
7
6.5
6
1 4 13 14 15
Element
Fig. 11. Maximum bending moment (1 – exposed, with zinc anode; 4 – exposed, without zinc
anode; 13; 14; 15 – control)
4 Conclusions
This research evaluated the structural performance of high-performance fibre reinforced

concrete elements affected by corrosion and its subsequent deteriorations. The concrete
compression strength was fcm = 146.44 MPa and the flexural tensile strength was fct,
fl = 12.34 MPa. The concrete was developed with common commercially available
cement and local sand, but with increased workability in order not to use any means of
vibration. This led to a reduction in compressive strength which degraded the exper-
imental concrete from the ultra-high performance concrete series (>150 MPa).
The structural performance was assessed on pre-cracked elements exposed to
accelerated corrosion and tested to failure in comparison with control, non-exposed
elements. Here are the most important findings:
• Half-cell potential measurements showed more than 50% chances of corrosion, but
the extracted tensile reinforcement showed neither significant signs of corrosion
products, nor a significant mass loss. The steel fibres might have acted as anodes in
the surrounding concrete taking up all the corrosion deterioration from the tensile
bar. The exception might have been in the element containing a sacrificial zinc
piece, which was found half-way destroyed, thus being the element taking up the
deterioration.
• A crack width of 0.1 mm was reached at a value of (0.64–0.84) of the maximum
bending moment, which indicates the ductility limitations of low scale elements.
However, there were no significant quantitative differentiation between the exposed
elements and the control elements.
• The values of the maximum experimental bending moment of all the elements were
in the same range, with 0.62% to 13.97% less than the greatest value experienced by
control element 13.
• Furthermore, the design resistant bending moment was MRd = 3.66 kNm, which
represents approximately 50% of the maximum bending moment for all elements,
exposed and control.
Although these findings are proof that the corrosion process did not affect the tested
elements, more specimens are in order for viable conclusions and the research will be
continued in this direction.
Acknowledgements. This research was partially supported by the project 21 PFE in the frame
of the programme PDI-PFE-CDI 2018.
References
1. Dong, Y.: Performance assessment and design of ultra-high performance concrete (UHPC)
structures incorporating life-cycle cost and environmental impacts. Constr. Build. Mater.
167, 414–425 (2018)
2. Zhou, M., Lu, W., Song, J., Lee, G.C.: Application of Ultra-High Performance Concrete in
bridge engineering. Constr. Build. Mat. 186, 1256–1267 (2018)
3. Wang, D., et al.: A review on ultra-high performance concrete: part II. hydration,
microstructure and properties. Constr. Build. Mat. 96, 368–377 (2015)
4. Zhou, Z., Qiao, P.: Durability of ultra-high performance concrete in tension under cold
weather conditions. Cement and Concr. Composites 94, 94–106 (2018)
5. An, M., Wang, Y., Yu, Z.: Damage mechanisms of ultra-high-performance concrete under
freeze–thaw cycling in salt solution considering the effect of rehydration. Constr. Buil.
Mater. 198, 546–552 (2019)
6. Fan, L., Meng, W., Teng, L., Khayat, K.H.: Effects of lightweight sand and steel fiber
contents on the corrosion performance of steel rebar embedded in UHPC. Constr. Buil.
Mater. 238, 117709 (2020)
7. Berrocal, C.G., Lundgren, K., Löfgren, I.: Corrosion of steel bars embedded in fibre
reinforced concrete under chloride attack: state of the art. Cem. Concr. Res. 80, 69–85
(2016)
8. Hou, L., Liu, H., Xu, S., Zhuang, N., Chen, D.: Effect of corrosion on bond behaviors of
rebar embedded in ultra-high toughness cementitious composite. Constr. Build. Mater. 138,
141–150 (2017)
9. Hou, L., Guo, S., Zhou, B., Chen, D., Aslani, F.: Constr. Build. Mater. 196, 185–194 (2019)
10. Nguyen, W., Duncan, J.F., Devine, T.M., Ostertag, C.P.: Electrochemical polarization and
impedance of reinforced concrete and hybrid fiber-reinforced concrete under cracked matrix
conditions. Electrochim. Acta 271, 319e336 (2018)
11. Azad, A.K., Ahmad, S., Azher, S.A.: Residual strength of corrosion – damaged reinforced
concrete beams. ACI Mater. J. 104(1), 40–47 (2007). Title no. 104 – M05
12. Mokhtar, K.A., Loche, J.M., et al.: Report no. 2–2. Concrete in marine environment.
MEDACHS, Interreg IIIB Atlantic Space – Project no. 197 (2006)
Effect of Corroded Steel Fibers on Mechanical
Behavior of Steel Fiber Reinforced Concrete
Minoru Kunieda1(&), Masaki Tsutsui2, and Le V. Tri2

1
Department of Civil Engineering, Gifu University, Gifu, Japan
kunieda@gifu-u.ac.jp
2
Graduate School of Natural Science and Engineering,
Gifu University, Gifu, Japan
Abstract. This paper presents the exposure test results of Steel Fiber Rein-
forced Concrete (SFRC) beams having crack width of 0.2 mm for 3 months, 1
year and 2 years. By flexural loading test, initial stiffness of the exposed SFRC
beams was decreased with increasing of exposure period. And corrosion of steel
fibers themselves was observed at crack surface. The corrosion depth from the
specimen surface was about 15–35 mm. There was no significant effect of fiber
content on corrosion depth. Corrosion of the re-bar across the exposed crack was
observed, and it may affect the reduction of stiffness of the beams due to loss of
bond properties between re-bar and SFRC.
Keywords: SFRC Cracking Corrosion of fiber Durability
1 Introduction
Fiber Reinforced Concrete (FRC) is used to improve mechanical properties of concrete

structures such as shear capacity, ductility in addition to flexural capacity. In ordinary
design concept of reinforced concrete structures, cracking in service period is allowed
(e.g. allowable crack width less than 0.2 mm). Durability of FRC after cracking is not,
however, well known.
Various kinds of short fibers are proposed for FRC, steel fibers are familiar with
practical applications. When Steel Fiber Reinforced Concrete (SFRC) is used after
cracking, corrosion of steel fibers across a crack may be concerned [1], even if the crack
width is less than allowable crack width in design. J. L. Granju [2] reported the test
results on cracked SFRC samples with 0.5 mm crack mouth openings exposed to
marine-like environment for 1 year. The results confirm the small sensitivity of SFRC to
corrosion. The factors affecting the corrosion of the fibers and the reasons for the increase
in flexural strength after corrosion are discussed. The researches on stainless steel and/or
other fibers having higher resistance against corrosion have been conducted [3].
This paper presents the results of exposure test of cracked SFRC beams with re-bar,
and flexural failure behavior of SFRC beams with corroded steel fibers was investigated.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_41
446 M. Kunieda et al.
2 Outline of Experiments
2.1 Specimen
Mix proportions of used SFRC are shown in Table 1. Water to cement ratio was 50%,
and high early strength Portland cement was used. Volume fractions of fibers were 0.5,
1.0 and 1.5. The length and diameter of used steel fibers, which had hoock in both ends,
were 30 mm and 0.62 mm, respectively. Compressive strength was measured by using
cylindrical specimens having the size of d100 x h200 mm (JIS A 1108 Method of test
for compressive strength of concrete), and flexural strength was measured by using
rectangular specimens having the size of h100 x w100 x l400 mm (JIS A 1106 Method
of test for flexural strength of concrete).
Compressive strength and flexural strength at the age of 28 days are summarized in
Table 2. The compressive strength of the specimen in Vf = 0.5% series was slightly
lower than those of other cases. Figure 1 shows the tension softening curves of each
series. The curves were obtained from the flexural tests was converted to tension
softening curves by means of modified J integral method [4]. The fiber bridging stress
was observed corresponding to fiber content. The difference of fiber bridging stress in
Vf = 1.0% and Vf = 1.5% series was not, however, significant.
As shown in Table 3, 4 specimens were prepared for each series having different
fiber content (only Vf = 1.5% series was 3 specimens). Two of them were tested after 3
months after the initial loading. One specimen was tested after 1 year, and rest one
specimen was also tested after 2 years.
Table 1. Mixproportions of SFRC

Vf (%) W/C (%) Unit content (kg/m3)
Water Cement Sand Gravel Fiber
0.5 50 168 335 851 913 39
1.0 50 168 335 851 898 79
1.5 50 168 335 851 888 118
Table 2. Compressive strength of SFRC (age: 28 days)

Vf (%) Compressive strength (N/mm2) Flexural strength (N/mm2)
0.5 28.5 4.8
1.0 37.8 6.2
1.5 34.1 6.6
2.2 Initial Loading Test

At the age of 28 days, four-point bending tests were carried out to induce initial cracks
in the SFRC beam. Figure 2 shows the image of loading test. Size of specimen was
Effect of Corroded Steel Fibers on Mechanical Behavior of SFRC 447
h200 x w100 x l1600 mm, and effective depth was 170 mm. Two deformed rebar
having a diameter of 13 mm (SD345, fsy = 395 N/mm2) was used. Loading span was
1400 mm, load and displacement at loading points were measured.
2.3 Treatment of Occurred Cracks

Target crack (major crack) in the specimen after initial loading tests was selected for
the exposure test. Figure 3 shows schematic image of the specimen with initial cracks,
and photo of the treated specimen. Epoxy resin was painted to the specimen surface
except for 50 mm around the target crack. The maximum crack width (residual crack
width after unloading) at tension side was 0.2 mm approximately.
5 5
4 4
Tensile stress (N/mm2
Tensile stress (N/mm2
3 3
2 2
1 1
0 1 2 3 4 5 0 1 2 3 4 5
(a) Vf =0.5% (b) Vf =1.0%
5
2
4
Tensile stress (N/mm
1
0 1 2 3 4 5
Crack width (mm)
(c) Vf=1.5%
Fig. 1. Tension softening curves

Table 3. Number of the specimens in this test program

Initial loading 3 months 1 year 2 years
Vf = 0.5% 4 2 1 1
Vf = 1.0% 4 2 1 1
Vf = 1.5% 3 2 1 0
Fig. 2. Test setup
2.4 Exposure Test and Re-Loading

The specimens with cracks were exposed at outside for 3 months, 1 year and 2 years,
and loading tests were conducted. The exposure test was conducted in the campus of
Gifu University (started from Nov. 2015). Note that, the specimens were affected by the
climate condition includes rain and temperature. The specimen was arranged so that the
side of the specimen is on the upper side. The NaCl solution (3.0%) was sprayed to the
specimen with interval of once a week during 3 months.
Epoxy coating
Fig. 3. Schematic image of exposed specimen (cracks were coated except for a target crack)
70 70
60 60
50 50
Load (kN)
Load (kN)
40 3 months (No.1) 40 3 months (No.1)

30 3 months (No.2) 30 3 months (No.2)
20 1 year 20 1 year
10 2 years 10 2 years
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Disp. (mm) Disp. (mm)
(a) Vf=0.5% (b) Vf=1.0%
70
60
50
Load (kN)
40
30 3 months (No.1)
20 3 months (No.2)
10 1 year
0
0 2 4 6 8 10
Disp. (mm)
(c) Vf=1.5%
Fig. 4. Load-displacement curves
3 Experimental Results
3.1 Load-Displacement Relations

Figure 4 shows the load-displacement curves obtained from the re-loading. Note that
the load displacement curves were displayed with the origin set to zero. The decrease of
initial stiffness was observed at the age of 1 year and 2 years, comparing to that of 3
months. In addition, no significant difference was observed in initial stiffness between 1
year and 2 years. In the re-loading test, the target crack, which was selected as an
exposed crack, was re-opened in all series.
3.2 Corrosion of Steel Fibers Across a Crack

Figure 5 shows exposed crack surface of FRC having fiber content of 1.0% at the age
of 3 months. In addition, exposed crack surfaces at the age of 1 year and 2 years are
also shown in Figs. 6, 7, 8 and 9.
As shown in Fig. 5, there was no corrosion of steel fibers in the crack surface at the
age of 3 months. In the case of exposure for 1 and 2 years, corrosion of steel fibers near
specimen surface (depth about 15–35 mm) was observed. Note that the depth was
measured by ruler directly. There was no significant effect of fiber content on the
observed depth of corrosion. The corrosion of fibers was not, however, so severe as to
cause the steel fibers to lose its cross section. It seems that this corrosion affects the
decrease of initial stiffness in load-displacement curves.
Fig. 5. Fibers at crack surface (3 months, Vf = 1.0%)
Fig. 6. Observed corrosion of fibers at crack surface (1 year, Vf = 0.5%)
Fig. 7. Observed corrosion of fibers at crack surface (2 years, Vf = 0.5%)

Fig. 8. Observed corrosion of fibers at crack surface (1 year, Vf = 1.0%)
Fig. 9. Observed corrosion of fibers at crack surface (2 years, Vf = 1.0%)
Figure 10 shows the photo of the re-bar across a crack. Corrosion on the surface of
re-bar was slightly observed. Although loss of cross-sectional area of re-bar was not
observed, and the corrosion may affect the decrease of initial stiffness in load-
displacement curves too. Further researches on the contribution of re-bar corrosion to
the decrease of the stiffness in load-displacement curves are needed.
Fig. 10. Observed corrosion of re-bar (2 years, Vf = 0.5%)

4 Conclusions
Experimental study of corroded Steel Fiber Reinforced Concrete (SFRC) beams sub-
jected to flexural was carried out, and following results were obtained.
• The exposure tests for SFRC beams having crack width of 0.2 mm was carried out
for 3 months, 1 year and 2 years. By flexural loading test, initial stiffness of the
exposed SFRC was decreased with increasing of exposure period.
• Corrosion of steel fibers themselves was observed at crack surface. The corrosion
depth from the surface was about 15–35 mm. There was no significant effect of fiber
content on corrosion depth.
• Corrosion of the re-bar across the exposed crack was observed. Although loss of
cross-sectional area of re-bar was not observed and it may affect the reduction of
stiffness of the beams due to loss of bond properties between re-bar and SFRC.
References
1. Kosa, K., Naaman, A.E.: Corrosion of steel fiber reinforced concrete. Mater. J. 87(1), 27–37
(1990)
2. Granju, J.L., Balouch, S.U.: Corrosion of steel fibre reinforced concrete from the cracks. Cem.
Concr. Res. 35(3), 572–577 (2005)
3. Fu, X., Chung, D.D.L.: Bond strength and contact electrical resistivity between cement and
stainless steel fiber: their correlation and dependence on fiber surface treatment and curing
age. Mater. J. 94(3), 203–208 (1997)
4. Uchida, Y., Rokugo, K., Koyanagi, W.: Determination of tension softening diagrams of
concrete by means of bending tests. J. JSCE 426, 203–212 (1991). (in Japanese)
Self-healing of Fibre Reinforced Concrete
Containing an Expansive Agent in Different
Exposure Conditions
K.-S. Lauch, C. Desmettre, and J.-P. Charron(&)
Department of Civil, Geological and Mining Engineering,

Polytechnique Montreal, Montreal, Canada
jean-philippe.charron@polymtl.ca
Abstract. While most studies about self-healing concrete investigated several

healing conditions such as water immersion, humid chamber and wet/dry cycles,
very few assessed the self-healing capacity of concrete in realistic outdoor
condition. Furthermore, self-healing capacity is usually determined with a single
testing procedure (mechanical or permeability measurements), sometimes with
visual observations. Hence, this project aimed to evaluate, through mechanical
and water permeability tests, as well as optical and microscopic observations on
the same specimens, the self-healing capacity of high performances fibre rein-
forced concretes (HPFRC) containing different admixtures. This paper focuses
on the water permeability evolution of two HPFRC (water to binder ratio of 0.43
and 0.75%-vol of steel macrofibres), one control and one containing calcium
sulfoaluminate-based expansive agent (CSA), under different laboratory and
outdoor conditions. Prisms were pre-cracked at 28 days by means of a 3-point
bending test and then exposed to laboratory conditions (air, water immersion
and wet/dry cycles) for 3 months and outside for one year. The results showed a
better self-healing performance of the CSA HPFRC in wet/dry cycles compared
to water condition. For the control HPFRC, the opposite is observed. Self-
healing of outdoor prisms is slower than in laboratory (6 months outside cor-
responds to around 2 weeks in water immersion for CSA). Self-healing per-
formance of the CSA HPFRC is better than the control mix in outdoor condition.
Keywords: Self-healing concrete Expansive agent Water permeability

Outdoor condition
1 Introduction
Most reinforced concrete structures prematurely reach the end of their service life
because of gradual degradations caused by the environment. Twenty-five percent of
Quebec Ministry of Transportation (QMT) bridges suffer from various disorders such
as rebar corrosion and concrete cracking and require major repairs [1], bridge upgrade
being of the order of $2634 M/year [2]. Concrete cracks, which are inevitable,
accelerate the penetration of aggressive agents such as CO2 or chlorides and hence
reduce the durability of the structures. As the durability and safety of structures are
rising concerns, concretes with self-healing capabilities are a promising solution in line
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_42
454 K.-S. Lauch et al.
with sustainable development. Such construction materials could not only increase the
service life of bridges, but also cut down the huge costs of maintenance and repair, and
reduce the traffic and disturbances around construction sites for the road users.
Cementitious materials have a natural self-healing capacity, called autogenous
healing. Hydration of anhydrous cement and carbonation are the main mechanisms
involved in this autogenous capacity, being efficient to heal very small cracks (10 to
100 µm) [3]. In practice, cracks are often wider. The cracks width limits stand at
0.3 mm in the Canadian standard CSA A23.3 for reinforced concrete in outdoor
environment and at 0.25 mm in the Canadian standard CSA S6 for highway bridges.
To improve and better control this self-healing capacity, different techniques have been
developed, such as the use of bacteria, encapsulated chemicals or mineral admixtures
[4]. The latter has the advantages of being available products, less complex, more
economical and having a long shelf life. Among mineral admixtures, crystalline
admixtures [5] and calcium sulfoaluminate (CSA)-based expansive agents [6] or a
combination of these two [7] have been investigated for their self-healing potential.
Sisomphon et al. [8] found that the addition of CSA allow crack closing, even for crack
widths of 0.25–0.40 mm.
As numerous testing methods are used to evaluate the self-healing capacity, it is
difficult yet to establish the optimal measurement techniques. Some studies consider
optical measurements [7], while others run water permeability [9], bending tests, or
non-destructive tests [5]. In most cases, each test is carried out on different specimens.
Besides, self-healing depends on many parameters such as concrete composition, age
of pre-cracking, crack width and the exposure condition. While most studied only the
water immersion as a healing condition [10], Sisomphon et al. found that wet/dry
cycles (12 h dry and 12 h wet) was the best condition for a mortar mix with CSA [7].
Furthermore, as pointed out by Li and Herbert [4], there is a lack in research of self-
healing under realistic outdoor conditions. While most studies showed promising
results in controlled laboratory conditions, very few investigated the self-healing
capacity under long-term outdoor conditions involving a wide and random range of
temperatures and precipitations. Self-healing can occur in real condition and be efficient
under multiple damage events, as demonstrated by Herbert and Li [11], Sherir et al.
[12] and Cuenca et al. [13].
Although the number of papers about self-healing concrete is increasing expo-
nentially, the research area is still relatively recent and not mature yet. This project has
been launched to respond to some issues mentioned. The project goal is to evaluate the
self-healing capacity of high performance fibre reinforced concretes (HPFRC) con-
taining different admixtures (control, CSA, crystalline admixture, superabsorbent
polymer), subjected to different conditions (laboratory and long-term outdoor expo-
sure). The presence of fibres limits the crack widths, while the admixture promotes the
healing of the cracks. The self-healing performance is assessed by mechanical tests,
water permeability measurements, optical observations and microscopic analyses on
the same specimens. This paper focuses only on the self-healing capacity of control and
CSA HPFRC, exposed to different laboratory conditions (air, water immersion, wet/dry
cycles) and outdoor exposure via water permeability tests.
Self-healing of Fibre Reinforced Concrete 455
2.1 Materials
A high performance fibre reinforced concrete (HPFRC) with a water to binder ratio of
0.43 and containing a blended General Use Portland cement with 8% by mass of silica
fume (type GUb-SF) was used. 0.75% volume of hooked-end steel macrofibres (lf =
35 mm and /f = 0.55 mm) was included to limit crack width of testing specimens.
Two HPFRC mixes, one control without any admixture and one with an expansive
agent (Denka Power CSA) at a dosage of 3.33% by mass of cement as recommended
by the manufacturer, were produced. The compositions of the control and the CSA
mixtures are summarized in Table 1.
A first batch was made to produce specimens intended for the long-term outdoor
exposure. For each mix, 12 cylindrical specimens (/100 mm 200 mm) for com-
pressive strength characterization at different ages, 4 prisms (75 mm 125 mm
450 mm, Fig. 1) to be pre-cracked under flexure before healing and 2 reference
uncracked prisms, were cast. Then a second batch was made for the specimens sub-
jected to laboratory exposure conditions. 12 cylinders and 14 prisms were produced per
mix.
The prisms included 2 uncracked prisms per exposure condition. The specimens
were demoulded at 24 h and then immersed in lime-saturated water for 28 days of
curing. The compressive strength (fc) and the Young’s modulus (Ec) were determined
at 28 days, 6 months and 12 months for outdoor specimens and at 28 days and 3
months for indoor specimens, in accordance with ASTM C39 and ASTM C469
respectively.
Table 1. Compositions of the HPFRC mixtures

Material Control CSA
Cement (kg/m3) 550 550
CSA (kg/m3) – 18.3
Water (kg/m3) 237 237
Superplasticizer (l/m3) 10 10
Viscosity agent (l/m3) 0.7 0.85
Sand (kg/m3) 779 771
Coarse aggregates (kg/m3) 631 624
Steel fibre (kg/m3) 58.5 58.5
2.2 Methodology to Assess Self-healing

The prisms (Fig. 1) were first notched at mid-span (20 mm deep) and pre-cracked at
the age of 28 days by means of a 3-point bending test according to EN 14651. The
specimens were loaded up to a crack mouth opening displacement (CMOD) of 0.9 mm
measured by an extensometer. After unloading, the residual CMOD was around

0.5 mm at the bottom of the notch, which corresponded to crack widths of around 0.1–
0.3 mm at the notch root (beginning of the actual crack), as measured by a digital
microscope. These values fall into the Canadian standards crack width limits for
exposed structures.
After pre-cracking, the initial water permeability of prisms was measured. Because
of the inherent variability of HPFRC properties, the characteristics of the flexural
cracks varied widely in width, length, and tortuosity, which led to a large scatter of
initial water permeability coefficients (Kwi). Therefore, prisms showing a fair dispersion
of initial water permeability (i.e. prisms with high and low initial permeability) were
attributed to each exposure conditions to compare objectively the impact of the
exposure on the healing process.
Prisms intended for outdoor conditions were kept in laboratory air condition
another 28 days after pre-cracking to allow a partial drying before they were placed
outside (winter period at that time). They were kept for one year in Montreal climate.
Prisms intended for laboratory conditions were submitted 3 months either to water
immersion, wet/dry cycles (3.5 days wet and 3.5 days dry) and air condition (only for
the control mix). The air and drying conditions took place in a climatic chamber at
21 ± 3 °C and 45 ± 10% relative humidity. Water permeability Kw was then mea-
sured after 3, 6, 9 and 12 months for outdoor prisms and at 11 days, 25 days, 2 months
and 3 months for indoor prisms. Before each water permeability measure, pictures of
the crack were taken with a digital microscope at 4 equidistant points along the crack,
on both prism faces (face 1 “front ! inlet” and face 2 “back ! outlet”, Fig. 1). These
visual observations were only used to assess the self-healing qualitatively.
Once the healing period terminated (3 months for indoor specimens and 1 year for
outdoor ones), prisms were reloaded until failure with the same 3-point bending test as
in the pre-cracking phase to assess the mechanical recovery provided by the healing
products. The two uncracked prisms kept in the same conditions as the others were
consecutively pre-cracked and loaded to failure as references. Following the reloading,
some samples were sawn at healed cracks for scanning electron microscopy (SEM) and
energy dispersive X-ray (EDX) analyses to identify the healing products. This paper
Face 1, inlet Face 2, outlet

Aluminium
box Aluminium
box
Notch filled
Flexural crack
with elastomer
Fig. 1. 3D diagram of the notched and cracked concrete specimen

focuses only on the water permeability measured on the control and CSA concrete
mixtures, subjected to the laboratory and outdoor conditions.
2.3 Water Permeability Device

A new water permeability set-up was developed to measure the permeability through
the pre-cracked prisms. This new set-up was inspired by a testing device previously
developed at Polytechnique Montreal [14]. A significant benefit from this new water
permeability set up is that it allows to measure transport properties on a prism than can
then be reloaded, enabling to investigate correlations between a self-healing state and
mechanical recovery.
The set-up consists of two water-filled aluminium boxes clamped on the two opposite
cracked lateral faces of the prism (Fig. 2). The boxes are linked to inlet and outlet tanks.
Water in the inlet tank is put under pressure, while the outlet tanks remains at the
atmospheric pressure to subject the specimen to a pressure gradient of 30 kPa (corre-
sponding to a 3-m water depth). This pressure gradient initiates a water flow through the
cracked specimen from the inlet to the outlet tank. The incoming and outgoing water flows
through the prism are recorded. The permeability coefficient Kw (m/s) is calculated after
5 min, insuring that the inlet and outlet flows are equal, using Darcy’s law (Eq. (1)), with
Q (m3/s) the water flow, L (m) the flow path length, A (m2) the specimen cross-section and
Dh (m) the drop in the hydraulic head across the specimen.
QL
Kw ¼ ð1Þ
ADh
Pressure regulator Patm

Water flow
Pressure
transmitter
Clamping
system
Notched and
cracked concrete
Aluminium specimen
boxes
Notch filled
with elastomer
Inlet tank Outlet tank
Fig. 2. New water permeability device

3 Results
3.1 Characterization Tests
All concrete batches (for outdoor and indoor exposures) and mixes (control and CSA)
reached similar slump flows of 550 ± 40 mm and Young’s modulus of 33
000 ± 1000 MPa. The evolution of the compressive strength fc (average and standard
deviation) is shown in Fig. 3. At a same time, fc was quite similar for the control and
CSA mixes. Furthermore, there were no significant differences regardless of the
exposure condition. Thus, the fresh state properties and compressive strength of the
HPFRC mixtures only had a limited impact on the self-healing capacity measured.
90
Control CSA 82.0
74.50
80 72.8 71.1 73.1
Compressive strength (Mpa)
65.2 66.1 69.5 66.9

70 63.7
57.3
60
50
40
30
20
10
0
28 days (indoor) 119 days 119 days 119 days 28 days 247 days
(water) (wet/dry) (air) (outdoor) (outdoor)
Fig. 3. Results of compressive strength at various times.
3.2 Water Permeability Measurements

3.2.1 Prisms in Controlled Laboratory Conditions
The average results and standard deviations of water permeability coefficients Kw
measured at different exposure times are shown in Fig. 4 and Fig. 5 for the control and
CSA HPFRC respectively. These values came from 4 to 5 specimens for the water and
wet/dry exposures and from 2 specimens for the air condition. In these figures, the
values obtained for the water and wet/dry conditions have been slightly shifted by 0.5
and 1 day respectively for clearer display purpose. The healing ratios (HR), calculated
for each specimen by Eq. (2) with Kwt the permeability coefficient at a time t and Kwi
the initial permeability coefficient, are averaged and illustrated in Fig. 6.

Kwt
HR ¼ 1 :100 ð2Þ
Kwi
8.E-06
Control water
7.E-06
Permeability coefficient Kw (m/s) Control wet/dry
6.E-06 Control air
5.E-06
4.E-06
3.E-06
2.E-06
1.E-06
0.E+00
0 10 20 30 40 50 60 70 80 90
Exposure time (days)
Fig. 4. Evolution of permeability coefficient for the control HPFRC.
8.E-06
CSA water
7.E-06
Permeability coefficient Kw (m/s)
CSA wet/dry
6.E-06
5.E-06
4.E-06
3.E-06
2.E-06
1.E-06
0.E+00
0 10 20 30 40 50 60 70 80 90
Exposure time (days)
Fig. 5. Evolution of permeability coefficient for the CSA HPFRC.
The control specimens presented average initial permeabilities of 2 10−6 to

3 10−6 m/s. At 11 days, Kw decreased rapidly to 1.5 10−7, 3.6 10−7 and
1.8 10−6 m/s for water, wet/dry cycles and air exposures respectively (Fig. 4), which
represents average self-healing ratios (HR) of 95%, 87% and 12% (Fig. 6). Then the
decreasing rate slowed down and Kw stabilized after 2 months to reach average values
of 5.3 10−8, 1.1 10−7 and 1.2 10−6 m/s after 3 months in water, wet/dry cycles
and air respectively, which represents HR of 98%, 95% and 46%. Self-healing of the
control HPFRC is greater in water than in wet/dry condition and, as expected, lower in
air exposure.
For CSA HPFRC, the self-healing kinetics was similar (fast self-healing rate during
the first 11 days and then the rate slowed down). Their average initial permeabilities
was around 2.8 10−6 m/s. At 11 days, Kw decreased to 1 10−6 and 1.6 10−7 m/s
100%
11 days
90%
25 days
80% 2 months
3 months
70%
60%
Healing ratio
50%
40%
30%
20%
10%
0%
Control water CSA water Control wet/dry CSA wet/dry Control air
Fig. 6. Evolution of healing ratios for control and CSA HPFRC for different exposure
conditions.
for the water and wet/dry cycles exposures respectively (Fig. 5), which represents
average HR of 62% and 87% (Fig. 6). Then Kw dropped to 5.4 10−7 and 9.1 10−9
m/s respectively for the water and wet/dry cycles exposures, which represents HR of
86% and 99%. Unlike the control HPFRC, self-healing of the CSA HPFRC is superior
in wet/dry than in water condition.
The standard deviations for the initial permeability measured from one specimen to
another for the control and CSA mixes are addressed in the Discussion section below.
Figure 6 also compares the average healing ratios and standard deviations of the
HPFRC mixes when submitted to a same exposure condition. When kept in water, the
control mix showed higher HR at all times than the CSA mix. After 3 months, the
control HPFRC had an average HR of 98% against 86% for the CSA HPFRC. In
wet/dry condition, both mixes showed an average HR of 87% at 11 days, then the CSA
mix healed better than the control mix (HR of 99% against 95% after 3 months). In air
condition, the HR of the control specimens were much lower than the other conditions,
but they increased with time (HR from 12% at 11 days to 46% after 3 months).
3.2.2 Prisms in Realistic Outdoor Conditions

Figure 7 shows the average permeability coefficients and their standard deviation at
different times (0, 3, 6 and 9 months) of the prisms exposed to outdoor conditions. The
results of the prisms kept in indoor laboratory conditions (water immersion), described
in the previous section, are also shown in Fig. 7 for comparison purpose. A different
time scale was used to compare the self-healing kinetics between the indoor and
outdoor prisms, in months for outdoor prisms and in days for indoor prisms. As for the
results of the indoor prisms, the Kw values of the control and CSA outdoor prisms have
been slightly shifted in the x-axis for clearer display purpose.
It should be noted that the initial water permeability coefficients Kwi of the outdoor
specimens varied from the indoor specimens. The difference in Kwi for the control and
CSA outdoor concretes was also large (Kwi for the control mix was almost 4 times
Time exposure for outdoor specimens (months)

0 2 4 6 8 10 12
7.E-06
Control water
6.E-06 CSA water
Permeability coefficient Kw (m/s)
Control outdoor
5.E-06
CSA outdoor
4.E-06
3.E-06
2.E-06
1.E-06
0.E+00
0 10 20 30 40 50 60 70 80 90
Time exposure for indoor specimens (days)
Fig. 7. Evolution of permeability coefficient for prisms exposed to outdoor and indoor
conditions.
higher than the CSA mix). This is because a second batch of the control and CSA
concretes had to be prepared for indoor specimens and some discrepancies were
obtained in the permeabilities between the two batches.
The average water permeability coefficients of the 4 prisms exposed to outdoor
conditions decreased from initial values of 5.2 10−6 and 1.3 10−6 m/s to
2.2 10−6 and 4.1 10−7 m/s at 9 months for the control and CSA HPFRC
respectively. These correspond to average healing ratios of 59% (control) and 81%
(CSA) after 9 months.
As observed for the indoor specimens, the self-healing kinetics (decrease of Kw) of
the outdoor specimens was faster at the beginning and then slowed down with time
until stabilization.
Table 2 shows the evolution of healing ratios for indoor and outdoor prisms. The
self-healing of the outdoor control prisms after 3 months was equal to 46%. The same
HR was obtained in less than 11 days for the indoor specimens exposed to water or
wet/dry cycles and after 3 months for the ones subjected to air condition (with a relative
humidity of 45%). The self-healing of the outdoor CSA specimens after 6 months was
66%, which is close to the HR of 62% obtained after 11 days of water immersion. After
9 months, HR for the CSA outdoor prisms reached 81%, which is close to the HR of
85% obtained after 25 days of water immersion or to the HR of 87% obtained after 11
days of wet/dry cycles. In conclusion, the self-healing obtained in 2 to 4 weeks in
laboratory conditions (water or wet/dry cycles) was accelerated compared to the
effective self-healing process in outdoor prisms for which 6 to 9 months were necessary
to reach the same self-healing level.
Table 2. Evolution of healing ratios for prisms exposed to outdoor and indoor conditions.
Indoor 11 days 25 days 2 months 3 months
Control water 95% 98% 98% 98%
Control wet/dry 87% 85% 93% 95%
Control air 12% 10% 35% 46%
CSA water 62% 85% 86% 86%
CSA wet/dry 87% 98% 99% 99
Outdoor 3 months 6 months 9 months 12 months
Control outdoor 46% 50% 59%
CSA outdoor 51% 66% 81%
4 Discussion
As mentioned in the methodology section, initial permeability coefficient (Kwi) of

prisms varied due to the inherent variability of the crack pattern (width, length, tor-
tuosity) within HPFRC specimens. The impact of the initial permeability coefficient
was reduced as much as possible in the research program by distributing fairly prisms
with high and low permeability to each exposure conditions, thus the average Kwi of
prisms for different exposure condition was similar. Then, the healing ratio was
evaluated on each specimen and averaged for prisms of the same condition. This
methodology should limit the impact of the Kwi variability of prims in the research
program and does not modify the general trends observed from one exposure condition
to another.
Whether for the indoor or outdoor prisms, the self-healing kinetics is globally
similar with a first phase of fast decrease of the water permeability (during around 11
days for the indoor and around 3 months for the outdoor specimens) followed by a
second phase where permeability slowed down and stabilized over time. For example,
in the first 11 days of exposure of prisms in laboratory conditions, water permeability
coefficients (Kw) dropped by 95% and 87% respectively for water and wet/dry con-
ditions for the control HPFRC mix. Afterwards and until 3 months of exposure, per-
meability decreased by 23% (water) and 43% (wet/dry). Such self-healing kinetics was
already reported in literature [15, 16] and is explained by the calcium carbonate pre-
cipitation mechanism. First, precipitation is rapid as calcium ions Ca++ are readily
available at the concrete surface. Then it is controlled by the diffusion of the calcium
ions through the concrete matrix, which is a slower process.
Another trend observed is that the variability of the initial permeability coefficients
(Kwi) of prisms (due to the variability of the cracks pattern) decreased sharply with the
evolution of the healing process, as shown with the errors bars in Fig. 4 and Fig. 5.
This trend is logical as all the specimens converged over time to the same completely
healed state. This can also be seen in Fig. 6 where the magnitude of error bars of the
healing ratios decrease slightly with time, except for the control specimens in air
condition. The control specimens immersed in water presented low variability of the
healing ratios, in contrast to the large variability of the specimens in wet/dry and air
conditions, which can be related to a more efficient healing in water. For the CSA
prisms, the opposite trend is observed as self-healing proceeds: high magnitude of error
bars is noted in water immersion and low magnitude in wet/dry condition, which can be
related to a more efficient healing in wet/dry condition. These trends are coherent with
the decreasing variability of permeability observed for the outdoor CSA prisms but not
for the control prisms (Fig. 7). This high variability for the outdoor control prisms may
also be explained by the still high permeability of these specimens after 9 months.
According to these previous observations, the control HPFRC prisms performed
better than the CSA HPFRC prisms in water condition, while it is the opposite when
exposed to wet/dry cycles. However, Sisomphon et al. [8] found a greater permeability
decrease of mortars specimens immersed in water made with CSA in comparison with
a control mortar (water to cement ratio of 0.25, pre-cracked at 28 days). When different
exposure conditions were tested, they found however that wet/dry cycles was the best
exposure for mechanical regain, for CSA and for control specimens [7]. The better
performance of the control HPFRC in water in this paper, in comparison to the study of
Sisomphon et al., could be explained by the silica fume (SF) content of HPFRC. It has
been shown that a mortar with SF showed outstanding self-crack closing ability [10].
Another reason would be the high alkalinity of the water. As lime is added in con-
tainers used for water and wet/dry conditions in laboratory, the water has a high pH
fostering healing [6]. The results of the SEM and EDS analyses planned in the project
will provide a better understanding of the different self-healing behaviours of the
control and CSA HPFRC mixes subjected to different exposure conditions.
Concerning the outdoor prisms, this paper showed a high healing ratio (81%) for
the CSA HPFRC mix at 9 months compared to the control HPFRC mix (59%). This
would suggest that the CSA specimens show a better self-healing capacity in real
environment than the control ones. This observation is in line with the results found
earlier for indoor prisms as CSA HPFRC showed the best performance in wet/dry
cycles, which better replicate real climate exposure. Herbert and Li [11] and Sherir and
al. [12] also confirmed with Engineered Cementitious Composites (ECC) the self-
healing capability of specimens in realistic environment. They assessed the mechanical
recovery and Resonant Frequency, thus no correlation can be established with per-
meability measurements shown in this paper. Cuenca et al. [13] investigated the self-
healing capacity of FRC with crystalline admixture (CA). They calculated a sealing
index (SI) based on image analysis of crack surfaces and found low SI for open air
exposure for both CA and control specimens (SI around 5 to 30% at 6 months). Sealing
of crack widths at specimen’s surfaces may not provide the same trend for self-healing
as permeability measurement represents the healing of the interior of the crack. Thus
results comparison with this study is not obvious.
5 Conclusions
The self-healing capacities of two high performance fibre reinforced concretes

(HPFRC), one control and one containing a calcium sulfoaluminate-based (CSA) ex-
pansive agent, subjected to different indoor laboratory conditions (ambient air, water
immersion and wet/dry cycles), and to a long-term outdoor condition (Montreal climate
for 1 year) have been investigated and compared. Self-healing was assessed by means
of water permeability measurements. Based on the results of this experimental pro-
gram, the following conclusions can be drawn:
• Self-healing capacity depends greatly on the mixture composition and the exposure
condition. While control HPFRC healed better in water immersion, CSA HPFRC
achieved greater healing in wet/dry cycles.
• The high healing ratios of the CSA HPFRC were also observed in realistic envi-
ronment. Thus, HPFRC with a CSA-based expansive agent seems to have a strong
self-healing capacity in outdoor conditions.
• The self-healing kinetics was slower for the outdoor exposure condition than for the
water and wet/dry laboratory exposures. The self-healing state of the CSA HPFRC
exposed 9 months in outside conditions corresponds approximatively to 25 days in
water or less than 11 days in wet/dry condition.
This paper only focused on the water permeability measurements carried out on
prisms. At the end of their healing conditions, all specimens will be reloaded to assess
their mechanical recovery and healing products will be identified with SEM/EDX
analysis. These results will be presented in future papers.
Acknowledgements. The authors would like to thank the Québec Research Fund on Nature and
Technologies (FRQNT) for the financial support and Denka for providing the admixture.
References
1. Quebec Ministry of Transport: General condition index. In: QMT statistics (2019)
2. Quebec Ministry of Transport: Quebec infrastructure plan PQI 2011–2016 (2013)
3. De Belie, N., et al.: A review of self-healing concrete for damage management of structures.
Adv. Mater. Interfaces, 28 (2018)
4. Li, V., Herbert, E.: Robust self-healing concrete for sustainable infrastructure. J. Adv. Concr.
Technol. 10, 207–218 (2012)
5. Ferrara, L., Krelani, V., Carsana, M.: A “fracture testing” based approach to assess crack
healing of concrete with and without crystalline admixtures. Constr. Build. Mater. 68, 535–
551 (2014)
6. Jiang, Z., Li, W., Yuan, Z.: Influence of mineral additives and environmental conditions on
the self-healing capabilities of cementitious materials. Cem. Concr. Res. 57, 116–127 (2015)
7. Sisomphon, K., Copuroglu, O., Koenders, E.A.B.: Effect of exposure conditions on self
healing behavior of strain hardening cementitious composites incorporating various
cementitious materials. Constr. Build. Mater. 42, 217–224 (2013)
8. Sisomphon, K., Copuroglu, O., Koenders, E.A.B.: Self-healing of surface cracks in mortars
with expansive additive and crystalline additive. Cem. Concr. Compos. 34, 566–574 (2012)
9. Roig-Flores, M., Moscato, S., Serna, P., Ferrara, L.: Self-healing capability of concrete with
crystalline admixtures in different environments. Constr. Build. Mater. 86, 1–11 (2015)
10. Jaroenratanapirom, D., Sahamitmongkol, R.: Self-crack closing ability of mortar with
different additives. J. Met. Mater. Miner. 21(1), 9–17 (2011)
11. Herbert, E., Li, V.: Self-healing of microcracks in Engineered Cementitious Composites
(ECC) under a natural environment. Materials 6, 2831–2845 (2013)
12. Sherir, M.A.A., Hossain, K.M.A., Lachemi, M.: Development and recovery of mechanical
properties of self-healing cementitious composites with MgO expansive agent. Constr.
Build. Mater. 148, 789–810 (2017)
13. Cuenca, E.T., Tejedor, A.; Ferrara, L.: A methodology to assess crack-sealing effectiveness
of crystalline admixtures under repeated cracking-healing cycles. Constr. Build. Mater. 179,
619–632 (2018)
14. Desmettre, C., Charron, J.-P.: Novel water permeability device for reinforced concrete under
load. Mater. Struct. 44, 1713–1723 (2011)
15. Escoffres, P., Desmettre, C., Charron, J.-P.: Effect of a crystalline admixture on the self-
healing capability of high-performance fiber reinforced concretes in service conditions.
16. Homma, D., Mihashi, H., Nishiwaki, T.: Self-healing capability of fibre reinforced
cementitious composites. J. Adv. Concr. Technol. 7(2), 217–228 (2009)
Characterisation of Strain-Hardening
Cementitious Composite (SHCC) Under Cyclic
Loading Conditions for Self-healing
Applications
Zixuan Tang(&), Chrysoula Litina, and Abir Al-Tabbaa
Department of Engineering, University of Cambridge, Cambridge, UK

zt235@cam.ac.uk
Abstract. The ground shaking in an earthquake often imposes cyclic loadings

on infrastructures placing them in danger. Concrete is quasi-brittle in tension
and easy to crack under cyclic loadings. Fibre reinforced strain-hardening
cementitious composites (SHCC) featuring high ductility, high energy absorbing
capacity and controlled multiple cracking have been proposed for seismic-
resistant applications. The fine cracks have been proved to not only improve the
durability but also enhance the autogenous self-healing ability. This study
focuses on investigating the material behaviour and self-healing potential of
SHCC under cyclic flexural loading conditions. Four-point flexural tests (in-
cluding monotonic and cyclic tests) were performed to study its mechanical
properties and cracking behaviour. The surface crack widths were measured by
optical microscopy. Results showed that 28-day air cured specimens exhibited a
deflection capacity of up to 9.6 mm and an average crack width of 28 lm under
monotonic flexural loading. Regarding the flexural stress-deflection curves, the
envelops of cyclic testing results were close to monotonic results with both
elastic and hardening phases. SHCC could still maintain fine crack widths
(below 60 lm) under cyclic loading conditions. SEM/EDX test was conducted
to investigate the fibre-matrix interface.
Keywords: Strain-hardening Self-healing Multiple cracking Cyclic

loading
1 Introduction
During an earthquake, the ground shaking initiated by seismic waves often applies
cyclic forces to infrastructures and threatens the integrity and stability of constructions.
Concrete is inherently brittle and is extremely weak under tensile or cyclic shear
loadings. Although shear reinforcement is often applied to avoid brittle failure and
enhance shear strength, large cracks will still occur affecting the durability of the
reinforced concrete section. Traditional fibre reinforced concrete (FRC) is still quasi-
brittle with low ductility under tensile stress, and the crack widths are not well con-
trolled. Therefore, a more durable material featuring high ductility and energy
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_43
Characterisation of SHCC Under Cyclic Loading Conditions 467
absorption capacity to bear cyclic loadings without brittle failure is required for seismic
resistance.
Moreover, substantial and timely repair work is usually required after a large
earthquake due to unavoidable damage, which is often costly and difficult. Cementi-
tious materials inherently have the ability to heal small cracks and regain transport and
mechanical properties mainly due to continuing hydration, carbonation and pozzolanic
reactions; called autogenous healing [1]. However, this mechanism can only heal very
small cracks (mostly < 100 lm) depending on 1) intrinsic features such as age and mix
design of the cement-based material, crack widths and depths; 2) external healing
conditions such as environmental exposure (e.g. availability of water and CO2 sur-
rounding cracks, thermal conditions), stress state and steadiness of the cracked state
(mechanical loadings) and healing duration [2, 3].
Hence, it is beneficial to develop a ductile material with a more consistent self-
healing ability against seismic damage or cyclic loadings. Fibre-reinforced strain-
hardening cementitious composites (SHCC), represented by Engineered Cementitious
Composites (ECC) can be a good candidate [4]. Designed based on micromechanics
models and using special tailoring techniques, SHCC incorporates a low fibre content
(typically 2%) and relies on a synergistic interaction between the fibres, matrix and
the interface to realize a tensile strain-hardening effect. SHCC exhibits a multiple
cracking behaviour under tension due to effective fibre bridging and load transfer
mechanisms. Accordingly, SHCC could achieve much higher strain capacity or
deformation ability than traditional FRC under tension or flexural stress. The high
ductility and fracture energy have made SHCC a promising seismic-resistant material
[5]. Moreover, the cracking control capacity of SHCC has also been proved to effec-
tively enhance the durability and promote autogenous self-healing of the cementitious
material in different environments [6–9].
This research aims at characterising and studying the performance of SHCC under
cyclic loading conditions and evaluating its potential for self-healing. Based on existing
studies on SHCC, M45-ECC [10] is one of the most investigated mixtures and has
exhibited overall good mechanical performance, featuring a high tensile strain capacity
(>3%), moderate tensile strength and fine crack opening (crack widths < 100 lm) [6].
In this study, this standard mix was selected to assess the mechanical behaviour and
crack patterns of SHCC under monotonic and cyclic flexural loading. Results of this
preliminary investigation will inform further self-healing studies.
2 Materials and Mix Proportions
2.1 Materials
CEM-I 52.5 N High strength Portland cement (C) supplied by Hanson UK was used as
the clinker source. Silica sand (S) with a maximum grain size of 250 lm was adopted
as fine aggregates. Fly ash (FA) with a fineness category of N according to BS EN 450-
1 was supplied by CEMEX UK as a supplementary cementitious material. No coarse
aggregates were adopted in order to maintain the micromechanical properties between
the matrix, fibres and interfaces. A polycarboxylate-based superplasticiser (SP) with a
468 Z. Tang et al.
feature of enhancing workability retention and a viscosity modifying admixture

(VMA) were incorporated to control the fresh properties of SHCC and to achieve
uniform fibre distribution, both of which were supplied by Sika UK. Polyvinyl alcohol
(PVA) fibres produced by Kuraray, Japan were coated with oil (1.2% by weight) to
reduce the excessively strong interfacial bonding with the matrix. The properties of the
PVA fibres are provided in Table 1.
Table 1. Properties of PVA fibres.

Density Tensile strength Elastic Elongation Diameter Length
(g/cm3) (GPa) modulus (GPa) (%) (lm) (mm)
1.3 1.56 41 6.5 40 12
2.2 Mixture Proportions

The mixture proportions were similar to [11]. The ratio of water to binder (W/B, or
W/(C + FA)), was set as 0.26, the sand to binder ratio (S/B) was fixed at 0.36, and
FA/C was 1.2. The fibres were added as 2% of the total mixture volume. The mix
proportions are shown in Table 2.
Table 2. The mix proportions of SHCC investigated.

Ingredients Cementitious Aggregates Chemical admixtures
materials/binders
C FA S W SP VMA Fibre
Mass ratio 1 1.2 0.8 0.58 1.69% 0.49% 4.80%
Unit weight (kg/m3) 542 650 434 314 9.18 2.67 26
2.3 Specimen Preparation and Curing

Cement, sand and fly ash were first dry mixed in a Hobart mixer, after which a pre-
mixed water solution containing SP and VMA was added and uniformly mixed for
3.5 min. Then a flow table test according to BS EN 12350-5 was performed to quantify
and check the workability. The flow spread was 290 ± 10 mm. At last, fibres were
added and mixed for 3 min. The fresh SHCC mixture was then cast into moulds. The
specimens included 40 40 40 mm3 cubes and Ø75 150 mm3 cylinders for
compressive tests, and 300 (length) 50 (width) 26 (height) mm3 prisms for four-
point flexural tests. Two curing environments were selected; high moisture and
ambient. For the former the specimens were stored in a high moisture container at
98 ± 2% relative humidity (RH) and 20 ± 2 °C until the required age was reached.
Conversely for the latter, the specimens were placed in ambient conditions with
50 ± 5% RH and a temperature of 20 ± 2 °C.
3 Methods
3.1 Physical and Mechanical Properties
3.1.1 Unconfined Compressive Strength and Density
Unconfined compressive strength tests were conducted on cubic specimens
(40 40 40 mm3) of different ages. Triplicate specimens were tested in a 250kN
testing cell with a vertical loading speed of 2400 N/s according to BS EN 196-1. The
density of SHCC was calculated by measuring and averaging the weights and volumes
of 24 cubic specimens.
3.1.2 Secant Modulus of Elasticity

Secant modulus of elasticity in compression was also tested based on BS EN 12390-13.
Three cylindrical specimens with a dimension of Ø75 150 mm3 were cured in a high
moisture condition for 7 days before testing. One day before testing, the top surfaces of
specimens were capped to make the surface smooth and flat. Using a 2000 kN testing
frame (Controls), two specimens were loaded till failure to achieve an average com-
pressive strength fc. Based on fc, a loading profile containing three preloading cycles
and three loading cycles within the elastic range were designed, and then another
specimen from the same batch was loaded to obtain the secant modulus of elasticity.
3.1.3 Four-Point Flexural Tests (Monotonic and Cyclic)

Prisms with a dimension of 300 (length) 50 (width) 26 (height) mm3 were loaded
using a 30 kN Instron machine under displacement control. The loading span was
80 mm and the supporting span was 240 mm. Three types of tests, i.e. monotonic
loading tests, one-side loading/unloading cyclic tests and reversed cyclic tests were
conducted. For monotonic loading tests, the loading rate regarding the crosshead dis-
placement (u) was 0.2 mm/min and all the specimens were loaded until failure. For
cyclic loading tests, the loading speed was a combination of 0.2 and 0.5 mm/min
considering the testing efficiency. The midspan deflection (v) of each specimen was
monitored by using a laser extensometer.
3.2 Microstructural Observation

3.2.1 Optical Microscopy
A Leica DM 2700 M stereoscope was used for measuring all the crack widths gen-
erated during the four-point flexural tests. Since flexural cracks are often in a wedge
shape, the maximum width of each crack was measured from the bottom surface of the
specimens. For specimens with a total crack number of no more than ten, each indi-
vidual crack was identified and nine evenly-spaced measuring points were positioned
along each crack. For those with a crack number greater than ten, since individual
cracks were difficult to classify and one crack often connected or branched to another,
three parallel lines along the specimen span were drawn (Fig. 1) and all the crack
widths crossed by the lines were measured and averaged.
470 Z. Tang et al.
80mm
Line 1
Line 2
Line 3
Fig. 1. Multiple cracking pattern of a 28-day air-cured SHCC specimen and the crack width
measuring method.
3.2.2 SEM-EDX Analysis

Scanning electron microscopy with an energy-dispersive X-ray spectroscopy detector
(SEM-EDX) was applied to observe the fibre-matrix interfaces, specifically the fibre
bridging behaviour and the substance on the fibre surface. Phenom ProSuite software
was used for EDX analysis. 0.5 0.5 0.5 mm3 specimens were cut from large
cracked SHCC specimens and completely dried by storing in an oven with a constant
temperature of 40 °C for one day before testing.
4.1 Physical and Mechanical Properties

4.1.1 Density, Unconfined Compressive Strength and Secant Modulus
of Elasticity
Results indicated that the development of compressive strength was not significantly
affected by the curing conditions after seven days of high moisture curing. In particular,
under consistent high moisture condition, the compressive strength was 37.52 (±0.83)
MPa at 3 days, 50.33 (±2.71) MPa at 7 days and 75.42 (±2.00) MPa at 28 days,
whereas for specimens cured in high moisture for 7 days and then cured in ambient
conditions (RH = 50 ± 5%) until 28 days the corresponding strength was 71.55
(±0.71) MPa at 28 days. The average density of SHCC specimens was 2050 kg/m3.
Moreover, the 7-day modulus of elasticity was obtained. Two specimens were first
tested to obtain the compressive strength, and then by loading the third specimen, the
initial modulus of elasticity (13.3 (±2.1) GPa) and the stabilised modulus of elasticity
(14.0 (±0.7) GPa) were calculated.
4.1.2 Behaviour Under Monotonic Flexural Loading

7-day old prisms were tested under monotonic flexural loading until failure, which all
exhibited deflection-hardening behaviour. The flexural stress-deflection curves are
presented in Fig. 2(a). After the initial crack, the stress continued to rise with some
fluctuations while the midspan deflection of SHCC increased significantly and multiple
fine cracks were generated. According to visual observation, as the deformation
increased, the crack number increased until a localised crack occurred. Then the width
of the localised crack expanded until the fibres could not sustain the tensile stress, and
finally the specimen softened with a continuous decreasing stress. Representative
salient points are summarised in Table 3. The modulus of rupture (MOR) is defined as
the flexural stress just before softening, and the deflection capacity is the corresponding
midspan deflection. The flexural stiffness is defined as the slope of flexural stress-
deflection curve when the flexural stress is between 1 and 3 MPa, where the slope is
almost linear [12].
Table 3. Characteristics of 7-day SHCC under monotonic flexural loading.

First cracking First cracking MOR Deflection Flexural stiffness
strength (MPa) deflection (mm) (MPa) capacity (mm) (MPa/mm)
4.46 (±0.40) 0.26 (±0.04) 7.90 3.79 (±0.44) 19.13 (±1.41)
(±0.91)
28-day cured specimens were also tested under monotonic flexural loading, and the
curing conditions included both high moisture and ambient curing. Under the former
curing condition, the 28-day specimens exhibited higher flexural strength than the 7-
day cured specimens due to the longer hydration duration. For the 28-day cured
specimens, the air-cured specimens showed significantly higher deflection capacity (up
to 9.6 mm), slightly lower flexural strength (11.7 MPa) and lower stiffness than the
moisture-cured specimens (5.5 mm and 13.2 MPa respectively). The explanation could
be that the matrix tended to develop a higher fracture toughness and was more difficult
to crack after high moisture curing; accordingly, based on micromechanical analysis,
the crack tip toughness would increase while deflection-hardening behaviour and
flexural ductility would decrease. Typical flexural stress-deflection curves under dif-
ferent curing conditions are given in Fig. 2(b).
10
14
8 12
Flexural stress (MPa)
10
6
8
4 6
Specimen No.1 4 28d high moisture

2 Specimen No.2 28d air
2
Specimen No.3 7d high moisture
0 0
0 1 2 3 4 5 0 2 4 6 8 10 12
Midspan deflection (mm) Midspan deflection (mm)
(a) (b)
Fig. 2. Flexural stress-deflection curves of (a) 7-day SHCC (high moisture curing); (b) SHCC
under different curing conditions or durations.
472 Z. Tang et al.
4.1.3 Behaviour Under Cyclic Loading-Unloading

7-day SHCC prisms were tested under loading-unloading cycles until failure. During
each cycle, the specimen was loaded until the crosshead displacement reached a certain
value um and then unloaded. When um < 1 mm, the displacement increment was
Du = 0.2 mm, and the loading-unloading rate was 0.2 mm/min; when um > 1 mm,
Du = 0.5 mm and the loading-unloading speed was 0.5 mm/min. The resulting stress-
deflection curve is given in Fig. 3. The envelop curve of the cyclic testing result was
close to the monotonic stress-deflection profile, showing that the specimen still expe-
rienced an elastic phase, a hardening phase and a softening phase. The resulting MOR
was slightly higher probably because a higher loading rate typically results in higher
fibre strength and composite strength [13]. The cracking pattern of multiple parallel fine
cracks was also similar to that for monotonic tests, as shown in Fig. 3.
80mm
12
10
0
0 1 2 3 4 5 6 7
Midspan deflection (mm)
Fig. 3. Flexural stress-deflection curve and cracking pattern of a 7-day SHCC specimen under
cyclic loading-unloading.
4.1.4 Performance Under Reversed Flexural Loading Cycles

7-day SHCC specimens were tested under symmetrically reversed cyclic loading up to
three cycles. The increase in the midspan deflection Dv was 1 mm between cycles. The
loading and unloading speed in both directions was 0.5 mm/min. Typical flexural
stress-deflection curves are shown in Fig. 4. Based on visual observation, in every
cycle, the reversed force could help close previous cracks in one half part of the
specimen while re-opening old cracks or generating new cracks in the other half
part. On the other hand, when the specimen was unloaded to zero stress after each
cycle, a residual deformation was generated. This shows that when the SHCC has been
deformed to crack and beyond elastic phase, plastic deformation would occur and the
material could not retrieve its original shape, and the cracks would not completely close
without external forces. This is because the process of fibre debonding, pull-out or
rupture could not be reversed. The indication is that if the cracked SHCC went through
autogenous self-healing, the cracks could be filled with healing products and the
bonding strength of fibres might recover to some extent, although the ruptured fibres
could no longer bridge the crack.
10
0
-3 -2 -1 0 1 2 3
-5
Cycle 1
Cycle 2
-10 Cycle 3
Midspan deflection (mm)
Fig. 4. Hysteresis loop under reversed cyclic loading.
4.2 Microstructural Observation

4.2.1 Crack Width Measurement
Under flexural loading, most cracks generated were approximately perpendicular to the
neutral axis. An example of micro-cracks under stereoscope is shown in Fig. 5. The
total crack numbers and non-localised crack widths of specimens subjected to mono-
tonic flexural loading are summarised in Table 4. Overall, the air cured specimens
exhibited a higher number of cracks and smaller crack widths than the high moisture
cured specimens. These were also in accordance with the flexural performance and
geometric properties of SHCC. A larger product of the average crack width and total
crack number tended to contribute to a higher deflection capacity. The crack widths
were 58 (±15) lm for cyclic loading-unloading tests and 38 (±14) lm for reversed
cyclic tests. The average crack widths for either monotonic or cyclic loading condition
were below 60 lm thanks to the bridging effect of fibres, indicating that these cracks
could potentially be completely self-healed in favourable environments [8].
474 Z. Tang et al.
Fig. 5. Micro-cracks under stereoscope.
Table 4. Crack number and widths of SHCC.

Curing condition 7-day high moisture 28-day high moisture 28-day air
Crack number 12 32 61
Crack width (lm) 51 (±19) 35 (±17) 28 (±17)
4.2.2 SEM/EDX Analysis

From SEM images, it is shown that PVA fibres in cracks either bridged the crack,
ruptured or were pulled out due to the crack opening. When the crack widths were
small (generally below 100 lm), fibres worked together to bridge the crack, and being
in the debonding stage were able to generate steady-state cracks. As a crack localised
and its width increased, fibres across this crack started to lose function. They tended to
rupture when the pull-out force exceeded their strength (Fig. 6(a)), or they were pulled
out, as shown in Fig. 6(b). This confirms previous findings [14].
The fibres were covered with cementitious materials according to both SEM images
and EDX analysis (Fig. 6(c)). Fibres are composed of carbon and oxygen. The
materials attached on the fibre surfaces exhibited an element combination of either Ca,
Si, O or Ca, Si, O, Al, showing that they were probably calcium silicate hydrate
(C-S-H), unhydrated cement particles or fly ash (in circular shapes). The latter two are
essential for continuing hydration and pozzolanic reaction, which promote autogenous
self-healing. This is in agreement with [15], i.e. the fibres could not only bridge cracks
mechanically, but also act as a scaffolding for autogenous healing products to pre-
cipitate on due to the high polarity of PVA fibres, thus the cracks could be filled more
efficiently. Larger cracks (up to 0.3 mm) in PVA-SHCC could also be more effectively
healed compared with those in normal concrete under favourable conditions.
Further work is underway to assess the self-healing performance of the investigated
mixes.
(a) (b) (c)
Fig. 6. SEM images of (a) fibres in a crack and ruptured fibres; (b) bridging fibres being
partially pulled out; (c) hydration products and unhydrated particles on fibre surface.
5 Conclusions
In the work herein we investigated the behaviour of SHCC under monotonic loading,
cyclic loading-unloading and reversed cyclic flexural loading conditions respectively
and studied the cracking behaviour and microstructural features. The main findings are
summarised below:
• The compressive strength of SHCC was 50.33 MPa at 7 days and 75.42 MPa at 28
days under high moisture curing.
• The SHCC specimens exhibited deflection-hardening and multiple cracking beha-
viour under either monotonic or cyclic flexural loading conditions. The envelop
curves of the cyclic testing results were close to the monotonic stress-deflection
profile, including elastic, hardening and softening phases.
• 28-day air-cured specimens showed higher deflection capacity and smaller crack
width than the 7-day and 28-day high-moisture-cured specimens, exhibiting the
highest potential self-healing. Further mechanical testing is required for valid
experimental proof.
• In the reversed cyclic tests, the reversed force could help close previous cracks at
one half part of the specimen while generating new cracks or re-opening old cracks
in the other half of the specimen.
• The specimens could not return to their original shape when unloading after plastic
deformation, showing that the process of cracking, fibre debonding, pull-out or
rupture was irreversible. However, if the cracked SHCC went through autogenous
self-healing, the cracks would be filled with healing products and the bonding
strength of fibres might recover to some extent, which also requires further vali-
dation and is the object of ongoing work.
• The fibres were covered with cementitious materials and hydration products
according to both SEM images and EDX analysis, confirming that PVA fibres with
a high polarity could not only bridge the cracks mechanically, but also act as a
medium for autogenous healing products to participate on, thus larger cracks could
potentially be healed.
476 Z. Tang et al.
Acknowledgements. The financial support from the EPSRC for the Resilient Materials for Life
(RM4L) Programme Grant (EP/P02081X/1) is gratefully acknowledged. The financial support
from the Cambridge Commonwealth, European and International Trust (CCEIT) for the first
author’s PhD research, is highly appreciated.
References
1. de Rooij, M., Van Tittelboom, K., De Belie, N., Schlangen, E.: Self-healing phenomena in
cement-based materials. Springer (2013)
2. Huang, H., Ye, G., Qian, C., Schlangen, E.: Self-healing in cementitious materials:
Materials, methods and service conditions. Mater. Des. 92, 499–511 (2016)
3. De Belie, N., Gruyaert, E., Al-Tabbaa, A., Antonaci, P., Baera, C., Bajare, D., Darquennes,
A., Davies, R., Ferrara, L., Jefferson, T., Litina, C., Miljevic, B., Otlewska, A., Ranogajec,
J., Roig-Flores, M., Paine, K., Lukowski, P., Serna, P., Tulliani, J.M., Vucetic, S., Wang, J.,
Jonkers, H.M.: A review of self-Healing concrete for damage management of structures.
Adv. Mater. Interfaces. 15(7), 28 (2018)
4. Li, V.C.: From micromechanics to structural engineering - the design of cementitious
composites for civil engineering applications. J. Struct. Mech. Earthq. Eng. 10(2), 37–48
(1993)
5. Li, V.C.: On Engineered Cementitious Composites (ECC) - a review of the material and its
6. Yildirim, G., Keskin, Ö.K., Keskin, S.B.I., Şahmaran, M., Lachemi, M.: A review of
intrinsic self-healing capability of engineered cementitious composites: recovery of transport
and mechanical properties. Constr. Build. Mater. 101, 10–21 (2015)
7. Li, V.C., Herbert, E.: Robust self-healing concrete for sustainable infrastructure. J. Adv.
Concr. Technol. 10, 207–218 (2012)
8. Wu, M., Johannesson, B., Geiker, M.: A review: self-healing in cementitious materials and
engineered cementitious composite as a self-healing material. Constr. Build. Mater. 28, 571–
583 (2012)
9. Van Zijl, G.P.A.G., Wittmann, F.H.: Durability of strain-hardening fibre-reinforced cement-
based composites (SHCC). Springer (2011)
10. Wang, S., Li, V.C.: Engineered Cementitious Composites with high-volume fly ash. ACI
Mater. J. 104(3), 233–268 (2007)
11. Yang, Y., Lepech, M.D., Yang, E.-H., Li, V.C.: Autogenous healing of engineered
cementitious composites under wet-dry cycles. Cem. Concr. Res. 39, 382–390 (2009)
12. Qian, S.Z., Zhou, J., Schlangen, E.: Influence of curing condition and precracking time on
the self-healing behavior of Engineered Cementitious Composites. Cem. Concr. Compos.
32, 686–693 (2010)
13. Yu, K., Li, L., Yu, J., Wang, Y., Ye, J., Xu, Q.: Direct tensile properties of engineered
cementitious composites: a review. Constr. Build. Mater. 165, 346–362 (2018)
14. Redon, C., Li, V.C., Wu, C., Hoshiro, H., Saito, T., Ogawa, A.: Measuring and modifying
interface properties of PVA fibers in ECC matrix. J. Mater. Civ. Eng. 13(6), 399–406 (2001)
15. Nishiwaki, T., Koda, M., Yamada, M., Mihashi, H., Kikuta, T.: Experimental study on self-
healing capability of FRCC using different types of synthetic fibers. J. Adv. Concr. Technol.
10, 195–206 (2012)
Corrosion Pattern and Mechanical Behaviour
of Corroded Rebars in Cracked Plain
and Fibre Reinforced Concrete
E. Chen1(&), Carlos G. Berrocal1,2, Ingemar Löfgren1,2,

and Karin Lundgren1
1
Division of Structural Engineering, Chalmers University of Technology,
Gothenburg, Sweden
teresa.chen@chalmers.se
2
Thomas Concrete Group AB, Gothenburg, Sweden
Abstract. This paper presents experimental results of corrosion pattern and

tensile behaviour of corroded rebars extracted from 4 uncracked and 18 pre-
cracked plain concrete and fibre reinforced concrete (FRC) beams. The beams
were pre-cracked through three-point bending to a target maximum crack width
of 0.1 and 0.4 mm, and then subjected to natural corrosion through cyclic
exposure to a 16.5% NaCl solution for more than three years. 3D-scanning was
used to characterise the pit morphology and evaluate the maximum local cor-
rosion level of extracted rebars. Under the same loading condition and crack
width, most rebars in FRC had smaller maximum local corrosion level than
those in plain concrete. Subsequently, tensile tests were carried out on the
extracted rebars, with Digital Image Correlation (DIC) technique adopted to
investigate the influence of pit morphology on the local strain development.
Finally, the time-dependent influence of transverse and longitudinal cracks on
the pit morphology which governs the ultimate strain of corroded rebars was
discussed. The time-varying nature of corrosion morphology should be con-
sidered when predicting the durability and long-term safety of conventional
reinforced concrete and FRC structures with reinforcing bars under chloride
environments.
Keywords: Fibre reinforced concrete Pitting corrosion Corrosion-induced

crack Tensile behaviour Durability
1 Introduction
Reinforced Concrete (RC) structures inevitably possess cracks, originating from

shrinkage, thermal gradients and/or mechanical loading. These cracks accelerate the
ingress of various adverse agents, leading to earlier corrosion initiation. However,
regarding to the effect of cracks on the corrosion propagation, contradictory results
have been obtained, as shown in the state-of-the-art report [1]. Some current codes
dictate the maximum allowable crack width in addition to the minimum cover depth to
limit the risk of corrosion. The restrictive requirement of controlling crack width often
results in congested reinforcement layouts for structures exposed to marine or de-icing
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_44
478 E. Chen et al.
salt environment. Fibre reinforcement combined with conventional reinforcing bars is

an attractive alternative to avoid deploying heavy steel bars. Fibres are an effective
means of crack control and can also improve the mechanical performance of concrete
structures [2]. Nevertheless, the long-term performance of fibre reinforced concrete
(FRC) combined with conventional steel bars under natural corrosion condition has not
be adequately studied so far.
A long-term experimental program on conventional RC beams and concrete beams
reinforced with steel bars and various types of fibres (with a volume fraction of <1%)
subjected to cyclic exposure of chloride has been initiated at Chalmers University of
Technology since 2013. Flexural cracks were induced in the beams through three-point
bending, with the maximum flexural crack controlled as 0.1, 0.2, 0.3 and 0.4 mm. The
effect of crack width and fibres on the corrosion initiation time of steel bars in the
beams has been examined in an earlier study by the authors [3]. Experimental results
showed a tendency towards earlier initiation of corrosion, with increasing crack widths.
A small improvement (in terms of delayed corrosion initiation) was observed when
fibres were added. In the subsequent study [4], the flexural behaviour of FRC beams
were found to display higher residual load capacity at reinforcement yielding than plain
RC beams after the same period of corrosion, whereas the relative loss of load capacity
as a function of the maximum local corrosion level was similar to that seen in plain RC
beams.
The present study investigated the remaining uncracked beams and cracked ones
with target crack widths of 0.1 and 0.4 mm prepared in [3]. The first purpose was to
examine the effect of cracks and fibres on the corrosion characteristics of rebars. The
other purpose of this study was to investigate the influence of pit morphology on the
mechanical behaviour of corroded rebars. In particular, a 3D-scanning technique was
used to characterise the pit morphology, including the maximum local corrosion level
(defined as the maximum cross-sectional area loss percentage) as well as different
geometry parameters. Subsequently, Digital Image Correlation (DIC) was used in the
rebar tensile tests to study the strain behaviour along the bars exhibiting various pit
morphologies. The results of this study provide basic data to develop a structural safety
model for FRC structures in a chloride environment and quantitatively appraise the
service life of conventional RC and FRC structures with reinforcing bars.
2 Experimental Description
2.1 Specimens
The beams had the dimension of 1100 180 100 mm, reinforced with three
U10 mm “as-received” ribbed rebars at a clear concrete cover of 30 mm. Four different
series of specimens were cast, including conventional reinforced beams referred to as
“plain” (PL) series, and three FRC series with different types of fibre reinforcement,
referred to as “steel” (ST), “hybrid” (HY) and “synthetic” (SY) series. A self-
compacting concrete mix with the same water/cement ratio of 0.47 was used for all the
series. The types of fibres used for the different FRC series were 35 mm end-hooked
steel fibres for the ST series, 30 mm straight polyvinyl alcohol (PVA) for the SY series
Corrosion Pattern and Mechanical Behaviour of Corroded Rebars 479
and a combination of steel fibres and 18 mm long PVA fibres for the HY series. The
mix proportions can be found in [4].
The beams were pre-cracked under three-point bending until the maximum crack
width reached a prescribed value. Three different loading conditions were considered:
loaded once and then unloaded, five loaded-unloaded cycles to promote greater damage
at the rebar-concrete interface, and sustained loading to keep the maximum crack width
as the target value. They were denoted as “unloaded” (U), “cyclic” (C) and “loaded”
(L). The beams never loaded were noted as “uncracked” (N). In one of the ST beams
under cyclic loading, the maximum crack width reached 0.8 mm instead of its target
crack width of 0.4 mm. For beams subjected to unloaded and cyclic conditions, the
cracks closed partially, and the remaining surface crack width ranged between 0.02 and
0.06 mm. After the pre-loading, all the beams were subjected to wet-drying cyclic
exposure of chloride solution (16.5% NaCl). After three years’ exposure, the specimens
were stored in the laboratory until testing.
Table 1 summarises the specimens investigated in this study. The uncracked and
six cracked PL beams, and two cracked ST beams were taken out for testing after being
stored in the laboratory for 18 months, while the other 13 beams were tested after 24
months’ storage. The influence of the six-month time difference of the storing time
following the exposure period was ignored as the corrosion rate during this period was
expected to be low compared to that in the exposure period.
2.2 Crack Mapping

Before extracting the rebars from the beams, both the pre-induced flexural cracks and
corrosion-induced cracks (longitudinal cracks) were mapped for all the beams. In
addition, the crack widths of beams tested after 24-month storage were measured using
a microscope with 0.02 mm resolution.
2.3 Corrosion Level Evaluation with 3D-Scanning

The extracted bars were cleaned with sand-blasting to remove corrosion products and
adhered concrete. The total corrosion level of each rebar was evaluated from the weight
loss. Then the rebars were strategically cut to obtain 500 mm segments with the most
severe pit of the rebar located in the middle region for 3D-scanning and tensile testing.
In two rebars with more than one pit of seemingly similar severity at a relatively close
position, they were cut to a length of 550 mm to incorporate both pits in the same
segment. As the weight loss of all the bar segments in this study was found to be less
than 3.5%, much smaller compared to the local steel loss, it was not discussed in this
study. Figure 1a–b shows the 3D surface mesh and longitudinal variation of the cross-
sectional area of one bar. The maximum local corrosion level lmax is defined as the
cross-sectional area loss percentage at the section with minimum remaining area (i.e.
the minimum cross-section):
0 Ac
Amin min
lmax ¼ min ð1Þ
A0
480 E. Chen et al.
Table 1. Specimens in this study.

Load conditions Series Target Pre- Exposure Storing
crack loading period period
widths time before
(mm) testing
Uncracked PL – 10- 3 years 18
week months
ST – age 24
months
HY – 24
months
SY – 24
months
Cracked Unloaded 1 PL 0.1 0.4 18
cycle months
ST 0.1 0.4 18
months
HY 0.1 0.4 24
months
SY 0.1 0.4 24
months
5 PL 0.1 0.4 18
cycles months
ST 0.1 0.4 24
(0.8) months
HY 0.1 0.4 24
months
Loaded PL 0.1 0.4 18
months
ST 0.1 0.4 24
months
where Amin min

0 and Ac are the original and remaining areas at the minimum cross-section.
The appearance of the minimum cross-section and its original uncorroded section,
definitions of pit length lp , pit dept xp and pit width wp are given in Fig. 1c.
2.4 Rebar Tensile Test with Digital Image Correlation System

After characterizing the corrosion morphology of rebars, tensile tests were carried out
on an MTS Universal Testing machine, according to the standard ISO EN 6892 [5].
The grip length at each end of the tested bar was 60 mm. The tensile load was applied
by controlling the displacement increment, at a speed of 0.5 mm/min in the elastic
stage and 2 mm/min afterwards. The DIC apparatus, ARAMIS Adjustable camera
system was used. The working principle of DIC is using two cameras to track the 3D
(a)
(b)
(c)
Fig. 1. 3D-scanning and corrosion evaluation in MATLAB.
coordinates of points within the measuring volume, and deriving the displacements,
strains, velocities or accelerations from the coordinates’ changes between images.
Therefore, the surface of the bars should have stochastic pattern displaying random
gray values and high contrast in a random neighborhood. This was done by spraying
white and black paint successively. After calibration of the cameras, the support rods of
the cameras were moved to include the most severe pit (which is also the failure zone)
within the capturing volume 100 75 55 mm3. The experimental set-up is shown in
Fig. 2. The local strain along the bar and engineering strain based on several different
strain gauges shown in Fig. 3 were computed.
Corrosion pit
DIC cameras
Fig. 2. The set-up for the tensile test of bars

482 E. Chen et al.
Corrosion pit
Fig. 3. Strain gauges defined in the DIC software
3 Results on the Crack and Corrosion Characteristics
3.1 Relation Between the Crack Pattern and Corrosion Pattern

The rebars in uncracked beams only presented light corrosion in the form of randomly
distributed tiny pits. In pre-cracked beams, several pits were formed on separate sites of
a rebar, with most pits located near the flexural cracks. However, in approximately half
of the flexural cracks, no corrosion was found. This is probably due to the mechanism
described in [6], which suggests that corrosion is induced at the widest crack or
weakest position first, which delays and suppresses corrosion in other cracks. It should
be noted that some flexural cracks less than 0.02 mm were found to be partially or fully
healed, or filled with white material after years of exposures. In the previous studies
[7, 8], increasing crack frequency decreased the local corrosion rate at each corroding
spot, because the crack distance limited the cathodic area available to contribute to the
macro-cell current. However, as the flexural crack spacing was not uniform and the
difference in crack frequency between plain series and fibre reinforced series was small
(less than 2x), the influence of crack spacing on the corrosion level of each pit along a
rebar was not examined. Only the maximum local corrosion level of the most severe pit
along each rebar (defined in Eq. (1)) was focused in this study.
The pit shapes were mainly elliptical, while some pits exhibited extended corrosion
next to the elliptical cavity which was probably caused by the presence of longitudinal
cracks. Figure 3 shows three examples of the relation between the crack pattern and
corrosion pattern. From Fig. 4, the pit length was found to be quite relevant to the
extent of longitudinal cracks, whereas the maximum local corrosion level did not have
an increasing relation with the longitudinal crack width. It should be aware that a rebar
may lose a large amount of cross-sectional area in a localised pit, but no large corrosion
cracks are produced thus not giving any warning of ongoing corrosion. This is of major
importance for structural condition assessment.
3.2 Maximum Local Corrosion Level

In this section, the effects of crack width, loading condition and fibre reinforcement on
the maximum local corrosion level of rebars are presented. One bar in PL-U0.4, one bar
in ST-L0.4 and one bar in SY-U0.1 were bent severely during the extraction, so their
corrosion levels were not evaluated. Figure 5 shows the maximum local corrosion level
Fig. 4. Mapping cracks and longitudinal variation of the cross-sectional area.
of bars in all beams. For three rebars in the same beam, the maximum local corrosion
levels varied.
Fig. 5. Maximum local corrosion level of all bars
From Fig. 5, the uncracked beams have much smaller lmax than pre-cracked ones;
however, no consistent relation of lmax with the target crack width of 0.1 and 0.4 mm
was observed. However, the three rebars in the ST beam cyclically loaded by accident
to an 0.8 mm crack width all had lmax greater than 25%, much more severe than other
beams.
For PL and ST series, most rebars in the unloaded condition had larger lmax than
those in cyclic and loaded conditions for PL-0.1, PL-0.4 and ST-0.4 groups. For ST-0.1
group, the loaded beam had the largest lmax among the three loading conditions.
Moreover, it was found that under the unloaded condition, elliptical pit shape with a
locally deep pit depth was dominant, whereas it was found that pits were longer but
with shallow depth for most loaded cases. The long pit length formed on bars in loaded
beams may relate to the greater extent of slip and separation between concrete and
484 E. Chen et al.
steel, while a more localised pit is likely to form where lesser interface damage is
caused under the unloaded condition.
Under the same loading condition and target crack width, the majority of rebars in
ST series showed a lower lmax than PL series. Most rebars in HY and SY series had
similar or lower lmax than PL series under the target crack width of 0.4 mm. However,
the opposite was true under the target crack width of 0.1 mm. This suggests fibre
reinforcement may be more effective in reducing the corrosion level when the achieved
crack width is larger. Further, HY beams showed most scattered lmax for the three bars
in the same beam in the four mixes. This may be due to uneven fibre distribution,
causing locally severe damage during pre-loading and severe pitting corrosion on one
of the bars in a beam. More specimens need to be studied before conclusions can be
drawn regarding the effect of fibres on the maximum local corrosion level.
4 Results on the Mechanical Behaviour of Corroded Bars

4.1 Load-Strain Curves and Strength Versus Maximum Local Corrosion
Level
The load-strain (F-e) curves of all the bars are shown in Fig. 6, in which the strain was
calculated based on the 50-mm long strain gauge across the failure zone. The colour of
each curve represents the bar’s maximum local corrosion level, lmax . From Fig. 6, with
increasing corrosion level, the load capacity and ductility of rebars decreased. The
uncorroded steel bars exhibit a distinct yield plateau before strain hardening. With
increasing lmax , the yield plateau becomes indiscernible and the strain hardening stage
is reduced at the same time, leading to a brittle failure at higher corrosion levels.
Fig. 6. Load-strain curves of all the bars. Fig. 7. Strength versus lmax .
As for most corroded rebars, the yield plateau disappeared, proof strength at a total
strain of 0.5% was taken to replace the yield strength. The relations of the proof and
ultimate strengths (ft0:5 and fu ) with the maximum local corrosion level are shown in
Fig. 7. It should be mentioned that the strength was calculated based on the minimum
remaining cross-sectional area measured before tensile testing. It is found that the proof
and ultimate strengths remained nearly constant with increasing corrosion level.
4.2 Strain Properties of Corroded Bars

4.2.1 Local Strain Distribution and the Influence of Pit Morphology
The local axial strain, computed as the maximum principal strain in the local coordi-
nate, was examined. Figure 8 shows examples of the local strain field and its histogram
at ultimate (i.e. at the maximum load) for different bars. Strain localisation was
observed in the corrosion pit. Moreover, the strain outside the pit was significantly
reduced with increasing maximum local corrosion level. On the right side of Fig. 8, the
local strain fields of two bars with similar maximum local corrosion levels but large
difference in the pit length are displayed as well as their pit morphologies. It can be
inferred that the local strain distribution along the bar is closely related to the pit
morphology. Longer pit with less uneven pit depth variation along the pit has larger
strain capacity than shorter pit with more uneven pit depth variation.
4.2.2 Ultimate Strain from Different Strain Gauges Versus lmax

Due to the uneven distribution of local strain along the corroded bars, the ultimate
strain is dependent on the gauge length and position. Figure 9 shows the ultimate strain
normalised by the value of uncorroded bars eu =eu0 versus lmax . With increasing length
for the gauges across the failure zone (lg = 25, 50 and 75 mm), the ultimate strain
decreases; while for the gauges outside the failure zone ((lg = 5, 10 and 25 mm (out)),
the ultimate strain is stable regardless of the gauge length, and it is close to the strain
calculated from the total elongation (lg = 380 mm). From the DIC, it was observed that
the corrosion pit induces strain localisation very early (before the yielding); while in the
uncorroded bar (lmax ¼ 0 in Fig. 8), the strain has not been localised at ultimate which
is before necking occurs. Therefore, for uncorroded bars, the choice of gauge length
does not influence the strain values as long as the gauge length is larger than the rib
spacing, while it greatly affects for corroded bars. This phenomenon is similar to the
“size effect” on concrete strain in the softening stage of the stress-strain curve due to the
crack localisation. This finding may explain the large scatter of the ultimate strain of
corroded rebars reported previously. However, how to consider the gauge-dependent
strain capacity of corroded rebars in assessing the structural behaviour of corroded RC
structures needs further studies.
486 E. Chen et al.
μ =17.2%, lp=93.6mm
μ =0
μ =15.7%
μ =18.6%, lp=12.2mm
μ =32.7%
Fig. 8. Local strain field at ultimate.
Fig. 9. Normalised ultimate strain from different strain gauges versus lmax .
5 Discussions
5.1 Time-Varying Corrosion Morphology and Strain Capacity

of Corroded Rebars
For practical RC structures with pre-existing cracks, pitting corrosion may occur first
near some of those cracks. However, as corrosion-induced cracks are generated and
propagate in the longitudinal direction of rebars, the corrosion pits will have a major
change of the morphology from a narrow elliptical shape to a wide long shape.
Since the corrosion morphology plays an important role in the strain capacity of
corroded bars, the time-varying nature of the corrosion morphology needs to be taken
into account to assess the strain capacity of rebars. As pits grow longitudinally with
corrosion-induced cracking, the ultimate strain of rebars may decrease more slowly, or
even increase a little. Further work on the corrosion morphology evolution with con-
crete cracking and corrosion time is required in order to predict the time-dependent
strain capacity of corroded rebars.
5.2 Implications for the Durability and Long-Term Safety of FRC

Structures
The rebar corrosion in conventional concrete and FRC differ in several aspects due to
the different crack characteristics. Compared to plain concrete, fibre reinforced concrete
has improved cracking resistance to mechanical loading and corrosion expansion. First,
adding fibres can delay the corrosion initiation as the internal cracking is more tortuous
and steel-concrete interfacial damage is reduced under service loading. In the corrosion
propagation phase, corrosion-induced cracking can also be more resisted by FRC.
Hence, the total corrosion amount will be reduced in FRC. However, the corrosion
morphology may be more localised in FRC before corrosion-induced cracks
develop. Further, the ultimate strain of corroded rebars depends on the corrosion
morphology: a localised deep pit leads to more brittle behaviour of bars in tension than
a longer pit. In addition, it has been known that the residual tensile strength and greater
shear resistance of FRC also contributes to improve the mechanical performance of
structures. Lastly, corrosion of fibres may occur for steel fibres crossing large cracks as
they are directly exposed to the environment. As a result, all these aspects should be
considered in the future to quantitatively evaluate the improvement of FRC combined
with steel bars over conventional reinforced concrete under long-term chloride envi-
ronments. Further theoretical studies on the influence of particular cracks in FRC on the
local corrosion rate, as well as experimental work on specimens exposed for longer
periods, are needed to assess the overall performance of FRC structures in a chloride
environment.
6 Conclusions
This study characterised the corrosion morphology and mechanical behaviour of rebars
in un- and pre-cracked plain and fibre reinforced concrete beams subjected to chloride
solution for more than three years. The main conclusions are summarised as follows:
• The pre-induced cracks in beams greatly increased the corrosion level of rebars
compared to uncracked beams. However, the maximum local corrosion level did
not show obvious correlation with the maximum flexural crack width of 0.1 and
0.4 mm. However, the three bars in one of the beams occasionally pre-cracked to
0.8 mm showed much more severe corrosion than other beams.
• Most bars in the fibre series beams had lower maximum local corrosion levels than
the counterparts of plain series; however, a few bars presented more severe
488 E. Chen et al.
corrosion than in plain series beams. Moreover, the corrosion levels in the three bars
in one beam had large scatter in HY series. This might be caused by the uneven
fibre distribution.
• The DIC measurements enabled the observation of strain localisation in the cor-
rosion pits of rebars subjected to uniaxial tensile loading and revealed that the local
strain distribution is strongly influenced by the pit morphology. Furthermore, the
ultimate strain of corroded bars was found to depend on the gauge length and
position.
• The evolution of corrosion morphology from localised pitting corrosion at an early
corrosion period to extensive pitting corrosion with the propagation of corrosion-
induced cracks needs to be considered to evaluate the durability and long-term
safety of conventional reinforced concrete and FRC with reinforcing bars.
Acknowledgements. This work has been supported by: the Swedish Transport Administration
(project grant TRV 2018/36506); the construction industry’s organisation for research and
development (project grant SBUF-13683); Chalmers University of Technology; Thomas Con-
crete Group; and Cementa AB (Heidelberg Cement Group).
References
1. Boschmann Käthler, C., Angst, U.M., Wagner, M., Larsen, C.K., Elsener, B.: Effect of cracks
on chloride-induced corrosion of steel in concrete-a review: Etatsprogrammet Varige
Konstruksjoner 2012–2015. ETH Zurich (2017)
2. Jansson, A., Flansbjer, M., Löfgren, I., Lundgren, K., Gylltoft, K.: Experimental investigation
of surface crack initiation, propagation and tension stiffening in self-compacting steel–fibre-
reinforced concrete. Mater. Struct. 45(8), 1127–1143 (2012)
reinforced concrete: influence of crack width, fibre type and loading conditions. Corros. Sci.
98, 128–139 (2015)
4. Berrocal, C.G., Löfgren, I., Lundgren, K.: The effect of fibres on steel bar corrosion and
flexural behaviour of corroded RC beams. Eng. Struct. 163, 409–425 (2018)
5. ISO, EN 6892-1, Metafile Metallic materials-Tensile testing-Part 1 (2016)
6. Suzuki, K., Ohno, Y., Praparntanatorn, S., Tamura, H. : Mechanism of steel corrosion in
cracked concrete. In: The Third International Symposium on Corrosion of Reinforcement in
Concrete Construction, Warwickshire (1990)
7. Schießl, P., Raupach, M.: Laboratory studies and calculations on the influence of crack width
on chloride-induced corrosion of steel in concrete. ACI Mater. J. 94(1), 56–61 (1997)
8. Arya, C., Ofori-Darko, F.: Influence of crack frequency on reinforcement corrosion in
concrete. Cem. Concr. Res. 26(3), 345–353 (1996)
Evaluation of the Self-healing Capability
of Ultra-High-Performance Fiber-Reinforced
Concrete with Nano-Particles and Crystalline
Admixtures by Means of Permeability
Hesam Doostkami(&), Marta Roig-Flores, Alberto Negrini,

Eduardo J. Mezquida-Alcaraz, and Pedro Serna
Instituto de Ciencia y Tecnología del Hormigón,

Universitat Politècnica de València, Valencia, Spain
hedoo@doctor.upv.es
Abstract. Self-healing is the capability of a material to repair its damage

autonomously. Ultra-High-Performance Fiber Reinforced Concrete (UHPFRC)
has potentially higher self-healing properties than conventional concrete because
of its lower water/binder content and controlled microcracking due to the high
fiber content. This work uses a novel methodology based on the permeability to
evaluate autogenous self-healing of UHPFRC and enhanced self-healing,
incorporating several additions. To this purpose, one UHPFRC was selected and
modified to include alumina nanofibers in 0.25% by the cement weight,
nanocellulose (nanocrystals and nanofibers), in a dosage of 0.15% by the cement
weight, and 0.8–1.6% of a crystalline admixture. The results obtained show that
the methodology proposed allows the evaluation of the self-healing capability of
different families of concrete mixes that suffered a similar level of damage using
permeability tests adapted to the specific properties of UHPFRC.
Keywords: Ultra High-Performance Fiber Reinforced Concrete Self-healing

Crystalline admixtures Nanocellulose Alumina nanofibers
1 Introduction
The presence of small cracks is unavoidable in reinforced concrete due to the brittleness
of its cementitious matrix. Cracks can be produced by several causes such as mechanical
loads, drying shrinkage, external loads, or thermal actions as freezing and thawing
cycles [1]. Generally, small cracks are not relevant from the structural point of view but
can reduce the durability of the structure through the entrance of water or aggressive
agents, such as chlorides. This problem can be especially relevant in structures with
expensive maintenance or repair, such as marine off-shore structures [2].
Concrete can heal its cracks autogenously by continuing hydration and carbonation
of its particles [3]. For these reactions, the presence of water is necessary to hydrate the
un-hydrated cement particles and to produce the precipitation of carbonate calcium
crystals as a result of the reaction between calcium and carbonate ions and carbon
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_45
490 H. Doostkami et al.
dioxide [4]. Whereas cracks under 300 µm are considered not to affect the mechanical
properties of reinforced concrete and are accepted in codes such as CEB_FIP code,
1990 Eurocode 2, 1992 or BS 8110-1, 1997, they can decrease the durability and
service life of concrete [4]. Additionally, some structures will require to be watertight,
and the presence of small cracks should be completely minimized. It has been reported
that the shape, tortuosity, and direction of cracks affect the permeability of concrete [5].
While crack openings until 50 µm have a negligible effect on water permeability,
cracks around 200 µm increase water permeability [6].
In addition to the entrance of water, the penetration of chlorides is also of interest
due to their role in corrosion of steel reinforcement. It has been reported [7] that small
cracks of sizes until 60 µm can fully heal in terms of resistance to chloride diffusion,
while for wider crack widths, the resistance produced by healing will only be partial.
The age of the healed crack was reported [8] as relevant for cracks <60 µm, in terms of
protection to chloride diffusion, while for larger cracks, it was not critical.
Ultra High-Performance Fiber Reinforced Concrete (UHPFRC) is a recently
developed concrete with high durability performance [9]. UHPFRC allows increasing
deformability and reduction of brittleness due to the improvement in the tensile
properties and achievement of post-cracking behavior. This behavior is obtained thanks
to the presence of fibers, optimizing the size distribution of its particles, reducing
water/binder ratio, and including silica fume. UHPFRC has a high compacity, low
permeability and capillarity absorption, and better control on crack opening [10].
Additionally, low water/binder ratios in the range of UHPC-UHPFRC can improve
autogenous healing capacities due to the high percentage of un-hydrated cement par-
ticles [11]. Studies performed on High-Performance Concrete (HPC) and High-
Performance Fiber-Reinforced Concrete (HPFRC) [12], analyzed the efficiency of
autogenous healing due to the presence of fibers under monotonic loading. Their results
showed that healed specimens with fibers had a lower water permeability and higher
healing ratios compared to healed samples without fibers.
One of the most common additives used to promote self-healing of concrete is
Crystalline Admixtures (CA). Some authors compared the self-healing efficiency
samples in several conditions, [13] studying the effect of CA combined or not with
calcium sulfoaluminate expansive additive. Their results showed better crack closure
for samples submerged in water and measured higher content of Ca2+ and higher pH in
those samples with CA, which favors calcium carbonate precipitation. These results are
consistent with those obtained by [14, 15], which showed the critical effect of water
presence for the reactions, obtaining healing ratios up to 95–98% in water immersion
conditions, measured through improvements in water tightness, even though the
improvement compared with their reference specimens was only scarce. Other authors
[16] reported that the presence of CA could speed up healing and are highly effective in
the restoration of stiffness and load-bearing capacity. The effect of CA on the self-
healing of high-performance fiber-reinforced concrete is potentially more critical
compared to normal strength concrete due to their higher binder content and lower w/b
ration, which favors delayed hydration reactions [17]. Some studies focusing on high
salinity conditions [18, 19], showed that CA promoted better strength recovery, higher
crack closure and a positive effect on reducing chloride penetration, even in air-
exposure [19].
Evaluation of the Self-healing Capability of UHPFRC 491
In addition to CA, other materials may improve concrete healing properties. This
work studies the self-healing capability of a UHPFRC upgraded to incorporate CA,
alumina nanofibers (ANF), and nanocellulose fibers and crystals (CNF, CNC respec-
tively), using water permeability tests adapted to the properties of UHPFRC, based on
the penetration of chlorides through water inside the cracks.
2 Materials and Methodology

2.1 Materials
An Ultra High-Performance Fiber Reinforcement Concrete (UHPFRC, labeled as C3)
was prepared as the reference concrete. This concrete mix was modified to include
crystalline admixtures and nanomaterials to study the effect of these materials on the
self-healing properties.
The materials used in this study include CEM I 42.5 R-SR5 from LAFARGE,
undensified silica fume from Elkem as part of the binder. The aggregates used are fine
silica sand (0/0.5 mm), medium-size silica sand (0/1.6 mm), and silica flour Quarzfin
U-S 500 from Sibelco. The mix incorporated short straight-shaped steel fibres 13/0.2
(length 13 mm, U = 0.2 mm, lf/df = 13/0.2 = 65, tensile strength > 2000 MPa), to
guarantee ductility and multiple micro-cracking. Superplasticizer Sika Viscocrete
20HE was used to obtain the self-compacting behavior of this concrete. The crystalline
admixture was provided by PENETRON. Nanocellulose was provided by API; the
materials used are cellulose nanocrystals (CNC) from API (U = 4–5 nm, L 50–
500 nm) and cellulose nanofibers (CNF) (U = 3–4 nm L 400–700 nm). Alumina nano-
fibres (ANF) were obtained from NAFEN (U = 4–11 nm, L 100–900 nm). Table 1
shows the seven concrete mixes, their composition and code used for this work.
From each mix, two couples of beams of size 100 150 750 mm3 were pre-
pared to evaluate the self-healing properties. Additionally, four cubes sized
100 100 100 mm3, and two prisms of size 100 100 500 mm3 were prepared
to control the variation of properties between groups through compressive strength and
flexural strength tests, respectively. All specimens were demolded 24 h after mixing
and were stored in a humidity chamber at 20 °C and 95% relative humidity.
Compressive strength was tested on cubes at the age of 28 days following EN
12390-3, and the results are included in Table 1. The results show a slight decrease in
the compressive strength in the mixes incorporating the self-healing admixture and the
nanomaterials. In this latter case, this decrease can be caused by a slight increase in
water content (3–12.5%) to maintain the self-compacting properties of the mix. Despite
the higher additional water introduced in those mixes with nanocellulose (C34 to C36),
compressive strength in these mixes is not reduced further.
Additionally, the two prismatic 100 100 500 mm3 specimens were tested in
four-point bending, and the tensile performance of each mix was evaluated following a
simplified inverse analysis method [20, 21]. Table 1 shows the values obtained through
this method, where the mixes with Penetron and ANF displayed a similar response
comparable to the reference mix.
Table 1. Mix designs of the seven UHPFRCs prepared.

3
(kg/m ) C3 C31 C32 C33 C34 C35 C36
Cement I 42.5 R-SR 800 800 800 800 800 800 800
Silica fume 175 175 175 175 175 175 175
Water 160 160 165 165 180 180 180
w/c 0.200 0.200 0.206 0.206
0.225 0.225 0.225
w/b 0.164 0.164 0.169 0.169
0.185 0.185 0.185
Silica sand – 0/1.6 mm 565 565 565 565 565 565 565
Silica sand – 0/0.5 mm 302 302 302 302 302 302 302
Silica flour – <50 m 225 225 225 225 225 225 225
Short steel fibres (13/0.2) 160 160 160 160 160 160 160
Superplasticizer 30 30 30 30 30 30 30
CA – 7.8 7.8 15.6– 7.8 7.8
ANF (solid content) – – 2.44 2.44– – –
CNC (solid content) – – – – 0.73 1.46 0.73
CNF (solid content) 0.73 – 0.73
fcm,28d (MPa) 139.94 119.39 130.33 120.42 114.23 124.16 119.39
ft (MPa) 9.49 7.61 9.41 9.59 7.32 7.25 7.39
ftu (MPa) 6.83 6.28 6.80 9.90 6.3 5.26 5.07
etu (‰) 2.36 1.81 3.53 6.25 2.71 3.00 2.79
2.2 Self-healing Methodology

2.2.1 Pre-cracking Methodology
To evaluate self-healing, 100 150 750 mm3 reinforced beams (Fig. 1) were
prepared. The reinforcement embedded in the beams consisted of 6ø6 stirrups and 2ø8
lower reinforcing bars. Four-points bending test was performed with these beams
following a setup with a central span of 600 mm for the bottom supports and 150 mm
of the distance between the two loading points. Two rows of DEMEC points were
glued at the middle span in order to control the strain variation at these two levels using
a before, during, and after every loading-unloading cycle. Complementary, an Linear
Variable Differential Transformer (LVDT) was placed at the bottom of the beam to
register the load-displacement evolution. Additional information about the test per-
formed can be read in a previous publication [22].
Fig. 1. Reinforced Beam model and DEMEC point locations, sizes are in millimeters.
These beams were pre-cracked after 28 days of age, utilizing a four-point bending
test. For this study, pre-cracking tests were divided into two levels of damage, named
Low and High Strain levels:
– Low Strain (LS): strain on the lower DEMEC line of 0.5‰ during loading;
– High Strain (HS): residual strain between 1 and 2‰ after unloading, measured at
the lower DEMEC row, a condition in which the rebars start to yield, beyond the
service limit state.
Whereas for those beams damaged at the LS, the strain value was very controllable
and the desired value was easy to reach, in those beams targeted to suffer high strain
values, it was challenging to achieve the target residual strain, since the beam is
working near its service limit, and thus, a small increase in load will produce enormous
changes in the residual strain. Figure 2 shows an example of the load-displacement
graphs obtained for the LS and HS levels.
14.00
12.00
10.00
Load (Ton)
8.00
6.00 HS
4.00 LS
2.00
0.00
0.00 0.50 1.00 1.50 2.00
Displacement (mm)
Fig. 2. Load Displacement curves obtained for concrete type C3 during the pre-cracking stage
of the beams for LS and HS damage levels.
2.2.2 Preparation of the Healing Samples

After pre-cracking the beams, all the prisms were sawn in four smaller prisms
15 15 5 cm3 as indicated in Fig. 3, since cracks are produced in the middle span
of the beam. These samples were named A, B, C, D, and were assumed to have the
same deformation due to their location in the beam (Fig. 1). Since the self-healing test
performed in this work is destructive, samples “A” were selected to evaluate the
properties before self-healing and samples “B” after self-healing. Samples named C
and D were used as extra samples to verify results when necessary.
Fig. 3. Diagram of the beam indicating the sawing lines to obtain the self-healing samples (left)
and tube location for the water permeability test based on chloride penetration (right).
To promote self-healing reactions, complete water immersion in deionized water

for 28 days was selected as the healing condition. Deionized water was selected to
avoid any influence of the chlorides and calcium content of local tap water in the test.
2.2.3 Crack Analysis

The crack size was measured in the sawed samples by using a PCE-MM200 micro-
scope for the LS group and using a concrete crack-meter for the HS group with visible
cracks. This quantification was performed at the two DEMEC lines. In this way, the
crack properties (number, size) can be measured for the two lines and compared with
the strain level obtained during (or after) the pre-cracking test.
It needs to be mentioned that, during these measurements, it was detected that some
small cracks (<0.05 mm) were healed partially or even entirely before starting the
healing conditions (Fig. 4).
2.2.4 Modified Water Permeability Test Through Chlorides Penetration

As detected in previous works [22], the water permeability of UHPFRC samples in
cracked conditions was harder to evaluate since they were able to resist the high-
pressure water permeability test, and water was not passing through their cracks.
Because of that, a modified water permeability test was proposed, based on the use of
sodium chloride and its reactivity to silver nitrate as an indicator to show the areas
where water was able to penetrate.
This method consisted of using a PVC tube (øe = 75 mm, length around 60 cm)
glued to the specimens with a resin (in this case, Sikaflex 11 FC). This tube was filled
Fig. 4. Partially healed cracks in beams that suffered low and high strain levels (left and right,
respectively).
with salt water with a concentration of 35 g NaCl/liter, and the salted water column
was left for 3 days over the samples. The water level was controlled every working day,
so it remained around 50 cm, and in case of visible evaporation, tap water was added to
maintain the same pressure level. After three days, the tube was removed, and the
specimen was cut transversely to the direction of the cracks with a circular saw for
concrete following the diagram in Fig. 3 right.
In order to measure the chloride penetration depth, first, specimens were left to dry,
and afterwards, a silver nitrate (AgNO3) solution with a concentration of 0.1 mol/l was
sprayed on both cracked surfaces where the cut was made. After the chemical reaction,
the area colored in white is the area of interest, where silver ions reacted with the
chloride ions, indicating the penetration.
The sections obtained were photographed using a digital camera to evaluate the
differences between samples that healed during 28 days in deionized water and
unhealed samples. These images were analyzed with photography software (Adobe
Photoshop), binarized to black and white pixels (Fig. 5), and the penetration of chlo-
rides in the section was compared with the penetration in each section. With this
method, cracks are detected, even if they are in the range of small microns. It has also
been detected that, in some samples, the penetration of chlorides occurred not only
through the crack but also through its porosity in the uncracked areas, P0 in Fig. 5.
Generally, this penetration due to porosity is extremely low for UHPFRC, and thus,
when evaluating the affected areas, most of the influence will come from the cracked
areas.
3.1 The Relation Between Strain and Crack Width

The average crack width obtained for each beam (xavg) was compared with the residual
strain of the beam obtained at the upper and the lower DEMEC lines (eavg) from the
pre-cracking test.
Fig. 5. Result of image analysis performed on a section after the water permeability test based
on chloride penetration.
Figure 6 shows the linear relation obtained between the residual strain and the
average crack width, as measured in all the beams in the low and top DEMEC lines. As
expected, those values obtained at the top DEMEC line were smaller in terms of crack
width and terms of residual strain. It can be seen that strains lower than 0.5‰ produced
cracks near 10 µm, and when increasing the strain, the crack size increased noticeably.
The crack size in the HS is dispersed, while in the LS level, the crack size range
remains limited. No significant differences were detected at this pre-cracking stage in
the formation and size of the cracks depending on the presence of crystalline admix-
tures and nanomaterials.
3.00
2.50
Residual Strain (‰ )
2.00
1.50
1.00
0.50
0.00
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00
Average Crack (μm)
HS (Top line) HS (Bottom line) LS (Top line) LS (Bottom line)
Fig. 6. Comparison of the residual strain and the average crack width obtained in all the beams.
3.2 Chloride Penetration

All the prisms after the cut were photographed before and after healing, and each photo
was analyzed to evaluate the penetration of chlorides. Figure 7 shows four examples of
the pictures obtained for the mix C3 after suffering the low or high strain levels, before
and after healing.
C3 Before healing After healing
Low
Strain
High
Strain
Fig. 7. Comparison between chloride penetration in C3 samples that suffered low strain (top
row) and high strain (bottom row) before and after healing.
Comparing both, the photos and the binarized pictures, in general, the healed
samples presented lower penetration than the samples before healing. Additionally, the
high strain specimens had larger and much clearer penetration areas than low strain
specimens. However, several difficulties have been detected to evaluate the penetration
in low strain specimens, and thus, the accuracy of this method should be discussed and
improved. Furthermore, another difficulty of this method is that some samples can be
contaminated during the wet-cut, depending on the human factor.
Despite these difficulties to quantify the penetration, the penetration results
obtained in this work can be discussed for each case. For instance, control samples
without self-healing admixtures (C3) partially healed at both low and high strain levels.
This decrease of chloride penetration indicates that inside the cracked sample, chlorides
are not able to penetrate deeper inside the matrix or this small penetration is not
detectable with the methodology used in this work. This result is an evidence of the
protection to the penetration of chlorides that autogenous healing is able to provide to
UHPFRC structures.
Similarly, for C31 and C32, containing Penetron CA and CA plus ANF, the depth
of penetration was reduced after healing when compared to the results before healing.
In this case, except for one sample with higher crack width (50 µm), which experi-
enced only partial healing, the cracks were almost entirely healed and no penetration
was detected after healing. However, in C33 samples, which also had CA plus ANF but
with higher contents of CA, healing did not occur to the same extent, probably due to
an excessive dosage of CA, obtaining larger depths of penetration.
In the case of mixes containing nanocellulose, C34 to C36, the results indicate that
for C34 and C35 samples, for the small strain limit, self-healing was able to stop
chloride penetration completely. However, for the high strain level, healing did not
occur, but the depth of penetration was partially healed. In the case of C36, whereas the
photos display potential healing, chloride penetration still occurred significantly in

depth for both low and high strain limits.
The obtained qualitative results are consistant with those reported in the literature,
showing the potential protection from chlorides penetration that autogenous and
enhanced self-healing are able to produce in concrete [2, 7, 18]. However, this
methododology needs further improvements to allow the exact quantification of the
areas penetrated by chlorides, and will be the focus of future studies.
4 Conclusions
The conclusions that can be drawn from this study are:

• The pre-cracking method used allows obtaining different crack width ranges for the
low and high strain level (around 10 µm and 25–50 µm respectively). For the low
strain, crack widths increased gradually with the strain, however, when reaching the
high strain levels, the crack size became more dispersed. No significant differences
were detected in the formation and size of the cracks depending on the presence of
crystalline admixtures and nanomaterials.
• The methodology followed allows a qualitative evaluation of self-healing and the
results indicate that: 1) healed samples presented lower penetration than samples
before healing, and 2) high strain specimens had larger and clearer penetration areas
than low strain specimens. However, the accuracy of this methodology needs to be
improved to allow the quantification of the penetration.
• Reference UHPFRC obtained partial healing after 28 days healing under deionized
water, showing that autogenous healing in this concrete type for the size of cracks
studied is able to provide some protection from the penetration of chlorides.
Additionally, upgrading concrete to contain small dosages of CA and ANF
improved the self-healing response, to obtain practically no penetration from the
chlorides, even for those samples with large cracks (25–50 µm). However, the
presence of nanocellulose or an excessive use of CA of have been reported to not
clearly improve, or even reduce the self-healing properties of this concrete.
Acknowledgements. The authors would like to acknowledge the European Union’s Horizon
2020 ReSHEALience project (Grant Agreement No. 760824).
References
1. Homma, D., Mihashi, H., Nishiwaki, T.: Self-healing capability of fibre reinforced
cementitious composites. J. Adv. Concr. Technol. 7(2), 217–228 (2009)
2. Maes, M., Snoeck, D., De Belie, N.: Chloride penetration in cracked mortar and the
influence of autogenous crack healing. Constr. Build. Mater. 115, 114–124 (2016)
3. Edvardsen, C.: Water Permeability and Autogenous Healing of Cracks in Concrete, vol.
96 (1999)
4. De Belie, N., et al.: A review of self-healing concrete for damage management of structures.
Adv. Mater. Interfaces 5(17) (2018)
5. Wang, H.L., Dai, J.G., Sun, X.Y., Zhang, X.L.: Characteristics of concrete cracks and their
influence on chloride penetration. Constr. Build. Mater. 107, 216–225 (2016)
6. Wang, K., Jansen, D.C., Shah, S.P., Karr, A.F.: Permeability study of cracked concrete.
Cem. Concr. Res. (1997)
7. Šavija, B., Schlangen, E.: Autogeneous healing and chloride ingress in cracked concrete.
Heron 61(1), 15–32 (2016)
8. Ismail, M., Toumi, A., François, R., Gagné, R.: Effect of crack opening on the local diffusion
of chloride in cracked mortar samples. Cem. Concr. Res. 38(8–9), 1106–1111 (2008)
9. Habel, K., Gauvreau, P.: Response of ultra-high performance fiber reinforced concrete
(UHPFRC) to impact and static loading. Cem. Concr. Compos. 30(10), 938–946 (2008)
10. Denarié, E., Brühwiler, E.: Strain-hardening ultra-high performance fibre reinforced
concrete: deformability versus strength optimization. Restor. Build. Monum. 17(6), 397–
410 (2014)
11. Granger, S., Pijaudier-Cabot, G., Loukili, A.: Mechanical behavior of self-healed ultra high
performance concrete: from experimental evidence to modeling. In: Proceedings of the 6th
International Conference on Fracture Mechanics of Concrete and Concrete Structures, vol. 3,
pp. 1827–1834 (2007)
12. Escoffres, P., Desmettre, C., Charron, J.P.: Effect of a crystalline admixture on the self-
healing capability of high-performance fiber reinforced concretes in service conditions.
13. Sisomphon, K., Copuroglu, O., Koenders, E.A.B.: Self-healing of surface cracks in mortars
with expansive additive and crystalline additive. Cem. Concr. Compos. 34(4), 566–574
(2012)
14. Roig-Flores, M., Moscato, S., Serna, P., Ferrara, L.: Self-healing capability of concrete with
crystalline admixtures in different environments. Constr. Build. Mater. 86, 1–11 (2015)
15. Roig-Flores, M., Pirritano, F., Serna, P., Ferrara, L.: Effect of crystalline admixtures on the
self-healing capability of early-age concrete studied by means of permeability and crack
closing tests. Constr. Build. Mater. 114, 447–457 (2016)
16. Ferrara, L., Krelani, V., Carsana, M.: A ‘fracture testing’ based approach to assess crack
healing of concrete with and without crystalline admixtures. Constr. Build. Mater. 68, 535–
551 (2014)
17. Ferrara, L., Krelani, V., Moretti, F.: On the use of crystalline admixtures in cement based
construction materials: from porosity reducers to promoters of self healing. Smart Mater.
Struct. 25(8), 1–17 (2016)
18. Cuenca, E., Cislaghi, G., Puricelli, M., Ferrara, L.: Influence of self-healing stimulated via
crystalline admixtures on chloride penetration. In: America Concrete Institute, vol. 2018(SP
326), pp. 1–10. ACI Spec. Publ. (2018)
19. Borg, R.P., Cuenca, E., Gastaldo Brac, E.M., Ferrara, L.: Crack sealing capacity in chloride-
rich environments of mortars containing different cement substitutes and crystalline
admixtures. J. Sustain. Cem. Mater. 7(3), 141–159 (2018)
21. López, J.Á.: Characterisation of the Tensile Behaviour of UHPFRC By Means of Four-Point
Bending Tests, March 2017
22. Negrini, A., Roig-Flores, M., Mezquida-Alcaraz, E.J., Ferrara, L., Serna, P.: Effect of crack
pattern on the self-healing capability in traditional, HPC and UHPFRC concretes measured
by water and chloride permeability. In: MATEC Web Conference, vol. 289, p. 01006 (2019)
Analytical and Numerical Models
Material Characterisation for Nonlinear Finite
Element Analysis (NLFEA)
P. J. van der Aa(&) and A. A. van den Bos
DIANA FEA BV, Delft, The Netherlands

engineering@dianafea.com
Abstract. The design of steel fibre reinforced concrete (SFRC) structures is

becoming more popular. It is beneficial to combine traditional reinforcement
with SFRC, because the design becomes more robust (more ductile) and the
traditional reinforcement and/or the concrete cross-section can be decreased.
The modelling of the material is often based on a smeared crack approach.
However good guidelines, which describe modelling strategy for SFRC are
lacking. Therefore, this paper will propose a new constitutive model for SFRC
and will go into detail about improving crack localization. Good crack local-
ization is often a problem with smeared crack approach. The modelling strategy
will be applied using the finite element software DIANA [1] and compared with
Keywords: Steel fibre reinforced concrete Smeared crack approach Crack

localization DIANA
1 Introduction
Designing reinforced concrete structures with the addition of steel fibres becomes more
popular. The advantages (e.g. decrease of traditional reinforcement and/or the increase
of the robustness of the structure) and the applications (e.g. industrial floors and
foundations) of steel fibres are known for most structural and civil engineers.
Unfortunately, good rules and guidelines are still lacking. Since steel fibres are only
activated after cracking of the concrete, a nonlinear analysis is obliged to take into
account the positive effect of the steel fibres. Besides performing a nonlinear cross-
section analysis (e.g. by using a multi-layer model) also redistribution of forces and
bending moments should be taken into account. Therefore it is strongly recommended
to perform a nonlinear finite element analysis (NLFEA) to describe the behaviour of
steel fibre reinforced concrete (SFRC), which automatically considers both phenomena.
This paper will provide clear guidelines, on how to perform a NLFEA for a SFRC
structure.
In this paper all NLFEA have been performed with the finite element software
DIANA [1].
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_46
504 P. J. van der Aa and A. A. van den Bos
2 Constitutive Model of Concrete and SFRC
The Eurocode 2 (EC2) [2] and the fib Model Code 2010 (MC2010) [3] do not present a
proper constitutive model for concrete and/or SFRC, which can be used for NLFEA.
The EC2 only presents the basic properties of concrete related to a concrete class, e.g.
Young’s modulus, tensile strength and compressive strength with the inclusion of a
compressive stress-strain relation with fixed strain values. The MC2010 presents some
extra properties such as the fracture energy and a tensile stress-strain relation, which is
energy based. Also for SFRC a tensile stress-strain/CMOD diagram is defined.
Nowadays concrete-like materials are primarily modelled using a smeared crack
approach in NLFEA. The input for a smeared crack model is a stress-strain relation,
which has to be provided by the user. As mentioned before, the MC2010 provides a
stress-strain relation for both the behaviour in compression as well as the behaviour in
tension. However, it is strongly discouraged to use these as input for a smeared crack
model, because this will lead to incorrect results. This will be elaborated in the next
sections.
2.1 Concrete
Figure 1 shows the stress-strain/crack opening relation of the MC2010. In this diagram,
there is a short hardening branch, which starts when 90% of the tensile strength is
reached. This behaviour has been introduced to capture microcracking close to the
macrocrack. This means that this macrocrack describes the full cracked zone and
therefore it can only be used for discrete crack approach and not for smeared crack
approach. Because in a smeared crack approach both microcracking and macrocracking
will be captured. Figure 2 shows the micro- and macrocracks of a concrete beam
subjected to bending. In the graph the stress-strain behaviour of the macrocrack is
plotted (orange line) and the stress-strain behaviour between the macro cracks (blue
dashed line = microcracks), which shows that microcracking is occurring.
Fig. 1. Stress-strain and stress-crack opening for uniaxial tension [3]

Material Characterisation for NLFEA 505
Fig. 2. Micro- and macrocracking in smeared crack approach
Therefore, the short hardening branch (which describes the influence of microc-
racking) should not be included. Moreover, not eliminating this branch will lead to an
overshoot of the tensile capacity, because the area underneath the graph, which rep-
resents the energy, will be overestimated. Therefore, for smeared cracking approach it
is recommended to use an elastic branch up to the tensile strength, followed by a
softening curve such as Hordijk [4], exponential, bilinear or something similar. The
stress-strain curve should be defined by either the fracture energy or the crack opening
and not in terms of (ultimate) strains. The strain in a finite element analysis is always
relative to the element size or band width; therefore predefining the strains in the stress-
strain relation will lead to an element size-dependent behaviour. So for a smeared
cracking approach it is recommended to use a Hordijk softening curve, which is defined
by the tensile strength and the fracture energy. Both the tensile strength and the fracture
energy are provided by the MC2010.
As mentioned before the stress-strain curve to describe the compressive behaviour
has predefined strain values in both the EC2 as well as in the MC2010. As explained
before, predefining the strains in the stress-strain relation will lead to an element size-
dependent behaviour for a smeared crack approach. Therefore, it is recommended to
use a stress-strain curve, which is based on the compressive fracture energy. Figure 3
shows a parabolic compression curve [5], which is depended on the compressive
fracture energy. The compressive fracture energy can be found in literature [6] and is
related to the fracture energy in tension.
Fig. 3. Compressive stress-strain relation [5].

By defining both the tensile behaviour and the compressive behaviour in terms of
energy, the material model becomes independent of the element size of the finite
element model.
2.2 SFRC
The MC2010 does present a stress-CMOD (or stress-strain) relation for the tensile
behaviour; however, it is strongly discouraged to use this relation for NLFEA. Figure 4
depicts the stress-strain diagram as is advised by the MC2010. A similar hardening
branch at 90% of the tensile strength can be seen (from point A to B), which should not
be used for a smeared crack approach as explained before.
Fig. 4. Stress-strain relation MC2010 for the tensile behaviour of SFRC [3]
Point D represents fFts (CMOD = 0.5 mm) and E represents fFtu (CMOD = 2.5 mm).
These two points are derived from Fr1 and Fr3, which follow directly from the 3-point
bending test. However, point C is not derived from the 3-point bending test and it is
assumed to be an extrapolation of D and E. However looking at Fig. 5 this assumption can
easily be questioned.
Below we present a method to determine the stress-CMOD diagram based on
combined tensile behaviour of concrete (which is described by a Hordijk softening
curve) and steel fibres (which is described by a root function). The combined behaviour
is depicted by the red graph in the figure below for a CMOD up to 0.5 mm.
The red graph can be simplified by a few points represented by the black graph of
Fig. 6 and Fig. 7. The first part of the graph is mostly influenced by the concrete beha-
viour. The first point is represented by the tensile strength of concrete fct. This point is not
influenced by the addition of steel fibres. The second point is named fFt0, which describes
the softening of the concrete and the first point of the steel fibre behaviour. This point is
calculated as 0.85 fFts = fFt0. The corresponding CMOD is determined by the Hordijk
softening curve, with an increased fracture energy (3 GF). Since the steel fibres
immediately start working after cracking, it makes sense to increase the fracture energy.
Point fFts (from fr1) and fFtu (from fr3) complete the stress-CMOD diagram, which follow
directly from the MC2010 Section 5.6 Fibres/fibre reinforced concrete.
Fig. 5. Combined behaviour of steel fibres and concrete.
Fig. 6. Simplified stress-CMOD behaviour for SFRC (detail).
3 Crack Localization with Smeared Crack Approach
Crack localization is very important when using the smeared crack approach. Poor
localization could lead to, for example a mesh-dependent finite element model, an
overshoot of the deformation capacity of the structure, an undershoot of the crack width
and/or an incorrect crack pattern.
To show these aspects, some projects from literature are used for illustration pur-
poses. Figure 8 shows the crack pattern at the bottom of a suspended slab on columns.
The full-scale test of Parmentier at Limelette (Belgium) is used as model input [7].
Fig. 7. Simplified stress-CMOD behaviour for SFRC (complete).
From the figure, one can see that some of the cracks are smeared out over 10 elements.
In the next section, a numerical investigation is performed to show and explain the
influence of the element type on the cracking behaviour of SFRC.
Fig. 8. Poor localization of the crack pattern in FEA.
3.1 Numerical Investigation

A simply supported beam with uniform distributed load (UDL) (see Fig. 9) will be
used to investigate the cracking behaviour of SFRC with traditional reinforcement for
different element types. The beam will be modelled in two ways, using shell elements
(which will be loaded in the out-of-plane direction) and solid elements (see Fig. 10).
Fig. 9. Simply supported beam with UDL and bending moment diagram.
Fig. 10. FE-model with shell elements and solids element.
The results of the model with the shell elements are depicted in Fig. 11 and Fig. 12. It
can be clearly seen that the crack width is completely smeared out over the middle
elements. The bending moment and curvature are related to each by the bending stiffness.
From Fig. 12 it can be seen that, at the location of the smeared cracks, the curvature has
significantly increased, which means that the bending stiffness has reduced.
Fig. 11. Crack width in the shell elements.
Shell elements assume that the straight line normal to the undeformed middle plane
remains straight, which means that the strain distribution over the height is always
linear. This means, in order to increase the bending moment, the curvature has to
increase. This can also be seen from Fig. 13, which plots the moment-curvature dia-
gram for 2 nodes. Both nodes have to follow the same moment-curvature diagram.
Fig. 12. Curvature in the shell elements.
Fig. 13. Moment-curvature of shell elements.
The results of the model with the shell elements are depicted in Fig. 14 and Fig. 15.
It can be seen that the cracks are nicely localizing in one integration point. From
Fig. 15 it can be seen that only at the location of the cracks the curvature significantly
has increased.
Fig. 14. Crack width in the solid elements.
For solid elements, there is no restriction in terms of deformation for the straight
line normal to the undeformed middle plane, which means that the strain distribution
over the height of the cross-section does not have to be linear. The consequence for the
moment-curvature relation can be seen in Fig. 16, which plots the moment-curvature
diagram for a cracked node and a node between to cracks. Both follow a different
moment-curvature path (Fig. 17).
Fig. 15. Curvature in the solid elements.
Fig. 16. Moment-curvature of shell elements.
3.2 Conclusion
From the numerical investigation it can be seen that elements are needed, which can
describe a nonlinear strain distribution over the height of the cross-section, in order to
find the correct crack pattern with good localization. For the sake of completeness the
full scale test [7] is simulated also with solids and a final comparison of the crack
pattern is depicted in Fig. 18, showing the better localization with the solid elements.
4 Experimental Verification of NLFEA
To verify the modelling approach for SFRC and NLFEA an experimental test will be
simulated. The international benchmark of Politecnico di Milano (Italy) 2018 is used.
The paper is still not published but the similar twin test is [8]. Tests of SFRC (fiber
only), RC (traditional reinforcement) and Hybrid (SFRC + RC) are performed. To
reduce the size of this paper, only the Hybrid solution is evaluated.
Fig. 17. Strain distribution in the crack (left) and between cracks (right).
Fig. 18. Crack patterns when using shell elements (left) and when using solid elements (right).
Fig. 19. Dimension (in mm) and FE-model.

4.1 Experimental Setup and Finite Element Model

The slab is supported by 4 points and is loaded in the centre depicted in Fig. 19.
Concrete class is C50 and 35 kg/m3 of steel fibres (Dramix® 4D – 65/60 BG) are
added.
The finite element model consists of solid elements. Symmetry has been applied in
one direction. The reinforcement has been modelled as beam elements with bond-slip
behaviour. The constitutive model presented in this paper has been applied. The
material parameters have been derived from the 3-point bending tests.
4.2 Results
Figure 20 shows the comparison of the load-deflection diagram for the experimental
test (Hyb1 and Hyb2) and the FE-simulation. The FE-results show good agreement
with the experiments. In addition, the crack pattern and failure mechanism are in good
agreement with the experimental results (see Fig. 21).
Fig. 20. Load-displacement experiment and FE-simulation
Fig. 21. Crack pattern at the bottom of the slab from experiment and FE-simulation
References
1. DIANA FEA BV. DIANA User’s Manual, Manie, J.: Delft, The Netherlands (2020). https://
dianafea.com/manuals/d103/Diana.html
2. EN 1992-1-1. Eurocode 2: Design of concrete structures – Part 1-1: General rules and rules
for buildings (2005)
3. Fédération internationale du béton. Model Code for Concrete Structures 2010 (MC2010)
(2013)
4. Hordijk, D.A.: Local approach to fatigue of concrete. Ph.D. thesis, Delft University of
Technology (1991)
5. Feenstra, P.H.: Computational aspects of biaxial stress in plain and reinforced concrete. Ph.D.
thesis, Delft University of Technology (1993)
6. Nakamura, H., Higai, T.: Compressive fracture energy and fracture zone length of concrete
(2001)
7. Parmentier, B., Van Itterbeeck, P., Skowron, A.: The flexural behaviour of SFRC flat slabs:
the Limelette full-scale experiments for supporting design model codes. In: Proceedings of the
FRC 2014 Joint ACI-fib International Workshop, ACI-fib, pp. Montréal (2014)
8. Di Prisco, M., Colombo, M., Pourzarabi, A.: Biaxial bending of SFRC slabs: is conventional
reinforcement necessary? Mater. Struct. 52, 1 (2019)
Comparison Between the Cracking Process
of Reinforced Concrete and Fibres Reinforced
Concrete Railway Tracks by Using Non-linear
Finite Elements Analysis
Jean-Louis Tailhan, Pierre Rossi(&), and Thierry Sedran
MAST-EMGCU, Univ Gustave Eiffel, IFSTTAR, 77447 Marne-la-Vallée,

France
pierre.rossi@ifsttar.fr
Abstract. This study concerns the use of numerical models to analyze and
compare the cracking process of two types of Railway Tracks: one in reinforced
concrete and the second one in FRC. The models used are the Probabilistic
Explicit Cracking models developed by IFSTTAR and fully validated in the
framework of previous works.
The main result obtained can be summarized as following:
• When in a given concrete structure, the tensile stresses are localized, the
traditional rebars are more efficient to control cracks than the presence of
fibres.
• On the contrary, when the tensile stresses are diffused, that is the case in
statically indeterminate mechanical situations, the fibres are significantly
more efficient than traditional rebars in regards to this cracks control. The
parts of the railway track slabs which are located between the zones
delimited by the joints in the foundation slab are concerned by this
conclusion.
Keywords: Fibre reinforced concrete Railway tracks Numerical models

Crack control
1 Introduction
Steel fiber reinforced concrete (SFRC) is increasingly used in structural applications.

One of the principal reasons of its gain in popularity is due to the existence of national
and international recommendations for the design of structures using this type of
material. These recommendations are efficient for designing simply supported struc-
tural elements subjected to bending. However, they do not possess a sufficient physical
base to propose relevant solutions for more complex structures such as statically
indeterminate. Hence, it is claimed that the most efficient approach for designing such
structures with respect to both safety and sustainable development is to use non-linear
finite element analysis.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_47
516 J.-L. Tailhan et al.
IFSTTAR has been developing, since 1985, a probabilistic discrete cracking model
to simulate the cracking process of concrete. Then, this model was extended to analyze
reinforced concrete and fibre reinforced concrete structures. All these numerical models
have been, today, fully and deeply validated [1–8].
The present work concerns the use of these numerical models to analyze the
cracking process of two types of Railway Tracks: one in reinforced concrete and the
second one in fibres reinforced concrete. In fact, this comparative analysis is performed
on a large mockup designed to represent, reasonably well, the reality.
2 Description of the IFSTTAR Railway Track Mockup
The mean objective behind the design of the IFSTTAR Railway Track Mockup is to
get a representative structural situation and to perform mechanical tests on it. Figure 1
presents the selected mockup geometry.
In this Fig. 1, BC5 represents the track slab, BC3 the foundation slab and the
Sylomer layer is to take into account the mechanical reaction of the ground.
Fig. 1. Geometry of the selected mockup.
Figure 2 presents the mechanical loading applied to the Railway Track Mockup.
Fig. 2. Loading and boundary conditions.

Comparison Between the Cracking Process of Reinforced Concrete 517
It is important to note that added to this external mechanical loading, the dead load
generated by the rails, the soles and the saddles as well as thermal loadings (positive
and negative gradients) are considered in all numerical simulations of this study.
3 Finite Elements Meshes
The objective of the work is to perform a comparison between the cracking process of,
both, Reinforced Concrete Railway Track (RCRT) and Fibres Reinforced Concrete
Railway Track (FRCRT) by using non-linear finite elements analysis. Figures 3 and 4
present the finite elements meshes related respectively to RCRT and FRCRT solutions.
Fig. 3. 2D Mesh of the RCRT solution.
Fig. 4. 2D Mesh of the FRCRT solution.
Concerning the RCRT, the total section of rebars at the top and at the bottom of the
track slab are, respectively, 1.53 10−3 m2 (7 rebars of diameter 14 mm + 4 rebars of
diameter 12 mm) and 3.38 10−3 m2 (14 rebars of diameter 16 mm + 5 rebars of
diameter 12 mm).
As observed on these figures, the meshes are two-dimensional (2D). Indeed,
although the mechanical problem is a three-dimensional (3D) one, it is unreasonable to
perform non-linear 3D simulations which should consume too much computational
time with the models used. The numerical simulations are then performed under 2D
conditions. Plane stresses conditions are chosen rather than plane strains conditions.
Indeed, the models used in this study which take into account volume effects (Sect. 4),
the width (i.e. the value of length measured in the direction perpendicular to the plane
of the figure) of the mockup has to be considered (to calculate the volume of the mesh
elements).
If the total width of the mockup was considered, the finite element volumes
obtained would not be enough small to get a fine and acceptable description of the
cracking processes (see Sect. 4 related to the volume effects taken into account by the
model). So, a more detailed description of the cracking processes could be achieved by
choosing smaller value of this width. This strategy is relevant only if 2D and 3D elastic
simulations lead to comparable information about the longitudinal distribution of the
elastic tensile stresses at the bottom of the track slab (where cracks will be created).
The width of the cast iron plates (whose the thickness is equal to 14 mm), under the
rail, seems to be the smaller one it is reasonable to consider in the present 2D numerical
analysis. This width is equal to 0.395 m (Fig. 1). This choice leads to underestimate the
real stresses diffusion under the rails and, so, to underestimate the global stiffness of the
mechanical system (all the mockup). To solve this problem, one possible solution is to
increase, artificially, the Young modulus of the materials constituting the mechanical
system, especially for the track and foundation slabs and the elastic layer (where the
stresses are diffused). To achieve this objective, the following steps of numerical
simulations have been performed:
• First step: 3D elastic simulation of the full mockup with the real stiffness of the
materials. Figure 5 presents the 3D mesh used for this numerical analysis.
• Second step: 2D elastic simulation (plane stresses conditions) with the full width of
the track slab mockup and with the real stiffness of the materials.
• Third step: 2D elastic simulations with smaller widths of the mockup and different
values of stiffness related to the different materials.
Remark: In all these numerical simulations (3D and 2D), the rebars are not
considered.
Fig. 5. 3D Mesh of the mockup.
This search (fitting approach) of the materials stiffness related to the 2D simulation
leads to the values summarized in Table 1. It is important to point out that this set of
Young modulus values does not correspond, of course, to a unique solution.
Table 1. Values of the materials Young modulus used in the elastic numerical simulations.
Young modulus (MPa)
3D Simulation 2D Simulation full wide 2D Simulation reduced wide
Rails 210000 210000 210000
Pad 7.702 7.702 7.702
Iron saddles 17200 17200 17200
Track slab 35000 35000 43750
Foundation slab 24000 24000 32500
Sylomer 1.2 1.2 1.625
Table 2 presents the values of the others materials properties used in the framework
of these elastic numerical simulations. They are independent of the type of linear
numerical simulation performed.
Table 2. Values of the others materials properties used in the elastic numerical simulations.
Poisson coefficient Density
Rails 0.3 7.8
Soles 0.4 1.2
Iron saddles 0.3 7.8
Track slab 0.25 2.5
Foundation slab 0.25 –
Sylomer 0.32 –
In the light of this linear elastic study, performing non-linear 2D finite elements
simulations on the proposed reduced width of the mockup (and with the modified
Young modulus values) can be considered as acceptable.
4 Numerical Models
4.1 Probabilistic Explicit Cracking Model of Concrete

The model was first developed at IFSTTAR (formerly LCPC) by Rossi [9–11] and
recently improved by Tailhan et al. [12]. It describes the behaviour of concrete via its
two major characteristics: heterogeneity, and sensitivity to scale effects [13]. The
physical basis of the model can be summarized as follow:
• The heterogeneity of concrete is due to its composition. The local mechanical
characteristics (tensile strength ft, shear strength sc) are randomly distributed.
• The scale effects are a consequence of the heterogeneity of the material. The
mechanical response directly depends on the volume of material that is stressed.
• The cracking process is controlled by defects in the cement paste, by the hetero-
geneity of the material, and by the development of tensile stress gradients.
The following points specify how the numerical model accounts for these physical
evidences:
• The model is developed in the framework of the finite element method, each ele-
ment representing a given volume of heterogeneous material.
• The tensile strength is distributed randomly over all elements of the mesh using a
Weibull distribution function whose characteristics depend on the ratio: volume of
the finite element/volume of the largest aggregate, and the compressive strength (as
a good indicator of the quality of the cement paste). The volume of the finite
element depends on the mesh, while the volume of the largest aggregate is a
property of the concrete.
• Remark: a Weibull distribution function is the best to take into account the rupture
in tension of a brittle and heterogeneous material as concrete.
• The shear strength is also distributed randomly over all elements using a distribu-
tion function: (1) its mean value is independent of the mesh size and is assumed
equal to the half of the average compressive strength of the concrete and (2) its
deviation depends on the element size, and is the same (for elements of same size)
as that of the tensile strength.
In what follows, only aspects related to 2D modelling are presented.
In 2D simulations, the cracks are explicitly represented by 2D non-linear interface
elements (quadratic elements) of zero thickness. These elements connect volume ele-
ments representing un-cracked plain concrete. Failure criteria of Rankin in tension and
Tresca in shear are used. As far as tensile or shear stresses remain lower than their
critical values, the interface element ensures the continuity of displacements between
the nodes of its two neighboring volume elements. The material cell gathering these
two volume elements and the interface element remains therefore elastic. Once one of
the preceding failure criteria is reached, the interface element opens and an elementary
crack is created. The tensile and shear strengths as well as the normal and tangential
stiffness values, related to this interface element, become equal to zero. In case of crack
re-closure, the interface element recovers its normal stiffness and follows a classical
Coulomb’s law.
Note that in this modelling approach, the creation and the propagation of a crack is
the result of the creation of elementary failure planes that randomly appear and can
coalesce to form the macroscopic cracks (Fig. 6).
4.2 Probabilistic Explicit Cracking Model of Fibre Reinforced Concrete

(FRC)
The creation of cracks in the concrete matrix is represented by an elastic perfectly
brittle behaviour, whereas the bridging effect of the fibres is described by the following
modelling approach.
Fig. 6. Schematization of the basic principles of the probabilistic model.
Normal and tangential stresses in the interface element linearly increase with normal
and tangential displacements when a “broken” interface element re-opens to take into
account the elastic bridging effect of the fibres inside the crack. Physically speaking, the
rigidity of the fibres (inside the cracks) is more important in tension than in shear. Thus,
the interface element rigidity is considered different for normal and tangential dis-
placements. In 2D, normal and tangential rigidities of the interface element are Kn’ and
Kt’ respectively. The post-cracking elastic behaviour exists until it reaches a threshold
value, f0, related to the normal displacement (Fig. 7). The mechanical behaviour of the
interface element changes once this threshold value is reached. The normal stress is
considered as linearly decreasing with the normal displacement in order to take into
account the damage of the bond between the concrete and the fibre, and fibre pullout.
The decreasing evolution is obtained by using a damage model.
Fig. 7. Principles of the numerical (probabilistic) mechanical model.
Finally, the interface element is considered definitively broken when the normal
displacement reaches a threshold value, fc (Fig. 7). This value corresponds to the state
where the effect of fibres is considered negligible. It is determined from a uniaxial

tensile test. At this point, its normal and tangential rigidities are set to zero.
The post-cracking energy dissipated by the bridging effect of the fibres is consid-
ered randomly distributed over the mesh elements. The random distribution chosen is a
log-normal distribution function with a mean value independent of the mesh elements
size [14] and a standard deviation, due to the heterogeneity of the material, increasing
as the mesh elements size decreases. The choice of a log-normal distribution function is
an arbitrary one. It is convenient to avoid having negative values of the post-cracking
energy when the element meshes are very small. To model a given structural element,
the distribution function is determined in the following manner:
• The mean value is directly obtained experimentally from a certain number of
uniaxial tensile tests on notched specimens, more specifically from the load-crack
opening experimental curves (it is the best way of determination). If the uniaxial
tensile tests have not been performed, that mean value can be determined also by
analysing bending test results using an inverse approach.
• The standard deviation, which depends on the mesh elements size, is determined by
an inverse analysis approach that consists of simulating the uniaxial tests with
different element mesh sizes. As the mean value of the post-cracking energy is
known from the experimental results, several numerical simulations are realized for
each mesh size to determine the standard deviation that best fits the experimental
results. The inverse analysis approach thus allows finding a relation between the
standard deviation and the finite element mesh size. As for the mean value, the
standard deviation can be determined by analysing bending test results using an
inverse approach.
The threshold parameters f0 and fc are determined by an inverse analysis approach
to best fit the simplified triangular stress-displacement curve representing the post-
cracking energy (Fig. 7) to the experimental tensile softening curve.
4.3 Rebars and Concrete/Steel Bond Modelling

The rebars are considered, in the 2D modelling approach, as equivalent plates mixing
concrete and steel rebars. The reinforcement ratio at the top of the track slab being
different of the one located at the bottom, two equivalent plates are considered in the
simulations. The heights of these plates are taken equal to the greater diameter of the
local rebars. The Young moduli of the plates are determined by using a classical
mixture rule (see Eq. 1).
E ¼ ðEa Aa þ Eb Ab Þ=ðAa þ Ab Þ ¼ ðAb =Aa þ Ab Þ½Eb þ ta Ea ð1Þ
Where:
Ea and Eb are respectively the steel and the concrete Young moduli;
Aa and Ab are respectively the steel and the concrete sections;
ta is a local reinforcement ratio (for the considered plate).
Table 3 presents the values related respectively to the height, the width, the Young
modulus and the Poisson coefficient of the equivalent plates in the simulations.
Table 3. Values related respectively to the high, the wide, the Young modulus and the Poisson
coefficient of the equivalent plates in the simulations.
High (m) Width (m) ta (%) E (MPa) Poisson coefficient
Top plate 0.014 0.395 3.81 41646 0.25
Bottom plate 0.016 0.395 8.81 50506 0.25
Numerically speaking, the equivalent plates are composed of one layer of volume
elements. Here again, the volume elements are all interfaced by interface elements.
These interface elements allow the crack crossing through the equivalent plates. They
are exactly the same as those used to simulate concrete cracking and follow the same
opening criteria (see Sect. 4.1). In this case, the volume of concrete considered for each
opening of interface element (see Sect. 4.1) is that of the cumulated volume of elements
surrounding it (it is, of course, an approximation).
After the opening of interface element, its residual stiffness is calculated consid-
ering these two following assumptions:
• The interface element section considered is that of the equivalent plate crossed by
the crack.
• A total rupture of adherence along the two volume elements concerned by the
interface element is assumed.
5 Non-linear Numerical Simulations
The parameters values (in relation with the size of finite element meshes, see Sect. 3)
used in the non-linear simulations are given in Table 4.
Table 4. Parameters values used for cracking of concrete and FRC.

Parameter Value
RT (Tensile strength) Mean value 4.87 [MPa]
Standard deviation 0.67 [MPa]
W (Post-cracking energy) Mean value 4.26 10−3 [MPa.mm]
Standard deviation 2.6 10−3 [MPa.mm]
f0 50 [µm]
fc 4 [mm]
The FRC considered in this study is taken from the literature and successfully used
in a previous numerical study [5]. This FRC contains 78 kg/m3 of steel fibres. These
fibres are hooked end ones.
In this paper are presented results related only to positive temperature gradient.
Indeed, negative temperature gradient has led to very less and smaller cracks compared
with those related to positive temperature gradient. Three simulations for each type of
reinforcement solution (RC and FRC) are performed.
In Fig. 8 is presented one example of cracking pattern of the FRC track slab.
Fig. 8. Detailed information of the cracking pattern of one of the three simulations of the FRC
track slab.
In Fig. 9 is presented one example of cracking patterns of the RC track slab.
Fig. 9. Detailed information of the cracking pattern of one of the three simulations of the RC
track slab.
Considering Figs. 8 and 9, the following comments can be made:

• When the tensile stresses in the track slab are localized (isostatic situation), like at
the vertical of the joints of the foundation slab, only one crack appears, and the
crack opening is smaller for the reinforced concrete track slab than for the FRC
track slab.
• On the contrary, when the tensile stresses in the track slab are diffused (statically
indeterminate situation), like between two joints of the foundation slabs, the cracks
openings are clearly smaller in the case of FRC track.
6 Conclusions and Perspectives
This study concerns the use of numerical models to analyze and compare the cracking
process of two types of Railway Tracks: one in reinforced concrete and the second one
in FRC. The models used are Probabilistic Explicit Cracking ones developed by
IFSTTAR and fully validated in the framework of previous works.
The main result obtained can be summarized as following:
• When in a given concrete structure, the tensile stresses are localized, the traditional
rebars are more efficient to control cracks than the presence of fibres. For that, it is
necessary, of course, to get an easy and optimal placement of the rebars in relation
with cracks orientation. That is the case in railway tracks.
• On the contrary, when the tensile stresses are diffused, that is the case in statically
indeterminate mechanical situations, the fibres are significantly more efficient than
traditional rebars in regards to this cracks control. The parts of the railway track
slabs which are located between the zones delimited by the joints in the foundation
slab are concerned by this conclusion.
For the practice, it can be argued that, for railway track slabs; the use of 78 kg/m3
de fibres is mechanically more efficient than the usual ratio of rebars. On the other
hand, if this technical solution is adopted, it is preferable to use, in zones of high
concentration of tensile stresses, some local rebars (low quantity) to ensure a necessary
level of safety to the structure.
In the future, it would be interesting to confirm the present results, to perform
experimental tests on a mockup with a slab track made of the FRC used in that study, as
it was previously done with the reinforced concrete during original development of
NBT concept.
It should be also interesting, to use the numerical models presented in this work to
propose an optimization of the structural design of the SFR track slab (percentage of
fibers, slab thickness…).
References
1. Phan, T.S., Tailhan, J.-L., Rossi, P.: 3D numerical modelling of concrete structural element
reinforced with ribbed flat steel rebars. Struct. Concr. 14(4), 378–388 (2013)
2. Tailhan, J.L., Rossi, P., Daviau-Desnoyers, D.: Numerical modelling of cracking in steel
fibre reinforced concrete (SFRC) structures. Cement Concr. Compos. 55, 315–321 (2015)
3. Rossi, P., Daviau-Desnoyers, D., Tailhan, J.L.: Analysis of cracking in steel fibre reinforced
concrete (SFRC) structures in bending using probabilistic modelling. Struct. Concr. 16(3),
381–388 (2015)
4. Rastiello, G., et al.: Macroscopic probabilistic cracking approach for the numerical
modelling of fluid leakage in concrete. Ann. Solid Struct. Mech. 7(1–2), 1–16 (2015)
5. Rossi, P., et al., Numerical models for designing steel fiber reinforced concrete structures:
why and which ones? In: Massicotte, B., Charron, J.-P., Plizzari, G., Mobasher, B. (eds.)
FRC 2014: ACI-fib International Workshop, FIB Bulletin 79 – ACI SP, vol. 310, pp. 289–
300 (2016)
6. Rossi, P., Tailhan, J.L.: Numerical modelling of the cracking behaviour of steel fibre
reinforced concrete (SFRC) beam on grade. Struct. Concr. 18(4), 571–576 (2017)
7. Rossi, P., Daviau-Desnoyers, D., Tailhan, J.L.: Probabilistic numerical model of ultra-high
performance fiber reinforced concrete (UHPRFC) cracking process. Cement Concr. Compos.
90, 119–125 (2018)
8. Nader, C., Rossi, P., Tailhan, J.L.: Numerical strategy for developing a probabilistic model
for elements of reinforced concrete. Struct. Concr. 18(6), 883–892 (2017)
9. Rossi, P., Richer, S.: Numerical modelling of concrete cracking based on a stochastic
approach. Mater. Struct. 20, 334–337 (1987)
10. Rossi, P., Wu, X.: Probabilistic model for material behaviour analysis and appraisement of
concrete structures. Mag. Concr. Res. 44(161), 271–280 (1992)
11. Rossi, P., Ulm, F.-J., Hachi, F.: Compressive behaviour of concrete: physical mechanisms
and modeling. J. Eng. Mech. 122(11), 1038–1043 (1996)
12. Tailhan, J.L., Dal Pont, S., Rossi, P.: From local to global probabilistic modelling of concrete
cracking. Ann. Solid. Struct. Mech. 1, 103–115 (2010)
13. Rossi, P., Wu, X., Le Maou, F., Belloc, A.: Scale effect on concrete in tension. Mater. Struct.
27, 437–444 (1994)
14. Rossi, P.: Experimental study of scaling effect related to post-cracking behaviours of metal
fibres reinforced (MFRC). Eur. J. Environ. Civil Eng. 16(10), 1261–1268 (2012)
Experimental/Computational-Based
Determination of Material Parameters
for Nonlinear Simulation of UHPFRC
David Lehký1(&), Martin Lipowczan1, Drahomír Novák1,

Radomír Pukl2, and Milad Hafezolghorani3
1
Brno University of Technology, Brno, Czech Republic
lehky.d@fce.vutbr.cz
2
Cervenka Consulting, Prague, Czech Republic
3
Dura Technology Sdn. Bhd, Chemor, Perak, Malaysia
Abstract. The paper describes development of a soft-computing-based iden-

tification software FRCID-4PB for material parameter determination of ultra-
high-performance fiber-reinforced concrete composite material. Such a deter-
mination is performed with the help of experimental data from four-point
bending tests used in inverse analysis based on artificial neural networks and
nonlinear computational modelling of tests. A new tensile softening model for
studied composite material has been proposed and tested with the help of sen-
sitivity analysis. The procedure also utilizes stratified statistical simulation
method for the preparation of a training set for the artificial neural network.
Trained network is then implemented into the software allowing routine and
user-friendly identification of material parameters from the test results. The main
aspects of the software implementation, its testing and application is described
and discussed in the paper.
Keywords: Ultra-high-performance fiber-reinforced concrete (UHPFRC)

Soft computing Parameter identification Computational modelling Artificial
neural network
1 Introduction
Over the last two decades, Ultra-High-Performance Fiber-Reinforced Cement-Based

Composites (UHPFRC) are enjoying a growing share of popularity, with a progres-
sively increasing number of implementations [1]. Bridges comprise the category which
has benefited the most from UHPFRC implementation. Since the millennium, over 200
bridges or bridge components have been constructed using UHPFRC all around the
world [2]. UHPFRC with their compressive strength ranging from 150 to 200 MPa and
their tensile strength achieving up to 15 MPa, substantial durability due to low porosity
in comparison with other high-performance materials, UHPFRC positions itself also as
a viable option for rehabilitating existing structures [3].
However, the complexity of this heterogeneous material is not yet fully understood.
An advanced assessment of mechanical fracture properties is of primary importance for
design, quality control and numerical simulation of engineering structures and its
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_48
528 D. Lehký et al.
components made of UHPFRC. The constitutive law of UHPFRC in tension can be

obtained straightforwardly by standard uniaxial direct tensile tests. However, tensile
tests are affected by the eccentricity both of the load and the specimen, the boundary
conditions (fixed-end and rotating-end conditions) and stress concentrations. As a
result, design procedures recommend indirect tensile tests such as three-point bending
tests or four-point bending tests. These are considerably simpler to conduct than the
direct tensile test and require less care in the preparation of the specimens. The major
criticism of flexural tests concerns the fact that the maximum tensile stress occurs at the
outer surface, which accentuates the effect of surface irregularities and may result in
low indicated values of tensile strength. Another great disadvantage of flexural tests is
that an inverse analysis is required to determine the constitutive law of the material in
tension. An inverse analysis methodology for UHPFRC, based on four-point bending
tests, is proposed e.g. in [4]. The model assumes a two-step solution, with a preliminary
inverse analysis to convert the load versus deflection curve into bending moment
versus curvature. In [5] an inverse analysis procedure has been proposed to evaluate
tensile behavior of UHPFRC from test results with notched UHPFRC beams subjected
to three-point bending test. The procedure is based on section analysis in which the
simplified diverse embedment model is employed to take into account UHPFRC stress
distribution along the section with a notch.
In this paper, material parameter determination is performed via soft computing-
based inverse analysis based on four-point bending test [6]. In this method, an artificial
neural network (ANN) is employed as a surrogate model of inverse function. Such
approach allows easy software implementation of the identification process [7].
The research and proposed methodology consist of several steps:
• Development of both a computational model of unnotched beam subjected to four-
point bending test and a suitable constitutive law for UHPFRC cementitious
composites, respectively. Both were created in accordance with experimental data
obtained in laboratory of DURA Technology Sdn. Bhd. Company.
• Sensitivity analysis which is performed in order to understand the effect of the
individual material parameters on the beam response in the four-point bending test.
The aims of the sensitivity analysis are (i) to identify the so-called dominant
parameters, (ii) to determine the parameters that do not influence the response and
thus can be excluded from the subsequent inverse analysis, (iii) to find out the real
ranges of individual parameters, and (iv) to propose a set of suitable response
parameters that will be used as inputs of the inverse analysis.
• Identification of the material parameters which is based on the combination of
experimental test, numerical simulations, sensitivity analysis and artificial neural
network-based inverse analysis. Part of this step was development of FRCID-4PB
software for user-friendly application of above-mentioned complex methodology.
• Verification of the software and the implemented artificial neural network by means
of the comparison of identified material model parameters, and also via the com-
parison of structural responses obtained for both the original (experimental) and
identified input parameters.
Experimental/Computational-Based Determination of Material 529
2 Computational Model
Laboratory tests of samples made of produced materials are important part for quality
control. They also provide a useful information on material properties and behaviour of
the composite. In case of FRC and UHPFRC materials, four-point bending (4PB) tests
on unnotched prism specimens are widely used. The outcome of each test is a force vs.
deflection diagram (F–d diagram), which is subsequently used as input data for
mechanical fracture parameter identification. The nonlinear computational model for
numerical simulations of the four-point bending beam tests were created in
ATENA FEM software [8, 9]. This software enables the response of the structural
member to be calculated, including material damage.
The constitutive law at material point plays a crucial role in nonlinear numerical
analysis. The realism of the structural response, damage and failure calculated by the
computational model depends strongly on the quality of the material model, which
decides the exactness and accuracy of the achieved results. Since UHPFRC is a rather
complex heterogeneous material with a strongly nonlinear response, the “3DNonlin-
earCementitious2User (3DNLC2User)” material model was utilized for the realistic
modelling of this composite. It can capture all the important aspects of the UHPFRC
material’s behaviour and response under tensile as well as compressive loading.
Concrete in tension is described within the smeared crack concept by the nonlinear
fracture mechanics with crack band approach. The main material model parameters are
tensile strength and the shape of the softening function, which characterizes the crack
mouth opening in relation to the remaining tensile stress. In the model, a real discrete
crack is represented as a band of localized strains where the strain corresponding to the
crack width is related to the size, shape and orientation of the finite element. The
softening function for the material law used in the smeared crack concept has to be
determined for every material point (or finite element) in such a way that the objective
crack opening law will be preserved. Only such an approach based on an energy-related
formulation assures an objective solution independent of the finite element mesh [8].
In the case of UHPFRC composite, in order for a softening function to be appro-
priate it must reflect all the typical aspects of UHPFRC cracking. The proposed
function for the optimal reproduction of the tensile softening of the investigated
composite material is shown in Fig. 1. The form of the function is described by tensile
strength and four additional parameters, C1 to C4. Its form attempts to capture all the
particular stages of damage to the FRC material, such as the reaching of ultimate
strength by the cement matrix (see the point in Fig. 1 where stress reaches ft), the
delayed activation of the fibers within the composite (see C1 parameter), and the
consequent stiffening of the composite with activated fibers after the bearing capacity
of the matrix is exhausted (see the stiffening part of the diagram till strain reaches C2
value).
Fig. 1. Stress–strain law utilized for the UHPFRC material.
3 Material Parameter Identification
3.1 Methodology
An artificial intelligence-based inverse procedure developed by Novák and Lehký [6]
transforms fracture test response data into the desired mechanical fracture parameters.
This approach is based on matching laboratory measurements with the results gained
by reproducing the same test numerically. The ANN is used here as a surrogate model
of an unknown inverse function between input mechanical fracture parameters P and
the corresponding response parameters R:
1
P ¼ fANN ðRÞ ð1Þ
The cornerstone of inverse analysis is an artificial neural network, which is of a

feed-forward multilayer type. The most important step in the whole procedure is the
creation of the network and its training—the adjustment of its synaptic weights and
biases. The set for training the ANN is prepared numerically via the utilization of an
FEM model using a stress-strain law as shown in Fig. 1. It simulates 4PB testing with
random realizations of material parameters. These are generated with the help of the
stratified sampling method (LHS) and by performing an inverse transformation of the
distribution function in order to reflect the probability distribution of the parameter.
The random responses obtained from the computational model and the corre-
sponding random realizations of parameters serve as input–output elements of the ANN
training set. After training, the ANN is ready to solve the main task, which is to provide
the best material parameters in order for the numerical simulation to achieve the best
agreement with the experiment. This is performed by simulating a network using the
previously measured responses as an input. This results in a set of identified material
parameters. The last step is result verification—the calculation of the computational
model using the identified parameters. Comparison with the experiment will show the
extent to which the inverse analysis was successful. More theoretical backgrounds and
detailed description can be found in [7].
3.2 Sensitivity Analysis

The second step of the procedure was dedicated to sensitivity analysis (SA). SA help to
identify what are termed the dominant parameters along with the parameters that do not
influence the response and thus can be excluded from the inverse analysis. An addi-
tional aim was to determine the real ranges of parameters and to propose a set of
suitable parameters that will be used as inputs for the inverse analysis. SA was per-
formed for six model parameters—tensile strength, compressive strength and four
parameters of the tensile softening model, C1 to C4, see Fig. 1.
The results of SA were extensively presented in [10]. The main outcomes are
summarized here:
• All studied parameters significantly influence the structural response with the
exception of compressive strength.
• The size of parameter C1, which is related to the delayed activation of fibers in the
cement matrix, controls the “depth” of the first curve drop. Due to its clear effect on
the first visible decrease of the F–d diagram when reaching the ultimate bearing
capacity of the cement matrix, and with respect to the constant drop of diagram for
studied composite material it is possible to exclude it from the identification process
and consider it to have a constant value of C1 = 0.8.
• Parameters C2 and C3 are closely related to the ultimate bearing capacity of the
beam—C3 influences the ultimate force and slope of the second part of the diagram,
when it is strengthened due to the joint action of the matrix and the fibers while C2
affects the amount of related deformation.
• Parameter C4 influences the maximum deformation that will be achieved during the
test. When C4 has a large enough value, the numerical test is not terminated at a low
deflection value. Parameter C4 was set to 0.15 based on the SA results and was
further excluded from identification.
3.3 Stochastic Model

Based on the SA results the identification set containing four material model param-
eters were selected—tensile strength, C2 and C3, and the modulus of elasticity, which
influences the slope of the initial linear part of the diagram. In order to prepare a
training set for ANN, all parameters were considered as random variables with rect-
angular distribution function and basic statistical moment according to Table 1.
Appropriate ranges of parameters were adjusted according to SA results. Note that Xmin
and Xmax stand for lower and upper boundary of rectangular distribution.
Table 1. Stochastic model utilized in preparation of the ANN training set.

Parameter Distribution Mean StD CoV Xmin Xmax
E (MPa) Rectangular 70000 5774 0.082 60000 80000
ft (MPa) Rectangular 11 1.732 0.157 8 14
C2 (–) Rectangular 0.003 0.001 0.385 0.001 0.005
C3 (–) Rectangular 1.3 0.173 0.133 1 1.6
The set for training the ANN is prepared numerically via the utilization of an FEM
model using a stress-strain law as shown in Fig. 1. It simulates 4PB testing with
random realizations of material parameters. These are generated with the help of the
stratified sampling method (LHS) and by performing an inverse transformation of the
distribution function in order to reflect the probability distribution of the parameter. As
mentioned above the random responses obtained from the computational model and the
corresponding random realizations of parameters serve as input–output elements of the
ANN training set.
4 Software FRCID-4PB
The above-described method of parameter identification, which combines nonlinear

simulations with the training of an artificial neural network, is relatively time con-
suming and of high complexity. Therefore, the whole procedure was implemented in
FRCID-4PB software. Screenshots of its first version is shown in Fig. 2. Software
contains already created and trained ANN which structure is depicted at the bottom of
Fig. 2. It consists of eight nonlinear neurons in the hidden layer and four linear neurons
in the output layer, each related to one identified parameter. There are also four inputs
to the network which correspond to the four response parameters derived from
experimental F–d diagram during initial data processing, see the top part of Fig. 2. The
training set consisted of 50 random realizations of parameters generated using LHS
method and corresponding responses obtained from the computational model.
The ANN training was carried out using Genetic algorithms combined with gradient
descent method. More details on ANN-based identification method can be found in [6,
7].
5 Verification and Application
The proposed method and software with implemented neural network were verified
using experimental data obtained in laboratory of DURA Technology Sdn. Bhd.
Company from Malaysia. Here, the results for four selected unnotched prismatic
specimen with the nominal dimensions of 100 100 500 mm made of
DURA UHPFRC material and subjected to four-point bending are presented. First, the
laboratory data (F–d diagrams) were loaded to FRCID-4PB software and pre-processed
in order to obtain corresponding response parameters. Second, response parameters
were introduced to the network as an input signal which was then propagated forwardly
through the network from the inputs to the output layer. Result of network simulation is
the set of identified parameters, see Table 2.
Identified parameters were subsequently used in computer simulation of laboratory
tests. A comparison of the resulting numerical responses with the original experimental
responses in the form of F–d diagrams is depicted in Fig. 3. Note that among important
features of the response belong the initial stiffness, the first drop of the diagram, the
slope of the hardening part, maximum load-bearing capacity, related deflection and
Fig. 2. FRCID-4PB software: loading and processing of experimental data (top) and parameter
identification (bottom).
slope of the final softening part of the diagram. The results for all tested samples show a
good agreement between the original and identified data in all these aspects.
Table 2. Set of identified material model parameters for four samples made of DURA
composite material.
Specimen Material parameter
E (MPa) ft (MPa) C2 (–) C3 (–)
FT1 58524 10.285 0.0018 1.1652
FT3 87914 7.913 0.0010 1.4555
FT4 81090 8.851 0.0010 1.4830
FT5 86869 10.762 0.0012 1.2166
Fig. 3. Comparison of original versus simulated structural responses for four selected samples.
6 Conclusions
The paper presents development of a soft-computing-based identification software

FRCID-4PB for material parameter determination of UHPFRC material using results of
four-point bending tests. The main conclusions which can be drawn from the com-
putational modeling of elements made of UHPFRC material and identification of its
parameters supported by sensitivity analysis are:
• The proposed stress–strain law and its implementation into the 3DNLC2User
material model reflects all the typical aspects of UHPFRC cracking.
• The dominant material parameters for description of material behavior when sub-
jected to four-point bending are modulus of elasticity E, tensile strength of cement
matrix ft, and four parameters C1–C4 of tensile softening model.
• A verification of software and implemented ANN performed on selected UHPFRC
samples proved the ANN’s efficiency and ability to identify material parameter
values leading to the accurate simulation of the response of the studied composite.
Acknowledgements. This work has been supported by IdeMaS project No. FV30244 awarded
by the Ministry of Industry and Trade of the Czech Republic and MUFRAS project No. 19-
09491S awarded by the Czech Science Foundation (GACR). This support is gratefully
acknowledged.
References
1. Buttignol, T.E.T., Sousa, J.L.A.O., Bittencourt, T.N.: Ultra High-Performance Fiber-
Reinforced Concrete (UHPFRC): A review of material properties and design procedures.
Rev. IBRACON Estrut. Mater. 10(4), 957–971 (2017)
2. Hafezolghorani, M., Voo, Y.L.: Design of 38m span post-tensioned ultra high performance
fiber-reinforced concrete (UHPFRC) composite bridge. IOP Conf. Series: Mater. Sci. Eng.
431 (042007) (2018)
3. Martín-Sanz, H., Chatzi, E., Egger, A.: Sustainable strengthening of structures using ultra
high performance fibre reinforced cement-based composites and computational verification.
In: Challenges in Design and Construction of an Innovative and Sustainable Built
Environment. Proceedings of the 19th IABSE Congress, Stockholm (2016)
4. Baby, F., Graybeal, B., Marchand, P., Toutlemonde, F.: UHPFRC tensile behavior
characterization: Inverse analysis of four-point bending test results. Mater. Struct. 46, 1337–
1354 (2013)
5. Lee, S.-C., Kim, H.-B., Joh, C.: Inverse analysis of UHPFRC beams with a notch to evaluate
tensile behavior. Adv. Mater. Sci. Eng. 2017 (6543175) (2017)
6. Novák, D., Lehký, D.: ANN inverse analysis based on stochastic small-sample training set
simulation. Eng. Appl. Artif. Intell. 19, 731–740 (2006)
7. Lehký, D., Keršner, Z., Novák, D.: FraMePID-3PB software for material parameters
identification using fracture test and inverse analysis. Adv. Eng. Soft. 72, 147–154 (2014)
8. Červenka, V., Červenka, J., Pukl, R.: ATENA—A tool for engineering analysis of fracture in
concrete. Sadhana 27(4), 485–492 (2002)
9. Červenka, V., Jendele, L., Červenka, J.: ATENA Program Documentation—Part 1: Theory.
Červenka Consulting, Prague, Czech Republic
10. Novák, D., Lehký, D., Pukl, R.: Fiber-reinforced cementitious composite: Sensitivity
analysis and parameter identification. Adv. Mater. Lett. 11(3), 1–5 (2020)
Mechanical Response of High Strength Fibre
Reinforced Concrete Under Extreme Loads
B. Luccioni1(&), F. Isla1, F. Fiengo1, R. Codina2, D. Ambrosini2,

J.C. Vivas3, Raúl L. Zerbino3, Graciela M. Giaccio4,
and María C. Torrijos3
1
CONICET, Structure Institute, National University of Tucumán, Tucumán,
Argentina
bluccioni@herrera.unt.edu.ar
2
CONICET, Engineering Faculty, National University of Cuyo, Mendoza,
Argentina
3
CONICET, LEMIT, Engineering Faculty, National University of La Plata,
La Plata, Argentina
4
CIC, LEMIT, Engineering Faculty, National University of La Plata, La Plata,
Argentina
Abstract. High Strength Fibre Reinforced Concrete (HSFRC) presents great

advantages when compared with conventional concrete under static loads and
thus, it constitutes a promising material to withstand extreme loads. An exper-
imental and numerical research carried out with the objective of developing
design criteria for HSFRC use in protective structures construction is presented.
The mechanical behaviour of HSFRC elements under extreme loads is experi-
mentally and numerically analysed. Numerical models represent useful tools for
the design of this type of HSFRC applications but they should be carefully
calibrated and validated with experimental results. HSFRC prisms and slabs
including different types of hooked-end steel fibres are tested under static, blast
and impact loads. Material models at the meso and the macro scale are devel-
oped, they are calibrated with characterization tests and validated with experi-
mental results. Experimental results are analysed with the aid of numerical
models showing the effect of fibre type and content under extreme load.
Numerical models are able to reproduce the blast and impact tests results and
give additionally information about the local and structural response under
impulsive loads that could be valuable for the design of protective structures.
Keywords: High strength concrete Steel fibres Blast response Impact

Numerical model
1 Introduction
Many researchers have proven that HSFRC has an improved behaviour with respect to
conventional concretes under low and high-speed dynamic loads. These advantages
make it suitable for the construction of new strategic structures or for the repair and
reinforcement of existing structures against impact loads or explosions. However, the
use of these materials in protective structures is still a challenge. The complexity of the
behaviour added to that of impact and impact with penetration require a deep
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_49
Mechanical Response of High Strength Fibre Reinforced Concrete 537
knowledge about the behaviour of the material. Moreover, robust prediction tools that
allow to reproduce the behaviour under this type of actions are needed for the proper
design of HSFRC elements.
Several types of tests are normally used to study the behaviour of concrete and fibre
reinforced concrete (FRC) under impact loads. Drop weight and pendulum impact tests
are used to produce impact at relatively low strain rates while Split Hopkinson Pressure
Barr tests and blast tests are usually performed to produce high strain rate loads [1].
The present knowledge about fibre reinforced concrete dynamic behaviour [2–4] under
high strain rates is still limited and there are still some contradictory results [5]. Tensile,
compressive and flexure strength increase with strain rate while toughness and post-
peak ductility are absent or reduced at high strain rates [1]. It was also proved that
tensile and compressive strengths, energy absorption, fracture toughness and impact
resistance increase with steel fibre content [1].
Some results about the behaviour of fibre reinforced concrete elements under blast
loads have been published [6–8] but experimental results concerning HSFRC and ultra-
high performance fibre reinforced concrete (UHPFRC) elements under close or contact
explosions are still very scarce [9, 10] and it is not possible to accurately predict
spalling damage [9] using available empirical methods [11].
HSFRC elements under blast loads have shown benefits in damage control,
accelerations and displacements when compared to conventional concrete elements.
UHPFRC elements have also presented higher ductility, lower permanent deformation,
higher load bearing capacity, crack control and greater ability to absorb energy without
fragmentation [12, 13] than conventional concrete panels. Blast resistance increases
with the increase of fibre volume and it was shown that different types of steel fibres
have similar effects improving blast strength. Fibres help controlling concrete spalling
and crack propagation [14]. The panels are less likely to fail presenting higher strength
and extension of damage than conventional concrete specimens [15].
Available constitutive models for the simulation of FRC and composites can be
classified in macro and meso models. Macro models simulate the behaviour of the
composite using an equivalent homogeneous model with average properties. Consti-
tutive laws are directly derived from tests results. The main advantage of this type of
models is that they use information that is relevant at the structural scale. As a
counterpart, they require large number of tests considering different fibres contents and
types of loads since the fibres are not explicitly taken into account. Some of these
problems can be avoided using meso-models that take into account the material
components: concrete matrix, fibres and sometimes fibre/matrix interface [16–19]. This
type of models has three key issues: the derivation of the fibres sewing forces through
cracks, the model used for the concrete matrix and the way in which the composite
behaviour is obtained from the components. The models differ in the way in which
these three aspects are derived and combined.
Sewing forces through cracks can be obtained from pull-out tests or using meso-
models [16, 20]. The fibres are explicitly modelled in some meso-mechanic approaches
using different types of discrete elements [21].
Except for a few models [13, 22], most of the meso-models have been developed
and calibrated under static loads and are not suitable to simulate dynamic behaviour
because they don’t take into account strain rate effects.
538 B. Luccioni et al.
The behaviour of HSFRC elements under impact and blast loads is frequently
simulated with explicit codes like using concrete macro-models [23–25]. The cali-
bration of these models is not a simple task. The use of meso-models to analyse fibres
contribution and the effect of fibres geometry on mechanical properties can contribute
reducing the required number of tests to calibrate these homogeneous models.
An experimental and numerical research carried out with the objective of devel-
oping design criteria for HSFRC and numerical simulation tools is presented in this
paper.
2 Experimental Tests
Experimental tests consisting on blast tests on slabs and impact tests on beams were
performed in order to study de behaviour of HSFRC under extreme loads and the effect
of fibres type and content on the corresponding response.
2.1 Materials and Mixtures

Five High Strength Concretes were used for the experimental study: a plain concrete
(P) and four HSFRCs varying the fibre length and the fibre dosages. All concretes were
prepared using the same base self-compacting matrix. Ordinary Portland cement
(730 kg/m3), silica fume (73 kg/m3), calcareous filler (48 kg/m3), natural siliceous
sand (490 kg/m3) and 12 mm maximum size granitic crushed stone (860 kg/m3) were
used as component materials. A set retarding plasticizer (1 kg/m3) and a high range
water reducer (8.9 kg/m3) were used as chemical admixtures. The water/(ce-
ment + silica fume) ratio was equal to 0.24.
Two types of hooked-end high carbon steel fibres (L30 and L60) with different
lengths (30 and 60 mm) and diameters (0.38 and 0.75 mm respectively) were added to
HSFRC. The tensile strength of the fibres was over 2500 MPa. Two fibres contents
were used 40 and 80 kg/m3. HSFRC are identified with the name of the fibre followed
by the fibre content (L30-40, L30-80, L60-40, L60-80).
To evaluate the compressive strength [26] and the static elastic modulus [27]
100 mm x 200 mm cylinders were cast. Mean values of 114 MPa and 40 GPa were
obtained for the compression strength and the elastic modulus respectively.
Prisms of 430 mm 50 mm 105 mm with a notch of approximately 19 mm in
the central section were used for static bending characterization [28]. They were tested
under three points loading. The crack mouth opening displacement (CMOD) was
registered. After the tests, specimens were separated in two parts and the fibres crossing
the central section were counted. In all cases, fibres were pulled out without failing.
Additionally, 40 pull-out tests of each type of fibres, including some straight fibres
(cut hooked-end), were performed. The fibres were aligned with the load and the
embedded length was half the fibre length. The pull-out rate was varied from 0.017 to
8.33 mm/s obtaining similar responses for the different velocities [29].
All the specimens were cured 28 days in moist room and then stored at the labo-
ratory environment during four months until the characterization, blast and impact tests
were performed.
2.2 Blast Tests

Square slabs of 550 mm side and 50 mm thickness were tested under blast loads. They
were supported on a highly reinforced steel frame leaving a free span of 460 mm.
Three different types of blast tests were performed on the slabs varying the explosive
masses and stand-off distances. In all cases, the explosive had cylindrical shape and the
detonator was located in the centre of the upper surface. In Test 1 49 g of equivalent
TNT were led on the slabs while in Tests 2 and 3 the explosive was supported on an
expanded polystyrene block over the slabs. For Tests 2 and 3, 244 g and 488 g with a
height of the explosive gravity centre over the slabs of 0.2525 m and 0.2725 m were
used respectively. More details about the blast tests can be found in Refs. [29] and [30].
2.3 Impact Tests

Prisms of 430 mm 75 mm 105 mm of P, L30-40 and L30-80 were used for drop
weight impact tests. The prisms were mounted with 320 mm of free span in the drop
weight machine illustrated in Fig. 1 where the different parts are indicated.
The projectile consisted of a 16 mm thick, 304 mm long, 5182 g steel plate with a
semi-cylindrical tup at the bottom end. Fall heights were varied between 33 mm and
1000 mm. Due to limitations of the measurement equipment range, the accelerations on
the bottom of the beam were measured for fall heights up to 100 mm, and the impact
force could be measured for heights up to 429 mm. For greater heights, only the crack
opening, the beam deflection and the damage pattern were registered. A piezoelectric
load cell (PCB, 203 model) was used to record the impact force. An accelerometer
(PCB, 353 B03 model) was used to measure accelerations. Both devices were con-
nected to a signal conditioner (PCB, 482C05 model), which, in turn, was connected to
two signal acquisition boards (Measurement Computing, USB-1608FS model). The
signal acquisition rate was 50 kHz.
Fig. 1. Impact test setup. (1) Projectile, (2) metal plates, (3) load cell, (4) metal bar that fixes the
specimen to the support, (5) metal rod that links the metal bar and metal supports, (6) metal
supports, (7) accelerometer, (8) HARRF beam, (9) projectile side view and (10) projectile semi-
cylindrical tup.
3 Numerical Models
The numerical simulation of HSFRC slabs under blast loads and HSFRC was per-
formed with hydrocodes using concrete macro-models calibrated to reproduce the
behaviour of HSFRC with different types and contents of steel fibres. A meso-model
developed by the authors [18] was first used in order to obtain the parameters required
for the macro-models.
3.1 Meso-Model for HSFRC

A meso-model was used to obtain the properties [19] for the homogeneous equivalent
models available in hydrocodes for the simulation of HSFRC elements under blast and
impact loads. HSFRC is assumed as a composite material made of a high strength
concrete matrix with dispersed fibres oriented in a finite number n of discrete orien-
tations. A modification of classical mixture theory is used to take into account ortho-
tropic behaviour of fibres and their slip respect to the matrix in a simplified way [18].
Concrete matrix is modelled with an elastoplastic model [31] and steel fibres are
assumed as orthotropic elastoplastic inclusions that can debond and slip from the
matrix. Constitutive equations of fibres are modified in order to include this inelastic
phenomenon without explicitly modelling the interface. The model requires concrete
properties, fibres material, geometry, distribution and orientation as input data. The
fibres bond-slip behaviour is obtained from a pull-out model [30]. The strain rate
dependence is explicitly taken into account in the constitutive models of the compo-
nents. The model was proved to accurately simulate the 2D static behaviour of steel
fibre reinforced concrete [18] and UHPFRC under high strain rates [19].
First, pull out tests of straight fibres obtained from both fibres used were numeri-
cally simulated. In this way, the sliding threshold, the residual tangential strength and
the friction coefficient were calibrated. Then, the load-displacement curves for hooked-
end fibres were numerically reproduced with the pull-out model. The results for aligned
fibres are presented in Fig. 2 where a good agreement between numerical results and
average experimental response is observed.
P beams results were used to calibrate concrete tension strength and fracture
energy. Then, the flexure response of HSFRC notched beams with different fibres
contents was simulated with the meso-model described and compared with experi-
mental results in order to check the model validity. The fibres were supposed to be
located in vertical planes, 40% in axial direction, 30% at 60º and 30% at -60º. This fibre
distribution is only valid for these beams and should be approximated in correspon-
dence with the actual fibre distribution in other cases. The pull-out responses for
different fibre orientations were obtained with the pull-out model. As illustration, the
flexure responses of HSFRC with different contents of fibres L60 are shown in Fig. 3.
These fibre contents correspond to the actual fibre contents verified after the tests.
Once the material properties and the model were validated, compression and ten-
sion response curves for the different HSFRC were obtained. These curves were used
as input data for the macro-models.
a) b)
Fig. 2. Pull-out tests of aligned fibres. a) Fibres L30, b) Fibres L60.
a) b)
c) d)
Fig. 3. Flexure tests. a) P, b) L60-0.25%, c) L60-0.5%, d) L60-0.75%.
3.2 Numerical Simulation of Blast Tests

Blast tests were numerically modelled with AUTODYN [32]. The model consisted of
the air volume, where the slabs were immersed and the blast wave was generated and
propagated, the slabs and the support. The ideal gas equation of state (EOS) was used
for the air while JWL EOS was used for TNT. Standard material parameters available
in AUTODYN material library were used for air and TNT [32]; The problem was
modelled from the beginning (explosive detonation) with a 3D model and using an
Euler Gudunov processor. A Lagrange processor with 3D brick elements of 5 mm side
was used for the slabs while shell elements were used for the supports. Euler Lagrange
interaction was defined between air and the slab and the supports.
Concrete RHT model [33] in combination with p-a equation of state (EOS) was
used for HSC and HSFRC. Based on previous numerical results, the effect of adding
different steel fibres and fibres contents was taken into account modifying the main
parameters related to tension strength and ductility (failure strain and erosion limit).
More details about this calibration can be found in Ref. [30].
3.3 Numerical Simulation of Impact Tests

The numerical simulation of impact tests was performed with LS-Dyna [34]. All ele-
ments except for the metallic rods were simulated using brick elements (ELFORM 1)
of 5 mm for the beam and the supports and 3 mm side for the projectile and load cell.
Truss elements were used for the rods.
K&C model [35] calibrated using the described meso-model was used for HSC and
HSFRC. Basically, the parameters of the compression and tension hardening curves for
the different fibres contents were obtained. A penalty formulation (soft 0) was used for
contact and friction was taken into account. Rayleigh damping with parameters
obtained from calibration with experimental results was considered.
4 Experimental and Numerical Results
4.1 Blast Tests Results

The damage patterns on front and rear face of some of the slabs after blast Tests 1 and 3
are shown in Figs. 4 and 5 where numerical results are also included for comparison.
The rest of the results can be found in Ref. [30].
Face Front Rear

P
L30-40
L30-80
Fig. 4. Experimental results and numerical simulation of blast Test 1.

Face Front Rear

P
L30-40
L30-80
Fig. 5. Experimental results and numerical simulation of blast Test 3.
In all tests performed on plain concrete brittle flexure failure of the slabs was
obtained due to the reduced slab thickness and the brittle nature of HSC. An important
concrete spalling at the rear face that reaches the front face was observed for contact
explosion (Test 1). The damage patterns obtained for HSFRC slabs were strongly
different from those of P slabs. No HSFRC slab tested under contact explosions (blast
Test 1) exhibited flexure failure and they preserved integrity after the tests (Fig. 4).
Flexure cracks can hardly be identified. The front faces of the slabs presented a small
eroded zone like a crater produced by the high pressures originated by the detonation of
the explosive in contact with concrete. The crater dimensions were reduced when fibre
content was increased but only a slight reduction of crater dimensions was obtained for
fibres L30 when doubling the fibre volume content. Slabs containing fibres L30 (shorter
fibres) exhibit smaller craters on front face and spalling on rear face than slabs with
fibres L60 with the same fibre content showing that for the same fibre content, shorter
fibres are more efficient controlling concrete cratering and spalling than longer fibres.
In blast Test 2 slab P was fractured while L30-40 and L60-40 slabs presented
flexure cracks at the rear face and slabs L30-80 and L60-80 exhibited almost imper-
ceptible flexure cracks. Erosion in a reduced zone was observed in the front face while
a bigger spalling zone was created at the rear face. In contrast with L60-40, slab L30-40
did not exhibit either crater or spalling showing the greater effect of shorter fibres
controlling these types of damage. Increasing the fibre, content flexure cracks and
spalling zone were reduced.
For the case of blast Test 3 (Fig. 5). slabs P and L30-40 were fractured but L60-40
presented wide flexure cracks suggesting that longer fibres are more efficient control-
ling flexure cracks than shorter fibres. As in the rest of the blast tests, fibres were
pulled-out without failing in all cases.
A good agreement between numerical and experimental final damage patterns was
obtained but the numerical simulations show that homogeneous models are not able to
reproduce HSFRC behaviour under the whole range of scaled distances to reproduce
flexure failure, cratering and spalling [30].
4.2 Impact Tests Results

Only P and HSFRCs with the shorter fibres have been tested under impact loads until
now. The comparison of numerical and experimental results for two fall heights H on
L30-80 beam are presented in Fig. 6. It can be observed that numerical acceleration
peaks (a_max = 474 g, a_min = −400 g) are greater than those registered in the tests
(a_max = 377 g, a_min = −334 g). Nevertheless, the envelopes of both records are
similar. The impact forces histories are satisfactorily reproduced
60 60
400 Exper Exper Exper
Impact load [kN]
Acceleration [g]
Impact load [kN]

Numer Numer Numer
200 40 40
0
20 20
-200
0 0
0 1 2 0 1 2 0 1 2
a) Time [ms] b) Time [ms] c) Time [ms]
Fig. 6. Impact (L30-80 beam) a) Acceleration H = 33 mm b) Load H = 33 mm c) Load

H = 300 mm.
The permanent deflections numerically obtained for the first impact and those
registered in the tests for three consecutive impacts from the same height are presented
in Fig. 7 for two fibres contents. The results for each height correspond to a different
specimen. For all the heights above 100 mm the fibres were pulled-out without failing.
The deflections for the first impact are well reproduced by the numerical model. The
comparison of Figs. 7a and b shows the deflection are reduced when the fibre content is
increased but the effect is more important for the repeated impacts.
15 15
1st impact (numer.) 1st impact (numer.)
1st impact (exper.) 1st impact (exper.)
Deflection [mm]
Deflection [mm]
10 2nd impact (exper.) 10 2nd impact (exper.)

3rd impact (numer.) 3rd impact (exper.)
5 5
0 0
0 250 500 750 1000 1250 0 250 500 750 1000 1250
Fall height H [mm] Fall height H [mm]
Fig. 7. Deflections a) Beam L30-40 b) Beam L30-80.

5 Conclusions
The effect of incorporating different contents of hooked-end high carbon steel fibres of
different lengths in a HSC on blast and impact response was experimentally and
numerically studied. The following conclusions are only valid for the particular con-
crete composition chosen for the testing. Different casting techniques or mix designs
could result in different fibre behaviour.
• The incorporation of fibres, not only increases residual loading capacity and
toughness under static loads but also significantly improves impact and blast
behaviour. Flexural crack thickness, erosion and spalling zones dimensions and
permanent deflections are reduced with the increase fibre content.
• For the same fibre content, shorter fibres lead to a greater improvement of the matrix
behaviour and they are more effective controlling cratering and spalling.
• Numerical models developed are able to approximately reproduce static, blast and
impact behaviour but further development should be performed mainly related to
erosion criteria and limits since it is difficult to reproduce cater and spalling under
blast loads and fracture under impact loads with the available erosion models.
Acknowledgements. The cooperation in the blast tests of Prof. Oscar Curadelli, Bioeng.
Gabriel Houri, Dr. Eng. Martín Domizio and Dr. Eng Hernán Garrido are specially acknowl-
edged. The authors wish to thank the financial support of National Agency for Scientific and
Technological Promotion, National Scientific and Technological Research Council (CONICET),
LEMIT-CIC, National University of Tucumán Research Council (CIUNT) and National
University of Cuyo. The companies Cementos Avellaneda S.A., Ferrocement S.A, N.V.
Bekaert S.A. and Sika Argentina S.A.I.C. that provided the cement, the silica fume, the steel
fibres and the chemical admixtures are also greatly acknowledged.
References
1. Soufeiani, L., Raman, S., Bin Jumaat, M.Z., Alengaram, U.J., Ghadyani, U.J., Mendis, P.:
Influences of the volume fraction and shape of steel fibers on fiber reinforced concrete
subjected to dynamic loading–a review. Eng. Struct. 124, 405–417 (2016)
2. Millard, S.G., Molyneaux, T.C., Barnett, S.J., Gao, X.: Dynamic enhancement of blast-
resistant ultrahigh performance fibre-reinforced concrete under flexural and shear loading.
Int. J. Impact Eng. 37, 405–413 (2010)
3. Wang, S., Zhang, M.-H., Quek, S.T.: Mechanical behavior of fiber-reinforced high-strength
concrete subjected to high strain-rate compressive loading. Constr. Build. Mater. 31, 1–11
(2012)
4. Tran, N.T., Tran, T.K., Kim, D.J.: High rate response of ultra-high-performance fiber-
reinforced concretes under direct tension. Cem. Concr. Res. 69, 72–87 (2015)
5. Caverzan, A., Cadoni, E., di Prisco, M.: Tensile behaviour of high performance fibre-
reinforced cementitious composites at high strain rates. Int. J. Impact Eng. 45, 28–38 (2012)
6. Wang, Z.L., Konietzky, H., Huang, R.Y.: Elastic–plastic-hydrodynamic analysis of crater
blasting in steel fiber reinforced concrete. Theor. Appl. Fract. Mech. 52, 111–116 (2009)
7. Pantelides, C.P., Garfield, T.T., Richins, W.D., Larson, T.K., Blakeley, J.E.: Reinforced
concrete and fiber reinforced concrete panels subjected to blast detonations and post-blast
static tests. Eng. Struct. 76, 24–33 (2014)
8. Foglar, M., Kovar, M.: Conclusions from experimental testing of blast resistance of FRC and
RC bridge decks. Int. J. Impact Eng. 59, 18–28 (2013)
9. Li, J., Wu, Ch., Hao, H., Wang, Z., Su, Y.: Experimental investigation of ultra-high
performance concrete slabs under contact explosions. Int. J. Impact Eng. 93, 62–75 (2016)
10. Nam, J., Kim, H., Kim, G.: Experimental Investigation on the blast resistance of fiber-
reinforced cementitious composite panels subjected to contact explosions. Int. J. Concr.
Struct. Mater. 11(1), 29–43 (2017)
11. Li, J., Hao, H.: Numerical study of concrete spall damage to blast loads. Int. J. Impact Eng.
68, 41–55 (2014)
12. Aoude, H., Dagenais, F.P., Burrell, R.P., Saatcioglu, M.: Behavior of ultra-high performance
fiber reinforced concrete columns under blast loading. Int. J. Impact Eng. 80, 185–202
(2015)
13. Ellis, B.D., Di Paolo, B.P., McDowell, D.L., Zhou, M.: Experimental investigation and
multiscale modeling of ultra-high performance concrete panels subject to blast loading. Int.
J. Impact Eng. 69, 95–103 (2014)
14. Mao, L., Barnettm, S.J., Tyas, A., Warren, J., Schleyer, G.K., Zaini, S.S.: Response of small
scale ultra high performance fibre reinforced concrete slabs to blast loading. Const. Build.
Mater. 93(15), 822–830 (2015)
15. Thoma, K., Stolz, A., Millon, O.: Performance and suitability of ultra-high-performance
concrete under a broad range of dynamic loadings. In: Advances in Protective Structures
Research. IAPS Special publication 1, London Taylor & Francis, London, pp. 65–96 (2012)
16. Caggiano, A., Martinelli, E.: A unified formulation for simulating the bond behaviour of
fibres in cementitious materials. Mater. Des. 42, 204–213 (2012)
17. Oliver, J., Mora, D., Huespe, A., Weyler, R.: A micromorphic model for steel fiber
reinforced concrete. Int. J. Solids Struct. 49, 2990–3007 (2012)
18. Luccioni, B., Ruano, G., Isla, F., Zerbino, R., Giaccio, G.: A Simple approach to model
SFRC. Constr. Build. Mater. 37, 111–124 (2012)
19. Luccioni, B., Isla, F., Forni, D., Cadoni, E.: Modelling UHPFRC tension behavior under
high strain rates. Cem. Concr. Compos. 91, 209–220 (2018)
20. Cunha, V.M., Barros, J.A.O., Sena-Cruz, J.M.: A finite element model with discrete
embedded elements for fibre reinforced composites. Compos. Struct. 94–95, 22–33 (2012)
21. Fang, Q., Zhang, J.: Three-dimensional modelling of steel fiber reinforced concrete material
under intense dynamic loading. Constr. Build. Mater. 44, 118–132 (2013)
22. Bragov, A.M., Karihaloo, B.L., Petrov, Y.V., Konstantinov, A.Y., Lamzin, D.A., Lomunov,
A.K., Smirnov, I.V.: High-rate deformation and fracture of fiber reinforced concrete.
J. App. Mech. Tech. Phys. 53(6), 926–933 (2012)
23. Li, J., Zhang, Y.: Evaluation of constitutive models of hybrid-fibre engineered cementitious
composites under dynamic loadings. Const. Build. Mater. 30, 149–160 (2012)
24. Mao, L., Barnett, S.J.: Investigation of toughness of ultra high performance fibre reinforced
concrete (UHPFRC) beam under impact loading. Int. J. Impact Eng. 99, 26–36 (2016)
25. Guo, W., Fan, W., Shao, X., Shen, D., Chen, B.: Constitutive model of ultra-high-
performance fiber-reinforced concrete for low-velocity impact. Compos. Struct. 185, 307–
326 (2018)
26. ASTM C39/ C39M–15 Standard Test method for Compressive strength of cylindrical
Concrete Specimens. Book of Standards. Volume 04.02. https://doi.org/10.1520/c0039_
c0039m
27. ASTM C469/ C469M–14 Standard Test method for Static Modulus of Elasticity and
Poisson´s Ratio of Concrete in Compression. Book of Standards. Volume 04.02. https://doi.
org/10.1520/c0469_c0469m
28. EN 14651:2005 Test method for metallic fibered concrete - Measuring the flexural tensile
strength (limit of proportionality (LOP), residual). CEN-European Committee for Standard-
ization, Brussels, pp. 1–17 (2005)
Effect of Steel fibers on static and blast response of high strength concrete. Int. J. Impact
Eng. 107, 23–37 (2017)
Experimental and numerical analysis of blast response of high strength fiber reinforced
concrete slabs. Eng. Struct. 175, 113–122 (2018)
31. Luccioni, B., Rougier, V.: A plastic damage approach for confined concrete. Comput. Struct.
83, 2238–2256 (2005)
32. ANSYS, AUTODYN Version 18.1 User’s Manual, 2017
33. Riedel, W., Kawai, N., Kondo, K.: Numerical assessment for impact strength measurements
in concrete materials. Int. J. Impact Eng. 36, 283–293 (2009)
34. LSTC, LS-Dyna Theory manual, Livermore Software Technology Corporation (2018)
35. Malvar, L.J., Crawford, J.E., Wesevich, J.W., Simons, D.A.: A plasticity concrete material
model for DYNA3D. Int. J. Impact Eng. 19, 847–873 (1997)
Numerical Damage Modelling
of Macro-synthetic Fibre Reinforced Concrete
Dayani Kahagala Hewage1(&), Christophe Camille1, Olivia Mirza1,

Fidelis Mashiri1, Brendan Kirkland1, and Todd Clarke2
1
Western Sydney University, Locked Bag 1797, Penrith,
New South Wales 2751, Australia
d.kahagalahewage@westernsydney.edu.au
2
BarChip Pty Ltd., Sydney, NSW, Australia
Abstract. Macro-synthetic fibre reinforcement for concrete applications is

gaining popularity in the construction industry owing to its’ advanced devel-
opment towards higher mechanical properties, electrical and corrosion resis-
tance. However, the main drawback of the effective application of macro-
synthetic fibre reinforced concrete (MSFRC) is the limited analysis procedures
adopted from the existing concrete behavioural models and guidelines. Indeed,
the behaviour of MSFRC is mainly characterised by the post-cracking
hardening/softening, which significantly differs from the brittle nature of plain
concrete.
Currently, material models which are available for the numerical modelling of
steel fibre reinforced concrete (SFRC) characterises the hardening and softening
behaviour immediately after the limit of proportionality. In regards with
MSFRC modelling, the initial frictional slippage of fibres causes an instanta-
neous reduction in the tensile stress (i.e. post-cracking phase), wherein which
the damage evolution requires a distinctive approach. Therefore, the paper
herein focuses on reviewing the adoptability of current models and evaluate the
sensitivity of damage parameters in macro-scale analysis. As a result, this paper
provides significant insights into the different parameters and calibrations
required towards the recognition of the MSFRC material model in the finite
element analysis.
Keywords: Macro-synthetic fibre ABAQUS Concrete damage plasticity

Damage parameter
1 Introduction
The use of fibre in construction materials dates back to Neolithic age as per records [1].
Cob and adobe are such construction materials in which straw, a natural fibrous
material, has been used with soil. With the expansion of the construction industry and
technological development, various types of fibres emerged. Currently, there are steel
fibres, glass fibres, asbestos fibres synthetic plastic fibres, carbon fibres, and natural
fibres, used in construction concrete. Accordingly, these fibres significantly improve
the tensile strength, ductility, severity of failure and energy absorption under all types
of loading (i.e. Static, Dynamic, Fatigue, and Impact) [2–7]
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_50
Numerical Damage Modelling of Macro-synthetic Fibre Reinforced Concrete 549
Thus far, the most common fibre types used in structural applications are steel
fibres and synthetic fibres. Steel fibre exhibits the highest modulus of elasticity and the
density where synthetic fibre possesses the lowest. Theoretically, higher tensile strength
and elastic modulus should result in higher compressive, tensile and flexural strength of
fibre reinforced concrete [5, 7, 8]. However, due to the corrosive nature of steel fibres,
there is a high tendency towards the utilisation of synthetic fibres [9, 10]. The devel-
opment of synthetic fibres with material properties similar to the steel fibres has also
increased the feasibility of using macro-synthetic fibre reinforced concrete (MSFRC),
lately.
Irrespective of the strong advantages, the main drawback of the effective applica-
tion of MSFRC is the limited analysis procedures and design guidelines which simulate
the constitutive behaviour characterised by the hardening/softening phase after the first
cracking. This behaviour significantly differs from the brittle nature of plain concrete.
During the crack propagation and damage evolution, distributed micro-cracks are
initiated. Afterward, the micro-cracks are developed to the localised macro-cracks
which can be simulated with concrete damage evolution. However in this stage, the
fibre bridging governs the damage evolution which has to be modelled based on fibre
pulling out mechanism [11].
There are several attempts in developing numerical and mathematical models of
fibre reinforced concrete behaviour based on single fibre pull-out behaviour and ran-
domly distributed fibres, the latter of which considers a homogeneous material [11–14].
Out of the two mechanisms, the single fibre pull-out models become complex due to
its’ associated fibre orientation. Even though the constitutive models developed based
on homogeneous material requires experimental stress-strain relationship and crack
mouth opening displacement values, it can be considered as an industrially viable
analysis technique.
ABAQUS; a finite element software package offers a range of models, namely,
concrete damage plasticity (CDP), Drucker-Prager, and Cap Plasticity in such analysis.
However, these models fail to incorporate the hardening after a drop of tensile strength
in the aspect of damage evolution. ATENA; another heavily used software package in
the industry is capable of capturing this behaviour, yet fail to simulate the impact
response of the MSFRC. Contrarily, ABAQUS lately provided a platform for impact
analysis which is termed as dynamic explicit analysis. Thus, the research presented
herein evaluates the adaptation of the CDP model of ABAQUS package, as it involves
the simple experimental procedure to determine the compressive and tensile behaviour.
Furthermore, the significance of the damage evolution parameters in the overall
structural performance of the MSFRC is also discussed.
2 Experimental Mechanical Properties
The mechanical properties of the plain concrete and fibre reinforced concrete are
obtained from the previous set of experiments carried out [15]. MSFRC samples with
the fibre volume ranging from 0.2% to 2% with two different lengths of fibres (i.e.
48 mm and 58 mm) were tested. However, fibre volume fraction of only 1% of 48 mm
550 D. K. Hewage et al.
fibre was selected to continue the study herein presented characterising the optimum
fibre dosage for the subjected polypropelene fibre.
2.1 Stress-Strain Relationship

The static compressive strength of the concrete samples was determined from the static
compressive strength test according to AS1012.9-2014 [16]. Subsequently, the stress-
strain relationship was obtained using the Eq. (1);

ec
rc b e0c
0 ¼ b ð1Þ
rc
b 1 þ eec0
c
0
where rc = maximum compressive strength, rc = concrete stress,ec = concrete strain,
0
ec . = strain corresponding to maximum stress and b. = material parameter [17]. The
tensile stress-strain relationship of the MSFRC was determined according to the fib
model code for concrete structures 2010 [18] using the test data of three-point bending
test which complies with EN14651-2005 [19]. The tensile stress-strain relationship of
the plain concrete (NC) and MSFRC are plotted in the Fig. 1.
6
Tensile Strength
4
(N/mm2)
0
0 0.005 0.01 0.015 0.02
Strain (mm/mm)
NC MSFRC
Fig. 1. Stress-strain relationship of plain concrete and MSFRC.
2.2 Damage Parameters

Damage parameter represents the level of damage at a given strain which without only
the behaviour of plasticity is simulated. Both compression and tension damage is
defined based on the ratio of inelastic/cracking strain to the total strain [20]. Due to the
intricacy of calculating the strain values, the damage parameters were obtained using
Eq. (2) and (3);
rc
dc ¼ 1 ð2Þ
r0c
rt
dt ¼ 1 ð3Þ
r0t
0
where, dc = compression damage parameter, rc = maximum compressive strength,
0
rc = concrete compressive stress,dt = tension damage parameter, rt = maximum ten-
sile strength and rt = concrete tensile stress
3 Finite Element Analysis
The impact behaviour of both plain concrete and MSFRC is evaluated using ABAQUS
dynamic explicit package. The analysis setup is exactly similar to the experiment
carried out by Erdem, et al. [21] allowing the model verification of plain concrete.
3.1 The Model

A concrete specimen of 150 mm 150 mm 710 mm was selected for the analysis.
Only weight of 8 kg was dropped at the height of 1000 mm in the experiment carried
out by Erdem, et al. [21]. However, the analysis was extended to weights of 16 kg and
32 kg to identify the behaviour under the impact failure. A steel plate and a neoprene
rubber layer have been used in the contact region to distribute the impact load uni-
formly across the top surface of the concrete. A fixed boundary condition for 50 mm
was applied from both the ends resulting in a clear span of 610 mm as displayed in
Fig. 2.
Impactor
Steel plate and

Neoprene rubber layer
Fig. 2. Finite element model setup.
The mesh type selected for the analysis was C3D8R which is commonly used by
the researches. As the impacts can result in excessive distortion, enhanced hourglass
controls were used for the mesh. A mesh element size of 25 mm was used for all the
parts in the model.
3.2 Sensitivity Analysis

Since the concrete damage plasticity (CDP) was selected as the material model, the
damage parameters govern the impact behaviour. Thus, 5 analysis cases were carried
out with and without damage parameter variables to evaluate the significance of the
damage parameters under an impact loading. Only plasticity was defined using the
stress-strain relationship for plain concrete (NC-P) and MSFRC (MS-P) for cases
without damage parameters. Both compression and tension damage was defined in the
rest of the cases (NC-D, MS-D1, MS-D2). The compression damage is similar for all
three cases, which was calculated using Eq. (2) and the tension damage was determined
using Eq. (3).
The MSFRC material shows a hardening effect after the first drop as displayed in
the Fig. 1, resulting in a decrease in the damage parameter with the increase of strain.
This decrease in the damage parameter is not supported by the ABAQUS program.
Thus the damage parameter for MS-D1 was calculated based on the tensile stress at the
drop and kept constant throughout. MS-D2 was defined using the stress at the strain of
0.02 and kept constant after the first crack in order to evaluate the significance of the
hardening effect in an impact loading. Table 1 summarizes the material models
adopted.
Table 1. Damage parameter for different cases.

Cracking strain Damage parameter
NC-P NC-D MS-P MS-D1 MS-D2
0 No Damage Parameter 0 No Damage Parameter 0 0
0.0005 0.9 0.7 0.5
0.004 0.95 0.7 0.5
0.02 0.98 0.7 0.5
4.1 Failure Load and Concrete Stresses

Under 8 kg, both the specimens have not failed, yet higher stresses compared to
MSFRC sample are depicted in the plain concrete sample at the support and the region
of the impact as seen in Fig. 3. However, under 16 kg, plain concrete beam displays a
crushing failure, where MSFRC beam has not reached any failure conditions resulting
in a lesser load transferring to the supports. Thus it can be concluded that MSFRC
elements are capable of sustaining a higher impact load. Both the specimens have failed
under the weight of 32 kg displaying a crushing failure as can be seen in Fig. 3.
8 kg
NC-D MS-D1
16 kg
NC-D MS-D1
32 kg
NC-D MS-D1
Fig. 3. Stress distribution of plain concrete and MSFRC at the impact
As von-mises stresses in an element cannot be experimentally evaluated under an

impact, obtaining the ultimate impact resistance of a material is highly debatable. Thus
far, ultimate impact resistance has been obtained by visual inspection of the failure of
different trial impacts, incurring a consequential time and financial loss. Accordingly, a
similar finite element analysis can be developed to evaluate the impact resistance under
defined maximum stresses or strain criteria. Defining such strain criteria for different
applications subsequently results in user-friendly designing procedure of MSFRC
elements under impact loading.
4.2 Effect on Impact Force

The impact force generated has been calculated using the acceleration of the impactor.
As seen in Fig. 4, the generated impact force significantly increases with the weight of
the impactor. However, the impact force has not significantly changed with the
introduction of the fibres under the 8 kg and 16 kg impactor loads. Only under the
32 kg of impactor load, a maximum reduction of 6% can be noticed. Thus a similar
loading protocols can be used for the experimental comparison of the impact behaviour
of plain concrete and MSFRC.
160
Impact Force (kN)

120
8 kg
80
16 kg
40 32 kg
0
Material Model
Fig. 4. Effect of MSFRC on the impact force
4.3 Effect on Residual Deflection

The deflection during the impact of all the cases under three impactor loads can be seen
in Fig. 5, 6 and 7. The finite element analysis depicts a higher initial and residual
deflection in the plain concrete samples compared to MSFRC under similar loading
conditions. Thus the MSFRC is absorbing higher energy without transferring stresses
to the supports in an impact. Thus the numerical or analytical models which simulate
the impacts on MSFRC should be modified to incorporate the energy absorption at the
time of contact.
0
0.44 0.48 0.52 0.56 0.6 0.64
Deflection (mm)
-0.2
-0.4
-0.6
Time (s)
Fig. 5. Effect of MSFRC residual deflection under 8 kg.
Before the failure load, a significantly higher reduction of residual deflection can be
noticed. i.e. under the impact load of 8 kg, the reduction of maximum deflection is only
20% where the reduction of the residual deflection is 33%. Under the impact load of
0
0.44 0.48 0.52 0.56 0.6 0.64
Deflection (mm) -0.2
-0.4
-0.6
-0.8
-1
Time (s)
0
0.44 0.48 0.52 0.56 0.6 0.64
-0.4
Deflection (mm)
-0.8
-1.2
-1.6
-2
Time (s)
16 kg, the reduction of initial and residual deflection is 13% and 27% respectively as
displayed in Fig. 6. Accordingly, it can be concluded that the fibres are contributing to
deflection recovery in MSFRC. Thus the progressive collapse under repeated impacts is
curtailed in MSFRC. However, the initial and residual deflection reduction is similar
beyond the failure impact load (under 32 kg) as seen in Fig. 7.
4.4 Effect of Damage Parameter Values

As seen in Fig. 5, the significance of defining the damage parameter is comparatively
low when the material has not reached the failure phase. However, beyond the failure, a
clear influence of defining damage parameters can be observed in both plain concrete
and MSFRC. In addition, no difference in the analysis results of deflection and the
impact force is seen between the MS-D1 and MS-D2. Accordingly, a 20% deviation of
the tension damage parameter values results in insignificant differential behaviour of
MSFRC under impact loads. Thus the tension damage parameter in such situations can
be defined using the maximum and minimum stress value of the stress-strain behaviour,
keeping it constant throughout.
5 Conclusions
From the finite element study of the impact behaviour of macro-synthetic fibre rein-
forced concrete, the emphasis on the following conclusions can be made:
• Impact stresses – The stresses generated in MSFRC under an impact is lesser
compared to the plain concrete, thus resulting in higher impact resistance.
• Impact force – The change of impact force exerted on the element is insignificant
with the introduction of macro-synthetic fibres thus enabling the utilisation of the
similar loading protocol in comparative experimental studies.
• Initial and residual deflection – Both initial and residual deflection is reduced in
MSFRC compared to plain concrete with a higher reduction of residual deflection
prior to the failure of MSFRC. Accordingly, a beneficial effect on the progressive
collapse under repeated impact loads can be achieved.
• Significance of the damage parameter values – Beyond the failure, defining the
damage parameter results in a significant change of the impact behaviour of plain
concrete and MSFRC. However, a 20% deviation of tension damage parameter
values is insignificant with respect to impact force and deflection.
Acknowledgements. The authors gratefully acknowledge the Australian Research Council’s

Industrial Transformation Training Centres Scheme (ARC Training Centre for Advanced
Technologies in Rail Track infrastructure; IC170100006) which provided the catalyst for
undertaking this research as well as Western Sydney University through the School of Com-
puting, Engineering and Mathematics for the support given to complete the work described
herein.
References
1. Costa, C., Cerqueira, Â., Rocha, F., Velosa, A.: The sustainability of adobe construction:
past to future. Int. J. Archit. Heritage 13, 1–9 (2018)
2. Di Maida, P., Sciancalepore, C., Radi, E., Bondioli, F.: Effects of nano-silica treatment on
the flexural post cracking behaviour of polypropylene macro-synthetic fibre reinforced
concrete. Mech. Res. Commun. 88, 12–18 (2018)
3. Ghadban, A.A., Wehbe, N.I., Underberg, M.: Effect of fiber type and dosage on flexural
performance of fiber-reinforced concrete for highway bridges. ACI Mater. J. 115(3), 413–
424 (2018)
4. Li, J., Wu, C., Liu, Z.-X.: Comparative evaluation of steel wire mesh, steel fibre and high
performance polyethylene fibre reinforced concrete slabs in blast tests. Thin Walled Struct.
126, 117–126 (2018)
5. Al-Masoodi, A.H.H., Kawan, A., Kasmuri, M., Hamid, R., Khan, M.N.N.: Static and
dynamic properties of concrete with different types and shapes of fibrous reinforcement.
6. Hao, Y., Hao, H.: Dynamic compressive behaviour of spiral steel fibre reinforced concrete in
split Hopkinson pressure bar tests. Constr. Build. Mater. 48, 521–532 (2013)
7. Xu, Z., Hao, H., Li, H.N.: Experimental study of dynamic compressive properties of fibre
reinforced concrete material with different fibres. Mater. Des. 33, 42–55 (2012)
8. Yao, W., Li, J., Wu, K.: Mechanical properties of hybrid fiber-reinforced concrete at low
fiber volume fraction. Cem. Concr. Res. 33(1), 27–30 (2003)
9. Hwang, J., Jung, M.S., Kim, M., Ann, K.Y.: Corrosion risk of steel fibre in concrete. Constr.
Build. Mater. 101, 239–245 (2015)
reinforced concrete: influence of crack width, fibre type and loading conditions. Corros. Sci.
98(C), 128–139 (2015)
11. Manca, M., Karrech, A., Dight, P., Ciancio, D.: Dual cohesive elements for 3D modelling of
synthetic fibre-reinforced concrete. Eng. Struct. 174, 851 (2018)
12. Abbas, A.A., Syed Mohsin, S.M., Cotsovos, D.M.: A simplified finite element model for
assessing steel fibre reinforced concrete structural performance. Comput. Struct. 173, 31–49
(2016)
13. Awinda, K., Chen, J., Barnett, S., Fox, D.: Modelling behaviour of ultra high performance
fibre reinforced concrete. Adv. Appl. Ceram. 113(8), 502–508 (2014)
14. Othman, H., Marzouk, H.: Applicability of damage plasticity constitutive model for ultra-
high performance fibre-reinforced concrete under impact loads. Int. J. Impact Eng 114, 20
(2018)
15. Camille, C., Kahagala Hewage, D., Mirza, O., Mashiri, F., Kirkland, B., Clarke, T.:
Mechanical properties of macro-synthetic fibre reinforced concrete under static loading for
railway transoms application. In: Presented at the 9th Australian Small Bridges Conference
2019, Gold Coast, Queensland (2019)
16. Methods of testing concrete Compressive strength tests - Concrete, mortar and grout
specimens, AS1012.9-2014 (2014)
17. Hsu, L.S., Hsu, C.T.T.: Complete stress-strain behaviour of high-strength concrete under
compression. Magazine of Concrete Research 46(169), 301–312 (1994)
18. C. International Federation for Structural, Ed. Model Code 2010: final draft,
(Bulletin/International Federation for Structural Concrete Draft model code, no. 65), vol.
1, p. 311. International Federation for Structural Concrete, Lausanne (2012)
19. BS EN 14651:2005: Test method for metallic fibre concrete. Measuring the flexural tensile
strength (limit of proportionality (LOP), residual) (2005)
20. Wahalathantri, B.L., Thambiratnam, D.P., Chan, T.H.T., Fawzia, S.: A material model for
flexural crack simulation in reinforced concrete elements using ABAQUS. Presented at the
Proceedings of the First International Conference on Engineering, Designing and
Developing the Built Environment for Sustainable Wellbeing, Queensland University of
Technology, Brisbane, Qld (2011). https://eprints.qut.edu.au/41712/
21. Erdem, R.T., Engin, G., Erkan, K., Muhiddin, B.: Impact behaviour of concrete beams.
Građevinar 66(6), 523–531 (2014)
Finite Element Analysis of Ultra High
Performance Fibre Reinforced Concrete
Beams Using Microplane Modelling
William. Wilson1,2(&) and Tomas O’Flaherty1,3

1
Department of Civil Engineering and Construction, Institute of Technology
Sligo, Sligo, Republic of Ireland
wwilson@roadstone.ie
2
Roadstone Ltd, Fortunestown, Tallaght Dublin 24, Republic of Ireland
3
Centre for Environmental Research Innovation and Sustainability, Institute of
Technology Sligo, Sligo, Republic of Ireland
Abstract. Ultra high performance fibre reinforced concrete (UHPFRC) exhibits

enhanced strength, ductility and durability properties in comparison to con-
ventional and high performance concretes. This research uses results from
experimental tests on small scale UHPFRC beams to validate three-dimensional
finite element modelling methods developed for concrete beams using micro-
plane theory. The effect of 0%, 1% and 2% fibre volumes were investigated on
beams with varying cross-section. Beams with and without conventional steel
reinforcement were tested in four point bending. The numerical models accu-
rately replicated the behaviour of the experimental specimens. The load-
deflection and failure modes in the numerical beam models provided an accurate
representation of the experimental behaviour, including the observed change in
failure mode from shear in the ultra high performance concrete (UHPC) beams
with conventional steel reinforcement to a flexural failure in their UHPFRC
counterparts.
Keywords: UHPFRC Finite element modelling Microplane ANSYS

Structural elements
1 Introduction
Ultra high performance fibre reinforced concrete (UHPFRC) is characterised by sig-

nificantly improved mechanical performance, extremely low permeability and tensile
strain hardening. Fibres prevent explosive failure, improve the tensile strength, and
make it possible to achieve higher levels of ductility [1]. UHPFRCs are likely to have a
compressive strength of 130 to 230 MPa while the flexural tensile strength is usually
between 25 and 60 MPa depending on the early age curing methods used [2]. The
enhanced strength of UHPFRC results in thinner sections in comparison to conven-
tional concrete.
To date most experimental studies on UHPFRC reported in the literature have
concentrated on obtaining the characteristic strengths of a single UHPFRC mix. In
addition, limited experimental results are available on the behaviour of beams in
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_51
Finite Element Analysis of Ultra High Performance Fibre 559
bending and in shear. The design of conventional concrete structures is well understood
and relatively straightforward once the concrete compressive and steel yielding
strengths are known. This is not the case for UHPFRC structures, as the material
behaviour is significantly different. Recommendations have been proposed by AFGC
[3] and the JSCE [4] to predict the flexural and shear strength of UHPC and UHPFRC
beams. However, these methods deal with only one mix strength and fibre dosage and
cannot be applied to a range of strengths or fibre dosages.
Although the limited number of experimental studies on the behaviour of UHPFRC
beams have aided in advancing our understanding, comparisons between experimental
and numerical behaviour are limited. Accurate models of UHPFRC are required to
ensure that the cost of projects can be reduced by eliminating the requirement of full
scale experimental tests. A substantial increase in the body of knowledge at a fraction
of the cost will be obtained by bridging the gap between experimental testing and
numerical modelling. Numerical modelling can also play an important role in the
development of design guidelines for UHPFRC structural elements. However, its
applicability is limited to a certain extent due to the lack of a universally accepted
concrete material model that could be used to model all types of reinforced normal
strength concrete (NSC) and UHPFRC structural members. Therefore, the purpose of
this research was to develop numerical models to determine the practicability and
accuracy of using the finite element method (FEM) to analyse the behaviour of
UHPFRC beams.
2 UHPFRC Mix
In this research, a UHPFRC mix consisting of fine sand as the only aggregate,
developed as part of a previous study [5] is further investigated. The final mix designs
used in this research are presented in Table 1. In addition, Table 1 presents the asso-
ciated strength properties for each mix, which were determined from the average test
results of at least three specimens. CEM I Class 42.5R was used due to its high rate of
strength development. Elkem 920D microsilica was used as it improves the early age
and final strengths, density and durability of concrete. A fine sand with a
10 lm/550 lm grading was used. A polycarboxylate polymer superplasticiser specif-
ically used to develop high early age strength, high water reductions and excellent
flowability was the admixture used. Dramix OL 13/.20 steel fibres with a length of
13 mm, a diameter of 0.20 mm, and a tensile strength of 2600 MPa were used.
The compressive strength, and associated stress-strain relationship and elastic
modulus was determined using 100 mm diameter by 200 mm high cylindrical speci-
mens. The loading rate up to 85% of the expected failure load was set at 0.5 MPa/s.
Loading then switched to displacement control using three linear variable displacement
transducers (LVDT’s) and loading continued at a rate of 1 lm/s. Therefore, the peak
and post-peak behaviour was accurately recorded. Tensile strength and behaviour was
determined using specially developed dogbone specimens with circular cross sections
that reduced from a diameter of 120 mm to 55 mm with a total throat length of
250 mm. The specimen shape ensured that fracture occurred away from the grips and
stress concentrations at changes in cross-section were minimised. Loading was
560 W. Wilson and T. O’Flaherty
controlled using displacement and preliminary test results and previous literature [6, 7]
indicated that the most suitable loading rate was 0.4 mm/min as a lower loading rate
induced creep and a higher loading rate resulted in premature failure.
Table 1. Material quantities and strength properties in each mix.

Mix UHPC-F UHPFRC-1F UHPFRC-2F
Cement (kg/m3) 810 810 810
MicroSilica (kg/m3) 203 203 203
Fine Sand (kg/m3) 1022 1022 1022
Admixture (l/m3) 42 42 42
Water (l/m3) 178 178 178
Steel Fibres (kg/m3) 0 (0%) 77.5 (1%) 155 (2%)
Cylinder Compression Strength (MPa) 126 130 147
Direct-Tensile Strength (MPa) 10.5 10.8 11.6
Elastic Modulus (GPa) 44.1 45.0 46.2
Table 2 and Fig. 1 present details of the 12 beams tested as part of the experimental
program. Fibre volumes of 0%, 1% and 2% were used to investigate the effect of fibre
volume on the structural behaviour. Two different beam sizes were selected to deter-
mine if size effects existed in UHPC and UHPFRC. Due to testing machine limitations
the maximum beam height was selected as 150 mm. All beams had a span of 1500 mm
and a height to width ratio of 1.5, which is the ratio typically used in UHPFRC beams
[8]. The concrete was poured from one end of the beam form and allowed to flow to the
other end until the form was full, while being appropriately compacted with a poker
vibrator, to ensure uniform fibre distribution.
Beams were loaded in four point bending. Loading points were located only
200 mm apart at the midspan of the beam and the distance from each support to loading
point was 650 mm. The loading rate was displacement controlled using two LVDTs
that were attached to the loading frame and measured deflection at the beam’s mid-
point. It was determined that the 120 mm and 150 mm deep beams should be displaced
at a rate of 3 lm/s and 5 lm/s respectively.
Table 2. Details of the 12 experimental beam specimens

Beam name Fibre Beam height Rebar Beam name Fibre Beam height Rebar
volume (%) (mm) ratio (%) volume (%) (mm) ratio (%)
UHPC-F-120 0 120 0 UHPC-F-120R 0 120 1.3
UHPC-F-150 0 150 0 UHPC-F-150R 0 150 1.3
UHPFRC-1F-120 1 120 0 UHPFRC-1F-120R 1 120 1.3
(a) (b)
Fig. 1. (a) Beam cross-sectional reinforcement arrangements, and (b) Beam flexural test setup.
4.1 Beam Peak Loads

The first crack loads, peak loads and failure modes of the beams are presented in
Table 3. The results for the 12 beams demonstrate that the 1% and 2% fibres achieved
higher first crack and peak loads than the plain UHPC beams. Furthermore, the results
indicate that the UHPFRC beams with 2% fibres and no conventional reinforcement,
obtained peak loads only approximately 30% lower than that of the plain UHPC beams
with conventional reinforcement. This indicates that by using 2% fibres the require-
ments for conventional steel reinforcement could be significantly reduced.
Table 3. First crack loads, peak loads and failure modes of beam specimens.
Beam Name First crack Peak load Failure Beam name First crack Peak load Failure
(kN) (kN) mode (kN) (kN) mode
UHPC-F-120 – 8.8 Flexure UHPC-F-120R 7.1 22.7 Shear
UHPC-F-150 – 16.5 Flexure UHPC-F-150R 14.9 40.0 Shear
UHPFRC-1F-120 9.2 11.4 Flexure UHPFRC-1F-120R 8.1 28.2 Flexure
4.2 Cracking Pattern and Failure Mode

The 120 mm and 150 mm deep beams exhibited similar failure patterns for all beam
types. The two plain beams, with no fibres or conventional reinforcement, failed due to
one single flexural crack at midspan. The four beams with fibre reinforcement but no
conventional reinforcement initially cracked in a similar location to the plain beams.
The steel fibres bridged the initial crack and allowed the beams to support additional
load. As loading continued and deflection increased these cracks increased in size and
further micro-cracks developed. The micro-cracks did not appear to open significantly
but continued to propagate upwards until failure occurred when one crack eventually
opened up and fibre pullout occurred. For the UHPC beams with conventional rein-
forcement, initial cracking occurred on the bottom face of the beam near midspan, as
expected. As a result of the improved flexural resistance, a sudden shear failure was
observed with a crack developing at the support and running along the rebar and then
diagonally towards the loading point, as shown in Fig. 2 (a), as no shear links were
provided. The addition of 2% steel fibres to these beams caused the failure mode to
change from a shear failure to a flexural failure, as highlighted in Fig. 2 (b), with. This
clearly demonstrates how steel fibres in UHPFRC could be used in place of conven-
tional shear links to prevent a sudden shear failure.
(a)
(b)
Fig. 2. Failure modes for beams (a) UHPC-F-150R failed in shear, and (b) UHPFRC-2F-150R
failed in flexure.
4.3 Load-Deflection Relationship

The load-deflection curves for the plain UHPC-F beams are displayed in Fig. 3 (a). The
UHPC-F specimens with secondary steel bar reinforcement experienced a sudden shear
failure. The shear failure can be attributed to the significantly increased tensile and
compressive stress capacity of UHPC. Although flexural cracks were visible at midspan
in the beams a change in stiffness denoting steel yielding was not apparent in the load-
deflection curves indicating failure occurred before steel yielding.
The load-deflection curves for the UHPFRC-1F beams are given in Fig. 3 (b). After
first crack, the two 1% steel fibre specimens without conventional reinforcement
underwent large rotations with only small increases in load. In both cases, once peak
load was reached a gradual decrease in the load carrying capacity occurred until the
opening of a macrocrack and fibre pullout resulted in flexural failure. The inclusion of
conventional reinforcement increased the peak load and significantly improved the
stiffness after first crack. For both the UHPFRC-1F-R specimens the load-deflection
curves indicate that steel yielding was followed by a minor increase in load before
flexural failure occurred. The load-deflection curves for the UHPFRC-2F beams are
given in Fig. 3 (c). The behaviour of the beams with 2% fibres is comparable to the
beams with 1% fibres. The main differences being both the increase in stiffness after
first crack and peak load achieved by the 2% fibre beams.
UHPC-F-150R UHPFRC-1F-150R UHPFRC-2F-150R

UHPC-F-120R UHPFRC-1F-120R UHPFRC-2F-120R
UHPC-F-150 UHPFRC-1F-150 UHPFRC-2F-150
UHPC-F-120 UHPFRC-1F-120 UHPFRC-2F-120
75 75 75
50 50 50
Load (kN)
Load (kN)
Load (kN)
25 25 25
0 0 0
0 10 20 30 40 0 10 20 30 40 0 10 20 30 40
Deflection (mm) Deflection (mm) Deflection (mm)
(a) (b) (c)
Fig. 3. Beam load-deflection curves for the (a) 0% fibre specimens, (b) 1% fibre specimens, and
(c) 2% fibre specimens.
5 Numerical Simulation
Several approaches are available for defining the stress-strain behaviour of concrete
under different states of stresses. One of the most popular methods is the constitutive
model for the triaxial behaviour of concrete developed by Williams and Warnke [5],
which can be used with the smeared crack approach developed using the smeared crack
constitutive model [6]. This type of model has previously been shown to be accurate at
modelling UHPFRC structures [7]. However, modelling the post-peak softening
behaviour of concrete, in particular UHPFRC, can be problematic.
Microplane modelling offers improvements in simulating the behaviour of non-
linear materials as it is particularly suited to replicating damage in materials, such as
cracking in concrete [8]. The microplane model was first developed in the 1980s based
on theory in which the material behaviour is modelled using uniaxial stress-strain laws
on different planes known as microplanes [9, 10]. The constitute law for a microplane
model is characterised in terms of stresses and strains in which the normal components
rN and eN are split into their normal and deviatoric parts, which can then be used to
define the incremental elastic relations [11]. Although microplane theory and a number
of different microplane model types have been developed that are suitable for mod-
elling concrete [9], it is difficult to find examples in the literature of microplane theory
applied to concrete. Therefore, in this research microplane theory was used to model
the behaviour of the concrete beams up to peak stress and their degradation post-peak
stress using ANSYS [10].
5.1 Model Parameters

The SOLID185 element and the microplane material model [10] were utilised to model
the concrete material behaviour. ANSYS has incorporated the microplane theory
previously discussed [11] and is defined using Young’s modulus, Poisson’s Ratio and
the six constants presented in Table 4. For concrete the damage function constants
k0, k1 and k2 are calculated using equivalent strain expressions given by de Vree et al.
[12] and Geers et al. [13]. The critical equivalent-strain-energy density parameter is
used to determine the tensile strength of the material. amic determines the residual
strength of the material and its value ranges from 0 to 1, where 0 implies no damage
and 1 implies complete damage. bmic determines the rate of material degradation, which
dictates how quickly material damage will propagate with increasing load. The
microplane model follows exponential decay so the higher the value of bmic the faster
degradation occurs. A parametric study of concretes ranging in strength from 20 MPa
to 200 MPa was used to determine suitable values for amic and bmic.
Table 4. Inputted concrete material properties for the FE microplane models

Constant Symbol Definition Inputted value
C1 k0 Damage function constant 0.7803
C4 cmic
0 Critical equivalent-strain-energy density 0.238 10−3
C5 amic Maximum damage parameter 0.95
C6 bmic Scale for rate of damage 700
REINF264 was used to model the fibre reinforcement, which was embedded in the
SOLID185 concrete element. BEAM188 was used to model the cross-sectional area of
the conventional steel reinforcement as a discrete material. The element is based on
Timoshenko beam theory that includes shear-deformation effects. The material prop-
erties for the steel fibres and the steel bar reinforcement were inputted using a bilinear
kinematic hardening curve, in which the Young’s Modulus, Poisson’s Ratio, yield
stress and tangent modulus were specified as given in Table 5. In addition, as fibres in
UHPFRC typically pullout before yielding occurs a pullout stress was inputted for
fibres in tension based on historical data.
Table 5. Inputted steel fibre and steel bar reinforcement properties.

Material Young’s Poisson’s Yield Stress Pullout Stress Tangent
Modulus (GPa) Ratio (MPa) (MPa) Modulus (GPa)
Steel Fibre 200 0.3 2600 650 0
Steel 200 0.3 555 – 2000
Reinforcement
5.2 Model Arrangement

To reduce computational solution times one-quarter models were used. A mesh sen-
sitivity analysis was conducted by solving the models using a coarse mesh, which had
the lowest quantity of elements possible without using irregular shaped elements. The
models were then reanalysed with the number of elements doubled. If the two meshes
yielded substantially different results, then further mesh refinement was required. This
process was repeated until nearly identical results were obtained for successive models.
The mesh used in the current study is shown in Fig. 4.
Force control loading was utilised as this prevented large decreases in load due to
cracking that occurs with displacement control loading, which is a numerical
idiosyncrasy rather than being reflective of the experimental behaviour. As force
control loading was used the models were not capable of replicating the post-peak
behaviour of the experimental specimens. The force was applied over the flat surface of
the semi-circular loading roller on the UHPFRC beam models. The supports could
move in the longitudinal direction and were constrained in the vertical direction. This
allowed the models to rotate about the support, and as the support was free to move in
the longitudinal direction the same effect was obtained as for the four point bending
experimental test setup.
Fig. 4. Mesh arrangement for one quarter model of the beams.
The fibres were inputted in three directions, x, y and z. It was observed from the
experimental beam specimens that a significant majority of fibres aligned with the
direction of flow. Therefore, 80% the fibres were inputted in the direction parallel to the
span and the remaining 20% evenly distributed between the beam depth and width.
6 Numerical Results
Table 6 presents the first crack and peak loads of the 12 numerical models and their
experimental counterparts. The majority of the beam models underestimated the first
crack load of the experimental specimens with an average difference of 10%. Some of
the differences between the experimental and model first crack loads may be related to
issues in distinguishing the first crack in the experimental specimens, as these were not
entirely clear during testing. The average difference between the peak loads of the
experimental specimens and beam models was 8%. The largest difference in the
experimental and numerical peak load results occurred in the plain concrete beams
because the first cracking load of the numerical models was considerably less than that
of the experimental specimens.
For the beams with fibres and/or conventional reinforcement there is improved
agreement between the models and experimental specimens as the first crack load has a
reduced effect on the peak load achieved by the models. This is because it is more
dependent on the fibre volume and steel reinforcement strength rather than the concrete
tensile strength. The results highlight that as fibre volume increases the difference in
peak load of the experimental and numerical models reduces.
Table 6. Comparison of experimental and FEM results.

First Crack Load Peak Load
Beam Name Exp FEM Difference Exp FEM Difference
(kN) (kN) (%) (kN) (kN) (%)
UHPC-F-120 – – – 8.8 7.6 14
UHPC-F-120R – – – 16.5 12.6 24
UHPC-F-150 7.1 8.6 21 22.7 18.4 19
UHPC-F-150R 14.9 13.9 7 40 35.9 10
UHPFRC-1F-120 9.2 7.6 17 11.4 9.3 18
UHPFRC-1F-120R 15.4 13 16 22.6 20 12
UHPFRC-1F-150 8.1 8.6 6 28.2 26.1 7
UHPFRC-1F-150R 15.8 14.3 9 55.9 52.4 6
UHPFRC-2F-120 8.5 8.4 1 14.9 14.9 0
UHPFRC-2F-120R 14.6 15.2 4 30.1 31.2 4
UHPFRC-2F-150 8.8 8.7 1 33.8 33 2
UHPFRC-2F-150R 16.8 14.7 13 64.7 64.7 0
Figure 5 presents a comparison of the load-deflection behaviour of the experi-

mental beam specimens and numerical models. Most models exhibited accurate
agreement with the experimental specimens in the initial linear loading phase but some
models were slightly stiffer. There are two reasons for why this occurs relating to
microcracks existing in the experimental specimens before loading commences and
slippage of the reinforcement in the experimental specimens.
In the models with no fibres or conventional reinforcement shown in Fig. 5 (a),
failure of the beams occurred after first crack or damage was initiated due to the rapid
upwards propagation of the crack. For the UHPC-F-R beams, Fig. 5 (b) highlights that
after first crack the behaviour of the experimental specimens and numerical models
diverges. Firstly, the numerical models were stiffer after first crack. Secondly, the
experimental specimens failed due to shear and the load-deflection behaviour did not
exhibit steel yielding, but both numerical models exhibited a certain amount of steel
yielding. However, the models did exhibit high shear stresses indicating shear was part
of the failure mode. As the numerical models exhibited steel yielding and flexural
cracks were visible in the experimental specimens, there is a possibility that steel
yielding may have occurred to some extent in the experimental specimens. However, as
experimental failure occurred suddenly the LVDTs did not record a change in stiffness.
Flexural cracks and the area of damage had also propagated towards the top surface of
the damage models.
25 50
UHPC-F-150 Exp UHPC-F-150R Exp
UHPC-F-150 FEM UHPC-F-150R FEM
20 UHPC-F-120 Exp 40 UHPC-F-120R Exp
UHPC-F-120 FEM UHPC-F-120R FEM
Load (kN)
Load (kN)
15 30
10 20
5 10
0 0
0.0 0.5 1.0 1.5 0 10 20 30
(a) (b)
50 100
UHPFRC-1F-150 Exp UHPFRC-1F-150R Exp
UHPFRC-1F-150 FEM UHPFRC-1F-150R FEM
40 80
Load (kN)
UHPFRC-1F-120 FEM Load (kN) UHPFRC-1F-120R FEM

30 60
20 40
10 20
0 0
0 10 20 30 0 15 30 45
(c) (d)
50 100
UHPFRC-2F-150 FEM UHPFRC-2F-150R FEM
40 UHPFRC-2F-120 Exp 80 UHPFRC-2F-120R Exp
UHPFRC-2F-120 FEM UHPFRC-2f-120R FEM
Load (kN)
Load (kN)
30 60
20 40
10 20
0 0
0 10 20 30 0 15 30 45
(e) (f)
Fig. 5. Comparison of experimental and numerical beam load-deflection curves, (a) UHPC-F,
(b) UHPC-F-R, (c) UHPFRC-1F, (d) UHPFRC-1F-R, (e) UHPFRC-2F, and (f) UHPFRC-2F-R.
For the UHPFRC-1F models without steel bar reinforcement, first crack occurred at
or near midspan. Stress redistribution resulted in the tensile stress capacity being
exceeded outside the region of constant moment where a second crack developed. As
loading increased failure occurred due to the fibres on the bottom surface near midspan
reaching the specified pull-out stress, resulting in large increases in deflection and
rotation. In Fig. 5 (c) it is clear that although the models of the UHPFRC-1F beams
supported lower peak loads than the experimental specimens, the post-first cracking
stiffness of the numerical and experimental specimens were in good agreement. The
lower numerical peak loads can be attributed to the rapid propagation of cracks in the
models in comparison to the experimental specimens. Figure 5 (d) presents the load-
deflection curves for the UHPFRC-1F-R beams. The models had lower steel yielding
loads in comparison to their experimental counterparts. This resulted in these models
also having lower peak loads in comparison to the experimental specimens. However,
behaviour for both the experimental and numerical specimens was comparable.
The behaviour of the UHPFRC-2F beam models was similar to the UHPFRC-1F
beam models except that the UHPFRC-2F models exhibited more extensive cracking
and damage at failure, similar to what was observed in the experimental specimens.
Good agreement occurred between the experimental specimens and beam models for
the UHPFRC-2F and UHPFRC-2F-R beams, as shown in Figs. 5 (e) and (f), respec-
tively. The inclusion of steel reinforcement resulted in essentially the entirety of
UHPFRC-1F-R and UHPFRC-2F-R beam models undergoing cracking. Numerical
plots indicated that failure occurred as a result of the extensive cracking damage and
eventual fibre pullout. The damage locations were more tightly spaced in comparison to
the UHPC-F-R models, which was a result of the steel fibres supporting additional load
and allowing further crack propagation. This demonstrates how it is possible to
numerically replicate increased redistribution of stresses in UHPFRC through
microcracking.
From the results presented it is clear that the beam models are in good agreement
with the experimental results. In general, the lower first cracking loads of the models
resulted in some differences in terms of loading at first crack, steel yielding and peak.
The load-deflection and failure modes observed in the numerical models were an
accurate representation of the experimental behaviour. The major difference noted was
for the UHPC-F-R beams models that exhibited steel yielding before shear failure
occurred but the experimental specimens did not clearly exhibit any steel yielding. The
numerical models replicated the ability of UHPFRC to redistribute stress due to micro
cracking.
7 Conclusions
This paper presents the findings of a study on the validation of the microplane material
model to determine the behaviour of UHPFRC beams as an alternative to experimental
work. The following conclusions can be drawn based on the results:
• UHPC beams with conventional longitudinal steel bar reinforcement and no shear
links can exhibit a sudden shear failure due to the increased tensile and shear
capacity of UHPC.
• UHPFRC beams with and without conventional reinforcement failed in flexure,
which illustrates that steel fibres can prevent shear failure occurring and could be
considered as a substitute for conventional shear web reinforcement.
• The numerical models provided excellent agreement to the experimental specimens
with an average difference in peak load of 8%.
• The numerical models demonstrated the change in failure mode from shear to
flexure in the steel bar reinforced beams when steel fibres were included.
• The numerical models demonstrated the ability of UHPFRC to redistribute stress
through microcracking, an important feature of the material.
The conclusions from this research will be applied to future UHPFRC numerical
modelling problems and aid in the development of design codes and guidelines
Acknowledgements. The authors appreciatively acknowledge Banagher Precast, Bekaert,

Kerrigan’s Quarry, Irish Cement, Sika Ireland and Roadstone Ltd for providing materials for this
research. The work was financially supported by the Institute of Technology, Sligo President’s
Bursary Award.
References
1. Kim, D.J., Park, S.H., Ryu, G.S., Koh, K.T.: Comparative flexural behavior of hybrid ultra
high performance fiber reinforced concrete with different macro fibers. Constr. Build. Mater.
25(11), 4144–4155 (2011)
2. Tam, C., Tam, V.W.: Microstructural behaviour of reactive powder concrete under different
heating regimes. Mag. Concr. Res. 64(3), 259–267 (2012)
3. Association Française de Génie Civil (AFGC), Ultra High Performance Fibre-Reinforced
Concretes: Interim Recommendations, Mariselle (2002)
4. Japan Society of Civil Engineers, Recommendations for design and construction of ultra-
high strength fiber reinforced concrete structures, draft (2004)
5. Wilson, W., O’Flaherty, T.: Ultra high performance fibre reinforced concrete for
infrastructure construction. In: The 9th International Concrete Conference 2016, pp. 489–
501 (2016)
6. Phillips, D.V., Binsheng, Z.: Direct tension tests on notched and un-notched plain concrete
specimens. Mag. Concr. Res. 45(162), 25–35 (1993)
7. Hassan, A.M.T., Jones, S.W., Mahmud, G.H.: Experimental test methods to determine the
uniaxial tensile and compressive behaviour of ultra high performance fibre reinforced
concrete (UHPFRC). Constr. Build. Mater. 37, 874–882 (2012)
8. Yang, I.H., Joh, C., Kim, B.-S.: Structural behavior of ultra high performance concrete
beams subjected to bending. Eng. Struct. 32(11), 3478–3487 (2010)
9. Caner, F.C., Bazant, Z.P., Wendner, R.: Microplane model M7f for fiber reinforced concrete.
Eng. Fract. Mech. 105, 41–57 (2013)
10. ANSYS Inc., “Mechanical APDL Documentation (2013)
11. Bažant, Z.P., Gambarova, P.G.: Crack shear in concrete: crack band microplane model.
J. Struct. Eng. 110(9), 2015–2035 (1984)
12. de Vree, J.H.P., Brekelmans, W.A.M., van Gils, M.A.J.: Comparison of nonlocal continuum
damage. Comput. Struct. 55(4), 581–588 (1995)
13. Geers, M.G.D., de Borst, R., Peerlings, R.H.J.: Damage and crack modeling in single-edge
and double-edge notched concrete beams. Eng. Fract. Mech. 65(2–3), 247–261 (2000)
Machine Learning Prediction of Flexural
Behavior of UHPFRC
Joaquín Abellán-García1,2(&), Jaime A. Fernández-Gómez1,

Nancy Torres-Castellanos2, and Andrés M. Núñez-López3
1
Department of Civil Engineering, Polytechnic University of Madrid (UPM),
Madrid, Spain
j.abellang@alumnos.upm.es
2
Escuela Colombiana de Ingeniería Julio Garavito, Bogotá, Colombia
3
I + D + I Cementos Argos, Medellin, Colombia
Abstract. To evaluate the possibility of predicting the flexural behaviour of

UHPFRC, four analytical models were developed, based on artificial neural
networks (ANN), to predict the first cracking tension or Limit of Proportionality
(LOP), its corresponding deflection (dLOP), ultimate strength or Modulus of
Rupture (MOR), and its corresponding deflection (dMOR) of UHPFRC under
bending test. The models that were composed of an input level, one output level,
and four hidden levels were developed through the R platform. The input level
applied the most significative Principal Components (PC) of a large dimension
of input dataset. To avoid overfitting K-fold validation and l2 regularization was
used. After the models were created, an improvement based on assembling of
models by incorporating the predicted values in the dataset of features. The
results indicated that the developed assembling models have a good accuracy for
the prediction of the behaviour of UHPFRC under three or four points bending
test, even when containing supplementary cementitious materials and hybrid
mixture of fibers.
Keywords: UHPFRC LOP dLOP MOR dMOR Machine learning PCA

L2 regularization K-fold validation
1 Introduction
Over the last twenty years, remarkable advances have taken place in the research on
ultra-high-performance fiber reinforced concrete (UHFRPC). UHPFRC is a new type
high-tech concrete that is characterized by its ultra-high compressive strength, low
permeability, and improved durability, ductility and toughness [1–5]. However, the
incorporation of expensive components such as fibers and superplasticizer based on
polycarboxylate ether makes the price of this material higher than conventional con-
crete. Besides, due to the absence of a coarse aggregate together with its need of
achieving high packing density, UHPFRC contents of binder used (usually cement,
quartz sand, quartz powder and silica fume) are relatively higher in relation to con-
ventional concrete, increasing its cost and carbon footprint [6]. As consequence,
notwithstanding the outstanding performance in mechanical properties and durability,
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_52
Machine Learning Prediction of Flexural Behavior of UHPFRC 571
UHPFRC has not been widely applied in construction primarily due to these high costs
and carbon footprint. In agreement to that, on the one hand, great efforts have been
made lately in obtaining a less costly and eco-friendlier UHPC and UHPFRC, by
employing supplementary cementitious materials (SCM) as replacement for cement
(partial), quartz powder (total) and silica fume (partial or total) [7]. Accordingly, some
industrial by-products such as fly ash (FA), ground granulated blast slag furnace
(GBSF), rice husk ash (RHA), recycled glass powder (GP), and fluid catalytic cracking
residue (FC3R) aside from other SCM such limestone powder (LP) and metakaolin
(MK) have been used as components of UHPC’s binder. Inter alia, RHA was studied as
SCM in UHPC with successful results in mechanical properties and durability since its
high amorphous silica content and the internal-curing process of the paste [8]. Blending
limestone powder in UHPC’s binder was proven to lead and improvement of the
hydration process at the early-age, having as consequence higher particle packing
density and improving mechanical properties [9]. As well, nano-CaCO3 as a compo-
nent of binder was investigated by Huang and Cao [10]. The outcomes of their research
showed that compressive strength raised a 17% in comparison with UHPC control
specimens without nano-CaCO3. Abellan et al. [11] analysed the partial substitution of
cement for FC3R. Their results showed that the partial substitution produced an
increase in the compressive strength at early ages due to the high content in reactive
silica-aluminate of FC3R, while a decreasing in the slump. In other study they
demonstrated the possibility of producing an eco-friendly and low cost UHPC by using
two sizes of recycled glass powder, micro-limestone powder in addition to cement and
silica fume (the latter limited to 100 kg/m3) [12]. These researches demonstrated that is
possible to produce UHPC/UHPFRC by using SCM in partial replacement of cement
and silica, without significantly worsening its mechanical properties and durability.
On the other hand, regarding to the final cost of UHPFRC, fiber content has the
greatest weight in the final cost of the concrete. Abellán et al. [1] presented an average
mixture of UHPFRC, calculated from 150 dosages from scientific literature, which had
the following characteristics in common: minimum compressive strength of 150 MPa
with no special curing condition, maximum size of aggregate in the range of 0.5 and
0.6 mm, and 2% of steel fiber volume fraction. As can be observed in Fig. 1, a fiber
volume fraction of 2% of steel fiber renders a 37% of the final cost of the concrete
dosage. For that reason, lessening the fiber content or substitute the fibers for others
less expensive while maintaining an adequate mechanical behaviour is the main
challenge to lessen costs and permit a widespread use of UHPFRC.
Fig. 1. Average dosage of 150 dosages from the scientific literature. Components and its
implication in cost [1].
572 J. Abellán-García et al.
However, the key properties of UHPFRC containing hybrid blending fibers as well
as several SCMs require to be studied experimentally owing to the fact of the unclear
combination effects of these components. In addition, it is important to highlight that
most of the times laboratory tests are labour-intensive, time-consuming, and costly [3].
To reduce the experimental campaign and its associated costs and times, probabilistic
models can be opportunely used to forecast the mechanical properties of concrete.
However, these models cannot be applied when the approaching problem involves too
many factors and the inter-relations amidst them are too complex, unknown, or both [3,
4, 13]. In the case like the UHPFRC using several types of fibers and SCM, due to its
large amount of components, the use of traditional techniques of approach is unsuc-
cessful in achieving the expected precision [13].
In the last decades, Artificial Neural Network (ANN) applications have spread due
to its great ability to map non-linear and unknown relationships between input and
output data pairs [14, 15]. That is the reason why those algorithmic procedures had
proved its capability to help in problematic and complex engineering issues [14–16].
The main objective of this paper is to develop four models through the R language
[17] to forecast the behaviour of UHPFRC under three and four point bending test,
when concrete incorporates SCM and different types of fibers (even hybrid blending of
fibers). A single model will be created for each response, i.e., Limit of Proportionality
(LOP), its corresponding deflection (dLOP), ultimate strength or Modulus of Rupture
(MOR), and its corresponding deflection (dMOR). In order to consider an appropriate set
of inputs nodes for the artificial neural networks, principal components were used
instead a large amount of data. Besides, l2 regularization was used to avoid overfitting.
A different parameter l2 is used for each of the four hidden layers of de models, and its
value is adjusted by using the k-fold validation procedure. Once the four models were
constructed, four assembled models were proposed to gain accuracy. To do that the
prediction of LOP, dLOP, MOR and dMOR (known as LOP*, d*LOP, MOR* and d*MOR)
were incorporated into the variable dataset. Thereafter, the new principal components
were calculated, and new four deep-learning neural network were created for each
response, by using again l2 regularization and k-fold validation.
2 Database
2.1 Data Collection

Datasets containing dosages and of LOP, dLOP, MOR and dMOR of UHPFRC were
collected from several international symposium proceedings on high and ultra-high-
performance concrete (including Kassel 2004, 2008, 2012 and 2016), PhD thesis, and
other published works, totalizing 615 observations from the scientist literature. Besides,
81 experimental tests with different combinations of SCM and fibers were performed.
As a result, the present investigation begun with a total of 696 observations for the
development of the machine-learning models.
2.2 Experimental Investigation

As aforementioned, 81 experimental flexural tests over UHPFRC were performed to
complete the data base. Dosages of UHPC used in this part were developed previously
by the researchers [11, 12, 18].
Fibers used in this work were commercially available in Colombia. Table 1 sum-
marizes the properties of the fibers employed.
Table 1. Properties of steel fibers used in this study.

Notation Form Material df (mm) lf (mm) lf/df Tensile strength (MPa)
S1 Straight Steel 0.20 13 65 2600
S2 Straight Steel 0.20 6 30 2600
H1 Hooked Steel 0.50 35 70 1600
H2 Hooked Steel 0.30 30 80 1600
T Twisted Steel 0.50 13 26 1700
The bending test performed during the experimental phase included both 3-point
bending test and 4-point bending test. Test set-ups were depicted in Fig. 2.
Fig. 3. Test set-up for 4-point bending test (left) and 3-points bending test (right)
2.3 Data Treatment

2.3.1 Dealing with the Outliers
Before performing regression analyses it is necessary to deal with outliers, as they can
greatly affect the resulting model [19]. Hence, a descriptive statistical analysis was
carried out on each variable to identify outliers [20] based on bivariate boxplots [21].
Once this procedure was completed, 64 observations were eliminated from the data
base, leaving 632 for training purposes.
2.3.2 Data Missing

Dealing with data missing was performed by using the caret function in R [22].
2.3.3 Data Normalization

After the outliers were removed from database; next step consists in data normalization.
The input data and output data generally comprise of different identities either having
no or minimum similarities. Pre-processing or normalization of data eliminates the
possibility of neural network bias towards the different identities and scales down all
the input and output data. Since sigmoid function was used as activation function,
linear scaling in the range [0, 1] has been used as depicted in Eq. (1).
x xmin
xnormsigmoid ¼ ð1Þ
xmax xmin
where xnorm is the normalized value of the variable x, xmax and xmin are the minimum
and maximum values of variable x respectively.
2.3.4 Data Dimension

It is important to highlight that after the outliers treatment the dimension of data base
was 632 183, which means that there were 179 independent factors to predict the
four responses (i.e., LOP, dLOP, MOR and dMOR). Some of those factors were: volume
of cement, cement to binder ratio, cement to paste volume fraction ratio, silica fume,
silica fume to binder ratio, silica fume to paste volume fraction ratio, supplementary
cementitious materials, quartz powder, sand, aggregate, water, superplasticizer, type of
cement, virtual packing density, water to binder ratio, water to cement ratio, water to
total powders ratio, number of type of fibers, total volume of fibers, equivalent fiber
ratio, volume of fiber 1, nature of fiber 1, tensile strength of fiber 1, length of fiber 1,
aspect ratio of fiber 1, volume of fiber 2, nature of fiber 2, tensile strength of fiber 2,
length of fiber 2, aspect ratio of fiber 2, volume of fiber 3, nature of fiber 3, tensile
strength of fiber 3, length of fiber 3, aspect ratio of fiber 3, aggregate to cement-
equivalent ratio, aggregate plus fiber to total powder ratio, fine aggregate ratio, the
difference between the minimum and maximum value of aggregate, compressive
strength, type of bending test (3-point or 4-point), specimen’s wide, specimen’s height,
specimen’s length, specimen’s span, among others.
So, there was a necessity to perform a dimensionality reduction technique such as
Principal Component Analysis (PCA) [21].
2.3.5 Preparing Training and Test Data Set

To facilitate training and testing of four-hidden-layer models, the obtained data was
randomized and divided into training and test datasets. 474 observations were used for
training purposes and the remaining 158 were used for testing of the trained models.
Experimental results were contained in both subsets.
3 Neural Networks
3.1 Introduction
The simplest form of neural network architecture was devised by Rosenblant in 1959:
the perceptron, [23]. Perceptron is composed of one neuron which receives information
of two inputs and produces one output [24].It is defined as a four-tuple entity (i.e.,
sensors that (i) receive inputs and (ii) multiply them by weights, (iii) a function col-
lecting all the weighted data to produce a measurement on the impact of the observed
phenomenon, and (iv) a constant threshold). Ascertaining these weights to yield a
particular output is called “training”, which is the procedure that allows the model to
learn [25]. Figure 3 shows the schematic diagram of perceptron structure.
Fig. 4. Schematic diagram of perceptron structure
For more complex applications, multi-layer perceptron (MLP) are used, which
contain one input layer, one output layer, and one or more hidden layers. The multi-
layer perceptron (feed-forward network) has been a commonly used neural network
architecture [13, 26, 27].
In the engineering field, MLP models have been employed in applications like
detection of structural damage, water resources engineering, traffic engineering,
structural system identification, material behaviour modelling, concrete mix propor-
tioning and concrete strength forecasting [27, 28]. In fact, the application of ANN to
forecast mechanical properties of pastes, mortars and concretes had become one of the
most fertile fields in the scientific literature of civil engineering production, [29].
However, until now few investigations have been conducted on forecasting the per-
formance of UHPC using neural networks. Among those little research, [4, 30] and [3]
investigations can be highlighted, as presented in Table 2.
Table 2. Previous application of ANN to forecast mechanical properties of UHPC.

Reference SCMs Architecture Number of Output(s)
observations
[30] SF 4-2-2-1 38 12 day compressive strength under heat
treatment
[4] SF, QP 7-15-3 53 Slump Flow, 28 day compressive strength, 2
day compressive strength under heat
treatment
[3] SF, FA 11-10-4 78 7, 28, 90 and 365 day compressive strength
3.2 Deep Learning

Deep learning is a class of machine learning algorithms based on artificial neural
networks that uses multiple layers to progressively extract higher level features from
the raw input [31]. This research used deep-learning algorithms with four hidden layers
to predict the flexural behaviour of UHPFRC.
3.3 L2 Regularization and k-fold Validation

One of the most common problems in data science is to avoid overfitting. Overfitting
refers to a model that models the training data too well. Overfitting happens when a
model learns the detail and noise in the training data to the extent that it negatively
impacts the performance of the model on new data [31].
To avoid overfitting this research employed l2 regularization, also known as
stopping and weight decay.
Besides, to assess the network while keeping adjusting the l2 value (lambda) for
each hidden layer, the training data could be divided into a training set and a validation
set. Nevertheless, owing to the fact that there are such few data points, the validation
set would end up being very small for a deep learning algorithm validation (as in our
case). For that reason, the validation scores might change a lot depending on which
data points were chosen to use for validation and which were chosen for training. In
other words, the validation scores might have a high variance regarding the validation
split. This would prevent from reliably evaluating the model [31]. The best practice in
such situations is to apply k-fold cross-validation (see Fig. 4). This procedure is based
on splitting the available data into k sub-sets, instantiating k identical models, and
training each one on k – 1 sub-sets while assessing the remaining sub-set. The vali-
dation score (in our case the root mean square error or RMSE) for the model used is
then the mean of the k validation scores obtained [31]. To select the best performance l2
values for the model, computational tests were performed by varying the l2 value in
each hidden layer in the range of 0.0 to 0.1 at step of 0.001and choosing the combi-
nation which achieved the minor value of RMSE. In this research k = 10 was
considered.
Fig. 5. K-fold validation [31]. T: Training set, V: Validation set.
4 Principal Component Analysis
Principal component analysis (PCA) is a statistical methodology that employs an

orthogonal transformation to convert a set of observations of possibly correlated
variables into a set of values of linearly uncorrelated variables known as principal
components. This transformation is defined in such a way that the first principal
component has the largest possible variance (i.e., accounts for as much of the vari-
ability in the data as possible), and each succeeding component in turn has the highest
variance possible under the constraint that it is orthogonal to the preceding compo-
nents. The resulting vectors (each being a linear combination of the variables and
containing n observations) are an uncorrelated orthogonal basis set [32]. In this
research, PCA was used as dimensionality reduction technique [21] owing to the high
length of the data set.
To select the number of principal components to consider in the models, the Kaiser
criterion was used [33].
5 Index of Predicction Accuracy
The models were trained on the training dataset by using k-fold validation to adjust
lambda values. The assessment of the models was performed by comparison of the
predicted values with the real ones in the validation set through computing mean
absolute error (MAE), normalized mean bias error (NMBE) and coefficients of multiple
determination (R2), using Eqs. (2–5) respectively.
1X n
MAE ¼ j ai ^
ai j ð2Þ
n i¼1
rP ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
i¼1 ðai ^
n
ai Þ
RMSE ¼ ð3Þ
n
P
1 n
ð ai ^ ai Þ
NMBE ð% Þ ¼ n i¼1 100 ð4Þ
ai
Pn
i¼1 ðai ^ ai Þ 2
R2 ¼ 1 P ð5Þ
i¼1 ð^
n
a i Þ2
where: a is the target or experimental value; ā represents the mean observed target, â is
the predicted value or model’s output and n is the total number of observations in the
current data.
MAE uses absolute differences between the measured and the predicted values. The
optimal value of MAE is zero [27]; RMSE is one of the most used error measures of
accuracy for statistical models [34]. RMSE compares the observed values to the pre-
dicted values and computes the square root of the average residual error, indicating
error in the units (or squared units) of the constituent of interest, which aids in analysis
of the results. RMSE values of zero indicate a perfect fit. However RMSE gives more
weightage to large errors [13]; The NMBE provides information on the mean bias in
the predictions from a model. A negative NMBE indicates over-prediction and a
positve NMBE indicates under-prediction of the model [35]; Coefficient of determi-
nation (R2) compares the accuracy of the model with the accuracy of a superficial
benchmark model wherein the prediction is the mean of all samples [26]. R2 statistics is
based on the linear relationships between the real and predicted values and may
sometimes supply biased results if those relationship are not linear or if there are many
outliers. A R2 value of unity points out a perfect association between the real and
predicted values. A combined use of those performance metrics can supply an unbiased
idea of forecasting ability of the machine learning models.
6.1 First Step: Single Models

6.1.1 Principal Component Analysis
After the PCA, only the first eleven PC satisfied the Kaiser rule. Therefore, the
dimension reduction lead to 11 input factors from 179.
6.1.2 Models Architecture and L2 Values

The number of hidden layers of the models was fixed at the beginning. After defined
the number of input nodes, the architectures were completely defined. Figure 5 depicts
the architectures used.
After the adjusting of l2 values during the k-fold validation process, the value of
regularization parameters for each model and each hidden layer was depicted in
Table 3.
Fig. 6. Selected architecture for the four models created in the first step. The output node (O1)
varies for each model, representing LOP, dLOP, MOR and dMOR.
Table 3. Values of lambda in l2 regularization obtained thought k-fold validation.

Model response l2 regularization value
1st hidden layer 2nd hidden layer 3rd hidden layer 4th hidden layer
LOP (MPa) 0.056 0.001 0.098 0.001
dLOP (mm) 0.014 0.005 0.038 0.015
MOR (MPa) 0.086 0.048 0.051 0.021
dMOR (mm) 0.049 0.062 0.005 0.055
6.1.3 Assessment of Prediction Accuracy for the Proposed Models

Table 4 depicts the index of prediction accuracy for each proposed model.
Table 4. Values of the index of prediction for each proposed model.

MAE RMSE NMBE R2
LOP (MPa) 3.451 3.672 9.215% 0.713
dLOP (mm) 0.017 0.019 −8.367% 0.654
MOR (MPa) 2.861 2.941 8.251% 0.783
dMOR (mm) 0.122 0.142 −10.034% 0.668
According to the results presented in Table 4 the accuracy of the model could be
considered acceptable for LOP and MOR, but poorly in the case of both deflections.
6.2 Second Step: Assembled Models

6.2.1 Principal Component Analysis
After the development of single models, the database is increased considering the
predictions of the response variables, i.e. LOP*, d*LOP, MOR* and d*MOR. Therefore, the
new database had 183 features instead 179. Once the PCA analysis was performed,
only the first ten PC satisfied the Kaiser rule. Hence, the dimension reduction lead to 10
input factors from 183.
6.2.2 Models Architecture and L2 Values

Unlike the previous models, in this second stage the models had only ten input data (i.e.
the PC selected by using the Kaiser rule), keeping four hidden layers and a single
output node. Values of lambda obtained through k-fold validation process are presented
in Table 5.
Table 5. Values of lambda in l2 regularization obtained thought k-fold validation.

1st hidden layer 2nd hidden layer 3rd hidden layer 4th hidden layer
LOP (MPa) 0.022 0.002 0.001 0.021
dLOP (mm) 0.012 0.024 0.021 0.025
MOR (MPa) 0.002 0.018 0.002 0.017
dMOR (mm) 0.009 0.089 0.045 0.001
6.2.3 Assessment of Prediction Accuracy for the Proposed Models

Table 6 depicts the index of prediction accuracy for each proposed model.
Table 6. Values of the index of prediction for each proposed model.

MAE RMSE NMBE R2
LOP (MPa) 1.934 2.002 3.215% 0.894
dLOP (mm) 0.011 0.012 −3.346% 0.752
MOR (MPa) 1.828 1.971 2.251% 0.907
dMOR (mm) 0.092 0.112 −3.034% 0.803
Now, according to the results presented in Table 6 the accuracy of the assembled
models could be considered good for LOP and MOR, and acceptable in the case of
deflections, which means an important improvement in both cases.
However, it can be observed that the R2 value was smaller than those obtained by
other authors consulted, who achieved correlation coefficients between 0.96 and 0.98
when predicting UHPFRC’s properties using ANN [3, 4]. The latter can be explained
owing to the fact that those researchers used data from their own experimental research,
and, generally, a few SCM and fibers. The wide range of data utilized in this work
incorporated some statistical noise into the system, viz., the use of different types of
cement and the characteristics of the SCM such as physical and chemical properties;
the used of different kind of fibers; the unreported detail of the use of vibration during
the pouring of the concrete, the different nature and shape of aggregates used in the
database, and the different technology of superplasticizers considered in the database,
among others.
7 Conclusions
In this paper, four machine deep learning models were developed to predict the
behaviour of UHPFRC under different configurations of bending tests. sing different
combinations of supplementary cementitious materials viz., silica fume, fly ash. To
avoid overfitting, k-fold validation with ten partitions as well as l2 regularization were
employed, thereby leading to higher confidence of the model when predicting on new
data. From the obtained results of this investigation, the following conclusions are
drawn:
• The assembled models for predicting the flexural behaviour of UHPFRC provided a
scientifically improvement in the forecast of LOP, dLOP, MOR and dMOR in com-
parison with the single machine-learning models.
• The proposed assembled models based on a wide range of experimental and pre-
vious work data can be very handy for forecasting the flexural behaviour of
UHPFRC in quick time. It could be helpful in the developing of UHPFRC as
decision support tool, predicting the LOP, dLOP, MOR and dMOR of a particular
UHPFRC design. This procedure will considerably decrease the effort, costs and
time to design an UHPFRC dosage for a customized bending behaviour without
performing multiple trials.
• The results of the combined use of the performance indexes, which included MAE,
RMSE, NMBE, and R2, provided an unbiased estimate which proved the adequacy
of the proposed assembled machine-learning models.
Acknowledgments. Special thanks go to APOLO from EAFIT university (Medellín-Colombia)

and OSIRIS from Escuela Colombiana de Ingenieria Julio Garavito (Bogotá-Colombia) for the
servers and computer related support. Also, to Cementos Argos SA. for donating most of the
materials used in the research described herein and provide the SEM images of the components.
The supply of recycled glass from Cristaleria Peldar SA, FCC from Ecopetrol SA, and GBSF
from Gerdau SA for this research is highly appreciated. The writers would also like to
acknowledge the support and suggestions of Escuela Colombiana de Ingeniería Julio Garavito
and Polytechnic University of Madrid (UPM).
References
1. Abellan, J., Torres, N., Núñez, A., Fernández, J.: Ultra high preformance fiber reinforced
concrete: state of the art, applications and possibilities into the latin american market. In:
XXXVIII Jornadas Sudam. Ing. Estructural, Lima, Peru (2018)
2. Abellan, J., Torres, N., Núñez, A., Fernández, J.: Influencia del exponente de Fuller, la
relación agua conglomerante y el contenido en policarboxilato en concretos de muy altas
prestaciones, In: IV Congr. Int. Ing. Civ., Havana, Cuba (2018)
3. Zhang, J., Zhao, Y.: Experimental investigation and prediction of compressive strength of
ultra-high performance concrete (UHPC) containing supplementary cementitious materials.
Hindawi Adv. Mater. Sci. Eng. 2017, 522–525 (2017). https://doi.org/10.1155/2017/4563164
4. Ghafari, E., Bandarabadi, M., Costa, H., Júlio, E.: Prediction of fresh and hardened state
properties of UHPC: Comparative study of statistical mixture design and an artificial neural
network model. J. Mater. Civ. Eng. 27, 04015017 (2015). https://doi.org/10.1061/(ASCE)
MT.1943-5533.0001270
5. ACI Committe 239, ACI – 239 Committee in Ultra-High Performance Concrete (2018)
6. Meng, W., Samaranayake, V.A., Khayat, K.H.: Factorial design and optimization of UHPC
with lightweight sand. ACI Mater. J. (2018). https://doi.org/10.14359/51700995
7. Abellán-García, J., Núñez-López, A., Torres-Castellanos, N., Fernández-Gómez, J.: Factorial
design of reactive powder concrete containing electric arc slag furnace and recycled glass
powder. Dyna. 87, 42–51 (2020). https://doi.org/10.15446/dyna.v87n213.82655
8. Viet, T.A.V., Ludwig, H.M.: Proportioning optimization of uhpc containing rice husk ash
and ground granulated blast-furnace slag. In: Schmidt, M., Fehling, E., Glotzbach, C.,
Fröhlich, S., Piotrowski, S. (Eds.) 3rd International Symposium. UHPC Nanotechnology
Construction Materials, Kassel, Germany, pp. 197–205 (2012)
9. Li, W., Huang, Z., Zu, T., Shi, C., Duan, W.H., Shah, S.P.: Influence of nanolimestone on the
hydration, mechanical strength, and autogenous shrinkage of ultrahigh-performance concrete.
J. Mater. Civ. Eng. 28, 1–9 (2016). https://doi.org/10.1061/(ASCE)MT.1943-5533.0001327
10. Huang, Z., Cao, F.: Effects of Nano-materials on the Performance of UHPC, 材料导报B:研
究篇. 26 136–141 (2012)
11. Abellán-García, J., Núñez-López, A., Torres-Castellanos, N., Fernández-Gómez, J.: Effect of
FC3R on the properties of ultra-high-performance concrete with recycled glass • Efecto del
FC3R en las propiedades del concreto de ultra altas prestaciones con vidrio reciclado. Dyna.
86, 84–92 (2019). https://doi.org/10.15446/dyna.v86n211.79596
12. Abellán, J., Fernández, J., Torres, N., Núñez, A.: Statistical optimization of ultra-high-
performance glass concrete. ACI Mater. J. 117, 243–254 (2020). https://doi.org/10.14359/
51720292
13. Chandwani, V., Agrawal, V., Nagar, R.: Modeling slump of ready mix concrete using
genetic algorithms assisted training of Artificial Neural Networks, Expert Syst. Appl.
42 (2015) 885–893. https://doi.org/10.1016/j.eswa.2014.08.048
14. Abellán-García, J., Fernández-Gómez, J., Torres-Castellanos, N.: Properties prediction of
environmentally friendly ultra-high-performance concrete using artificial neural networks.
Eur. J. Environ. Civ. Eng 1–25.10.1080/19648189.2020.1762749 (2020)
15. Abellán-García, J.: Four-layer perceptron approach for strength prediction of UHPC, Constr.
Build. Mater. 256 (2020). https://doi.org/10.1016/j.conbuildmat.2020.119465
16. Khashman, A., Akpinar, P.: ScienceDirect non-destructive prediction of concrete compres-
sive strength using neural networks prediction of concrete compressive strength using neural
networks. Proc. Comput. Sci. 108, 2358–2362 (2017). https://doi.org/10.1016/j.procs.2017.
05.039
17. R Core Team, “R: A Language and Environment for Statistical Computing,” Vienna, Austria
(2018). https://www.r-project.org/
18. Abellán, J., Torres, N., Núñez, A., Fernández, J.: Quality optimization of low-cost UHPC
using micro limestone powder and glass flour, Comput. Concr. (n.d.)
19. Atkinson, A., Riani, M.: Robust Diagnostic Regression Analysis. Springer, US, New York
(2000)
20. Härdle, W.K., Simar, L.: Applied Multivariate Statistical Analysis. Springer-Verlag GmbH,
Berlin (2012)
21. Everitt, B., Hothorn, T., MVA: An Introduction to Applied Multivariate Analysis with R
(2015)
22. Max Kuhn Contributions from Jed Wing, A., Weston, S., Williams, A., Keefer, C.,
Engelhardt, A., Cooper, T., Mayer, Z., Benesty, M., Lescarbeau, R., Ziem, A., Scrucca, L.,
Tang, Y., Candan, C., Hunt, T., Max Kuhn, M., Package “caret” Classification and
Regression Training Description Misc functions for training and plotting classification and
regression models, CRAN - R Repos (2017). https://cran.r-project.org/web/packages/caret/
caret.pdf
23. Rosenblatt, F.: The Perceptron: A probabilistic model for information storage and
orgnization in the brain. Cornerr Aeronaut. Lab. 65, 386–408 (1958)
24. Ghafari, E., Al.: Optimization of UHPC by Adding Nanomaterials. In: Proceedings of
Hipermat 2012, in 3rd International. Symposium. UHPC Nanotechnology Construction.
Material., Kassel Uni, Kassel, Alemania, pp. 71–78 (2012)
25. Estebon, M.D., Perceptrons : An Associative Learning Network, Virginia Tech (1997)
26. Gupta, S.: Using artificial neural network to predict the compressive strength of concrete
containing nano-silica. Civ. Eng. Archit. 1, 96–102 (2013). https://doi.org/10.13189/cea.
2013.010306
27. Aderaw, M., Muse, S., Abiero, Z.C.: Artificial neural network based modelling approach for
strength prediction of concrete incorporating agricultural and construction wastes. Constr.
Build. Mater. 190, 517–525 (2018). https://doi.org/10.1016/j.conbuildmat.2018.09.097
28. Adeli, H.: Neural networks in civil engineering: 1989–2000. Comput. Civ. Infrastruct. Eng.
16, 126–142 (2001)
29. Chandwani, V., Nagar, R.: Applications of artificial neural networks in modeling
compressive strength of concrete: a state of the art review. Int. J. Curr. Eng. Technol. 4,
2949–2956 (2014)
30. Taghaddos, H., Mahmoudzadeh, F., Pourmoghaddam, A., Shekarchizadeh, M.: Prediction of
compressive strength behaviour in RPC with applying an adaptive network-based fuzzy
interface system. In: Proceeding International Symposium Ultra High Performance Concr.,
Kassel, Alemania (2004)
31. Chollet, F., Allaire, J.J.: Deep Learning with R. Manning Publications Co, New Jersey
(2018)
32. James, G., Witen, D., Hastie, T., Tibshirani, R.: An Introduction to Statistical Learning with
Applications in R (2007). https://doi.org/10.1016/j.peva.2007.06.006
33. Kaiser, H.F.: The application of electronic computers to factor analysis. Educ. Psychol.
Meas. 20, 141–151 (1960)
34. Moriasi, D.N., Arnold, J.G., Liew, M.W.V., Bingner, R.L., Harmel, R.D., Veith, T.L.:
MODEL evaluation guidelines for systematic quantification of accuracy in watershed
simulations. Am. Soc. Agric. Biol. Eng. 50, 885–900 (2007)
35. Srinivasulu, S., Jain, A.: A comparative analysis of training methods for artificial neural
network rainfall – runoff models. Appl. Soft Comput. 6, 295–306 (2006). https://doi.org/10.
1016/j.asoc.2005.02.002
Study of Dimensioning Aspects of FRC Based
on the Beam Flexion Theory
Iva E. Pereira Lima1(&) and Aline S. Ramos Barboza2

1
Civil Engineering/Structures from Federal University of Alagoas,
Alagoas, Brazil
ivaemanuellyl@gmail.com
2
Structural Engineering from the University of São Paulo and Professor Federal
University of Alagoas, Alagoas, Brazil
Abstract. The fiber reinforced concrete (FRC) have a higher load capacity in
the post-cracking stage, however, this increase only occurs if the concrete is
dosed and applied properly and, for this, there are standards that establish design
aspects for use of the FRC. However, in these documents, the portion of the
resulting resistant strength of the fibers is not fully described in their equations
and is determined for specific situations. Therefore, and with the objective of
obtaining the positioning and the resultant of the fibers of a model of fluid
concrete reinforced with fibers, this work presents a study related to the deter-
mination of the resultant of the fibers in the stretched part of the concrete. This
determination was obtained based on the bending theory in beams, through the
three-point bending test standardized by EN 14651 (2007) [8], where both
beams with steel fibers and polymeric fibers were used. From this, the equation
obtained analytically was compared with those determined by international
codes and it was realized that, when comparing the three methods used, the
intermediate value corresponded to the determined analytically, where the
results showed variations due to the different forms of arrangement that each
method uses.
Keywords: Fiber reinforced concrete Flexural test Beam flexural theory

Determination of the resultant of the fibers
1 Introduction
The use of fiber reinforced concrete (FRC) has been gradually increasing worldwide
and has been undergoing several advances since the year 1970 [2, 12, 14]. From this
advance, research with fiber-reinforced concrete began to show the application of fibers
as a structural material and, together with this type of reinforcement, studies showed
that the negative aspects of FRC could be reversed with the use of additives, where the
fluidity of the composite allows the distribution of fibers to occur in a uniform way,
being possible to use the flow of concrete launching to guide the fibers [7, 10, 17]
The use of this type of concrete for structural purposes has been consolidated over the
last 15 years with the development of concepts of fracture mechanics to describe its
residual tensile strength [6, 10]. As regards the world scale, this growth trend is
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_53
Study of Dimensioning Aspects of FRC 585
exemplified by the appearance of several international normative instructions [1, 5, 9].

What is related to Brazil, the lack of Brazilian standards related to the FRC means that
professionals who work with this type of material in the country, base their design on
foreign standards and, among these international codes, the material can be highlighted
described both in the FIB Model Code 2010 (2013) [9] and in the ACI 544.8R (2016) [1].
In these regulations, the design aspects that should be considered to design a
structural design produced with FRC elements are described, where this design must be
based on the study of the strength provided by the fiber reinforcement, in which the
crack openings in the part are observed traction of concrete from constitutive laws.
These laws must be defined for the analysis of the post-cracking behavior of the FRC
and can be deduced from the results of bending tests. However, for the study of these
laws, the aforementioned regulations do not consider the orientation of fibers in con-
crete as an interference factor in the direction of the resistant force, as investigated in
this work, as they consider that the maximum load corresponds to the cracking load,
when using fibers dispersed in the matrix, and base their studies on this hypothesis.
Therefore, and knowing the significant change that the preferential orientation of
the fibers can provide to the concrete, when using a fluid matrix for example, in order to
obtain the positioning and the resultant of the fibers of a fiber reinforced concrete
model, we present a study related to the determination of the resulting resistant strength
of the fibers in the tensile part of the concrete. This result was obtained based on the
bending theory in beams, performed through the three-point bending test standardized
by EN 14651 (2007) [8], where a sample of concrete reinforced with steel fibers
(SFRC) was used and another sample of concrete reinforced with polymeric fibers
(PFRC).
2 Theoretical Reference
2.1 Sizing Aspects of the FRC

The Fib Model Code 2010 (2013) [9] is a document published by Féderation Inter-
nationale du Béton (FIB), whose main objective is to serve as a parameter for future
structural concrete standards and to present new research developments about new
cementitious materials with structural purpose. And the American Concrete Institute
(ACI) aims to develop norms/codes that serve as a parameter for the development,
dissemination and adoption of technical resources for the best use of concrete. This
institute (ACI) prepared the ACI 544.8R (2016) [1], which is a report referring to the
indirect method to obtain the stress/strain response of fiber reinforced concrete (FRC).
In order to use the FRC for structural purposes, both the Fib Model Code 2010
(2013) [9] and the ACI 544.8R (2016) [1] describe some design aspects that must be
considered to conceive a structural project produced with fiber reinforced concrete
elements, where this project must be based on the study of the residual strength
provided by the fiber reinforcement. As regards the Fib Model Code 2010 (2013) [9],
the linear model identifies two reference values, the fFts and the fFtuk , in which the first
is related to the service limit state and the second with the ultimate limit state (Eq. (1)
586 I. E. Pereira Lima and A. S. Ramos Barboza
and (2)). From the resistance values related to residual strength, the resulting resistant
strength of the fibers in the tensioned part of the FRC must be calculated (Eq. 3).
fFts ¼ 0; 45fR1 ð1Þ

wu
fFtuk ¼ fFts ðfFts 0; 5fR3 þ 0; 2fR1 Þ 0 ð2Þ
CMOD3
fFtuk
fFtud ¼ ð3Þ
1; 5
Where: fFts : residual voltage characteristic related to the service limit state; fR1 : residual
flexural strength corresponding to a crack opening of 0,5 mm; fFtuk : final residual stress
characteristic determined by the linear model; wu : maximum crack opening limit;
CMOD3: crack opening corresponding to 2,5 mm; fFtud : final tensile stress on the FRC.
As regards ACI 544.8R (2016) [1], the verification of safety through flexion also
occurs when defining the final residual strength of the FRC project, as shown by the Fib
Model Code 2010 (2013) [9]. The final tensile stress in the FRC (Ft2 ) is the indicator
that must be analyzed in the study of the ultimate limit state and it can be determined
with the aid of Eq. (4) to (8).

ð1 k Þðb 1Þðgb g þ 2Þ
Ft2 ¼ b h rcr ð4Þ
2b
Where: b: coefficient that must obey the following interval 1\b a; g: coefficient
which, for simplification for cementitious composites, must be within the following
range 0; 0 g 0; 4; b: specimen width; h: specimen height; rcr : initial cracking stress
of fiber-reinforced concrete; Ft2 : final tensile stress on the FRC.
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2b 1
k ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð5Þ
2b 1 þ b
etrn
a¼ ð6Þ
ecr
Where: etrn : ultimate tensile strain on FRC; ecr : initial cracking deformation of the FRC.
ecr ¼ e1 ¼ e2 ð7Þ
In accordance with ACI 544.8R (2016) [1], it is necessary that the e2 ¼ 0:12%.
rcr ¼ r1 ¼ C1 ð1; 6 d Þftm;fl ð8Þ
Where: d: working height of the specimen; ftm;fl : proportionality limit corresponds to

the resistance calculated from the maximum load within the crack opening range
between 0 and 0.05 mm.
3 Methodology
3.1 Materials and Dosage Definition
The matrix used in the production of concretes is composed of the following materials:
fine and coarse aggregates, cement, mineral addition (RBMG), water and superplas-
ticizer additive. For the manufacture of SFRC, the reinforcement used was steel fibers,
and for PFRC, it was polymeric fibers. Regarding the dosage, it was developed using
the method of Gomes et al. (2003) [11], where these values are represented in Table 1
and consumption is represented in mass.
Table 1. Consumption of materials by m3 - FRC.

Materials Pasta (kg)
Cement 377.90
RBMG 188.95
Aggregate kid 774.53
Big aggregate 813.75
Initial water 151.16
Complementary water 37.75
Absorption water 11.17
Superplasticizer 6.30
SFRC
Steel fibers 25.98
PFRC
Polymeric fibers 3.78
3.2 Application of the Standardized Test by EN 14651 (2007)

For the application of the three-point flexion test standardized by EN 14651 (2007) [8],
the Shimadzu test machine, model AG-X Plus, was used and this equipment was
coupled with the ITOM and Trapezium X software (Fig. 1a). From this, Trapezium X
provides data that relate test time to applied load, while ITOM provides data that relate
time to crack opening. In order to make the two sets of data compatible, a scan of the
values was carried out, which makes it possible to relate the load parameters to the
crack opening. Therefore, it was possible to perform a graphical analysis of the results,
where two curves were generated, one referring to the load versus crack opening and
the other related to the load versus test time (Fig. 1b).
Through Fig. 1, it is observed that it is possible to obtain the load supported by the
composite and that, from established times, its associated loads are obtained. In
addition, during the tests, pictures were taken every 10 s regarding the crack opening
progress of the beams, with the objective of checking the cracking time, where it was
verified through the analysis of the images, the cracking time for each tested beam and,
from that, the corresponding crack load was obtained.
Fig. 1. Test EN 14651
4 Results and Discussions
4.1 Results Obtained with the Standard Test EN 14651 (2007)

Regarding the SFRC elements, 8 beams were tested and treated as a single batch. The
comparative study, corresponding to the crack load value with the maximum load result
obtained by the test, was carried out from a statistical characterization of the data,
where for each load sample analyzed, the mean, standard deviation was verified,
variance, coefficient of variation and Skewness coefficient (Table 2).
Table 2. Results obtained from maximum load and cracking load for SFRC beams.
Sample identification Pmáx (kN) Pfiss (kN)
V1 17.99 19.83
V2 16.24 17.33
V3 20.15 20.50
V4 20.15 20.33
V5 25.42 24.33
V6 16.45 16.83
V7 16.17 14.67
V8 17.58 16.67
Average 18.77 18.81
Standard deviation 3.13 3.04
Variance 9.80 9.21
Coefficient of variation (%) 16.68 16.14
Skewness coefficient 1.02 0.38
Observing Table 2, it appears that, in general, the two parameters analyzed have
similar values, however, the maximum load has higher values of standard deviation,
variance and coefficient of variation, which implies that this sample has a greater
variability of the data. In addition, by assessing the Skewness coefficient, it can be seen
that both coefficients were positive and this means that the values of x greater than the
mean are more dispersed than the smaller ones. Therefore, and in order to assess the
dispersion of these two parameters, Table 3 shows the results regarding the variation of
the maximum load in relation to the crack load.
Table 3. Variation between the two evaluated variables (steel fibers).

Sample identification Pmáx (kN) Pfiss (kN) Variation (%)
V1 17.99 19.83 −9.28
V2 16.24 17.33 −6.29
V3 20.15 20.50 −1.71
V4 20.15 20.33 −0.89
V5 25.42 24.33 +4.48
V6 16.45 16.83 −2.26
V7 16.17 14.67 +10.22
V8 17.58 16.67 +5.46
From Table 3, it can be seen that what refers to the largest variations, they occur in
beams V1 and V7, where it is observed that the crack load is greater than the maximum
load around 9%, and the maximum load exceeds that of cracking by around 10%,
respectively, showing that even for the largest dispersions, the differences can still be
considered small. Still analyzing the same Table, it can be seen that, in general, the crack
load presents a higher value, when compared to the maximum load, and this occurs due
to the high elasticity modulus that the steel fibers have and the efficient adhesion SFRC
fiber matrix. This fact can be explained because this type of fiber is not very deformable
and collaborates to stiffen the composite in the post-cracking, showing that, even after
the matrix rupture (Pmáx), there may be a redistribution of efforts provided by the fibers,
due to the adhesion matrix fiber, which promotes an increase in the rupture load of the
matrix and refers to the crack resistance of the composite (Pfiss ) [2–4]. The procedure
described above was repeated for beams with PFRC (Tables 4 and 5).
Table 4. Results obtained from maximum load and crack load for PFRC beams.
Sample identification Pmáx (kN) Pfiss (kN)
V1 16.92 6.33
V2 14.42 5.33
V3 14.92 4.42
V4 14.42 4.17
V5 15.17 4.17
V6 14.67 5.17
V7 14.33 5.33
V8 15.17 4.83
Average 15.00 4.97
Standard deviation 0.84 0.73
Variance 0.71 0.54
Coefficient of variation (%) 5.62 14.73
Skewness coefficient 1.31 0.46
Observing Table 4, it appears that, for all the beams analyzed, the maximum load
values are higher. In addition, this parameter also presents higher values of standard
deviation and variance, although the values present a small difference, however, the
coefficient of variation regarding the crack load is greater, which implies that this
sample has a greater variability of the data. In addition, evaluating the Skewness
coefficient, it is noticed that the values of x greater than the mean are more dispersed
than the smaller ones. Therefore, in view of the differences found, and in order to assess
the dispersion of these two parameters, Table 5 presents the results regarding the
variation of the maximum load in relation to the crack load.
Table 5. Variation between the two variables evaluated (polymeric fibers).

Sample identification Pmax (kN) Pfiss (kN) Variation (%)
V1 16.92 6.33 +167.30
V2 14.42 5.33 +170.54
V3 14.92 4.42 +237.56
V4 14.42 4.17 +245.80
V5 15.17 4.17 +263.79
V6 14.67 5.17 +183.75
V7 14.33 5.33 +168.86
V8 15.17 4.83 +214.08
From Table 5, it can be seen that all variations are significant, showing discrep-
ancies greater than 100%, which shows that the crack loads are much lower than the
maximum loads. Still observing the data in the same Table, it can be seen that the
maximum load values are higher, and this is due to the low modulus of elasticity that
the polymeric fibers present. After the matrix breaks (Pmáx), polymeric fibers are
unlikely to contribute to an increase in the strength of the composite, as this type of
fiber is more fragile than the matrix because it presents a modulus of elasticity less than
the concrete module [3, 15].
The fibers that exhibit this type of behavior can generate a composite with less
stiffness, which can provide a crack resistance of the composite (Pfiss ) less than the load
corresponding to the rupture of the matrix (Pmáx). This also proves the hypothesis of
Toledo (1997) [16], where this author affirms that, for elements that present softening
behavior and show a preferential orientation of the fibers, generally, the maximum load
can assume a value higher than the crack load.
4.2 Definition of Neutral Line Positioning

After the previous study, it was found that the samples showed discrepancies and, from
that, we sought to define the position of the neutral line of the structure. This parameter
can be obtained by defining the indicators a and c (Fig. 2), where the first corresponds
to the traction height of the plastic response and the second is equivalent to the traction
height of the elastic response. For this, the height related to the first crack, which
represents the height of traction related to the plastic response, was experimentally
defined (Table 6), and a validation study of that length was made. After that, the total
length corresponding to the two mentioned traction heights was verified.
Fig. 2. Structural element studied [13].
Table 6. Height related to the first crack for the SFRC and PFRC beams, respectively.
Sample Height Height Sample Height (px) Height
identification (px) (mm) identification (mm)
V1 656 42.50 V1 771 50.00
V2 769 49.80 V2 744 48.20
V3 808 52.40 V3 744 48.20
V4 823 53.30 V4 726 47.00
V5 780 50.50 V5 735 47.60
V6 702 45.50 V6 767 49.70
V7 810 52.50 V7 756 49.00
V8 764 49.50 V8 787 51.00
Average 49.50 Average 48.84
Standard deviation 3.74 Standard 1.34
deviation
According to the data provided in Table 6 for the SFRC, it is observed that the
height of the first crack is equivalent to an average of 49.50 mm with a standard
deviation of 3.74 mm (49.50 ± 3.74 mm). Still analyzing the data illustrated in
Table 6, with respect to the PFRC, it appears that the height of the first crack is
equivalent to an average of 48.84 mm with a standard deviation of 1.34 mm (48.84 ±
1, 34 mm). In order to establish a value that represents both steel fiber and polymeric
beams, the two specified intervals were used, in which the value for the height of the
first crack corresponding to 50 mm was established.
With this definition, a numerical validation study of this length was made, in which
it was observed the type of request that the piece was undergoing at that point. Thus, it
was found that, from the stress evolution points along the crack, all stress values were
positive, and that means that, for the tensile height corresponding to 50 mm, only
tensile stress occurs, which validates the behavior of the traction height regarding the
plastic response. Therefore, this same analysis was performed for other predetermined
heights (62.5 mm, 75 mm, 87.5 mm, 100 mm, 112.5 mm and 125 mm) and it was
found that the equivalent centroid height at 75 mm corresponds to the positioning of
the neutral line of the FRC structure, since there was no variation in the order of
magnitude of the stresses, which shows that the tensile stress was equivalent to the
compressive stress.
4.3 Determination of the Final Tensile Strength of Concrete

After establishing the positioning of the neutral line of the structure, as shown in the
previous section, from the bending theory in Euler-Bernoulli beams and by means of
equilibrium equations, the resulting resistant strength of the fibers in the tensile part of
the fibers was numerically determined. concrete. The equation for determining this
force is shown below and, for that, Fig. 3 was used as an aid. In addition, the beam has
a cross section of 150 mm 150 mm (base height) and the following parameters
were used k = 0.8 for fck 50 MPa e g = 0.85 for concretes of grades up to C50.
Fig. 3. Scheme regarding the distribution of resistances.

Mrd þ ð0; 007651i 0; 000803Þ 1939; 66Mrd þ 0;01125i0;000422
0;0000058 f yd A s
fFtud ¼
ð0; 01125i 0; 000422Þ
ð9Þ
Where: fFtud : final tensile stress on the FRC; Mrd : resistant structure moment; fyd : yield
strength of steel; As : steel area of longitudinal reinforcement.
As can be seen, from Eq. 9, the final tensile stress in fiber reinforced concrete can
be determined. In order to verify the Equation found and validate it, was compared with
the Equations determined in accordance with the FIB Model Code 2010 (2013) [9] and
in accordance with ACI 544.8R (2016) [1], respectively. For this, the same parameter
values were adopted, which are necessary in the equations, and to what the first
standard refers, Eqs. 1 to 3 were used, and, in relation to the second standard, Eqs. 4 to
8 were used. employed. The results found are shown in Tables 7 and 8.
Table 7. SFRC final tensile stress in accordance with the different analyzes performed.
Analytical study International Variation
(MPa) standardization (MPa) (%)
Fib Model Code 2010 4,20 3,49 16,85
(2013)
ACI 544.8R (2016) 4,20 4,91 −16,95
Table 8. PFRC final tensile stress according to the different analyzes performed.
Analytical study International Variation
(MPa) standardization (MPa) (%)
Fib Model Code 2010 3,36 1,91 43,19
(2013)
ACI 544.8R (2016) 3,36 4,18 −24,45
Analyzing the data in Table 7, it can be seen that, when comparing the final tensile
stress of the SFRC obtained by the analytical study with that determined in accordance
with the FIB Model Code 2010 (2013) [9], the value by the first method is higher than
the second around 17%. However, when contrasting the result of the analytical study
with the one defined in accordance with ACI 544.8R (2016) [1], the first resulted in a
value lower than the second around 17%. In parallel, it can be seen from Table 8 that,
when comparing the final tensile stress of the PFRC obtained by the analytical study
with that determined in accordance with the FIB Model Code 2010 (2013) [9], the
value for the first is greater than the second around 43%. However, when contrasting
the result of the analytical study with that defined in accordance with ACI 544.8R
(2016) [1], the first resulted in a value lower than the second around 24%.
Therefore, it is clear that the result defined by the Fib Model Code 2010 (2013) [9]
results in a lower value and that established by ACI 544.8R (2016) [1] in a higher
value, when compared with the analytical, and this can be justified by the adoption of
the parameters established by their respective equations. In summary, the parameters
that were inserted in the equation determined by the FIB Model Code 2010 (2013) [9]
are, in general, related to the flexural strength of the composite, while the parameters
requested by the equation defined in ACI 544.8R (2016) [1] both the flexural strength
of the composite and the intrinsic characteristics of the material are related.
5 Conclusions
– From the investigation carried out, it was noted that with regard to the SFRC, the
crack load presented a higher value and this occurred due to the high modulus of
elasticity that the steel fibers present, whereas in relation to the PFRC, it can It
should be noted that the maximum load values were higher, and this was due to the
low modulus of elasticity of the polymeric fibers.
– Therefore, it was noted that the maximum load did not correspond to the crack load,
where it was sought to define analytically the positioning of the neutral line of the
structure. After this definition and from the comparison with the referenced stan-
dards, it was observed that the final tensile stress in the FRC obtained by the
experimental procedure resulted in a higher value, in relation to that determined
according to the Fib Model Code 2010 (2013) [9], and at a lower value compared to
that defined according to ACI 544.8R (2016) [1].
– Still, it was noticed that when comparing the three methods used, the intermediate
value corresponded to the one determined experimentally, where the values pre-
sented variations due to the different forms of arrangement that each method uses.
Therefore, it can be said that the present work achieved the goals suggested initially,
where the resulting resistant strength of the fibers in the tensile part of the concrete
was determined.
References
1. American Concrete Institute. ‘ACI 544.8R-16’: Indirect method for obtaining the
stress/strain response of fiber-reinforced concrete (FRC). Dubai (2016)
2. Bentur, A., Mindess, S.: Fiber Reinforced Cementitious Composites. Elsevier Applied
Science, London and New York (1990)
3. Caetano, L.F., Graeff, A.G., Garcez, E.O., Bernardi, S.T., Silva Filho, L.C.: Fiber-reinforced
cementitious matrix composite. Porto Alegre (2004)
4. Consiglio Nazionale Delle Ricerche. ‘CNR’: DT 204 guide for the design and construction
of fiber-reinforced concrete structures. Rome (2006)
5. Deutscher Beton- Und Bautechnik-Verein E.V.: DBV Guide to good practice steel fibre
concrete. Berlin (2001)
6. Di prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in Fib Model Code 2010:
Principles, models and test validation. Structural Concrete (2013)
7. Di prisco, M., Plizzari, G., Vandewalle, L.: Fibre reinforced concrete: new design
8. European Committee for Standardization: ‘EN 14651’: Test Method for Metallic Fiber-
Reinforced Concrete – Measuring the Flexural Tensile Strength (Limit of Proportionality
(LOP), Residual). CEN, London (2007)
9. Fédération Internationale du Béton – FIB. Fib model code for concrete structures 2010.
Switzerland (2013)
10. García-Taengua, E., Sonebi, M., Crossett, P., Taylor, S., Deegan, P., Ferrara, L., Pattarini,
A.: Self-compactability and strength criteria for concrete mixes with mineral additions and
fibres. In: ‘Concrete – Innovation and Design, fib Symposium’, Copenhagen May 18–20
(2015)
11. Gomes, P.C.C., Gettu, R., Agulló, L.: A new methodology for obtaining high strength self-
compacting concrete with mineral additives. In: V EPUSP Symposium on Concrete
Structures. São Paulo (2003)
12. Lopes, M.M.: Partial replacement of bending reinforcement in concrete beams. Engineering
Graduate Program - Federal University of Rio de Janeiro. Rio de Janeiro (2005)
13. Mobasher, B., Yao, Y., Soranakom, C.: Analytical solutions for flexural design of hybrid
steel fiber reinforced concrete beams. Eng. Struct. 100, 164–177 (2015)
14. Naaman, A.E.: Fiber reinforcements for concrete: looking back, looking ahead. In: RILEM
Proceedings. PRO 15. RILEM Publications SARL (2003)
15. Tanesi, J.: The Influence of Polypropylene Fibers on Shrinkage Crack Control. Polytechnic
School of - University of São Paulo, São Paulo (1999)
16. Toledo, R.D.: Composite Materials Reinforced With Natural Fibers: Experimental
Characterization. Pontifical Catholic University of Rio de Janeiro, Rio de Janeiro (1997)
17. Vassaneli, E., Micelli, F., Aiello, M., Plizzari, G.: Crack width prediction of FRC beams in
short and long term bending condition. Mater. Struct (2012)
Load-Carrying Capacity of SFRC Suspended
Slabs with Different Support Conditions
Olugbenga B. Soyemi1 and Ali A. Abbas2(&)

1
The Federal Polytechnic Ilaro, Ilaro, Nigeria
2
abbas@uel.ac.uk
Abstract. Traditionally, steel fibres have been used in ground- and pile-
supported concrete slabs. More recently, the structural use of steel fibres in
suspended concrete slabs is on the rise. In the present research study, the
structural behaviour of suspended steel-fibre-reinforced concrete (SFRC) slabs
was investigated by means of non-linear finite-element analysis (NLFEA). The
numerical models were calibrated (at the material level) and validated (at the
structural level) using existing experimental data. NLFEA-based parametric
studies were carried out to extend the range of Vf beyond the one considered in
the experiential investigations. The aim was to determine the correlation
between the slab load-carrying capacity and the corresponding fibre dosage
increase. This is useful for design purposes and can provide guidance on the
amount of fibres needed to achieve a certain prescribed load resistance level. It
was found that the addition of steel fibres improved the load-carrying capacity
and changed the failure mode to a flexural ductile one. Other critical structural
performance indicators were also examined such as deflections, ductility, energy
absorption and cracking patterns. So, in essence both the ultimate and ser-
viceability limit states were investigated. Comparisons with current design
guidelines were also made and recommendations for improvements proposed.
Keywords: Suspended slabs Load-carrying capacity Design Ductility

Validation NLFEA
1 Introduction
Concrete in its ordinary state comprise of cement, fine and coarse aggregate, water and
where required, an additive. It has emerged the most used human-made construction
material of all time (Beddar 2004). Significant improvement has been recorded both in
the art and usage of concrete over the years compare to when it was first used for
masonry work by the Roman in which pozzolanic mortar was used as binder. Nor-
mally, concrete is strong in compression (Ali A. Abbas, Syed Mohsin, Cotsovos, and
Ruiz-Teran 2014; America Concrete Institute 2002; J.A.O. Barros and Cruz, 2001;
British Standard Institute, 1997 2004; Crowther 2009; T. S. Lok and Xiao 1999) and
possesses about one-eighth of this strength in tension. At the application of loads, the
concrete member bends and cracks at the tension face. In suspended structural mem-
bers, where tension is paramount in the design, the members are constructed using
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_54
Load-Carrying Capacity of SFRC Suspended Slabs 597
concrete reinforced with longitudinal steel rebar or welded mesh fabric. An increasing
trend, both in the UK and overseas, in the method of construction, is the replacement of
some or all the reinforcement with fibres (A. Abbas, Syed Mohsin, and Cotsovos 2014;
Ahmed, 2014; America Concrete Institute 1999; ArcelorMittal, 2011; J.A.O Barros,
Salehian, Pires, and Gonçalves, 2012; Beddar 2004; Blanco, Pujadas, Cavalaro, de la
Fuente, and Aguado 2014; Cameron, 2002; Cerioni, Meda, & Plizzari, 2004; Chanh,
1990; Crowther, 2009; Destrée, 2001; J. Hedebratt & Silfwerbrand, 2008; McCraven,
2002; Mohsin, 2012a; Rana, 2013; RILEM, 2002b; Singh, 2015; Swamy, 1974; Zollo,
1997).
1.1 Historical Background

Once cracking stresses are reached, an unreinforced concrete matrix fails in a brittle
manner. This is due to its brittleness and lack of post-cracking resistance. Fibres have
been found to enhance the ductility of concrete, by carrying stresses beyond matrix
cracking and thus improve the structure’s integrity and ability to accommodate
deformations. The use of SFRC in construction of different types of structural system
has recently been explored. This type of slab is generally designed as pile-supported or
Elevated SFRC (Destree 2004)
In 1874, an American, A. Berard patented an idea in which grains of steel left-overs
(waste) are mixed to the concrete to create a material that is more ductile (Berard 1874).
Today, different types of fibres are used in the construction industry for structural and
non-structural works. Examples are steel, carbon, polyethene, natural fibres and so on,
but the most researched on is the steel fibre. Earlier researches revealed the advantages
of adding steel-fibres in concrete in structural members and these include [i] enhancing
the post-cracking behaviour of the concrete regarding more ductility which leads to an
increase in flexural strain capacity, [ii] the significant reduction in construction time
(since they are conveniently cast in large pours compared to time to lay the traditional
reinforcement); [iii] improve tensile strength and energy absorption and, [iv] reduction
in crack-width giving rise to enhanced damage tolerance (T.-S. Lok and Pei 1998;
Oslejs, 2008; RILEM, 2002a). Though, in some early works on SFRC provisions were
made for rebar to resist peak moments over piles or supports, notably in the corner
panel, and to control cracks due to early-age thermal stresses and then it was concluded
that it will be better to not have steel-fibres only reinforced concrete (Chanh, 1990;
IstructE 1999). However recent experimental works on full-size specimen have shown
that there can be fibres-only reinforced concrete (J.A.O Barros et al. 2012; Destree
2004; Destrée 2007b; Destrée and Jürgen 2008; Jerry Hedebratt and Silfwerbrand
2014; Khan, Ansari, Saroj, Haider, and Kulkarni 2016; Lee, Cho, and Choi, 2017;
Najim, Saeb, and Al-Azzawi, 2018; Ragi 2015; Vollum 2007)
Earlier studies carried out by (J. A. Barros, Cunha, Ribeiro, and Antunes 2005;
Beckett 2004; Cameron 2003; Destrée 2001; H Thooft 1999) only consider a low range
of volume fibres fraction and are restricted to certain types of structural arrangement
due to time and financial limitations. These studies are mainly experimental, which
does not have the advantage of implementing full parametric examinations such as the
one carried out in this present research work. In time past, in order to control cracking
and improve on the tensile strength of ground supported slabs, fibre were been added
598 O. B. Soyemi and A. A. Abbas
(America Concrete Institute 2002; J.A.O. Barros and Figueiras, 2001; Beddar 2004;
British Standard Institute 2006; Chanh, 1990; Chen, 2005; Crowther 2009; Eddy 2008;
Edgington, 1973a; Ghaffar, Chavhan, and Tatwawadi, 2014; Jansson, 2011; Narayan,
2014; Swamy and Lankard, 1974; Zollo 1997). There have been tremendous works on
numerical investigations on ground-supported slabs (Ali A Abbas, Pavlović, and
Kotsovos 2004; J.A.O. Barros and Figueiras, 2001; Cerioni et al. 2004; Maya, Fer-
nández Ruiz, Muttoni, and Foster, 2012) and minimal numerical investigation works on
piled floors.
1.2 Design Methods

The existing design methods are either based on analytical material properties of SFRC
established from tests on small beam specimens or empirically based on full-scale
testing. General design guidelines are based on both elastic and yield-line analyses to
calculate peak moments (Eddy 2008; The Concrete Society TR63 2007; Vollum 2007)
alongside some national guides (America Concrete Institute 2002; British Standard
Institute 2006; INRC 2007; Institute, 2011; NZS 2006). Proprietary design guidance
were also provided by Bekaert and ArcelorMittal, which are the principal suppliers of
steel fibres (Dramix® and TAB-Structural system) in the UK. The Bekaert design
method (Viney 2007) relies on the elastic moment coefficient of the Dutch Code NEN
6720 and is thus similar to TR 63 guidelines (Concrete-Society 2007). Only uniformly-
distributed loads are considered, and conventional steel rebar is provided within the
slab at the point of contact with the piles to resist hogging moments, but no rebar is
provided in the interior of the slab (Viney 2007).
Yield line analysis is also incapable of providing information regarding deforma-
tions (deflections and slopes) or cracking, which are critical for suspended floors to
ensure the safe operation of mechanical handling equipment (e.g. the lift truck is
susceptible to slight tilts of the floor). Current design provisions are unsatisfactory as
they adopt simple span/effective depth limits which were developed for flat slabs in
buildings where the aim is to minimised cracking in finishes and partitions. Therefore,
there is a need for appropriate limits to be derived to suit stringent surface regularity
requirements in industrial floors. Current design guidance accounts only for uniformly
distributed loads and concentrated load (whether single or multiple), often encountered
in practice, are not considered. Although most floors are constructed using proprietary
designs, the latter are excluded from the published guidance and require critical
examination. The treatment of crack control due to restrained early-age and drying
shrinkage is not precise, especially in the ArcelorMittal method which eliminates
conventional reinforcement. Crack control measures are critical to piled floors since the
slab is often cast in large pours with no internal joints (Beckett, 2004; Goodchild 2004).
There is an acknowledged scatter in the thicknesses recommended by current design
guidelines as they utilise a range of material and loading parameters, effective spans
between piles and safety factors resulting in thicknesses that could differ by 25% or
even more, with proprietary designs providing the smaller slabs sections as there will
be no provision for cover. (Jerry Hedebratt, 2012; Hulett and Sketchley 2009). The
performance of existing slabs is also a source of concern. There is a need for a unified
and well-defined design method.
2 Case Studies
This paper presents two case studies on single SFRC elevated round-panel with dif-
ferent boundary conditions, which are based on the experimental investigations of de
Montaignac et al (2011) and Soranakom et al (2007). The post-cracking behaviour of
SFRC structural elements was the focus of the research work by de Montaignac et al
(2011). Round-plates were examined (see Fig. 1) consisting of six specimens for each
SFRC mix with varying amount of steel fibres [0.75–1.25%]. For this study, the SFRC
round-plate [labelled F35-1.0] chosen was made from 1% volume fraction of fibres, has
the mechanical properties with compressive strength 46.9 MPa, modulus of elasticity
33.5GPa, Poisson ratio of 0.23 and the post-cracking tensile strength of 2.6 MPa. The
hooked-end steel-fibre used in the concrete matrix has an aspect ratio of 67.2 and a
length of 35 mm. The strength of the fibre is 1200 MPa.
Fig. 1. Dimensions of round-plate adapted from (de Montaignac, Massicotte, Charron, and
Nour, 2011)
The second case study examined the experimental work of Destree & Mandl (2008)
as reported by Soranakom et al (2007), where load-displacement responses were
established experimentally using SFRC round panels (see Fig. 2) with 1500 mm span
and 150 mm thick, in inverse analysis to determine the properties of material (Young
modulus, tensile stress, Poisson’s ratio). The round panel was placed on a continuous
simple support. The panel test was mainly for characterisation of the materials. One
cubic meter of the SFRC were made of 350 kg Cement (type I), aggregate grading of
0–16 mm, 60 kg fly ash, 1.25% superplasticizer by volume, Undulated steel wire,
TABIX 13/50 steel-fibres [1.3 mm diameter, 50 mm length and 900 MPa tensile
strength and 0.50 water-cement ratio.
At 28 days, the average crushing strength of cube test was 43.7 MPa. The panels
were loaded to failure. Their crack patterns are shown in Fig. 3 [a-c]. The load-
displacement response is the core of the round panel tests, which is used for material
characterisation purposes. The research works used inverse analysis to determine the
properties of material of steel-fibre reinforced concrete [Young’s modulus, Poisson’s
ratio and tensile stress crack width parameters] from the load-deflection response of the
round-plate test [from which the load carrying capacity of the round-panel is known].
Fig. 2. Experimental setup of Destree & Mandl (2008)
Fig. 3. Crack Patterns of Soranakom et al. round plate (2007)
3 Numerical Modelling
The FEA was carried out using ABAQUS to simulate the behaviour of the SFRC
panels. ABAQUS has a materials library that has several models of the elastic and
inelastic behaviour of various materials which include concrete, cast iron, metals, rock
and soils. There are models embedded in it that can be used to model concrete and
masonry structures. ABAQUS/Standard utilises the FEM to indirectly solve a system of
equations at each solution “increment” for the analysis of framework, shell, and, solid
models [ABAQUS, 2012].
3.1 Material Properties

Material properties for all elements are specified; however, high-quality material data
were difficult to obtain, particularly for the more complex material models like material
damage properties. The validity of the results is essentially limited by the accuracy and
extent of the material data. Here linear elasticity and nonlinear plasticity of rein-
forcement steel and isotropic elasticity in combination with damaged plasticity model
of concrete is discussed to depict a mechanical constitutive model.
3.2 Constitutive Model for SFRC in Compression

From various experimental and numerical results, the compressive strength of steel-
fibre reinforced concrete [with volume fraction of fibre 1%] and that of plain
concrete are about the same, but a higher percentage of fibre fraction tends to produce a
slight increase in compressive strength but they are generally assumed to be the same
(America Concrete Institute, 1999 2002; J.A.O. Barros and Cruz, 1999, 2001;
Joaquim A. O. Barros, Lourenço, Soltanzadeh, and Taheri, 2013; J.A.O Barros et al.
2012; British Standard Institute, 2006; Carr, 2000; Chanh, 1990; Chen 2005; Edg-
ington, 1973b; Fall, 2014; Fathima and Varghese, 2014; Garcia and Borrell, 2010;
Ghaffar et al. 2014; Ghosh, Bhattacharjya, and Chakraborty 2007; Hadi 2008; Jansson,
2011; Kooiman, 2000; T.-S. Lok and Pei 1998; Mohsin 2012b; N., 2009; Naaman and
Gopalaratnam 1983; Neal 2000; RILEM, 2002b; Rossi, 1999; Swamy 1974; Van
Chanh, 2004; Wang, Mindess, and Ko, 1996; Xie, Guo, Liu, and Xie, 2015; Zollo
1997).
3.3 Constitutive Model for SFRC in Tension

Different constitutive models for post-cracking behavaviour of SFRC were considered,
and Lok and Xiao (1999) model was selected as the most suitable to simulate the
experimental result and mimic tension behaviour.
3.3.1 Lok and Pei (1998)

The proposed constitutive model was based on the stress-strain relationship. Its tensile
behaviour incorporates a bilinear strain softening. The model is defined by the prop-
erties of the composite material and that of steel-fibres. The first phase consists of a
parabolic curve to point of first crack. Phase two [2] adopted a bilinear descending
process to simplify the post-cracking phase. The main challenge of this model is that it
is limited to low fibre dosage.
Fig. 4. Tensile stress-strain relationship adapted from Lok and Pei (1998)
3.3.2 Barros and Figueiras (1999)

In their work on “Flexural Behaviour of SFRC: Testing and Modelling, the outcomes
of tests performed on specimens, using the 3-point bending test for steel fibre rein-
forced concrete (SFRC) were presented. The tensile constitutive model is depicted in
Fig. 5.
Fig. 5. Proposed tensile stress-strain diagram adapted from (Barros and Figueiras, 1999)
The fracture energy model was used to evaluate the post-peak tensile behaviour,
and a numerical model was developed. A bilinear tensile stress-strain relationship was
used to depicts the strain softening behaviour of the beam. The model is based on
stress-strain relationship of the SFRC
3.3.3 RILEM TC 162-TDF (2002)

In a 3-point bending test on several beams, the RILEM TC 162-TDF Recommendation
(2002) proposed a stress-strain relation for SFRC with compressive strengths up to
C50/60. The design of SFRC according to the r e method [Fig. 6]
Fig. 6. Stress-Strain Diagram adapted from (RILEM, 2002)
3.3.4 Lok and Xiao (1999)

Based on their previous works which resulted in the accurate prediction of the steel-
fibre reinforced concrete (SFRC) flexural response from a constitutive model, unam-
biguous expressions for the first crack and the ultimate flexural strength are derived in
the work.
There was good agreement when the Predictions obtained by using the simplified
ultimate strength expression are compared with results obtained from both the above
analytical procedure and experimental data. The approaches serve as an effective tool
for flexural strength assessment of SFRC.
The three constitutive models were used in the analysis of a simply supported
SFRC beam with the results displayed in Fig. 8.
The Lok and Xiao (1999) constitutive model is chosen for works done based on its
agreement with experimental results and also due to its incorporation of high volume
fibre fraction (0-3%) while Barros and Figueiras (1999) model have a maximum
volume fibre fraction of 1%.
Fig. 7. Stress-Strain constitutive model adapted from (Lok and Xiao, 1999)
Fig. 8. Load-displacement curve based on different constitutive models
50.00
40.00
Load [kN]
30.00
20.00 Experimental
10.00 ABAQUS-CDP
0.00
0.00 10.00 20.00 30.00 40.00
Displacement [mm]
Fig. 9. Load-Displacement Curve de Montaignac et al Round-Plate (2011)
3.4 Validation and Calibration Using Experimental Data

For the ABAQUS-CDP, the plasticity parameters are given as follows: w = 400, e =
0.1, fb0=fc0 = 1.16, K=0.667 and l= 0.006. The compression and tensile stresses and
damaged parameters provided at 80% tensile damaged. The displacement loading of
160.00
140.00
120.00
Load [kN]
100.00
80.00
60.00 Experimental
40.00 ABAQUS-CDP
20.00
0.00
0.00 10.00 20.00 30.00 40.00
Displacement [mm]
Fig. 10. Load-Displacement curve for Soranakom et al Round Plate (2007)
70.00
60.00
Vf-1.0%
50.00
Load [kN]
40.00 Vf-1.25%
30.00 Vf-1.50%
20.00
Vf-1.75%
10.00
0.00 Vf-2.00%
0 10 20 30 40 50 Vf-2.50%
Displacement [mm]
Fig. 11. Load-displacement curve for a range of Vf
4 mm was applied at the centre of the beam at a slow incremental rate of 0.04 mm.
This is to reduce convergence challenges. The convergence criteria based on the
residual forces tolerance is maintained at 5%. The bond stress used for this analysis is
3.0 MPa and 25 mm mesh size.
The flexural behaviour of the round panel was modelled using ABAQUS Concrete
Damaged Plasticity (CDP) nonlinear model. To really capture the panel, the full panel
was modelled using 3D brick model with 25 mm mesh. The loading on the round panel
was modelled using static analysis and displacement based loading. The compressive
and shear stresses are assumed to develop within the elastic range thus guided by the
elastic properties of the material [young’s modulus and Poisson’s ratio]. The tension
behaved elastically until it cracks and then conform to the tensile-strain (Loks and Xiao
1999) relationship afterward. Figures 4 and 5 show the Load-Displacement curve for
de Montaignac et al. (2011) and Soranakom et al. (2007) round plates. The young’s
modulus controls the rising elastic response of the curve from the origin up to the first
250.00
200.00 Vf=1.00%
Load [kN]
150.00 Vf=1.25%
100.00 Vf=1.50%
50.00 Vf=1.75%
0.00 Vf=2.00%
0.00 10.00 20.00 30.00 40.00 Vf=2.50%
Displacement [mm]
Fig. 12. Load-displacement curve for a range of Vf for Round-Panel
crack. The point that the load-displacement curve diverges from the linear elastic
response correlates to the tensile strength parameter ft . The maximum load is governed
by the first softening of the slope and the ft while the post cracking response is
governed by the tensile stress-strain parameters (Loks and Xiao 1999). The load-
displacement curve was plotted in Fig. 5, which shows the experimental result with an
ultimate load of 138.6kN. The NLFEA gave an ultimate load of 140.09kN. This is
1.08% higher than the experimental value. The response of ABAQUS CDP was in
agreement with the experimental results, thou stiffer. This shows the conservativeness
of the model. At the peak load, the concrete cracked and the fibre held the composite
together while undergoing to softening behaviour.
4.1 Parametric Studies

The FEA is used to extend the experimental studies and cover the full practical range of
the key parameters, i.e the fibre volume ratio and the characteristic strength of the
SFRC. The result of the experimental round-plate SFRC with Vf = 1.00% is used as a
control. The experimental work examined the effect of adding fibres to suspended
panels up to 1.5%. This case study extended the fibre volume ratio to full range as
practically obtainable in the industry, 2.5%.
4.2 Case Study1: 3-Point Support

For the round-plate with three supports. the load-displacement curves for the varying
volume fibre fraction are shown in
Fig. 6. These are also interpreted in Table 1,
namely: the load at first crack Py signifying the yield load and its associated dis-
placement ½dmax the maximum load ½Pmax signifying the load-carrying capacity and
associated displacement ½dmax , the ultimate load ½Pu signifying the residual strength
(taken as the lesser of the load at failure or 85% of the maximum load to ensure its
practical usefulness) and associated displacement ½du and the ductility ratio l defined

as l ¼ du dy For comparison purposes, the SFRC with Vf = 1.00% is taken as the
control specimen. The maximum load is the measure of the strength of the slab while
the ultimate load is a measure of the residual strength. Likewise, the displacement at
yield and ultimate loads are dy and du respectively. The ductility of the slab is measured
as the ratio of dy to du .
Table 1. Results for SFRC round plate

l
Vf Py dy Pmax [kN] PPmaxc
max dmax Pu du Py
l ¼ dduy le
Pye
% [kN] [mm] [mm] [kN] [mm]
Control 1.00 38.21 0.51 38.21 0.51 32.48 1.75 3.43
1.00 38.21 0.51 38.21 1.00 0.51 32.48 1.75 1.00 3.43 1.00
1.25 39.08 0.51 40.51 1.06 1.01 34.43 3.28 1.02 6.43 1.87
1.50 40.31 0.51 43.24 1.13 1.53 36.75 3.22 1.05 6.31 1.84
1.75 41.68 0.52 47.73 1.25 1.71 40.57 6.15 1.09 11.83 3.45
2.00 42.77 0.52 52.44 1.37 2.13 44.57 6.32 1.12 12.15 3.54
2.50 57.11 1.16 63.91 1.67 4.82 54.32 6.62 1.49 5.71 1.66
*Pmaxc represent the control peak load
In the SFRC with fcu = 30 MPa, the result shows that round-panel with 1% fibre
volume ratio yielded 38.21 kN, while that of 2.50% fibre volume is 57.11 kN in
control. As the fibre volume is increased, there is a gradual increase in the yield load.
There is also an increase in the stiffness as the fibre volume increases. The strength of
the slab was also enhanced as the fibre volume increases. An 18% and 22% increment
in strength was obtained in the slab when the fibre was increased to 1.25% and 1.5%
respectively. Incidentally, all the FEA results for the displacement at the yield load give
0.51 mm.
4.3 Case Study2: Continuous Support

In the case of the round-panel with continuous support along the edge, the load-
displacement curves for the varying fibre volume fractions are shown
in Fig. 6. These
are also interpreted in Table 2, namely: the load at first crack Py , its displacement
½du , the maximum load ½Pmax , its displacement ½dmax , the ultimate load ½Pu , its
displacement ½du and the ductility ratio ½l. The load-displacement curves of the slab
with different amount of fibre volume ratio shows that the maximum loads were
reached for vf = 1.00−1.50% before 3 mm central displacement while for vf = 1.75
−2.50%, they were reached after 3 mm. These resistances are due to the amount of
fibres in the concrete matrix. For all the specimens under study, the first crack was
noticed before 5 mm displacement. The fibre volume ratios influence can be observed
clearly as the fibre content increases.
Table 2. Results for SFRC round plate

l
Vf Py dy Pmax [kN] ½dmax PPmaxc
max Py
Pyc
Pu ½du l ¼ du
dy ls
% [kN] [mm] [mm] [kN] [mm]
Control 1.00 132.91 1.11 132.9 1.11 113.0 14.5 13.06
1.00 132.91 1.11 132.9 1.11 1.00 1.00 113.0 14.5 13.06 1.00
1.25 143.79 1.85 143.8 1.85 1.08 1.08 122.2 17.5 9.46 0.72
1.50 155.51 3.03 155.5 3.03 1.17 1.17 132.2 26.8 8.84 0.68
1.75 166.82 3.10 170.4 14.36 1.28 1.26 144.8 23.7 3.05 0.23
2.00 184.19 3.40 189.7 15.24 1.43 1.39 161.2 21.1 2.36 0.18
2.50 215.64 3.60 221.8 14.74 1.67 1.62 188.5 22.1 2.42 0.19
It can be seen that the load-carrying capacity ½Pmax of the round-panels were
improved by the increment in steel fibres. The enhancement was evident in all the
round-panels as the values of fibre volume increases which lead to higher strength. The
curves also show that stiffness was enhance with fibres between 1.00–1.50%. The
round-panels deflects more with 1.750–2.50% fibre volume.
5 Conclusions
It was found that the load-carrying capacity of the plates are enhanced by the addition
of steel fibres and ensures a more ductile structural response (thus avoiding a brittle
failure mode), which is preferred in design. Likewise, the strength increases by *7%
over the previous result of the round-panel with isolated supports and *11% for the
round-plates with continuous support under monotonic loading. This also leads to a rise
in ductility and strength enhancement.
References
beam-column joints. Eng. Struct. 59, 261–283 (2014a)
Abbas, A.A., Pavlović, M., Kotsovos, M.D.: Permissible-stress design of ground-floor slabs.
Proc. ICE-Struct. Build. 157(6), 369–384 (2004)
Abbas, A.A., Syed Mohsin, S.M., Cotsovos, D.M., Ruiz-Teran, A.M.: Shear behaviour of steel-
fibre-reinforced concrete simply supported beams. Proc. ICE – Struct. Build. 167(9), 544–558
(2014b). https://doi.org/10.1680/stbu.12.00068
America Concrete Institute. State-of-the-Art Report on Fibre Reinforced Concrete: American
Concrete Institute (2002)
ArcelorMittal. Steel Fibre Reinforced Concrete (SFRC) for Industrial Floors Especially Slabs
without Joints and Slabs on Piles. (2011). www.arcelormittal.com (Ed.): ArcelorMittal
Barros, J.A., Cunha, V.M., Ribeiro, A.F., Antunes, J.: Post-cracking behaviour of steel fibre
reinforced concrete. Mater. Struct. 38(1), 47–56 (2005)
Barros, J.A.O., Figueiras, J.A.: Model for the analysis of steel fibre RC slabs on grade. Comput.
Struct. 79, 97–106 (2001)
Beddar, M.: Fibre-reinforced concrete: past, present and future. Concrete 38(4), 47–49 (2004)
Berard, A.: Improvement in artificial stone: Google Patents (1874)
Blanco, A., Pujadas, P., Cavalaro, S., de la Fuente, A., Aguado, A.: Constitutive model for fibre
reinforced concrete based on the Barcelona test. Cemt. Concr. Composites 53, 327–340
(2014). https://doi.org/10.1016/j.cemconcomp.2014.07.017
British Standard Institute. BS EN 1992, Eurocode 2, Design of concrete structures, Part 1–1:
General rules and rules for buildings, British Standard Institute London (2004)
British Standard Institute. BS EN 14889-1-2006 - Fibres for concrete - Steel fibres, British
Standard Institution, London (2006)
Cameron, G.: Steel-fibre-reinforced pile-supported car park in Belfast. Concrete (2002)
Cameron, G.: Providing a floor slab fit for rolls-royce. Concrete 37, 36–37 (2003)
Chen, S.:Strength of steel fibre RC GFS. Proceedings of the Institution of Civil Engineers
Structures & Buildings, 157 (2005)
Concrete-Society. TR63 - Guidance for the design of Steel Fibre Reinforced Concrete (2007).
(ISBN 1-904482-32-5)
Crowther, D.: Fibre-reinforced concrete - origin of the species. Concrete 43(3), 22–23 (2009)
de Montaignac, R., Massicotte, B., Charron, J.-P., Nour, A.: Design of SFRC structural elements:
post-cracking tensile strength measurement. Mater. Struct. 45(4), 609–622 (2011)
Destree, X.: Structural application of steel fibre as only reinforcing in free suspended elevated
slabs: conditions-design-examples. In: Paper presented at the 6th International RILEM
Symposium on Fibre Reinforced Concretes (2004)
Destrée, X.: Steel fibre reinforcement for suspended slabs. Concrete 35(8), 58–59 (2001)
Destrée, X.: Steel-fibre-reinforced pile-supported slabs. Concrete, 41(4) 26–27 (2007a)
Destrée, X.: Structural steel-fibre-reinforced concrete construction. Concrete, 41(8) 23–24
(2007b)
Destrée, X., Jürgen, M.: Steel fibre only reinforced concrete in free suspended elevated slabs:
Case studies, design assisted by testing route, comparison to the latest SFRC standard
documents. Tailor Made Concrete Structures, pp. 437–445 (2008)
Eddy, D.: Fibre-only suspended ground-floor slabs - do they work. Concrete 48(8), 14–15 (2008)
Ghosh, S., Bhattacharjya, S., Chakraborty, S.: Compressive behaviour of short-fibre-reinforced
concrete. Mag. Concr. Res. 59(8), 567–574 (2007). https://doi.org/10.1680/macr.2007.59.8.
567
Goodchild, C.: Design guidance on large area pours for suspended slabs. Concrete 38(6), 19–20
(2004)
Hadi, M.N.S.: Behaviour of FRC slabs. Paper presented at the Structural Engineering and
Construction Conference, London (2008). http://ro.uow.edu.au/engpapers/2754
Hulett, T., Sketchley, C.: Design of pile-supported concrete industrial floors. Concrete 43(3), 45–
46 (2009)
INRC. Guide for the Design and Construction of Fiber-Reinforced Concrete Structures NRC.
Italian National Research Council, Rome (2007)
Institute, A.C.: Building Code Requirement for Structural Concrete (ACI 318-11) and
Commentary, pp. 1–509. America Concrete Institute, Farmington Hills, Michigan (2011)
IstructE. Interim Guidance on The Design of Reinforced Concrete Structures Using Fibre
Composite Reinforcement. The Institution of Structural Engineers, London (1999)
Lok, T.-S., Pei, J.-S.: Flexural behavior of steel fibre reinforced concrete. J. Mater. Civ. Eng.
ASCE 10(2), 0086–0097 (1998)
Lok, T.S., Xiao, J.R.: Flexural strength assessment of steel fiber reinforced concrete. J. Mater.
Civ. Eng. ASCE 11(3), 0188–0196 (1999)
Loks, T.S., Xiao, J.R.: Steel-fibre-reinforced concrete panels exposed to air blast loading. Proc.
Instn. Civ. Engrs Struct. Bldgs 134, 319–331 (1999)
Mohsin, S.M.B.S.: Behaviour of Fibre-Reinforced Concrete Structures Under Seismic Loading.
(PhD), Imperial College London (2012)
Naaman, A.E., Gopalaratnam, V.: Impact properties of steel fibre reinforced concrete in bending.
Int. J. Cem. Composites Light. Concr. 5(4), 225–233 (1983)
Neal, F.: Design of steel fibre floors - problems and pitfalls. Struct. Eng. 78(11), 14–16 (2000)
Rilem, T.: RILEM TC 162-TDF - test and design methods for SFRC - principles and
applications. Mater. Struct. 35, 262–278 (2002)
Rilem, T.: RILEM TC 162-TDF - Test and design methods for Steel Fibre Reinforced Concrete -
Bending test - final Recommendation. Mater. Struct. 35, 579–528 (2002b)
Swamy, R.N.: The Technology of Steel Fibre Reinforced Concrete For Practical Applications.
Paper presented at the Proc. Instn Civ. Engrs (1974)
Swamy, R.N., Lankard, D.R.: Some practical applications of steel fibre reinforced concrete.
Proceedings, 56, 143–159 (1974)
The Concrete Society TR63. TR63 - Guidance for the design of Steel Fibre Reinforced Concrete,
vol. 63, pp. 1–109. The Concrete Society, UK (2007)
Thooft, H.: Design of steel fibre reinforced floors on foundation piles. paper presented at the
specialist techniques and materials for concrete construction In: Proceedings of the
International Conference Held at the University of Dundee, 8-10 September 1999. Scotland,
UK on (1999)
Thooft, H.: Structural behaviour of steel fibre reinforced pile-supported concrete floors. Concrete
34(8), 50–54 (2000)
Viney, T.: Piled floor design - the Dutch method. Concrete 41(2), 10–11 (2007)
Vollum, R.: Design of steel-fibre-reinforced pile-supported slabs. Concrete 41(2), 12–14 (2007)
Destrée, X., Mandl, J.: Steel fibre only reinforced concrete in free suspended elevated slabs: Case
studies, design assisted by testing route, comparison to the latest SFRC standard documents.
Tailor Made Concrete Structures, 437–443 (2008)
Discrete Element Simulation of the Fresh State
Steel Fiber Reinforced Self-compacting
Concrete
A. Najari1(&), A. Blanco2, A. de la Fuente1, and S. H. P. Cavalaro2

1
Department of Construction Engineering,
Universitat Politecnica de Catalunya, Barcelona, Spain
alireza.najari@upc.edu
2
School of Architecture, Building and Civil Engineering,
Loughborough University, Loughborough, UK
Abstract. The distribution and orientation of fibers in Steel Fiber Reinforce-

ment Self-Compacting Concrete (SFRSCC) are paramount given its influence in
the mechanical properties of the material/structural elements.
A two-way coupled model based on Discrete Element Method (DEM) to
simulate the flow of clumps of particles with the high aspect ratio as fibers and
two-phase particles as concrete is developed and introduced briefly in this paper.
The framework is capable of modelling of the movement (translation and
rotation) including the separation and contact detection of the particles. The
interaction of the particles is treated as a dynamic process with a developing
state of equilibrium whenever the internal forces are in balance. Newton’s laws
of motion provide the fundamental relationship between particle motion and the
forces causing that motion.
Several experiments were conducted and analyzed by means of the inductive
test method (a non-destructive method to assess steel fiber content and orien-
tation). Orientation numbers coming from the inductive test method were
compared with the simulation to verify the ability of the model to properly
represent the flow of the fresh state SFRSCC. Through comparison with the
experimental data, it is shown that the numerical model predicts the final dis-
tribution and orientation of the fibers sufficiently accurate and in a reasonable
amount of time. The results obtained represent a step forward, showing that it is
possible to apply advanced numerical tools for a preliminary assessment of the
performance of SFRSCC, which might have positive implications through
improved reliability of the design procedures.
Keywords: Steel Fiber Reinforcement Self-Compacting Concrete DEM

Simulation Orientation and Distribution of the Fibers
1 Introduction
Steel fiber reinforced self-compacting concrete (SFRSCC) is a type of cementitious

composite construction material that emerged recently [1, 2]. Since SFRSCC is a
complex material, its composition makes it extremely difficult to predict the final
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_55
Discrete Element Simulation of the Fresh State Steel Fiber 611
detailed behavior of structural elements. It cannot be assumed to be an isotropic

material any more since the fibers orient and disperse during the flow. The knowledge
of final orientation and distribution of fibers in the structural elements could provide a
basis for understanding how the fibers influence the final mechanical properties of the
structural elements [3].
Experimental work leading to the knowledge of orientation and distribution of
fibers is often a very time and resource consuming procedure. At first, someone has to
do the casting of the elements. The cast element has to be left to harden, cut into pieces
and only then a computational tomography (CT) scanner can be used to obtain the 3D
image of the fibers in the elements. However, only small parts of the elements can be
CT scanned due to the fast overheating of the device [4]. An alternative method to
assess the fiber content and orientation in SFRC element is inductive method. Despite
several advantages, it only gives average values without providing any information on
the scatter or probabilistic distribution of the fiber orientation [5]. A completely dif-
ferent approach might be to use a transparent yield stress fluid such as Carbopol [6] to
replace the fluid matrix of self-compacting concrete. All these approaches are not
simple and, therefore, only a limited amount of information is obtained from such
experiments. On the other hand, numerical simulations are limited only by the com-
putational power. A simulation tool capable of simulating a flow of self-compacting
concrete together with fibers and the largest aggregates could provide a sufficient
alternative for obtaining the required information.
The objective of this study is to propose and validate a numerical approach for the
simulation of the casting process of SFRSCC in fresh state. For this subject, a set of
constitutive laws that represents the interaction of particles (aggregates) surrounded by
a layer of cement paste and clumps of particles as high aspect ratio steel fibers in fresh
state self-compacting concrete is proposed. Then an experimental program related to
the casting process of the SFRSCC in was conducted. Finally, calibration of the
constitutive law with the experimental results are performed.
2 Applications
Any simulation tool should be validated preferably against both analytical solutions
and as many experiments as possible. One of the main objectives of this study is to
show the capability of the model to properly describe the complex behavior of steel
fiber orientation and distribution in the self-compacting concrete. To do so, the ori-
entation of fibers as a result of the simulation model is compared with the orientation of
fibers in the real experiment.
3 Experimental Investigation
The production and characterization of the SFRSCC was performed at the Laboratory
of Technology of Structures Luis Agulló (UPC). Six cubic sample compositions with
the dimension of 15 cm were defined and produced (see Fig. 1). The compositions
used were selected in order to reproduce those typically found in practice. After the
612 A. Najari et al.
production, inductive test was conducted to evaluate the fiber orientation of the mixes.
The material composition and the notation used to refer each element is detailed in
Table 1.
Fig. 1. Casting of the cubic elements
Table 1. Material composition and notation used to refer each element.

Components Characteristics Content ðkg=m3 Þ
Cement CEM II 42.5R 425
Sand 0–5 mm Granite 1325
Sand 5–10 mm Granite 150
Gravel 10–20 mm Granite 250
Water 200
Superplasticizer Pozzolith 7003 4.5
Hydration activator Master paste 3850 5.85 6.75 7.95
Fibers Steel fiber 60 90 120
Reference code C1A, C1B C2A, C2B C3A, C3B
3.1 Test Method

To evaluate the orientation of the steel fibers, the “inductive test method” [7] improved
by Cavalaro et al. [8] has been used. The inductive test method allows assessing the
amount and orientation of steel fibers in cubic specimens. The method is based on the
ferromagnetic properties of the steel fibers that are able to alter the magnetic field
around them. Therefore, if an inductive coil acting as a sensor is placed wrapping the
specimen, the fibers will affect the inductance of the sensor. An increase of the
inductance occurs when the fibers are located in a position parallel to the direction of
the magnetic field (parallel to the axis of the coil), whereas those located in a per-
pendicular direction cause practically no variation [9].
The testing procedure for a cubic specimen is very simple: the cubic specimen is
located on a non metallic surface with the concrete-pouring face upwards (axis Z), for
instance. The specimen is then wrapped by the inductance generated by a coil, and the
increase of the inductance is measured with an impedance analyzer (see Fig. 2).
The same procedure is repeated with the specimen turned towards the axes Y and X.
Cylindrical specimens can also be measured using same technique by spinning the
specimen around specific directions. Further information about the method may be
found in [10].
Fig. 2. Equipment of the inductive method: analyzer and cylindrical sensor.
4 Simulation
The method used in this study is a variation of the DEM, which allows the modelling of
movement (translation and rotation) including the separation and contact detection of
particles [11]. The interaction of the particles is treated as a dynamic process with a
developing state of equilibrium whenever the internal forces are in balance. A general
particle-flow model simulates the mechanical behavior of a collection of arbitrarily
shaped particles in a system. The term particle denotes a body that occupies a finite
amount of space and are assumed to be rigid. Hence, the mechanical behavior of such a
system is described in terms of the movement of each particle and the inter-particle
forces acting at each contact point. Newton’s laws of motion provide the fundamental
relationship between particle motion and the forces causing that motion.
4.1 Concrete Particle Definition

In general, DEM that are capable of simulating the flow analysis of a concrete mixture
can be divided into three main sub groups [12] as shown in Fig. 3.
• Single-phase element model, (see Fig. 3-a). The mixture is considered to be the
concrete as a single-phase material, represented by the assembly of spherical
elements.
• Separate single-phase element model, (see Fig. 3-b). The mixture is divided into a
mortar phase and an aggregate phase, represented by a combination of separate
single-phase elements as mortar and aggregate.
• Two-phase element model, (see Fig. 3-c). The mixture is divided into a mortar
phase and an aggregate phase, represented by the assembly of the two-phase ele-
ments consisting of the inner core as aggregate covered with a layer of the mortar.
Concrete Coarse agg. Mortar Coarse agg. Mortar
Single-phase element Separate single-phase elements Two-phase element

(a) (b) (c)
Fig. 3. The discrete element models for a concrete mixture: (a) single-phase model; (b) separate
single-phase model; (c) two-phase model.
A two-phase element model is used in this study to represent not only coarse
aggregates, but a layer of cement paste or fine mortar covering them as well. Although
more complex models are also possible to simulate the real shape, a two-phase element
model is suitable due to the high accuracy and the medium computational cost required
to evaluate the contact between fiber elements and concrete elements. In fact, using
more advanced models could impose unnecessary calculation time, which would not
add significant value on the results.
4.2 Fiber Definition

Steel fibers may be defined as a clump of 131 separate spherical particles (see Fig. 4).
This number balances the accuracy and the computation cost. The fiber selected for the
simulation is a double hooked-end steel fiber, which was also used in the experimental
program. The properties of the fiber are presented in Table 2.
Fig. 4. Numerical simulation of 4D steel fiber consist of 131 spherical elements.
Table 2. Properties of 4D steel fiber.

Shape Double Hooked 4D
Length 60 mm
Diameter 0.75
Aspect ratio 80
Tensile strength 1,800 N=mm2
4.3 Rheological Model

The constitutive relations are developed in order to describe the interactions between
two neighboring particles and clumps in simulating fresh state SFRSCC. The normal
and tangential components of the contact force reflect the different types of interaction
amongst the particles. The normal component represents the state of the contact,
compression or tension mode, and gives quantitative information about the normal
force. This component of the force is active throughout the entire simulation, even if
the material does not move, here compression under gravitational force. The tangential
component of the force represents the effect of the friction and begins acting when the
particles start to move. Then both force components are acting simultaneously.
Studies from other authors [13] showed that the motion of the fresh state self-
compacting concrete in the normal direction has elastic and viscous components
working simultaneously. The Kelvin rheological model is generally used to represent
this condition. This model is composed by an ideal spring arranged in parallel to a
dashpot. The spring accounts for the elastic behavior of the mortar whereas the dashpot
accounts for the effect of its viscosity.
Although there is a scarce amount of studies in the field of simulation of fiber
elements in the DEM, the author suggested to use Bingham rheological model in the
normal direction. The Bingham model consists of a Saint Venant element and a dashpot
connected in parallel. The deformation of this model is not possible before reaching the
yield stress. When the yield stress is achieved, the model exhibits visco-plastic
deformation. A complete list of simulated interactions in normal and tangential
direction are presented in Table 3.
Table 3. Interaction between elements in normal and tangential direction.
Contact Type Ball-Ball Ball-Clump Clump-Clump Ball-Wall Clump-Wall

Normal Direction
Schematic
Drawing
Contact Model Kelvin Bingham Bingham Kelvin Bingham

Key Parameters
Tangential Direction
Schematic
Drawing
Bingham- Bingham- Bingham- Bingham- Bingham-

Contact Model
Hook Hook Hook Hook Hook
Key Parameters
Figure 5 shows schematically the force–displacement relation between two concrete

elements as introduced for the contact model in the normal direction. Based on this
model, the inter-element interaction can be generally divided into six different zones.
Force (F n )
Compression Mode
Zone A Zone B Zone C Zone D Zone E Zone F
2 Distance between
particles (Sn )
Tension Mode 1
Interaction through aggregates Interaction through paste No
Fig. 5. The interaction force (Fn) - distance (Sn) plot between two discrete elements in normal
direction. dp and Srup are the thickness of the paste layer and rupture distance, respectively.
In zone A, in which the inter-particle distance Sn < 0, the inner phases of elements
are in direct contact. In zone B, in which the inter-particle distance varies between zero
and the sum of the two paste layer thicknesses around each particle 0 < Sn < 2dp , the
interaction happens through contact between paste layers. In zone C, D and E with 2dp
Sn < Srup , the cohesion force due to the presence of the paste bridge decreases by
gradually increasing the inter-particle distance. In zone F, with Sn Srup , two particles
are completely separated and there is no interaction between them anymore. The other
n n
parameters, i.e. FB1 , FB2 , which depend on the size of particles and paste consistency,
can be obtained from the force-distance formula.
In tangential direction, the elastic-plastic behavior
t is considered
where the elastic
n
behavior is restricted by the Coulomb criterion ( Fmax j¼ ljF ). l (frictional coeffi-
cient) is considered equal to 0.7 and 0.5 for inside phases contact and outside phases
contact, respectively.
5 Results
This section presents results obtained by inductive test method. These experimental
results are then compared to results obtained from their respective numerical simula-
tions. The following section studies the average orientation of the steel fibers in each
direction by the means of the orientation number.
The casting process of the cubic specimen was simulated using ITASCA Particle
flow code (PFC) [14] and the interaction model described in the previous section. The
formwork of the cast specimen was modelled using discrete element method. The
numerically obtained orientation of the steel fibers was subsequently compared to the
experimental results presented in the following section.
Figure 6 shows the resulting distribution of the steel fibers with 60, 90 and 120
kg=m3 dosage of steel fiber, respectively. The orientation of the steel fibers is depicted
as the 3D view of the orientation numbers in x, y and z directions. The steel fibers tend
to orient in the regions close to the vertical walls of the formwork, which is primarily
caused by the wall effect [15]. It is geometrically impossible to have a rigid fiber
located normal to the formwork at a distance less than half-length of the fiber. Hence,
the rigid fiber tends to orient according to the surrounding flow with a restriction
imposed by the wall. In the bulk of the specimen, the steel fibers remain in a more or
less random orientation state in x and y direction due to the shear induced orientation
and the extensional stress induced orientation [16].
Fig. 6. Representation of the orientation of fibers inside the cubic mold. Fiber dosage: a)
60 kg/m3 b) 90 kg/m3 c) 120 kg/m3
Table 4 presents a comparison of the orientation numbers in each direction

obtained from the simulation and the inductive test experiment. There seems to be a
direct correlation between the results from the experimental study and the simulation in
each direction. The orientation of fibers in individual regions is represented by means
of second order orientation tensors [16]. The orientation tensors were visualized in the
form 3D ellipsoids and in this paper in the form of 2D ellipses (see Fig. 7). The
orientation of the ellipses represents the mean orientation of fibers in the region. Aspect
ratio of the ellipses represents the alignment of fibers in the region. High aspect ratio of
the ellipses corresponds to a high alignment of the fibers and the circular shape of the
ellipses corresponds to a uniformly random orientation of fibers.
The main outcome of the experiment was a set of 3D ellipsoids representing the
orientation of fibers in the cubic element. Figure 7 presents top views and side views of
the elements with orientation ellipses result from both numerical simulation and
inductive test analysis. The fibers tend to orient normal to the flow direction which
forms a circular pattern. The longer the distance from the inlet the more the orientation
of fibers increase. Due to the limited size of the element, the fibers seem to be more
randomly distributed in the x-y direction. A reasonable explanation might be a

boundary effect as described in [17]. An apparent Navier’s slip [18] occurred during the
experiment and simulation.
Table 4. Comparison of the simulated fiber orientation number with experimental results.
Reference code Exp Num Error Exp Num Error Exp Num Error
cos(hx) cos(hy) cos(hz)
C1-A 0.5098 0.534 4.70% 0.5082 0.5206 2.40% 0.4611 0.4587 0.50%
C1-B 0.5316 0.40% 0.5262 1.10% 0.4151 10.50%
C2-A 0.5222 0.5256 0.70% 0.5373 0.5646 5.10% 0.4126 0.4249 3.00%
C2-B 0.5234 0.40% 0.5334 5.80% 0.4161 2.10%
C3-A 0.5482 0.5778 5.40% 0.537 0.5476 2.00% 0.3798 0.3643 4.10%
C3-B 0.5385 7.30% 0.5437 0.70% 0.3838 5.10%
Fig. 7. Graphical comparison of the fibers in each specimen. Red dashed-lines are representing
numerical values and black solid-lines are representing experimental values.
6 Conclusion
The work presented in this paper involves the combined use of experiments and
numerical modelling techniques to investigate an important aspect of the steel fiber
reinforced self-compacting concrete. The good fit obtained between numerical and
experimental results confirm the model and the constitutive law reproduce the fresh
state behavior of the SFRSCC in the casting procedure. The results obtained represent a
step forward, showing that it is possible to apply advanced numerical tools for a
preliminary assessment of the performance of fiber reinforced concrete, which might
have positive implications through improved reliability of the design procedures. The
proposed use of discrete element method simulations to determine fiber coordinates
within cast self-compacting concrete is a novel means for connecting methods of
processing to the mechanical performance of fiber reinforced cement composites.
The ability to simulate the casting process of steel fiber reinforced self-compacting
concrete, including the movement of individual fibers, has profound implications
toward the effective use of these materials within the civil infrastructure. As an
important aspect, the results of such flow simulations can serve as input to mechanical
models of material behavior. Together with physical experimentation, this coupled
simulation of concrete casting presents opportunities for improving material perfor-
mance for both ordinary and high-performance applications.
Moreover, the framework can be used to approach the material in a more general
way and to investigate and understand the other underlying processes of the casting
process such as segregation of the fibers and aggregates in a suspension, blocking of
fibers behind reinforcement, and Influence of formwork surface on the orientation of
the fibers.
Acknowledgements. The first author thanks the AGAUR FI Catalan government for the
scholarship granted (2017 FI_B 00748). Likewise, this work was only possible thanks to the
support from the Laboratory of Technology of Structures and Materials “Lluis Agulló” of
Polytechnic University of Catalonia.
References
1. Di Prisco, M., Plizzari, G., Vabdewalle, L.: Fibre reinforced concrete: new design
perspectives. Mater. Struct. 42(9), 1247 (2009)
2. Walraven, J.: High performance fiber reinforced concrete: progress in knowledge and design
codes. Mater. Struct. 42(9), 60 (2009)
3. Sarmiento, E.V.: Influence of concrete flow on the mechanical properties of ordinary and
fibre reinforced concrete., Master thesis, Universitat Politècnica de Catalunya (2011)
4. Švec, O., Skocek, J., Stang, H., Olesen, J.F., Thrane, L.N.: Application of the fluid dynamics
model to the field of fibre reinforced self-compacting concrete. In: Proceedings of the
International Conference on Numerical Modeling Strategies for Sustainable Concrete
Structures (SSCS). Aix-en-Provence, France (2012)
5. Cavalaro, H.P., Carreno, R.L., Torrents, J.M., Aguado, A., Garcia, P.J.: Assessment of fibre
content and 3D profile in cylindrical. Mater. Struct. 49(1–2), 577–595 (2015)
6. Oldrich, Š., Skocek, J., Olesen, J.F., Stang, H.: Fibre reinforced self-compacting concrete
flow simulations in comparison with L-Box experiments using Carbopol. In: Proceedings of
8th Rilem International Symposium on Fibre Reinforced Concrete: Challenges and
Opportunities. RILEM Publications, pp. 1–8 (2012)
7. Torrents, J.M., Blanco, A., Pujadas, P., Aguado, A., Juan-Garcıá, P., Sańchez-Moragues,
M.A.: Inductive method for assessing the amount and orientation of steel fibers in concrete.
Mater. Struct. 10, 1577–1592 (2012)
8. Cavalaro, S.H., Carreno, R.L., Torrents, J.M., Aguado, A.: Assessment of fibre content and
3D profile in cylindrical SFRC specimens. Mater. Struct. 49(1–2), 577–595 (2015)
9. Blanco, A.: Characterization and modeling of SFRC elements, Doctoral thesis, Universidad
Politecnica de Catalunya, pp. 85–86 (2013)
10. Cavalaro, S.H.P., Lopez, R., Torrents, J.M., Aguado, A.: Improved assessment of fibre
content and orientation with inductive method in SFRC. Mater. Struct. 48(6), 1859–1873
(2014)
11. Roussel, N., Gram, A.: Simulation of Fresh Concrete Flow, In State-of-the-Art Report of the
RILEM Technical Committee 222-SCF. New York London, RILEM (2014)
12. Mechtcherine, V., Gram, A., Krenzer, K., Schwabe, J.H.: Simulation of fresh concrete flow
using discrete element method (DEM). Mater. Struct. 47(4), 615–630 (2014)
13. Spangenberg, J., Roussel, N., Hattel, J.H., Thorborg, J., Geiker, M.R., Stang, H., Skocek, J.:
Prediction of flow induced heterogeneities and their consequences in Self Compacting
Concrete (SCC). In: Proceedings of SCC. Springer, Montreal, Canada (2010)
14. Itasca Consulting Group Inc., PFC3D 5.0 Documentation, Itasca Consulting Group Inc.,
Minneapolis, MN, USA (2017)
15. Blanco, A.: Doctoral Thesis Characterization and Modelling of SFRC Elements. Universidad
Politecnica de Catalunia, Barcelona, Spain (2013)
16. Advani, S.G., Tucker, C.L.: The use of tensors to describe an predict fiber orientation in
short fiber composites. J. Rheol. 8, 751 (1987)
17. Jacobsen, S., Vikan, H., Haugan, L.: Flow of SCC Along Surfaces, Dordrecht. Springer,
Netherlands (2010)
18. Thrane, L.N.: Form Filling with Self-Compacting Concrete. PhD thesis, Technical
University of Denmark (2007)
Crack Width Simulation and Nonlinear Finite
Element Analysis of Bursting and Spalling
Stresses in Precast FRC Tunnel Segments
Under TBM Thrust Jack Forces
Mehdi Bakhshi(&) and Verya Nasri
AECOM, New York, USA

mehdi.bakhshi@aecom.com
Abstract. Tunnel boring machine (TBM) thrust jack force is one of the gov-
erning load cases for the design of precast Fiber-Reinforced Concrete
(FRC) tunnel segments. Due to the concentration of the jacking forces and
partial loading of the circumferential joints, the bursting stresses are formed
deep under the thrust pads while the spalling stresses are generated in the areas
between the thrust pads, and between the thrust pads and the end faces of the
segments. Under high thrust conditions and in conjunction with eccentricity of
the thrust pads, bursting and spalling tensile stresses can exceed the tensile
strength of FRC which results in segment cracking. Serviceability Limit State
(SLS) design includes limiting the crack width to allowable values. This paper
presents results of three-dimensional nonlinear finite element analyses (3D
NFEA) of TBM thrust jack forces by modeling the actual parallelogram shape of
segments, the actual size and positions of thrust pads, and maximum jacking
forces for an ongoing construction project in North America. Tensile response of
FRC as a multi-linear stress-strain curve is obtained following an inverse
analysis procedure conducted on nominal response of standard beam tests.
Presented NFEA simulation is recommended as a reliable and optimized pro-
cedure for design of FRC segments loaded by TBM thrust jack forces.
Keywords: Analysis Crack width Fiber Finite element Lining

Nonlinear Precast Segment Tunnel
1 Introduction
One-pass lining is the predominant segmental lining system for tunnels bored by Tunnel
boring machine (TBM). In this system, damage to precast concrete segments undermines
the serviceability and the waterproofing functionality of the tunnel lining (Fig. 1(a)).
Through studying previous tunneling projects it has been indicated that segment damages
occur by the construction loads or the increase in earth pressure due to either the con-
solidation settlement or future construction works in proximity of the tunnel after tun-
neling completion [1]. Most damages to segments is, however, caused during the
construction phase when the TBM occasionally produces large thrust forces that will lead
to cracking of concrete [2–4]. Information obtained from more than 50 tunnel
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_56
622 M. Bakhshi and V. Nasri
construction sites on segment damages during bored tunnel construction shows that more
than 65% of all segment damages occur during the excavation when segments are loaded
by TBM thrust forces [5]. In addition, about 60% of damages occurring more than once
within 20–30 rings and about 50% of damages occurring more than once within 2–3 rings
are the types of cracks that are caused by TBM thrust jack forces [5]. As shown in Fig. 1,
these include spalling and longitudinal cracks along the tunnel in the mid-section of the
segment (between locations of applied jacking forces), spalling of segment sides and
corners, and bursting cracks of the segment deep under jacking pads [6]. In addition, it was
found out that about 70% of longitudinal and spalling cracks appears at the curve
alignment [5]. This is due to non-uniform thrust jack forces applied on segments to help
TBM steer into the curve. When TBM deviates from its the theoretical position it can be
steered back by using the hydraulic jacks. This correction can also lead to non-uniform
jack forces along the lining and consequently, some segments may be exposed to higher
jack forces than other ones and damaged [3]. In addition to high thrust jack forces,
eccentricity of jack forces with respect to segment thickness centerline and asymmetric
positioning of jacking pads with respect to segment ends in the circumference of the ring
(thrust jacks are irregularly spaced or segments sizes are unequal) are considered as other
major causes of cracking during the thrust phase [6]. One can think of limiting magnitude
of jack forces, reducing the eccentricity of jack forces along the ring thickness and
Fig. 1. Cracks in segments during TBM thrust phase: a) spalling and longitudinal cracks, b)
spalling of segment sides and corners, c) bursting cracks deep under jacking pads [6].
Crack Width Simulation and Nonlinear Finite Element Analysis 623
uniform distribution of jack pads with respect to segment ends as mitigation methods.
Nonetheless, in practice almost none of these methods can be guaranteed. Therefore, such
extreme conditions should be considered during the design procedure of segments. In
reinforced concrete segments, TBM thrust phase, among other methods can be most
accurately analyzed by elastic finite element method (FEM). As a result, the tensile
demand are compared with the provided capacity by rebar for verification of design [7, 8].
Nowadays most of precast tunnel segments are made of fiber-reinforced concrete (FRC).
For FRC segments, nonlinear finite element analysis (NFEA) and fracture mechanics are
the only methods that can account for the non-linearity of FRC and are capable of
accurately predicting the crack width in extreme service load conditions.
This paper presents details and procedures of modeling the TBM thrust phase with
3D NFEA simulations. It includes results of analyses on precast FRC tunnel segments
that is currently being used for construction of a light-rail tunnel in North America. In
the simulations, the actual parallelogram shape of segments and the actual size and
position of thrust pads are modeled, and accurate magnitudes of nominal and maximum
exceptional machines thrusts are considered. Different combinations of thrust pad
eccentricities and thrust jack forces are considered in this study. A 3D NFEA program,
DIANA, is adopted because of the 3D nature of applied forces and boundary conditions
and the strong capability of this solver in implementation of a smeared crack band
model. In order to accurately simulate the crack width, the tensile and compressive
response of FRC material as multi-linear stress-strain curves are introduced to the
solver as an input. The tensile and compressive material models are obtained following
an inverse analysis procedure recommended by ACI 544.8R-16 [9] on standard beam
tests. For the first time in performing this type of analysis the actual nominal or
characteristic flexural response of FRC, and not the average response, is used for back-
calculation of tensile and compressive stress-strain curves. Nominal flexural response
of FRC segments is derived considering ‘five percent fractile’ concept as results of
statistical analyses on more than 200 standard beam tests [10] performed in the segment
precast plant. Presented NFEA simulation can be considered as a reliable and optimized
procedure for design of FRC segments loaded by TBM thrust jack forces.
2 Project Ouline and Segmental Ring System
A new 3.6 km long bored tunnel is being built below Montréal international airport
runways as part of the Réseau Express Métropolitain (REM) which soon will be the
fourth largest electric and fully automated light-rail transit network in the world. This
7 m diameter tunnel is being excavated using a hybrid TBM that progresses through
saturated soft ground and competent rock and will be supported by one-pass FRC
lining system [11]. Considering 6.478 m internal diameter, 300 mm was selected as the
thickness of segments resulting in an internal diameter to thickness ratio of 21.5, a

value well within the well-known recommended range of 20–23 [12]. Optimized length
of the ring or segment width was selected as 1700 mm. As shown in Fig. 2, a 6 + 1
rhomboidal segmental ring system was selected. It consists of a trapezoidal reverse key
segment slightly larger than other five full-size rhomboidal segments and one small
trapezoidal key segment in a ring, slightly smaller than 1/3rd of the length of ordinary
segments. Segment slenderness (segment curved length-to-thickness) is between 10.7
and 11.4 which is near the maximum value used in the world for FRC segments. The
universal ring system was selected with a ring taper of 25 mm on each side (total ring
taper of 50 mm) to negotiate alignment curves as tight as 240 m in radius. Longitudinal
and circumferential joints were designed as completely flat joints which are advanta-
geous to curved joints and coupling joints for transferring loads between the segment
joints. Bolt connection was designed for the longitudinal joints and dowels were
chosen for connecting rings in the circumferential joints.
3 Concrete Mixture and Mechanical Response of FRC
An optimum concrete mix design was selected for REM Airport tunnel segments to
achieve a high early-age strength as well as a high performance and a durable concrete
lining for 125 years of service life. Specified 28-day compressive strength and required
average strength for this mix are 49 MPa and 60 MPa, respectively. Maximum water to
cementitious materials (w/cm) ratio is 0.33, and silica fume and ground granulated blast-
furnace slag are added by 5 and 22% of cementitious materials, respectively. Minimum
cementitious material content is 485 kg/m3 and the maximum aggregate size is 14 mm.
Precast segments are reinforced with 42 kg/m3 of double hooked-end high-strength
(1800 MPa) steel fiber with a length of 60 mm and an aspect ratio of 80 (Dramix® 4D
80/60BG) as well as to 2 kg/m3 of monofilament propylene microfibers (Duomix® M6
Fire) with 6 mm in length and 18 lm in diameter to control explosive spalling in fire.
Concrete mix proportions for this mix is presented in Table 1. Results of chloride ion
penetrability test per ASTM C1202 [13] in terms of charges passed are in the range of
213 and 314 Coulombs which is near the lower limit for category of ‘Very Low’ and
upper limit for category of ‘Negligible’. As far as beam tests, the FRC mixture was
thoroughly studied in the segment precast plant prior to final production with more than
200 flexural tests performed following ASTM C1609-19 [10] standard procedure. Test
results have been statistically analyzed and nominal flexural response of FRC at 28 days
has been obtained (Fig. 3). Nominal strengths are strength levels below which with 90%
confidence there is 95% probability of the actual strength exceeding the nominal
Fig. 2. Segmental ring: a) elevation view; b) developed plan on intrados.

Table 1. Mix design for concrete used in precast segments of REM Airport Tunnel.
Mix Ingredients or Characteristics Quantity
Cement (kg/m3) 356
Silica Fume (kg/m3) 24
Slag or GGBFS (kg/m3) 107
Fine Aggregate (kg/m3) 765
Coarse Aggregate (kg/m3) 1026
Water (kg/m3) 160.6
Superplasticizer (liter/m3) 3.811
Air Entraining Admixture (liter/m3) 0.15
Steel Fiber (kg/m3) 42
Micro Polypropylene Fiber (kg/m3) 2.0
w/cm Ratio 0.33
Total Unit Mass (kg/m3) 2485
Fig. 3. Nominal response of FRC mix at 28 days tested according to ASTM C1609 [10].
strength. As shown in Fig. 3, this margin is 1.645 standard deviations (SD) below the
average strength level and is defined as ‘nominal strength’ or ‘characteristic strength’.
When designing for TBM thrust jack forces, complete tensile and compressive stress-
strain responses are needed for analysis. An inverse analysis approach recommended by
ACI 544.8R-16 [9] was adopted to back-calculate stress-strain relationship of FRC from
nominal experimental load-deflection response in bending. As shown in Fig. 4, a tri-
linear tensile model and a bi-linear compressive stress-strain model predicts the nominal
load-deflection response with a good accuracy. The nominal tensile cracking strength is
2.77 MPa, the nominal ultimate tensile strain capacity is 0.03 and nominal constant
residual tensile strength is determined as 1.6 MPa.
Fig. 4. Nominal response of specified FRC mixture: (a) experiment vs. simulation for standard
beam test, (b) back-calculated tensile stress strain curve, (c) back-calculated compressive
response.
4 TBM Thrust Jack Forces
After assembly of a complete ring, TBM advances by thrusting against the most recent
assembled ring. As part of this process, the TBM jacks bear against the jacking pads
placed along the exposed circumferential joint. High compression stresses develop
under the jacking pads and result in the formation of significant bursting tensile stresses
deep within the segment. Furthermore, spalling tensile stresses develop between
adjacent jack pads along the circumferential joint, and between jack pads and edge of
segments. For this project, nominal machine thrust is Fn = 50,192 kN (at 350 bar) and
Fig. 5. a)TBM thrust jack configuration along the segmental ring circumference, b) position of
thrust pad along segment circumferential joint, c) geometry of thrust pads (dimensions are in
mm).
maximum exceptional machine thrust is Fmax = 58,796 kN (at 410 bar). As shown in
Fig. 5, total number of thrust cylinders is 19 and each thrust pad’s internal and external
radii are 3314 mm and 3464 mm respectively, exerting pressure on segments over an
arch angle of 13.5o in the circumferential direction.
Exact size of each jack pad considering slight tapering of the ring is Apad =
119,819 mm2. In this project, nominal and maximum exceptional eccentricities of jack
pad in the radial direction are considered en = 30 mm and emax = 40 mm, respectively.
Figure 6 shows detail of circumferential joints with gasket recess on the exterior, and
stress relief recess on the interior side of tunnel. As shown in this figure, maximum
thickness of segment in contact with TBM thrust pads is 200.4 mm. Consequently, high
eccentricities of jack pads in radial direction toward extrados may result in reduced
contact area with segments and increased pressure on the contact zone. The design
procedure considers all possible combinations of nominal and maximum exceptional
thrust jack forces with nominal and maximum exceptional thrust pad eccentricities. This
results in a total number of 10 jack force scenarios. Following load and resistance factor
design (LRFD) method, each jack force scenario is defined by an appropriate load factor
of 1.0, 1.05 or 1.2 in order to consider the variability in the prediction of loads and
eccentricities. For example, the likelihood of exceeding maximum exceptional eccen-
tricity and maximum exceptional machine thrust at the same time is practically zero, and
therefore, a load factor of 1.0 is applied on these thrust force scenarios (load case scenarios
9 and 10 in Table 2). The machine operational nominal thrust is almost never exceeded
unless in exceptional cases when the TBM is stuck in the ground. Therefore, for other load
case scenarios in which exceptional machine thrust is considered, namely load case
scenarios 6, 7 and 8, the likelihood of occurrence for these jack force scenarios regardless
of thrust pad eccentricity (zero or nominal) is extremely low. Hence considering a load
factor of 1.05 is appropriate for these cases. For thrust force scenarios in which the
machine operational nominal thrust is considered, and the thrust pad eccentricity is less
than or equal to the nominal value of 30 mm, the likelihood of occurrence of a situation
when one of the two nominal values (either thrust or pad eccentricity) are exceeded are not
low. Hence, applying a load factor of 1.20 is recommended for thrust case scenarios 1
through 3. Finally, for thrust cases of combination of nominal machine thrust and max-
imum exceptional thrust pad eccentricities (40 mm), namely cases 4 and 5 in Table 2, the
likelihood of exceeding eccentricity is zero. But it is possible in rare situations for the
exceptional machine thrust to be mobilized and therefore a load factor of 1.05 is advisable.
Because eccentricities impact the contact area of thrust pads and segments, in all sce-
narios, contact areas and resulting pressure on contact zone were obtained and presented
in Table 2. The eccentricity toward the extrados (upward in Fig. 6) is shown with a
positive sign in this table and the eccentricity toward intrados (downward in Fig. 6) is
shown with a negative sign. One quick conclusion from this table is that the maximum
contact pressure applied on circumferential faces by TBM thrust jacks is 34 MPa which
provides a basis for unfactored specified compressive strength of the concrete mixture.
Fig. 6. Segment circumferential joint: a) details of joint profile, b) Thrust pads positioning.
Table 2. Different thrust force scenarios considered in the design.

Case Unfactored Jack Force on Eccentricity, Load Contact Factored Pressure Applied on
Each Pad, Fp (kN) e (mm) Factor, Area, contact Area, Pcontact (MPa)
c Acontact
(mm2)
1 Fn/19 = 2642 e=0 c = 1.20 0.119819 1.20(2642)/0.119819 = 26.46
2 Fn/19 = 2642 en = −30 c = 1.20 0.119819 1.20(2642)/0.119819 = 26.46
3 Fn/19 = 2642 en = +30 c = 1.20 0.099801 1.20(2642)/0.099801 = 31.76
4 Fn/19 = 2642 emax = −40 c = 1.05 0.119819 1.05(2642)/0.119819 = 23.15
5 Fn/19 = 2642 emax= +40 c = 1.05 0.091725 1.05(2642)/0.091725 = 30.24
6 Fmax/19 = 3094 e=0 c = 1.05 0.119819 1.05(3094)/0.119819 = 27.12
7 Fmax/19 = 3094 en = −30 c = 1.05 0.119819 1.05(3094)/0.119819 = 27.12
8 Fmax/19 = 3094 en = +30 c = 1.05 0.099801 1.05(3094)/0.099801 = 32.56
9 Fmax/19 = 3094 emax = −40 c = 1.00 0.119819 1.00(3094)/0.119819 = 25.83
10 Fmax/19 = 3094 emax = +40 c = 1.00 0.091725 1.00(3094)/0.091725 = 33.74
5 Analysis of Thrust Forces
A 3D NFEA program, DIANA [14], was used for numerical simulation and crack
analysis of the segments for TBM thrust forces. DIANA is capable of adopting FRC
materials model in the format of multi-linear stress-strain curves shown in Figs. 4(b)
and (c). These parameters are used in conjunction with the Total Strain Crack con-
stitutive model to introduce and assign a nonlinear concrete material model to the
segments. As for geometry, the exact shape of segments in rhomboidal or parallelo-
gram shape with longitudinal joint taper angle of 8.5o as shown in Fig. 2, and exact
positioning of thrust pads in actual asymmetric pattern with respect to segment cen-
terline and segment ends were imported to DIANA. Recesses along the segment
thickness due to the gasket and stress relief grooves (shown in Fig. 6) were accurately
simulated (Fig. 7(a)). Boundary condition for three non-loaded end faces of segments
are set by a compression-only constraint (blocking movement of joints) normal to the
joint faces. Segments are modeled by solid linear hexahedron (dominant) elements and
meshed with element grading size varying from 8 mm near the thrust pads to 200 mm
towards the opposite boundary parallel to the thrust pads (Fig. 7(b)). The fine meshed
model (1.8 million hexahedron dominant elements) with higher density near the
jacking forces helps ensure that the model covers the crack initiation and propagation in
various dimensions of the segment including the thickness. Uniform surface pressure
with the intensity shown in Table 2 was used to impose the effect of jacking forces
acting on the pads. In order to perform a nonlinear analysis, an incremental loading
method was introduced to capture the evolution of the plasticity in the model. The
accuracy of the results was ensured with simultaneous convergence criteria based on
the Quasi-Newton iterative method.
Fig. 7. a) Three-dimensional FEM model, b) Fine mesh that was used for simulations.
6 Results of Analysis and Discussion
Figures 8 through 13 show typical results of 3D NFEA simulations for jack force case
5 presented in Table 2 including the circumferential bursting and spalling stresses (S1),
the radial bursting stresses (S2), the longitudinal compressive stresses (S3), the tensile
strains that caused cracking (Eknn), extent of strains causing cracks (Eknn shown from
interior face), and most importantly the crack width opening dimensions (Ecw1), which
represents the opening of the crack normal to the direction of the crack. As shown in
Fig. 8, results of NFEA confirm that under high TBM thrust jack forces and high
eccentricities of thrust pads, the circumferential spalling or bursting stresses can reach
2.77 MPa which is the nominal tensile cracking strength of FRC segments. Therefore,
cracking of segments under such extraordinary conditions are inevitable (Figs. 11, 12,
13). However, due to high toughness (energy absorption capacity) and high residual
post-cracking strengths of FRC materials shown in Figs. 3 and 4, crack width
dimensions are not significant. Figure 9 shows that the maximum radial bursting
stresses (S2) which are predominantly present along the breadth of segments are
slightly below the nominal splitting strength or back-calculated tensile strength of the
FRC materials. This indicates a very well fine-tuned and optimized design of fiber
reinforcement for resisting radial bursting stresses. Figure 10 shows the maximum
longitudinal compressive stresses (S3) are limited to applied thrust jacking forces
which is an expected result considering the geometry and boundary conditions. Fig-
ures 11 and 12 show that the tensile cracks extend beyond just the circumferential
joints as the load application surface. However, Fig. 13 indicates that maximum crack
opening dimension for extended crack zone beyond the circumferential joint is only
0.035 mm. A summary of maximum circumferential bursting or spalling stresses,
maximum radial bursting stresses and maximum crack widths for all different thrust
force cases on Table 2 are shown in Table 3. Considering only the order of factored
pressure applied on contact area in Table 2 and disregarding effect of thrust pad
eccentricity, it was expected that resulting crack widths from biggest to smallest to be
in the order of case 10, case 8, case 3, case 5, case 6/7, case 1/2, case 9 and case 4.
Apart from case #10, results of NFEA are exactly in this same order as the values of
estimated maximum crack width decrease from 0.08 mm for case 8 to 0.077, 0.053,
0.045, 0.035, 0.042, 0.028, 0.024 and 0.02 mm in that aforementioned case order.
Another observation is that whenever the magnitude of factored pressure applied on
contact area are the same for two jack force scenarios with no eccentricity (e = 0) and
with negative eccentricity (eccentricity toward intrados, e.g. e = -30 mm), the resulting
maximum crack width from NFEA is bigger for cases with no eccentricity. For
example, when results of maximum crack width are compared between cases 6 and 7 or
between cases 1 and 2, results are significantly higher for case 1 and case 6 (e = 0). We
cannot draw a similar conclusion for the case of positive eccentricity (eccentricity
toward extrados) because there is no two cases that have the same factored pressure
applied on contact area but one with no eccentricity and one with positive eccentricity.
That being said, significant difference in results of maximum crack width between case
10 and case 8 with the same analogy can be attributed to 33% larger eccentricity of case
10 while the magnitude of factored applied pressure for case 10 is only 3% higher
compared to case 8. Based on results of this study, while effect of applied jack pressure
on the simulated crack width are clearly illustrated, more analyses are needed in the
future to better understand effect of thrust pads eccentricities. In general, the resulting
crack widths for all cases are in the range of 0.020–0.08 mm. These maximum sim-
ulated crack widths are below the allowable crack width limits set forth by all standards
and guidelines [15] in the range of 0.15–0.30 mm, and therefore, it validates the design
for serviceability limit state of cracking for the load case of TBM thrust forces. While
this paper presents details and procedures of modeling the TBM thrust phase with 3D
NFEA, future studies are needed for further validation of this procedure through the
comparison with the experimental full-scale test results.
Table 3. Results of NFEA simulation for the thrust jack force cases presented on Table 2.
Case 1 2 3 4 5 6 7 8 9 10
Max Circumferential 2.77 2.77 2.77 2.77 2.77 2.77 2.77 2.77 2.77 2.77
Bursting/Spalling
Stress (MPa)
Maximum Radial 2.68 2.59 1.98 2.66 1.92 2.66 2.58 1.84 2.67 2.17
Bursting Stress
(MPa)
Max Crack Width 0.042 0.028 0.077 0.020 0.053 0.045 0.035 0.080 0.024 0.059
(mm)
Fig. 8. Circumferential bursting/spalling stresses for case 5.

Fig. 9. Radial bursting stresses for case 5.
Fig. 10. Longitudinal compressive stresses for case 5.
Fig. 11. Tensile stresses that resulted in cracking under TBM thrust jack forces (case 5).
Fig. 12. Extent of cracks through the segment width for case 5.
Fig. 13. Simulated crack width dimensions for case 5.
7 Conclusions
Damages to segments mostly caused during the construction phase by TBM thrust
forces impair the serviceability and the waterproofing functionality of the one-pass
lining. Precast segments should be designed for maximum thrust forces and eccen-
tricities considering actual positioning of thrust pads in the construction to mitigate
such damages. The design procedure as presented in this paper should consider all
possible combinations of the nominal and maximum exceptional thrust jack forces and
thrust pads. An appropriate load factor should be applied to each combination of thrust
force and thrust pad eccentricity to take into account the variability in the prediction of
these parameters. In addition, nominal stress-strain response of FRC material should be
modeled to achieve reliable results, which can be obtained as a result of back-
calculation procedure on nominal standard beam tests data. Following these steps, 3D
finite element is recommended as one of the preferred methods of analysis.
Presented results of finite element analyses show that once the extreme thrust force
conditions are simulated, maximum circumferential bursting or spalling stresses at or in
the proximity of the load application surface (circumferential joint) reach the tensile
strength of FRC. This indicate that nonlinear finite element analysis (NFEA) is required
because of the capability of modeling the nonlinearity of FRC material in the post-crack
zone. Results of 3D NFEA confirm that high tensile stresses leading to cracking are
generated in the areas between the thrust pads, and between the thrust pads and the end
faces of the segments. While tensile cracks may extend beyond the circumferential joints,
crack opening dimension beyond the joint surface is negligible. Maximum simulated
crack width for the presented project with designed FRC mix is below the allowable crack
width limits which validates the design for serviceability limit state of cracking.
References
1. Koyama, Y.: Behaviors of segmental linings constructed by closed-type shield, pp. 113–127.
IS-Kyoto Short Course Text, Kyoto (2001)
2. Maidl, B., Herrenknecht, M., Maidl, U., Wehrmeyer, G.: Mechanised Shield Tunnelling, 2nd
edn. Ernst & Sohn, Berlin (2012)
3. Mayer, P.M., Klotz, C., Franzius, J.N., Hörrle, D.: A tool to improve risk analysis of TBM
tunneling processes’, in ‘Safe Tunnelling for the City and for the Environment’. In:
Proceedings of the ITA-AITES World Tunnel Congress (WTC), Budapest, May (2009)
4. Gruebl, F.: ‘Segmental rings (critical loads and damage prevention)’, In: ‘International
Symposium on Underground Excavation and Tunnelling’, Bangkok, February 2006, pp. 9–
19 (2006)
5. Sugimoto, M.: ‘Segmental rings (critical loads and damage prevention)’. In: Proceedings of
the International Symposium on Underground Excavation and Tunnelling, Bangkok,
February 2006, pp. 67–74 (2006)
6. SIG WG N. 2, ‘Report No. 1 - Damages of Segmental Lining’, Italian Tunnelling
Society/Società italiana gallerie (SIG) (2019)
7. Bakhshi, M., Nasri, V.: ‘Practical aspects of segmental tunnel lining design’. In:
‘Underground—The Way to the Future’, Proceedings of the ITA-AITES World Tunnel
Congress (WTC), Geneva, May-June 2013, pp. 563–570 (2013)
8. Bakhshi, M., Nasri, V.: ‘Design considerations for precast tunnel segments according to
international recommendations, guidelines and standards’, ‘Tunnelling in a Resource Driven
World’. In: Proceedings of the Tunnelling Association of Canada (TAC), Vancouver,
October 2014
9. ACI 544.8R-16, ‘Report on Indirect Method to Obtain Stress-Strain Response of Fiber-
Reinforced Concrete (FRC)’, American Concrete Institute (ACI), 2016
10. ASTM C1609, ‘Standard Test Method for Flexural Performance of Fiber-Reinforced
Concrete (Using Beam With Third-Point Loading)’, ASTM International (2019)
11. Nasri, V., De Nettancourt, X., Patret, P., Mitsch, T.: ‘Design and construction of Montreal
express link tunnels and underground stations’. In: Proceedings of the Rapid Excavation and
Tunneling Conference (RETC), Chicago, June, 2019 (Society for Mining, Metallurgy &
Exploration, Englewood, 2019), pp. 166–177 (2019)
12. Bakhshi, M., Nasri, V.: Guide for optimized design of tunnel segmental ring geometry. In:
Proceedings of the North American Tunneling Conference (NAT), Washington D.C., June,
2018 (Society for Mining, Metallurgy & Exploration, Englewood, 2018), pp. 541–549 (2018)
13. ASTM C1209, ‘Standard Test Method for Electrical Indication of Concrete’s Ability to
Resist Chloride Ion Penetration’, ASTM International (2019)
14. Diana FEA, ‘User’s Manual – Release 10.3’ (2019). https://dianafea.com/
15. Bakhshi, M., Nasri, V.: ‘Design of segmental tunnel linings for serviceability limit state’, in
‘See Tunnel - Promoting tunnelling in South East European Region’. In: Proceedings of the
ITA-AITES World Tunnel Congress (WTC), Dubrovnik, May 2015
Finite Element Modelling of UHPFRC
Flexural-Reinforced Elements
Eduardo J. Mezquida-Alcaraz(&), Juan Navarro-Gregori,

and Pedro Serna
Instituto de Ciencia y Tecnología del Hormigón (ICITECH),

Universitat Politècnica de València, Camino de Vera S/N, 46022 València, Spain
edmezal@alumni.upv.es
Abstract. In our research group a simplified method to obtain the constitutive

tensile parameters of Ultra High Performance Fibre Reinforced Concrete
(UHPFRC) from a four point bending test has been developed: 4P-IA. After the
numerical validation of the method, a finite element modelling (FEM) of beams
of different size has been carried out taking into account two different approa-
ches: a smeared approach and a discrete approach. Moreover, important effects
have been included in order to improve the model such as the shrinkage that
generates internal stresses due to the influence of the reinforcement. After the
FEM simulation, very reliable results can be observed for both approaches at
service and ultimate load levels.
Keywords: Ultra High Performance Fibre Reinforced Concrete Beams

Finite element modelling Four point bending test Experimental program
Tensile parameters
1 Introduction and Objectives
Ultra High Performance Fibre Reinforced Concrete (UHPFRC) is a type of concrete

that exhibits high compressive strength (more than 120 MPa) and high tensile strength
(more than 7 MPa). Generally it is characterized by a strain-hardening constitutive
behaviour followed by a micro-cracked phase before the macro-crack apparition, and a
relatively high energy absorption capacity [1]. The strain-hardening behaviour will
depend, in this case, on the quantity and the type of fibres, the strength of concrete
matrix, bond between concrete matrix and the fibres, the specimen size and its
geometry, the pouring system, etc.
Tensile characterization of UHPFRC is still a challenge for the researchers. Due to
its simplicity, the four point bending test (FPBT) is considered one of the most
common test to characterize this type of concrete. However, in order to obtain the
tensile constitutive parameters considering this test, it is necessary to deal with inverse
analysis methodologies. Currently, there are different inverse analysis methods that can
be used to obtain these tensile parameters [2–10].
The development of numerical models arises of great importance in order to
determine the validity and precision of the inverse analysis methods. In our research
group a simplified inverse analysis method (4P-IA) has been developed for the
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_57
640 E. J. Mezquida-Alcaraz et al.
determination of UHPFRC tensile properties that exhibits strain-hardening constitutive

behaviour from the load-deflection curve obtained from a FPBT [9]. From this, a finite
element model (FEM) has been carried out for the numerical validation of the sim-
plified method for concretes with strain-hardening [11, 12] and, in addition, its possible
adaptation and application to UHPFRC that exhibits strain-softening is also solved
[13].
After the development of the simplified method and the numerical model for the
characterization of UHPFRC, the objective of this research is their progressive appli-
cation at the structural level in order to cover the field of design and analysis of
UHPFRC structures. Therefore, this document constitutes a first extrapolation of the
numerical FEM developed in [11–13] to structural elements such as reinforced beams.
In its first part, an experimental program and the FEM of short reinforced beams
considered as D regions have been developed. Next, a new experimental campaign and
the consequent modelling of reinforced beams of greater entity that could already be
considered as B regions are presented.
In order to go towards the structure reality and to study the response of the model in
case of reinforced UHPFRC elements two experimental programs with reinforced
beams at different scale have been developed. The first experimental program has been
set with reinforced short beams close to a D region. For the second program two larger
beams have been casted.
2.1 Short Beams

To test the FEM developed and the applicability of the inverse analysis procedure in
reinforced specimens a campaign of short beams close to a D-region has been done (see
Fig. 1). This will constitute a first step towards the application of the model at a
structural level.
Fig. 1. Short Beams campaign

Finite Element Modelling of UHPFRC Flexural-Reinforced Elements 641
These specimens were casted using UHPFRC with 160 kg/m3 of steel short fibres.
The dimensions of the beams are 750 150 100 mm. Figure 2 shows the geometry
and the disposition of the reinforcement of the short beams developed. This is, six
/6 mm stirrups, three in each side symmetrically positioned to avoid shear failure, a
couple of /8 mm longitudinal reinforcement bars situated at the top of the section, two
in each side symmetrically positioned, and two /8 mm longitudinal reinforcement bars
situated at the bottom of the section.
Fig. 2. Reinforced short beam. Geometry.
A set of 8 UHPFRC reinforced short beams have been tested using the FPBT. The
position of the load and support roller can be seen in Fig. 2. A LVDT sensor has been
placed to obtain the load-deflection at mid span curve. Figure 3 shows the FPBT set up
for these short beams and the LVDT sensor measurements (stress-deflection at mid
span curve).
Fig. 3. FPBT set up of reinforced short beams (left) and stress-deflection at mid-span curve
(right)
2.2 Long Beams

To test the FEM developed and the applicability of the inverse analysis in wider scale
specimens, an experimental program of two large reinforced beams have been carried
out (Fig. 4).
Fig. 4. UHPFRC beams of 4.50 m
The two beams has been casted using UHPFRC with 160 kg/m3 of steel short
fibres. The dimensions of these two beams are 4.50 0.30 0.10 m. As it can be
seen in Fig. 5 the reinforcement consists in two longitudinal bars B500S of /20 for one
beam and /16 for the other situated at the bottom face, with a covering of 20 mm in
both cases.
Fig. 5. Dimensions and position of the reinforcement of the beams (in meters)
The beams have been tested in a four point bending test with 4 m span between
supports. The loading plates are situated at 1.5 m from the supports keeping a sepa-
ration between them of 1 m, as it can be seen in Fig. 5. Two LVDTs have been situated
at the supports level to register the possibility of a vertical displacement there. To
measure the deflection at mid-span two LVDTs have been situated, one in the front face
and the other in the back face. Moreover, two horizontal sensors have been situated at
mid span at two levels: 2.5 cm from the top and 3 cm from the bottom (at reinforce-
ment level). The same has been done at the back face to measure the compressive and
tensile strains.
Figure 6 shows the curves obtained as a result of the experimental measurements
from vertical LVDTs at mid span section in both faces (H2 and H3), the average and
the relative deflection (y) corrected by the measurements of the supports’ sensors, for
the /20 mm beam.
Fig. 6. Load-deflection curve at mid span section for the /20 mm beam
3 Numerical Model
A finite element model (FEM) has been developed to modelling the reinforced
UHPFRC specimens using the Finite Element software DIANA FEA [14]. This model
starts from a FEM defined by the authors in [11–13]. In this case the objective is to
adapt the capabilities of this FEM even with reinforced concrete beams.
To model tensile UHPFRC constitutive behaviour, two different approaches are
used: a Smeared Cracking Approach and a Discrete Cracking Approach. In the
smeared approach, the constitutive model for UHPFRC is based on a fixed total strain
crack model expressed according to the crack opening curve. For the discrete
approach, the constitutive model for UHPFRC is founded on the discrete cracking
model as interface behaviour. This behaviour is forced only on the central specimen
section. The rest of the specimen is modelled using the Smeared Cracking Approach.
The reinforcement has been modelled using Von Mises strain-hardening elasto-
plastic behaviour for the steel with a bond-slip behaviour between the reinforcement
and the UHPFRC matrix.
To characterize the material properties both in compression and in tension,
experimental tests have been done (Fig. 7). For compression, compression tests from
the 100 mm cubical specimens have been done and, for tension, the 4P-IA has been
applied to obtain the tensile properties from a set of FPBT from 100 100 500 mm
specimens. These results are implemented in the constitutive behaviour of the FEM
described before. So, the effect of the fibres is considered in the constitutive tensile
behaviour. Another effect that it is important to be considered is the shrinkage and its
consequences in concrete and reinforcement, just because of the presence of the
reinforcement. In this case, it has been taken into account as a material function when
the UHPFRC model is defined in the program using the total strain crack model. As a
basis, the shrinkage function from the EN 1992-1-1 Eurocode 2 [15] has been used.
From it, the values obtained are incremented in a percentage to be adapted to the
UHPFRC response. It is, the total shrinkage strain is expressed as expression (1), so the
value obtained at the day when the test is done is incremented between 120 (multi-
plying by 2.20) and 140% (multiplying by 2.40) to take into account the influence of
shrinkage in UHPFRC, obtaining shrinkage values between 0.6 and 0.9 mm/m, as it is
reflected in [16].
Fig. 7. Compression test (left) and FPBT (right)
ecs ¼ ecd þ eca ð1Þ
Both, the reinforced short beams and the long beams have been modelled with a 2D
quadratic plane stress element when the Smeared and Discrete Cracking Approaches
are used. Besides, when the Discrete Cracking Approach is used, a quadratic 2D line
interface element is placed on the central beam section. Using this interface element, it
is possible to model the discrete behaviour of the crack at the mid-span cross-section.
Figure 8 and Fig. 9 show the mesh used for the finite element model. The reinforce-
ment has been modelled in a discrete manner using a truss bond-slip element.
The load has been applied to the steel load plates by gradual increasing displace-
ment. Non-linear analysis has been carried out by an incremental-iterative solution
procedure.
Fig. 8. Modelling of the geometry of the FPBT for the reinforced short beams
Fig. 9. Modelling of the long beams
After the application of the model in the different specimens from the two experimental
programs defined above, the results obtained are described and analysed here.
4.1 Short Beams

Figure 10 shows the experimental stress-deflection at mid-span curve for two short
beams from the same batch. If the numerical stress-deflection and mid-span curves
obtained as a result from the application of the FEM using both approaches: the
smeared approach and the discrete approach are represented, it can be seen how the
numerical model adjusts very well the experimental response. The elastic and the multi-
crack stages are represented accurately. Besides, unlike the behaviour obtained with the
unreinforced specimens in the FPBT in [12], it can be observed that the differences
between the two hypotheses of the FEM are not significant. It means that the presence
of the reinforcement causes a notorious influence in the mechanical behaviour of the
specimens and leads to a close response between the smeared approach and the discrete
approach of the FEM.
Fig. 10. Stress-deflection at mid-span curve for FEM with the two constitutive hypotheses.
In this case, the model has been applied without taking into account the shrinkage
effect. As it can be seen, the model in both cases adjusts very well to the experimental
curve without this effect. It can be explained because the amount of reinforcement is
not relevant compared to the concrete section and it does not produce a big opposition
to the natural shrinkage deformation.
In Fig. 11 the response at failure stage of the analysis in terms of stresses in
concrete and in the reinforcement bars is shown for the discrete approach of the model.
As it can be appreciated, it is possible to distinguish the apparition of the forced crack
using the discrete approach in the mid-span section, and there, as a tie, the influence of
the reinforcement bar. As concrete reaches its ultimate tensile strength and, conse-
quently, the discrete crack is opening, the tension stress is being transmitted to the
reinforcement bars situated at the bottom. In this figure, it can be possible to see how it
is growing up, even reaching its yielding point. Moreover, as concrete reaches its
ultimate tensile stress along the depth of the section, the crack is progressing and,
again, it is possible to see the deformation produced in the position of the longitudinal
reinforcement.
Fig. 11. Concrete and reinforcement stresses and interface tractions in x direction
4.2 Long Beams

Once the model has been run for the long beams, several results can be analysed.
Figure 12 shows the equivalent bending stress (r)-displacement at mid (d) curves for
the experimental test compared to the response for different characteristics of the
model: when the model is applied taking into account the shrinkage calculated as EN
1992-1-1 Eurocode 2 [15] (“model 0%sh”) and the model incrementing a 120% the
value of the shrinkage (“model 120%sh”) for the two long beams tested. Moreover, for
the JE-1 (/20 mm) the model response without taking into account the shrinkage has
been represented as a reference (Fig. 12 (left)). As it can be seen in Fig. 12, the model
adjust very well the experimental curve when an increment of 120% of the shrinkage at
49 days is considered.
Fig. 12. r-d curves for JE-1 (left) and JE-2 (right)
Both beams were tested in a FPBT at 49 days after being casted. At that moment
the total shrinkage strain obtained applying EN 1992-1-1 Eurocode 2 [15] is ecs. If this
value is incremented at 120% and taken in consideration in the model, the result is the
r − d response depicted in Fig. 12. Table 1 summarizes these values for the two beams
tested. The values of shrinkage 120ecs are consistent with the values of UHPC
shrinkage considered by the other researchers in [16].
Table 1. Shrinkage values at 49 days

Beam Testing day ecs (mm/m) 120%ecs (mm/m)
JE-1 (/20 mm) 49 0.367 0.807
JE-2 (/16 mm) 49 0.374 0.823
Figure 13 shows the stresses in the beam and in the longitudinal reinforcement for
the JE-1 (/20 mm) beam at day 49, before starting the experimental test. Here it can be
observed the levels of stress due only to the shrinkage effect and its important effect.
Fig. 13. Shrinkage stresses at UHPFRC beam (up) and longitudinal reinf. (down) at 49 days.
For more detail, Fig. 14 shows the section stress profile of the mid span section of
the two beams due to the incremented shrinkage at 120%, it is, the UHPFRC shrinkage
at 49 days from their casting. As it can be seen, the stress generated by the shrinkage is
very important due to the influence of the reinforcement bars embedded in UHPFRC
matrix. Seeing this, it can be deduced that the UHPFRC in the vicinity of the rein-
forcement is highly tensioned even close to the ft value before the test starts. That
means that, concrete there, is pre-stressed near the collapse and, consequently, its
capacity is reduced due to this effect. The effect in the reinforcement is the opposite.
The longitudinal bars are compressed and, as a result, they have a pre-compression
before the test that derives in a more bearing capacity for the bending test. Moreover, if
the two graphs are compared, it can be seen that the JE-1 beam reaches more stress both
in compression (h = 300 mm) and tension (h = 0 mm) than the JE-2 beam. Further-
more, the reinforcement is less compressed in JE-1 than in JE-2. These results are
completely logical if the different amount of reinforcement between JE-1 (2 /20 mm)
and JE-2 (2 /16 mm) is considered. Therefore, after the analysis of these results, it can
be concluded that the shrinkage effect is very important in this kind of UHPFRC and,
consequently, its influence in the structural response of the beam is not negligible at all.
Fig. 14. Shrinkage stresses at mid span section for JE-1 (left) and JE-2 (right) at 49 days
5 Conclusions
After the application of the FEM in reinforced beams, the following conclusions can be
extracted:
• The model has been developed using two different approached: smeared approach
and discrete approach. It has been proved by the authors in previous work that
these approaches cause different response in specimens without reinforcement when
they are tested in a FPBT. However, because of the presence of reinforcement, in
reinforced specimens, both approaches have a very close response and the results
are very similar and accurate if they are compared to the experimental response.
• UHPFRC is a type of concrete that because of its particular dosage suffers a high
level of shrinkage. This can generate internal stresses before the test that, depending
on the amount of reinforcement, are necessary to take into account. When the
shrinkage is considered in the FEM, in case of the short beams, the internal stresses
due to the shrinkage has not effect because the amount of reinforcement compared
to concrete section is not very significant. However, in case of large beams with
higher level of reinforcement, the effect of shrinkage can generate internal stresses
in the vicinity of the reinforcement and, consequently, the bearing capacity of the
beam can be very influenced. According to the FEM simulation, the internal tensile
stresses in the UHPFRC matrix can be close to the elastic limit (ft).
• Therefore, after the analysis of these results, it can be concluded that the shrinkage
effect is very important in this kind of UHPFRC and, consequently, its influence in
the structural response of the beam is not negligible at all.
Acknowledgements. This work forms part of project “BIA2016-78460-C3-1-R” supported by

the State Research Agency of Spain.
References
1. Yokota, H., Rokugo, K., Sakata, N.: (JSCE-2008) Recommendations for Design and
Construction of High Performance Fiber Reinforced Cement Composites with Multiple Fine
Cracks (HPFRCC) (2008). http://dx.doi.org/10.1016/j.dci.2010.01.003
2. Rigaud, S., Chanvillard, G., Chen, J.: Characterization of bending and tensile behavior of
ultra-high performance concrete containing glass fibers. In: Parra-Montesinos, G.J.,
Reinhardt, H.W., Naaman, A.E. (eds.) High Performance Fiber Reinforced Cement
Composites 6 HPFRCC 6, pp. 373–380. Springer Netherlands (2012). https://doi.org/10.
1007/978-94-007-2436-5_45
3. Baby, F., Graybeal, B., Marchand, P., Toutlemonde, F.: Proposed flexural test method and
associated inverse analysis for ultra-high-performance fiber-reinforced concrete. ACI Mater.
J. 109, 545–555 (2012). https://doi.org/10.14359/51684086
characterization: inverse analysis of four-point bending test results. Mater. Struct. 46, 1337–
1354 (2012). https://doi.org/10.1617/s11527-012-9977-0
5. Qian, S., Li, V.C.: Simplified Inverse Method for Determining the Tensile Strain Capacity of
Strain Hardening Cementitious Composites. J. Adv. Concr. Technol. 5, 235–246 (2007).
https://doi.org/10.3151/jact.5.235
6. Ostergaard, L., Walter, R., Olesen, J.F.: Method for determination of tensile properties of
engineered cementitious composites (ECC). In: Construction Material Proceedings ConMat
2005 Mindess Symposium, p. 74 (2005)
7. Soranakom, C., Mobasher, B.: Closed-form moment-curvature expressions for homogenized
fiber-reinforced concrete. ACI Mater. J. 104, 351–359 (2007). https://doi.org/10.14359/
18824
8. Gröger, K., Johannes, V., Nguyen, W.: Bending behaviour and variation of flexural
parameters of UHPFRC. In: 3rd International Symposium on UHPC Nanotechnology High
Performance Construction Material, Kassel University Press GmbH, pp. 419–426 (2012)
9. López, J.Á., Serna, P., Navarro-Gregori, J., Coll, H.: A simplified five-point inverse analysis
method to determine the tensile properties of UHPFRC from unnotched four-point bending
tests. Compos. Part B Eng. 91, 189–204 (2016). https://doi.org/10.1016/j.compositesb.2016.
01.026
11. Mezquida-Alcaraz, E.J., Navarro-Gregori, J., López, J.A., Serna-Ros, P.: Numerical
Validation of a Simplified Inverse Analysis Method to Characterise the Tensile Behaviour
of UHPFRC (2018)
12. Mezquida-Alcaraz, E.J., Navarro-Gregori, J., Angel Lopez, J., Serna-Ros, P.: Validation of a
non-linear hinge model for tensile behavior of UHPFRC using a finite element model.
Comput. Concr. 23, 11–23 (2019)
13. Mezquida-Alcaraz, E.J., Navarro-Gregori, J., Serna-Ros, P.: Numerical validation of a
simplified inverse analysis method to characterize the tensile properties in strain-softening
UHPFRC. In: IOP Conference Series Material Science Engineering, p. 12006. IOP
Publishing (2019)
14. DIANA (Software). User’s Manual – Release 10.2, TNO DIANA, Netherlands (2017).
https://dianafea.com/manuals/d102/Diana.html
15. The European Union Per Regulation 305/2011, D. 2004/18/EC Directive 98/34/EC, EN
1992-1-1: Eurocode 2: Design of concrete structures, Eur. Stand. (2004)
16. Fehling, E., Schmidt, M., Walraven, J., Leutbecher, T., Fröhlich, S.: Ultra-High Performance
Concrete UHPC: Fundamentals, Design, Examples. Wiley, New York (2014)
Punching Shear Response of RC Slab-Column
Connections Strengthened with UHPFRC -
Finite Element Investigation
Demewoz W. Menna(&) and Aikaterini S. Genikomsou
Department of Civil Engineering, Queen’s University, Kingston, Canada

{18dwm,ag176}@queensu.ca
Abstract. Reinforced concrete (RC) flat slabs can be retrofitted with a top thin
layer of ultra-high performance fiber reinforced concrete (UHPFRC) to enhance
their structural performance and service life. UHPFRC can also protect the
concrete slab from impact and percolation of chemicals that could otherwise
cause deterioration and corrosion of reinforcement. This study examines the
punching shear performance of RC flat slabs retrofitted by fully and partially
covering with a thin layer of UHPFRC. 3D non-linear Finite Element Analysis
(FEA) is used to determine the response of the composite slabs. Both the
concrete and the UHPFRC are modelled using the concrete damage plasticity
model in ABAQUS software. The maximum shear resistance, deformation and
crack patterns of the slabs with different areas of the UHPFRC layer are
examined. The UHPFRC layer significantly increases the punching shear
capacity and delays the crack propagation in all analyzed slabs. Incorporating a
40 mm UHPFRC layer on a 120 mm thick RC slab-column connection results
in a 39% increase in ultimate punching load capacity. The partial cover of
UHPFRC is a good suggestion since it results in a more economical solution and
to a more ductile connection.
Keywords: Punching shear Reinforced concrete slab-column connection

Ultra-high performance fiber reinforced concrete (UHPFRC) Finite element
analysis Strengthening
1 Introduction
Reinforced concrete flat slab systems are preferred for their easy and aesthetically
appealing construction. However, flat slabs may be susceptible to brittle punching shear
failure [1]. A significant number of slab-column connections require punching shear
retrofit due to change in building use, deterioration of concrete, errors in the initial
design, or to comply with more stringent design provisions [2]. Various techniques
have been used to strengthen slabs for punching shear. These retrofit techniques include
post-installed steel shear bolts [3], carbon fiber reinforced polymer laminates [4], and
installed FRP fans through holes around the critical punching zone [5]. However, the
cost and complexity of the above strengthening methods result in a demand for a more
efficient technique. Strengthening reinforced concrete slabs using a thin layer of ultra-
high performance fiber reinforced concrete (UHPFRC) recently became an efficient
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_58
652 D. W. Menna and A. S. Genikomsou
technique for its simplicity and multiple advantages. This method increases the
punching shear capacity, flexural performance, and impact resistance [6].
UHPFRC is a relatively new generation of concrete containing very low
water/binder ratio, high powders content, and optimum fiber volumetric ratio with
ultra-high-strength and toughness. The compressive and post crack tensile strengths of
UHPFRC are greater than 150 MPa and 8 MPa respectively [7]. The optimum per-
centage of fibers prevent the percolation of corroding chemicals and limit propagation
of shrinkage and plastic cracks. Hence, UHPFRC has a wide range of applications in
rehabilitating existing structures.
Recent studies evaluated the performance of two-way and one-way RC slabs ret-
rofitted with UHPFRC top layer. Bastien-Masse and Brühwiler [6] experimentally
investigated punching shear resistance of reinforced concrete slabs (without shear
reinforcement) strengthened with a post-installed UHPFRC layer. The overlay acts as
tensile reinforcement and improved the rigidity, and punching shear capacity of slabs
by out of plane bending. 50 mm thick UHPFRC layer increased the punching shear
resistance by more than 69%. The performance of one way reinforced concrete slabs
strengthened with a thin UHPFRC layer over the tensile zone is experimentally studied
by Yin, Teo [8]. The results show that the UHPFRC overlay improved the rigidity and
delayed shear crack formation. As the thickness of UHPFR increases, the type of failure
shifts from shear crack and deboning (between regular concrete and overlay) to more
ductile flexural cracks. However, one of the major challenges in using UHPFRC for
retrofitting slab column connections is the absence of reliable design criteria and
methodology. In addition, since UHPFRC is a relatively expensive material a more
economical and optimum use is intrinsic. Thus, more numerical study in conjunction
with experimental research is required to fully exploit this innovative method.
The goal of this study is to investigate the punching shear performance of RC flat
slabs strengthened with full and partial cover of UHPFRC using 3D finite element
analysis (FEA). The FEA models are verified using experimental results from literature.
Test results of a conventional reinforced concrete slab (SB1) from [3] and a RC slab
retrofitted with UHPFRC (PBM1) from [6] are analysed to verify the finite element
model. A new optimum application of UHPFRC layer to strengthen flat plates under
punching shear is proposed and evaluated using the verified models. Effects of
increasing the volume of retrofitting layer on maximum punching capacity, deforma-
tion and crack propagation are presented and discussed.
2 Analysed Specimens
Ten reinforced concrete flat slabs are modelled and analysed to evaluate their post-
retrofit punching shear capacity. The specimens include two groups, small size
(SB) and larger size (PBM) slabs. SB series consists of square 1800 mm by 1800 mm
wide and 120 mm thick regular reinforced concrete slab. The SB slabs have a central
150 mm by 150 mm column extending 150 mm from the top and bottom surfaces of
the slabs. The control specimen (SB1) was tested by Adetifa and Polak [3], with a
reinforcement ratio equal to 1.1%. The load is applied at the central column and
supported at the edges. The PBM group has 3000 mm by 3000 mm rectangular slabs
Punching Shear Response of RC Slab-Column Connections 653
with 210 mm thickness. The control specimen (PBM1) is strengthened by a full cover
of 50 mm thick UHPFRC, tested by Bastien-Masse and Brühwiler [6]. The slab is
loaded using eight load cells arranged at 1500 mm radius from centre and support is
provided with 260 mm by 260 mm plate located at the centre. The regular concrete for
PBM series has ø16 @ 150 mm tension and ø10 mm @ 150 mm compression
reinforcements.
Figure 1 shows the geometry and loading of the analysed specimens, where, Lc and
hc denote the width and depth of the reinforced concrete slab respectively. Lu and hu
denote the width and depth of retrofitting UHPFRC layer, respectively.
Fig. 1. Orientation and test setup for (a) PBM group and (b) SB group.
Dimensions of the analysed specimens are given in Table 1. SBF and PBM1 are
fully covered with the overlay while the other specimens are partly covered. The width
of the UHPFRC layer varies between the specimens. During a full floor strengthening
the slabs with the proposed orientation for partial use of UHPFRC will ensure conti-
nuity of the topping and create redundancy for load redistribution.
Table 1. Dimensions of analysed specimens.

Specimen Normal concrete (mm) UHPFRC (mm)
Lc hc Lu hu
SB1a 1800 120 – –
SBF 1800 40
SBP1 600 40
SBP2 900 40
SBP3 1100 40
PBM1b 3000 210 3000 50
PBMF0 – –
PBMP1 500 50
PBMP2 1000 50
PBMP3 1500 50
a b
Tested by Adetifa and Polak [3], and Tested by Bastien-Masse and Brühwiler [6].
The material properties of the analysed specimens are presented in Table 2, where;
fc and ft are the compressive and tensile strengths respectively. The tension and
compression reinforcements used in regular concrete for SB series have similar yield
strength of 455 MPa. The yield strength of reinforcements used in PBM is 546 MPa
and 518 MPa for tension and compression bars respectively.
Table 2. Material properties of specimens.

Specimen Normal concrete strength UHPFRC strength (MPa)
(MPa)
fc (compressive) ft (tensile) fc (compressive) ft (tensile)
SB series 44 2.2 138 7.5
PBM series 36.7 2.8 130 9.7
3 Finite Element Modeling and Analysis
ABAQUS [9], a non-linear finite element modelling package is used for analysing all the
specimens. Regular concrete and UHPFRC are modelled using 8-noded hexahedral with
reduced integration (C3D8R) elements. 2-noded linear truss (T3D2) elements are used
for modelling reinforcement bars. 20 mm maximum mesh size is used for concrete and
UHPFRC. Also, more than two elements are used over the thickness of the UHPFRC
layers. Loading and boundary conditions of specimens simulate the actual experiments.
Displacement controlled loading is applied at the bottom of the central column for SB
series and restrained from vertical displacement at the edges. For PBM series (shown in
Fig. 2) the displacement is applied on radially distributed eight plates and vertically
supported at the centre. Symmetry is considered and a quarter of the specimens is
modelled because the geometry, loading and boundary conditions are symmetrical. For
all specimens, quasi-static analysis is conducted using ABAQUS/Explicit.
Fig. 2. Quarter symmetrical model for PBMP3.
Concrete damage plasticity (CDP) model is used to simulate both normal concrete
and UHPFRC. Previous studies proved that the CDP model can predict the punching
shear response of reinforced concrete slab-column connections [10] and the behaviour
of UHPFRC [11]. The present study adopted the damage plasticity parameters and
modelling methods from the aforementioned studies [10, 11].
The concrete damage plasticity model accounts for tension and compression
stiffness degradation. The concrete material parameters required in CDP include;
modulus of elasticity, Poisson’s ratio, and compressive and tensile strengths. The
uniaxial tensile and compressive stress-strain relationships of regular concrete and
UHPFRC used in SB series are shown in Fig. 3. For regular concrete, the compressive
stress-strain relationship is based on the Hognestad type parabola [10]. The compres-
sive stress rc is calculated from compressive strength fc0 and the strain corresponding to
the compressive strength eo (where eo = 2fc0 /Esec) as shown in Eq. (1). The secant
p
modulus of elasticity ðEsecÞ is considered equal to 5000 f 0 c . In Eq. (1), ec represents
the strain corresponding to the calculated compressive stress rc. The uniaxial stress-
strain relationship for regular concrete in tension is computed using a bilinear stiffening
response. Before cracking the stress-strain is assumed as linear elastic until it reaches
the tensile strength ft0 . After ft0 , the cracked concrete is characterized by softening,
(Fig. 3b).

ec ec 2
rc ¼ 2fc0 ð1Þ
eo eo
For the UHPFRC both the compressive and tensile uniaxial stress-strain responses
are adopted from experimental results and a calibrated model by Shafieifar et al. [11]. In
addition to the stress-strain relationships, the CDP model requires to define other plas-
ticity parameters. The eccentricity n represents the difference between the plastic
potential function and its asymptote. The ratio of biaxial to uniaxial compressive strength
(rb0/rco) is related to the plane stress yield surfaces. The shape of yield surface in the
deviatory plane is defined by K. The dilation angle W describes the inclination of the
plastic strain increment vector. Visco-plastic regularization of concrete is defined by the

viscosity parameter [10, 11]. The CDP parameters are adopted from previous studies and
a dilation angle of 40° is used for regular concrete and 56° for UHPFRC. The other
parameters n, rb0/rco, K, and viscosity are similar in regular concrete and UHPFRC and
set to be 0.1, 1.16, 0.667, and 0 respectively based on [10, 11].
150 10.0
Compressive stress (MPa)
125 8.0
Tensile stress (MPa)

100
6.0
75 UHPFRC UHPFRC
Regular Concrete 4.0
50 Regular Concrete
25 2.0
0 0.0
0 0.003 0.006 0.009 0.012 0 0.01 0.02 0.03 0.04 0.05
Strain (mm/mm) strain (mm/mm)
(a) (b)
Fig. 3. Stress-strain response for regular concrete and UHPFRC used in SB series (a) compres-
sive and (b) tensile.
3.1 Model Calibration

The FEA models are first calibrated using experimental results of the two control
specimens, SB1 [3], and PBM1 [6]. The FEA model reasonably predicted the behaviour
of the slabs under punching shear. Both of the simulations failed in punching shear with a
sudden drop in resistance (typical punching shear failure), similar to the actual experi-
ments. Figure 4 compares PBM1 specimen experimental and FEA results for crack
distribution at failure. In the concrete damage plasticity model, cracks initiate when the
principal plastic strain becomes positive [10]. Thus, the crack propagation in the model is
represented by the maximum principal stress and agrees with the test results.
Fig. 4. Crack pattern of PBM1 at failure for (a) experimental (reproduced from [6]) and
(b) FEA.
Experimental and FEA load versus deflection (displacement of centre relative to the
edges) plots of SB1 and PBM1 specimens are presented in Fig. 5. For SB1, numerical
results of the maximum punching load (VR) and the corresponding displacement at this
load (UR) are 87.16% and 98.31% of the test result, respectively. In PBM1, VR and UR
are 99.89% and 73.86% of the experimental results, respectively. The FEA overesti-
mated PBM1 loads corresponding to initial displacements and underestimated the final
displacement. This is probably due to the assumption in the model that the RC slab and
the UHPFRC are perfectly bonded. The initial slabs are also assumed to be straight
while there may have been displacements due to imperfection in fabrication. Overall,
the FEA model reasonably predicted the behaviour of the control specimens both in
terms of load-deflection and crack propagation and can be used to study the effec-
tiveness of the proposed retrofitting orientation.
350 1400
PBM1
300 SB1 1200 Experiment
Experiment
FE Analysis
FE Analysis
250 1000
Load (kN)
Load (kN)
200 800
150 600
100 400
50 200
0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 16
Displacement (mm) Displacement (mm)
Fig. 5. Load deflection curves-Comparison between experimental and FEA results of SB1 and
PBM1.
All specimens failed in punching shear. The maximum punching load (VR) and the
displacement at the maximum load (UR) are shown in Table 3, where, VR-NC and
UR-NC denote the maximum punching resistance and its corresponding displacement
for specimens without UHPFRC overlay (reference specimens). The reference speci-
mens for SB series and PBM series are SB1 and PBMF0, respectively. For both SB and
PBM groups, the UHPFRC layer increased punching shear resistance of retrofitted
specimens.
The UHPFRC increased the bending rigidity by resisting the tensile stress that
allows the composite to carry more load. Fully covering the smaller size (1500 mm by
1500 mm width and 120 mm thick) reinforced concrete slab with 40 mm thick
UHPFRC resulted in a 39% increase on the punching load resistance. Likewise, the
fully covered PBM slab has 65% higher punching load resistance than the un-retrofitted
slab. However, for the fully covered specimens, the displacement at the maximum
punching load is lower than the un-retrofitted slab because the UHPFRC layer act as a
tensile reinforcement and increased rigidity.
Table 3. FEA results and comparison between retrofitted and un-retrofitted specimens
Specimen VR (kN) UR (mm) VR/VR-NC UR/UR-NC
SB1a 220.54 11.70 1.00 1.00
SBF 305.47 8.30 1.39 0.71
SBP1 251.43 9.07 1.14 0.77
SBP2 258.81 11.64 1.17 0.99
SBP3 304.05 8.90 1.38 0.76
PBM1b 1087.80 10.34 1.65 0.46
PBMF0 659.74 22.68 1.00 1.00
PBMP1 747.23 19.85 1.13 0.88
PBMP2 937.09 15.89 1.42 0.70
PBMP3 1018.61 10.94 1.54 0.48
a
Tested by Adetifa and Polak [3], and b Tested by
Bastien-Masse and Brühwiler [6].
4.1 Optimum Use of UHPFRC for Retrofitting RC Slab-Column

Connections
Figure 6 shows the load-displacement responses of the analysed specimens. The pro-
posed configuration for retrofitting slab column connections by partially covering with
UHPFRC seems effective in terms of economy and ductility. Covering 56% and 75%
of the 120 mm thick RC slab (SB1) with 40 mm thick UHPFRC increased the
punching shear capacity by 14% and 17% respectively. The capacity of the fully
covered specimen is recorded after 85% of the area is covered by UHPFRC in SB
specimens. Likewise, more than 96% of the fully covered capacity is reached by
covering three-fourth of PBM1. Using 56% and 75% area coverage enhanced punching
shear capacity of the 210 mm thick slab (PBM series) by 42% and 54% respectively.
In both SB and PBM groups, the partially covered specimens have a higher dis-
placement (at maximum load) than the fully covered specimens. This higher ductility is
probably due to the relatively lower rigidity of the partly covered specimen than the full
covered slabs.
The crack distribution in the specimens is also significantly affected by the
UHPFRC overlay. The UHPFRC layer delayed the initial crack propagation. However,
the strain on the retrofitted slab at maximum load is higher than the corresponding
strain of un-retrofitted slab, as shown in Fig. 7. Also, propagation of cracks far from the
punching shear cone is limited in retrofitted specimens.
350 1400
PBM1
300 1200 PBMF0
PBMP1
PBMP2
250 1000 PBMP3
Load (kN)
Load (kN)
200 800
150 SB1 600

SBP1
100 SBP2 400
SBP3
50 SBF
200
0 0
0 2 4 6 8 10 12 14 16 18 0 4 8 12 16 20 24 28
Displacement (mm) Displacement (mm)
Fig. 6. Load deflection responses of specimens.
Fig. 7. Maximum principal strain distribution (visualized crack pattern) at maximum punching
load for (a) SB1 and (b) SBP1.
5 Conclusions
Punching shear performance of reinforced concrete slab-column connections

strengthened by a thin UHPFRC top layer are evaluated using 3D nonlinear FEM. The
following conclusions can be drawn from this study:
• Retrofitting reinforced concrete slabs by placing UHPFRC on the tension side
enhances the punching shear resistance. The maximum punching load capacity of
1500 mm by 1500 mm reinforced concrete slabs increased by 39% after post
installing 40 mm thick UHPFRC layer. Likewise, 50 mm thick UHPFRC overlay
increased punching shear capacity of 3000 mm by 3000 mm slab by 65%.
• Partially covering the reinforced concrete slabs using UHPFRC with the proposed
configuration is a promising retrofit technique since it results in an economical and
more ductile connection. The capacity of the fully covered specimen is recorded
after 85% of the area is covered by the UHPFRC in SB1 slab, and more than 96% of
the fully covered capacity is reached by covering three-fourth of the PBM1 slab.
• The UHPFRC layer affects the displacement and crack propagation of the analysed
specimens. Fully covered specimens show smaller displacement at the maximum
punching load. However, higher ductility is recorded for partly covered specimens
than the fully covered slabs.
Acknowledgements. The authors would like to thank the MasterCard Foundation Scholars
Program (https://mastercardfdn.org/foundation/) for the financial support provided to the first
author.
References
1. Azizi, R., Talaeitaba, S.B.: Punching shear strengthening of flat slabs with CFRP on grooves
(EBROG) and external rebars sticking in grooves. Int. J. Adv. Struct. Eng. 11(1), 79–95
(2019)
2. Polak, M.A., Bu, W.: Design considerations for shear bolts in punching shear retrofit of
reinforced concrete slabs. ACI Struct. J. 110(1), 15 (2013)
3. Adetifa, B., Polak, M.A.: Retrofit of slab column interior connections using shear bolts. ACI
Struct. J. 102(2), 268 (2005)
4. El-Salakawy, E., Soudki, K., Polak, M.A.: Punching shear behavior of flat slabs strengthened
with fiber reinforced polymer laminates. J. Compos. Constr. 8(5), 384–392 (2004)
5. Meisami, M.H., Mostofinejad, D., Nakamura, H.: Strengthening of flat slabs with FRP fan
for punching shear. Compos. Struct. 119, 305–314 (2015)
6. Bastien-Masse, M., Brühwiler, E.: Experimental investigation on punching resistance of R-
UHPFRC–RC composite slabs. Mater. Struct. 49(5), 1573–1590 (2015)
7. Habel, K., Denarié, E., Brühwiler, E.: Structural response of elements combining ultrahigh-
performance fiber-reinforced concretes and reinforced concrete. J. Eng. Struct. 132(11),
1793–1800 (2006)
8. Yin, H., Teo, W., Shirai, K.: Experimental investigation on the behaviour of reinforced
concrete slabs strengthened with ultra-high performance concrete. Constr. Build. Mater. 155,
463–474 (2017)
9. ABAQUS Analysis user’s manual 6.10-EF, Dassault Systèmes Simulia Corporation,
(Providence, RI, USA, 2010)
10. Genikomsou, A.S., Polak, M.A.: Finite element analysis of punching shear of concrete slabs
using damaged plasticity model in ABAQUS. Eng. Struct. 98, 38–48 (2015)
11. Shafieifar, M., Farzad, M., Azizinamini, A.: Experimental and numerical study on
mechanical properties of Ultra High Performance Concrete (UHPC). Constr. Build. Mater.
156, 402–411 (2017)
Numerical Evaluation the Effect of Specimen
Thickness on Fibre Orientation
in Self-consolidating Engineered
Hai Tran Thanh1(&), Jianchun Li1, and Y. X. Zhang2

1
Centre for Built Infrastructure Research, Faculty of Engineering
and Information Technology, University of Technology Sydney,
Sydney, NSW 2007, Australia
Hai.TranThanh@student.uts.edu.au
2
School of Computing, Engineering and Mathematics,
Western Sydney University, Sydney, NSW 2751, Australia
Abstract. The orientation of distributed synthetic fibres in the matrix of

engineered cementitious composites (ECCs) governs their capability of bearing
stress and bridging micro-cracks. Understanding the orientation of synthetic
fibres in ECCs matrix of different thickness specimens, therefore, is necessary.
In the present work, the effect of specimen thickness on the orientation of
synthetic fibres in self-consolidating (SC) ECC, a typical member of family ECC
materials, is numerically investigated through the simulation of the casting of
fresh SC-ECC into different thicknesses of moulds. The moulding of fresh SC-
ECC, which is discretised by a limit number of separated mortar and fibre
particles, is simulated using the mesh-free smoothed particle hydrodynamics
(SPH) method. The synthetic fibre utilised in SC-ECC is considered as flexible
fibre and virtually connected by a drag force between two adjacent fibre parti-
cles. The SPH allows tracking the movement of mortar and fibre particle during
their flow, thus providing the real image of flexible synthetic fibres orientation in
specimens during the casting process. A simple method is proposed to evaluate
the orientation of flexible synthetic fibres at various sections of specimens after
the SC-ECC stop flowing in the moulds. The results of this study reveal that thin
specimens tend to have higher fibre orientation factors than thick specimens.
Synthetic fibres tend to parallel with the longitudinal direction of specimens at
the bottom of the formwork and rotate freely at the top surface of specimens.
Keywords: Self-consolidating ECC Flexible synthetic fibre Fibre

orientation Thickness effect
1 Introduction
Short synthetic fibres such as polyvinyl alcohol (PVA), polyethylene (PE) or

polypropylene (PP) fibres have been micromechanical incorporated with cementitious
matrix to form a unique class of high-performance fibre reinforced cementitious
composites. This type of materials, named as Engineered Cementitious Composites
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_59
662 H. Tran Thanh et al.
(ECCs) [1], exhibits a high tensile ductility with tensile strain capacities up to 5% at a
small volume of fibres fraction (less than 2%).
In ECCs, the inclination of distributed synthetic fibres in the cementitious matrix is
recognised to play a critical role in the capability of bearing stress and bridging micro-
cracks [2]. The larger inclined fibre tends to premature rupture than the smaller one due
to the snubbing effect and the reduction of its apparent strength [3]. This intrinsic
nature of synthetic fibres raises the concern of how the size of specimens affects the
orientation of synthetic fibres, which will then influence the mechanical properties of
ECC structural members. However, studies of size-effects on the orientation of syn-
thetic fibres utilised in ECCs are still rarely.
Unlike steel fibres which are rigidly rotated and translated in the mixture during the
mixing and moulding process, the rotation and translation of synthetic fibres in ECC
materials are more complex in nature. Synthetic fibres are tiny in diameter (ranging
around 0.012 to 0.039 mm) and flexible to be bent or coiled in the mix during the
process of producing. This property could lead to a wide scatter of the orientation of
synthetic fibres in the matrix, even casting by the same batch and same technique. As
one of the crucial factor, such variation can be explained for the fluctuation of
mechanical properties in the experimental study of Rokugo et al. [4], but the effect of
specimens thickness on the fibre orientation is unclear. In another study, Lu and Leung
[5] theoretical studied the influence of the specimen thickness on the tensile perfor-
mance of ECC by limiting the rotation of fibres at near the top and bottom surface of
different thickness specimens. The results concluded that thick specimens exhibited
lower tensile strength and ductility than thin specimens as fibres tend to rotate more in
thicker specimens. However, fibres were assumed to be straight and identically
distributed.
In this context, numerical simulation provides an ideal tool to investigate the
influence of specimen thickness on the orientation of flexible synthetic fibres utilised in
ECC materials, gaining an insight into the dispersion of fibres orientation in ECCs
matrix. Recently, the authors developed a model for simulating the flow of self-
consolidating (SC) ECC, which was validated by the slump cone test in 2D simulation
[6], the V-funnel and U-box tests in 3D simulations [7]. The developed model provided
the real image of the fresh SC-ECC, the actual distribution of synthetic fibres and their
orientations in the mix during and after stop flowing. Hence, to examine the influence
of specimen thickness on the orientation of synthetic fibre, the mouldings of fresh SC-
ECC into three different thicknesses are simulated by using the developed 3D model in
this paper. The orientation of synthetic fibres at various sections of specimens is
evaluated after fresh SC-ECC stop flowing in the formworks. The results of this study
reveal the significant effect of element thickness on the orientation of flexible synthetic
fibres.
2 Simulation the Moulding of Fresh SC-ECC
In the moulding process of fresh SC-ECC, the mixture starts flowing into formwork
once the gravitation force overcomes its yield stress and then restrained by its viscosity
and the boundary friction. Due to the large deformation of this process, the mesh-free
Numerical Evaluation the Effect of Specimen Thickness on Fibre Orientation 663
smoothed particle hydrodynamic (SPH) is a powerful method for solving the

Lagrangian form of the governing equations during its flow. In SPH approach, the
entire domain of fresh SC-ECC is discretised by a limit number of separated mortar and
fibre particles which possessing the same continuum properties, excepting their
assigned number. Then field variables of a current particle a is approximated by a
summation of all the values of surrounded particles b in its dominant range kh based on
a kernel function W (Fig. 1). A drag force T, between two neighbouring fibre particles
of synthetic fibre, is added in the momentum equation each calculation step. This
tension force only exerts attractively when two adjacent fibre particles recede from each
other. Meanwhile, synthetic fibre is considered as being bent or coiled if its two
adjacent particles are approaching each other.
(
rab
EA rab
1 if rab rab
o
Tab ¼ o
rab rab ð1Þ
0 if rab \rab
o
o
where rab is the initial distance between particles a and b, and EA is the tensile rigidity
of utilised fibre.
Fig. 1. Interpolation of field variable of a current particle a by a kernel function W and its
support domain. [7]
By implementing the constitutive rheology Cross model [8] to describe the relation
of the shear stress and strain tensors during the flow of fresh SC-ECC, a model has been
developed for simulating the flow of self-consolidating ECC using PVA fibre by
authors. The model was validated by the slump cone test, the V-funnel and U-box tests
(Fig. 2) in 2D and 3D simulations [6, 7]. In this model, the yield stress of SC-ECC was
first assumed, and then the plastic viscosity was estimated using an enormous error
trial-corrector SPH modelling [6]. The values of these two-rheology parameters, the
density of mixture and properties of PVA fibre utilised as input data are shown in
Table 1. It should be mentioned that the cubic spline kernel was used for the inter-
polation, and the repulsive force boundary particle was employed in this model.
A kinematic coefficient was also applied to account for the friction between the mixture
and wall-boundary particles [7].
Fig. 2. Simulation of fresh SC-ECC in two standard tests: (a) V-funnel; (b) U-box [7]
In the present work, the developed 3D model is used to numerical investigate the
influence of the thickness of beams on the orientation of synthetic fibres in SC-ECC
using PVA fibre through simulating the casting process of SC-ECC into beam speci-
mens. The moulding of beams with three different thicknesses, T, of 30 mm, 50 mm
and 100 mm are simulated with the same position of funnel mouth (Fig. 3) and
identical initial dispersion of orientated fibres. The results of these simulations are then
used to quantify fibre orientation distribution in different thickness specimens.
100
H mm
mm
depth
Flow
350 mm
Fig. 3. Moulding of beam specimens

3 Initial Configuration
The funnel position is at the end of beams with the same distance from its bottom outlet
to the bottom of formwork in three cases. Mortar and synthetic fibre particles are then
created in the funnel-boxes, which the height value, H, is calculated corresponding to
three thickness beams. In all cases, synthetic fibres are created straightly with the same
random distribution and orientation, reflecting the fresh mixes are in the same batch
with similar dispersion of orientated synthetic fibres. The details of initially created
mortar and fibre particles are shown in Fig. 4 and highlighted as follows:
• Mortar particles are initially arranged in a square grid form with the x-axis, y-axis
and z-axis spacing equal to 4 mm.
• Fibre particles are generated randomly with a range of inclined angles from 0 to 90
degrees.
• Synthetic fibres are straight in the beginning, and the distance between the two
neighbouring particles within a fibre is chosen to be equal to 2 mm.
Table 1. Properties of SC-ECC mix and PVA fibres.

SC-ECC PVA fibre
Density (kg/m3) 2,065 Young modulus (Gpa) 42.8
Plastic viscosity (Pa.s) 17 Diameter (mm) 0.039
Yield stress (Pa) 165 Length (mm) 12
Fig. 4. Initial mortar and synthetic fibre particles for beams thickness T: (a) T = 30 mm;
(b) T = 50 mm; (c) T = 100 mm
The number of fibre in each case is calculated based on the number of mortar
particles and its 2% volume fraction. Table 2 shows the number of boundary, mortar
and synthetic fibre particles involving in three computations.
Table 2. The number of involved particles in simulations

Beams T = 30 mm T = 50 mm T = 100 mm
Boundary particles 13,260 15,318 20,512
Mortar particles 18,216 29,808 59,064
Fibre volume fraction (%) 2 2 2
Number of fibres 364 596 1,181
Fibre particles 2,548 4,172 8,267
Total particles 34,024 49,298 87,843

4.1 Simulation Results and Cutting Specimens
The results of the moulding simulations of three-thickness beams when material par-
ticles of fresh SC-ECC stop flowing are shown in Fig. 5. For facile observation of fibre
orientation distribution, only fibre particles are plotted. Due to the limitation in time
and the bottom of funnel mouth is near the surface of T = 100 mm beam, some mortar
and fibre particles still move slowly at the funnel mouth when the simulation stops after
20 s computational pre-determination time. However, this limitation does not affect the
result of fibre orientation of T = 100 mm beam in this work.
Fig. 5. Moulding results of three thickness beams: (a) T = 30 mm; (b) T = 50 mm;
(c) T = 100 mm
Three beams are then cut virtually by multi-vertical and horizontal planes to
quantify the orientation of synthetic fibres at cutting sections. Considering that a short
straight line links two neighbouring fibre particles within a fibre, a PVA fibre will be
represented by six continuity lines (Fig. 6b). With the initial distance between the two
nearby fibre particles equals 2 mm in this study, the spacing of cut planes in vertical
and horizontal directions are selected to be equal to 2 mm and 4 mm, respectively.
For vertical planes, this value is comparable with the reported crack spacing in an
experimental study of SC-ECC using PVA fibre [9].
(a) Specimen Vertical cut plane (b)
Horizontal cut plane
Fig. 6. (a) Vertical and horizontal cutting planes; (b) Fibre particles intersection at multi-cut
planes
4.2 Fibre Orientation Factor

The inclined angle of a synthetic fibre j at a vertical section is determined by the cosine
of the angle aj between two adjacent fibre particles of a fibre with the normal of the
vertical plane (x-axis for vertical plane yz). The fibre orientation factor, hi at plane i, is
the mean value of the cosine angle of n intersected fibres at that plane. It is cleared that
only two neighbouring fibre particles in the opposite side of a vertical cut plane are
accounted in the orientation factor of fibre at that plane.
1 Xn
hi ¼ cosaj ð2Þ
n j¼1
It can be seen from Fig. 7 the variation of fibre orientation factors along the
specimens. The thin specimen tends to have higher fibre orientation factor than the
thick specimen. This result reveals that synthetic fibres in the matrix of thin specimen
tend to parallel more with the longitudinal direction than in the thick specimen.
4.3 Fibre Ratio Along Depths of Specimen

To clarify the influence of bottom formwork and free surface of a specimen on the
orientation of synthetic fibres in SC-ECC matrix, a fibre ratio along depths of a
specimen is proposed in this work. It is defined as the ratio of the number of fibre
particles locating between two adjacent cut planes in the horizontal direction with the
total amount of fibre particles. It is clear that if fibres tend to parallel with the bottom or
free surface of a specimen, the number of fibre particles counted in these areas would
larger than that of other parts of this specimen.
As shown in Figs. 8, the bottom of formwork significantly affects the orientation of
synthetic fibres with resulting in fibres tend to align with the longitudinal direction of
specimens in this zone. Thus, the ratio of fibre in this area is remarkable larger than in
Fig. 7. Distribution of fibre orientation factor along three different thickness beams
other regions of specimens. Additionally, this influence is more noticeable in thin

specimens in comparison with thick specimens. Meanwhile, synthetic fibres tend to
rotate freely at the top surface of specimens.
Fig. 8. Fibre ratio along depths of three different thickness beams
5 Conclusions
The mouldings of fresh SC-ECC into the formworks of three different thickness beams
are simulated using the developed 3D simulation model with the approach of SPH.
From analysing the orientation of flexible synthetic fibres in simulated specimens, the
following conclusions can be drawn:
• The orientation factors of synthetic fibres in SC-ECC vary along the beam
specimens.
• Thin specimens tend to have higher fibre orientation factors than thick specimens.
• The orientation of synthetic fibres is remarkably affected by the bottom of the

formwork. Synthetic fibres are likely paralleled with the longitudinal direction of
specimens in this region.
• Synthetic fibres tend to rotate freely at the top surface of specimens.
References
1. Li, V.C.: From micromechanics to structural engineering-the design of cementitous
composites for civil engineering applications. J. Struct. Mech. Earthq. Eng. JSCE 10(2),
37–48 (1993)
2. Kanakubo, T., Miyaguchi, M., Asano, K.: Influence of fiber orientation on bridging
performance of polyvinyl alcohol fiber-reinforced cementitious composite. ACI Mater. J. 113
(2), 131–141 (2016)
3. Kanda, T., Li, V.C.: Interface property and apparent strength of high-strength hydrophilic
fiber in cement matrix. J. Mater. Civil Eng. 10(1), 5–13 (1998)
4. Rokugo, K., Uchida, Y., Moriyama, M., Lim, S.: Direct tensile behavior and size effect of
strain-hardening fiber-reinforced cement-based composites (SHCC). In: 6th International
Conference on Fracture Mechanics of Concrete and Concrete Structures (2007)
5. Lu, C., Leung, C.K.: Theoretical evaluation of fiber orientation and its effects on mechanical
properties in Engineered Cementitious Composites (ECC) with various thicknesses. Cem.
Concrete Res. 95, 240–246 (2017)
6. Thanh, H.T., Li, J., Zhang, Y.: Numerical modelling of the flow of self-consolidating
engineered cementitious composites using smoothed particle hydrodynamics. Constr. Build.
Mater. 211, 109–119 (2019)
7. Thanh, H.T., Li, J., Zhang, Y.X.: Numerical simulation of self-consolidating engineered
cementitious composite flow with the V-funnel and U-box. Constr. Build. Mater. 236, 117467
(2020)
8. Shao, S., Lo, E.Y.M.: Incompressible SPH method for simulating Newtonian and non-
Newtonian flows with a free surface. Adv. Water Resour. 26(7), 787–800 (2003)
9. Kong, H.-J., Bike, S.G., Li, V.C.: Constitutive rheological control to develop a self-
consolidating engineered cementitious composite reinforced with hydrophilic poly (vinyl
alcohol) fibers. Cem. Concr. Compos. 25(3), 333–341 (2003)
Numerical Modelling of Fiber-Reinforced
Concrete Shear-Critical Beams
Santiago Talavera-Sánchez, Juan Navarro-Gregori(&),

Francisco Ortiz-Navas, and Pedro Serna
ICITECH, Universitat Politècnica de València, Valencia, Spain

juanagre@cst.upv.es
Abstract. The numerical modelling of shear-critical reinforced concrete beam

has been a traditional challenging problem due to the varying mechanisms
involving the shearing problem. The inclusion of fibres improves significantly
the shear strength of the beams but include in the equation a new unknown. In
this study, a group of 16 beams with varying parameters including steel or
macro-synthetic fibres, presence or not of transverse reinforcement, different
shear-span-to-depth ratios, and different transverse reinforcement ratios are
studied. The numerical modelling makes use of the nonlinear finite element
analysis following the smeared crack approach and a total strain based crack
material model. The numerical modelling has shown that a nonlinear finite
element model can predict with great accuracy the behaviour and the strength in
those beams with presence of transverse reinforcement. For the case of members
without transverse reinforcement, the predictions of the shear strength are
acceptable but some doubts remain unclear since these beams are characterized
by a failure originated in a major critical crack. A parametric analysis based on
the crack bands employed is discussed to give recommendations for a good
numerical modelling of FRC shear-critical beams.
Keywords: Shear Beams Fibres Numerical model
1 Introduction
One of the attempts to reduce the brittleness of concrete has been the inclusion of
dispersed fibres in its matrix. In fact, the use of macro fibres significantly improves the
ductility, toughness and tensile strength of concrete members such as beams, slabs,
tunnel segments, pavements, or columns among others. However, with the increase of
knowledge about fibres, it has been proved that macro fibres result a good mechanism
to increase the shear strength of structural concrete members. Moreover, fibres can
prove their effectiveness to substitute partially or totally the conventional reinforce-
ment. However, the effect of fibres on shear behaviour continues on debate as many
factors are involved like compressive concrete strength, longitudinal reinforcement
ratio, aggregate size, prestressing, load conditions, shear span-to-effective-depth ratio,
fibre type, effective depth, or fibre properties (content, shape, slenderness and material).
In fact, some of these factors are still on the debate stage in traditional reinforced
concrete (RC) elements. In order to increase knowledge about shear, many
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_60
Numerical Modelling of Fiber-Reinforced Concrete 671
investigations have been performed in the last decades to explain the contribution of the
factors involved, as well as their influence on shear performance and modes of failure.
Given the background, one of the classical beam series performed by Bresler-Scordelis
[1] were 12 RC beams tested in order to study the behaviour of shear-critical beams.
Nearly forty years later, the classical Bresler-Scordelis program was re-examined by
Vecchio and Shim [2]. Experimental results were used to be a reference for the cali-
bration of nonlinear finite element models of RC shear-critical beams.
Several investigations using non-linear finite elements analysis (NLFEA) have been
developed in the last years to validate and analyse experimental results in order to
evaluate the controlling parameters in shear behaviour of fibre reinforced concrete
(FRC) elements.
Within this context, a recent experimental campaign based on the beams by
Bresler-Scordelis was carried out. Some of the beams were once again re-examined [3].
However, the program was extended using steel fibre-reinforced concrete (SFRC) and
polypropylene fibre-reinforced concrete (PFRC). In this way, the experimental cam-
paign covered parameters such as presence or not of transverse reinforcement, steel or
polypropylene fibres, span conditions. Hence, a combination of factors that could
significantly influence the mechanisms of shear transfer in FRC beams was considered.
In this work, it is developed the numerical simulations of the experimental pro-
gramme carried out. It will be used a NLFEA software to model the shear response of
the beams tested. The numerical model follows the smeared crack approach and a total
strain based crack material model. The tensile properties employed are derived from
inverse analysis of prismatic beams tested according to EN14651 [4].
Research performed by Bresler-Scordelis consisted of 12 beam (four series of three

beams). All beams series had the same total depth of 552 mm. However, each series
varied in amount of longitudinal and transverse reinforcement, width and shear-span-
to-depth ratio (a/d). Bresler-Scordelis beams covered a wide range of parameters in
order to achieve different modes of failure (MOF). The present research reproduces
some of the beams by Bresler-Scordelis. These beams are 0A1, 0A2, A1, A2, B1, and
B2. Additionally, two new cross-sections were added: 0B1 and 0B2, which have the
same cross-section than beams B1 and B2 but without transverse reinforcement.
Moreover, as the main objective of the present research is to study the shear behaviour
of FRC elements, the selected RC beams were extended to specimens manufactured
with SFRC and PFRC. So, a total of 24 beams were tested (8RC, 8PFRC and 8SFRC)
at the ICITECH facilities. Details of the cross-section properties of only those beams
with fibres tested as well geometries are shown in Fig. 1 and Table 1. Beams are
grouped in four series (A1, A2, B1 and B2), and each series contained various levels of
shear reinforcement (non-reinforced, stirrups, steel fibres (ST), polypropylene fibres
(PP) and a combination between stirrups and fibres).
To prevent failure due to insufficient anchorage of longitudinal bars, length of
beams was extended 700 mm from the support (see Fig. 1). Moreover, in beams with
conventional reinforcement, bars of 8 mm anchored with 135° were used as transverse
672 S. Talavera-Sánchez et al.
Fig. 1. Beams cross-section (left) and beam length (right).
Table 1. Cross-section details of the ICITECH beams.
reinforcement. It should be pointed out that the longitudinal and transverse rein-
forcement ratios were similar to those used by Bresler-Scordelis. All beams were tested
at 3PBT, adopting a shear-to-span-ratio (a/d) equal to 3.87 and 4.84, respectively.
Two types of crushed limestone gravel, three types of sand and Portland cement
type CEM I 42.5 N were used in all concrete batches. The water to cement ratio was
0.5 and in order to obtain a good workability, superplasticizer was used in ratios of
1.4% to 1.5% of cement weight. Further details about concrete doses used in this
investigation can be found in Ortiz et al. 2018 [3].
Two type of macro fibres were employed in the present research. Polypropylene
fibres, 48 mm long with a nominal aspect ratio (length/diameter) equal to 57, were
added at 10 kg/m3 in the PFRC specimens. Tensile strength and modulus of elasticity
were 400 MPa and 4.7 GPa, respectively. Concerning to steel fibres, double hooked-
end steel cold-drawn fibres were used at a dose of 30 kg/m3 in SFRC elements. Fibres
were 0.9 mm in diameter and 60 mm long with a nominal aspect ratio
(length/diameter) equal to 65. Tensile strengths and modulus of elasticity were
2300 MPa and 2100000 MPa, respectively. Steel reinforcement bars were evaluated
according to EN-ISO 6892-1[5]. The yielding stress fy and the ultimate stress fu of
rebars were 518 and 660 MPa for 8 mm bar, 526 and 670 for 10 mm bar, 529 and 640
for 12 mm bar, 579 and 662 for 20 mm bar, and 579 and 662 for 25 mm bar.
In order to obtain the mechanical properties (compression strength, modulus of
elasticity and residual tensile strength) of the three types of concrete, 72 cylindrical
specimens of 150 300mm and 82 prismatic beams of 150 150 600 were
prepared and cured at indoor in a precast concrete plant during the beams’ manufacture.
Compression strength, modulus of elasticity and residual tensile strength were deter-
mined according to EN 12390-3 [6], EN 12390-13 [7] and EN 14651 [4], respectively.
Table 1 lists the compression strength and modulus of elasticity of each beam.
Moreover, Fig. 2a and Fig. 2b show the flexural tensile strength vs. CMOD response
of SFRC and PFRC, respectively. In both figures, the main campaign of prismatic
specimens has been plotted as well as their corresponding mean and characteristic
values. Similar mean values were achieved in both types of concrete. However,
because of the variability achieved in SFRC, some differences in terms of characteristic
values were found with PFRC. The obtained results, fulfil the requirements of Model
Code 2010 [8] for the use of fibres in structural elements since fR,1/fL > 0.40 and
fR,3/fR,1 > 0.50 in terms of mean values.
9.0 9.0
Strees SFRC Strees PFRC
8.0 [MPa] 8.0 [MPa]
7.0 7.0
6.0 6.0
Mean
5.0 Mean 5.0
4.0 4.0 Characteristic
3.0 Characteristic 3.0

2.0 2.0
1.0 1.0
CMOD [mm] 0.0 CMOD [mm]
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Fig. 2. Residual tensile strength of steel (left) and polypropylene (right) fibres
Concerning to test set-up, beams were tested in a three-point loading scheme.

A servo-hydraulic jack was used to apply the load at an average deflection of
0.30 ± 0.10 mm/min. The support system was constituted by a pinned and roller
support (350 350 20 mm) to allow the horizontal displacements of beam due to
bending and shearing deformations (Fig. 3). Nine potentiometric displacement trans-
ducers (PTs) were employed to capture deflection at mid-span (two PTs 50 mm),
vertical movements in the supports (two PTs 125 mm), shear crack opening (two PTs
of 125 mm), average strain in the compression and tension chords at mid span (2 PTs
50 mm), and compression chord between the loading application point and the sup-
ports (two PTs 50 mm).
2000000 16
Inverse SFRC SRFC Test
Inverse Analysis
MC2010 Analysis PFRC
1500000 12
σ (N/m2)
P (kN)
1000000 8
500000 4
0 0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0 0.5 1 1.5 2 2.5 3 3.5 4
Strain (ε) CMOD (mm)
Fig. 3. Characteristic tensile properties of FRC derived from bending tests.
3 Test Observations
Similar to Bresler-Scordelis beams, three different modes of failure were observed in

FRC beams. Diagonal tension (DT) in all those beams without stirrups, while in the
case of beams with stirrups shear-compression (VC) and flexural-compression
(FC) were observed. Table 2 shows maximum load (Pu) and the experimental mode
of failure (MOF) achieved for each beam tested.
Table 2. Experimental and numerical beam results

Beam MOF Pu MOF Pcalc,25 Pu/ MOF Pcalc,60 Pu/
Exp. [kN] Calc,25 [kN] Pcalc,25 Calc,60 [kN] Pcalc,60
OAP1 D-T 446.68 D-T 424.92 1.051 D-T 503.56 1.384
OAS1 D-T 466.47 D-T 354.01 1.318 D-T 391.64 1.176
AP1 V-C 695.30 F-C 708.59 0.981 V-C 704.92 0.969
AS1 F-C 682.43 V-C 696.61 0.980 F-C 695.61 0.956
OAP2 D-T 486.26 D-T 471.13 1.032 D-T 334.19 0.766
OAS2 V-C 411.00 D-T 345.61 1.189 D-T 315.85 1.349
AP2 V-C 678.60 F-C 674.86 1.006 F-C 673.37 0.965
AS2 V-C 658.07 F-C 672.03 0.979 V-C 659.94 1.000
OBP1 D-T 361.73 D-T 344.98 1.049 D-T 261.42 1.384
OBS1 D-T 309.86 D-T 289.75 1.069 D-T 263.53 1.176
BP1 V-C 555.56 V-C 599.82 0.926 V-C 573.44 0.969
BS1 V-C 546.22 V-C 569.56 0.959 V-C 571.37 0.956
(continued)
Beam MOF Pu MOF Pcalc,25 Pu/ MOF Pcalc,60 Pu/
Exp. [kN] Calc,25 [kN] Pcalc,25 Calc,60 [kN] Pcalc,60
OBP2 D-T 295.60 D-T 340.48 0.868 D-T 385.87 0.766
OBS2 D-T 326.89 D-T 263.03 1.243 D-T 242.27 1.349
BP2 F-C 487.37 F-C 504.01 0.967 F-C 504.82 0.965
BS2 V-C 492.74 F-C 498.50 0.988 V-C 492.69 1.000
Stirrups Avg 0.973 0.983
CV (%) 2.42 1.87
No Avg 1.102 1.206
Stirrups
CV (%) 12.83 21.40
Total Avg 1.038 1.095
CV (%) 11.41 19.29
4 Numerical Simulations
The numerical simulation of the 16 fibre-reinforced concrete beams included in this

paper has been carried out in two main stages.
First, an iterative inverse analysis to derive the tensile properties of the SFRC and
PFRC from the bending tests has been carried out. Thus, a stress-strain curve response
has been obtained assuming two characteristic lengths or crack bands of 25 mm and
60 mm. The influence of crack band is studied throughout the paper. It is important to
mention that tensile properties used in the numerical simulations are the characteristic
values, which is in the same vein as [2] in RC elements. Figure 3 shows the tensile
properties for both fibres (for 25 mm crack band). In Fig. 3 (left) is also shown the
prediction of the MC2010 simplified post-crack constitutive law, which is close to the
obtained by inverse analysis.
Next, a nonlinear numerical analysis using NLFEA software DIANA has been
carried out. The material properties introduced were the ones derived from the iterative
inverse analysis. All beams were meshed (Fig. 4) using 8-node quadrilateral plane
stress elements, and in some cases 6-node triangular elements for refinement. The
reinforcement employed for the longitudinal and transverse reinforcement were mod-
elled as discrete truss elements. Moreover, 8-node quads were employed to model the
loading plate and the bearing supports. An interface element with null shear defor-
mation was employed in the loading plate to assure a uniform transmission of the
applied load to the FRC beams.
The material model used for the longitudinal and transverse reinforcement was a
linear perfectly plastic material according to the yield stress of the rebars used. The
material model used for the fibre-reinforced concrete was a 2D smeared total strain
crack model. In compression, the parabolic curve based on the EN-1992-1 was used,
while in tension a multilinear stress-strain curve derived from inverse analysis was
implemented. It is important to mention that in the inverse analysis the tensile strains
were adjusted to two different crack bands (25 and 60 mm). Two mesh sizes were
Crack band = 25mm F.E. = Interface element
Crack band = 60mm F.E. = Interface element
Fig. 4. Typical finite element meshes.
employed with its corresponding tensile curve (Fig. 4). The numerical solution strategy
was the arc-length using tangent-stiffness.
The numerical modelling of all the 16 beams has been carried for the two crack
bands. The results of the calculated ultimate loads, as well as the numerical mode of
failure obtained is reported in Table 2. In general, the prediction of the shear strength of
those beams including transverse reinforcement is very accurate. Moreover, a signifi-
cant difference for the two crack bands used was not observed.
Figure 5 shows the load-deflection curves obtained for the FRC beams with stirrups
corresponding to series A1, this is, beams AS1 and AP1. The prediction of the max-
imum load is excellent. Moreover, the prediction along the increasing load is very
accurate. There is insignificant difference between the estimations done with crack
bands of 25 mm and 60 mm. In the experimental test and in the numerical simulations
was observed that the longitudinal reinforcement reached yielding.
Figure 6 shows the load-deflection response obtained for series B1 with transverse
reinforcement (BS1 and BP1). The mode of failure was a shear-flexure compression
(V-C) both in the test and the numerical simulations. The accuracy in the load-
deflection response is accurate along the loading process.
Figure 7 shows the crack pattern predicted in the model at maximum load for
beams AP1 and BP1 for crack band of 25 mm. As it is a smeared model the crack
pattern obtained in the model is on average similar to the real obtained in the exper-
imental test, but the model is unable to depict the real cracks that were generated in the
test, just the average response.
Now, it is discussed the numerical predictions for FRC beams without transverse
reinforcement. By way of example, Fig. 8 shows the response obtained for beam OAP1.
The prediction is very accurate up to a certain load level where the shear-critical diagonal
crack took place in the test. At this point, the model may stop the numerical procedure
because of convergence problems or may keep a stiffness higher than the one observed in
the test. As stated before, this point meets when the critical diagonal crack takes place.
Therefore, the smeared model poorly performs when this singular and localized effect
occurs. However, the prediction up to this point is very accurate in terms of load-
deflection response. During the tests, horizontal cracks were detected at the height of the
lower longitudinal reinforcement, where the discrete crack starts, which could indicate
800
AP1
700
600
500
Load (kN)
400
Test
300
Numerical (cb=25mm)
200
Numerical (cb=60mm)
100
0
0 5 10 15 20 25 30 35 40 45 50
Mid-span deflection (mm)
800
700
AS1
600
500
Load (kN)
400
Test
300
Numerical (cb=25mm)
200
Numerical (cb=60mm)
100
0
0 5 10 15 20 25 30 35 40 45 50
Fig. 5. Load-deflection beams AP1 and AS1.
600 Test BP1

500 Numerical (cb=25mm)
Load (kN)
400
300
200
100
0
0 5 10 15 20 25 30
600 Test BS1

500
Numerical (cb=25mm)
400
Load (kN)
300
200
100
0
0 5 10 15 20 25 30
Fig. 6. Load-deflection response beams BP1 and BS1.

AP1
BP1
Fig. 7. Load-deflection response beams AP1 and BP1.
600
OAP1
500
400
Load (kN)
300
Test
200
Numerical (cb=25mm)
100
Numerical (cb=60mm)
0
0 2 4 6 8 10 12 14 16 18 20
Fig. 8. Response of FRC beam without transverse reinforcement (Beam OAP1)

that the failure of the reinforcement anchorage occurred at this moment, a phenomenon to
be introduced in the model. Thus, dowel action effect may give a better response in the
simulations of FRC beams without transverse reinforcement.
The use of the smeared cracking approach in the numerical modelling of FRC
elements with low or inexistent amounts of transverse reinforcement should be
addressed with caution to avoid erroneous predictions. This is probably because this
type of model does not present the numerical ‘memory’ of the damage already
occurred, i.e. the cracks that have occurred may close again and acquire load, which
would be physically unacceptable. It is therefore recommended to calculate with
reduced load steps in order to avoid incompatible or unrealistic load-discharge and
redistribution processes.
In order to validate the results obtained for use in real structural applications, it can
be concluded that in the case of elements with a longitudinal and transverse rein-
forcement level greater than the minimum demands, satisfactory and stable results
should be expected.
In elements without transverse reinforcement, where a greater instability in the
results is obtained, probably due to the fragility of the model due to uncontrolled
cracking, results are always obtained on the safety side for small steps, with some
exceptions. In this case, it would be recommended to continue with specific investi-
gations in order to adjust the calculation parameters, even resorting to more complex
NLFEAs including other aspects like the discrete crack approach or random fields [9]
for the material properties characterization throughout the FRC element.
5 Conclusions
On the basis of the results obtained and the checks carried out, the following con-
clusions can be drawn:
• The use of the characteristic tensile properties obtained from inverse analysis gives
satisfactory results in terms of global stiffness prediction for the group of FRC
beams analysed.
• The influence of the crack band (25 and 60 mm) on the results obtained are not
conclusive. Thus, its importance is unclear on the numerical predictions in the
specimens with and without transverse reinforcement.
• For all FRC beams with transverse reinforcement, the analyses carried out show
satisfactory results in terms of maximum load and deflection response, which
probably derives from the fact that the shear reinforcement allows for crack control,
with a smeared crack pattern similar to the 2D material model employed in the
numerical simulations. Moreover, there is little evidence of major crack redistri-
butions and/or localized failures in these beams.
• For all FRC beams without transverse reinforcement, the numerical simulations
predict well the load deflection-response up to the point of the formation of the
critical diagonal crack. In the numerical simulations, the crack in the beam tries to
rotate towards the support but is unable to generate the discrete crack that really
appears in the experimental test.
• The mode of failure of the beams analysed is similar in all cases, except for the
models without transverse reinforcement.
• The use of the smeared crack approach in the modelling of FRC elements with low
or inexistent amounts of transverse reinforcement should be carried out with
caution.
• It is expected that in shear-critical FRC structural applications, with longitudinal
and transverse reinforcement ratios over the minimum demands, satisfactory and
stable results should be obtained in the numerical simulations.
Acknowledgements. This study forms a part of the project BIA2016-78460-C3-1-R supported

by the State Research Agency of Spain.
References
1. Bresler, B., Scordelis, A.C.: Shear strength of reinforced concrete beams. J. Am. Concr. Inst.
60, 51–72 (1963)
2. Vecchio, F.J., Shim, W.: Experimental and analytical reexamination of classic concrete beam
tests. J. Struct. Eng. 130(3), 460–469 (2004)
3. Ortiz Navas, F., Navarro-Gregori, J., Leiva Herdocia, G., Serna, P., Cuenca, E.: An
experimental study on the shear behaviour of reinforced concrete beams with macro-synthetic
fibres. Constr. Build. Mater. 169, 888–899 (2018)
4. European Committee for Standardization. EN 14651, Test method for metallic fibres concrete.
Measuring the flexural tensile strength (2005)
5. European Committee for Standardization-International Organization for Standardization. EN-
ISO 6892-1-2009: Metallic materials-Tensile testing Part 1: Method of test at room
temperature (2009)
6. European Committee for Standardization. EN 12390-3: Testing hardened concrete. Part 3:
Compressive strength of test specimens (2009)
7. European Committee for Standardization. EN 12390-13: Testing hardened concrete. Part 13:
Determination of secant modulus of elasticity in compression (2014)
8. International Federation for Structural Concrete (fib). Model Code 2010, final drafts, vol.
1 and 2, Wilhelm Ernst and Sohn (2013)
9. Rossi, P., Daviau-Desnoyers, D., Tailhan, J.L.: Analysis of cracking in steel fibre reinforced
concrete (SFRC) structures in bending using probabilistic modelling. Struct. Concr. 16(3),
381–388 (2015)
Experimental Analysis of Crack Development
of an UHPC Wall Element Under Shear
Loading
V. Příbramský1(&), M. Kopálová2, and L. Dlouhý2

1
Department of Concrete and Masonry Structures, Faculty of Civil Engineering,
Czech Technical University in Prague, Thákurova 7/2077,
166 29 Praha 6 Prague, Czech Republic
v.pribramsky@scia.net
2
SCIA CZ s.r.o, Evropská 2591/33d, 160 00 Praha 6 Prague, Czech Republic
Abstract. UHPC (Ultra High-Performance Concrete) is an innovative material

that enables design of lightweight and structurally optimized, long-lasting
structures. In this paper, an experimental analysis of precast webs of a “butterfly
web” box-girder bridge on scaled-down specimens is presented. Pretensioned
beam specimens were analysed in 2 variants–with continuous web and with
lightened web. Based on the experimental results both variants are compared
and numerical and material models suitable for UHPC modelling in software
SCIA Engineer are presented. In SCIA Engineer the modified Mazarz material
damage model is implemented which is applicable for material with residual
strength typical for FRC and UHPFRC.
Keywords: UHPC Lightened beam FRC shear resistance Structural

optimization
1 Introduction
This paper describes an experimental analysis of initiation and propagation of cracks in

webs of a bridge with a box cross-section. The web is composed of thin UHPC precast
walls rigidly connected to the bottom and top slab of the bridge cross-section. The
shape of these UHPC wall elements was optimized in order to achieve the most
favourable stress distribution. The optimal shape was determined from optimisation
based on the principal tensile and compressive stress distribution with distinctive shape
of the tensile and compressive diagonal, which are caused by shear force transfer
between the top and bottom slab. The web of the box cross-section is lightened and
suitable for prefabrication so these elements may be used on bridges over 100 m span
(concept shown on Fig. 1). The design of lightened web elements is based on concepts
of so called “butterfly web” bridges [1].
Precast panels connected with top and bottom slab by composite action do not
transfer shear force as a continuous web would and their behaviour is well differen-
tiated with tensile and compressive diagonals and is closer to a Warren truss system [2].
Based on the research in this paper, the viable description of the behaviour is the
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_61
682 V. Příbramský et al.
Fig. 1. “Butterfly web bridge” concept visualisation.
analogy of a Vierendeel beam, where discrete web elements act as vertical members
with rigid connection to the top and bottom slab (flange) in the longitudinal direction.
Vierendeel frames are more appropriate to describe the behaviour of UHPC webs, due
to different UHPC properties in plain tension and tension due to bending.
2 Analysis Methods Implemented in SCIA Engineer

for Uhpfrc Analysis
Physical non-linear analysis represents a very powerful tool for analysing any kind of
structure in civil engineering created not only from UHPFRC but also from other
materials. Generally, there could be significant differences of results compared to a
linear analysis, especially in case of hyper-static structures. In case of linear analysis,
only the E-modulus and Poisson ratio of the material are considered for the preparation
of the stiffness matrix. There is no stress redistribution based on increasing the strain in
the component. On the other side the non-linear analysis provides a stress redistribution
and increase of the bearing capacity after reaching the ultimate strain until the collapse
mechanism. This mechanism is based on different values of stress and strain, dependent
on a predefined non-linear stress-strain relationship in the material diagram.
Two approaches are usually used for considering fibres in non-linear analysis. The
first one uses the fibre and concrete matrix independently. The second and more
common one considers the steel fibre directly in the behaviour of the concrete material.
The second, more exact option was selected in this paper.
A typical shape of the stress-strain diagram is best described with a parabolic
behaviour in compression in the same way as for standard concrete. The tensile
behaviour displays a narrow peak expressed by the mean tensile strength. From the
point of strain, this peak works very well as a crack localizer. After the crack formation,
the toughness of steel reinforced fibre concrete allows to keep a certain level of tensile
stress with increasing of the strain up to failure.
Experimental Analysis of Crack Development 683
2.1 Modified Mazarz Material Damage Model

The nonlinear calculation is based on a very efficient damage material model called the
Mazars model [3]. This material model is very well applicable for a material diagram
with peaks and descending stress-strain diagrams typical for steel fibre concrete. The
combination of elasticity and damage behaviour is combined in this material model.
Moreover, the damage description was initially considered as isotropic and directly
affecting the stiffness matrix. The original Mazars model was updated to the “modified
Mazars model” based on [4] which has been derived from [6]. This leads to a very
simple anisotropic damage model better respecting different behaviours of the fibre
concrete in tension and compression and mainly to the decomposition of the stress
tensor (r) to the tension (rt ) and compression (rc ) parts
r ¼ r t þ rc ð1Þ
based on principal stresses and using 2nd order tensors ei .
X
3
r¼ r i ei ei ð2Þ
i¼1
It is suitable to define 4th order projection tensors for tension Pt and compression Pc ,
rt ¼ Pt : r
ð3Þ
rc ¼ Pc : r
where these tensors are obtained as
X
dim
Pt ¼ hri iðei ei Þðei ei Þ
i¼1 ð4Þ
P ¼ dik djl ei ej ek el P
c t
The item hi is the MacAuley bracket in the formula above. Additionally, item
“dim” corresponds to 2D or 3D dimensional problems.
The resultant stress tensor is based on the so called “effective damage parameter” in
tension (d t ) and in compression (d c ) which determines changes of the stiffness
depending on the elastic estimation of stress (rtrial ) from the loading.
r ¼ð1 d t Þ rt þ ð1 d c Þ rc ¼ ½ð1 d t Þ Pt þ ð1 d c Þ Pc : rtrial ð5Þ
This elastic estimation of stress can be expressed using the constitutive tensor
(C) as follows:
rtrial ¼ C : e ð6Þ
The damage parameters are calculated based on the equivalent Mazars strains (et ;
ec .) which help for determining of actual values of stress from the stress-strain diagram
of the material.
rðet Þ rðec Þ
dt ¼ 1 ; d c ¼ 1 trial ð7Þ
rtrial ðet Þ r ðec Þ
As the stress-strain diagram of UHPC typically has a descending branch of stress-

strain in tension, it is not possible to use the tangential constitutive tensor (C.) but it is
recommended to use the secant one (Cs ) to fulfil its positive definition which is finally
calculated as below.
Cs ¼ ½ð1 d t Þ Pt þ ð1 d c Þ Pc : C; ð8Þ
Additionally, the effect of cracks must be considered during application. Here an

analogy with the thermodynamic variable is applied for the two main damage states
which are cracking of concrete in tension and crushing in compression. In case of
plotting the surface failure of the steel fibre reinforced concrete the very well-known
curve described by Kupfer [5] for biaxial loading is obtained which is also typical for
regular reinforced concrete. The standard Newton-Raphson method is used for solving
of this physical non-linear problem.
3 Experimental Analysis of UHPC Web Elements
3.1 Description of Tested UHPC Specimens

For experimental analysis of slender structural members of UHPC under shear loading
2 types of beam specimens were designed – one with a full, continuous web and one
with a longitudinally lightened web. These beams have an I-cross-section with suffi-
ciently designed flanges with longitudinal pretensioned tendons in order to mitigate
effects of bending. The topology of these beams is apparent on Fig. 2 below. The aim
of this experimental setup is to verify the behaviour of lightened specimens and via-
bility of the application of similar larger scale precast and pretensioned web elements in
greater magnitude and on bridges of span over 100 m.
The dimensions of the beams were chosen with respect to prefabrication, manip-
ulation and transportation possibilities. The beams were 2,3 m long and 0,39 m high.
The bottom flange was prestressed with 2 straight tendons with initial prestressing
stress 900 MPa.
In the structural analysis software SCIA Engineer the specimen was modelled with
a loading mechanism as half (from support to midspan) with corresponding boundary
conditions in order to reduce the analysis time. Furthermore, the FEM mesh was refined
in areas where cracks were expected to develop.
Fig. 2. Shape of UHPC specimens.
3.2 UHPC Beam Specimens and UHPC Mixture Characteristics

In total, 6 beam specimens were casted and loaded in 4-point bending tests, 3 speci-
mens with continuous web labelled P1, P2, P3 and 3 beams with lightened web labelled
V1, V2, V3. The first beam from each set (P1, V1) was loaded till failure. On the other
2 specimens in each set first a cyclical loading was applied and after that the specimens
were loaded till failure. The compressive strength of the UHPC mixture was measured
on cubes 100 100 100 mm and the average compressive strength was
157,1 MPa. The tensile strength in bending was measured on prismatic specimens of
40x40x160 mm and was in average 28,0 MPa. The modulus of elasticity was measured
on cylindrical specimens 300 mm high with diameter 150 mm and the average mod-
ulus of elasticity was 51,6 GPa.
4 Results and Experiment Evaluation
In this section the results are presented of the experiment for both types of beam
specimens. The results are presented in the form of force – deflection diagrams and are
approximated by a numerical model in SCIA Engineer 18 where a non-linear material
model with damage was used. Basic characteristics of the material model were set
according to experimental results on cubes (strength in compression), prisms (tensile
strength in bending) and cylinders (modulus of elasticity). These specimens had the
same age in the time of testing as the bigger beam specimens. For both beam specimens
the numerical model took into account the effects of longitudinal prestressing with
estimated short-term and long-term losses of 15%. The age of the specimens at the time
of testing was in average 90 days.
4.1 Results on Beam with Full Web

The beams with continuous web were tested in 4-point bending in 2 separate scenarios.
In the first scenario the first beam (P1) was loaded by a continuous increase of dis-
placement till failure. In the second scenario the 2nd and 3rd beams (P2 and P3) were
loaded with a cyclic loading pattern. Five loading cycles of approximately 0–100 kN
were applied and the beams were loaded till failure. The force – displacement diagram
of this experiment is presented on Fig. 3 below.
0
-20
-40
-60
Applied force [kN]
-80
-100
-120
-140
-160
-180
-200
-220
-240
-260
-280
-300
-320
0 2 4 6 8 10 12 14
P1 Mid span deflection [mm]
P1_2
P2 - Damaged by microcracks
P3 - Damaged by microcracks
SCIA - UHPC Material model with microcracks
SCIA - UHPC Material model
Fig. 3. Force–displacement diagram: beams with continuous webs
From the force – displacement diagram is apparent the softening of the beams with
applied cyclical loading to the magnitude of the applied load of approximately 150 kN
when compared to the beam which was loaded without the cyclic scenario. Further-
more, in the specimens with applied cyclical loading the first visible cracks appeared at
much lower magnitude of the acting force. Cause of this behaviour is the initiation of
microcracks when cyclical loading was applied. From the Fig. 3 it is apparent, that this
effect has a significant effect on both the mean and residual tensile strength of the
UHPC.
For the numerical model of beams with continuous webs were used material
models of UHPC with characteristics shown below in Fig. 4. Magnitudes of mean and
residual stresses were obtained iteratively in order to achieve a behaviour consistent
with the measured force – displacement diagram in Fig. 3. Magnitudes of residual
strain in the material model are dependent on the FEM mesh size. In this case the size
of the mesh in areas of crack initiation was set to 10 mm and the crack width was
limited to 5 mm. This corresponds to the upper limit of residual strain for the mesh
elements to be 50% (for better readability of values in Fig. 4 only the section below 3%
strain is displayed).
30
UHPC beam with continuous
25 web
Stress σ [MPa]
UHPC with continuous web
20 with microcracks
15
10
5
0
0 100 200 300
Tensile strain ε [1e-4]
Fig. 4. Stress–strain diagrams of UHPC–tension
The type of failure and shape of the developed cracks is apparent on Fig. 5 below.
Photography taken during the experiment is compared to the shape of the developed
crack in program SCIA Engineer 18.
Fig. 5. Analysis of crack shape: photo (a); development of macrocracks (b) in the numerical
model
4.2 Results on Beam with Lightened Web

The beams with lightened web were tested in 4-point bending in 2 separate scenarios the
same way as the beams with a continuous web. In the second scenario a cyclic loading
pattern was applied on specimens V2 and V3. Five loading cycles of approximately 0–
40 kN were applied and the beams were loaded till failure. The force – displacement
diagram of this experiment is presented on Fig. 6 below.
From the force – displacement diagram the softening of the beams with applied
cyclical loading to magnitude of applied load of approximately 50 kN is apparent when
0
-10
-20
-30
-40
-50
Applied force [kN]
-60
-70
-80
-90
-100
-110
-120
-130
-140
-150
-160
-170
-180
0 2 4 6 8 10 12 14
Mid span deflection [mm]
V1
V2 - Damaged by microcracks
V3 - Damaged by microcracks
SCIA - UHPC Material model with microcracks
SCIA - UHPC Material model
Fig. 6. Force–displacement diagram: beams with lightened webs
compared to the beam which was loaded without the cyclic scenario. Furthermore, in
the specimens with applied cyclical loading the first visible cracks appeared at a
significantly lower magnitude of the acting force.
For the specimen which was not cyclically loaded a much better performance can
be seen in comparison with the beam which was subjected to cyclical loading. The
increase of the load bearing capacity between the scenarios is much greater than in the
case of specimens with a continuous web. This effect is very important and is influ-
enced by the fact, that tensile strength of UHPC in tension is several times greater when
subjected to bending rather than plain tension. The tension strength in bending was
experimentally verified on prismatic specimens 40 40 160 mm, which represents
cross-section with comparable dimensions as the thickness of the web of the beam
specimens to mitigate the size effect. For the numerical model of beams with lightened
webs material models of UHPC were used with characteristics as shown below in
Fig. 7.
30
UHPC lightened beam
25
Stress σ [MPa]
UHPC lightened beam with
20 microcracks
15
10
5
0
0 100 200 300
Tensile strainε [1e-4]
Fig. 7. Stress–strain diagrams of UHPC–tension
The type of failure and shape of the developed cracks is apparent on Fig. 8 below
and again shows a good correlation between the observed crack distribution and the
numerical model.
Fig. 8. Analysis of crack shape: photo (a); development (b) of macrocracks in the numerical
model
4.3 Comparison of Solid and Lightened Web Beams

For an objective comparison of beams with continuous and lightened webs which were
loaded till failure (loading scenario 1) it is prudent to determine the values of the
maximal applied loading on such lightened specimens, that would require an identical
amount of UHPC as the specimen with a continuous web. Such modified specimen
would have a lightened web with thickness 48,4 mm increased by 38%. On Fig. 9
below the comparison is shown of the magnitude of the applied forces on the level of
macrocrack initiation and with a maximal applied load.
-450
Continuous web thk. 35mm
Lightened web thk. 35mm
Applied force [kN]

-350 -310.8
Lightened web thk. 48.4mm
-245.9
-250
-182.5 -167.4 -177.7
-150 -121
-50
Applied force when Maximal applied force on

macrocrack appear the specimen
Fig. 9. Comparison of specimen behaviour
The recorded effect of reduction of both the mean and residual strength of UHPC
beam specimens loaded with cyclical loading when compared to a beam loaded straight
till failure without cyclical loading is more severe in the case of the lightened specimen.
This effect is caused by a higher magnitude of tensile stresses in the specimens when
the cyclical loading is applied. In localised areas of the lightened web specimen the
magnitudes of tensile stresses rise to 16,5 MPa when compared to the tensile stresses in
the specimen with a continuous beam 9,0 MPa. Localised tensile areas in the lightened
beam with a greater magnitude lead to higher initiation of microcracks and more severe
damage of the specimens, before any macrocracks are visible. On Fig. 10 below are
shown the distributions of the principal tensile stresses from the combination of self-
weight, prestressing and the amplitude of the applied cyclical load.
Fig. 10. Maximal tensile stresses in web of (a) lightened beam and (b) beam with continuous
web during cycled loading
5 Discussion
The comparison of a beam with lightened and continuous web is made on specimens
with the same amount of material required. The beam with continuous web has a web
thickness of 35 mm (as the real specimen) and the lightened beam has a recalculated
web thickness of 48,4 mm. The first visible macrocracks appear in the case of a beam
with lightened web on a load level 15% lower and maximal loading 22% lower than the
beam with continuous web. Despite the fact that the magnitudes are lower, there are the
following important aspects:
• The lightened beam shape may yet be further optimized (increase thickness on
edges of the wall segments, where macrocracks first appear).
• Possibility of prefabrication of separate web segments to ensure superior quality.
• Due to clear and consistent shear force transfer by the web segments, these seg-
ments may be provided with prestressing tendons in the direction of the tensile
diagonal to mitigate tensile stresses.
• Contradictory to analysis performed prior the experiments and based on available
studies of butterfly-web bridges [1,2] where web segment behaviour was described
as an approximation with tensile and compressive diagonal. More suitable seems
the analogy with a Vierendeel beam. Wall segments are thus approximated as frame
members and their action is bending in longitudinal direction. Given the excellent
UHPC properties in tensile strength in bending, this behaviour is most convenient.
6 Conclusions
In this paper the results were presented of an experimental analysis of UHPC beams
with lightened webs and numerical verification in the program SCIA Engineer. These
specimens demonstrate on a smaller scale the behaviour of precast web segments of a
bridge with box cross-section, which are connected to the top and bottom monolithic
slabs of the cross-section by composite action. Discrete behaviour of the web segments
was compared to the continuous behaviour of a beam with solid web with constant
thickness. The conclusion of the analysis is viability of application of UHPC precast
webs especially due to the excellent properties of the UHPC in tension under bending
action. When these web segments are provided with efficient prestressing to eliminate
tensile stresses on their edges, the precast web segments provide superior performance
compared to beams with a continuous web.
Acknowledgements. This work was supported by SGS grant of CTU in Prague, Czech
Republic, grant No. SGS19/036/OHK1/1T/11.
References
1. Kasuga, K., Nagamoto, N., Kata, K., Asai, H.: Study of a bridge with a new structural system
using ultra high strength fiber reinforced concrete. In: Proceedings of 3nd FIB Congress,
Washington (2010)
2. Kata, K., Ashizuka, K., Miyamoto, K., Nakatsumi, K.: Design and construction of butterfly
web bridge. In: Third International Conference on Sustainable Construction Materials and
Technologies, Kyoto (2013)
3. Mazars, J.: A description of micro and macroscale damage of concrete structure. Eng. Fract.
Mech. 25, 729–737 (1986)
4. Němec, I.: Trcala M. Rek V; Nelineární mechanika, Vutium, Brno (2018)
5. Kupfer, H., Hilsdorf, H.K., Rusch, H.: Behaviour of concrete under biaxial stresses.
ACI J. 66, 656–666 (1969)
6. Wu, J.Y., Li, J., Faria, R.: An energy release rate-based plastic-damage model for concrete.
Int. J. Solids Struct. 43(3–4), 583–612 (2006)
Assessment of the Shear Behaviour of Fibre
Reinforced Concrete Through Numerical
Modelling of Shear-Friction Theory
Álvaro Picazo1, Marcos G. Alberti2, Alejandro Enfedaque2,

and Jaime C. Gálvez2(&)
1
Departamento de Tecnología de la Edificación, E.T.S de Edificación,
Universidad Politécnica de Madrid, Av. Juan de Herrera, 6, 28040 Madrid, Spain
2
Departamento de Ingeniería Civil: Construcción,
E.T.S de Ingenieros de Caminos, Canales y Puertos,
Universidad Politécnica de Madrid,
C/ Profesor Aranguren, s/n, 28040 Madrid, Spain
jaime.galvez@upm.es
Abstract. Few published studies have dealt with the mechanisms of rein-
forcement of fibre reinforced concrete (FRC) when it is subjected to shear
stresses. Possibilities for noticeable improvements and steel rebars reduction
have been reported although additional research is needed in order to determine
and quantify the shear resisting mechanisms of FRC.
The significance of this research relies on the use of several types of FRC,
previously characterized under flexural tests, to perform push-off tests. The tests
were complemented with digital image correlation (DIC) techniques in order to
obtain displacements and crack openings that will permit the development of the
shear-friction theory adapted to FRC.
The experimental campaign was performed with specimens of dimensions
270 150 150 mm3. The specimens were manufactured with six types of
concrete: two moderate-strength concrete matrixes with 6 and 7.5 kg/m3 of
polyolefin fibres, two medium-strength concrete (vibrated and self-compacted)
reinforced with 10 kg/m3 of polyolefin fibres and two steel fibre reinforced
concrete types with 50 and 70 kg/m3 of steel hooked fibres. The results showed
that FRC follows an analogous behaviour compared with reinforced concrete
and the shear-friction theory. The relations between the displacements and the
crack openings were achieved as well as deformation maps in the push-off tests.
Keywords: Shear-friction theory Fibre reinforced concrete Steel fibres

Polyolefin fibres Digital image correlation
1 Introduction
The most widely used reinforcement to improve concrete tensile strengths is the use of
steel rebars, commonly named reinforced concrete. Another reinforcement technique
involves adding fibres in the concrete mass. This reinforcement is known as fibre
reinforced concrete (FRC). The contribution of the fibres can be considered in the
structural design, allowing the replacing of steel rebars for elements subjected to
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_62
694 Á. Picazo et al.
bending stresses or shear stresses [1, 2]. Such substitution may imply significant
reduction the labour of the structure.
Structural use of FRC has mainly been developed through the use of steel fibres,
which involved extensive research on steel fibre reinforced concrete (SFRC) [3]. Sub-
sequently, other fibres of different materials were used: aramid, carbon, glass and
polyethylene. In recent years, the structural use of polyolefin fibres has also shown
structural capacity becoming an attractive alternative. These polymer fibres have char-
acteristics that may improve certain ones of steel fibres such as the absence of corrosion,
lower final weight of the structure, lower cost of maintenance of the machinery [4]. This
material is usually known as polyolefin fibre reinforced concrete (PFRC).
All fibre types used for concrete reinforcement can also help cracking at an early age
of concrete [5]. In addition, SFRC and PFRC have shown to meet the requirements of
current concrete structural codes [6, 7]. Such requirements are based on residual tensile
strengths obtained by three or four point bending tensile tests, such as that of EN-14651
[8]. However, this bending strength may not be suitable to be considered for other type
of stresses such as elements subjected to shear as some research has shown [9].
Published research about the shear behaviour of FRC has mainly been based on two
forms of study: on beams subjected to bending and shear [9], and on specimens
subjected to direct shear stress [10–12]. Nowadays, it is still important to address this
research topic in order to understand the resistant mechanisms according to the shear-
friction theory [13, 14]: the aggregate interlock and the dowel action of the fibres or
stirrups that new the crack. In any event, fibres mechanisms of reinforcement signifi-
cantly improve the behaviour of the material subjected to bending and shear stresses
[15–17]. Thus, the cracking patterns should be investigated, in order to provide
information on the different cracking modes of the FRC.
With this background, the objective of this study was focused on analysing the
different cracking patterns that occur in the FRC subjected to shear stress and obtaining
the displacements that occur in the resistant section. Experimental results were also
sought to analyse the pre and post-crack phase of FRC. To achieve these objectives,
push-off specimens obtained from residual halves of specimens tested prior to bending
[12] were tested. Moreover, a digital image correlation (DIC) system [18] was used to
monitor the faces of the specimens supplying cracking and displacement data. This
information can be relevant for future development of shear-friction models.
The experimental campaign was based on push-off tests. The specimens for these tests
can be defined as a Z-shaped prism on which a vertical load is applied. The load
produces a shear stress on the ligament section [19] as shown in Fig. 1(a). These
specimens were obtained from the residual halves of tensile bending tests samples. The
extreme side areas that may have been affected were sawn, leaving a central area
formed by a prism of approximately 270 150 150 mm3, as shown in Fig. 1(b). In
this prismatic specimen two opposite notches 75 mm deep and 9 mm high were made,
leaving a resistant area in the central section of 150 75 mm2. The two cantilevers of
Assessment of the Shear Behaviour 695
the specimens had to be externally reinforced by a SikaWrap® carbon fibre sheet

bonded with epoxy resin so that the bending stresses, previously analysed with a finite
element model, did not exceed the material strength.
Fig. 1. (a) Push-off specimen and (b) Push-off specimens obtained from the remaining halves of
a three-point bending test.
The specimens were manufactured with six types of concrete: two moderate-
strength concrete matrix with 6 and 7.5 kg/m3 of polyolefin fibres (HVBP6 and
HVBP7.5), two medium-strength concrete (vibrated and self-compacting concrete)
reinforced with 10 kg/m3 of polyolefin fibres (HVP10 and HACP10M) and two steel
fibre reinforce concrete types with 50 and 70 kg/m3 of steel hooked fibres (HACA50
and HACA70). The main mechanical properties of these concrete types were known, as
they were analysed in previous studies [20, 21]. These characteristics are shown in
Table 1, where is shown that all the FRCs analysed reached the minimum conditions
Table 1. Residual tensile flexural strength and compression strength.

HACA50 HACA70 HVBP6 HVBP7.5 HVP10 HACP10M
fLOP (MPa) 7.59 11.32 2.76 2.57 4.21 5.22
fR1 (MPa) 7.22 11.09 1.43 1.70 1.98 2.41
95.13%fLOP 97.97%fLOP 51.81%fLOP 66.15%fLOP 47.03%fLOP 46.17%fLOP
fR3 (MPa) 5.21 9.43 1.75 2.15 2.87 3.87
68.64%fLOP 83.30%fLOP 63.41%fLOP 83.66%fLOP 68.17%fLOP 74.14%fLOP
fcm (MPa) 62.2 58.8 21.7 20.1 39.7 51.5
net by the regulations for structural applications, according to Eqs. (1) and (2). The mix
proportions are shown in Table 2.
fR1 [ 40%fLOP ð1Þ
fR3 [ 20%fLOP ð2Þ
In order to finish with the preparation of the specimens it was necessary to paint a
random monochrome pattern on the front and back face of each sample. DIC, as apears
in Fig. 2 enables, in the calculation of relative distances between two points on the
specimen surface at different times of the test. This distance, indicated in pixels, must
be transformed into units of length before the results can be analysed. This technique
does not modify any condition of the test, as there is no physical contact between the
system and the specimen [22]. This system has previously been used in other exper-
imental campaigns [18, 23].
The tests were carried out by using an Instron® (8803) machine with a maximum load
capacity of 0.5 MN, being controlled by setting the actuator position speed of 0.001 mm
per second. Two five Mpx high-definition cameras were used for DIC in order to record
the front and back sides of the specimens. In order to measure the opening or closing of the
notches during the tests, two linear variable differential transformer (LVDT) were
available to detect any rotation of the specimen during the test.
The specimens were carefully placed in the test machine in order to concentrate the
load on the ligament section. Two steel bars of dimensions 200 10 10 mm3 were
laid on the top and bottom of the specimens and aligned, by laser, with the end of the
specimen notches, as shown in Fig. 1(a).
Table 2. Concrete mix dosage (kg/m3).

Cement 52.5 425 425 – – 375 375
Cement 32.5 – – 312 312 – –
Gravel 492 486 519 519 450 320
Grit – – 198 198 300 213
Sand 947 947 875 875 916 991
Water 199 199 216 216 188 188
w/c 0.47 0.47 0.69 0.69 0.50 0.50
Limestone 210 210 – – 100 200
Superplasticizer (% 1.39 1.39 – – 0.82 1.25
cement)
Fibres PF48 – – 6 7.5 – –
Fibres PF60 – – – – 10 10
Fibres SF35 50 35 – – – –
Fibres SF50 – 35 – – – –
Compaction Self- Self- Vibrated Vibrated Vibrated Self-
weight weight weight
Fig. 2. Displacements on digital image correlation.
3 Results
As a result of the tests, time, displacement of the test machine, applied load and shear
slide of the ligament section could be obtained. In addition, DIC and the timing of the
tests were correlated so the load and displacements were synchronised with DIC. The
comparison of results among different concrete types could be performed, having the
load and ligament section values, so the data could be compared in terms of tangential
stress through use of Eq. (3).
s ¼ Load=Section ð3Þ
Table 3 shows the residual strengths obtained in the tests. The values are at least the
average values of two specimens of each type of concrete. Table 3 shows the residual
stresses for shear displacements of 0.5 and 2.5 mm and the maximum strength. It is
worth noting that the specimens had residual shear load-bearing capacities up to shear
displacements of 7 mm, which could provide a significant safety factor to structures.
Table 3. Average maximum shear stress (MPa) and residual strength at 0.5 mm and 2.5 mm
shear displacement.
smax 12.16 17.37 5.19 4.93 7.93 9.70
s at 0.5 mm 8.34 13.06 2.78 4.27 2.81 2.60
s at 2.5 mm 0.89 1.61 1.04 1.44 1.70 1.15
4 Digital Image Correlation Analysis

4.1 Shear-Friction Mechanism
The shear displacement and crack opening data were obtained based on the relative
position between two points in the image, at different times in the test. To obtain these
results, five DIC measurement elements were available in each ligament section, in
accordance to Fig. 2. For the analysis of the results, it was necessary to perform the
conversion from pixels (unit in which the DIC provides the data) and length units.
With the values obtained from displacement to shear and crack opening, different
curves were generated that showed the behaviour of FRC in the shear-friction face. In
these graphs, the behaviour of the SFRC and PFRC were similar and analogue to that
developed by Walraven [24] on precracked specimens, which concluded the rela-
tionship between the crack openings and the shear displacements, based on a “k”
coefficient, according to the Eq. (4), where “x” is crack opening and “D” is the shear
displacement.
x ¼ k D2=3 ð4Þ
Figure 3 shows the theoretical curve, obtained analytically according to Eq. (4) for
a reinforced concrete with traditional stirrups reinforcement, by setting the value of
k = 0.5. In addition, data for HACA50, HACP10M and HVBP6 has been included.
The similar behaviour of the three concrete types can be observed and it can be seen
that the PFRC shows a greater crack opening than the SFRC for the same shear
displacement. This behaviour is maintained throughout the development of the tests.
In addition, the theoretical crack opening [24] is always lower than that shown by
any of the FRC. This is because the specimens used in this study were not pre-cracked,
as those tested by Walraven. Another factor to be considered in the analysis of the
results is that in the theoretical case the stirrups were disposed to better support the
shear stress, perpendicular to the cracking section. Fibres, by definition, present a
random situation in the concrete matrix, so it is not possible to place them in the most
favourable position to withstand shear stress.
4.2 Crack Pattern

The mechanisms of shear fracture in FRC have been studied by means of DIC, being
able to obtain relevant data on the appearance and development of cracks generated in
the specimens during the tests. The frames were obtained with one image per second
rate. This speed allowed that each image to be associated with the correspondent load
and displacements through the test time. In this way the different cracking patterns were
obtained for each concrete.
When HVBP6 and HVBP7.5 specimens were studied, it turned out that the cracks
were generated at the tip of the upper or lower notch. The crack grew vertically
between the tips of the notches while small cracks were generated with a certain
inclination. Figure 4(a) shows the cracks of the HVBP6, seeing a crack at the time of
its appearance. Micro-cracks can also be seen in areas marked with red.
3.5
HVBP6
3.0 HACP10M
HACA50
2.5
Walraven k=0.5
w (mm)
2.0
1.5
1.0
0.5
0.0
0 1 2 3 4 5
D (mm)
Fig. 3. Crack opening vs shear slide.
(a) (b) (c)

Bending crack
Shear crack
Shear crack
Bending crack
Bending crack
Fig. 4. Image of the specimen cracking: (a) HVBP6 at moment of cracking, (b) HVP10 al the
end of the test and (c) HACA70 at moment of cracking.
In the case of HVP10 specimens, the cracks were generated in the central area of
the resistant section, along with bending cracks in the middle area of the notch. During
the development of the tests these bending cracks were reduced to be imperceptible at
the end of the tests. In all cases a single crack was not generated, but a multi-fissure of
inclined cracks were formed that subsequently joined together to form a crack that
connected the ends of the notch. Figure 4(b) shows a HVP10 specimen at the final of
the test.
The behaviour of both types of SFRC was similar. During the loading branch the
first crack occurred, at which point the stiffness of the specimens varies, leading to a
variation in the slope of the load-displacement curve. The test load continued to
increase up to the point of maximum load and strength, where significant development
of cracking occurred. Figure 4(c) shows the initial cracking of an HACA70 specimen
during the loading branch. The main crack grew to communicate the tips of both
notches. It is worth noting that the generation of bending cracks in the central area.
These cracks decreased to be only visible by DIC and were foreseeable according to a
finite element model made in a previous investigation [12].
5 FRC Pre and Post-Cracking Behaviour
In Fig. 5, it can be seen the behaviour of the six concrete types, with the data obtained
from the test machine. All concretes showed a first linear upstream branch, associated
with the strength of the unreinforced concrete matrix, except in the case of HACA
where the maximum load values were slightly higher than that of unreinforced con-
crete. HVBP6 and HVBP7.5 were matrices with moderate compressive strength, so
they showed the lowest strength values at the limit of proportionality.
20
Average Shear (MPa)
15
HVBP6
HVBP7.5
HVP10
10 HACP10M
HACA50
HACA70
5
0
0.0 0.5 1.0 1.5 2.0 2.5
Shear slide (mm)
Fig. 5. Experimental curves average shear vs shear slide up to 2.5 mm.
In the first part of the curves, HACA showed greater strengths than PFRC, as for
small crack openings they maintain superior behaviour. The discharge branch in
HACA was less abrupt than that of HACP10M and HVP10. It is worth noting that in
this discharge branch there was a significant behaviour of HVBP, which showed good
post-cracking behaviour by the action of the fibres and a smooth curve.
From 1 mm of shear displacement, the behaviour of the six concrete types was
remarkably similar. A residual shear stress was maintained, ensuring the ductile
behaviour of the concrete element. From 2 mm of shear displacement, PFRC showed
residual strengths greater than HACA50. This was consistent with the superior beha-
viour that occurred in PFRC subjected to flexural tests for large crack openings in
comparison with SFRC [25].
6 Conclusions
• Push-off tests have been complemented by DIC that have led to relevant results
regarding the behaviour of FRC subjected to shear stress.
• Images of the different cracking modes and patterns of the specimens were
obtained. DIC analysis allowed the visualisation of the cracks before it was visible
to the naked eye and it was possible to associate that instant of the test with the
correspondent load and displacement value.
• The behaviour of FRC was compared to the shear-friction theory developed by
Walraven for pre-cracked specimens. The application of the theoretical formulation
on FRC was adequate, so the theory was considered applicable to the case of FRC.
• FRC showed ductile fracture behaviour, resulting in significant displacement-to-
shear values without the specimen collapse.
• Push-off tests provided relevant information regarding Mode II of fracture. This
data, together with those obtained in bending tensile tests, made possible to analyse
the same fracture material in Mode II and I.
Acknowledgements. The authors gratefully acknowledge the financial support provided by

Ministry of Science and Competitiveness of Spain by means of the Research Fund Project
PID2019-108978RB-C31.
References
1. Li, V.C., Ward, R., Hamza, A.M.: Steel and synthetic fibers as shear reinforcement (1992)
2. Kwak, Y.-K., Eberhard, M.O., Kim, W.-S., Kim, J.: Shear strength of steel fiber-reinforced
concrete beams without stirrups. ACI Struct. J. 99(4), 530–538 (2002)
3. Brandt, A.M.: Fibre reinforced cement-based (FRC) composites after over 40 years of
development in building and civil engineering. Compos. Struct. 86(1), 3–9 (2008)
4. Alberti, M.G., Enfedaque, A., Gálvez, J.C.: On the mechanical properties and fracture
behavior of polyolefin fiber-reinforced self-compacting concrete. Constr. Build. Mater. 55,
274–288 (2014)
5. Banthia, N., Gupta, R.: Influence of polypropylene fiber geometry on plastic shrinkage
cracking in concrete. Cem. Concr. Res. 36(7), 1263–1267 (2006)
6. International Federation for Structural Concrete (fib), The fib Model Code for Concrete
Structures 2010. Lausanne, Switzerland: International Federation for Structural Concrete
(2010)
7. Spanish Minister of Public Works, Spanish Structural Concrete Code EHE-08. Madrid,
Spain: Spanish Minister of Public Works (2008)
8. European Committee for Standardization, C: Test Method for Metallic Fiber Concrete.
Measuring the Flexural Tensile Strength (Limit of Proportionality (LOP), Residual),
EN14651:2007 + A1 (2007)
9. Cuenca, E., Echegaray-Oviedo, J., Serna, P.: Influence of concrete matrix and type of fiber
on the shear behavior of self-compacting fiber reinforced concrete beams. Compos. B Eng.
75, 135–147 (2015)
10. Soetens, T., Matthys, S.: Shear-stress transfer across a crack in steel fibre-reinforced
concrete. Cem. Concr. Compos. 82, 1–13 (2017)
11. Zeranka, S.: Steel Fibre-Reinforced Concrete: Multi-Scale Characterisation Towards
Numerical Modelling. Stellenbosch University, Stellenbosch (2017)
12. Picazo, Á., Gálvez, J.C., Alberti, M.G., Enfedaque, A.: Assessment of the shear behaviour of
polyolefin fibre reinforced concrete and verification by means of digital image correlation.
Constr. Build. Mater. 181, 565–578 (2018). https://doi.org/10.1016/j.conbuildmat.2018.05.
235
13. Walraven, J.C., Stroband, J.: Shear friction in high-strength concrete. Spec. Publ. 149, 311–
330 (1994)
14. Taylor, H.P.: The fundamental behavior of reinforced concrete beams in bending and shear.
Spec. Publ. 42, 43–78 (1974)
15. Hamrat, M., Boulekbache, B., Chemrouk, M., Amziane, S.: Flexural cracking behavior of
normal strength, high strength and high strength fiber concrete beams, using digital image
correlation technique. Constr. Build. Mater. 106, 678–692 (2016)
16. Altoubat, S., Yazdanbakhsh, A., Rieder, K.-A.: Shear behavior of macro-synthetic fiber-
reinforced concrete beams without stirrups. ACI Mater. J. 106(4), 381 (2009)
17. Krassowska, J., Kosior-Kazberuk, M., Berkowski, P.: Shear behavior of two-span fiber
reinforced concrete beams. Arch. Civ. Mech. Eng. 19(4), 1442–1457 (2019). https://doi.org/
10.1016/j.acme.2019.09.005
18. De Wilder, K., Lava, P., Debruyne, D., Wang, Y., De Roeck, G., Vandewalle, L.:
Experimental investigation on the shear capacity of prestressed concrete beams using digital
image correlation. Eng. Struct. 82, 82–92 (2015)
19. Echegaray, J.: Upgrading the push-off test to analyze the contribution of steel fiber on shear
transfer mechanisms (2014)
20. Alberti, M.G., Enfedaque, A., Gálvez, J.C.: Comparison between polyolefin fibre reinforced
vibrated conventional concrete and self-compacting concrete. Constr. Build. Mater. 85, 182–
194 (2015)
21. Rodríguez, I.: Diseño y caracterización de hormigones autocompactantes reforzados con
fibras de acero. TFM (2013)
22. Mirzazadeh, M.M., Green, M.F.: Fiber optic sensors and digital image correlation for
measuring deformations in reinforced concrete beams. J. Bridge Eng. 23(3), 04017144
(2018)
23. Alam, S., Loukili, A., Grondin, F., Rozière, E.: Use of the digital image correlation and
acoustic emission technique to study the effect of structural size on cracking of reinforced
concrete. Eng. Fract. Mech. 143, 17–31 (2015)
24. Walraven, J.C.: Aggregate interlock: a theoretical and experimental analysis (1980)
25. Alberti, M.G., Enfedaque, A., Gálvez, J.C.: Improving the reinforcement of polyolefin fiber
reinforced concrete for infrastructure applications. Fibers 3(4), 504–522 (2015). https://doi.
org/10.3390/fib3040504
Modeling the Compressive Behavior of Steel
Fiber Reinforced Concrete Under High Strain
Rate Loads
Honeyeh Ramezansefat(&), Mohammadali Rezazadeh,

Joaquim A. O. Barros, Isabel B. Valente, and Mohammad Bakhshi
Structural Division, Department of Civil Engineering, ISISE,

University of Minho, 4800-058 Guimarães, Portugal
Honeyrscivil@gmail.com
Abstract. Concrete is a strain-rate sensitive material and shows relatively low

ductility and energy dissipation capacity under high strain rate loads (HSRL)
such as blast and impact, representative of terrorist attacks and accidents.
Experimental research in the literature has evidenced that introducing steel
fibers, into the concrete mixtures can significantly improve the concrete
behavior under HSRL. Besides the experimental research, development of
design models is an important aspect to provide more confidence for engineers
to use SFRC in structural elements when subjected to HSRL. The existing
design codes (e.g. CEB-FIP Model Code 1990 and fib Model Code 2010)
propose models for the prediction of the strengths of concrete under different
HSRL, but they are only function of strain rate. In this regard, the current paper
deals with the improvement of design models in the fib Model Code 2010 for
the prediction of the compressive behavior of SFRC by considering the effects
of the important parameters such as volume fraction, aspect ratio and tensile
strength of steel fibers, and concrete compressive strength, besides the strain rate
effect. The developed artificial neural network mathematical model is calibrated
and its predictive performance is assessed using a database collected from the
existing compressive impact tests results on SFRC specimens.
Keywords: Steel fiber reinforced concrete Design model Drop-weight

impact test High strain rate load
1 Introduction
Concrete is a strain-rate sensitive material with low ductility and energy dissipation
capacity under high strain rate loads such as blast and impact, representative of terrorist
attacks and accidents. The experimental studies in the literature evidence that intro-
ducing various types of fibers, especially steel fibers, into the concrete mixtures can
significantly improve the concrete behavior under high strain rate loads [1, 2]. In this
context, SFRC has high impact resistance and energy dissipation capacity due to the
reinforcement mechanisms provided by fibers bridging the crack surfaces. These
mechanism, mainly those of fiber pull-out and snubbing effect at the fiber exit point,
limit crack propagation and enhances the energy dissipation capacity.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_63
704 H. Ramezansefat et al.
The impact behavior of steel fiber reinforced concrete (SFRC) composite materials
under compressive drop-weight tests is complex due to the nonlinear relationship
between the impact force and other variables. In this regard, some efforts have been
spent in order to establish an analytical model for the impact behavior of SFRC under
impact tests [2]. In this area, proposing a design formulation to accurately predict the
behavior of SFRC composite materials under impact loads considering the influence of
effective parameters (i.e. volume fraction of steel fibers (vf ), aspect ratio of steel fibers
(L=D), tensile strength of steel fibers (fts ), and concrete compressive strength (fcs ), and
shape of fibers is an issue that needs to be addressed. Moreover, there are still many
other issues, such as sophisticated equipment to accurately measure the impact forces in
these test types, deep analysis of failure mechanism, stress propagation in specimens,
effect of the distribution of inertial forces in specimens and inertial force measurement.
Meanwhile, such type of problem can be solved by some mathematical methodologies
in the field of machine learning, for example, artificial neural network (ANN). ANNs
have been used to solve some sophisticated problems in the field of civil engineering
and provided excellent results [3].
In this regard, the present study focuses on proposing an analytical model with
design framework for predicting the compressive behavior of SFRC composite mate-
rials under high strain loads using ANN method and considering the effective
parameters.
2 Architecture of the ANN
A neural network is a concept born due to the scientific interest about artificial intel-
ligence (AI) during in the middle of 1950s. Artificial neural networks are inspired by
the architecture of the human central nervous system which consists in a large number
of cells (neurons) working in parallel in order to facilitate decision-making in the most
rapid way [4]. These simple units are connected each other by electrical stimulus
named synapse. Similarly, the ANN offers the synaptic activity through a matrix of
weight (numeric values), which is typically updated by the like-human learning pro-
cess. The direct advantage of this approach is that neuron-computing devices do not
have to be programmed, but on the contrary, the random choice of initial weights
induces the learn-making from the process of adjusting the weights itself by reaching
the minimum error of prediction. ANN composes of an input layer including the
variables and an output layer. A defined number of layers are also added in between
these input and output layers, called hidden layers. Commonly only one hidden layer is
considered. However, the use of two or more hidden layers can sometimes significantly
improve the performance of ANN [3–5].
In this study, the proposed mathematical model was developed in Python pro-
gramming language to estimate dynamic increase factor (DIF) of SFRC composite
materials under high strain rate compressive load. A large database was collected from
the literature to use for the development of empirical model. This database includes the
experimental results of compressive drop weight tests, split Hopkinson pressure bar
(SHPB) tests and quasi-static tests on SFRC, which are the most current tests to assess
the impact compressive behavior of this type of material. The analysed database
Modeling the Compressive Behavior of Steel Fiber 705
included 80 SFRC samples tested under a strain rate load of less than 30 s−1 and 296
SFRC samples under a strain rate load of higher than 30 s−1, thus totally 376 SFRC
samples was included in the database and utilized in this study. The effective param-
eters adopted in the proposed model including volume fraction of steel fibers (vf ),
aspect ratio of steel fibers (L=D), tensile strength of steel fibers (fts ), and concrete
compressive strength (fcs ), were all reported for the specimens used in this database.
The database was randomly divided into two sub-databases for training (80%), and
test (20%). Following the data division and preprocessing, the architecture of the
hidden layer (including the number of hidden layers and the corresponding number of
hidden nodes) was established by the usual method of trial and error (learning and
training stages). Several trials were carried out, utilizing the whole database, in order to
find the optimum number of nodes in the hidden layer offering the highest coefficient of
correlation (R2) with experimental data [3].
In brief, the network was designed by using four neurons as input layer, two
neurons for hidden layer and one neuron as output layer. Transfer function in the
hidden layer was sigmoid and in the output layer was linear. After designing the
network and standardizing the value of the input parameters to improve the ANN-
mathematical model and make the training faster, the network would be trained Fig. 1.
Vf (%) 1
L/D (mm/mm) 2 N1
k
fts (MPa) 3 N2
Output
Hidden Layer
fcs(MPa) 4 Layer
Input
Layer
Fig. 1. Architecture of the comprehensive version of the proposed network.
3 Proposed ANN-Mathematical Model for Prediction

of Dynamic Increase Factor (DIF)
The strength of concrete at high strain rate loads can increase significantly. To char-
acterize the effects of strain rate on the compressive strengths of concrete, the dynamic
increase factor (DIF), i.e. the ratio of dynamic to static strength, is generally proposed as
a function of strain rates. In this regard, the CEB-FIP Model Code 1990 (MC1990) [6]
and CEB-FIP Model Code 2010 (MC2010) [7] are proposed design formulations for
estimating the DIF of concrete as a function of strain rate of loading. This section aims to
first evaluate the performance of the proposed formulations in MC1990 and MC2010 to
predict the compressive DIF of SFRC materials, and then, compare their predictive
performances with the corresponding performance of the developed ANN-mathematical

model to predict the compressive DIF of SFRC. This model was developed based on
modifying the proposed formulation in MC2010. In addition, a simplified closed form
formulation derived from the developed ANN-mathematical model is proposed in the
next section to predict the DIF of SFRC materials under compression.
Based on the MC1990, for a given strain rate, the compressive strength under high
rates of loading are estimated from Eqs. (1) and (2). These power functions were
established using appropriate underlying theory derived from thermodynamics and
fracture mechanics analysis [6].
fcd =fcs ¼ ðe_ c =_ec0 Þ1:026a for e_ c 30s1 ð1Þ
fcd =fcs ¼ cs ðe_ c =_ec0 Þ1=3 for e_ c [ 30s1

log cs ¼ 6:156 a 2
ð2Þ
a ¼ 1=ð5 þ 9fcs =fc0 Þ
where fcd is the dynamic compressive strength under high rates of loading, fcs is static
compressive strength, e_ c is compressive strain rate and e_ c0 ¼ 30:106 s1 , and
fc0 ¼ 10 MPa.
These formulations were proposed for two domains of strains, the first ranging from
low to intermediate (_ec 30s1 ) and the other from intermediate to high rates
(_ec [ 30s1 ). In a log (DIF) versus log(_e) the relationship is bilinear with a change in
slope around 30 s1 . On the other hand, the proposed DIF in MC2010 for the com-
pressive strength is given by:
fcd =fcs ¼ ðe_ c =_ec0 Þ0:014 for e_ c 30s1 ð3Þ
fcd =fcs ¼ 0:012ðe_ c =_ec0 Þ1=3 for e_ c [ 30s1 ð4Þ
where fcd is compressive strength under high rates of loading, fcs is the mean value of
compressive strength of concrete, e_ c is compressive strain rate in the range of 30
106 s1 to 300 s1 and e_ c0 ¼ 30:106 s1 .
These constitutive relations are valid for normal concrete, while for SFRC they
need to be updated due to the steel fiber effects in concrete, by considering the effective
parameters in the formulation. By inspiration of Eqs. (3) and (4), the alterations were
conducted on the power of Eq. (3) proposed by MC2010 (k1 in Eq. (5)) for the range of
e_ c 30s1 and on the constant coefficient of Eq. (4) proposed by MC2010 (k2 in
Eq. (6)) for the strain rates beyond the 30. In this regard, k1 and k2 parameters were
derived from experimental database using Eqs. (5) and (6), and were adopted as output
variable in ANN-mathematical model.
fcd =fcm ¼ ðe_ c =_ec0 Þk1 for e_ c 30s1

ð5Þ
k1 ¼ lnðfcd =fcm Þ=lnðe_ c =_ec0 Þ
fcd =fcm ¼ k2 ðe_ c =_ec0 Þ1=3 for e_ c [ 30s1

. ð6Þ
k2 ¼ ðfcd =fcm Þ ðe_ c =_ec0 Þ1=3
k1 and k2 parameters were considered as a function of volume fraction of steel fibers

(vf ), aspect ratio of steel fibers (L=D), tensile strength of steel fibers (fts ), and concrete
compressive strength (fcs ) for SFRC materials. Consequently, these four variables
(vf , L=D,fcs , fts ) were adopted in the input layer in the ANN-mathematical model. Two
neurons were assumed in the hidden layer, and k1 and k2 parameters were considered as
the output layer.
The different numbers of neurons in the hidden layer and the different transfer
functions for hidden and output layers were accepted in the ANN-mathematical model
to find an optimal network architecture offering highest coefficient of correlation (R2)
with experimental data. On the other side, since the main objective of this study is to
propose a closed form design formulation derived from the ANN-mathematical model
for the compressive DIF of SFRC materials, the neurons number adopted in the hidden
layer was minimized.
The performance of the proposed ANN-mathematical model was verified against
the experimental results for two range of strain rates e_ c 30s1 and e_ c [ 30s1 . The
plot of the experimental compressive DIF versus the corresponding ANN-mathematical
model predictions for the database is shown in Figs. 2 and 3, and also, compared with
the DIFs obtained from MC1990 and MC2010.
The best-fit line approximately aligns with the 45 benchmark proving a proper
correlation between the experimental results and the predictions of the proposed ANN-
mathematical model for the train and test data. The coefficient of correlation (R2 ) of
DIFs obtained from the ANN-mathematical model with the experimental results for
training and test data are, respectively, R2 ¼ 0:67 and 0.70 for strain rates e_ c 30s1 ,
and R2 ¼ 0:72 and 0.78 for strain rates e_ c [ 30s1 (Figs. 2 and 3).
However, the coefficient of correlation (R2 ) of DIFs obtained according to MC1990
and MC2010 using Eqs. (1) and (3) for strain rates e_ c 30s1 are, respectively, 0.13
and 0.07 for training data and 0.028 and 0.01 for test data. In addition, for strain rates
e_ c [ 30s1 , R2 of DIFs of MC1990 and MC2010 (Eqs. (2) and (4)) are, respectively,
0.41 and 0.23 for training data and 0.47 and 0.33 for test data. This comparison
demonstrates that the proposed ANN-mathematical model performs significantly better
than the current commonly used model MC1990 and MC2010 in the prediction of
compressive DIF of SFRC materials.
Moreover, the experimental compressive DIF is plotted versus strain rate in Fig. 4
and compared with the compressive DIF obtained from MC1990, MC2010 and ANN-
mathematical model. This figure also evidences the good predictive performance of the
developed ANN-mathematical model for the compressive DIF of SFRC materials in
comparison with the proposed formulations in MC1990 and MC2010.
Fig. 2. Comparison of DIFs obtained from MC1990 and MC2010 and ANN-mathematical
model with corresponding experimental values for e_ c 30s1 .
Fig. 3. Comparison of DIFs obtained from MC 1990 and MC 2010 and ANN-mathematical
model with corresponding experimental values for e_ c 30s1 .
Fig. 4. Comparison of experimental DIF with DIF obtained from MC 1990, MC 2010 and
proposed ANN-mathematical model.
4 Design Formulation Based on the ANN-Mathematical

Model
It is worth noting that since the proposed ANN-mathematical model has been trained, it
is ready to be adopted for predicting the DIF of SFRC under high strain rate com-
pressive loads. Although the simulated results from the proposed ANN-mathematical
model have a good agreement with the experimental data, it is inconvenient for
engineers to use the networks for engineering design purposes. Since, the engineers
need to have the fundamental knowledge of ANN and Python to be able to use the
proposed model. Therefore, to make the proposed model more practical for direct
application without performing ANN analysis, a functional-form equation could be
explicitly derived from the trained networks by using input parameters and transfer
functions and combining the weight matrix and the bias matrix (Fig. 5) [3, 5, 8]. The
sigmoid transfer function was used in the hidden layer (see Eq. (7)) and linear transfer
function (see Eq. (8)) was used in the output layer. The procedure to develop the user-
friendly equations based on the ANN-mathematical model to determine k1 and k2
parameters is represented in Fig. 5.
f ðxÞ ¼ 1=ð1 þ ex Þ ð7Þ
f 0 ðxÞ ¼ x ð8Þ
Fig. 5. Architecture of the proposed ANN-mathematical model
The equations derived from the ANN-mathematical model to determine k1 and k2

parameters to be used in Eqs. (5) and (6) for predicting the compressive DIF of SFRC
composite materials are as follows:
0:3 0:3
k1 ¼ þ 0:006 ð9Þ
1 þ ea 1 þ eb
a ¼ 11:70vf þ 0:12ðL=DÞ 0:0007fts 0:09fcs þ 4:80

b ¼ 11:05vf þ 0:22ðL=DÞ 0:0013fts 0:09fcs 0:94 ð10Þ
k2 ¼ 1 0:26
þ ea 1 þ eb þ 0:009
0:26
a ¼ 49:43vf þ 0:04ðL=DÞ 0:0012fts 0:02fcs þ 1:68

b ¼ 55:20vf þ 0:04ðL=DÞ 0:0013fts 0:02fcs 1:55
It worth to note that the proposed formulations are valid for the range of the
parameters reported in the Tables 1 and 2.
Table 1. Range of the variables based on the collected database for e_ c [ 30s1 .
Input Description Unit Range
vf Volume fraction of steel fibers % [0.3−6]
L=D Aspect ratio of steel fibers mm/mm [20−125]
fts Tensile strength of steel fibers MPa [750−4300]
fcs Concrete compressive strength MPa [32−200]
Table 2. Range of the variables based on the collected database for e_ c \30s1 .
Input Description Unit Range
vf Volume fraction of steel fibers % [0.5−3]
L=D Aspect ratio of steel fibers mm/mm [39−81]
fts Tensile strength of steel fibers MPa [900−2850]
fcs Concrete compressive strength MPa [32−160]
5 Conclusion
This study develops a new artificial neural network (ANN) mathematical model based
on modifying the proposed formulation in fib model code 2010 (MC2010) to predict
the dynamic increase factor (DIF) of SFRC composite materials under high strain rate
compressive loads by considering the effective parameters (i.e.: volume fraction of steel
fibers (vf ), aspect ratio of steel fibers (L=D), tensile strength of steel fibers (fts ), and
concrete compressive strength (fcs )). The compressive DIF of SFRC materials obtained
from the developed ANN-mathematical model is compared to the corresponding DIF
obtained from MC2010 and MC1990. A simplified closed form formulation derived
from the developed ANN-mathematical model is proposed to predict the DIF of SFRC
materials under compression. An overview of the developed model points out the
following conclusions:
• A good predictive performance for the proposed ANN-mathematical model in terms
of the compressive DIF of SFRC materials for both strain rates of e_ c 30s1 and
e_ c [ 30s1 is demonstrated by comparing with the relevant experimental results.
• The proposed formulations in MC1990 and MC2010 for the compressive DIF of
concrete cannot predict well the compressive DIF of SFRC material, due to the lack
of parameters to consider the impact of steel fibers in concrete in these formulations.
• The proposed simplified closed form formulation with a design framework derived
from the developed ANN-mathematical model can provide useful estimations of
SFRC strength at high strain rate compressive loads (with the aim of designing the
impact resistance of SFRC elements subjected to high strain rate loads such as blast
and impact, representative of terrorist attacks and accidents which is an ongoing
research project in the University of Minho).
The main objective of the next step of this research study is to extend the ANN-
mathematical model to achieve a higher degree of accuracy in predicting the com-
pressive DIF of SFRC materials by adding other important parameters, i.e. the shape of
steel fibers, fiber orientation factor, and fiber efficiency factor, in the developed ANN-
mathematical model.
Acknowledgements. The study reported in this paper is part of the project “PufProtec - Pre-
fabricated Urban Furniture Made by Advanced Materials for Protecting Public Built” with the
reference of (POCI-01-0145-FEDER-028256) supported by FEDER and FCT funds. The second
author also acknowledges the support provided by FEDER and FCT funds within the scope of
the project StreColesf (POCI-01-0145-FEDER-029485).
References
1. Nili, M., Afroughsabet, V.: Combined effect of silica fume and steel fibers on the impact
resistance and mechanical properties of concrete. Int. J. Impact Eng. 37(8), 879–886 (2010)
2. Soufeiani, L., Raman, S.N., Jumaat, M.Z., Alengaram, U.J., Ghadyani, G., Mendis, P.:
Influences of the volume fraction and shape of steel fibers on fiber-reinforced concrete
subjected to dynamic loading–a review. Eng. Struct. 124, 405–417 (2016)
3. Pham, T.M., Hao, H.: Prediction of the impact force on reinforced concrete beams from a
drop weight. Adv. Struct. Eng. 11, 1710–1722 (2016)
4. Cascardi, A., Micelli, F., Aiello, M.A.: An Artificial Neural Networks model for the
prediction of the compressive strength of FRP-confined concrete circular columns. Eng.
Struct. 140, 199–208 (2017)
5. Pham, T.M., Hadi, M.N.: Predicting stress and strain of FRP-confined square/rectangular
columns using artificial neural networks. J. Compos. Constr. 18(6), 04014019 (2014)
6. Comité Euro-International du Béton, CEB-FIP Model Code 1990, Redwood Books,
Trowbridge, Wiltshire, UK (1993)
7. Fédération Internationale du Béton fib/International Federation for Structural Concrete (du
Béton, Fédération Internationale), CEB-FIP. Model code 2010: Final draft, Lausanne,
Switzerland (2010)
8. Yousif, D.S.: New model of CFRP-confined circular concrete columns: ANN approach.
IJCIET 4(3), 98–110 (2013)
Structural Design
Post-cracking Strength Classification
of Macro-synthetic Fibre Reinforced Concrete
for Sleeper Application
Christophe Camille1(&), Dayani Kahagala Hewage1, Olivia Mirza1,

Fidelis Mashiri1, Brendan Kirkland1, and Todd Clarke2
1
Western Sydney University, Locked Bag 1797, Penrith, NSW 2751, Australia
c.camille@westernsydney.edu.au
2
BarChip Pty Ltd, Sydney, NSW, Australia
Abstract. Nowadays, timber and concrete are among the most extensively used
materials for railway sleepers, characteristically considered as a crucial track
component. However, due to recent concerns regarding the inferior quality,
degradation, durability, high-cost and environmental impact of the conventional
materials, this paper focuses on macro-synthetic fibre reinforced concrete
(MSFRC) as a more sustainable alternative. Despite the encouraging strength
characteristics of the innovative material, its practical implementation as a
composite sleeper remains fairly limited due to the unknown post-cracking
behaviour categorised through the fibre reinforced concrete (FRC) constitutive
laws. Hence, the paper herein investigates the characteristic flexural residual
strength (i.e. serviceability & ultimate) of MSFRC classified in terms of strength
intervals and residual strength ratios as defined in the Fib Model Code.
Experimental data will be adapted onto the design stress-strain relationship with
an insightful comparison of the post-cracking propagation branch relative to
different fibre volume content and fibre types. Therefore, this paper herein will
present beneficial and non-beneficial behaviours of the compliant macro-
synthetic fibre reinforced concrete towards railway structural applications.
Keywords: Macro-synthetic fibre Sleeper Constitutive laws Flexural

residual strength
1 Introduction
Currently, one of the most essential railway track components is the sleeper, which
functions to securely maintain the track geometry while redistributing the wheel loads
(i.e. vertical, lateral & longitudinal) at permissible stresses onto the supporting ballast
[1–3]. Typically known as transoms or bridge ties when in use on railway bridges,
sleepers are the transverse components underneath the rails and are characteristically
made from timber, steel and concrete as represented in Fig. 1. Nevertheless, with the
increase in speed and axle loads of modern trains, these traditional sleepers are sub-
jected to premature failures triggered through material degradation, corrosion, cracking
and fracture.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_64
718 C. Camille et al.
Fig. 1. Traditional sleepers: (a) Timber, (b) Steel & (c) Prestressed concrete.
In recent years, with the Australian Government Green Infrastructures initiatives as

well as numerous developments in the field of railway sleepers which have emerged,
the performance and sustainable characteristics of composite materials have been
highlighted as an alternative to the existing timber, steel and prestressed concrete
sleepers. Indeed, these composite technologies predominantly incorporate polymer-
based (i.e. fibre-reinforced foamed urethane, FFU) as well as fibre reinforced concrete
(FRC) materials which could be engineered with superior mechanical properties to
further reduce the disposal and maintenance cycles of the railway track components
[3, 4]. However, with limited information on the durability aspects of polymer-based
materials in operational conditions, FRC is considered a more feasible and cost-
effective alternative towards the improvement of the widely used prestressed concrete
sleepers where the major concerns are first-crack flexural strength and toughness.
Nowadays, the commonly available fibres for commercial use are steel and syn-
thetic fibre materials whose effectiveness within the cementitious matrix is critically
dependent on the inherent properties of the fibre. Despite steel fibres possessing
superior mechanical characteristics (i.e. elastic modulus & tensile strength), synthetic
fibres offer major benefits in regards to the corrosion resistance and environmental
aspects [5, 6]. Therefore, such implementation of macro-synthetic fibres (i.e. Class II:
Ø > 0.3 mm [7]) in prestressed concrete sleeper is predominantly aimed at the partial
or complete replacement of steel to further improve the design efficiency in terms of
weight, adaptability, structural performances, corrosion resistance and sustainability
[8, 9]. Yet, most studies which investigated the behaviour of fibre reinforced concrete
concluded that fibre content within the moderate range of 0.5% to 5.0% has limited
effect on the mechanical properties of uncracked concrete [10, 11]. In fact, the sig-
nificant benefits of macro-synthetic fibres are observed across the failure plane pro-
viding bridging and confinement of cracks towards considerably improving the energy
absorption capacity (i.e. toughness), ductility and post-cracking behaviour [5, 12].
Furthermore, the orientation and distribution of the synthetic fibres in the cementitious
material are critical towards inducing a more isotropic behaviour as well as ductile
fracture mechanisms of FRC as highlighted in Fig. 2. Even though limited research on
Post-cracking Strength Classification of Macro-synthetic Fibre 719
FRC sleeper has been conducted, comparable reduction in cracks propagation are
expected when implementing macro-synthetic fibre reinforcement in sleepers.
Fig. 2. Schematic illustration of the fibre/matrix bridging mechanism.
Accordingly, in an effort to understand and categorise the benefits of fibre rein-

forced concrete for structural applications, the post-cracking tensile behaviour (i.e. r-e
relationship) must be known [13]. Currently, there are limited accepted standards for
the design of macro-synthetic fibre reinforced concrete (MSFRC) which further
highlights the lack of research towards this innovative material [6, 13]. These standards
include predominantly the Italian National Code (CNR-DT 204/2006), American
Concrete Institute (ACI 544.4R-18) and Fib Model Code 2010 that introduces a
constitutive law model in uniaxial tension for the post-cracking behaviour of MSFRC
as further discussed in this paper. Therefore, the study herein presented will focus on
the reinforcing capabilities (i.e. post-cracking) of macro-synthetic fibre reinforced
concrete towards efficiently reducing the cost while optimising the load-bearing
capacity, failure mechanisms and sustainability for sleeper applications.
2 Methodology

A total of fifteen batches (i.e. 1 plain & 14 MSFRC mixes) were cast with a high
strength concrete mix which was designed and further adjusted to achieve a charac-
teristic compressive strength of 50 MPa as specified for railway sleeper applications
[14]. The mix proportion herein implemented for the evaluation of macro-synthetic
fibre reinforced concrete is provided in Table 1.
The nominal maximum size of coarse aggregates used was 10 mm to further
increase the fibre-matrix interface towards ensuring adequate packing density as well as
homogenous distribution of the fibres. In addition, a high-range water reducer (super-
plasticizer) was incorporated to further improve the workability, dispersion and prevent
fibre clumping throughout the mixes.
Table 1. Mix proportions used for MSFRC.
Component Dosage (kg/m3)

Cement (30% fly ash) 446
Coarse aggregate 870
Nepean paving sand (coarse) 510
Newcastle sand (fine) 350
Water 175
High-range water reducer (HRWR) 4.4 – 5.3
Water-cement ratio ~ 0.39
The macro-synthetic fibres herein investigated are predominantly made from high-
performance polypropylene base materials whose physical and mechanical character-
istics are provided in Table 2. Accordingly, both BC48 and BC58 are characterised by
quite similar mechanical properties including a continuously embossed surface as
presented in Fig. 3. In this study, the macro-synthetic fibres were introduced at various
fibre volumes ranging from 0.2%–2.0% (i.e. 1.82–18.2 kg/m3) to enable the evaluation
of MSFRC post-cracking behaviour for sleeper applications.
Table 2. Properties of the macro-synthetic fibres [15].
Tensile Elastic Specific

Length Diameter
Fibre type Base material strength modulus gravity
(mm) (mm)
(MPa) (GPa) (kg/m3)
Virgin
BC48 48 > 0.3 640 12 890 – 910
polypropylene
Bi-component
BC58 58 > 0.3 640 10 910 – 920
polymer
For each fibre-dosage combination, three cylindrical and four notch beams speci-
mens of size Ø100 200 mm and 150 150 575 mm were cast respectively to
further assess the compressive and residual flexural strengths at 28 days.

The testing procedures undertaken for both the compression and flexural specimens
comply with AS1012.9-2014 [16] and EN14651-2005 [17] respectively, thus ensuring
the accuracy and reliability of the experimental results herein presented. Accordingly,
the specimen’s geometry and experimental setup are displayed in Fig. 4.
Fig. 3. Macro-synthetic fibre reinforcement: BC48 (left), BC58 (right).
Fig. 4. Specimens geometry and testing configuration (mm).
Prior to testing, notches were cut at mid-span across the tensile side of the beam
specimens to be further assessed in a closed-loop displacement control three-point
bending test arrangement (i.e. 1 constrained DOF at each roller). As a result, crack
initiation and propagation occur within a predefined section of the beam for which the
load versus crack mouth opening displacement (CMOD) was continuously recorded
(i.e. 5 Hz). Correspondingly, the flexural strengths (i.e. pre & post-cracking) of the
macro-synthetic fibre reinforced concrete specimens were evaluated from the load-
CMOD curves as expressed through Eq. (1):
3:FJ :L
fR;j ¼ 2
ð1Þ
2:B:Hsp
where FJ is the load corresponding with CMOD; L = span length; B = average beam
width and Hsp = distance between tip of notch to the top of specimen. Similarly, the
effective fracture energy (Gf) of the notched specimens in a three-point bending con-
figuration was evaluated through Eq. (2) as recommended by RILEM Technical
Committee 50 [18]. The fracture energy of synthetic fibre reinforced concrete is
characterised by the area under the load-deflection curve which is also typically defined
as the energy required to open a unit crack surface area [19].
" #
dmax
Z
Gf ¼ F ðdÞdd þ m:g:dmax =½ðd ao Þ:B ð2Þ
0
where dmax corresponds to the maximum deflection; m = beam mass; g = 9.81 ms−2;
d = beam depth and ao = notch depth.
In comparison to plain concrete, it is predominantly the tensile bridging mechanism
of the macro-synthetic fibres which enables a substantial increase in the energy
absorption capacity and post-cracking behaviour of the MSFRC as further demon-
strated in the herein presented study.
2.3 Post-cracking Flexural Strength Behaviour

In the study herein presented, the classification of FRC’s post-cracking strength comply
with the Fib Model Code 2010 which assumed a linear-elastic relationship [9]. Indeed,
this approach defined two main characteristic flexural strength values that are crucial
towards the structural design of MSFRC for serviceability (fR1k) and ultimate (fR3k)
conditions. For instance, these particular strength parameters correspond with CMOD
of 0.5 and 2.5 mm respectively. Table 3 presents the classification of fibre reinforced
concrete in terms of strength interval and residual strength ratios.
Table 3. Residual strength class & ratio in accordance with Fib Model Code 2010 [9].
Strength interval (MPa) 1.0, 1.5, 2.0, 2.5, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, …
a if 0.5 ≤ fR3k / fR1k < 0.7
b if 0.7 ≤ fR3k / fR1k < 0.9
Residual strength ratio
c if 0.9 ≤ fR3k / fR1k < 1.1
(fR3k / fR1k)
d if 1.1 ≤ fR3k / fR1k < 1.3
e if 1.3 ≤ fR3k / fR1k
In addition, a stress-crack opening constitutive law can also be defined towards

characterising the post-cracking strength of macro-synthetic fibre reinforced concrete as
a linear softening/hardening behaviour. The linear model identifies two reference
values defined as the serviceability (fFts) and ultimate residual strengths (fFtu) evaluated
using Eq. (3) and Eq. (4) respectively [9].
fFts ¼ 0:45:fR1 ð3Þ

wu
fFtu ¼ fFts :ðfFts 0:5:fR3 þ 0:2:fR1 Þ 0 ð4Þ
CMOD3
where wu = maximum acceptable crack opening (taken as 2.5 mm), fR1 and fR3 cor-
responding to the loads at CMOD of 0.5 and 2.5 mm respectively.
As a result, the stress-strain relationship of FRC (shown in Fig. 5) in uniaxial
tension can be derived wherein which the various cases (i.e. I, II & III) may describe
the post-cracking behaviour through the softening or hardening of the material. In fact,
the first branch represents the pre-peak behaviour of plain concrete followed by the
second branch which characterises the post-peak crack propagation until intersecting
with the residual post-cracking behaviour (third & fourth branch). Hence, resuming the
contribution of the fibres throughout the material towards a characteristic softening or
hardening [9].
Fig. 5. Stress-strain relationship of fibre reinforced concrete [9].

At 28 days of curing, the macro-synthetic fibre concrete specimens were tested for their
ultimate compressive strength (fc) for which most of the specimens incorporating BC48
and BC58 at different fibre content were observed to exceed the minimum 50 MPa
sleeper’s requirement. Figure 6 summarises the average compressive strength beha-
viour of MSFRC specimens.
BC48 specimens were observed to peak at 62.7 MPa (i.e. 0.2% fibres) character-
ising an increase of approximately 18% over plain concrete which only achieved an
ultimate capacity of 53.4 MPa. This particular behaviour of BC48 specimens at 0.2%
fibre volume ratio can be justified by the slightly lower spread-flow obtained. On the
other hand, specimens reinforced with BC58 (up to 1%) experienced on average a 6.6%
increase in compressive strength as compared to BC48. The ultimate strengths of the
MSFRC encompassing BC58 fibres were observed to be closely consistent up to 1.0%
of fibres, reaching a peak of 61.8 MPa at 0.8%. However, as the fibre dosage exceeds
70
Compressive Strength, fc (MPa)
60
50
40
30
20
10
0
Fibre Volume Ratio, % (kg/m3)

BC48 BC58 Control
Fig. 6. Compressive strength behaviour of MSFRC at 28 curing days.
1%, the reported strength decreases due to the balling effect of longer fibres which also
involved undesirable air voids, poorer workability and segregation. Therefore, it is
recommended that the design mix is improved towards a more workable concrete
especially at higher fibre dosage, although the addition of macro-synthetic fibres has no
consensual benefits on the ultimate compressive performances. Yet, the reinforcing
fibres significantly improved the failure mechanisms towards more ductile concrete
specimens with the ability to absorb energy without shattering into pieces.

In general, concrete is considered to undergo quasi-brittle failure wherein which the
specimen experiences a complete loss of loading capacity once the failure has been
initiated. Therefore, the flexural behaviour presented in this study fundamentally
identifies the important benefits of incorporating macro-synthetic fibres into concrete
towards improved post-cracking characteristics as presented in Fig. 7.
Although the addition of macro-synthetic fibres had negligible impact on the pre-
cracking strength, both types of fibre are observed to significantly increase the residual
flexural (post-cracking) capacity as the fibre volume ratio increases up to 2.0%. The
main difference across the two types of fibre used in this study is the residual strength
which peaked at around a CMOD of 2.5 mm for BC48 as compared to 3.5 mm for
BC58. Correspondingly, BC58 specimens tend to possess a greater ductility due to the
longer and more flexible fibres. Overall, the specific post-cracking behaviour of the
MSFRC is justified by the mechanics of crack formation and propagation where both
fibre types interact with the concrete matrix. As a result, BC 48 and BC58 adequately
stabilised and suppressed crack growth towards more ductile failure modes, a desirable
material behaviour for sleeper’s application.
Table 4 classifies the post-cracking flexural strength of macro-synthetic fibre
reinforced concrete with various fibre dosages and types. Characteristically, these
residual strength classes could be used for design purposes wherein which the fibre
9
8
Flexural strength (MPa)
7
6
5
4
3
2
1
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
CMOD (mm) CMOD (mm)
Fig. 7. Residual flexural-CMOD behaviour for BC48 (left) & BC58 (right).
Table 4. Residual flexural strength classification of MSFRC.
Residual flexural strength (MPa) Residual strength class

Fibre dosage BC48 BC58 BC48 BC58
% kg/m3 fR1 fR3 fR1 fR3
0.2 1.82 1.26 1.18 1.02 0.62 - -
0.4 3.64 1.32 1.87 1.36 1.71 1e 1d
0.6 5.46 1.96 2.98 1.88 2.87 1.5e 1.5e
0.8 7.28 2.54 3.70 1.92 3.07 2e 1.5e
1.0 9.10 3.69 5.92 2.70 4.76 3e 2e
1.5 13.7 4.53 7.12 3.86 6.87 3e 3e
2.0 18.2 5.09 7.70 3.75 6.65 4e 3e
reinforcement partially or completely replace conventional reinforcement in sleepers at

ultimate limit state.
3.3 Fracture Energy

The effective fracture energy (Gf) of MSFRC derived in this study characterises the
energy dissipated across the notched specimens up to a corresponding mid-span
deflection of approximately 3.4 mm (i.e. CMOD = 4 mm). Indeed, this value was
designated to consistently evaluate and compare the Gf of the macro-synthetic fibre
reinforced concrete specimens within its design residual strength classification as
previously discussed. Figure 8 illustrates the changes in Gf with an increase in the fibre
volume ratio of the macro-synthetic fibres.
5.0
Fracture energy, Gf (N/mm)
4.0
3.0
2.0
1.0
0.0
Fibre Volume Ratio, % (kg/m3)
Control BC48 BC58 Linear (BC48) Linear (BC58)
Fig. 8. Effect of fibre type and volume fraction on the effective fracture energy (vertical bars
represent standard deviation).
In general, a linear increase in the effective fracture energy was observed for both
BC48 and BC58 specimens when tested up to a fibre volume fraction of 2.0%. This
linear behaviour of polymer fibres has already been observed previously, thus sup-
porting the accuracy of the results. As a result, the macro-synthetic fibre reinforcement
herein investigated has the ability to improve the toughness of concrete due to its
relatively low elastic modulus and flexible fibres which can be uniformly distributed
throughout the cement matrix [19]. Undeniably, this atypical behaviour of concrete is
desired in the railway industry wherein which the sleeper does not need to be imme-
diately replaced once cracked.
3.4 Stress-Strain Relationship

The design stress-strain (r-e) relationships were plotted for both BC48 and BC58 at
various fibre volume ratios as represented in Fig. 9. The r-e results obtained in this
study characterises the post-cracking stages (i.e. softening or hardening) of macro-
synthetic fibre reinforced concrete when subjected to uniaxial tension as predefined
through the crack-opening constitutive law.
Indeed, both BC48 and BC58 are observed to enhance the post-cracking perfor-
mances as compared to plain concrete, predominantly characterising a Case I (Fig. 5)
constitutive relationship. For instance, up to a fibre volume fraction of 0.6%, both types
of fibres exhibited a fairly improved softening behaviour as opposed to plain concrete.
Nevertheless, when reinforced with a fibre volume ratio of 0.8% up to 2.0%, an
uncharacteristic hardening branch (i.e. 4th branch) is developed due to the high fibre
Fig. 9. Design stress-strain relationship for BC48 (top) and BC48 (bottom) at various fibre
dosages.
dosage bridging the crack surfaces. Correspondingly, the stress-strain relationship of

macro-synthetic fibre reinforced concrete in the post-cracking region is characterised
through an improved Case I exhibiting hardening of the material as presented in Fig. 9.
4 Conclusions
In this paper various impacts of macro-synthetic fibre reinforcement on the post-

cracking performances of concrete are discussed from which the following conclusions
were achieved:
• Compressive strength – The experimental results demonstrated that the macro-
synthetic fibres provide evident bridging effect characteristically reducing the crack
propagation towards more ductile failure modes. However, the addition of fibres did
not noticeably influence the ultimate compressive strengths.
• Flexural behaviour & fracture energy – The addition of fibres was observed to
insignificantly impact the ultimate flexural strength (i.e. pre-cracking) of the spec-
imens. Nevertheless, the fibre reinforcement improved the fracture mechanisms
towards a more ductile behaviour, naturally minimising the loss in capacity sus-
tained after the initial cracks. In general, higher fibre dosages exhibited better
performance in terms of residual flexural strength (i.e. post-cracking), ductility and
toughness towards structural design implementation involving the partial or com-

plete substitution of traditional steel reinforcement.
• Stress-strain behaviour – The incorporation of the macro-synthetic fibres into the
concrete matrix especially at high fibre content generate an atypical yet desired
hardening behaviour once the specimens endured a loss in capacity after the first
crack.
Acknowledgements. The authors gratefully acknowledge the Australian Research Council’s

Industrial Transformation Training Centres Scheme (ARC Training Centre for Advanced
Technologies in Rail Track Infrastructure; IC170100006) which provided the catalyst for
undertaking this research as well as Western Sydney University through the School of Com-
puting, Engineering and Mathematics for their support to the authors work described herein.
References
1. Kumar, D.K., Sambasivarao, K.: Static and dynamic analysis of railway track sleeper. Int.
J. Eng. Res. Gener. Sci. 2(6), 1–10 (2014)
2. Tzanakakis, K.: The Railway Track and Its Long Term Behaviour, 1st edn. Springer Tracts
on Transportation and Traffic. Springer-Verlag, Berlin Heidelberg (2013)
3. Sharma, R.C., Palli, S., Sharma, S.K., Roy, M.: Modernization of railway track with
composite sleepers. Int. J. Vehicle Struct. Syst. 9(5), 321–329 (2017)
4. Ferdous, W., Manalo, A., Van Erp, G., Aravinthan, T., Kaewunruen, S., Remennikov, A.:
Composite railway sleepers–recent developments, challenges and future prospects. Compos.
Struct. 134, 158–168 (2015)
6. Juhász, K.P.: The effect of the synthetic fibre reinforcement on the fracture energy of the
concrete. In: IOP Conference Series: Materials Science and Engineering, vol. 613,
p. 012037, 4 Nov 2019
7. Fibres for concrete-Part 2: Polymer fibres-Definitions, specifications and conformity,
EN14889–2:2006 (2006)
8. Kohoutková, A., Broukalová, I.: Optimization of fibre reinforced concrete structural
members. Procedia Eng. 65, 100–106 (2013)
9. Fib Model Code for Concrete Structures 2010, 9783433604083 (2013)
10. Yehia, S., Douba, A., Abdullahi, O., Farrag, S.: Mechanical and durability evaluation of
fiber-reinforced self-compacting concrete. Constr. Build. Mater. 121, 120–133 (2016)
11. Enfedaque, A., Alberti, M.G., Gálvez, J.C., Domingo, J.: Numerical simulation of the
fracture behaviour of glass fibre reinforced cement. Constr. Build. Mater. 136, 108–117
(2017)
12. Merta, I., Tschegg, E.K.: Fracture energy of natural fibre reinforced concrete, (in English).
Constr. Build. Mater. 40, 991 (2013)
13. Jafarifar, N., Pilakoutas, K., Angelakopoulos, H., Bennett, T.: Post-cracking tensile
behaviour of steel-fibre-reinforced roller-compacted-concrete for FE modelling and design
purposes. Materiales de Construcción 67(326), e122 (2017)
14. Railway Track Materials-Part 14: Prestressed concrete sleepers, AS1085:14:2012 (2012)
15. BarChip, 20 October 2018. https://barchip.com/product/
16. Methods of testing concrete-Methods 9: Compressive strength tests-Concrete, mortar and

grout specimens, AS 1012.9:2014 (2014)
17. Test method for metallic fibered concrete-Measuring the flexural tensile strength (limit of
proportionality (LOP), residual), EN14651:2005 (2005)
18. RILEM, Recommendation TC 50-FMT-Determination of the Fracture Energy of Mortars
and Concretes by Means of Three-Point Bend Tests on Notched Beams. Mater. Struct. vol.
18 (1985)
19. Kosior-Kazberuk, M.: Post-cracking Behaviour and Fracture Energy of Synthetic Fibre
Reinforced Concrete. Mater. Sci./Medziagotyra, Article 22(4), 542–547 (2016)
Structural Behaviour of Steel-Fibre-Reinforced
Lightweight Concrete

a.abbas@uel.ac.uk
Abstract. This paper evaluates the influence of adding hooked-end 3D fibres to

recycled lightweight concrete beams with and without shear reinforcement.
Lightweight concrete used is from recycled fly ash and can lead to reductions in
the mass of the structure. However, it is typically more brittle than normal
weight concrete. To enhance ductility, steel fibres were used and the complete
and partial substitution of stirrups with fibres was investigated. The flexural and
shear behaviour of SFRLC beams under 3-point loading with adequate, sub-
standard and no shear reinforcement are studied. The experimental study
includes Vf = 0%, 1% and 2% as well as stirrups spacing of 120 mm, 240 mm
and ∞. The results in this paper are given in terms of strength, ductility, crack
patterns and nature of failure. It was found out that fibre reinforcement is capable
of partially replacing stirrups, however, when stirrups were completely removed
the failure became brittle. Nonlinear finite element analysis (NLFEA) using
ABAQUS is also used to validate a SFRLC constitutive model whose properties
were derived from compressive and pullout-tests.
Keywords: Recycled lightweight concrete beams 3D hooked-end fibres

3-point-bending test Shear design Stirrups replacement Ductility Load
bearing capacity Validation NLFEA ABAQUS
1 Introduction
The use of structural lightweight aggregate concrete brings a number of advantages as

compared to the conventional normal weight concrete such as thermal insulation and
fire resistance. Besides, the lightweight aggregate used in this work is recycled and
offers reduction in CO2 emissions as well as an alternative to the depleting gravel and
quarry resources (Gerritse 1981). More importantly, the improved strength-to-weight
ratio results in smaller cross sections which in turn leads to a decrease in reinforcement
and savings in material transport costs due to lower inertial and gravity loads. The
usage of lightweight concrete can therefore be ideal and competitive in industry for the
growing need for taller and longer span structures, especially in seismic and dynamic
zones (Libre et al., 2011; Campione, 2014; Dias-Da-Costa 2014; Mo et al., 2017).
A study on Lytag (Lytag, 2014) showed that lightweight concrete can bring about 34%
savings in CO2 as well as a reduction of up to 48% of concrete and reinforcement when
compared to conventional gravel concrete. These advantages however come as a trade-
off for the increased brittleness of lightweight concrete. The latter is the case due to the
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_65
Structural Behaviour of Steel-Fibre-Reinforced Lightweight Concrete 731
lightweight material having poor aggregate interlock which translates into lacking
toughening mechanisms in tension post-crack. This causes lower tensile strength which
ultimately lowers shear resistance in structures such as beams and slabs. Also, the
porous nature of lightweight concrete leads to lower modulus of elasticity (Chen et al.,
2010; Badogiannis and Kotsovos 2013) which leads to excessive deflections and
cracking (Lim et al., 2006; Wu et al., 2011). Nonetheless, the brittle nature of light-
weight concrete can be addressed by incorporating traditional reinforcement such as
rebars and stirrups. However, the latter solution can become inherently counterpro-
ductive and impractical when reduction in structural elements is sought by employing
lightweight concrete especially at critical zones such as joints. Therefore, fibre rein-
forcement which has long proven its effectiveness in controlling and bridging tensile
and shear cracks in the past for both lightweight and normal weight concretes can
become an adequate solution (Gao et al., 1997; Campione and La Mendola, 2004;
Abbas et al., 2014a; Di Prisco et al., 2013; Grabois et al., 2016; Mo et al., 2017). For
over 40 years the usage of steel fibres in concrete mixes has been experimented with
and used (Ritchie and Kayali 1975), however, comprehensive studies on fibrous
lightweight concrete is still scarce with most work being merely theoretical (Swamy
et al., 1993; Kang and Kim 2010; Di Prisco et al., 2013; Iqbal et al., 2015; Grabois
et al., 2016). It should be noted that, at present, there is no international standards
specific for steel fibre reinforced lightweight concrete (SFRLC) with current guidelines
being usually adapted from fibrous normal weight concrete.
This work is a continuity to an accompanying paper that evaluates the behaviour of
plain and fibrous lightweight concrete on the material level in both tension and com-
pression. The latter paper’s constitutive model is used to validate the reinforced con-
crete beams on the structural level by studying the shear and flexural behaviour via
NLFEA using ABAQUS (Habbit et al., 2000). The parameters used in this study
include fibre dosage (Vf), and stirrups spacing (s). The first parameter will be either 0%,
1% or 2% while the second will be 120 mm, 240 mm or ∞. The choice of the beam
spans and transversal reinforcement was as such so that a/d = 3 with d as the effective
depth and a as the shear span. This anticipated Type II shear failure behaviour
according to Kani’s valley (Kotsovos and Pavlovic 2013) and hence a brittle failure is
predicted for the beams unless adequately shear reinforced. It should be noted that a
stirrups spacing of 120 mm is adequate to Eurocode 2 shear design and is therefore
expected to fail in a ductile flexural manner while the spacing of 240 mm is inadequate
to Eurocode 2 beam design and is consequently expected to fail in brittle shear manner.
No stirrups spacing i.e. s = ∞, is an even more severe case of substandard reinforced
concrete beam design. The fibre type used in this study will be the conventional
hooked-end 3D fibres while the target compressive strength chosen was 30 MPa for all
the reinforced concrete beams tested. This fibre type was chosen for experimental tests
because hooked-end 3D fibres are currently the most commonly used in industry
(Abdallah et al., 2018a), and they provide the least tensile upgrade to plain concrete as
compared to 4D and 5D. Also, a compressive strength of 30 MPa which consequently
provides the lowest tensile strength possible for structural lightweight concrete.
2 Methodology
The methodology of the work presented in this paper consists of two main parts. The
first will involve experimental testing of the structural reinforced concrete beams under
3-point loading. The main focus of this part will be on the shear and flexural strength,
ductility and cracking pattern. The second part will cover nonlinear numerical finite
element analyses of the structural beams using ABAQUS’ concrete damaged plasticity
by adopting the proposed model suggested in a previous paper as well as a comparison
with 3 available SFRC models mentioned previously.

As previously noted, an accompanying paper with more work detailing the material
properties of plain and fibrous lightweight concrete was conducted. Figure 1 below
summarises the complete proposed model adopted for fcm = 30 MPa with Vf = 0%,
1% and 2% and used for the structural reinforced beams on ABAQUS. Also, for over
100 compression tests, Poisson’s ratio was recorded between 0.15 and 0.20 with no
particular pattern in relation to strength, Vf, or fibre type. It should be noted that this
model was input into ABAQUS concrete material relationships.
The proposed constitutive model for fcm=30MPa and 3D fibres

5
0
-5 -3 -1 1 3 5 7 9
-5
Stress (MPa)
-10
Vf=0%
-15 Vf=1%
Vf=2%
-20
-25
-30
Strain ε (‰) Crack width ω (mm)
Fig. 1. The proposed constitutive model for SFRLC
The study in this paper focuses on reinforced concrete beams. As mentioned before,
2 parameters are studied, namely; Vf and stirrups spacing s. 7 beams are tested, 3 for
s = 120 mm, 3 for s = 240 mm with Vf = 0%, 1% and 2%, and one beam for s = ∞
with Vf = 1%. The latter aimed at studying the possibility of fibres to completely
replace stirrups. The choice to skip lower dosages such as Vf = 0.5% is so, since most
common structures such as tunnels and pavements require at least a dosage of Vf = 1%
for efficient crack control and thickness reduction (The Concrete Society, 2007).
3.1.1 Materials
absorption was estimated to be around 15% per mass of LYTAG. Natural river sand
with a 4.75 mm maximum size was used as the fine aggregate of the concrete. The sand
had a water absorption coefficient of 0.09% and specific gravity of 2.65 complying with
BS EN 12620. The properties of the fibres used are summarized in Table 1 below. It
should be noted that to prevent the possibility of balling, fibres were collated from the
manufacturer. The reinforcement steel bars and stirrups modulus of elasticity was 210
GPa while the yield tensile stress for 6 mm bars was 512 MPa and that for 12 mm bars
was 548 MPa.

Fibre type ru (MPa) lf (mm) df (mm)
3D 65/60 1160 60 0.9
3.1.2 Mix Design

The mix design used is summarized in Table 2 below. The mix design of the Lytag
summarised below. These were directly adopted from Lytag (2011) manuals.

LC30/33 370 592 668.8 175
confirmed by Lytag manual. For this reason and as suggested by Lytag manual 5
(2011), excess water to saturate Lytag aggregates was added 30 min before mixing
(Fig. 2).

For the 3 point- flexural beam tests, the LVDTs were glued using high strength epoxy
after the concrete surface in touch with the LVDT was ground in the mid-section at the
front of the beam to enable the LVDTs to fully adhere onto the concrete. The LVDT’s
are connected to a computer software. For the purpose of estimating the vertical
deflection accurately, a steel bar inspired by a technique similar to JSCE-SF4 recom-
mendations was made. The beams are placed onto two frictionless steel supports. This
was deemed adequate as the loading was symmetrical (Fig. 3). A displacement con-
trolled constant loading of 0.2 mm/min was adopted using the hydraulic machine
which has a load capacity of 500 kN. It should be noted that the machine is also
supplied with a calibrated displacement transducer.
Fig. 3. On the left, vertical LVDT touching the Japanese Yoke and on the right, structural beam
under 3-point loading
A relatively low longitudinal reinforcement ratio in tension of q = 1.8% with 2

T12 rebars was chosen to avoid possible over-reinforcement of section in tension as
fibres were added. The longitudinal reinforcement ratio was similar to that adopted by
Kodur et al. (2018) for SFRC beams. For compression reinforcement, 2 T6 rebars
were used to support the compression flange of the lightweight concrete beams to avoid
the possibility of concrete crushing as Vf was increased to 2%.
Figures 4, 5 and 6 below show the 3 sections used for the structural beams.
Fig. 4. Cross section for beam with S = 240 mm
Fig. 5. Cross section for beam with S = 120 mm

Fig. 6. Cross section for beam with S = ∞
4 Numerical Study
Three dimensional nonlinear finite element analyses will be carried out using ABA-
QUS (Zienkiewicz and Taylor, 2005). This software has shown to be successful at
predicting the behaviour of SFRC in tension, compression, flexure and shear as well as
the cracking pattern and mode of failure of plain and reinforced elements to a good
level of accuracy (Tlemat et al., 2006; Syed Mohsin 2012; Abbas et al., 2014b;
Behinaein et al., 2018). The approach adopted in this work will involve modelling both
plain and fibrous concrete. The fibres will not be modelled discretely, instead they will
be introduced directly into the constitutive tensile and compressive models of the
concrete in ABAQUS. This methodology was opted for since modelling fibres dis-
cretely can be time consuming and will produce a nonflexible FE-based model difficult
to be adopted by designer engineers. Moreover, although modelling fibres discretely
using probabilistic techniques such as Monte Carlo is aimed to account for the random
distribution of fibres (Cunha et al., 2011), its usage can itself be unrealistic as the
prediction might as well be likely to completely miss the actual distribution and
location of fibres in a particular structural element. Hence, modelling fibres as part of
the concrete matrix as shown in a number of design guidelines can offer an easier and
perhaps safer prediction of the behaviour of composite material (Lok and Xiao 1999;
RILEM TC 162-TDF 2003; Barros et al., 2005; Di Prisco et al., 2013). However,
unlike the discrete 3D modelling, it should be noted that the homogenous fibrous
concrete modelling is not aimed to detect local failures on the mesoscale level explicitly
such as fibre rupture and concrete fracturing at the IFZ (Zhang et al., 2018).
Generally, there are two approaches that can be used in FEM to predict the tension
stiffening behaviour of fibrous concrete: the discrete crack approach (r-x) (Ngo and
Scordelis, 1967) or smeared crack approach (r-e) (Rashid, 1968). Although more
accurate at post-crack, the discrete crack approach can be impractical and numerically
intensive to use (Tlemat et al., 2006), while the more accepted smeared crack approach
that assumes the crack is smeared over an element can be more useful for design. In this
work, a smeared crack approach with (r-x) is adopted using ABAQUS option to
define cracking displacement rather than cracking strain. ABAQUS derives the strains
based on the characteristic length which is defined as the mesh size for hex elements
using e = x/lc. The values of crack width is input into ABAQUS directly from the
pullout test. As compared to the strain, the r-x approach offers a number of advantages
since it represents the actual behaviour of the fibrous material, is member size-effect
independent and can be directly applied to FEM (ABAQUS) from the pullout tests used
(De Montaignac et al., 2011). On ABAQUS, mesh sensitivity is not an issue for r-x
relation as compared to r-e. Also, r-x relationship can provide the necessary infor-
mation needed to design for service limit state including fatigue and shrinkage.
Concrete damaged plasticity (CDP) was calibrated and chosen to model the notched
beam specimens to check the validity of the r-x model. The calibration of CDP on the
material level for both cylinder compression test and pullout test can be found with
more detail elsewhere (Al-Naimi and Abbas 2019). Table 3 below summarises the
CDP parameters used.
Table 3. Parameters usage for CDP

Dilation angle Eccentricity fb0/fc0 K Viscosity
25 0.1 1.16 0.666 0
The explicit dynamic solver is adopted. The explicit dynamic analysis was found to be
a more computationally efficient tool at solving the problems used in this project as
compared with the implicit solver which had a tendency to terminate. The following
properties of the finite element model were based on a comprehensive convergence study.
To ensure a quasi-static solution the ratio between kinetic energy and internal
energy was kept below 1% throughout the analysis and any spikes in energy even
below 1% were monitored as they could represent an indication of model failure. The
analysis was ran using a displacement induced loading rate of 0.1 mm/step in a
smoothened step with a mass scaling of 50.
The finite element model investigated to validate the structural beams is shown
below. Due to the symmetrical setting only a quarter of the beam was modelled with
symmetry boundary conditions along the Z and X directions. The beam was restricted
from moving in the Y direction by applying a displacement rotation boundary con-
dition (in the initial step) along the middle line of the support. The displacement-
induced load was applied (in the explicit dynamic step) on the surface of the loading
steel plate in a similar manner to the experiment. To estimate the load, the reaction
force was calculated by summing up the load along the boundary condition line on the
bottom of the support then multiplied by 4. The displacement, however, was calculated
by taking the average Y displacement of the surface or line of the beam in middle point
of the total span of the full beam. With regards to meshing the concrete, hexagonal
brick element C3D8R of 20 mm element size with reduced integration and hourglass
control was used. For the reinforcement stirrups and rebars, wire elements T3D2 were
embedded in concrete with true stress-strain material properties based on tensile tests of
the rebars. The supports and load blocks were made to be rigid and a tie coupling to the
concrete was applied Fig. 7.
Fig. 7. FE model used for the reinforced concrete beams with boundary conditions for beam of
S = 120 mm (ABAQUS)

Figure 8 below shows the cracking pattern for the beams tested. The beam with no
stirrups fails in shear in an identical patter to a-. Aside from a- which fails in shear and
f- which fails in flexure all the beams show a shear-flexure failure.
Fig. 8. The tested structural beams

Figure 9 below shows the load-deflection curves for the 7 reinforced concrete
beams. For all the beams illustrated in the figure, it can be seen that the addition of
fibres increases the load bearing capacity and ductility. The beam which failed in shear
(solid brown line), failed in a ductile manner once fibres are added (dashed brown line).
The interval o-t indicates can be seen between 0 kN to 10−12 kN for all the beams
tested confirming that the effect of fibres on stiffness in lightweight concrete can only
be activated after cracking takes place. This was detailed in an accompanying paper. At
point t tension cracking takes place followed by tension steel rabar yielding at y. The
maximum load is denoted by the letter m while the ultimate load is denoted by the letter
u characterized by either complete failure (for beams failing in shear) or 85% of
maximum load at post peak (for beams failing in flexure).
Fig. 9. Load-deflection for the tested beams
Table 4 below details the load and deflection at yield y, peak p, and ultimate u. It is
evident that all the 3 types of loads increase with the increase of either fibre dosage or
reduction of stirrups spacing. It is interesting to see that the maximum load bearing
capacity of beam S = 240 mm, Vf = 1% is not very far from that of beam
S = 120 mm, Vf = 0% while similarly the maximum load bearing capacity of beam
S = 240 mm, Vf = 2% is close to that of beam of S = 120 mm, Vf = 1%. For all these
beams, the higher the Vf the higher the ductility regardless of stirrup spacing with the
beam with S = 120 mm, Vf = 2% recording the highest while the beam with
S = 240 mm, Vf = 0% recording the lowest. It can be deduced from the latter that the
maximum load bearing capacity can be maintained while the ductility improved by
doubling the stirrups spacing and adding fibre volume fraction of approximately 1%.
The beam with S = ∞, Vf = 1% bears a maximum loading slightly higher than that of
S = 240 mm, Vf = 1%. This slight difference can be attributed to variances in the
compressive strength. A high fcm consequently leads to stronger fibre-matrix interfacial
bond and thus a higher residual tensile strength of the beam. This is seen with the beam
of S = ∞, Vf = 1% which records an average fcm of 36 MPa as compared to that of
S = 240 mm, Vf = 1% which has an fcm of 31 MPa. However, it is vital to notice that
the beam with S = ∞, Vf = 1% still failed in a brittle shear manner and recorded a
ductility as low as that of S = 240 mm, Vf = 0%. The latter beam also failed in shear
which proves the adequacy of the adopted methodology. Figure 10 below shows both
the strength and ductility ratios. Both control ductility and load bearing capacity were
chosen as those of the adequately designed beam with S = 120 mm, Vf = 0%.
Table 4. Load bearing capacity and ductility details.

S Vf dy Py dp Pm du Pu Max crack Nature of lu
(mm) (%) (mm) (kN) (mm) (kN) (mm) (kN) spacing (mm) failure
240 0 0.67 27.7 1.37 32.1 1.86 27.3 79.4 S 2.77
1 0.76 30.1 3.90 59.3 4.96 50.4 60.4 SF 6.61
2 0.62 32.2 5.17 78.4 9.59 66.8 53.1 SF 15.41
120 0 0.77 28.3 3.95 51.8 5.31 44.1 79.0 SF 6.91
1 0.69 31.1 4.32 78.1 5.79 66.4 60.1 SF 8.41
2 0.50 33.8 3.69 92.7 24.3 78.8 53.3 F 48.24
∞ 1 0.77 30.2 3.11 62.5 3.56 53.4 60.2 S 4.64
*S=Shear F=flexure SF=shear flexure
Ductility Ratio μu (Solid line) and Strength Ratio P0 (dashed line)

8
S=120mm
7
S=240mm
6 S=∞
Ductility Ratio (μu)
Strength Ratio (P0)
5 S=120mm
S=240mm
4
S=∞
3
0
0 1 2
Vf (%)
Fig. 10. Strength and ductility ratios

5.2 Numerical Results

Figure 11 below illustrates the cracking patterns of the beam with S = 120 mm,
Vf = 2% which failed in flexure and that of S = ∞, Vf = 1% which failed in shear on
ABAQUS.
Fig. 11. At ultimate deflection: on the left: beam failing in flexure, on the right: beam failing in
shear
Validation of the tested SFRLC beams is illustrated in Fig. 12 below. It can be seen
that the model adopted closely resembles the load-deflection behaviour of the SFRLC
beams tested. Once high spikes were noticed in the KE/IE graph on ABAQUS, the
simulation was stopped since this was an indication of the instability of the solution
beyond that point as it represents a non-quasi-static behaviour. Both load bearing
capacity and ductility were accurately and safely predicted by ABAQUS.
Behaviour of fibrous beams using the proposed constitutive material model on ABAQUS
100
90 EXP (S=120mm. V=1%)

ABAQUS (S=120mm, V=1%)
80
EXP (S=120mm. V=2%)
70 ABAQUS (S=120mm. V=2%)
EXP (S=240mm. V=1%)
60
ABAQUS (S=240mm. V=1%)
Load (kN)
50 EXP (S=240mm, V=2%)

ABAQUS (S=240mm, V=2%)
40
EXP (S=∞, V=1%)
30 ABQUS (S=∞, V=1%)
EXP (S=∞, V=2%)
20
ABAQUS (S=∞, V=2%)
10
0
0 2 4 6 8
Displacement (mm)
Fig. 12. Validation of the SFRLC beams using the suggested constitutive model
6 Conclusions
• The methodology used was successful at predicting the behaviour of fibrous
lightweight concrete beams with and without web reinforcement.
• The addition of fibres can alter the behaviour of beams from brittle shear to ductile
flexural such as the case with the deficiently reinforced beam of S = 240 mm and
Vf = 0%. However, the latter statement only remains valid provided that a sort of
minimum reinforcement is provided.
• It is safe to deduce that adding fibres of Vf = 1% permits the stirrups spacing to be
doubled without compensation of strength or ductility. In fact the behaviour of
beam with S = 240 mm, Vf = 2% was more favourable in terms of ductility than
the beam with S = 120 mm, Vf = 1%, however the maximum load bearing capacity
Pm remained nearly identical for both beams.
• The numerical NLFE results show excellent correlation between the proposed
constitutive model predictions and experimental results.
References
Gerritse, A.: Design considerations for reinforced lightweight concrete. Int. J. Cem. Compos.
Lightweight Concr. 3(1), 57–69 (1981)
Libre, N., Shekarchi, M., Mahoutian, M., Soroushian, P.: Mechanical properties of hybrid fiber
reinforced lightweight aggregate concrete made with natural pumice. Constr. Build. Mater. 25
(5), 2458–2464 (2011)
Campione, G.: Flexural and shear resistance of steel fiber-reinforced lightweight concrete beams.
J. Struct. Eng. 140(4), 04013103 (2014)
Dias-da-Costa, D., Carmo, R., Graça-e-Costa, R., Valença, J., Alfaiate, J.: Longitudinal
reinforcement ratio in lightweight aggregate concrete beams. Eng. Struct. 81, 219–229 (2014)
durability-related properties of lightweight concrete reinforced with steel fibres. Mater. Struct.
50(1), 46 (2017)
Lytag: Ramboll Frame Comparison Study (2014). https://www.aggregate.com/our-businesses/
lytag, Accessed 31 Dec 2019
Chen, H., Huang, C., Tang, C.: Dynamic Properties of Lightweight Concrete Beams Made by
Sedimentary Lightweight Aggregate. J. Mater. Civ. Eng. 22(6), 599–606 (2010)
Badogiannis, E., Kotsovos, M.: Monotonic and cyclic flexural tests on lightweight aggregate
concrete beams. Earthquakes Struct. 6(3), 317–334 (2014)
Lim, H.S., Wee, T.H., Mansur, M.A., Kong, K.H.: Flexural behavior of reinforced lightweight
aggregate concrete beams. In: Asia-Pacific Structural Engineering and Construction
Conference. Kuala Lumpur: APSEC, pp. 68–82 (2006)
Wu, C., Kan, Y., Huang, C., Yen, T., Chen, L.: Flexural behavior and size effect of full scale
reinforced lightweight concrete beam. J. Mar. Sci. Technol. 19(2), 132–140 (2011)
lightweight concrete. Cem. Concr. Compos. 19(4), 307–313 (1997)
confined with transverse steel reinforcement. Cem. Concr. Compos. 26(6), 645–656 (2004)
beam–column joints. Eng. Struct. 59, 261–283 (2014a)
Di Prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in fib Model Code 2010:
Ritchie, A., Kayali, O.: The effects of fiber reinforcement on lightweight aggregate concrete. In:
Neville, A. (ed.) Proceedings of RILEM Symposium on Fiber Reinforced Cement and
Concrete, The Construction Press Ltd, pp. 247–256 (1975)
Swamy, N., Jones, R., Chiam, A.: Influence of steel fibers on the shear resistance of lightweight
concrete i-beams. ACI Struct. J. 90(1), 103–114 (1993)
Kang, T., Kim, W.: Shear Strength of Steel Fiber-Reinforced Lightweight Concrete Beams,
pp. 1386–1392. Korea Concrete Institute, Oklahoma (2010)
Iqbal, S., Ali, A., Holschemacher, K., Bier, T.: Mechanical properties of steel fiber reinforced
high strength lightweight self-compacting concrete (SHLSCC). Constr. Build. Mater. 98,
325–333 (2015)
Habbitt, Karlsson and Sorensen Inc. Abaqus User’s Manual, vol. II, version 6.1, pp. 11.5.1.1–
11.5.1.14, USA (2000)
Lok, T.-S., Xiao, J.R.: Flexural strength assessment of steel fiber reinforced concrete. J. Mater.
Civ. Eng. 11(3), 188–196 (1999)
Kotsovos, M.D., Pavlovic, M.N.: Ultimate Limit-State Design of Concrete Structures: A New
Approach, pp. 31–164. Thomas Telford, London (1999)
Mater. 178, 484–494 (2018)
The Concrete Society. Guidance for the Design of Steel-Fibre-Reinforced Concrete. Technical
Report No. 63. Cement and Concrete Industry (2007)
Lytag, : Technical manual. Lytag Ltd, London (2011)
Zienkiewicz, O.C., Taylor, R.L.: The Finite Element Method for Solid and Structural Mechanics,
6th edn. Butterworth-Heinemann, Oxford, UK (2005)
Tlemat, H., Pilakoutas, K., Neocleous, K.: Stress-strain characteristic of SFRC using recycled
fibres. Mater. Struct. 39, 365–377 (2006)
Syed Mohsin, S.M.: Behaviour of Fibre-reinforced Concrete Structures under Seismic Loading.
PhD thesis, Imperial College London, London, UK (2012)
Abbas, A., Syed Mohsin, S., Cotsovos, D., Ruiz-Teran, A.: Shear behaviour of steel-fibre-
reinforced concrete simply supported beams. Proc. Inst. Civ. Eng.-Struct. Build. 167(9), 544–
558 (2014b)
Behinaein, P., Cotsovos, D.M., Ali, A.A.: Behaviour of steel-fibre-reinforced concrete beams
under high-rate loading. Comput. Concr. 22(3), 337–353 (2018)
Cunha, V., Barros, J., Sena-Cruz, J.: Tensile behavior of steel fiber-reinforced self-compacting
concrete. In: Fiber-Reinforced Self Consolidating Concrete: Research and Applications (ACI
SP-274), Detroit, USA: American Concrete Institute, pp. 51–68 (2010)
RILEM TC 162-TDF.: r-e design method: final recommendation. Mater. Struct. vol. 36,
pp. 560–567 (2003)
Barros, J., Cunha, V., Ribeiro, A., Antunes, J.: Post-cracking behaviour of steel fibre reinforced
(2005)
Zhang, Y.J., Huang, Z.J., Yanga, S.L., Xua, X.W.C.: A discrete-continuum coupled finite
element modelling approach for fibre reinforced concrete. Cem. Concr. Res. 106(2018), 130–
143 (2018)
Ngo, D., Scordelis, A.C.: Finite element analysis of reinforced concrete beams. J. ACI 64(3),
152–163 (1967)
Rashid, Y.R.: Ultimate strength analysis of prestressed concrete pressure vessels. Nucl. Eng. Des.
7(4), 334–344 (1968)
De Montaignac, R., Massicotte, B., Charron, J.-P., Nour, A.: Design of SFRC structural
elements: post-cracking tensile strength measurement. Mater. Struct. 45(4), 609–622 (2012)
Al-Naimi, H., Abbas, A.: Ductility of steel-fibre-reinforced lightweight concrete. In: Eccomas
Proceedia COMPDYN (2019), Crete, Greece, pp. 4009–4023, 24–26 June 2019. www.
eccomasproceedia.org, Accessed 31 Dec 2019
Kodur, V., Solhmirzaei, R., Agrawal, A., Aziz, E.M., Soroushian, P.: Analysis of flexural and
shear resistance of ultra-high performance fiber reinforced concrete beams without stirrups.
Eng. Struct. 174, 873–884 (2018)
Experimental Analysis of Beams Produced
in Self-compacting Concrete Reinforced
with Different Contents of Steel Fibres
Ana R. L. Pires(&), Rafael R. Polvere, Sidiclei Formagini,

FAENG/UFMS, Federal University of Mato Grosso do Sul,

Campo Grande, Brazil
anaraquellealpires@gmail.com
Abstract. The insertion of steel fibres into the concrete matrix can provide
improvements in the control of crack propagation and increase on stiffness of
structural elements. This research aims to evaluate the behavior of reinforced
concrete beams in different percentages of carbon steel fibres subjected to
bending. So, three reinforced beams were made with two 12.5 mm bars on the
lower edge, in the dimensions of (12.5 25.5 180 cm), which were tested
by four-point flexural test. One beam was used as reference, without the pres-
ence of steel fibres, and another’s were added carbon steel fibres in different
volumetric fractions, 0.5% and 1.0%. During the flexural tests, measurements
were made of the vertical displacements in the middle of the span, also the
deformations in the longitudinal reinforcement and in the compressed region of
the concrete. Results showed that the insertion of the metal fibres gave a better
performance on crack control, lower deflections and deformations on the
compressed region and longitudinal armour of the concrete as well as promoted
slight gain in the load capacity of the beam.
Keywords: Fibrous self-compacting concrete Reinforced concrete beams

Four-point flexural
1 Introduction
The use of fibrous self-compacting concrete seeks to improve the properties of con-
ventional concrete (CC). In its fresh state, self-compacting concrete (SCC) provides
greater functionality to the material due to its fluidity and moderate viscosity. The fresh
mixture has the ability to fill all the spaces of a mold in a uniform manner, under its
own weight without the need for vibration [1]. Thus, there is a more uniform dispersion
of the fibers in the structural elements, in addition to a better alignment of the fibers
along the flow of fresh concrete [2]. The SCC has a structure denser than a CC, due to
the large amount of fines material. Thus, there is a reduction in the presence of voids to
the point of providing a better adhesion between concrete and steel [3, 4].
The introduction of steel fibres into the SCC provides more benefits and application
possibilities, which makes the material more efficient both in the fresh and hardened
state. The enhanced properties with the incorporation of fibers are tensile mechanical
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_66
746 A. R. L. Pires et al.
parameters, ductility and energy absorption capacity under impact [5, 6]. In addition,
this incorporation can delay the propagation of cracks, which provides a longer life and
durability to the material [7]. Steel fiber reinforced concrete (SFRC) is commonly used
in structural engineering, in road pavements and lining.
The fibers exert their main effect in the state of post-crack, intercepting the pro-
gressions of the microcracks, in order to avoid a sudden rupture [8]. Connecting
bridges are formed that serve to transfer the tensile forces through the crack. The ability
of SFRC to transfer tensile forces across cracked sections decreases the crack spacing
and increases the tension stiffening, both of which contribute to an improved crack
control [9]. According to Chenkui and Guofan [10], the fibers tend to settle at the
interface between the large aggregate and the mortar matrix, where the crack is more
likely to develop. When steel fibers are added to reinforced concrete, the steel bars are
more protected against external agents since the fibres control the cracking of the
concrete matrix when the traction is requested [11].
There are several studies that evaluate the influence of steel fibres in the cracking of
structural elements. In SFRC, Abdalkader et al. [12] showed that the addition of 1.0%
resulted in about 81% reduction in maximum crack width compared to concrete beam
without fiber. In steel fiber reinforced expanded-shale lightweight concrete (SFRELC)
beams [13], the fibers effectively restricted the extension and opening of cracks. And
also, the flexural capacity of reinforced SFRELC beams increased 23.3% with steel
fiber volume fraction varied from 0% to 1.6%. Yoo and Moon [14] have reported to the
implication of steel fibers on the flexural behavior of reinforced concrete beams with
reinforcement ratios below the minimum reinforcement ratio and first evaluates fea-
sibility of replacing some of conventional steel rebars to discontinuous steel fibers.
They observed that by including steel fibers, the crack propagation into the compres-
sive zone was effectively limited and more flexural cracks, indicating better stress
redistribution, were formed by the steel fibers. In SFRC beams subjected to bending,
Ashour et al. [15] observed their mechanical behavior and verified that with the
increase in the volume of steel fibres, they became larger: the moment of cracking of
the beam; the moment of beginning of flow of the longitudinal reinforcement; and, the
last moment. OH [16] studied the behavior of reinforced concrete beams with addition
of fibrous reinforcement, varying the volume of steel fibres from 0 to 2%, achieving
better results of ultimate resistance to bending, ductility and the ability to absorb energy
with the increase of the amount of fibres. In real-scale beams, Meda et al. [17] have
evidenced that the fibres significantly enhance the behavior at service conditions by
increasing the stiffness in the cracked-stage and, therefore, by limiting the crack
openings and deformations.
The aim of this article is to evaluate the influence of the different content addition of
carbon steel fibres in the self-compacting concrete, in relation to the mechanical per-
formance of beams submitted to simple bending by means of an experimental analysis.
Special attention is given for the general ductility and distribution of the cracks of
beams. In addition, the deflections and deformations in the longitudinal bars and
compressed region of the concrete are analyzed for each beam according to the applied
load.
Experimental Analysis of Beams 747
In this study, three steel fibers reinforced self-compacting concrete (SFRSCC) beams
with dimensions of 12.5 25.5 180 cm3 (base, height and length) were produced.
The first beam was used as reference (RB), with SCC without the incorporation of steel
fibres and the others (B1 and B2) were made with the presence of fibres in different
volumetric fractions (Vf ) of 0.5% e 1.0%, respectively.
2.1 Materials and Concrete Mixture

The dosage and characterization were obtained in a previous study [18], using normal
Portland cement type CP II 32, silica fume, natural quartz sand, gravel sand, gravel of
basaltic origin (characteristic maximum size of 12.5 mm) and a polycarboxylate based
superplasticizer (FORT FLOW 2558). A water/cement ratio of 0.5 was adopted to
achieve a concrete compressive strength of around 40 MPa.
Steel fibres chosen have hook anchorage at the ends, having a length of 30 mm and
a diameter of 0.60 mm (aspect ratio L/D = 50). The reinforcement steel of the beams
was of type CA-50 and CA-60 for transverse reinforcement, with resistance of 50
kN/cm2 and 60 kN/cm2, respectively. Bars with nominal diameters of 6.3 mm, 5.0 mm
and 12.5 mm were used.
2.2 Reinforced Concrete Beams

First, the preparation of the beams involved cutting and installing the steel reinforce-
ment (Fig. 1). The longitudinal reinforcement of the beams was composed of two steel
bars with a diameter of 12.5 mm (N1) and the transverse reinforcement consists of 14
stirrups made of steel with a diameter of 5 mm (N2). The stirrups were spaced every
10 cm between the supports and the load application, where the shear action is present.
In the central section, between the two loading application points were placed two
stirrups with spacing of 20 cm. In addition, a constructive reinforcement with a
diameter of 6.3 mm (N3) was used.
For the measurement of deformations (steel and concrete), at each point of interest
four electric strain gauges were used as shown in the diagram in Fig. 2. Strain gauges
were placed on the longitudinal bars (E1) and compressed region of concrete (E2).
The strain gauges were placed in the longitudinal and transverse direction (½
Wheatstone Bridge circuit configuration). After bonding, the terminals were connected
to the waiting conductors for later connection to the data acquisition system. Strain
gauges and their terminals have been protected with silicone to prevent direct contact of
the terminals with the metal surface of the armature. In the end, self-fusing tape was
used to harden and increase the insulation resistance, and tape was used to isolate the
strain gauges and avoid contact with water during concreting.
Prior to concreting, the beam molds were lubricated with mineral oil to facilitate the
demoulding process of the parts. The cover of the reinforced beams was ensured by the
use of spacers on the sides and bottom of the beam form. The beams were removed
after 5 days of concreting (during this period only an upper surface remained exposed
Fig. 1. Geometry of the beam (measurements in mm).
Fig. 2. Strain gauges Placement.
to the environment). The concrete beam remained exposed to relative laboratory

conditions. Four days before the test, the joint compound was applied to the faces of
each beam to obtain a better response to application requests.
In addition to the beams, for each concrete produced were moulded: cylindrical
specimens (/10 20 cm) for simple compression tests, tensile compression diametral
and modulus of elasticity; and, prismatic specimens with dimensions of 15 15
50 cm3 for performing four-point flexural tests.
2.3 Four-Point Flexural Test

To study the mechanical behavior of beams (RB, B1 and B2) with different percentages
of steel fibers, a four-point bending test was performed. The static scheme used sim-
ulates a bi-supported beam configuration in which the four-point bending test beams
had a total length of 180 cm and a clearance of 5 cm between the support shaft and the
end, with a clearance of 170 cm. Concentrated loads were applied equidistant at
56.5 cm from the supports (Fig. 3).
Fig. 3. Static scheme.
Fig. 4. Illustration of flexural test (measurements in cm).

In each beam, a 5.0 5.1 cm checkered mesh was demarcated on one of the side
faces. The strain gauges of the compressed concrete region were glued at 5,1 cm from
the upper edge, 72 h before the flexural test. The beams were tested by a four-point
bending test, with load applied to the thirds of the span, checking the evolution of the
cracks, reading the deflection by dial gauges and reading the deformations using strain
gauges. Deformations, deflections and loads were recorded on HBM’s QuantumX data
acquisition system.
The method of the test was divided into the following steps: installation of support
devices, concrete beam positioning, transmission beam support devices installation,
transmission beam positioning, hydraulic piston positioning, load cell positioning,
plumb check, marking and positioning of measuring devices. The position of the
instrumentation used in the assay can be identified in the illustration of Fig. 4. Vertical
loads were incrementally applied by two hydraulic pistons installed in a reaction frame.
The marking of the cracks on the side faces of the beam was performed using load
increments of approximately 3 kN.
3.1 Mechanical Properties of Concrete Mixtures

Table 1 shows the mechanical properties of concrete used at 28 days from a previous
study [18], referring to the simple compression tests - fcm, splitting test tensile - fct,sp,
elastic modulus - Em, and tensile strength tests at flexion - fct,f. The concrete nomen-
clature are SCC (standard concrete), SFRSCC 0.5% (0.5% of steel fibres in volume
addition) and SFRSCC 1.0% (1.0% of steel fibres in volume addition).
Table 1. Mechanical properties of the concrete used at 28 days.

Properties SCC SFRSCC0.5% SFRSCC 1.0%
fcm [MPa] 41.8 45.6 41.5
fct,sp [MPa] 3.69 5.71 5.06
Em [MPa] 26.7 30.5 29.4
fct,f [MPa] 3.79 3.87 5.15
In concretes with fibre incorporation there was an increase of the modulus of

elasticity and tensile strength by diametral compression because the fibers addition can
limite the transverse strain of concrete. The results of the compression tests indicated a
decrease in the simple compressive strength for SFRSCC 1.0%. It is also observed that
the tensile strength in prism bending is greater the higher the content of steel fibres.
This can be explained by the fact that with a higher content there is a greater number of
steel fibres acting as a stress transfer bridges along the crack, which increases the post-
cracking reinforcement of the concrete.
3.2 Cracks Patterns and Ultimate Strength

The incorporation of steel fibres in the self-compacting concrete increased the ultimate
loads (Pu .), as well as the loading of the first crack (Pcrack .) of beams B1 and B2.
Figure 5 illustrate the evolution of cracks in each beam tested. In beams B1 and B2, the
flexural test was terminated shortly before the rupture. As a result, the load registered at
the end of the test was considered as the last load of the beam, since it no longer
presented growth.
(a) RB
(b) B1 (c) B2
Fig. 5. Evolution of cracks in the beams.
In all three samples the first crack occurred in the central region, between the
applied loads where the bending is maximum. The rupture of the RB beam occurred in
a ductile form by flexion, with the crushing of the concrete in the compressed side and
another side with yield strain of the reinforcement flow in the tensioned region. At the
moment of rupture, there was a greater presence of multiple cracks in relation to the
beams with the addition of metallic fibres. In addition, significant typical shear cracks
appeared in the region between the concentrated loads and the supports. The inclined
cracks almost reached the upper edge of the beam.
It can be noticed that cracks result smaller with increasing fiber content; they
become more diffused, also with few shear inclined cracks. It’s noted that in steel fibre
reinforced beams, the propagation velocity of the cracks decreases in the concrete,
which makes the material, according to [19] exhibit a pseudo-ductile behavior, that is,
it presents a certain bearing capacity post-cracking. The reduction of the inclined cracks
in the fibrous concrete beams may indicate an action of these fibres in the absorption of
part of the shear, sewing these cracks that appear in the beam regions between the load
and the support. It’s also visually observed that the crack spacing is also influenced,
becoming slightly larger as the fiber volume is increased.
3.3 Load-Deflection Relationship

The experimental results, in terms of load (P) versus deflection for mid-span mea-
surements of self-compacting without fibres concrete beams are shown in Fig. 6. And
also, Table 2 shows the deflections at the critical points - cracking (dcrack ), yield (dy )
and ultimate (du ).
140
120
100
Load P(kN)
80
60
RB (SCC)
40
B1 (SFRSCC 0.5%)
20
B2 (SFRSCC 1.0%)
0
0 5 10 15 20
Deflection (mm)
Fig. 6. Load-Deflection Curves of the beams.
The B1 beam had a last load 4.3% higher than the RB beam made without the
addition of steel fibres. And B2 beam indicated a 10.3% increase in the last load. This
shows that the increase in ultimate load was proportional to the increase in the amount
of steel fibres used. This shows that the increase in ultimate load was proportional to
the increase in the amount of steel fibres used. In relation to the appearance of the first
crack, the B2 beam showed an approximately 63% increase in the first crack load
compared to the crack initiation load applied to the RB beam. The B1 beam presented a
cracking load 5.3% lower than the RB. In addition, there are increases of 8.3% and
23% of the yield load for RB beam compared to beam B1 and B2, respectively.
Table 2. Deflections and loads at critical points.

Beams Pcrack (kN) dcrack (mm) Py (kN) dy (mm) Pu (kN) du (mm)
RB 19 1.0 97 8.7 116 12
B1 18 0.85 105 8.0 121 9.7
B2 31 1.20 119 7.85 128 18
The concrete beam with higher volumetric fraction of steel fibres indicated greater
contributions to reduce tensile stresses in reinforcement bars.
It is noted that B2 presents its first crack with a deformation 41.2% greater than that
of RB beam. And, the ultimate load of the reference beam, occurred when the beam
had deflection of approximately 12 mm, while in B2 for the last load the deflection was
of approximately 18 mm, which represents an increase of 50%. B2 beam presented
more ductile behavior, corresponding to a greater ability to absorb energy.
3.4 Load-Deformation Relationship

Figures 7 and 8 show the curves obtained with the measurements of the compressed
concrete deformation in the beams and deformation in the longitudinal bars,
respectively.
140
120
100
Load P(kN)
80
60
RB (SCC)
40
B1 (SFRSCC 0.5%)
20
B2 (SFRSCC 1.0%)
0
-1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0
Deformation [‰]
Fig. 7. Load-Deformation Curves of the compressed concrete region.
Through the analysis of the curves obtained with the measurements of the com-
pressed concrete deformation in the beams (Fig. 7), it was found that the incorporation
of steel fibres in the self-compacting concrete gave the beams B1 and B2 smaller
deformations in the compression region for the beams with the same load level when
compared to RB. This can be explained by the fact that the modulus of elasticity
presented a lower value for the concrete of the RB beam, as presented in Table 1. It’s
observed that the higher the fiber percentage, the better the concrete behavior in the
140
120
Load P(kN) 100
80
60
RB (SCC)
40
B1 (SFRSCC 0.5%)
20
B2 (SFRSCC 1.0%)
0
0 1 2 3 4 5 6
Deformation [‰]
Fig. 8. Load-Deformation Curves of longitudinal bars.
compressed region. The effect of fiber addition was significant as beam B2 withstood
greater deformation with higher loading than other beams. It’s noteworthy that the
strain gauges fixed in the concrete detached at the end of the bending test on the beams
RB and B1.
The results of the measurements made in the longitudinal reinforcement (Fig. 8)
differed from some indications in the literature and other experimental studies, since the
deformations in steel are greater for the fiber reinforced beams in the initial phase given
the same loading level. However, in the final phase of the carrying capacity, beams B1
and B2 achieved higher load capacities for the same longitudinal bar deformation. This
may indicate that part of the tensile stresses were absorbed by the fibres immersed in
the concrete, which absorb the tensile stresses during bending, thus relieving the action
of the main reinforcement.
4 Conclusions
In this research, the beams made with the incorporation of fibres presented a better
structural behaviour. The following conclusions can be drawn, based on the analysis of
the experimental results:
• A greater amount of load was achieved in the fibrous concrete beams, with less
deformation in the compressed concrete region. Smaller deflections mid-span to the
same load level on beams B1 and B2. In addition to lower deflections mid-span to
the same loading level.
• The post-cracking behaviour in the beams with the presence of fibres showed
greater ductility, absorbing more energy. Factors that may be interesting for the
design of structures subjected to earthquake and dynamic load.
• The self-compacting fibrous concrete beams presented a better crack control, with
higher first crack load for the B2 beam. Visually, a better distribution of these cracks
is also observed, with smaller openings. And the beam of 1.0% of fibres showed a
significant increase of 10.3% higher than the reference beam in the final load.
The costs of steel fibre self-compacting concrete structures and flexural rein-
forcement may be higher, but overall, with the use of steel fibres you can get great
advantages such as: lower maintenance cost and longer durability the structure.
References
1. Okamura, H., Ouchi, M.: Self-compacting concrete. J. Adv. Concr. Technol. 1, 5–15 (2003)
2. Vandewalle, L., Heirman, G., Van-Rickstal, F.: Fiber orientation in self compacting fibre
reinforced concrete. In: Proceedings of the 7th International RILEM Symposium on Fibre
Reinforced Concrete: Design and Applications (2008)
3. Almeida Filho, F.M.: Contribuição ao estudo da aderência entre barras de aço e concretos
autoadensáveis. In: São Carlos, Tese (doutorado)–Escola de Engenharia de São Carlos,
Universidade de São Paulo, p. 310 (2006)
4. Hossain, K.M.A., Lachemi, M.: Bond behavior of self-consolidating concrete with mineral
and chemical admixtures. J. Mater. Civ. Eng. ASCE 20(9), 608–616 (2008)
5. Katzer, J., Domski, J.: Quality and mechanical properties of engineered steel fibres used as
reinforcement for concrete. Constr. Build. Mater. 34, 243–248 (2012)
6. Paja˛K, M., Kühn, T.: The influence of steel fibers on the dynamic response of self
compacting concrete. Key Eng. Mater. 711, 179–186 (2016)
7. Falkner, H., Huang, Z., Teutsch, M.: Comparative study of plain and steel fibre reinforced
concrete ground slabs. Concr. Int. 17(1), 45 (1995)
8. Sorelli, L., Meda, A., Plizzari, G.A.: Steel fibre concrete slabs on grade: a structural matter.
ACI Struct. J. 103(4), 551 (2006)
9. Bischoff, P.: Tension stiffening and cracking of steel fiber-reinforced concrete. J. Mater. Civ.
Eng. 15(2), 174–182 (2003)
10. Chenkui, H., Guofan, Z.: Properties of steel fibre reinforced concrete containing larger
coarse aggregate. Cem. Concr. Compos. 17, 199–206 (1995)
11. Noghabai, K., Noghabai, K.: Effect of various types of fibers on bond capacity–
experimental, analytical, and numerical investigations. Struct. Appl. Fiber Reinforced 182,
109 (1999)
12. Abdalkader, A., Elzaroug, O., Abubaker, F.: Flexural cracking behavior of steel fiber
reinforced concrete beams. J. Sci. Technol. Res. 06(08), 6 (2017)
13. Zhao, M., Li, C., Su, J., Shang, P., Zhao, S.: Experimental study and theoretical prediction of
flexural behaviors of reinforced SFRELC beams. Constr. Build. Mater. 208, 454–463 (2019)
14. Yoo, D., Moon, D.: Effect of steel fibers on the flexural behavior of RC beams with very low
reinforcement ratios. Constr. Build. Mater. 188, 237–254 (2018)
15. Ashour, S.A., Wafa, F.F., Kamal, M.I.: Effect of concrete compressive strength and tensile
reinforced ratio on the flexural behavior of fibrous concrete beams. Eng. Struct. 22, 1133–
1146 (2000)
16. Oh, H.B.: Flexural analysis of reinforced concrete beams containing steel fibers. J. Struct.
Eng. ASCE 118(10), 18 (1992)
17. Meda, A., Minelli, F., Plizzari, G.: Flexural behaviour of RC beams in fibre reinforced
concrete. Compos. Part B: Eng. 43(8), 2930–2937 (2012)
18. Polvere, R.R., Pires, A.R.L., Formagini, S., Cheung, A.B.: Mix Design and Properties of
Self-Compacting Fibrous Concrete. RILEM-fib X International Symposium on Fibre
Reinforced Concrete, Valencia (2020). (in press)
19. Figueiredo, A.D.: Concreto com fibras de aço. Boletim Técnico da Escola Politécnica da
USP. Departamento de Engenharia de Construção Civil e Urbana. BT/PCC/260. São Paulo
(2000)
Innovation in Durable
Segments for CSO Tunnels
Ralf Winterberg1(&), Michael R. Garbeth2, and Brian Glynn3

1
BarChip Inc., Tokyo, Japan
rwinterberg@barchip.com
2
Super Excavators Inc., Menomonee Falls, USA
3
Black & Veatch Corporation, Kansas City, USA
Abstract. The Blacksnake Creek and storm water runoff in St. Joseph, MO,
was piped along with sewage in a 100-year old pipe not large enough to carry all
the storm water and sewage to the wastewater treatment plant and it overflowed
to the Missouri River after most rainstorms. The Blacksnake Creek Storm Water
Separation Project will convey storm water directly to the Missouri River. This
will reduce water quantity in the existing sewer during storms and the quantity
of combined storm water and wastewater overflowing to the river. A 2.74 m ID
and 2.0 km long segmental tunnel was constructed as part of the Separation
Project. This project is an America’s First, using segments solely reinforced with
macro synthetic fibre (MSF). The project further features a challenging umbil-
ical TBM launch in a small diameter secant pile shaft. This paper addresses the
solutions to the technical challenges of the project, the design of the segments
and the benefits associated with the use of MSF.
Keywords: Macro synthetic fibre Fibre reinforced concrete Segmental

lining TBM Utility tunnel Sewer tunnel
1 Introduction
The Blacksnake Creek Storm Water Separation Improvement Project was required as
part of the City of St. Joseph, Missouri’s Combined Sewer Overflow (CSO) Long Term
Control Plan in order to improve water quality as mandated by the Federal Clean Water
Act. The Blacksnake Creek used to be directed into the City’s combined sewer system
through a double box culvert, and the creek flow was conveyed through the sewer
system to the Water Protection Facility and unnecessarily treated 365 days of the year.
During wet weather events, storm water runoff exceeded the capacity of the combined
sewer system and caused combined sewer overflows to the Missouri River. The
overflows were a mix of storm water and sanitary sewage and resulted in adverse water
quality problems. To combat the discharge, the project intercepts and redirects
Blacksnake Creek stream flows away from the City’s combined sewer system to a new
and dedicated storm water conveyance system that flows directly to the Missouri River,
thus reducing the frequency, volume, and impacts of combined sewer overflows to the
river. The project was designed by Black & Veatch, Kansas City, MO.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_67
758 R. Winterberg et al.
The tunnel alignment and profile, shown in Fig. 1 and Fig. 2 respectively, heads
west from the Drop Shaft located near the Second Harvest Food Bank, following
underneath the Highland Avenue right of way, and curves slightly to the south as it
nears Interstate 229. The alignment crosses below both Interstate 229 and the
Burlington Northern Santa Fe (BNSF) Railroad Tracks before terminating at an Energy
Dissipation Structure. From the Energy Dissipation Structure, flows are conveyed by
Roy’s Branch, which is a tributary to the Missouri River.
Fig. 1. Tunnel alignment - plan view
The tunnel alignment is approximately two kilometres in length and terminates at a

junction structure located between MacArthur Drive and the BNSF Railroad Tracks.
The TBM launch shaft is located at the western end of the alignment, near the BNSF
Railroad Tracks.
Fig. 2. Tunnel alignment - profile view

Innovation in Durable Segments for CSO Tunnels 759
The USD 27 million Tunnel Bid Package (TBP), awarded to the contractor Super
Excavators, Inc., was a critical portion of the project, as the topography of the area did
not allow for construction of a new and dedicated storm water canal via traditional
trench excavation without significant impacts to the public. The TBP not only com-
prised the segmental tunnel, but also an 11 m ID baffle drop structure, a 15 m rein-
forced concrete box culvert and an energy dissipation structure at the Missouri end of
the alignment.
2 Geological Conditions
The tunnel was to be constructed in soft ground and mixed ground conditions,
including soils and shale, and combinations of both at the interfaces. The soils to be
encountered varied from silty clay, silty sand, clayey sand, and sandy clay. The shale
encountered is low in strength, has low to medium-low durability, is easy to excavate,
and has low hydraulic conductivity. Groundwater along the tunnel varied along with
the topography from 12 m to 45 m of groundwater head.
The general geography and topography of the project site is located in north-
western St. Joseph, Missouri between St. Joseph Avenue and the Missouri River. The
overburden along the tunnel alignment consists primarily of fill material and alluvial
deposits with alternating beds of silt, clay, silty-clay, and clayey silt. The minimum
overburden on the project occurs at the launch shaft with approximately 9.0 m of cover.
Progressing away from the launch shaft at the peak of the topography on Highland
Avenue between Main Street and 2nd Street there is 55 m of cover and at the receiving
shaft near St. Joseph Ave. this reduces to 18 m. The primary rock conditions pre-
dominantly consisted of shale and claystone along the tunnel horizon. A limestone bed
was present above the tunnel crown for approximately 1,280 m. The percentage of the
alignment for the various different ground conditions consisted of soft-ground (23%),
rock (63%), and mixed-ground for 14% of the alignment.
The tunnel transitioned between soft ground and rock along each end of the tunnel
alignment and mixed-faced conditions were anticipated for reaches along these sub-
surface transitions. The soil to rock transitions on each end of the alignment were
gradual. It was anticipated that the tunnel groundwater inflow will not exceed a steady
state of flow greater than 50 GPM (gallons per minute, approx. 190 litres per minute).
Figure 3 shows the main parts of TBM Carrie. Considering the ground conditions
along the tunnel alignment and the contract specifications, the tunnel had to be exca-
vated with an EPB tunnel boring machine in order to mitigate risks during tunnelling in
soft or mixed ground below the water table.
Fig. 3. EPBM ‘Carrie’ assembly and testing on site
3 TBM Launch
A cradle was placed inside the secant pile launch shaft prior to the arrival of the TBM.
This cradle consisted of two I-Beams cast in a concrete floor to fit the curvature of the
machine. All the shaft and tunnel utilities were established inside the shaft prior to
lowering any machine components. A special electrical and hydraulic umbilical
assembly was designed and used to help launch the TBM. Note the secant pile launch
shaft diameter is 15 m where the entire TBM assembly is 78 m in length.
The entire machine was methodically aligned in two rows consisting of Row #1,
Main TBM Parts: cutter head, telescopic shield, gripper shield, and tail skin; and Row
#2, Backup System: all the gantries and backup equipment including the electrical
substation (Fig. 4 left). The TBM cutter head, gripper shield and telescopic shields
Fig. 4. TBM Carrie’s main parts (left) and launch cradle in the secant shaft (right)
were nearly completely assembled at the surface due to the ease of assembly at the
surface versus in the shaft. A 500 ton mobile crane was mobilized to the project site and
used to lower the heavy components into the shaft.
Once the major components were lowered down into the shaft, one crew began
hooking up the electrical and hydraulic lines while another crew continued to work on
assembly of the gantries. The gantries included such items as: the operator’s cabin,
electrical substation, air compressor, main drive motors, grease pumps, ground con-
ditioning system, ventilation cassette and cable festoons.
The special hydraulic and electrical umbilicals were used to link the gantries on the
surface to the main TBM body in the launching cradle on the bottom of the shaft. At
this point of time, the TBM could begin the functional testing and troubleshooting
process.
After finishing assembly of the initial configuration for the machine, the excavation
has commenced. During this process, the primary propulsion cylinders pushed off of
beams welded to the I-Beam in the cradle. This allowed for the cutter head, telescopic
shield and gripper shield to advance forward and penetrate the secant shaft. A short
screw conveyor was utilized to control the material in front of the TBM (Fig. 4 right).
After the TBM was launched, and most of the components were installed and in the
ground behind the TBM, the next step was to disconnect the complete electrical and
hydraulic cables/hoses, which allowed the following tasks to be completed:
• Remove the short launch screw conveyor
• Install the longer main screw conveyor
• Install the tail shield
• Install the segment transporter belly pan
• Re-hook up the electrical and hydraulic cables/hoses.
During this stage, a precast concrete segmental tunnel liner “half-ring” structure
was used to push and advance the TBM further into the excavation (Fig. 5 right).
Once the tailskin reached the secant shaft wall the full ring installation was ini-
tialized. A support bracket was braced off the secant piles to hold the first segment ring
in place while the TBM advanced forward. The gantries were lowered down into the
shaft one by one until all nine were advanced directly behind the TBM. Then, full
production tunnelling began.
The initial tunnel drive installation has proven that the fibre reinforced segments are
robust enough to withstand the temporary load cases, i.e. transportation, hoisting,
handling and installation, without experiencing damage to the segments. In addition,
the half-ring segment sections used in the shaft for the launching process (Fig. 5 right)
were inspected after the initial push and no damage occurred to the segments.
Fig. 5. Launching the TBM (left) and half-ring installation for TBM propulsion (right)
4 Segmental Lining
4.1 Introduction
In the past 20 years Fibre Reinforced Concrete (FRC) has become widely utilized in
segmental tunnel linings due to the improved mechanical performance, robustness and
durability of the segments. Further, significant cost savings can be achieved in segment
production and by reduced repair or reject rates during temporary loading conditions
[1]. The replacement of traditional rebar cages with fibres further allows changing a
crack control governed design to a purely structural design with more freedom in
detailing.
Macro Synthetic Fibres (MSF) are non-corrosive and thus ideal for segmental
linings in critical environments. The market share of FRC tunnel segments compared to
traditionally reinforced segments continuously grows. Recent publications such as the
ITA WG2 Report [2], the ITAtech guideline [3], the British PAS 8810 [4], and the FIB
state-of-the-art report [5] have given more credibility to the use of this reinforcement
type as well as its basis for design to support its application in tunnel lining segments.
The use of MSF in tunnel segments has been increasing on a global scale in the
tunnel industry, mainly for the durability benefits that the synthetic fibres provide, as
compared to steel fibres or rebar cages. A major reference for the use of MSF as sole
structural reinforcement for precast tunnel segments is the Santoña–Laredo General
Interceptor Collector in Northern Spain [6]. This project demonstrated very robust and
satisfactory performance of the MSF reinforced segments even under difficult condi-
tions [1].
The tunnel industry in the US has also started to embrace this technology. Based on
the successful completion of the final lining of the starter and end tunnels, and suc-
cessful trials with tunnel segments for the Euclid Creek tunnel in the NEORSD Clean
Lake Project [7], the precast concrete segment manufacturer proposed to use MSF for
the Blacksnake tunnel project.
4.2 Ring Geometry and Segmentation

The lining for the Blacksnake Creek Storm Water Separation Tunnel consists of a
precast concrete segmental tunnel lining. The segmental lining is composed of six
trapezoidal fibre reinforced concrete segments with rubber gasket frames inserted
between the segments and adjacent rings to create a watertight lining. When assembled,
the six trapezoidal segments form a ring with an internal diameter of 2743 mm (Fig. 6).
Fig. 6. Ring segmentation (left) and tunnel ring assembly (right)
The universal rings of 1219 mm nominal length have a taper of 12.7 mm in order
to accommodate a turning radius of 300 m on the alignment. Kinematic control during
installation is provided by means of dowels and alignment indicators on the segments.
The radial joints are connected with galvanized steel bolts.
The precast concrete segments were produced by CSI Tunnel Systems Inc.’s plant
in Macedonia, Ohio, and were shipped by truck to the project site in St. Joseph,
Missouri (Fig. 7). Segments were installed within the tunnel during the excavation
process by completing a segment ring with each advancement. Each individual seg-
ment was placed using a mechanical segment erector located in the trailing shield of the
TBM, and the segments were manually bolted together. Annular grouting was per-
formed through grout ports built into each segment after the rings were erected.
BarChip 54 Macro Synthetic Fibre was used to reinforce the precast concrete
segments as an alternative to steel fibre or traditional deformed steel bar reinforcement.
This alternative was chosen to reduce material costs while ensuring compliance with
American Iron and Steel provisions of the Contract Documents. The use of synthetic
Fig. 7. MSFRC segments in the Macedonia factory (left) and storage on site (right)
fibre reinforcement further eliminates the risk of corrosion and increases the service life
of the tunnel lining. This, in turn, reduces future maintenance cost of the tunnel. The
Blacksnake tunnel is the first tunnel in North America to be lined with precast concrete
segments that are solely reinforced with synthetic fibre reinforcement.
To better accommodate the fibre reinforced concrete solution, smaller segments
were adopted in order to limit the segments’ aspect ratio. The tunnel has an internal
diameter of 2.743 m with a segment thickness of 190.5 mm. The segmentation was
selected to be six trapezoidal segments (Fig. 6), each having a developed centre arc
length of 1.536 m. This yields a segment aspect ratio of 8.1, which is well below the
acknowledged limit of 10 to ensure segment robustness for temporary load cases [3].
4.3 Lining Structural Design

The MSFRC segments are made using concrete class f’c 7.0 ksi (48 MPa) at 28 days of
age, with a specified stripping strength of f’c 2.0 ksi (14 MPa). The required residual
strength was specified to be minimum 3.2 MPa at any deflection at or beyond span/600
(0.75 mm through 3.0 mm) according to ASTM C1609 [8]. That means the same high
residual strength value was specified for both ULS and SLS in order to guarantee crack
width control in service as well as ultimate residual strength.
BarChip 54 is the chosen fibre, based on the successful experience from the full
scale segment trials carried out in the Euclid Creek tunnel project [7]. A dose rate of
7.0 kg/m3 of this fibre proved sufficient in the pre-production trials to exceed the
3.2 MPa residual strength specification as detailed above.
The design approach adopted for the FRC segmental lining in ultimate limit state
(ULS) is the use of Normal Force-Bending Moment interaction diagrams or Moment-
Thrust Capacity Limit Curves (Fig. 8). The factored design load couples acting on the
section must remain within the N–M envelope [9]. The FRC material properties are
herein derived from the ASTM C1609 beam tests, which are eventually used as the
basis to determine the stress-strain relationship of the concrete on the tension side. The
idealized stress-strain diagram enables setting up the capacity limit curves, which are
obtained by equilibrium iterations on the given cross-section.
Fig. 8. N-M capacity envelope for ULS
All checks for temporary load cases as well as the checks for serviceability limit
states (SLS) have been performed with finite element analysis [10]. These included
numerical analysis of the jacking forces on the segmental lining during TBM forward
thrust (Fig. 9).
Fig. 9. SLS verification using FEA (Atena)
The material model adopted in the FEA employs the increased fracture energy
provided by the fibres [3, 11]. This advanced material model has proven to yield
accurate numerical simulations, e.g. as shown in the numerical analysis of the full scale
segmental ring testing for the Shanghai Metro extension at the Tongji University in
Shanghai [11].
5 Tunnel Excavation
The initial tunnel drive was launched in soft ground, and then transitioned into mixed
ground conditions fairly quickly. To control alignment and grade during the launching
process, and throughout the tunnel drive, a TACS system was implemented. The TACS
system provides continuous information about how the machine axis is aligned with
respect to its designed alignment, and in what direction it is moving. The system also
determines how the six-piece segmental lining will be installed, dependent on the
position of the rings with the taper, which determines the straightness or curvature of
the tunnel drive.
The TBM used a bulkhead type mechanical segment erector inside the tail skin to
erect the six piece segmental ring. It utilized a ball type pick up system, where the ball
head bolt is screwed into a threaded socket in the centre of the segments (Fig. 10 left).
The erector has a safe working load of 130% of the segment weight. The rotational
speed is fully variable even when loaded between 0−2 RPM. The erector is capable of
reaching segments over the two inner most rows of the tail seal brushes.
Fig. 10. Mechanical erector inside the tail skin (left) and completed rings in the tunnel (right)
The excavation of the Blacksnake Creek Tunnel project progressed through three
distinct geologies. These were identified in the Geotechnical Baseline Report as being
soft ground, mixed-ground, and rock conditions. The project tunnelling started out in
soft ground conditions, transitioning into mixed-ground and then rock conditions. As
the tunnel progressed towards the receiving shaft, the ground transitioned back from
rock to mix-ground and then soft ground conditions.
The project team did not experience any significantly challenging tunnelling con-
ditions during the transitions between the soil/rock interfaces. However, the production
rate increased considerably when excavating in rock conditions. The rock conditions
were generally consistent for a longer period of time, thus tunnelling became repetitive
for the crew.
The biggest challenge for the TBM was controlling steering throughout the
alignment in all three distinct geologies due to machine difficulties. The TBM had to be
used in exceptionally high pressure mode to overcome steering complications at var-
ious times. During these times the fibre reinforced segments remained robust and
generally did not step in, crack, or become damaged due to the high pressures. During
recent tunnel inspections though, a few segments with fine longitudinal cracks were
observed, which were likely due to the machine pushing off the segments in that period.
However, the fibres were able to control developing crack widths to remain within the
specified serviceability limits of the segmental lining design (wmax = 0.2 mm). No
segment had to be rejected.
The primary tunnel backfill grout was modified by adding a non-chloride accel-
erator. This accelerated the set time for the tunnel grout and prevented the segments
from floating in the rock section of the alignment. This was important because the
survey brackets were attached to the segment inside the tunnel and prevented them
from moving.
Fig. 11. TBM breakthrough in October 2019

The TBM exited from the rock geology back to mixed and soft ground conditions
without any issues. However, abnormally high ground water pressure was encountered
for a brief period of time in the soft ground. Tunnelling slowed down throughout this
brief section while the team overcame the water pressure. The segments were able to
withstand the additional water pressure and no leakage occurred.
It was anticipated based on the performance of the TBM, robustness of the seg-
ments, and ground conditions that were expected to be encountered, that the
advancement should be between 14.6 to 18.2 meters per 10 hour shift in full pro-
duction excavation. The tunnelling advancement was sporadic throughout the drive
with the best day of production at 19.5 meters per 10 hour shift in the rock conditions.
This equates to installing a bit more than 1.5 rings per hour.
Breakthrough was in October 2019 (see Fig. 11) and scheduled project completion
is spring 2020.
6 Conclusions
• Launching the EPB TBM with an extensive amount of umbilical cords, in a rela-
tively small size shaft compared to the overall length of the TBM, and a unique
launching process has been challenging but a lot of knowledge has been gained for
future launching of this TBM on other projects.
• This extraordinary project confirms that a strong partnership between all parties
involved in this project yields an innovative, technically and economically leading
solution that will deliver a quality project to the owner.
• The use of MSF reinforced concrete segments on the Blacksnake Creek project has
demonstrated robust, durable and dependable performance in tunnel projects even
under very difficult conditions during the TBM launch phase. The durability and
performance of the MSF reinforced segmental lining has outperformed
expectations.
• Ongoing research and continuous developments on macro synthetic fibre and macro
synthetic fiber reinforced concrete have made it today being a modern and cost-
efficient construction material. Eliminating durability issues with regard to corro-
sion of the primary reinforcement yields significant advantages for the design, since
it is no longer governed by serviceability limits.
• The successful completion of this America’s First project is expected to build
further confidence in MSF reinforced segmental linings. These types of utility
tunnelling projects (e.g. sewage and irrigation, power, or gas transfer tunnels) are
widely existing in the world market and they present a huge opportunity for MSF
reinforced concrete linings, benefiting from the proven advantages.
References
1. Winterberg, R., Justa Cámara, R., Sualdea Abad, D.: Santoña–Laredo General Interceptor
Collector – Challenges and Solutions. In: Proceedings FRC 2018, Fibre Reinforced
Concrete: from Design to Structural Applications, Desenzano, Italy, 28–30 June 2018 (2018)
2. ITA WG2 2016. Twenty years of FRC tunnel segments practice: lessons learnt and proposed
design procedure. ITA Report No. 16, ITA Working Group 2 Research, April 2016. www.
ita-aites.org
3. ITAtech 2016. Guidance for Precast Fibre Reinforced Concrete Segments – vol. 1 Design
Aspects. ITAtech Report No. 7, ITAtech Activity Group Support, April 2016. www.ita-aites.
org
4. PAS 8810 2016. Tunnel design – Design of concrete segmental tunnel linings – Code of
practice. The British Standards Institution, BSI Standards Ltd (2016)
5. FIB 2017. Precast Tunnel Segments in Fibre-Reinforced Concrete, State-of-the-art report fib
WP 1.4.1. fib Bulletin 83, Fédération internationale du béton (fib), October 2017
6. Winterberg, R., Mey Rodríguez, L., Justa Cámara, R., Sualdea Abad, D.: Segmental lining
design using macro synthetic fibre reinforcement. In: Proceedings FRC 2018, Fibre
Reinforced Concrete: from Design to Structural Applications, Desenzano, Italy, 28–30 June
2018 (2018)
7. Wotring, D.A., Vitale, M.G., Gabriel, D.A.: Synthetic-fiber-reinforced concrete segmental
lining - laboratory and field testing program and results. In: Proceedings World Tunnel
Congress 2016, San Francisco, USA, 22–28 April 2016 (2016)
8. ASTM C1609/C1609M-12. Standard Test Method for Flexural Toughness of Fiber-
Reinforced Concrete (Using Beam with Third-Point Loading), ASTM International, West
Conshohocken, PA (2012)
9. Nitschke, A., Winterberg, R.: Performance of macro synthetic fiber reinforced tunnel linings.
In: Proceedings World Tunnel Congress 2016, San Francisco, USA, 22–28 April 2016
(2016)
10. JKP 2017. Design report on MSF reinforced concrete segments for the Blacksnake Creek
tunnel. JKP Static, Budapest, Hungary, November 2017
11. Juhasz, K.P., Nagy, L., Winterberg, R.: Full-round numerical analysis of traditional steel bar
and macro synthetic fibre reinforced concrete segments for the shanghai metro extension. In:
Proceedings World Tunnel Congress 2015, Dubrovnik, Croatia, 22–28 May 2015 (2015)
Incorporation of Rate-Dependent Fracture
Properties in the Design of Precast Concrete
Tunnel Segment with Hybrid Reinforcement
Stefie J. Stephen(&) and Ravindra Gettu
Department of Civil Engineering, Indian Institute of Technology Madras,

Chennai, India
stefie.j.s@gmail.com
Abstract. The design of precast tunnel segment under service load condition is
mostly based on the axial force and bending moment (P-M) interaction diagram.
This diagram is usually derived using flexural or tensile parameters obtained
from short-term testing procedure. Considering the fact that the tunnel segments
are subjected to prolonged loading, the effect of the long-term loading rate on
these structures needs to be incorporated. In this paper, the assumption and
model used for deriving the P-M diagram is discussed and then the procedure to
include the rate-dependent tensile constitutive model in the design is presented.
The rate-dependence of the P-M interaction diagram of the tunnel segment with
conventional reinforcement and different type and dosage of fibres is discussed.
Keywords: Fibre reinforced concrete Hybrid tunnel segment Rate-

dependent tensile constitutive model P-M interaction diagram
1 Introduction
Tunnels excavated using the tunnel boring machine (TBM) are most commonly sup-
ported by precast concrete segments, where conventional steel bars (rebars) are pro-
vided as reinforcement (ITA Report, 2016). A combination (hybrid) of rebars and fibre
reinforcement has been used in many cases to withstand the bursting and spalling
stresses developed during the manufacture and erection of segments (Caratelli et al.,
2011; Briffaut et al., 2016). These tunnel linings are expected to last for at least 100
years, taking into consideration the huge initial cost. However, the cracking and
spalling of concrete due to improper grouting, vibrational movement of vehicles,
prolonged high overburden pressure, etc. puts the segments in risk in the long run
(Asakura and Kojima, 2003). Even though the fibres in concrete help in mitigating such
damage (Buratti et al., 2013; Tiberti et al., 2014), consideration of the rate-dependence
of the FRC response could improve the design and performance, as addressed in this
paper. The incorporation of rate effect in the derivation of P-M interaction diagram is
demonstrated for a hybrid reinforced concrete (HRC) tunnel segment with conventional
steel reinforcement and different types and dosages of fibres.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_68
Incorporation of Rate-Dependent Fracture Properties 771
2 P-M Interaction Diagram
The P-M interaction diagram can be developed for an HRC segment with a particular
cross section, if the tensile property of the FRC and the steel rebar and compressive
property of the FRC is known. In this section, the models used, and the assumption
made by Yao et al. (2018) are briefly explained. The behaviour of FRC is simulated by
the stress and strain diagrams for tension and compression using Fig. 1. The tension
model consists of a linear stress-strain diagram up to tensile strength ft, followed by a
post-cracking constant residual tensile stress, rp = µft = µEecr, with parameter µ
(0 µ < 1) representing the residual tensile strength parameter as a fraction of the
tensile strength (for plain concrete, µ will be taken as 0) and ecr representing tensile
cracking strain (ft/E). The parameters of the tensile model are obtained from the flexural
test response of beams, as per RILEM TC 162-TDF (2003). The compressive beha-
viour is simulated by an elastic perfectly plastic model, with a linear stress-strain
diagram up to the yield compressive strength rcy == x ft with parameter x (x > 1).
When the compressive strain is above the yield strain, the stress is maintained constant
up to concrete crushing (ec = ecu). The ultimate compressive strain, ecu, is assumed to
be 0.35% (RILEM TC 162-TDF, 2003). The normalized yield compressive strength x
is the compressive-tensile strength ratio. The compression and tension models termi-
nate at the normalized ultimate compressive strain kcu and tensile strain btu, respec-
tively. For steel rebars, the tensile model is assumed as elastic perfectly-plastic, with a
linear stress-strain diagram up to yield tensile strength of rebar, fsy = jft, where, j is the
normalized yield tensile strength, which is the ratio between the yield tensile strength of
steel rebar and the tensile strength of concrete. The tensile yield strain of steel is taken
as, esy = fsyEs, where Es is Young’s modulus of concrete. When the tensile strain of
rebar is above the yield tensile strain, the stress maintained constant.
0.002 0.0035
a) b) 0.025 c
c)
Fig. 1. Constitutive models for a) FRC in compression, b) FRC in tension and c) steel bars in
tension (Yao et al., 2018)
772 S. J. Stephen and R. Gettu
Based on the above-mentioned model and assumptions, the moment and axial force
capacity can be calculated for different strain profiles. If the forces and moments due to
the groundwater and soil loads are within the envelope of the developed interaction
diagram, then the design is safe (Fig. 2).
Fig. 2. Procedure to design a concrete tunnel segment for service loads
3 Rate Dependence of the Tensile Constitutive Model
The tunnel segments under service must withstand slow soil settlement and ground-
water movement. Even though these structural elements are designed for such mag-
nitude of loads, the variation of moment carrying capacity with loading rate is not
generally considered in analysis and design. The rate-effect model described in Stephen
and Gettu (2019) accounts for the tensile strength and residual bridging strength at
different loading rates. The effect of loading rate on the tensile stress-crack opening
(r-w) curve was determined by performing tests on notched beams made of fibre
reinforced concrete under CMOD rates covering 5 orders of magnitude. The inverse
analysis was performed to obtain r-w curves from the experimental P-CMOD
response, and the trends of tensile strength of the material was modelled as:
:
ft;x ¼ ft;std þ m log10 CTODx ð1Þ
:
where, ft;x = tensile strength of concrete at the CTOD rate of CTODx ; ft;std = tensile
strength of concrete at the standard loading rate and m is a parameter that depends on
the type and dosage of fibre in concrete.
The rate-dependence of the bridging stress was modelled as:
:
ri;x ri;std ¼ ft;conc;x ft;conc;std þ c log10 CTODx ; where i ¼ 1; 2; 3 ð2Þ
where c is a constant considered to be independent of the fibre type and dosage, and
ft;conc;x and ri:x are the tensile strength of the plain concrete and the bridging stress at the
:
CTOD rate of CTODx , respectively. The proposed Eqs. (1) and (2) consequently
comprise the model for rate-dependence of the fracture of FRC that can be used in
analysis and design. This rate-effect model is utilized to determine the fracture properties
required to derive the interaction diagram for the Terrassa tunnel segment, details for
which can be found in de la Fuente et al. (2012). The axial force-bending moment
interaction diagram is derived here based on the closed-formed solution proposed by Yao
a)
b)
Fig. 3. Load rate effects on moment-force interaction diagram for tunnel segment under service
load condition
774 S. J. Stephen and R. Gettu
c)
d)
Fig. 3. (continued)
et al. (2018) for concretes provided in Stephen and Gettu (2019). Note that the notation
used for the different cases indicates the strength of the concrete, type of fibre incor-
porated (polymer fibre – PF; steel fibre - SF), its dosage in kg/m3 and the CMOD rate of
the test in lm/s; e.g., M40SF30_0.01 denotes the 40 MPa grade concrete with steel fibres
at 30 kg/m3 dosage, tested at 0.01 lm/s. It is found that the incorporation of PF and
lesser dosage of SF have a greater impact on the rate-sensitivity of the segment moment
carrying capacity (Fig. 3). For instance, at zero axial force, the moment carrying capacity
reduces by 40% for M40PF3.75 and M40SF10, and less than 10% for higher dosage of
steel fibres, when the loading rate is changed from 100 µm/s to 0.01 µm/s. So, for a
hybrid tunnel segment reinforced with polymer fibres and low dosage of steel fibres, it
appears mandatory to include rate effects in the derivation of the interaction diagram.
4 Conclusions
In this paper, the design methodology to incorporate rate-dependent fracture properties

has been proposed. The significance of the methodology is discussed in terms of
variation of section capacity with loading rate. The modification factors are obtained
from the rate effect model to alter the design inputs and determine the section capacity.
Further, the load carrying capacity of tunnel segment was investigated at different
loading rates and was found to be decreasing with decrease in loading rate. The
segments reinforced with polypropylene fibres and low dosage of steel fibres are sig-
nificantly affected by long-term loading. With an increase in steel fibre dosage, the
long-term effects decreases. Therefore, it is crucial to incorporate rate effect model in
tunnel segments reinforced with low dosage of steel fibres or polymer fibres.
References
Asakura, T., Kojima, Y.: Tunnel maintenance in Japan. Tunn. Undergr. Sp. Technol. 18, 161–
169 (2003). https://doi.org/10.1016/S0886-7798(03)00024-5
Briffaut, M., Benboudjema, F., D’Aloia, L.: Effect of fibres on early age cracking of concrete
tunnel lining. Part II: Numerical simulations. Tunn. Undergr. Sp. Technol. 59, 221–229
(2016). https://doi.org/10.1016/j.tust.2016.08.001
Buratti, N., Ferracuti, B., Savoia, M.: Concrete crack reduction in tunnel linings by steel fibre-
reinforced concretes. Constr. Build. Mater. 44, 249–259 (2013). https://doi.org/10.1016/j.
Caratelli, A., Meda, A., Rinaldi, Z., Romualdi, P.: Structural behaviour of precast tunnel
segments in fiber reinforced concrete. Tunn. Undergr. Sp. Technol. 26, 284–291 (2011).
https://doi.org/10.1016/j.tust.2010.10.003
de la Fuente, A., Pujadas, P., Blanco, A., Aguado, A.: Experiences in Barcelona with the use of
fibres in segmental linings. Tunn. Undergr. Sp. Technol. 27, 60–71 (2012). https://doi.org/10.
1016/j.tust.2011.07.001
ITA Report, 2016. Twenty years of FRC tunnel segments practice : Lessons learnt and proposed
design principles ITA Working Group 2. International Tunnelling and Underground Space
Association, Avignon. France (2016)
RILEM TC 162-TDF, 2003. Final recommendation of RILEM TC 162-TDF: test and design
methods for steel fibre reinforced concrete sigma-epsilon-design method. Mater. Struct, vol.
36, pp. 560–567 (2003). https://doi.org/10.1617/14007
Stephen, S.J., Gettu, R.: Rate-dependence of the tensile behaviour of fibre reinforced concrete in
the quasi-static regime. Mater. Struct. 52, 107 (2019). https://doi.org/10.1617/s11527-019-
1405-2
Tiberti, G., Minelli, F., Plizzari, G.: Reinforcement optimization of fiber reinforced concrete
linings for conventional tunnels. Compos. Part B Eng. 58, 199–207 (2014). https://doi.org/10.
1016/j.compositesb.2013.10.012
Yao, Y., Bakhshi, M., Nasri, V., Mobasher, B.: Interaction diagrams for design of hybrid fiber-
reinforced tunnel segments. Mater. Struct. Constr. 51, 1–17 (2018). https://doi.org/10.1617/
s11527-018-1159-2
Codes and Standards
Developments and Standardisation of Flowable
Concrete Reinforced with Fibres for Structural
Design, Update of fib TG 4.3
Steffen Grünewald1,2(&), Liberato Ferrara3, and Frank Dehn4

1
Delft University of Technology, Delft, The Netherlands
S.Grunewald@tudelft.nl
2
Ghent University, Ghent, Belgium
3
Politecnico di Milano, Milan, Italy
4
Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany
Abstract. The fib Model Codes aim at integrating in a single document the

relevant knowledge for the structural design with concrete. Fibre reinforced
concrete is already integrated in fib Model Code 2010 (fib MC2010) as a general
category of materials. The group of flowable concrete consists of clusters of
different types of concrete among others Self-Compacting Concrete, Ultra High
Performance Concrete and Strain-Hardening Cementitious Composites. Being
highly flowable is the distinguishing characteristic, flowable concrete might
contain or not contain fibres. Although the fibre contribution on the structural
level can be assessed on short-term, the structural behaviour also depends on the
behaviour of the fibres and the matrix in which they are embedded. fib Task
Group 4.3 worked on identifying and characterising different types of flowable
concrete and discusses in a fib bulletin the most relevant aspects with regard to
mix design, manufacturing, material performance and structural behaviour of
flowable concrete which can allow innovative applications to be developed and
realised. This paper discusses recent developments with regard to flowable
concrete in a broader perspective and addresses the progress with regard to
standardisation.
Keywords: Concrete Fib Model Code Flowable concrete Fibres

Structural design
1 Flowable Concrete
Flowable concrete (FC) distinguishes itself first of all from conventional vibrated con-
crete (CVC) by its rheological characteristics, which are obtained by selective con-
stituent choice and mix design. Main types of FC developed during the past years are
self-compacting concrete (SCC, with or without fibres), fibre reinforced concrete
(FRC) with high flowability, strain-hardening cementitious composites (SHCC), high
performance fibre reinforced cementitious composites (HPFRCC), high performance
concrete (HPC) and ultra high performance concrete (UHPC). The trend of increase of
the workability of CVC facilitates the casting operation but also makes it difficult to
distinguish between FC and CVC. The different types of FC are often treated separately
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_69
780 S. Grünewald et al.
by standards and recommendations; the fib Model Code aims at integrating in a single

document the relevant knowledge with regard to performance of and structural design
with concrete, which also includes all types of FC. FC is an innovative concrete type and
often it is a material tailored for a specific application. Mainly, FC can be distinguished
from CVC through three dimensions (strength, fibre dosage and flowability), see Fig. 1.
In the different dimensions all before-mentioned FC types can be located and compared
with CVC. fib TG 4.3 [1] considers FC to be flowable with a slump of at least 200 mm;
depending on the material density and equation applied this translates into a yield stress
of roughly 600–900 Pa. Although, this flowability criterion with regard to the yield
stress seems to be high, FC containing a high fibre dosage with a slump of 200 mm has a
very low yield strength when tested without fibres. Such concretes require special mix
design considerations and are also included in the group of FC. Categories of fibre
dosages are: 1) non-structural applications (Vf 0.3 vol.%), 2) low to moderate fibre
dosage (> 0.3 & 1.0 vol.%), 3) significant fibre contribution (> 1.0 & 2.5 vol.%)
and 4) very high fibre dosage (>2.5 vol.%). Currently, the highest strength class of
normal concrete is C50/60 [2]; under discussion is the definition of high, very high and
ultra-high-strength concrete types.
Fig. 1. Three main differentiators of different concrete types.
Table 1 lists potential benefits per differentiator when adjusting the mix design of
FC compared to CVC hereby enabling the design and production of special applica-
tions. The achievement of potential benefits depends on component selection, mix
design, execution and structural design.
Table 1. Potential benefits when applying FC.

+Strength +Fibres +Flowability
- Compressive strength * - Maximum tensile strength * - Remote casting
- Tensile strength * - Post-cracking strength * - Casting of complex shape
- Modulus of elasticity * - Other mechanical properties * - Casting with dense reinforcement
- Other mechanical properties * - Crack width + - Production efficiency *
- Durability * - Replacement reinforcement - Aesthetic appearance improved
- Service life * - Combine with reinforcement - Fibre orientation can be beneficial/tailored
- Material volume + - Durability * - Mix design SCC promotes higher fibre dosage
- Increased early age strength - Self-healing
- Weight structure + - Material volume +
- Weight structure +
- Slender elements
- Thin elements
Developments and Standardisation of Flowable Concrete 781
2 Standardisation of Fibre Reinforced Concrete
Significant progress has been made during the past years for FRC and FC with regard
to material understanding, optimization of material and casting operation, the devel-
opment of test methods and structural design approaches. Important knowledge and
experience already have been implemented in new recommendations and standards.
For example, fib MC2010’s FRC provisions [2] have been partially implemented in the
Swedish Standard on SS812310 [3]. The AFGC guideline on ultra high performance
concrete [4] has been transferred to EN-complementary documents [5–7] addressing
concrete characteristics, execution aspects and structural design aspects. In the US, the
American Concrete Institute TC 544-Fibre Reinforced Concrete has published in 2018
the ‘Guide to design with fiber-reinforced concrete’ in which fib provisions are also
included [8]. Until now, most standards have a limit in scope e.g. with regard maxi-
mum strength class, level of flowability or applications covered. For the new fib Model
Code, a broader approach has to be considered in order to present FRC as a unique
group of concrete, flowable fibre concrete being a subgroup of it. Within fib COM 4
Concrete & Concrete Technology three task groups are working on topics related to
FRC and/or FC, which are: TG 4.1 (Fibre-reinforced concrete), TG 4.2 (Ultra high-
performance fibre-reinforced concrete) and TG 4.3 (Structural design with flowable
concrete).
At the moment of writing this paper, TG 4.1 and TG 4.3 are finalising drafts of
bulletins with regard to their work. Both documents support the preparation of the next
Model Code. fib TG 4.3 is finalising a state-of-the-art report and already published 9
papers during the past years addressing structural design with flowable concrete [9–17].
3 Flowable Concrete in the Fresh State
3.1 Mix Design

FC is obtained by the reduction of friction between solid particles and the optimization
of the paste (content and characteristics) complements the composition of the aggregate
skeleton with regard to yield strength and viscosity. Paste in FC has three functions in
the fresh state:
• Filling the interstices of the granular skeleton;
• Reducing the friction between the solid particles through an excess of the fluid
which separates the solids by means of an adequately “thick” layer of either water
or cement paste (dependent on the model considered);
• Keeping the aggregates suspended and preventing segregation.
The addition of fibres in concrete affects its characteristics in the fresh state, due to
the larger surface area of fibres, which requires more fluid phase to properly envelope
and lubricate the fibres. Stiff fibres also decrease the packing density of the granular
skeleton. Fibres can cause significant interparticle friction and interlocking between
fibres and between fibres and aggregates [18]. Fibres have an elongated shape; the
slenderness ratio (L/D) is calculated by dividing the length L by the diameter D. An
equivalent diameter can be calculated for fibres with cross-sections that are not round
or which are deformed (i.e. with hooked-ends or wave-shaped) by weighting fibres and
calculating the diameter with the assumption that the fibres have a round circular cross-
section. The effect of fibres on the workability of concrete depends among others on the
dosage, their flexibility, surface characteristics and how they interact with other
components of concrete during the flow. Fibres often are bundled with water-soluble
glue (effectively reducing the slenderness ratio of fibres before mixing) in order to
prevent the formation of intertwining fibres (balling) during the mixing process. The
size of the fibres relative to the aggregates determines their distribution. The explained
mechanisms and principles result in the fact that the mixture composition of CVC and
FC can differ considerably; Fig. 2 shows examples of mix designs of different types of
concrete.
100
90
80
% by volume
70 Cement
60 Filler
50 Water
40 Air
30 Sand
20 Coarse
10
0
Normal SCC HPC UHPC
concrete
Fig. 2. Examples of compositions for normal concrete/CVC, SCC, HPC and UHPC [1].
3.2 Flow Conditions

In order to take into account material behaviour for the structural design FC needs to
remain homogenous during the mixing and casting process. Two flow phenomena are
discussed by fib TG 4.3, which are 1) flow induced particle migration and 2) flow
induced anisotropy. A more detailed discussion with regard to flow conditions can be
found in [1].
3.2.1 Flow Induced Particle Migration

Concrete components can segregate or rise to the top of the concrete surface through
dynamic or static segregation. The first is associated with flow induced particle
migration originating from various phenomena whereas the second one is only asso-
ciated with the settling process due to the density difference between the components
when the material is at rest before or after casting. Yield stress and thixotropy seem to
dictate static segregation for a given granular skeleton. Depending on the considered
concrete flow, different phenomena can dominate and dictate particle migration within
the material. Different types of flow induced particle migration were identified:
• Shear induced particle migration;
• Gravity induced particle migration;
• Granular blocking;
• Heterogeneity in particle distribution induced by the wall effect;
• Induced heterogeneities during concrete casting or testing.
3.2.2 Flow Induced Anisotropy

Fibres can improve mechanical properties of cementitious materials in the hardened
state. This influence depends primarily on the fibres (shape, constitutive material,
volume fraction) but also on the casting process. Flow can induce a preferred orien-
tation of the fibres which influences the casting flow and, after setting, affects the
mechanical properties of the resulting fibre reinforced composite. An accurate pre-
diction of a population of fibre orientation should consider several parameters which
determine the macroscopic orientation in the hardened state material, such as the
geometry of the mould, the presence of free surfaces, the method of casting, or
interactions between fibres. The two categories of flow relevant in industrial practice
are: 1) free surface flow referred to as ‘slab casting’ and 2) confined flow referred to as
‘wall casting’.
4 Flowable Concrete in the Hardened State

4.1 Strength and Stiffness
As the fibre dosage of FC typically is up to 1.0 vol.% in ordinary FRC and 1.5−2.5 vol.
% in UHPC, the influence of fibres on engineering properties such as compressive
strength, modulus of elasticity, Poisson ratio and tensile cracking strength is very
limited or relatively small. This does not mean that the fibres do not influence such
properties when cracking is initiated. Since fibres also affect the workability, differ-
ences between mixtures with and without fibres can be caused by entrapped air or less
than adequate compaction after the addition of the fibres. However, as mentioned in
Sect. 3.1, in order to obtain FC within defined boundary conditions (e.g. bar spacing to
pass through, fibre content, required flowability) an adjustment of the mix design is
required. This can include: increase in paste content, decrease of aggregate content, size
and content, increase or decrease of water-powder ratio. Such changes affect the
stiffness and time-dependent deformation behaviour of concrete (see also Sect. 4.3).
4.2 Fibre Contribution

The behaviour of fibres during pull-out is determined by the characteristics of the fibres,
the matrix and the interface. Single and bundled (filament) fibres behave differently,
since the bond with the outer fibres in a bundle is better compared to fibres inside a
bundle. A brittle material/structure reinforced with brittle fibres (i.e. glass textiles)
loaded in tension can perform ductile, since the pull-out of inner fibres takes place by
friction with other fibres. According to [19], mainly three types of interactions deter-
mine the pull-out behaviour of a single fibre: 1) physical and chemical adhesion, 2)
friction and 3) mechanical anchorage induced by deformation of fibre or anchorage.
Dependent on the pull-out behaviour, the dosage, distribution and orientation of the
fibres, softening or hardening in direct tension is obtained (Fig. 3).
Fig. 3. Softening (a) and hardening (b) behaviour in direct tension [2].
A material with direct tension-softening behaviour still can be hardening in bending

or when applied in a statically indeterminate structure. In many cases, the production
with FRC is a trade-off between processing requirements, desired mechanical perfor-
mance and economical aspects. Since fibres increase the material costs, an optimum
fibre dosage has to be determined. The difference in contribution of softening or
hardening FRCs to the structural performance can be taken into account. Figures 4a
and 4b show the influence of the type and dosage of steel fibres on the compressive and
tensile strengths of HPFRCC; the abscissa value (fibre factor) is the product of fibre
volume and aspect ratio (ratio length to diameter) of the steel fibres.
4.3 Time-Dependent Deformations

The risk of cracking is also a consequence of an interrelation of the time-dependent
viscoelastic properties, the tensile strength and the modulus of elasticity. The time- and
load-dependent behaviour of FC can be rather different compared to CVC. As Fig. 2
shows, the paste contents and mixture compositions of FC can vary throughout a wide
range. The different contributions to shrinkage occur at different moments after casting
(chemical, autogenous and drying shrinkage). Higher cement contents can cause larger
deformations due to temperature-effects in an early stage; a high heat release potential
might not be obtained in practice because of a very low water-binder ratio and low
degree of hydration. SHCC has a high paste content at the highest w/c-ratio of all types
of FC. Accordingly, drying shrinkage and creep are relatively high. Strain caused by
shrinkage and creep can add up, but creep also results in relaxation, which reduces the
shrinkage strain. A heat treatment of UHPC (e.g. at 90 °C for 96 h) significantly
reduces the creep and shrinkage beyond the day of heat curing. Predictive model
Fig. 4. Influence of the fibre factor of steel fibres on a) compressive strength (cylinders of
d/h = 100/170 mm) and b) maximum tensile strength (dog-bone specimens) [20]
formulae provided in standards and recommendations may not be applicable for FC

[14]. A detailed discussion of the time- and load-dependent behaviour for FC can be
found in [14] and the fib bulletin (to be published) on Structural Design with Flowable
Concrete [1]. Main conclusions of [14] are summarised hereafter:
Drying shrinkage: The total shrinkage increases at increasing paste volume [21].
When the water-powder ratio is kept constant, the relation between shrinkage and paste
volume is approximately linear and can be regarded as the dominating parameter in
drying shrinkage. Drying shrinkage in UHPC is very low (in the range of 0.1‰
concrete deformation) compared to autogenous shrinkage [22]. Although the shrinkage
of SHCC is high (range of 1‰ at a relative humidity of 60%), its tensile strain capacity
seems to be higher than the drying shrinkage deformation [23].
Autogenous shrinkage: Influence of paste volume: At a constant water/powder-ratio,

the autogenous shrinkage of SCC increases at increasing paste volume [24]. As a result,
CVC shrinks less than SCC when the binder composition and strength class are the
same. Autogenous shrinkage of UHPC is considerably larger compared to CVC [25].
In SHCC a decrease in paste volume by the addition of fine-grained aggregates (grain
size of 0.15–0.30 mm) can significantly reduce autogenous shrinkage [26].
Creep: Even when results of different studies are not consistent, there seems to be a
general agreement that the creep coefficient and the specific creep are normally slightly
higher for SCC compared to CVC [21]. Specific creep of UHPC is in the range of 0.01–
0.035‰/MPa, while the creep coefficient is in the range of 0.5–1.2 [27]. Because of a
higher paste volume SHCC (HPFRCC) shows large creep deformations, but due to a low
modulus of elasticity, creep coefficients can be even smaller than the ones of CVC [28].
5 Model Code Provisions
Essential information for the structural design with FRC is the magnitude of post-
cracking (residual) strength at different values of crack mouth opening displacement
(CMOD) of 0.5, 1.5, 2.5 and 3.5 mm. The CMOD is generally measured from flexural
tests performed on notched beams. fR1 (at CMOD = 0.5 mm) and fR3 (at CMOD =
2.5 mm) are main parameters considered for design (Fig. 5). Fibre reinforcement can
substitute (also partially) conventional reinforcement in the ultimate limit state, if the
following relationships are fulfilled: fR1k/fLk > 0.4 and fR3k/fR1k > 0.5. Combining
both criteria means that fR3 has to be at least 20% of the flexural strength fLk, which
assures a minimum ductile behaviour in bending. Dependent on the ratio fR3k/fR1k
categories a-e (a 0.5 < fR3k/fR1k < 0.7, b 0.7 fR3k/fR1k < 0.9, c 0.9 fR3k/fR1k <
1.1, d 1.1 R3k/fR1k < 1.3 and e 1.3 R3k/fR1k) have been defined which allows
specifying concrete performance e.g. for ready-mix concrete producers. Accordingly,
the design engineer can derive structural performance without being involved in the
design and testing of the concrete. The orientation and distribution of the fibres but also
the variation of results within a test series also determine the flexural performance that
can be taken into account for design. With flow induced anisotropy FRC can be tensile
hardening in one direction and tensile softening in the other which has to be properly
considered when performing a structural analysis, mainly of statically redundant
structures, including slabs where biaxial stress states can be common.
b)
a)
Fig. 5. Deriving design residual strengths a) typical load F-CMOD curve for plain concrete and
FRC and b) translation to simplified post-cracking constitutive laws [2].
fib MC2010 [2] includes provisions with regard to the following aspects related to
the design with FRC (Table 2), provisions that take into account e.g. the contributions
of fibres in bending, shear, punching shear and for crack width calculations. Where
relevant the contribution of fibres can be included in the calculation of minimum
reinforcement (bending, shear, punching shear, skin reinforcement) as well as on
anchorage/introduction length of rebars.
Table 2. Provisions in fib MC2010 with regard to the design of FRC.

Section MC Design aspect
5.6 Fibres/fibre reinforced concrete
5.6.2/5.6.3 Material behaviour and classification
5.6.4 Constitutive laws
5.6.5 Safety factors
5.6.6 Fibre orientation
7.7 Verification of safety and serviceability of FRC structures
7.7.2 Design principles
7.7.3 Verification of safety
7.7.3.1 Bending and/or axial compression in linear members
7.7.3.2 Shear in beams
7.7.3.2 Torsion in beams
7.7.3.4 Walls
7.7.3.5 Slabs, members with/without reinforcement, punching
7.7.4 Verification of serviceability
7.7.4.1 Stress limitation
7.7.4.2 Crack width in members with conventional reinforcement
7.7.4.3 Minimum reinforcement for crack control
Bending and/or axial compression in linear members: fib Model Code 2010

includes simplified stress/strain relationships with which the cross-sectional analysis
can be executed. The difference between a strain-hardening and a strain-softening
material is shown by Fig. 6.
Fig. 6. Ultimate limit state for bending moment and axial force, use of simplified stress/strain
relationships [2].
Shear resistance: Due to the increase in fracture energy of FRC compared to normal
concrete and the opening of more and smaller cracks, the aggregate interlock contri-
bution of concrete also increases. Accordingly, fibres are able to increase the shear
resistance of concrete. The part of Eq. (1) multiplied by 7:5 fFtuk =fck is the additional
shear resistance provided by the fibres.
13 !
0:18 fFtuk
VRd;F ¼ k 100 q1 1 þ 7:5 fck þ 0:15 rcp bw d ð1Þ
c fctk
Punching shear resistance: The punching shear resistance is the sum of the contri-
butions of punching shear reinforcements of steel and fibres, which includes the
contribution VRd,c, see Eqs. (2) and (3).
fFtuk
VRd ¼ VRd;F þ VRd;s ) VRd;F ¼ VRd;c þ b0 dv ð2 þ 3Þ
cF
Crack width verification: The design crack width wd in FRC elements can be
determined with Eq. (4). The term with fFtsm (negative contribution) reduces the value
of wd; it is the tensile contribution of the fibres which reduces the crack opening. By
virtue of the action of the fibres, which generate a residual tensile strength fFtsm, the
force to be reintroduced by bond is reduced.
( )
1 /s fctm fFtsm 1
wd ¼ 2 k c þ ðrs b rsr þ gr 2sh Es Þ ð4Þ
4 qs;ef sbm Es
fib TG 4.3 identified and discussed differences between CVC and FC with regard to
mixture composition, execution aspects and structural performance. Fibre orientation
has to be considered; with an appropriate safety approach the potentially different
characteristics can be taken into account.
6 Conclusions
fib TG 4.3 aims at integrating different types of flowable concrete in a state-of-the-

report; the result will be published in an oncoming fib bulletin. The following con-
clusions were drawn:
• Significant steps have been made with regard to the development of performance
tests, non-destructive test methods and standardisation;
• Innovative and/or economical concrete solutions and structures can be developed
with tailor-made flowable concrete;
• Mix design, fibre orientation and distribution are essential aspects to consider in
order to take into account the structural performance;
Acknowledgements. The authors want to thank the members of fib TG 4.3 for their contribu-
tions; with regard to this publication T.A. Hammer, A. Leemann, N. Roussel and W. Schmidt are
gratefully acknowledged.
References
1. fib bulletin: Structural design with flowable concrete. State-of-the-art report (to be published)
2. fib Model Code 2010: fib Model Code for Concrete Structures 2010. Wilhelm Ernst and
Sohn, ePDF ISBN: 978-3-433-60408-3 (2013)
3. SS812310:2014: Fibre Concrete-Design of Fibre Concrete Structures, Swedish Institute for
Standards (2014)
4. AFGC: Ultra High Performance Fibre-Reinforced Concretes–Recommendations, Associa-
tion Française de Génie Civil (2013)
5. NF P 18-470: Bétons fibrés a ultra-hautes performances - Spécification, performance,
production et conformité, AFNOR, Paris (2016)
6. NF P 18-710: Calcul des structures en béton-Regies spécifiques pour les bétons fibrés a ultra-
hautes performances (BFUP), AFNOR, Paris (2016)
7. NF P 18-451: Execution des structures en béton-Régies spécifiques pour les BFUP, AFNOR,
Paris (2018)
8. ACI TC 544: Guide to design with Fiber-Reinforced Concrete’, 544.4R-18, ACI (2018)
9. Grünewald, S., Ferrara, L., Dehn, F.: fib task group structural design with flowable concrete.
In: Hirt, M.A., Radic, J., Mandic, A. (eds.) Codes in Structural Engineering; Developments
and Needs for International Practice, IABSE Dubrovnik, pp. 1333–1340 (2010)
10. Grünewald, S., Ferrara, L., Dehn, F.: fib-recommendation structural design with flowable
concrete. In: Proceedings of the 3rd International Congress Incorporating the PCI Annual
Convention and Bridge Conference, Chicago, Precast/Prestressed Institute PCI, pp. 1–9
(2010)
11. Ferrara, L., Grünewald, S., Dehn, F.: Design with higly flowable fiber-reinforced concrete:
overview of the activity of fib TG 8.8. In: Khayat, K., Feys, D. (eds.) Proceedings SCC 2010
Design, Production and Placement of SCC, Springer, ISBN 978–90-481-9663-0, pp. 395–
406 (2010)
12. Grünewald, S., Ferrara, L., Dehn, F.: Structural design with flowable concrete–a fib-
recommendation for tailor-made concrete. In Khayat, K., Feys, D. (eds.) Proceedings SCC
2010 Design, Production and Placement of SCC, Springer, ISBN 978-90-481-9663-0,
pp. 13–24 (2010)
13. Grünewald, S., Bartoli, L., Ferrara, L., Kanstad, T., Dehn, F.: Translation of test results of
small specimens of flowable fibre concrete to structural behaviour: a discussion paper of fib
task group 4.3, fib Bulletin No. 79. In: Proceedings FRC 2014 ACI-fib Int. Workshop Fibre-
Reinforced Concrete: From Design to Structural Applications, ISBN: 978-2-88394-119-9,
pp. 81–90 (2016)
14. Leemann, A., Hammer, T.A., Grünewald, S., Ferrara, L., Dehn, F.: Time- and load-
dependent behaviour of flowable concrete: progress report of fib task group 4.3. In:
Braestrup, M., Stang, H. (eds.) fib Symposium Concrete-Innovation and Design, Lausanne:
fib, pp. 1–11 (2015)
15. Schmidt, W., Grünewald, S., Ferrara, L., Dehn, F.: Design of concrete for high flowability:
Progress report of fib task group 4.3. In: Braestrup, M., Stang H. (eds.) fib Symposium 2015
Concrete-Innovation and design, Lausanne, fib, pp. 1–10 (2015)
16. Grünewald, S., Ferrara, L., Dehn, F.: Structural design with flowable concrete. In: Nunes, S.,
Sousa Coutinho, J., Faria, R. (eds.) Proceedings 4th Congresso Ibero-Americano sobre Betao
Auto-compactavel, BAC2015, Faculdade de Engenharia, Universidade do Porto, pp. 19–32
(2015)
17. Grünewald, S., Liberato, F., Dehn, F.: Flowable fibre-reinforced concrete: Progress in
understanding and development of design standards. In: Khayat, K.H. (ed.) Proceedings 8th
International RILEM Symposium on SCC, RILEM Publications, ISBN: 978-2-35158156-8,
pp. 467–478(2016)
18. Bayasi, M.Z., Soroushian, S.P.: Effect of steel fiber reinforcement on fresh mix proportions
of concrete. ACI Mater. J. 89(4), 369–374 (1992)
19. Bentur, A., Mindess, S.: Fibre Reinforced Cementitious Composites, 2nd edn. Taylor &
Francis, New York (2007). ISBN 978-0-415-25048-X
20. Sato, Y., Mier, J.G.M. van, Walraven, J.C.: Mechanical characteristics of multi-modal fiber
reinforced cement based composites. In: Lyon, Rossi, P., Chanvillard, G. (eds.) Befib 2000,
Cachan Cedex, pp. 791–800, ISBN: 2-912143-18-7 (2000)
21. Leemann, A., Lura, P., Loser, R.: Shrinkage and creep of SCC-the influence of paste volume
and binder composition. Constr. Build. Mater, vol. 25, pp. 2283–2289
22. Koh, K., et al.: Shrinkage properties of ultra-high performance concrete. Adv. Sci. Lett. 4,
948–952 (2011)
23. Weimann, M.B., Li, V.C.: Hygral behavior of engineered cementitious composites (ECC).
Int. J. Restor. Build. Monuments 9, 513–534 (2003)
24. Rozière, E., et al.: Influence of paste volume on shrinkage cracking and fracture properties of
self-compacting concrete. Cem. Concr. Compos. 29, 626–636 (2007)
25. Loukili, A., Khelidj, A., Richard, P.: Hydration kinetics, change of relative humidity, and
autogenous shrinkage of ultra-high-strength concrete. Cem. Concr. Res. 29, 577–584 (1999)
26. Billington, S.L., Rouse, J.M.: Time-dependent response of highly ductile fiber-reinforced
cement-based composites. In: International Symposium Brittle Matrix Composites, Warsaw,
pp. 47–54 (2003)
27. Flietstra, J.C., Ahlborn, T.M., Harris, D.K., De Melo e Silva, H.: Creep behaviour of UHPC
under compressive loading with varying curing regimes. In: 3rd International Symposium on
UHPC, Kassel, Germany, pp. 333–40, ISBN: 978-3-86219-264-9 (2012)
28. JSCE HPFRCC: Recommendations for design and construction of High Performance Fiber
Reinforced Cement Composites with multiple fine cracks (HPFRCC). Concrete Engineering
Series 82, Japan Society of Civil Engineers (2008)
Assesment of Codal Provision for SFRC Beam
in Minimum Shear
Kranti Jain and Bichitra S. Negi(&)
Department of Civil Engineering, National Institute of Technology,

Srinagar, Uttarakhand, India
bichitran@gmail.com
Abstract. Natural fibres (straw, chip, horse tail, goat hair and plume, etc.), was
being used from ancient time for construction purposes. Inspired from ancient
time, artificial fibres (vitreous, synthetic, carbon and steel fibre, etc.) are com-
monly used nowadays in order to improve the mechanical properties of
concrete.
Literature has suggested that concrete matrix with steel fibre commonly
known as Steel fibre reinforced concrete (SFRC) have enhanced flexural as well
as shear behaviour as compared to conventional reinforced concrete (RC). Also
it is recommended by ACI building code that steel fibre can use as minimum
shear reinforcement in RC beams. In the present study, the experimental results
for assessment of ACI building code provisions allowing the use of deformed
steel fibres as minimum shear reinforcement in RC beams are presented. Hooked
end steel fibres of length 35 mm and 60 mm have been used in concrete
matrices at volume fraction of fibres ranges from 0.75 to 1.5% and 0.5 to 1%
respectively. The experimental performance of the fibrous concrete beams has
been compared with the Indian Standard and ACI codal provisions. It is found
that in all the cases, even though at a volume fraction of 0.5%, lower than the
ACI code-specified lower limit of 0.75%, the measured shear strengths were
higher than the predicted values as per the ACI building code, Indian standard
code, as well as from the lower bound value of 0.3√fc′MPa (3.5√fc′ psi). Also
confirmed multiple diagonal cracking having crack width lower than the spec-
ified permissible values. On the basis, it is suggested that there is a need of
relook into codal provisions of ACI.
KEYWORDS: Steel fibre reinforced concrete (SFRC) Reinforced concrete

(RC) Minimum shear reinforcement Deformed steel fibres
1 Introduction
Concrete and mortar are well known for its brittle nature therefore less able to resist
tensile stresses and propagation of cracks. Various researchers have been carried out to
overcome this problem. The inclusion of deformed fibres in concrete matrix mainly
improves its residual strength, toughness and ductility. The improved material prop-
erties of fibre-reinforced concrete (FRC) further improves the flexural and shear
behaviour of FRC structures [1–4]. It has been already recognized that the fibre rein-
forcement is an effective way to enhance toughness of concrete for every type of failure
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_70
792 K. Jain and B. S. Negi
modes. Beam without web reinforcement fails suddenly in shear because of its brittle
nature. Research over last few decades has clearly established the potential use of fibre
for enhancing the shear capacity of reinforced concrete (RC) beams [5–7].
The effect of fibre reinforcement on shear strength is contributed of two main
factors: 1) a direct factor imposed by post cracking strength at inclined shear crack; 2)
an indirect factor which cause increase in concrete contribution to shear strength by
improving aggregate interlock and dowel action of reinforcement [2]. Many researchers
has been carried out in past few decade to focus on shear behaviour of SFRC beam.
The result showed that the inclusion of steel fibre boost the shear capacity of concrete
and enhance the shear crack distribution, therefore they are capable of replacing ver-
tical stirrups in RC structural members.
It is well known that shear failure loads vary widely about the value given by the
design equations. Most of the current design codes [8–10] suggested the requirement of
minimum amount of transverse reinforcement if applied shear force exceeds a definite
fraction of inclined cracking shear. Such reinforcement sometime may also have such
value that it may able to resist unexpected tensile force and overload. When a flexural
member is subjected to any load like fatigue loading, there exists the possibility that
inclined diagonal tension cracks may form under statis loading even at lower load. And
therefore codes suggested to provide minimum shear reinforcement even calculation
does not shows requirement of shear reinforcement.
According to Dinh et al. [11] fibre in concrete matrix enhance shear resistance by
transferring tensile stresses across inclined diagonal cracks and overcome further
increase in crack width and spacing and also enhance contribution due to aggregate
interlock. Parra-Montesinos [12] shows that the steel fibre can be used as shear rein-
forcement in reinforced concrete (RC) beam on the basis of past literature. A lower
bound for shear strength of FRC of value 0.3 √fc was considered adequate with fibre in
concrete at volume fraction greater than or equal to 0.75% After this research and data
from past literature, a new provision was included in 2008 ACI building code [13]
which allows use of deformed steel fibres as minimum shear reinforcement in normal
strength concrete when fibre volume fraction is greater than or equal to 0.75%.
This investigation explores the use of deformed steel fibres as minimum shear
reinforcement in RC beams. The experimental program comprises of test on large
SFRC beam which are designed to fail in shear under monotonically increasing loads in
three point loading configuration. The behaviour of SFRC beams are compared with
traditional beams with minimum shear reinforcement as per the recommendations of
ACI building code [8, 13] and IS 456:2000 [9].
Hooked end steel fibre of length 60 mm (aspect ratio = 80) and 35 mm (aspect
ratio = 65) having ultimate tensile strength of 1050 MPa and 1100 MPa respectively
were used in concrete mixture. 35 mm long steel fibre were added at volume fraction
(Vf) of 0.75%, 1% and 1.5% whereas 60 mm long steel fibre were added at volume
fraction of 0.5%, 0.75%, and 1%. 1.5% volume fraction of 60 mm long steel fibre were
not used because of possibility of balling effect of fibre at volume fraction more than
Assesment of Codal Provision for SFRC Beam in Minimum Shear 793
1%. ACI 2008 building code [13] recommended a minimum Vf of 0.75% whereas the
Vf of 60 mm long steel fibre were kept below 0.75% to expand the scope of investi-
gation. The detail of plain and SFRC mixture is documented in Table 1. For evaluation
of shear behaviour, control specimens of plain concrete mixture were used.
Table 1. SFRC mixture proportions

Ingredient Weight (kg) per m3
V*f = 0% Vf = 0.5% Vf = 0.75% Vf = 1% Vf = 1.5%
Cement 396 394 394 394 394
Fine aggregate, FA 870 862 860 857 852
Course aggregate, CA 656 648 646 643 638
(4.75 mm–10 mm)
Course aggregate, CA 353 349 348 346 344
(10 mm–12.5 mm)
Superplasticiser – – – – 0.79
Water 225 233 233 233 232
W/C ratio 0.57 0.59 0.59 0.59 0.59
Steel fibres – 39 59 79 118
CA/(FA + CA) 0.54 0.54 0.54 0.54 0.54
Shear behaviour of SFRC beams was investigated by testing singly reinforced

beam having length 1770 mm and cross-Sect. 150 mm 300 mm. The beam have
simply supported span of 1470 mm and failed under monotonically increasing three-
point loads. In order to ensure the correctness of results, nominally identical companion
beams were cast for every specimen. The geometry of beam specimens and test set up
configuration is presented in Fig. 1.
Each beam was designed in such a way that it will fail in shear in the tested span
whose shear span to effective depth ratio was 3.5. The shorter span was provided with
sufficient stirrups so that the failure always occurs in tested span. In order to obtain
shear failure prior to flexural failure, all beam specimen were intentionally over rein-
forced with 2.67% (100Ast/bd) of tension reinforcement at an effective depth of
251 mm.
Based on the distribution of transverse reinforcement in the tested span, the beams
were classified as follows:
• To observe typical brittle failure, no transverse reinforcement was provided in the
tested span.
• Minimum shear reinforcement in the tested span (6 nos of 8 mm diameter two-
legged closed rectangular stirrups) as per ACI building code [8].
• Minimum shear reinforcement in the tested span (4 nos of 8 mm diameter two-
legged closed rectangular stirrups) as per IS 456:2000 [9].
• Minimum shear reinforcement in the form of steel fibre only.
Tested span, a/d = 3.5
Steel bearing plate at the load

150 point, 100 x 150 x 50 150
150
300
Steel flats Load-point

LVDT arrangement for monitoring LVDT
inclined cracking
Steel bearing plate at support Mid-span Test beam, 150 x 300 x 1770
(typical), 100 x 150 x 50 LVDT
875 (Tested span) 595
1470
1770
(a) Front elevation of a typical beam
150
Effective depth = 251

300
Ast, typical
(All dimensions in mm) (b) Cross-section
Fig. 1. Beam specimen geometry and test set-up configuration
The assembled reinforcement cages are shown in Fig. 2, and it should be noted that
the longitudinal reinforcement bars were provided with sufficient hooked extensions at
their ends as per ACI building code [8] in order to minimize chance of anchorage
failure. The mechanical properties of the reinforcement bar used in the specimens are
tabulated on Table 2.
Before casting of beam, the prefabricated steel reinforcement cages were placed in
the formwork keeping clear cover of 25 mm from bottom. The target cylindrical
compressive strength of normal strength concrete was 26 MPa and initial slump was in
the range of 150 mm to 175 mm for plain concrete mixture whereas 40 mm to 100 mm
for fibrous concrete mixture. The beams and the control specimens are demoulded after
24 h of casting and moist curing was done for 10 days.
Subsequently, the specimens were air-cured in the laboratory until testing, which
was carried out after a nominal interval of 28 days from the day of casting.
The summary of beam specimens along with the experimental results obtained from
shear test was tabulated in Table 3.
The diagonal crack would be accompanied by web deformations, which would be
indicated by a sudden and significant change in the displacements measured by the
(a) Control beam (no shear reinforcement in the tested span)
(b) Minimum shear reinforcement per the ACI Building Code [8]
(c) Minimum shear reinforcement per the IS 456:2000 [9]
(d) Steel fibres as minimum shear reinforcement
Fig. 2. Details of assembled reinforcement cages
Table 2. Mechanical properties of reinforcing bar

Bar diameter (mm) Yield strength, MPa Ultimate strength, MPa % Elongation
8 558 723 8
10 553 646 25
16 566 692 26
Table 3. Beam specimens and their experimental results

Beam Concrete mixture and f′c Pu vu vu/(f Failure mode+
ID detailing of transverse MPa MPa MPa ′c)0.5
reinforcement
A-I Plain concrete, no 24.50 125.41 1.35 0.27 DT + ST
A-II transverse reinforcement 25.50 185.01 1.99 0.39 DT + ST
B-I Plain concrete, transverse 28.1 394.76 4.24 0.80 SC
B-II reinforcement per ACI 25.43 374.33 4.02 0.80 SC
B-III 318-08 (ACI 2008) 26.51 356.78 3.84 0.74 SC
C-I Plain concrete, transverse 28.15 402.20 4.32 0.81 SC
C-II reinforcement per IS 26.10 378.52 4.07 0.80 SC
C-III 456:2000 (BIS 2000) 25.64 355.88 3.83 0.76 SC
D-I N -35-0.75* 28.10 278.77 3.00 0.57 DT + ST + SC
D-II 25.28 195.63 2.10 0.42 DT + ST + SC
E-I N -35-1.00* 27.90 270.03 2.90 0.55 DT + ST + SC
E-II 26.20 305.20 3.28 0.64 DT + ST + SC
F-I N -35-1.50* 28.10 274.47 2.95 0.56 DT + ST + SC
F-II 27.33 323.25 3.48 0.66 DT + ST + SC
G-I N -60-0.50* 27.5 159.94 1.72 0.33 DT + ST + SC
G-II 24.94 190.59 2.05 0.41 DT + ST + SC
H-I N -60-0.75* 27.75 224.69 2.42 0.46 DT + ST + SC
H-II 27.33 251.56 2.70 0.52 DT + ST + SC
I-I N -60-1.00* 26.25 286.52 3.08 0.60 DT + ST + SC
I-II 27.12 258.80 2.78 0.53 DT + ST + SC
* N- in first place stands for normal strength, numeral value at second place stands for length of
fibre, numeral value at third place stands for volume fraction of fibre.
DT- Diagonal tension failure; ST- Shear tension failure; SC- Shear compression failure.
LVDT arrangement. For testing the beams specimen, hydraulic ram was used to apply
monotonically increasing load in 10–15 increments until failure and beam deflections at
different point was monitored by LVDTs. The data of testing was recorded by com-
puter aided data acquisition system. The crack pattern, number of cracks, crack widths
and failure modes were carefully monitored and noted for each specimen.
3 Result and Discussion
Various failure modes are observed in tested span of different beams specimen and the
peak load crack patterns were shown in Fig. 3 for the purpose of comparision. During
testing, the initial inclined cracking first observed in shorter span but it was sufficiently
reinforced in shear. Therefore the diagonal crack induced reduction in shear stiffness of
the shorter span and inclined crack appeared in the longer span, where the shear failure
in the beam specimen has to occur.
(a) Control beam (no shear reinforcement in the tested span)
(b) Minimum shear reinforcement per the ACI Building Code [8]
(c) Minimum shear reinforcement per the IS 456:2000 [9]
(d) 1% Vf of 35 mm long hooked-end steel fibres
(e) 1% Vf of 60 mm long hooked-end steel fibres
Fig. 3. Peak load crack patterns of the beam specimens

The crack pattern on the tested span was different based on detailing of transverse
reinforcement. Beam with no transverse reinforcement shows failure by diagonal
tension (DT) in combination with shear tension (ST), Fig. 3(a). Diagonal tension was
initiated by propagation of a single major crack was toward the load point and along the
longitudinal reinforcement. At failure the crack is penetrated to concrete compression
zone without crushing the concrete. In Shear tension, the inclined crack propagated
along the longitudinal reinforcement and towards the support, which tends to reduce
the anchorage capacity of the reinforcement.
In opposite to Fig. 3(a), multiple diagonal cracks were observed in other tested
span, Fig. 3(b), (c), (d), and (e). Figure 3(b), (c) shows shear compression (SC) failure
with crushing of concrete in the compression zone near the load point. Additional
multiple cracking with one single major crack were observed in fibrous beams which
provide earlier warning before collapse, Fig. 3(d), (e).
The measured peak loads and normalized peak average shear stresses are tabulated
on Table 3 and the peak average shear stress values are plotted in Fig. 4, wherein they
are compared with codal predictions and the lower bound value suggested by Parra-
Montesinos [12]. There was large variation in normalized shear stress value for the
plain and fibrous concrete beams as compared to others, although none of the beam had
value lower than the lower bound value of 0.3 of Parra-Montesinos [12]. Among all the
fibrous concrete 50% Vf of 60 mm long steel fibre has relatively less normalized shear
strength of value 0.33.
4.50
4.00 Min.-IS 456 (Experimental)
3.50 Min.-ACI 318 (Experimental)
3.00
Shear stress (MPa)
Hooked fibres 35 mm long

2.50
Hooked fibres 60 mm long
2.00
SFRC Predicted (Parra
1.50 Montesinos)
Min. IS 456 (Predicted)
1.00
0.50 Min. ACI 318 (Predicted)
0.00
0 0.25 0.5 0.75 1 1.25 1.5
Fibre volume fraction (%)
Fig. 4. Shear stress v/s fibre volume fraction
The minimum shear reinforcement requirement in ACI building code [8] and IS
456:2000 [9] correspond to the nominal shear strength of 0.34 MPa and 0.4 MPa,
respectively and values are plotted in Fig. 4, which shows that every beam specimens
had shear strength more than the lower bound value of 1.53 MPa. Lower bound value
pffiffiffiffi
is obtained by the putting fc’ = 36 MPa in the equation 0:3 fc0 suggested by Parra-
Montesinos [12].
Figure 4 also shows that the experimental shear strength of the specimens with the
code recommended minimum shear reinforcement were much more than the predicted
value, which shows that the conservative nature of the codal recommendations related
to shear design.
4 Conclusions
• The shear strength of 60 mm long steel fibrous concrete beam with 0.5% of Vf
which was lower than the lower bound limit recommended by ACI 318 building
pffiffiffiffi
code shows multiple cracking and shear strength more than the 0:3 fc0 .
• The failure mode of steel fibre reinforced beam was comparable to that of those
beam designed for minimum shear reinforcement in accordance to ACI 318
building code and IS 456:2000.
• The ACI 318 building code and IS 456:2000 specified predicted value of shear
strength i.e. 0.344 MPa and 0.4 MPa, respectively was much less than the exper-
imental results obtain by testing codal designed beams specimen.
• The service load crack width of steel fibrous concrete beam was very less compared
to allowable value of 0.3 mm for normal exposure condition as well as from the
beam specimen designed in accordance to codal provision.
References
1. Narayanan, R., Darwish, I.Y.S.: Use of steel fibers as shear reinforcement. ACI Struct. J. 84
(3), 216–227 (1987)
2. Altoubat, S., Yazdanbakhsh, A., Rieder, K.A.: Shear behaviour of macro-synthetic fiber-
reinforced concrete beams without stirrups. ACI Struct. J. 106(4), 381–389 (2009)
3. Dinh, H.H., Parra-Montesinos, G.J., Wight, J.K.: Shear behaviour of steel fibre-reinforced
concrete beams without stirrup reinforcement. ACI Struct. J. 107(5), 597–606 (2010)
4. Kwak, Y.-K., Eberhard, M.O., Kim, W.-S., Kim, J.: Shear strength of steel fiber-reinforced
concrete beams without stirrups. ACI Struct. J. 99(4), 530–538 (2002)
5. Batson, G.B., Jenkins, E., Spatney, R.: Steel fibers as shear reinforcement in beams.
ACI J. Proc. 69(10), 640–644 (1972)
6. Dupont, D., Vandewalle, L.: Shear capacity of concrete beams containing longitudinal
reinforcement and steel fibers. In: Banthia, N., Criswell, M., Tatnall, P., Folliard, K. (eds.)
Innovations in Fiber Reinforced Concrete for Value, SP-216, pp. 79–94. American Concrete
Institute, Farmington Hills (2003)
7. Li, V., Ward, R., Hamza, A.M.: Steel and synthetic fibers as shear reinforcement. ACI
Mater. J. 89(5), 499–508 (1992)
8. ACI Committee 318: Building Code Requirements for Structural Concrete (ACI 318-14) and
Commentary (318R-14). American Concrete Institute, p. 503. Farmington Hills, Michigan
(2014)
9. IS 456: Indian Standard Plain and Reinforced Concrete – Code of Practice,‖ Fourth
Revision. Bureau of Indian Standards, BIS, New Delhi, 100 p. (2000)
10. BSI: EN 1992-1-1:2004. Eurocode 2: Design of concrete structures. Part 1-1: General rules
and rules for buildings. BSI, London (2004)
11. Dinh, H.H., Parra-Montesinos, G.J., Wight, J.K.: Shear strength model for steel fiber
reinforced concrete beams without stirrup reinforcement. J. Struct. Eng. ASCE 137(10),
1039–1051 (2011)
12. Parra-Montesinos, G.J.: Shear strength of beams with deformed steel fibers. Concr. Int. 28
(11), 57–66 (2006)
13. ACI: ACI 318R-08: Building code requirements for structural concrete and commentary.
ACI, Farmington Hills (2008)
Laboratory Investigations on the Installation
of Fasteners in Fiber Reinforced Concrete
Panagiotis Spyridis1(&), Lars Walter1, Julia Dreier2,

and Dirk Biermann2
1
Chair of Fastening Technology, TU Dortmund, Dortmund, Germany
panagiotis.spyridis@tu-dortmund.de
2
Institute of Machining Technology, TU Dortmund, Dortmund, Germany
Abstract. A significant aspect regarding the use of post-installed anchors in

concrete is related to constructability, and mainly the characteristics of possible
geometrical configurations. Specifically for fibre reinforced concrete (FRC), as
fibre strengths or dosages increase, these configurations may change. One of the
critical parameters is the minimum allowable distance from the edge, which can
significantly influence the flexibility of using post-installed anchors, or even the
possibility to install them in the first place. This geometrical parameter strongly
depends on the drilling and anchoring procedure, as various technologies are
available. This paper will present and compare the various available technolo-
gies for fastenings, and it will focus on setting experiments in fibre reinforced
concrete by use of challenging means for drilling and setting, i.e. hammer
drilling, mortar injection and expansion anchors. The influence of fibres in the
installation and tightening of post-installed anchors is particularly analysed. The
results of the study reveal the possibilities for extended applications of fasten-
ings in fibre reinforced concrete, as compared to plain concrete.
Keywords: Concrete Fastenings Installation Steel fibres
1 Introduction
As a consequence of the increasing trend for the use of fibre reinforced concrete
(FRC) in several types of engineering structures, advanced knowledge and standardi-
sation in the field become more and more essential. In the last years, FRC became a
favourable solution in a variety of significant structural components, including tunnel
structures, high-performance floor slabs (e.g. pile-supported and plan foundations,
seamless slabs, slabs on grade), immersed structures, watertight and containment
infrastructure components, and various sprayed concrete applications such as structural
strengthening of buildings and bridges, or geotechnical structures. The increasing
interest in design guidance for FRC is now reflected, besides numerous scientific
publications, in respective guidelines and standards, application specific guidance
documents and most importantly in the inclusion of provisions for the design with fibre
reinforced concrete in the upcoming edition of Eurocode 2.
At the same time, the relevance of fastenings in the last years is also made evident
by the fact that new specifically dedicated design guides and norms have been recently
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_71
802 P. Spyridis et al.
published. The new Eurocode 2 – Part 4 “Design of fastenings for use in concrete” is
the most prominent standardisation effort, while new design and specification guides
and standards are also emerging globally for fastenings to concrete.
1.1 Overview of Fastenings Engineering and Technology

Depending on the construction approach, two main distinctions are made between cast-
in-place and post-installed anchors, while for post-installed anchors further distinction
can be made between the installation type as undercut, expansion, plastic, bonded,
power actuated anchors and concrete screws. These technologies are also strongly
associated to three main load bearing mechanisms in tension:
– Mechanical interlock, when the anchor is engaged in the surrounding material
through contact
– Friction, which is developed through expansion of the anchor towards the walls of
the borehole
– Bond, by means of an appropriate chemical reaction that connects the anchor rod to
the borehole, which is accomplished through a combination of adhesion and micro-
keying.
Figure 1 graphically presents some of these anchor types, while details of the
anchoring principles and assembly are discussed below. Figure 2 graphically represents
the main anchoring mechanisms.
Cast-in-place anchors: they can be inserts of several different shapes (ribbed bars,
threaded sleeves, headed studs, etc.) that are installed in advance during the assembly
of the formwork.
Undercut anchors: in most cases for the installation of undercut anchors it is
required that either the borehole is formed appropriately by means of special drilling for
an expanding part of the anchor to nest, or that the anchor can itself form an appropriate
tooth in concrete.
Expansion anchors: the functional concept of such systems is to expand a section
(bolt type) or a longer part of the body (sleeve type) so that they can develop com-
pression against the walls of the boreholes and activate frictional forces. These
expansion systems can be either torque-controlled, in which expansion forces are
mechanically driven through screwing, or displacement controlled, where expansion is
achieved by hammering a steel cone-shaped part through the outer steel sleeve of the
anchor or vice versa. Plastic anchors employ the same installation and functional
principle, by means of a plastic expansion sleeve.
Bonded anchors: they are usually divided to capsule systems (the components of
the bonding mortar are enclosed in separate compartments of a capsule and are mixed
in the borehole by driving in the anchor) and injection systems (the mortar is filled in
the borehole in liquid phase before setting the anchor). Another classification of bonded
anchors can be made on the bonding material, which is usually a two-component
mortar consisting of organic (polyester, vinylester, epoxy) substances, inorganic
(cement-based) ones or even a combination. Special bonded anchor systems (usually
intended for use in cracked concrete) may employ functional principles of expansion
anchors because of their particular cone-shaped body, or of undercut anchors by means
of specially shaped boreholes.
Laboratory Investigations on the Installation of Fasteners 803
Concrete screws: the installation of concrete screws demands initially drilling a

borehole in the concrete surface and then driving in the screw by use of the drilling
machine. These elements have properly prepared threads in order cut indentations in
the walls of the borehole and provide a firm interlock throughout their length.
Power actuated anchors: they are nails of high-strength steel directly installed in
concrete by use of a setting tool applying pneumatic, electrical or explosive powder
energy.
1.2 Status in Design and Product Harmonisation for Concrete Fastenings

The design of fastenings, particularly for safety critical applications, is primarily based
on the “European Technical Product Specification” (ETPS) for each fastening product.
Essentially, the ETPSs deliver information about the fabrication, installation, and
associated performance for each product. For concrete fastening products, the ETPS is
the “European Technical Assessment” (ETA) which is prepared on the basis of a
“European Assessment Document” (EAD). The EAD (also known as ETAG until
recently) typically prescribes a testing campaign for the product assessment. The ETPS
also confirms the “Declaration of Performance” by the product manufacturer and
allows for the product’s CE marking. A similar procedure is followed in the U.S. with
the International Code Council (ICC) and the Evaluation Service Reports (ESR).
To that end, fastening products are primarily characterised for their structural safety,
resistance to fire, and durability. Another important aspect in the technological devel-
opment of fastening products is the facilitation of safe installation, because installation
defects can substantially decrease the fastening’s structural performance. This is dis-
cussed on the example of bonded anchors by [2], while it is also evident by the fact that [1]
recommends an increased resistance partial safety factor for anchorages with potentially
low installation quality. Typically, the installation characteristics in order to guarantee the
performance declaration are dictated in a product’s ETA. These characteristics include
among other things, the borehole preparation, geometry, and cleaning, the applicable
drilling tools and the drill bit characteristics, the allowable substrate and fixture properties,
the tightening torque, and the minimum allowable anchor spacing and the anchor’s
distance from the concrete edge. The above provisions and the associated approval
procedures typically refer to unreinforced or rebar reinforced concrete, while the industry
often comes confronted with the design of fastenings in FRC.
Fig. 1. Various fastening systems in concrete: (a) concrete screw, (b) undercut anchor,
(c) expansion anchor - sleeve type, (d) expansion anchor – bolt type, (e) cast-in-place ribbed and
deformed bar, (f) cast-in-place headed stud, (g) bonded anchor with threaded rod, (h) special
bonded anchor.
Fig. 2. Distinctive load bearing mechanisms of axially loaded anchors in concrete.
1.3 Overview of the Paper

The focus of this paper lies in the influence of fibre in concrete with respect to the
installation safety and defects of post-installed anchors. An overview of the fastenings
engineering discipline is provided and issues related to the topic of this contribution are
highlighted. Furthermore, outstanding previous studies on the performance of fasten-
ings in fibre reinforced concrete are summarised. The designated tests and results
carried out at the TU Dortmund are presented, including tests on bonded anchors,
setting tests of expansion anchors close to an edge, and hammer drilling surveys.
Finally the findings are summarised with the intention to contribute to practical
engineering design and constructability issues. Table 1 presents the geometrical and
mechanical properties of the used concrete specimens.
Table 1. Geometrical and mechanical properties of the specimens

Concrete type Plain (A1) Fiber reinforced (A2)
Concrete strength class C35/45 C35/45
Geometry w h l [mm] 900 350 1100 900 350 1100
Fibre content [kg/m3] – 50
Cube compressive strength (150 mm, 6 no. per batch)
Average compressive strength [N/mm2] 49.55 51.32
Coefficient of variation [%] 5.9 2.8
4-point flexural strength (150 150 700 mm, 3 no. per batch)
Average flexural strength [N/mm2] 4.71 6.05
Coefficient of variation [%] 3.8 1.6
2 Previous Studies for Fastenings in Concrete
Extensive research has been performed on fibre reinforced concrete worldwide, ranging
from material and fibre analysis to large scale applications, design methods, and
codified standards. A substantial number of studies addresses the rebar bond perfor-
mance in FRC. [4] provides a borad literature review on bond strength and splitting
failure for rebar, in relation to various fibre properties and dosages. Rebar pull-out tests
and large scale beam bending tests are addressed, which show a capacity increase of up
to 55% depending on the fibre content and geometry and up to 150% for special mixes.
Furthermore, significant studies have been carried out with respect to the influence of
fibre reinforcement on the load bearing of fastenings in fibre reinforced concrete. [5]
and [6] present investigations of headed bolts with diameters of 8 to 12 mm in various
embedment depths, in shelled steel fibres reinforced concrete with normal strength. The
presented tests show a rise in tensile breakout capacity of up to 35% depending on the
test parameters. Steel fibre reinforced concrete and bonded anchors of typical sizes used
in building construction have been tested per ASTM-E488 and presented in [7], stating
a minor, “non-significant” influence of the fibres in the load bearing resistance. 95 axial
tests on different types of anchors (expansion, undercut, bonded) with a 10 mm
diameter and short embedment (50, 60 mm) are presented in [8]. In these tests waved
sheet cut and hook-ended steel fibres with a dosage: 50 kg/m3 had been mixed in
C20/25. The study concludes that there is negligible difference in resistance mean
values, whereas anchorages in FRC have lower characteristic values (higher dispersion)
than in normal concrete. [9] presents investigations on transverse loads toward a free
concrete edge in normal strength concrete with hooked ended steel fibres at 25–
35 kg/m3. The tests have been carried out according to European standards and showed
that the concrete edge breakout strength can be highly increased by adding steel fibers
to the concrete, in particular for small edge distances (up to 50% increase). Normal
strength concrete with steel hooked ended fibres at a dosage of 25 and 60 kg/m3 has
been used for tests (also per European standards) in various types of M10 and M12
anchors with embedments of 65 and 70–75 mm respectively as reported in [10]. The
tests did not show that the failure load is substantially increased, but that the failure
mode changes with a higher fibre content, from pure concrete failure to pull-out or
mixed failure modes. Normal and High-strength concrete reinforced with 80 kg/m3
hooked-ended fibres has been used in tensile tests reported in [11], for large diameter
headed anchors. The failure was concrete cone mode and most of the tests also
exhibited radial cracking/splitting, for a load increase of 25% for normal and 40% for
high strength concrete, and significantly higher ductility, as compared to plain concrete.
The effect of fibres on anchor group and load redistribution within the individual
anchors is included in [12], presenting tensile tests in concrete with a strength of
65 MPa reinforced with 50 kg/m3 hooked-end fibres. All tests led to concrete cone
failure, indicating an 18% increase in failure load for single anchors and 25–40% for
anchor groups, as compared to tests in concrete without fibres. Further investigations
presented in [13], additionally conclude that concrete cone, concrete edge, and pre-
sumably side blow out capacity can be increased by up to 25%, subject the dosage of
fibres. However, this approach is deemed reliable, only if the anchorage load is initiated
at an embedment depth (axial) or edge distance (shear load) at a distance of 1.7 times
the fibre length (Lfiber) from the concrete surface. The study also proposes that, based
on the findings, the critical edge distance for tension loads (nominal distance where the
concrete boundary effects cease) may be reduced yet not below 1.7 Lfiber. Tests on cast-
in-place anchor channels under either tensile or shear loads [14] have also indicated
higher ultimate loads and larger corresponding displacements for the latter mix. As
seen in the studies above, the improved ductility of fastenings in FRC is becoming
evident.
When it comes to the load capacity of fastenings, conclusions from the previous
studies are varying, witnessing beneficial, neutral, or even adverse influence of the fibre
reinforcement. In all cases, the item of installation reliability and possibly the miti-
gation of consequences of defects in installation is not specifically elaborated with
respect to fibre reinforcement. The investigations presented below, attempt to address
two significant aspects of installation, and common causes of defects in installation,
namely the borehole cleaning for bonded anchors, and the torqueing level of expansion
anchors close to a free edge.
3 Experimental Investigations on Borehole Cleaning
To assess the influence of the steel fibres on the bond strength, nine confined pull-out
tests were carried out in unreinforced and concrete reinforced with 60 mm long steel
hooked end fibres at a dosage of 50 kg/m3. The concrete cube strength was 49.5 MPa
and 51.3 MPa for the plain and fibre reinforced concrete respectively. For the tests M12
threaded rods and an injectable hybrid mortar with approval for concrete applications
were used. The tests were calculated to induce a pull-out type of failure, in which the
bond and shear strength of the injection mortar are the determining factors. The tight
confinement by means of a 40 mm thick steel plate, on which the hydraulic cylinder is
located, exerts pressure on the concrete surface in the immediate vicinity of the anchor
during the test to prevent concrete breakout. To pursue a constant stress distribution
over the length of the anchor, without stress peaks toward the surface, the upper 20 mm
of the rods was wrapped with PTFE tape in several layers, which prevented the mortar
from locking into the thread of the anchor rod and also prevents the formation of a
load-bearing mortar layer. The boreholes were clustered into two groups. Series 1 were
cleaned following a procedure of blowing out the dust twice with compressed air,
brushing twice with a steel brush, and finally blowing out twice again. Series 2 did not
undergo particular removal of the drill dust from the borehole surface, and it was only
ensured that the sediment did not reduce the intended anchoring length; this procedure
understandably leads to a substandard and potentially defective installation. The dril-
ling was carried out by a hammer drill with a four-edge drill bit in all tests. The tests
were displacement-controlled with a loading rate between 0.01 and 0.02 mm/s.
The load bearing curves achieved are shown in Fig. 3. The graph developments
allow to qualitatively characterise the failure situation for each bonded anchor [15]. In
adequately cleaned boreholes, pull-out tests have failed at the interface between the
anchor rod and the mortar, due to shearing of the mortar material. A somewhat different
pattern is exhibited for tests in defect installation with uncleaned boreholes. Tests
A1-3x and A2-1x show a failure between mortar and borehole wall, whereby the bond
strength at the concrete interface is higher than the sliding friction forces during pull-
out. A mixed failure can be implied by tests A1-1x, A1-4x, A2-2x and A2-3x, where
the load bearing throughout the test relies on both friction and concrete bond strength.
Failure in tests A1-2x, and A2-4x is likely involving material shearing failure, while the
latter appears to be entirely unaffected by the lack of borehole preparation. One average
though, a significant loss of load bearing capacity is shown in tests with defective
boreholes. In normal concrete, the reduction is in the range of 33%, and in FRC the
reduction rises to 39% (excl. test A2-4x). Considering these differences, as well as the
dispersion of load levels and modes of failure, it is statistically arduous to attribute a
correlation between the presence of fibres in the mix and the borehole interface char-
acteristics. Regarding adequately prepared boreholes, the pull-outs in unreinforced
concrete exhibit a slightly higher mean value but also variation coefficient (49.4 kN
and 9%) as compared to the ones in FRC (48.4 kN and 2%).
Fig. 3. Load displacement curves of confined pull-out tests.
4 Experimental Investigations on Torque
Fasteners were set close to a free edge, and they exerted transverse stresses on their
substrate by torqueing the head of the anchor. As a consequence, a lateral concrete
breakout occurs as the expansion force increases. Understandably, this depends on the
material capacity of the concrete, the level of torque, and the distance from the edge. As
per each product’s specification, a minimum spacing and edge distance, and the torque
to be applied are defined, beyond which these effects become critical. A steel torque-
controlled expansion anchor was used for this test, but an edge distance smaller than
the minimum allowable was implemented in order to mimic the situation of a deviant
installation. In particular, the minimum edge distance for the product used was 70 mm
for an anchor spacing of 160 mm [3], while the anchors were positioned at a spacing of
200 mm, but an edge distance of only 30 mm from the edge. This distance is assumed
to be the boundary of the wall-effect for the fibre distributions where fibres are expected
to be isotopically distributed, acc to. The recommended torque for this application is
50 Nm, which can be observed on the dial of the torque wrench. As a fixture below the
washer a 5 mm thick metal plate on top of a PTFE sheet was used. The concrete mixes
compared are identical to the ones discussed in the previous section. The failure mode
mostly observed in plain concrete tests resembled a side blow-out failure rather than a
concrete edge failure, while all tests in FRC exhibited cracking with a small angle to
the concrete surface at the fixture level (see Fig. 4). Only the applied torque was
measured through an electronic torque sensor installed between the key and the bolt
head. The rotation was measured in 45° steps and a rotation pace as constant as
possible. Thus, a maximum value was recorded for each 45° rotation. A curve
enveloping all measurement results connects the maxima of the applied torque and
allows comparisons between the individual tests and the two test specimens. Assuming
that the sleeve expansion responds symmetrically and uniformly to the tightening of the
anchor, a consideration of the torque over the rotation course is representative of the
horizontal deformation.
A direct comparison of the tests in non-reinforced concrete with those in fibre-
reinforced concrete shows that with the same edge distance, approximately twice as
much torque can be applied until the concrete develops visible cracking. A further
crack propagation leads to a reduction in the torque required to expand the anchor
sleeve further. As a result, the measurement curve of each test shows strong differences
in gradient. The orientation of the wings of the expansion sleeve in relation to the
concrete edge was found to have a major influence on the test results, regardless of the
fibre inclusion. Expansion of a single wing orthogonally to the edge leads to a lower
failure torque compared to expansion e.g. in parallel to the edge direction.
Measurement curves in non-reinforced concrete are smoother than those in fibre-
reinforced concrete, as the fibres in concrete provide an increased crack resistance and
influence the crack expansion, while consecutive crack bridging of the fibres is also
discerned in the graphs. One torque-time development for each type of concrete is
illustrated in Fig. 5. The contribution of fibres can simultaneously increase the trans-
verse stresses required for visible cracks to develop, but also to an enhanced post-
cracking load bearing. For the reduced edge distance, anchor torqueing reached a mean
of 16.5 Nm with a variation coefficient of 27%. The applied torque in fibre reinforced
concrete reached a mean of 51.7 Nm with a coefficient of variation of 38%. Conclu-
sively, the torque to failure, applied on anchors in both types of concrete for the given
reduced edge distance, cannot safely reach the minimum value of 50 Nm required in
order to comply with the product’s approval document and therefore fulfil the anchor’s
designated performance. Nonetheless, the inclusion of fibres led to a multiple capacity
of the concrete against a failure formation caused by the expansion transverse forces,
already at a very small distance to the specimen’s surface. An interesting observation in
tests with FRC, is that where an increased number of steel fibres lies in the immediate
vicinity of the sleeve, the expansion cone can be pulled through the sleeve without
completely expanding it, and the sleeve bends back toward it’s original shape (Fig. 4 -
right). Except for this, crack bridging between the anchors tested in FRC was observed,
which suggests that the spacing recommended for plain concrete may not directly apply
for anchors in plain concrete.
5 Analysis of Steel Fibre Separation by Hammer Drilling
According to [16], a distinction is made between four different drilling methods in rock
drilling technology: “rotary drilling”, “percussive drilling”, “combined drilling” and
“thermal drilling”. Hammer drilling of concrete is assigned to percussive drilling, since
the material separation is achieved by the notch effect of the drilling tool [16]. The
drilling tool is provided with an impact impulse. An impact impulse is applied on the
drill bit. This energy must be sufficient to exceed the compressive strength of the
concrete and thus lead to material disintegration [17]. Due to the notching mechanism,
tensile and compressive stresses occur in the effective area. In the primary cutting zone,
the concrete is crushed and many micro-cracks occur, while in the secondary cutting
zone, cracks occur in the concrete due to the axial cutting force, which is responsible
for the formation of chips [18]. When machining reinforced concrete by hammer
drilling, contact with steel rebar reinforcement should generally be avoided. In the case
of ductile, carbon-steel fibre reinforcement, drilling progresses but no chip formation
occurs. As a result of the percussive principle of action, plastic deformations occur with
the drillings action. The material is deformed in front of the cutting edges of the tool
until it is sheared off. Hammer drilling of ductile carbon-steel reinforcement is therefore
more similar to forging than to cutting [19]. Plastic deformations of the steel fibres can
be observed with hammer drilling such a concrete material. As a result of the plastic
deformation, enlargements of the cut surface of the steel fibres occur, see Fig. 6 (left).
In addition to cross-sectional expansion, the steel fibres are partially detached from the
concrete and local breakouts occur in concrete, as seen in Fig. 6 (middle). Some fibres
are not separated at all or they experience only a slight material separation. These fibres
are plastically deformed in the direction of the drilling tool rotation (Fig. 6 - right).
Fig. 4. Aspects of testing torque application for the installation of expansion anchors: circular
side breakout due to torque on expansion anchor – unreinforced concrete (left); edge breakout
due to torque – fibre reinforced concrete (middle); cone pull-through with bend-in of the sleeve
due to restraint from fibres.
Fig. 5. Indicative results of torque measurement of expansion anchors being set in unreinforced
(left) and reinforced concrete (right) close to a free edge.
Fig. 6. Steel fibre separation by hammer drilling: (left) plastic deformation of steel fibres;
(middle) chipping-off of the concrete surrounding the fibre; (right) plastic deformation in the
direction of rotation.
6 Conclusions
The investigations presented herein focus on the aspect of fastenings installation in

reinforced concrete. Furthermore, the inclusion of fibres in concrete as regards the
performance of anchors with substandard installation configurations and defects also
comes to question. It is noted, that the investigated parameters are limited, and a
quantification and generalisation of the findings can only be facilitated by further
research. Based on the described set of tests, the following conclusions can be
summarised.
• The bond performance of bonded anchors in normal strength concrete, installed by
injection of a hybrid mortar according per the recommended procedures is not
strongly influenced by steel fibres in the mix.
• Substandard preparation of the boreholes for injection anchors leads to a significant
reduction of the confined pull-out resistance. A somewhat different performance of
bonded anchors under pull-out is observed between plain and steel fibre reinforced
concrete.
• The positive influence of steel fibres on the resistance to lateral concrete failure due
to installation of torque-controlled expansion anchors became evident. This effect
was observed for anchors positioned at already very small distances to the free edge
(30 mm).
• In consequence to the above, a reduction of the minimum edge distance may be

considered for anchors in fibre concrete. The minimum spacing in such a case needs
to be accordingly increased.
• The installation performance of the expansion anchor shows a strong interaction
with localised random occurrences such as the orientation of the sleeve wings to the
free edge, or a concentration of fibres adjacent to the anchors.
• An analysis of the steel fibres after hammer drilling shows that the fibres are
plastically deformed and undergo enlargements of the cut surface. In addition, the
concrete surrounding the fibres breaks out in many cases. Some fibres are plastically
deformed in the direction of the hammer drill rotation and exhibit no or only slight
material separation.
Acknowledgements. The authors express their appreciation for the support of Hilti AG and
Bekaert GmbH with the provision of material used in this study. The views expressed in this
paper are those of the authors and do not necessarily coincide with those of the organisations
mentioned above.
References
1. Technical Committee 250/European Committee for Standardisation (CEN/TC 250). EN
1992-4:2018 Eurocode 2. Design of concrete structures. Design of fastenings for use in
concrete (2018)
2. Grosser, P., Fuchs, W., Eligehausen, R.: A field study of adhesive anchor installations.
Concr. Int. 33(1), 57–63 (2011)
3. European Technical Assessment, ETA-02/0042 of 07/09/2015: Torque-controlled expansion
anchor, made of galvanised steel, for use in concrete: sizes M8, M10, M12, M16, M20 and
M24
4. ACI Committee 408: ACI 408R-03 Bond and Development of Straight Reinforcing Bars in
Tension. American Concrete Institute, (2003)
5. Al-Taan, S.A., Mohammed, A.A.: Tensile strength of short headed anchors embedded in
steel fibrous concrete. Al-Rafidain Eng. J 18, 35–49 (2010)
6. Al-Taan, S.A., Mohammed, A.A., Al-Jaffal, A.A.: Breakout capacity of headed anchors in
steel fibre normal and high strength concrete. Asian J. Appl. Sci. 5, 485–496 (2012)
7. Gesoglu, M., Ozturan, T., Ozel, M., Guneyisi, E.: Tensile behavior of post-installed anchors
in plain and steel fibre-reinforced normal-and high-strength concretes. ACI Struct. J. 102(2),
224 (2005)
8. Holschemacher, K., Wittmann, F., Klug, Y.: Structural behaviour of fastenings in steel fibre
reinforced concrete. In: Yuan, Y.S., Shah, S.P., Lü, H.L. (eds.) International Conference on
Advances in Concrete and Structures, RILEM Publications, pp. 939–946 (2003)
9. Grosser, P.R.: Load-bearing behavior and design of anchorages subjected to shear and
torsion loading in uncracked concrete (2012)
10. Kurz, C., Thiele, C., Schnell, J., Reuter, M., Vitt, G.: Tragverhalten von Dübeln in
Stahlfaserbeton. Bautechnik 89, 545–552 (2012)
11. Nilforoush, R., Nilsson, M., Elfgren, L.: Experimental evaluation of tensile behaviour of
single cast-in-place anchor bolts in plain and steel fibre-reinforced normal- and high strength
concrete. Eng. Struct. 147, 195–206 (2007)
12. Bokor, B., Tóth, M., Sharma, A.: Influence of steel fiber content on the load-bearing capacity
of anchorages in concrete’. In: Proceedings, 3rd International Symposium on Connections
between Steel and Concrete (ConSC2017), Stuttgart (2017)
13. Toth, M., Bokor, B., Sharma, A.: Anchorage in steel fiber reinforced concrete–concept,
experimental evidence and design recommendations for concrete cone and concrete edge
breakout failure modes. Eng. Struct. 181, 60–75 (2019)
14. Mahrenholtz, C., Ayoubi, M., Müller, S., Bachschmid, S.: Tension and shear performance of
anchor channels with channel bolts cast in Fibre Reinforced Concrete (FRC). In: IOP
Conference Series: Materials Science and Engineering, vol. 615, no. 1, p. 012089. IOP
Publishing (2019)
15. Meszaros, J.: Das Tragverhalten von Einzelverbunddübeln unter zentrischer Kurzzeitbelas-
tung. Universität Stuttgart (2001)
16. DIN 20301: Gesteinsbohrtechnik Begriffe, Einheiten, Formelzeichen. Beuth, Berlin (1999)
17. Müller, C.: Einfluss der Stahlfaserzugabe auf die Spannungs-Dehnungs-Linie von Beton
unter Druckbeanspruchung. Influence of steel fibers on the stress strain relationship of
concrete under compression. Beton-und Stahlbetonbau 112(4), 228–237 (2017)
18. Weinert, K., Michel, O., Gillmeister, F.: Schneidengestalt und Leistung von Hammer-
bohrwerkzeugen. BMT, Baumaschine + Bautechnik (2), 74–78 (1994)
19. Moseley, S.G., Peters, C., Momeni, S.: Near-surface phenomena occurring in cemented
carbides with different binders during rotary-percussive drilling of reinforced concrete: FE
simulation and microstructural investigations. Int. J. Refract Metal Hard Mater. 86, 105–115
(2020)
Quality Control
Applications of Statistical Process Control
in the Evaluation of QC Test Data for Residual
Strength of FRC Samples of Tunnel Lining
Segments
Chidchanok Pleesudjai1, Devansh Patel1, Mehdi Bakhshi2,

Verya Nasri2, and Barzin Mobasher1(&)
1
School of Sustainable Engineering and Built Environment,
Arizona State University, Tempe, AZ, USA
barzin@asu.edu
2
Lead Tunnel Engineer, AECOM, New York, NY, USA
Abstract. Statistical process control procedures are widely applied to improve

the production efficiency of industrial products. Application of quality control
procedures in monitoring the production, delivery, and construction process are
essential, especially when the historical data collected on various projects can be
used to gain better insight to the operational procedures. Such information will
promote interaction; reduce the liabilities and benefits the owners and suppliers
if any of the design parameters that are below the minimum requirement can be
identified as soon as possible. The longer time it takes to detect discrepancies in
the data, the more the penalty, project delays, and the higher the associated costs
to the owners and suppliers. A detailed analysis of the test results of flexural
closed-loop control test data conducted in accordance to ASTMC1609 test is
conducted for QC requirements of precast segment in tunnel lining project.
More than 378 production samples are recorded. The data set contained 1 day
(demolding age) and 28 days of molding. The statistical process control and the
range of the data are studied in the context of material properties as well as back-
calculation of the tensile parameters. The number of parameters evaluated
includes flexural strength and deflection at ultimate flexural strength, the
residual strength results at L/600, L/300 and L/150. A series of statistical
analysis procedures to analyze the correlation of the parameters and addressed
the combination of control charts for early detection of spurious shifts in the
mean of the test data (out-of-control signal).
Keywords: Statistical process control Fiber reinforced concrete Tunnel

lining Quality control Time series control process
1 Introduction
Statistical process control identifies the sources in the variations in manufacturing and
thus improves the process efficiency. QC methods that actively monitor the variations
in the mean of a manufacturing process can be extremely beneficial in monitoring the
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_72
816 C. Pleesudjai et al.
delivery and construction of concrete precast sections. Control charts process the data
based on the historical records and provide insight into operational procedures. They
may be instrumental in monitoring strength or durability data to optimize the operating
process by identifying potential problems that may exist. A process typically operates
under normal conditions (i.e., in control) and produces a product that satisfies the
minimum specified strength. However, changes in normal operating conditions may
cause the process to go out of control. An out-of-control process will eventually fail to
meet the minimum specified metrics due to low-quality material, operator errors, or
production changes. Early detection when this change occurs and identification of root
causes that push a process to go out of control can lead to early corrective actions.
Failure to implement and correct a cause will lead to operations that are continuously
out of control, resulting in significant and unacceptable losses. There are several
approaches for assessing the quality of concrete used in precast applications. Sykora
recommended a standard deviation and coefficient of variation correction for testing
error to determine the uniformity of the concrete strength [1].
Control charts are widely used to monitor and improve process performance.
The CUSUM chart or the EWMA control charts are far more efficient as a moving
average control tool in detecting small process shifts (1.5r or less) in comparison to
Shewhart control charts [2]. Hybrid control charts for active control and monitoring of
concrete strength were developed by Laungrungrong et al. [3] to support a tertiary role
of improving efficiency, product formulation, and mitigating conflict through a rational
acceptance/rejection criteria.
The minimum strength and sample size (n) cannot be used as the only parameters to
achieve the acceptable producer’s risk (rejecting “good” concrete) and consumer’s
risks (accepting “poor” concrete). Models can be extended to consider multiple metrics
for evaluation that are not directly tied into a single performance indicator such as a
single strength criterion. This approach is useful for comparing multiple acceptance
metrics, or variations within a plant, but is not recommended for comparing plant to
plant variations. Leshchinsky suggested a metric that is a combination of two or more
methods to improve the reliability and accuracy of concrete quality [4].
The proposed methodology presents an approach to evaluate the flexural or residual
flexural strength using the combination of either the CUSUM or EWMA control chart
with a standard run chart. The CUSUM and EWMA control charts often behave
similarly in practice, although different weight functions can apply to current and recent
past data values [5]. The CUSUM method applies a constant weight factor to the entire
historical set of data. In the EWMA method, uses an exponential weight factor applied
to the data; giving current or recent past observations more weight than older data
values. The combination of a CUSUM (or EWMA) chart and a run chart are being used
to determine the best conditions for monitoring concrete residual strength as they
pertain to the full-scale test results. The performance of the CUSUM-run chart and the
EWMA-run chart are compared to determine which chart is more useful for the con-
crete industries [6].
Applications of Statistical Process Control 817
2 Exploratory Data Background
All 378 tests of ASTM C1609 were conducted as a part of the QC requirements
mandated by AECOM for Hartford CSO tunnel project in 2018. This specific Fiber
Reinforcement Concrete (FRC) was studied for serviceability requirement by limiting
time-dependent to guarantee that the effect of creep on crack opening will satisfy
residual nominal flexural strengths. As a part of processing data, the 378 tests were
digitized and replotted in terms of load vs. deflection and flexural strength vs.
deflection with consistent units by ASU graduate student.
The test results were individually plotted one by one as a series of test data of
demolding age and 28-day aged from June to November of 2018. Those data have been
identified and tagged accordingly. At the same time, some of the samples were 8–10 h
on the day of casting. Those samples are identified as 1-day test, accounting for 107 test
results. The other 271 data results were tested on ages 28 days. Individual samples were
analyzed for the following parameters, Flexural Strength and deflection at 1st crack,
ultimate flexural strength and associated deflection, residual strength and toughness at
deflections associated with L/600, and L/150. The effective dimensions of the specimen
were 450 150 150 as per ASTM C1609 standard. A typical load-deflection graph
from one of the data pool is shown below in Fig. 1a. The time history of the com-
pression strength test results is shown in Fig. 1b with an average strength of 57 MPa.
(a) (b)
Fig. 1. a. A typical load vs deflection plot from ASTM1609, b. Compressive Strengths test
result at 28 days
The demolding age test is generally conducted to find the early strength of the
specimen, which is around 8 h. An average plot of the Flexural stress vs. deflection
data of all the specimens with ±1.64 standard deviations is shown in Fig. 2b. The
procedure of finding average value and standard deviation is by first interpolating all of
the time history data as discretized deflection, and then averaging the points at
deflection steps. The discretized deflection size in this paper is 0.01 mm.
Similarly, for 28 days test, an average plot of the load vs. deflection data of all the
specimens with ±1.64 standard deviations is shown in Fig. 2b shows the scattered
plot and the average curve. The average load at a first crack is around 5 MPa, which is
almost twice the average load of the demolding age specimens, at 2.5 MPa. After the
first crack, it can be observed an increase in residual stress indicating higher bond
strength between concrete and fibers after 28 days.
3 Priliminary Statistic Analysis
Statistical analyses were conducted to evaluate nominal peak strength f1 and nominal
residual strengths, fD D D
600, f300 and f150 of the response of the studied FRC mixture in
ASTM1609 as it is used in specifications for acceptance and/or rejection of samples. The
proposed approaches are based on using control charts in the evaluation of variations in
the test results. The control chart is a graph used to study how a process changes over
time. A control chart always has a central line since an average, an upper line and lower
line for the control limit. These lines are determined from historical data [7]. The
residual stress from L/600 to L/150 are the primary concern in designing fiber reinforced
concrete. In the studied data, the QC specifications report determined residual stress at
2.5 MPa as the levels of Lower Control Limit (LCL) for the demolding age and 4.0 MPa
for 28 day age. It is true that variation exists in every manufacturing environment and
damage the quality characteristic product [10]. In Fig. 2a–2b present some of individual
sample fail from LCL. Normal distribution is made as an assumption in both concrete
age. However, the actual data distribution was plotted with normal distribution in
Fig. 3a–3b. Hence, it can say that Avg.1.64SD correspond to 90% of confidential
interval. In Fig. 2a, the green line is lower than LCL by mean that confidential interval to
pass QC specification in demolding age sample is lower than 90%.
(a) (b)
Fig. 2. a. Average ASTM1609 testing plot with 107 individual samples, at demolding age, b.
Average ASTM1609 testing plot 271 individual samples, at 28 days
(a) (b)
Fig. 3. a. Normal Probability distribution, residual stress L/300, b. Normal Probability

distribution, residual stress L/600
Otherwise, the run chart is an alternative way to present the dynamic movement of
a process changes over a period of time. Data are plotted in time order as in Fig. 4a–4b.
In Fig. 4a, 15% of flexural strength at L/600 (0.75 mm), 3–4% of flexural strength at
L/300 (1.5 mm) and L/150 (3 mm) were lower from QC specification. Whereas 28
days data, lower than 1% falling from QC criteria strength. However, only the run chart
cannot indicate a small out of control process. Combining the run chart with a mon-
itoring scheme that provides information about the stability process is required. The
next section will explain two types of control process which depending on memory
data. It calls cumulative sum (CUSUM) and exponential weighted moving average
(EWMA).
(a) (b)
Fig. 4. a. Demolding age flexural strength Run chart, b. 28 days flexural strength Run chart
CUSUM and EWMA

A cumulative sum (CUSUM) control is the chart for individual observations monitors
sample measurement by distinguishing between a significant change (real change in
slope) and non-significant change (error of aberration). The general form of the
cumulative sum is given by
X
n
Ci ¼ ðxj l0 Þ ð1Þ
j¼1
Where Ci presents sum of the derivation from target values(l0) over the domain.
Day recommended using the mean strength as the target value in the CUSUM chart [8–
10]. As a result of CUSUM, the sensitive value from the target can be detected. A one-
sided upper l0, is C+ and other sided lower l0, is C−. They were plotted on CUSUM
chart will indicate the signal of process. When C+ and C− exceed Upper control level
(UCL) or Lower control level (LCL) that means that the process signals the out-of-
control. UCL and LCL are be chosen by commonly value of H which define by hr.
Ciþ ¼ max½0; xi ðl0 þ KÞ þ Ci1

þ
ð2Þ
Ci ¼ max½0; ðl0 þ KÞ xi þ Ci1

ð3Þ
The constant K is the reference value, where calculated by K = kr. Given r is a

standard deviation. In the paper using k = 0.5. The value k can be adjusted where
present how sensitive we expect to observe the signal out of control process. The out of
control signal can show the inaccurate result when we have not reset upper and lower
− +
CUSUMs to zero (Ci−1 = 0, Ci−1 = 0) which out of control signal was detected. The
comparison between CUSUM chart with and without resetting the signal is given in
Fig. 5a and 5b. At the observation 10 was observed being exceeded LCL in Fig. 4a
which did not reset C−i−1 = 0. The cumulative sum in observation 11–31 present the
biased resulting above upper control level (UCL). Whereas in Fig. 5b, the signal was
reset to zero by mean the historical data before out of control signal are independent on
recent CUSUM. For more detail discussion on using CUSUM concrete production see
in ACI 214 which provides examples of applying the CUSUM chart and discusses
some of the difficulties with the CUSUM analysis (American Concrete Institute
Committee 214, 2002) [9].
Nominal Peak Strength,(MPa) Nominal Peak Strength,(MPa)

10.0 4
3 UCL=2.617
7.5
Cumulative Sum
Cumulative Sum
2
5.0 UCL=2.62 1
2.5 0
-1
0.0
-2
-2.5 -3
LCL=-2.62 LCL=-2.617
-5.0 -4
1 11 21 31 41 51 61 71 81 91 1 11 21 31 41 51 61 71 81 91
Observation Observation
(a) (b)
Fig. 5. a. CUSUM of Peak strength at 1 day without reset signal, b. CUSUM of Peak strength at
1 day with reset signal
The exponentially weighted moving average (EWMA) control chart. The normal
equation is given by:
X
i1
zi ¼ k ð1 kÞ j xij þ ð1 kÞi z0 ð4Þ
j¼0
iP
1
Where k ð1 kÞ j xij is the weight assigned to the neighbor observations. The
j¼0
weight factor against observation has been shown in Fig. 6. It can be observed when
k = 0.3, high contribution assigned to the close observations but less contribution to
the far observations comparing to k = 0.1 and 0.2. UCL and LCL in EWMA chart are
identified in Eq. 5–7. In this paper, h was defined as 3. The process is out of control
chart is detected when the EWMA exceeds the UCL and LCL.
The test data were analyzed based on the assumption that the data were collected on a
constant time interval within the five months production period of June 2018–October
2018. While all of the testing on 1 day and 28 days were conducted in accordance with
the fabrication schedule, it is possible the climatic changes during this period could
have resulted in various curing and conditioning of the samples. The Fig. 7 shows
frequency of the data collection based on the production schedule which can be
interpreted as a uniform distribution of sampling during the period of study. It is
observed that the CUSUM and EWMA charts perform in a similar manner in order to
capture the variations in the test data. When CUSUM is used, a constant weighted
factor is applied to the data during the observation period. The EWMA approach
applies an exponential weight factor to the data that enhances recently obtained data
compared to the older data sets. This indicated that EWMA gives a higher priority and
a more significant concern in the current observation domain (neighbor observation)

than the older data.
UCL ¼ l0 þ hrb ð5Þ
LCL ¼ l0 hrb ð6Þ

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
iffi
k h 2i
b¼ 1 ð1 kÞ ð7Þ
2k
P
i1
Fig. 6. Exponentially weighted factor to the neighbor observation, k ð1 kÞ j
j¼0
Fig. 7. Sample collection data points based on the production schedule of 5 months
In the current evaluation, the control chart was applied to 270 specimens within
subgroup = 1 (all of the available data) for 28 days and 99 specimens within the
subgroup = 1 for 1-day test data. The peak strength chart is shown in Figs. 8a and 8b
and indicates that using the 1-day CUSUM chart three observations were out of lower
control limits (LCL), while EWMA control chart did not indicate any out of lower
control results, possibly because the EWMA relies on neighbor observation than the
older data. Note that the upper control limits can be ignored since exceeding the
strength results is not typically a penalizing factor. The same observation can be made
with the 1-day chart of the Nominal strength at L/600 and at L/150 as shown in
Figs. 9a–9b and 10a–10b which indicate that the process is out of control. Similar plots
for the peak strength of the 28 day data are shown in Figs. 11a and 11b. At peak
strength of 28 days, CUSUM indicates observation points 55–101 showing a signal
control process significantly falls below the LCL. The same event is also captured by
the EWMA control chart. As a result, one can further evaluate the sample response in
this period to determine the causes such as fiber distribution, mix properties, or curing
conditions may have affected the residual strength of specimens. Similarly, in Fig. 12a
and 12b nominal residual strength L/600 values also indicate a similar reduction trend
in the data range of 55–101 observation points. Whereas, the EWMA chart of nominal
strength L/150 in Figs. 13a and 13b, did not show out of control process while both
peak strength and L/600 nominal strength fell lower than LCL. This may imply that at
the high deflection levels, fiber reinforcement plays a more dominant role than the
concrete matrix in affecting the residual strength. In other words, fiber properties have
less variation than concrete matrix.
4 4.75
3 UCL=2.617
4.50 UCL=4.444
__
Cumulative Sum
2
X=3.994
1 4.25
EWMA
0
4.00
-1
-2 3.75
-3 LCL=3.543
LCL=-2.617 3.50
-4 1 11 21 31 41 51 61 71 81 91
1 11 21 31 41 51 61 71 81 91
Observation
Observation (b)
(a)
Fig. 8. a. CUSUM of Peak strength at 1 day, h = 4, k = 0.5, b. EWMA of Peak strength at 1

day, h = 3 and k = 0.1
4
3.8
3 UCL=2.360 UCL=3.662
__
Cumulative Sum
2 3.6 X=3.256
1
EWMA
3.4
0
3.2
-1
-2 3.0
LCL=2.850
-3 LCL=-2.360
1 11 21 31 41 51 61 71 81 91
2.8
1 11 21 31 41 51 61 71 81 91
(a) (b)
Fig. 9. a. CUSUM of L/600 stress at 1 day, h = 4 and k = 0.5, b. EWMA of L/600 stress at 1
day, h = 3 and k = 0.1
4 UCL=3.064 4.50 UCL=4.391

3
2 4.25
Cumulative Sum
__
1 X=3.864
EWMA
0 4.00
-1
-2 3.75
-3
-4 3.50
LCL=-3.064 LCL=3.337
-5
1 6 11 16 21 26 31 36 41 46 1 6 11 16 21 26 31 36 41 46
Observation
Observation
(a)
(b)
Fig. 10. a. CUSUM of L/150 stress at 1 day, h = 4, k = 0.5, b. EWMA of L/150 stress at 1 day,
h = 3 and k = 0.1
4 UCL=2.739 7.8 UCL=7.517

3 7.6
Cumulative Sum
2 7.4
1 7.2
__
EWMA
0 7.0
6.8 X=7.045
-1
-2 6.6
LCL=-2.739
-3 6.4 LCL=6.574
-4 6.2
1 6.0
2 8 55 8 2 10 9 136 16 3 190 217 24 4 1 28 55 82 109 136 163 190 217 244
Observation
Observation
(a)
(b)
Fig. 11. a. CUSUM of Peak strength at 28 day, h = 4, k = 0.5, b. EWMA of Peak strength at 28
day, h = 3 and k = 0.1
5.0 UCL=3.42 6.8

6.6 UCL=6.558
Cumulative Sum
2.5 6.4
6.2
EWMA
6.0 __
0.0 X=5.969
5.8
5.6
-2.5 5.4
5.2 LCL=5.380
-5.0 LCL=-3.42 5.0
1
1 8 5 2 9 6 3 0 17 4
2 5 8 10 13 16 19 2 24 28 5 5 8 2 109 136 163 19 0 2 17 2 44
(a) (b)
Fig. 12. a. CUSUM of L/600 stress at 28 day, h = 4, k = 0.5, b. EWMA of L/600 stress at 28
day, h = 3, k = 0.1
5 UCL=3.382 6.8
4 6.6 UCL=6.446
Cumulative Sum
3 6.4
2 6.2 __
EWMA
1 X=5.844
0 6.0
-1 5.8
-2 5.6
-3 5.4 LCL=5.241
-4 LCL=-3.382 5.2
1 8 5 2 9 6 3 0 17 4 1
2 5 8 10 13 16 19 2 24 22 4 3 6 4 8 5 106 12 7 148 169 19 0
(a) (b)
Fig. 13. a. CUSUM of L/150 stress at 28 day, h = 4, k = 0.5, EWMA of L/150 stress at 28 day,
h = 3 and k = 0.1
5 Conclusions
By combining the control chart and preliminary statistic analysis section, using
CUSUM or an EWMA approach, early detection of shifts in the process mean become
possible. Otherwise, the run chart and average Load-Deflection plot can be identified as
the quality to reject/accept a concrete batch but they can not measure stability in
manufacturing process. The CUSUM and EWMA can address the stability of the
production process by analysis variation in peak strength, residual strength results at
L/600 and L/150 over a period of time domain. Analysis of the combination of the
control charts can identify where the process is in or out of control or capable in
segment fabrication. The consumer can decide on accepting or rejecting a sample set.
In addition, the tunnel lining producers can use the chart to determine if the monitored
process is out of control, and it provides an opportunity to identify the possible causes
for the situation.
References
1. Sykora, V.: Amendment of appendix X2 in ASTM C 917 evaluation of cement uniformity
from a Single Source. ASTM-Cem. Concr. Aggregates 17(2), 190–192 (1995)
2. Montgomery, D.C.: Introduction to Statistical Quality Control. John Wiley & Sons, NY
(2008)
3. Laungrungrong, B., Mobasher, B., Montgomery, D.: Development of rational pay factors
based on concrete compressive strength Data. SPR 608 Arizona Department of Transporta-
tion, AZ (2008)
4. Leshchinsky, A.M.: Combined methods of determined control measures of concrete quality.
Mater. Struct. 24(3), 177–184 (1991)
5. Bakhshi, M., Laungrungrong, B., Bonakdar, A., Mobasher, B., Borror, C.M., Montgomery,
D.C.: Economical Concrete Mix Design Utilizing Blended Cements, Performance-Based
Specifications, and Pay Factors, FHWA-AZ-13-633, Final Report 633, May 2013
6. Laungrungrong, B., Mobasher, B., Montgomery, D.C., Borror, C.M.: Hybrid control charts
for active control and monitoring of concrete strength. ASCE J. Mater. Eng. 22(1), 77–87
(2010)
7. https://asq.org/quality-resources/control-chart
8. Day, K.W.: Concrete Mix Design, Quality Control and Specification, pp. 82–85. Taylor &
Francis, New York, NY (2006)
9. ACI 214R-11 Guide to Evaluation of Strength Test Results of Concrete (2012)
10. Zaman, B., Abbas, N., Riaz, M.: Mixed CUSUM-EWMA chart for monitoring process
dispersion. Int. J. Adv. Manuf. Technol. 86, 3025–3039 (2016)
Using Decades of Data to Rethink
Proportioning and Optimisation of FRC
Mixes: The OptiFRC Project
Emilio Garcia-Taengua(&)

e.garcia-taengua@leeds.ac.uk
Abstract. Fibres enhance the mechanical properties of concrete, but residual

flexural strength parameters present significant variability. The proportioning
and optimisation of FRC should integrate fresh and hardened state properties as
well as their variability, and this is the urgent challenge addressed by this project
funded by the ACI Foundation. It follows a meta-analytical, multivariate
approach, based on the creation of an exhaustive database with information on
hundreds of FRC mixes compiled from papers published over two decades and
adopting a data analytics perspective. First, the relationships between the rela-
tive amounts of the mix constituents, fibre geometry and dosage are modelled.
Then, the strong correlations between residual flexural strength parameters, limit
of proportionality, and compressive strength are exploited to develop efficient
predictive tools. The final outcome will be a software package called
“OptiFRC”. This will integrate capabilities to access the database compiled, to
visualise and utilise the models for the optimisation of FRC mix proportionings,
and to calibrate and use the derived quality control charts. This paper presents an
overview of the project, reports on its progress to date and summarises the
ongoing developments and anticipated impact.
Keywords: Characterisation Optimisation Proportioning Residual flexural

strength Variability
1 Introduction
Fibre reinforced concrete (FRC) is, by definition, any concrete made primarily of
hydraulic cements, aggregates, and discrete reinforcing fibres [1]. Fibres enhance
numerous mechanical properties of concrete, particularly tensile, flexural strength and
toughness in the cracked state. The use of FRC has evolved from the small scale to the
larger scale of routine production and field applications, with tens of millions of cubic
yards produced every year [2]. However, the perception that fibres can be treated as an
add-on to conventional concrete mixes is still prevalent amongst practitioners. For
example, ACI 544.3R-08 [3] stated that, when fibres are incorporated to the mix, “some
mixture adjustments may be required”, or “more paste may be needed to provide better
workability”. This last statement also puts the focus on fresh state performance and its
importance when proportioning FRC mixes. These aspects can no longer be addressed
relying only on simple adjustments, considering the family of special concretes that has
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_73
828 E. Garcia-Taengua
developed around FRC. Fresh state performance is crucial to fibre-reinforced self-

compacting concrete (FRSCC) [4] or ultra-high performance fibre-reinforced concrete
(UHPFRC) [5]. Their relevance and increasing prevalence in industry requires that the
proportioning of FRC mixes is regarded as a multi-objective optimisation problem,
integrating all dimensions shown in Fig. 1.
Fig. 1. FRC proportioning redefined as a multi-objective optimisation problem.
In terms of hardened state performance, the consideration of fibres as reinforcement

is now embedded in the ACI 318 code, and their structural contribution is accounted
for in design equations [6]. However, the mechanical properties of FRC, particularly
the residual flexural strength parameters, present considerable variability [7]. This is
highly relevant, as these parameters are the basis of FRC characterisation and speci-
fication. Figure 2 shows the flexural test set-up configurations according to standards
EN 14651:2005 [8] and ASTM C1609/1609M [9], and an example of stress-strain
curve, where the limit of proportionality and the residual flexural strength parameters
are defined.
Fig. 2. Characterisation of the flexural behaviour of FRC.

Using Decades of Data to Rethink Proportioning and Optimisation 829
The residual flexural strength parameters (fR1, fR2, fR3, fR4), limit of proportionality
(fL) and compressive strength (fc) present strong correlations between them and are
mutually interdependent. Treating these parameters as if they were independent vari-
ables means that part of the information obtained from characterisation tests is con-
founded with noise or experimental error. However, their variability and sensitivity to
changes in the mix design have been analysed very rarely, and only for specific fibres
in specific mixes, in separate studies. Bearing in mind that the prediction and sys-
tematic control of variability is key to the effective quality control of FRC production, a
meta-analysis of the variability of FRC mechanical properties was urgently needed.
2 Overview of the Project
2.1 Main Objectives

With the aim of addressing the aforementioned challenges, the project “Optimization of
Fiber-Reinforced Concrete using Data Mining”, funded by the Concrete Research
Council through the ACI Foundation, has the following objectives:
• To compile, pre-process, and publish an exhaustive database with mix propor-
tionings and experimental results of FRC mixes available in literature.
• To analyse the properties of FRC as a multivariate phenomenon, using data mining
techniques.
• To develop predictive models, based on the database compiled, that provide
accurate estimates of the residual flexural strength parameters of FRC and their
variability.
• To implement the outcomes from the above in a software (OptiFRC) that assists in
the visualisation and interpretation of the database and the predictive models
developed.
2.2 Methodology and Work Programme

An overall view of the methodology followed in this project is summarised in the
flowchart shown in Fig. 3. It is based on the methodological framework previously
applied by the author to the meta-analysis of self-compacting concrete mixes [10, 11].
2.2.1 Construction of the Database

Detailed information on different FRC mix proportionings and their characterisation
tests results has been extracted from previous literature and compiled in a database. The
sources of information considered for this study are those papers published in ACI
Structures, ACI Materials, and journals indexed in ScienceDirect® since 1999,
resulting from the search with the terms “fiber-reinforced concrete” or “fibre-reinforced
concrete”. The database is prepared in mineable format and consists of two datasets in
row-to-row correspondence: one with the FRC mix proportionings, and another with
the characterisation tests results. The first dataset comprises the relative amounts of the
mix constituents (in kg/m3), maximum aggregate size (in mm), fibres aspect ratio and
length (in mm), and fibre material. The second dataset includes values for the slump
Fig. 3. Overview of the methodology and correspondence with OptiFRC modules.
and average compressive strength at 28 days, in addition to fL, fR1, fR2, fR3, and fR4. To
ensure sufficient statistical power for the subsequent analysis, the database size was
defined to include between 500 and 3,000 complete cases [12].
2.2.2 Meta-analysis of FRC Mix Designs

The objective of this meta-analysis is to model the relationships between the relative
amounts of the constituents of FRC mixes and to quantify their variation with respect to
the type of fibres used, their geometry and dosage. Data mining techniques have been
used to carry out a thorough analysis of the latent structure of correlations, namely:
multiple linear regression, surface response methodology, and the fitting of conditional
probability distributions [13]. Initially the analysis has been segmented by fibre type,
and this paper presents the most salient aspects of the preliminary analysis of the
database of steel FRC (SFRC). This will be followed by a global analysis of the entire
dataset without disaggregation by fibre types, which is work in progress at the moment
of writing this paper.
2.2.3 Meta-analysis of FRC Residual Flexural Capacity

The structure of correlations and association patterns between the residual flexural
strength parameters (fR1, fR2, fR3, and fR4), limit of proportionality (fL) and compressive
strength (fc) will be analysed by means of dimension reduction techniques [14] to
compress the information represented by these six interdependent variables into a
reduced set of independent factors. Multiple linear regression and analysis of variance
will then be applied to find equations to relate them to the FRC mix proportioning, not
only in terms of average values but also to estimate their variability and sensitivity to
changes.
2.2.4 Development of Multivariate Quality Control Charts

and the OptiFRC Software
The models obtained as outlined in the previous section and the corresponding prob-
ability distributions will be used to define quality control charts that can be used in the
monitoring of continuous production of FRC mixes. In addition to univariate tools
based on Shewhart charts, CUSUM and EWMA, new multivariate charts based on
Hotelling’s T2 statistic [15] will be derived. All these outcomes will be implemented in
a standalone software package that can be used to access the database compiled, to
visualise and utilise the models for the optimisation of FRC mix proportionings, and to
use the corresponding quality control charts. It will consist of three modules: Mixes,
Trends, and Control. The correspondence of these modules with the outcomes of the
different stages of the work programme is also shown in Fig. 3.
3 Meta-analysis of SFRC Mixes
At the time of preparation of this paper, the SFRC database consisted of 770 different
cases, each case corresponding to a different SFRC mix with detailed information of
their proportioning and the characterisation tests results. They have been extracted from
more than 100 papers published between the years 2000 and 2019. The following
sections present an overview of their analysis.
3.1 Binder Type and Content

The piechart, histogram and box-and-whisker plot shown in Fig. 4 present an at-a-
glance description of the binder type and contents in SFRC mixes. In almost 42% of the
cases, the original sources did not report the type of cement. However, in 94% of the
cases for which this information was available, the cement type was either CEM I or
CEM II, and CEM I in particular was the most prevalent.
In Fig. 4 (right), it can be observed that there were two distinct frequency distri-
butions of binder contents, and therefore two sub-populations could be defined in terms
of the total binder content. In 90% of the cases, the binder content was not higher than
710 kg/m3, and the median was 450 kg/m3. The first value can be regarded as the
maximum, whilst the median can be considered as a typical, representative value, as it
corresponds to the value below which 50% of the cases are found.
Fig. 4. Overview of the binder composition and distribution of total binder content values.
In terms of the binder composition, 47.3% of the cases in the database incorporated
supplementary cementitious materials (SCMs) in addition to cement. Figure 5 shows
the histograms and box-and-whisker plots for the cement content (left) and the SCMs
content (right). For the cement content, the median was 400 kg/m3, and in 90% of the
cases it was not higher than 660 kg/m3. Regarding the SCMs content, in 90% of the
cases it was not higher than 230 kg/m3. However, SCMs contents higher than
100 kg/m3 were associated with cement contents higher than 660 kg/m3.
Fig. 5. Distribution of cement and SCMs contents.
3.2 Fibres and Aggregates

The histogram corresponding to the fibres volume fraction, in percentage, is shown in
Fig. 6 (above). In 75% of the cases, the fibres volume fraction was not higher than 1%,
and in 90% of the cases it was lower than 1.6%. When the fibre content was higher than
this, in the majority of cases it ranged between 1.8% and 2%, very rarely exceeding this
last value. In terms of the fibre dimensions, the aspect ratio ranged between 45 and 80
in 90% of the cases in the database (data not shown).
An interesting correlation between the grading of the aggregates and the fibres
content was observed. The contour plot in Fig. 6 (below) shows how the relationship
between the gravel-to-sand ratio and the total aggregates content changes for increasing
fibre contents. Cases with moderate fibre contents, that is, below 0.75% in volume, were
associated with gravel-to-sand ratios between 1.5 and 2 but were not limited to any
particular range of aggregate contents. However, increasing fibre contents were linked
with a decrease in both the gravel-to-sand ratio and the total aggregates content. This can
be attributed to the fact that the introduction of higher fibre contents requires the mix
design to be substantially more cohesive and to have a higher relative volume of paste.
Fig. 6. Distribution of fibre contents and their relationship to aggregates content and grading.
3.3 Compressive Strength and Limit of Proportionality

The SFRC database represents a wide spectrum of mixes in terms of their average
compressive strength at 28 days, as the histogram in Fig. 7 (left) shows. Similarly to
what was observed in relation to the total binder content, there are two distinct clusters
in terms of compressive strength. The main cluster represented 90% of the cases in the
database and comprised mixes with compressive strength not higher than 110 MPa,
well distributed around a median of 55 MPa.
In terms of the limit of proportionality, all the cases in the database conformed to
the same frequency distribution, as the histogram in Fig. 7 (right) shows. The values of
this parameter were below 10 MPa in most of the cases.
Fig. 7. Distribution of compressive strength and limit of proportionality.
3.4 Residual Flexural Strength Parameters

The residual flexural strength parameters were observed to follow very similar distri-
butions, and only the histogram for fR1 values is shown in Fig. 8. It can be seen that this
histogram shows a distribution which is very similar to that observed for the limit of
proportionality (Fig. 7), the main difference being its range or width of the distribution,
with 90% of the values below 15 MPa.
Fig. 8. Distribution of fR1 values and bivariate scatterplots for fR1, fR2, fR3, and fR4.
More interesting are the bivariate linear correlations between fR1, fR2, fR3, and fR4
and how clear they are. The scatterplots for fR2 versus fR1, fR3 versus fR2, and fR4 versus
fR3 are shown in Fig. 8. These plots confirm that, from a multivariate perspective, any
of these parameters can be expressed as a direct function of the others.
As a preliminary analysis concerned with the sensitivity of these parameters with
respect to the mix proportioning, a multivariate regression analysis was carried out for
fR1. The following variables were considered: cement, SCMs, sand and gravel contents
in kg/m3; cement-to-water and SCMs-to-water ratios; fibres volume fraction in per-
centage, fibre length in mm, and fibre aspect ratio. It is worth noting that not all the
variables related to the mix proportioning were used in this regression analysis, as this
was intended to be a simple, exploratory model. Despite that and the fact that the data
come from many different sources, this analysis yielded a model with a high goodness
of fit, the R-squared being 0.83.
Fig. 9. Contour plots for fR1 in SFRC mixes, based on a preliminary model.
Figure 9 shows some of the contour plots corresponding to the model obtained for
fR1, and in all cases the vertical axis represents the fibre volume fraction, in percentage.
Figure 9(a) shows that residual flexural strength can be improved not only by
increasing the dosage of steel fibres, but also by increasing the cement content. This
model predicts that, on average, the effect of increasing the fibre volume fraction in
0.1% is equivalent to that of increasing the cement content in 50 kg/m3, from the point
of view of fR1 values.
The relationship between these parameters and water content is completed with the
contour plot in Fig. 9(b), which shows that the effect of any amount of steel fibres on
fR1 is affected by the cement-to-water ratio of the mix. In fact, this model predicts that,
on average, the cement-to-water ratio which can be considered optimal in enhancing
the effect of fibres is 2.5, which is equivalent to a water-to-cement ratio of 0.40.
A similar observation can be made with respect to the sand content. According to
the contour plot shown in Fig. 9(c), this model predicts that, on average, sand contents
close to 950 kg/m3 are best in maximising the performance of fibres in terms of net
increase in fR1 per unit volume of fibres. In fact, any of these contour plots can be used
as double-entry graphs to obtain different pairs of values associated to the same pre-
dicted performance, as the contour lines correspond to constant fR1 values.
Finally, the effect of the fibres dimensions is partially explored through the inter-
action between aspect ratio and volume fraction, in Fig. 9(d). This contour plot reflects
the fact that residual flexural strength of SFRC can be improved by increasing either
fibre content or aspect ratio, or a combination of both. The relationship between them
is, however, not linear, and therefore different combinations of aspect ratio and volume
fraction can be adequate depending on the fR1 value adopted as target.
Of course, the contour plots presented in Fig. 9 are only partial examples of a
preliminary model based on SFRC data, but they illustrate how the software tool
resulting from the OptiFRC project will present the information and predictive models
for optimisation purposes.
This paper presents an overview of the ongoing project “Optimization of Fiber-

Reinforced Concrete using Data Mining” and its progress to date. The most salient
aspects can be summarised as follows:
• The main objectives of this ongoing research project are: to compile an exhaustive
database of FRC mix proportionings and their properties; to analyse their variability
using data mining techniques; to develop robust models for estimating the residual
flexural strength; and to implement these developments in an optimisation tool
(software OptiFRC).
• Detailed information on hundreds of FRC mixes has been compiled from papers
published in indexed journals and conference proceedings published since 1999.
The database contains: relative amounts of mix constituents, maximum aggregate
size; fibres aspect ratio, length, material, and fibre content; slump and compressive
strength; and fL, fR1, fR2, fR3, and fR4 values.
• A preliminary analysis has been carried out on a dataset of SFRC mixes consisting
of 770 cases, extracted from more than 100 papers published over the last two
decades. It was found that the fibre volume fraction was typically below 1%, and
very rarely exceeded 1.6%.
• Cements type CEM I and CEM II were found to be the most common in SFRC
mixes. In almost half of the cases compiled, SCMs were used. In 90% of the cases,
the total binder and cement contents were below 710 kg/m3 and 660 kg/m3,
respectively. In 50% of the cases, total binder and cement contents were not higher
than 450 kg/m3 and 400 kg/m3, respectively.
• SFRC mixes with fibre contents up to 0.75% in volume were associated with
gravel-to-sand ratios of 1.5 and above. Increasing fibre contents were linked with a
decrease in both the gravel-to-sand ratio and the total aggregates content.
• Regarding the average compressive strength at 28 days, it was not higher than
55 MPa in 50% of cases, and values higher than 110 MPa corresponded to only
10% of the SFRC mixes compiled. All residual flexural strength parameters fol-
lowed a lognormal distribution, and it was confirmed that the linear correlation
between them was very strong.
• A preliminary analysis of the fR1 values with respect to the SFRC mix proportions
revealed that, to optimise the efficiency of fibres in improving the residual flexural
strength, a water-to-cement ratio of 0.40 and sand contents around 950 kg/m3 were
optimal.
Acknowledgements. The author wishes to thank the Concrete Research Council, the ACI
Foundation and the American Concrete Institute (ACI) for the financial support awarded to the
project “Optimization of Fiber-Reinforced Concrete using Data Mining” (2019–2021). The
endorsement of fellow members of ACI Committee 544 as well as the support of industrial
partners AECOM, OwensCorning and Banagher Concrete is also acknowledged.
References
1. ACI Committee 544: ACI 544.1R-96 Report on Fiber Reinforced Concrete. American
Concrete Institute (1996)
2. ACI Committee 544: ACI 544.2R-17 Report on the Measurement of Fresh State Properties
and Fiber Dispersion of Fiber-Reinforced Concrete. American Concrete Institute (2017)
3. ACI Committee 544: ACI 544.3R-08 Guide for Specifying, Proportioning, and Production
of Fiber-Reinforced Concrete. American Concrete Institute (2008)
4. Ferrara, L., Park, Y.D., Shah, S.P.: A method for mix-design of fiber reinforced self-
compacting concrete. Cem. Concr. Res. 37(6), 957–971 (2007)
5. Habel, K., Viviani, M., Denarié, E., Bruhwiler, E.: Development of the mechanical
properties of an ultra-high performance fiber reinforced concrete. Cem. Concr. Res. 36(7),
1362–1370 (2006)
6. ACI Committee 544: ACI 544.4R-18 Guide to Design with Fiber-Reinforced Concrete.
American Concrete Institute (2018)
7. Cavalaro, S.H.P., Aguado, A.: Intrinsic scatter of FRC: an alternative philosophy to estimate
characteristic values. Mater. Struct. 48(11), 3537–3555 (2014). https://doi.org/10.1617/
s11527-014-0420-6
8. European Committee for Standardization: EN 14651:2005. Test method for metallic fibre
concrete. Measuring the flexural tensile strength (limit of proportionality (LOP), residual)
(2005)
9. ASTM International: ASTM C1609/C1609M. Standard Test Method for Flexural Perfor-
mance of Fiber-Reinforced Concrete (Using Beam with Third-Point Loading) (2012)
10. Garcia-Taengua, E., Marti-Vargas, J.R.: Multivariate analysis of the fresh state parameters of
self-consolidating concrete. In: Proceedings of the 8th International RILEM Symposium on
Self-Compacting Concrete SCC2016, pp. 221–231 (2016)
11. Garcia-Taengua, E.: Fundamental fresh state properties of self-consolidating concrete: a
meta-analysis of mix designs. In: Advances in Civil Engineering, p. 5237230 (2018)
12. Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E.: Multivariate Data Analysis. Pearson
Education Inc., Upper Saddle River (2010)
13. Button, K.S., Ioannidis, J.P.A., Mokrysz, C., Nosek, B.A., Flint, J., Robinson, E.S.J.,
Munafò, M.R.: Power failure: why small sample size undermines the reliability of
neuroscience. Nat. Rev. Neurosci. 14, 365–376 (2013)
14. Jolliffe, I.T., Cadima, J.: Principal component analysis: a review and recent developments.
Philos. Trans. Roy. Soc. A 374(2065), 20150202 (2016)
15. MacGregor, J.F., Kourti, T.: Statistical process control of multivariate processes. Control
Eng. Pract. 3(3), 403–414 (2016)
Case Studies: Structural and Industrial
Applications
Fire Resistance of Steel Fibre Reinforced
Concrete Elevated Suspended Slabs: ISO Fire
Tests and Conclusions for Design
Xavier Destrée1(&), Andrejs Krasnikovs2, and Sébastien Wolf3

1
R&D ArcelorMittal Fibres, Bissen, Luxembourg
xavier.destree@scarlet.be
2
Academy of Sciences, Riga, Latvia
3
ArcelorMittal Fibres, Bissen, Luxembourg
Abstract. Since 2004, a number of projects of multi-storey buildings have been

completed in various countries where steel fibres are the only reinforcing of the
cast in-situ elevated slabs. A number of full-scale tests at cold stage of these steel
fibre reinforced concrete slabs have been conducted to show a very considerable
safety margin between the most onerous service loading conditions and the
ultimate cases at final collapse.
The positive influence of steel fibre reinforcing on the material properties of
the concrete under fire conditions is already well known and detailed in the
literature dating back 25 years. In 2015, the ACI 544-6R 15 document, “Report
on Design and Construction of Steel-Fiber Reinforced Concrete Elevated Slabs”
outlines in detail the design method of such slabs but mentions also that further
research is needed in the Fire Resistance area.
The purpose of the paper here is precisely to address the Fire Resistance of
S.F.R.C suspended elevated slabs and walls in full scale under service loadings.
The paper outlines the setting-up of the tests of a number of elevated slabs and
walls, subjected to the real loading conditions, the observations and then to
conclude and confirm by the high positive impact on the design as well as the
high safety of these slabs and walls.
Keywords: Fire resistance Steel fiber Concrete Slabs Walls
1 Introduction
The steel fibre reinforced concrete as a material has been investigated by many authors
in the last 35 years under high temperature starting from 200 °C to 1100 °C.
The A.C.I. report 544.5R-13 on “Physical properties and durability of fibre-
reinforced concrete” concludes that “Reinforcement in the form of steel fibres only,…,
generally improves the performance of structural concrete members under extreme
temperature and fire” and further “… fibres have been successful in extending the safe
time of fire exposure for many practical and proven applications. Extending the safe
time of fire exposure allows firefighters more time both to evacuate structures and to
extinguish the fire safely”
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_74
842 X. Destrée et al.
Kodur (1995) observed that “results indicate that the compressive strength at ele-
vated temperatures of steel-fibre-reinforced concrete is higher than that of plain con-
crete. It is concluded that the mechanical properties of fibre-reinforced-concrete are
more beneficial to fire resistance than those of plain concrete. The steel fibres prevented
early cracking and also contributed to the compressive strength of concrete at elevated
temperature”
One reason is that the ductility of normal strength concrete with steel fibres is
relatively stable despite the increase of temperature according to Widhianto et al.
(2014).
The presence of steel fibres increases also the ultimate strain as outlined also by the
Canadian Steel Construction Institute in the Bulletin N° 26 where it is concluded that
“With addition of steel-fibre reinforcing, fire endurance ratings up to 3 h are possible.
Much like the bar-reinforcing, the steel fibres confine the concrete core by resisting the
splitting forces generated as the concrete degrades and approaches its ultimate load
carrying capacity”.
Bednar and Wald (2012) concluded that “The material test of fibre concrete
demonstrated that its behaviour of fibre concrete is ductile enough to create a mem-
brane action if the cracks are developed at yield lines. The ductility of fibre concrete at
600 °C was higher than at 20 °C. The shear resistance of the fibre reinforced is sig-
nificantly higher compared to concrete”. These effects give a great benefit, especially in
redundant structural members such as slabs and walls.
Nurchasanah and Massoud (2016) conclude it is expected that in future concrete
having steel fibre will act as fire protective considerably (Ref. 5).
Klingsch (2014) recommended that steel fibers should always be added to HPC and
UHPC mixes since they increase the ductility of the concrete.
2 Purpose of the Tests
A multi-storey typical residential building in Riga-Latvia, whose scaled model can be

seen in Fig. 1, was taken as first case to design a slab and wall structure of twelve levels
built in steel fibre reinforced concrete. The whole frame is built using tunnel forms
where the walls and the slab of one level are cast in one single operation as shown in
Fig. 2. After 18 h, the tunnel forms are removed and installed to form the next level up.
Each level of the building is designed of 180 mm thick walls to carry the cast in situ
slabs spanning 5,40 m, a span to depth ratio of 30 and where both slabs and walls are
only reinforced by 50 kg/m3 of ArcelorMittal HE+ 1/60 steel fibres of 1 mm diameter,
60 mm length, 1500 N/mm2 steel wire tensile strength, provided with hooked ends. In
the slab depth with 50 mm coverage from the bottom face, one rebar of 16 mm
diameter is included at every 1,5 m distance apart in order to meet the Anti-Progressive
Collapse condition following the ACI 544 6R15 provision J.1, p. 36. (ref 7) in case a
wall should disappear by accident or explosion (Fig. 3).
A full-scale test at cold temperature, reported by Kleinman and Destrée (2012)
demonstrated that the first negative moment crack live load of such a slab was
6,5 kN/m2 and the ultimate loading was of 11 kN/m2 to be compared to the most
onerous SLS service loading is of 3 kN/m2.
Fire Resistance of Steel Fibre Reinforced Concrete Elevated Suspended Slabs 843
Fig. 1. Overview of the project in Riga, Latvia
Fig. 2. Tunnel formed level installation

Fig. 3. Typical level cross sectional
The last question to resolve is about the fire resistance or rating of both slabs and
walls and therefore the testing program has been defined, budgeted and decided.
3 Design of the Slab Under the Ambient Temperature
The 180 mm slab of a 5,40 m multiple continuous span (center to center), under a
3 kN/m2 total variable loads, typical of residential or office applications, undergoes a
maximum envelope positive Mmax = M. (A0—A1) of the edge span where p are the
permanent loadings and s the variable loadings (Fig. 4).
Fig. 4. Maximum bending moment value of a multiple span system
The initial deflection is calculated to be of 2,4 mm as observed in reality under the

test load as shown in Fig. 5, with a modulus of elasticity E = 31.500 N/mm2.
Fig. 5. Calculated deflection in the current loading system

M max = 15,99 kN.m/m generates a flexion stress f SLS = 2.96 N/mm2 under
unfactored loadings. The slab is also provided with a set of Anti-Progressive Collapse
Rebars according to ACI544-6R15 and the Canadian CSA-A23,3-04, in order to
prevent the slab to fall down in case of a supporting wall collapse, by accident,
explosion or terrorist attack. The APC rebar, in the span-direction only, are 16 mm
diameter rebars in the bottom of 16 mm diameter every 1.50 m distance
apart. These APC rebars are bottom rebars with 50 mm coverage thus a very small of
0,07% of reinforcing or only of 3.6 kg per cubic metre of steel rebar, a very little
proportion indeed as shown at installation in Fig. 6.
Fig. 6. SFRC slab pouring condition with APC rebars.
The principal reinforcing is obtained by the addition in the ready mixed concrete of
50 kg/m3 ArcelorMittal steel fibres of 1 mm diameter by 60 mm length made out of a
1500 N/mm2 constituent steel wire. The steel fibre (HE+ 1/60 type) are provided with
hooked ends.
The concrete matrix, supplied as a C25-30 with 280 kg CEM I and W/C < 0,50
with 16 mm maximum aggregate size was of 23,83 kN/m3 density but indeed showed
a 52,6 N/mm2 average compressive strength at 28 days! As seen in Fig. 5, the
flowability was such that no poker vibrating was needed. Suspended elevated slabs in
steel-fibre reinforced concrete can be designed today following three available standard
documents:
• The ACI (American Concrete Institute) 544–6R15 “Report on Design and con-
struction of Steel Fiber-Reinforced Concrete Elevated Slabs”.
• The FIB Model Code 2012 (in its chapters 5 and 7)
• The SS-EN 812310 in Sweden, “Design of Fibre Concrete Structure”.
4 Wall Design
The walls were also of 180 mm thickness like the slabs, 3 m height and 4 m length,
built in the same concrete as the slabs including 50 kg/m3 HE+1/60 steel fibres but
without any APC rebars so that the walls were without any rebars, only steel fibres.
The walls are subjected under the test fire to a 330 kN per lineal meter of vertical
loading intensity, typical of a ten floors high building outer walls.
The fire test program includes three identical walls subjected to a 330 kN/m lineal
loading under fire.
5 Temperature Regime of the Fire Resistance Test
The test itself has been organized and carried-out by the F.I.R.E lab in Poprad (Slo-
vakia - SK) with a furnace set to follow the ISO 834, temperature vs. time diagram
according to the following formula:
T ¼ 345 log ð8t þ 1Þ þ 20 ð1Þ
As shown in Fig. 7, the ISO 834 and ASTM E119 standards are almost similar, the
ISO being still of a slightly higher temperature from 60 min to 180 min fire duration
and is thus slightly more stringent than the ASTM E119.
Fig. 7. ISO and ASTM fire curves
Such an ISO/ASTM fire is also called cellulosic fire and is typical of houses and
condos. The hydrocarbon fire is much more severe but only applicable where oil and
gas are involved like in petrochemical plants.
6 The Fire Resistance Test Set-up
The test set-up is according to EN 1365-2 and EN 13381.3.2015 standards, “Test

methods for determining the contribution of the fire resistance of structural members”
where a statically determinated 180 mm thick slab of 4 m net span, 3 m width is
subjected to two line loadings so that the flexural moment is constant and maximum
over 2,40 m length as shown in Fig. 8 (scheme) and in Fig. 9 (detail) including the
location of the thermocouples and deflectometers.
Fig. 8. Statical system for the slab
Fig. 9. Location of the thermocouples and the deflectometers in the salb

The fire test program includes four identical slabs. The first slab has been fire tested
under an initial bending stress of 2.5 N/mm2, indeed the same positive moment and
stress to occur in a residential use condominium slab under its most onerous loading
condition. The three next slabs are loaded even more up to 3.5 N/mm2 bending stress,
thus 30% more than the most onerous loading in service condition in reality.
7 Fire Tests Observations and Results
The first slab under 2.5 N/mm2 flexion strength could sustain the loading during more
than four hours so that the test had to be stopped. The slab at 240 min fire duration was
still stable and the deflection increase was still smaller than the 11 mm per minute as
per the standards. Thus, no collapse was observed, even after 4 h! The slab was still
flame tight at 240 min and its upper surface temperature below the 140 °C.
The observed fire resistance has been exceeding far more the maximum standard
requirement of 2 h for a slab! The three next slabs were fire tested under a flexion stress
of 3,5 N/mm2 and met the fire resistance standard requirements up to respectively
192 min, 149 min and 204 min when the slabs collapsed, thus more than the 2 h limit
although the flexion stress rate was 30% higher than any possible 3 kN/m2 variable
loading in a real structure!
It is important here to repeat that the tested slabs were tested as statically deter-
minated as specified in the EN 1365-2 fire rating standards and this unlike in reality,
where the slabs are statically indeterminated and can benefit of the plastic moment
redistribution.
Quite above any expectations for statically determinated slabs, a large number of
bottom cracks (ca. 16 cracks) are observed and have grown up from the bottom as it is
visible in the Fig. 10. Up to 192 min, 149 min and 204 min, the deflection was smaller
than the 250 mm standard limit and its increase has been under 11 mm per minute limit
as also required by the standard.
The steel fibre concrete has been quite able to prevent the explosive spalling from
the bottom so that a layer of burnt but insulating concrete remained attached around
and by the fibres, indeed a very effective protection of the concrete core. The beneficial
effect is shown in Fig. 11 and Fig. 12 where the temperatures at various depths from
the top surface (30 mm depth) to the bottom (150 mm depth) so that it can be observed
that at 150 mm depth from the top surface, at 4 h fire, the temperature remains as low
500 °C when the concrete offers still 35% of its initial strength at cold.
In the top surface, the temperature did not exceed 70 °C, a lot colder than the
maximum 140 °C according to the standard limit.
Fig. 10. View of the slab over the heating room under at 192 min
Fig. 11. Recorded Temperatures at various depths during the fire test.
One wall only shown on Fig. 13. of the three walls, at the time of this report, could
have been tested under fire but it nevertheless showed full compliance after 4 h fire to
all performance criterions.
Fig. 12. Temperature vs. time from bottom in red (1150 °C at 4 h) to top (100 °C at 4 h) in blue
and at 30 mm-orange); 60 mm-grey; 90 mm-yellow; 120 mm-blue; and 150 mm-green.
Fig. 13. The tested wall at 241 min fire under 330 kN/m vertical loading
The wall has also been stable and flame tight, with a cold surface temperature at
70 °C under the 140 °C limit. The wall met and exceeded as well all standard
requirements of resistance to fire. Indeed, the fire test has been stopped at 242 min
without any observed failure to meet the EN1365-3 requirement. More details are given
in the Fig. 14 here below:
Fig. 14. The wall official report table to report the 4 h resistance to fire.
8 Conclusions
Three identical statically determinated steel fibre reinforced concrete slabs and one wall
have been tested to confirm the full compliance under fire to EN 1365-2 and EN 1365-3
thus well over the 2 h regarding slabs and over 4 h regarding the only wall tested. Two
more walls shall be tested in the same experimental programme.
Steel Fibre Reinforced Concrete used as the main reinforcing of elevated suspended
slabs and designed according to the provisions of the ACI 544 6R15, has fulfilled and
exceeded the standard safety requirements of EN 1365-2.
The thermal spalling has been prevented by the 50 kg/m3 steel fibres reinforcing, so
that the concrete core remained insulated from the very high temperatures in the
bottom. The prevention of spalling under fire is clearly one of the reasons why the steel
fibre reinforcing is a viable and safe method for free suspended elevated slabs as
designed following the ACI 544-6R15 report provisions.
References
Lie, T.T., Kodur, V.K.R.: Mechanical properties of fibre-reinforced concrete at elevated
temperatures. Internal Report N° 687, National Research Council Canada (1995)
Kleinman, C., Destrée, X.: Steel fibre as only reinforcing in free suspended one-way elevated
slab: design conclusions of a tunnel formed slab and walls based upon full scale testing
results. In: BEFIB 2012, 8th RILEM International Symposium on fibre Reinforced Concrete:
Challenges and opportunities (2012)
Klingsch, E.W.: Explosive spalling of concrete in fire. Doctoral thesis, ETH Zürich (2014)
Widhianto, A., Darmavadi, D.: Fire resistance of normal and high-strength concrete with contains
of steel fibre. Asian J. Civ. Eng. 15(5), 655–669 (2014)
Nurchasanah, Y., Massoud, M.A.: Steel fiber reinforced concrete to improve the characteristics of
fire-resistant concrete. In: Applied Mechanics and Materials, vol. 845, pp. 220–225 (2016)
Bednar, J., Wald, F.: Membrane action of composite fibre concrete slab in fire. Proc. Eng. 40,
498–503 (2012). Steel Structures and Bridges
ACI 544-6R15: Report on design and construction of steel fiber-reinforced concrete elevated
slabs (2015)
Structural Behavior of a Traditional Concrete
and Hollow Tiles Mixed Floor Reinforced
with HPFRC
D. Sirtoli1(&), P. Riva1, and P. Girardello2

1
Department of Civil Engineering, University of Bergamo, Bergamo, Italy
davide.sirtoli@unibg.it
2
Kerakoll S.p.A., Sassuolo, Italy
Abstract. The use of high performance fiber reinforced concrete (HPFRC) in

structural strengthening is widely accepted. The presence of steel fibers inside the
cementitious matrix provides the hardened matrix enhanced resources in terms of
strength and toughness, avoiding brittle crisis in favor of hardening/softening post-
peak behavior. These performances open different scenarios in structural reha-
bilitation, e.g. RC column jacketing, shear strengthening of RC beam, slab rein-
forcement. However, the quality of the intervention is highly dependent on the
technique considered for the element strengthening and on the HPFRC perfor-
mance, which is strongly related to the mix design. This paper presents the results
of an experimental campaign carried out on the flexural response of 4.5 1 m2
one-way floor elements reinforced with a commercial HPFRC material. The
mechanical performances of the HPFRC adopted have been qualified following
the Italian guidelines published in January 2019 for the identification and quali-
fication of FRC products. The effect of the FRC strengthening was evaluated by
comparing the performance of unreinforced and retrofitted floor elements, both
using traditional r.c. and HPFRC topping. The effectiveness of the intervention is
evaluated in terms of flexural strength, slab deflection and visible damage.
keywords: Steel fiber-reinforced concrete Slab strengthening Retrofitting
1 Introduction
High performance fiber reinforced concrete (HPFRC) is a well-known composite

material with good performance in compression thanks to the concrete matrix com-
bined with the improved post-cracking behavior provided by the fibers crack opening
control. Thanks to its performance, FRC has been used in several structural inter-
ventions for a multitude of purposes [1]. As a dispersed reinforcement it found
application in industrial floors and pavements [2] or building slabs [3]. Its combination
with steel bars grants resources against impact and fatigue, useful for building struc-
tural elements in seismic zones [4]. However, the effectiveness of the intervention is
strongly dependent to the correlation between its design and the FRC mechanical
properties. For both these aspects, codes of practice were produced, with specific
indications on the definition, realization and application of FRC. In terms of design
details and how to treat them, one of the reference codes is represented by the Model
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_75
Structural Behavior of a Traditional Concrete and Hollow Tiles 853
Code 2010 [5], based on researches of Di Prisco [6] and others. On the other hand, in
order to guarantee the quality of the material introduced into the market, each nation
issues guidelines, as a grant for the final user and protection for the producer. Italy, in
2019, published its own reference document for the identification, qualification, cer-
tification of technical evaluation and acceptance criteria of fiber reinforced concrete [7].
This paper has the aim to investigate the mechanical behavior of a traditional one-
way slab, made of a mixed reinforced concrete and clay brick void formers, reinforced
with FRC which properties were determined following the recently published Italian
guidelines. The performance of an un-reinforced sample is compared to four samples
strengthened with different solutions. The obtained data are used to evaluate the
effectiveness of the intervention, giving indications on the considered technique.
2.1 Samples Details

The structural elements subjected to the flexural strength test are traditional reinforced
concrete and hollow-core bricks mixed floors of 4.5 1 m dimension in length and
width respectively, reinforced with a topping layer of fiber reinforced concrete
(FRC) of different thicknesses. The details of the traditional floor are reported in Fig. 1.
A summary of the FRC mechanical performances is given in Table 1 with a typical
graph for the flexural test performed in accordance with EN 14651 proposed in Fig. 2.
These data are taken from an experimental campaign organized following the indica-
tions of the Italian guidelines for the identification, qualification, technical evaluation
certification and acceptance control for FRC [7]. Five different samples were delivered
to the laboratory, four strengthened and one plain. The basic sample is composed of
hollow-core bricks of 38 16 25 cm dimensions, concrete with strength class C12-
15 and steel bars of B450C class and 12 mm diameter. The floor was strengthened
following two techniques: the first is an FRC topping with three different thicknesses of
2, 3 and 4 cm while the second is composed of a welded mesh incorporated in a 4 cm
topping layer made of HPC. Before casting the strengthening material, the upper
surface of the basic sample was prepared by increasing its roughness in order to
improve the adhesion between the old and new materials. No primer of shear stud
connectors was used in any of the specimens to ensure shear stress transmission
between the original topping and the retrofitting topping. The obtained results on the
strengthened samples are compared with the one gathered from the un-strengthened
floor sample in order to evaluate the efficacy of the strengthening technique by ana-
lysing the improvements on the structural element performances. The sample weight
was considered in the calculation, adding 9.26, 11.42, 12.50, 13.58 and 13.58 kN to the
total load for NR, R2, R3, R4 and RT respectively.
The investigated samples of this experimental campaign are identified as follow:
1_NR: un-strengthened sample;
2_R2: strengthened sample with FRC topping of 2 cm;
5_RT: strengthened sample with HPC topping (Magma) of 4 cm combined with a
welded mesh.
854 D. Sirtoli et al.
Table 1. Mechanical performances of the FRC used in the experimental campaign.
Compressive strength [MPa]

Rcm Rck fck Strength class
109.3 ± 1.71 106.50 88.39 C80/95
Flexural strength [MPa]

ffct,L fR,1 fR,3
fR,3k/fR,1k HC
m c m c m c
6.95±0.63 5.91 12.21±1.74 9.54 9.88±1.62 7.33 0.768 8b
Tensile strength
fFts [MPa] ε [%]
m c m
7.40±1.20 5.32 0.0412
Fig. 1. Geometrical details of the floor section.
Fig. 2. Typical CMOD curve reported in the Italian guidelines [7].

2.2 Samples Preparation and Storage

The floor samples were realized by an external company under the indication of the
University of Bergamo. Once the floor samples were realized, the roughness of the top
surface of the samples to be strengthened was improved by hydro-sand blasting in
order to have better bond between old and new concrete topping material. Eventually,
the topping FRC was casted and left as is until the day of the test, without particular
curing attention, other than a sheet covering to avoid excessive drying for the first 24 h.
The samples were sent to the laboratory before the 28 days of maturation, where they
remained in laboratory environment conditions for, at least, 24 h before performing the
test.
2.3 Set-up
The test done on the floor samples is a four-point bending test in deflection control.
A general view of the used set-up is shown in Fig. 3. The two supports are realized
using metallic cylinders of 7 cm diameter and length sufficient to cover the whole
transverse section in order to sustain all the three joists. These cylinders are lodged in
special metallic plates designed to block longitudinal movement while the sample is
still free to rotate during the test, creating a hinge restraint. Between cylinders and joists
a neoprene layer is placed to uniformly distribute the load on the sample transverse
section. For the load application, a 500 kN electromechanical jack to apply the force
vertically to the sample was used. A system composed of three metallic beams is used
to distribute the single force of the jack on the two load lines defined for the test. Two
beams lay transversally on the slab element, representing the load points and the
supports for the third beam, which is placed longitudinally to the sample, distributing
half the jack load to each single supporting beam. A ball bearing was placed between
jack and longitudinal beam, in order to grant an equal load distribution. Between the
two supporting beams and the upper floor surface a neoprene layer is placed in order to
avoid any load concentration. The data acquisition has started in a configuration where
the sample is completely unloaded. Afterwards, the two metallic supporting beams are
positioned on the slab sample and loaded by the third. This own weight of the loading
system is not recorded in the test, but considered on the final data elaboration.
Considering the test protocol, just in the case of the un-retrofitted sample three
cycles at incremental deflection level were considered in order to stabilize the floor
response. For the other samples, a monotonic test was performed at a constant
deflection rate of 0.05 mm/sec until collapse or complete plasticization.
Fig. 3. Set-up for the flexural test performed on the floor samples
2.4 Instrumentation
During the test, the displacements of the floor in seven different positions were
recorded together with the applied force. In Fig. 4 the points where the instruments are
applied with their identification name are shown.
Fig. 4. Instruments used during the test and their identification (De: Deflection; Dr: Drop; l: left;
r: right)
Four potentiometric wire instruments with maximum stroke of one meter were used
to register the deflections (De) in the mid floor section and under the two loading
points. Considering “De mid” and “De r” the wire was placed at the mid-joist intrados.
Instead, in Sect. 2, two points were used to record the deflection, one at each slab side,
in order to analyse the likely rotation occurring on the transverse plane due to local
failure or bad load redistribution.
As under the two supports a series of neoprene layer were placed in order to
balance the slab and equally redistribute the load to each of the three joists, three
displacement transducers LVDT were placed at each floor end (Dr). Even in this case,
on one side of the floor a single LVDT “Dr l” was used, placing it at the mid-section
extrados over the support. On the other side of the floor, instead, two LVDT instru-
ments were placed close to the lateral joists, symmetrically respect the mid-point, over
the support on the slab extrados. As in the case of the deflection investigation, even
here the two instruments are aimed to record bad redistribution or local failure of the
lateral joists, besides the vertical settlement due to the compression of the neoprene
layer.
In both cases of deflection and support settlement investigation, the behavior of the
sections in which two instruments were considered are described by their mean value,
though unexpected behavior is identified.
2.5 Considered Formulations

2.5.1 Results Interpretation
The set-up defined for the flexural tests on the floor elements yields to the static scheme
reported in Fig. 5. The two extradoxal loads are redistributed on the whole width of the
specimen, equal to one meter, while their distance from the supports is equal to one
meter. Considering the maximum bending moment of the scheme to the left, setting it
equal to the maximum moment of the right one, the equivalent distributed load is
obtained. It represents that particular load which results in the same maximum bending
moment and shear obtained with the test configuration. This operation is useful in order
to compare the results obtained in the laboratory tests with the more realistic config-
uration of the buildings, where uniformly distributed loads are used for the structural
design. Particular attention must be paid when comparing the results obtained in both
configurations as the redistribution of the bending moment and shear is different. The
reinterpretation of the acquired results must be done taking into account this aspect.
Fig. 5. Considered static schemes.

2.5.2 Theoretical Strength

Theoretical flexural and shear strength of the analysed samples were obtained from the
formulations reported in the main design codes and guidelines available. Being an
experimental test on a real scale element, reduction coefficients in materials perfor-
mance “c” are set equal to 1. Due to the slab transverse section dimensions, theoretical
results for the single joist were doubled in order to take into account the presence of
other two half joists at the floor sides. Considering the theoretical bending moment
calculation for both strengthened and un-strengthened slab samples, the HPFRC Reluis
guidelines were considered, though not published yet. Differently, for the shear
strength, the indication reported in the Italian standard NTC 18 were considered. In the
following extracts from the used codes are illustrated.
Bending Moment
The bending moment evaluation is done referring to the scheme of Fig. 6, where two
different configurations are available, depending on whether the position of the neutral
axis lays in the FRC topping (Fig. 6a) or in the old concrete layer (Fig. 6b). For both
normal and fiber reinforced concrete, a stress-block distribution is considered in the
compressed section. As the FRC topping layer acts mainly in the compressed zone and
its thickness is limited, the possible tensile contribution is neglected. Firstly, the neutral
axis position is evaluated, assuming its starting position inside the FRC topping layer,
verifying thereafter its correctness. The neutral axis position is calculated through
normal forces equilibrium:
AS fyd
AS fyd ¼ 0:8 x fFcd B ) x ¼ ð1Þ
0:8 fFcd B
If x < 1.25 hreinf the resistant bending moment is calculated with the next
formulation:

MRd ¼ 0:8 x fFcd B d þ hreinf 0:4 x ð2Þ
If x > 1.25 hreinf the neutral axis position must be calculated again with a new
equilibrium:

AS fyd ¼ hreinf fFcd þ 0:8 x hreinf fcd B ð3Þ
Then, the reinforced floor resistant bending moment is calculated:

hreinf 0:8 x hreinf
MRd ¼ hreinf fFcd B dþ þ 0:8 x hreinf fcd B d
2 2
ð4Þ
Shear
All the calculation for the resistant shear are reported in chapter 4.1.2.3.5.1 of the
NTC18 standard.
Fig. 6. Stress and strain diagrams for the FRC reinforced section: (a) neutral axis inside the FRC
thickness; (b) neutral axis inside the old concrete section.
2.6 Samples Real Conditions

For a better interpretation of the results reported hereafter and obtained from the test
described above, the actual characteristics of the analyzed samples are provided.
Considering floor R3, its collapse mechanism highlights the presence of a diagonal hole
in longitudinal direction of diameter Ø5 cm and 50 cm length on one of the side joists,
exactly under the point where the load is applied. Moreover, demolishing the mid-
section for a final inspection, just three of the four steel reinforcement bars were found,
missing one of the two in the mid joist. Another aspect which was underlined by the
collapse mechanism, even in sample R4, was the thin concrete cover to the longitudinal
reinforcement, which turned out to be less than 1 cm in both extradoxal and lateral
directions. Eventually, sample RT reported a mistake in the reinforcement overlap in
one of the side joists where the two steel bars were placed head to head and not
overlapped, creating a section with no flexural resources.
3 Results
Results are reported in Fig. 7 in terms of shear/bending moment – mid deflection for
some key-points describing the general behavior of the samples. The same results are
summarized in Table 2, showing the deflections as a function of the sample length,
while in Table 3 are reported data in terms of strength, derived from codes formulations
(rd) and laboratory test (ex).
Figure 7 shows clearly the improvements due to the slab strengthening. Compared
to the black line, which represents the un-reinforced sample, the other curves showed
higher stiffness, maintained for higher load levels. Sample NR show a stiffness decrease
at around 40–50 kN, which is the moment in which a longitudinal crack appears on the
top surface of the sample from one of the two supports. The test was stopped before
collapse, when the curve started to flatten, reaching 61.3 kN load which corresponds to
15.33 kN/m equivalent distributed load.
Considering the curves representing samples strengthened with FRC topping layer,
two of three samples evidenced a brittle collapse at a load of 68.56 and 68.26 kN. The
peculiarity in this situation is that the two samples with thicker FRC layer evidenced
this collapse, while sample R2 reached a maximum load of 72.64 kN deforming
thereafter at a constant stress level. It must be underlined that sample R2 did not show
any problem during the sample check-up; on the contrary, samples R3 and R4 were
both characterized by thin concrete covers, with R3 characterized by many other
defects. Considering the strength resources of the slab samples, they are composed by
the sum of the basic resistance mechanisms, which remain constant between samples,
and the additional resources provided by the strengthening technique, which improves
mainly the flexural resistance. As can be seen by the failure mode shown by R3 and R4,
it is characterized by the typical damage and crack pattern of a shear mechanism, thus,
not much affected by the strengthening technique. With this assumption, it is not
rational to have a sample R2 which do not show any failure during the whole test;
however, taking into account all the defects recorded in samples R3 and R4 it is
plausible that specimen R2 had enough resources to resist the shear collapse until it
reaches its maximum flexural resistance, which proved to be close to the shear at which
the other two samples collapsed (36 kN vs 34 kN for R2 and R3-R4 respectively).
Considering that, it can be remarked that the shear resources of the basic floor on which
a strengthening intervention will be applied are strongly connected with the entity of
this intervention. Improving flexural strength of a floor element by increasing the
thickness of the FRC topping layer must be considered together with the study of the
shear resources of the un-strengthened system, requiring to improve the latter if it is
necessary to increase further the resources in bending.
The traditional strengthening technique used for sample RT showed a good per-
formance, with stiffness lower than R4 but higher compared to the other samples. This
can be related to the overlap defect evidenced in the sample, which can affect the floor
performance in terms of strength and stiffness. In any case, the sample reached its
maximum flexural capacity without showing any failure, performing the test until a
mid-span deflection of 60 mm.
Focusing the attention on the first 10 mm of deflection (Fig. 7), it is possible to
relate the slope of each single curve to the sample tested. As expected, the lowest slope
is found in sample NR (un-strengthened), which is the less stiff of the five samples, in
agreement with the design data. It is followed by R2, which did not show any defect in
the sample and during the test. The first consequence of the floor defects is visible in
sample R3, where the curve slope is closer, nearly overlapping that of sample R2,
while, from design data, we expected it a slightly stiffer response. The last two samples
are characterized by the same thickness of the topping layer, thus, supposed to have
similar stiffness. However, sample R4 showed a higher stiffness than RT, in which the
reinforcement overlap mistake was found in one of the joists composing the floor.
40
NR
Shear [kN]; Flexural moment [kNm] 36 R2
32 R3
R4
28 RT
24
20
16
12
8
4
0
0 5 10 15 20 25 30 35 40 45 50 55 60
Mid deflection [mm]
Fig. 7. Shear/bending moment-mid deflection graph for the investigated floors
Table 2. Load and bending moment results corresponding to different mid-deflection levels.
Mid deflection NR R2 R3 R4 RT
[-] [mm]
L/1000 4 [kN] 7.80 9.51 10.39 12.48
11.45
[kN/m] 3.90 4.75 5.20 6.245.72
L/500 8 [kN] 11.00 14.17 14.29 18.09
16.17
[kN/m] 5.50 7.09 7.15 9.048.08
L/400 10 [kN] 12.53 16.51 16.16 20.50
18.30
[kN/m] 6.27 8.25 8.08 10.259.15
L/300 13.33 [kN] 15.04 20.39 19.38 24.72
21.70
[kN/m] 7.52 10.20 9.69 12.36
10.85
L/200 20 [kN] 19.98 23.23 25.87 32.62
28.09
[kN/m] 9.99 11.61 12.94 16.31
14.05
L/150 26.67 [kN] 23.71 30.08 32.05 33.82
[kN/m] 11.86 15.04 16.02 16.91
L/100 40 [kN] 29.53 36.00 38.95
[kN/m] 14.76 18.00 19.47
Maximum force – [kN] 30.65 36.32 34.28 34.13 40.54
[kN/m] 15.33 18.16 17.14 17.07 20.27
Table 3. Theoretical and experimental shear and flexural strength results for the analyzed
samples (c = 1).
NR R2 R3 R4 RT
As mm2 Design 452.4 452.4 452.4 452.4 452.4
Effective 452.4 452.4 339.3 452.4 339.3
Mrd kNm Design 28.81 34.38 36.41 38.45 38.45
Effective 28.81 34.38 27.35 38.45 28.88
Mex kNm Test 31.00 35.95 34.42 34.13 40.79
Vrd kN Design 32.02 34.80 36.16 37.48 37.48
Effective 32.02 34.80 32.85 37.48 34.06
Vex kN Test 31.00 35.95 34.42 34.13 40.79
4 Final Remarks
In the present paper the flexural behavior of strengthened floor elements was investi-
gated by means of a four-point bending test. Five different samples were analysed, four
of which strengthened with different toppings and one left un-strengthened as refer-
ence. The technique used to reinforce the floors is represented by FRC topping with
incremental thicknesses for three samples, while, for the fourth, a welded mash was
drowned in a HPC topping layer.
• The collected results report an increment of the stiffness with increasing topping
thickness, with some exceptions due to samples construction defects. For instance,
samples R3 and RT showed lower stiffness compared to what expected from design
values; indeed, these samples were found with construction errors like hole pres-
ence in the joists, lack of steel reinforcement, wrong longitudinal bar overlap and
thin concrete cover.
• Considering the general behavior, it turns out that the maximum effective topping
thickness is strongly dependent to the shear strength of the basic floor. As the used
strengthening technique is not effective in shear but in bending, once the shear
capacity of the basic floor is reached a further thickness increment of the FRC
topping layer becomes useless. In this situation designers must consider shear
strengthening intervention in combination to the FRC topping in case they want to
increase the flexural capacity of the structural element.
Acknowledgements. We acknowledge Kerakoll S.p.A. company for its financial support to the
experimental campaign.
References
1. Brandt, A.M.: Fibre reinforced cement-based (FRC) composites after over 40 years of
development in building and civil engineering. Compos. Struct. 86(1–3), 3–9 (2008)
2. Sorelli, L.G., Meda, A., Plizzari, G.A.: Steel fiber concrete slabs on ground: a structural
matter. ACI Struct. J. 103(4), 551 (2006)
3. Hedebratt, J., Silfwerbrand, J.: Full-scale test of a pile supported steel fibre concrete slab.
4. Beschi, C., Meda, A., Riva, P.: Column and joint retrofitting with high performance fiber
reinforced concrete jacketing. J. Earthq. Eng. 15(7), 989–1014 (2011)
5. Code, C.F.M.: CEB-FIB model code 2010–Final draft. Thomas Thelford, Lausanne,
Switzerland (2010)
6. Di Prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in fib model code 2010:
7. dei Lavori Pubblici, C.S.: Linea Guida per la identificazione, la qualificazione, la
certificazione di valutazione tecnica ed il controllo di accettazione dei calcestruzzi
fibrorinforzati FRC (Fiber Reinforced Concrete), Italy (2019)
Pedestrian Bridge over Las Vegas Avenue
in Medellín. First Latin American
Infrastructure in UHPFRC
Joaquín Abellán-García1,2(&), Andrés M. Núñez-López3,

and Samuel E. Arango-Campo3
1
Department of Civil Engineering, Polytechnic University of Madrid (UPM),
Madrid, Spain
j.abellang@alumnos.upm.es
2
Escuela Colombiana de Ingeniería Julio Garavito, Bogotá, Colombia
3
I+D+i Cementos Argos, Medellin, Colombia
Abstract. To introduce a new product into a new consolidated market is

always a challenge, even for those products like ultra-high-performance fiber
reinforced concrete (UHPFRC) which have proven their value, and it is even
harder if this product does not count with local regulation that supports it, like
the case of UHPFRC in Colombia. This is the reason that made local devel-
opments, both dosages and applications, of vital importance for UHPFRC wider
use. That is why is important to emphasize the construction of the pedestrian
bridge over Las Vegas Avenue in Medellin, the first infrastructure built in
UHPC in Latin America, using a dosage with local materials developed by the
company Argos SA. The new pedestrian bridge that connects the campus of the
EAFIT University in Medellin with its language building is the first infras-
tructure work in Latin America in which UHPFRC was used. Initially the design
included a metallic structure, however, Argos proposed to the University to
change the steel for an UHPFRC dosage developed by Argos SA. The decision
to opt for this material represented a saving of 34% in the total cost of the
pedestrian bridge construction.
Keywords: UHPFRC Precast keystone Pedestrian bridge Fiber reinforced

concrete
1 Introduction
Over the past few decades, the advancements of the high-performance cementitious
composites have made colossal progress in building sector worldwide, pointing to a
new and high-tech material with enhanced material properties such as compressive
strength, durability, toughness and ductility. According to that, a new type of ultra-
high-performance cement based composite material, known as ultra-high performance
fiber reinforced concrete (UHPFRC), was defined by ACI committee 239R-18 as a
“concrete that has a minimum specified compressive strength of 150 MPa with spec-
ified durability, tensile ductility and toughness requirements; fibers are generally
included to achieve specified requirements” [1]. Such outstanding mechanical
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_76
Pedestrian Bridge over Las Vegas Avenue in Medellín 865
properties can be ascribed to its low porosity and high packing density due to its low
water to binder ratio, special mixture design and mixing procedure, leading to an
extremely optimized grain-binder matrix [2–11]. Long-term durability is another out-
come of the low porosity and high packing density [2, 9]. The applications of UHPFRC
are increasing in the last years in Europe, North America, Japan, Korea, China and
Australia. Among others, uses of UHPFRC include the construction of pedestrian
bridges, liner tunnel segments, special pre-stressed and pre-cast concrete elements,
concrete structure rehabilitation, precast deck panel bridge joints, urban furniture,
overlay on damaged pavements and industrial floors, and architectural applications [2,
7, 9, 12–14]. Therefore, notwithstanding the superb performance in strength and
durability, UHPFRC has not been widely applied in emerging construction markets
mainly due to the lack of local codes for its design and construction, as well as to lack
of knowledge and confidence local agents, builders and developers. The development
of UHPFRC projects in those markets will demonstrate economic for departing from
traditional practice under standard concrete.
During the last years, the city of Medellin has stood out for being a leader in
innovation in both Colombia and South America, both in the engineering and con-
struction sector. In particular, significant efforts were made in the development of
special concrete and construction processes that contribute significantly to the progress
of the country. With this intention, the company Argos S.A, with its research group,
has been a pioneer in the production of UHPFRC in Colombia; have already devoted
some years to research and have at present a UHPFRC dosage patent with the com-
mercial name of Advanced Concrete ®, developed from local raw materials available
in Colombia.
This paper presents the construction of the pedestrian bridge over Las Vegas
avenue in Medellin (Colombia). This building was not only the first infrastructure
project developed with UHPFRC in Colombia, but the first in all Latin America [2].
This pedestrian bridge, which connects the EAFIT University campus with its language
building, was constructed by using Advanced Concrete ®.
2 General Characteristics of the Pedestrian Bridge
The pedestrian bridge of EAFIT University, has double curvature - a horizontal and a
vertical one - 110 linear meters, a main span of 43 m over Las Vegas avenue and is
supported by five points: four columns and a head beam constructed in self-compacting
concrete with compressive strength of 42 MPa.
The section of the bridge is made up of a main drawer beam from which a can-
tilever emerges that together with the pre-slabs forms the support through which people
walk over the avenue. It is important to denote that the asymmetric geometry of these
sections implies efforts that a conventional concrete could not support.
This project was carried out using the premanufactured keystone bridge con-
struction process with non-adhered post-tensioning. This involved the construction of
29 prefabricated keystones, which were rise to their position with cranes, there spun
with steel cables and tensioned to form the structural system. The structural system of
the footbridge is composed of the main girder with a plug-socket joint system.
Each keystone weighs about ten tons and was built using 3.7 m3 of Advanced
Concrete ®. In addition, in this work 120 bags were used for the construction of the
joints, totalizing a consumption of 110 m3 of UHPFRC.
The AFGC recommendations [15] were followed for the structural analysis and
design of the UHPFRC elements.
Table 1 summarizes the characteristics of the pedestrian bridge (Fig. 1).
Table 1. Characteristics of the EAFIT pedestrian bridge.

Total length 110 m
Total weight 300 ton
Main span 43 m
Number of structural supports 5
Number of premanufactured keystones 29
Fig. 1. Overall view of the pedestrian bridge.
Figure 2 depicts a cross section of the pedestrian bridge over the pile cap. The
drawing includes dimensions and some reinforcement details for the main elements of
the structure.
Fig. 2. Cross section of the pedestrian bridge over the pile cap.
3 Material
Advanced Concrete ®, an UHPFRC developed from local raw materials available in

Colombia, was used. Table 2 summarizes the characteristics of the Advanced Concrete
® used in the construction of the pedestrian bridge.
Table 2. Characteristics of UHPFRC at 28 days.

Compressive strength 150 MPa
Direct tensile strength 10 MPa
Flexural strength 36 MPa
Modulus of elasticity 40 GPa
4 Pedestrian Bridge Construction Using Precast UHPFRC

Keystones
The construction procedure consisted in several phases which are described below:
4.1 Foundation and Pillars

The foundation was executed by reinforced concrete piles excavated in situ.
4.2 Construction of the Keystones

Starting in December 2016, the construction of 29 keystones using Advanced Concrete
® was carried out at TITAN SA field office in Medellín, Colombia. Pre-mixed concrete
was delivered through the concrete batch plant. To enhance the packing density of
UHPC, a slight vibration was applied to the mold during the pouring, prolonging it
until one minute after completed, although the concrete was visibly self-compacting
and flowing easily as it seemed in Fig. 3. Besides, a to improve the aesthetical aspect of
concrete's surface, a vibrating ruler was used as depicted in Fig. 4.
Fig. 3. UHPFRC flowing into the mould during construction of keystones.
Fig. 4. Pouring process and use of vibrating ruler on free surface.
After 24 h from pouring the keystones were demolding and stocked by using an
overhead crane (see Fig. 5). As depicted in Fig. 5, an auxiliary steel structure was used
to avoid damages during the stockage process.
Fig. 5. Keystone manipulation and displacement by overhead crane (left). Auxiliary structure to
avoid damages in keystone during the stockage (right).
4.3 Construction of Piles and Bridge Stirrups

Taking advantage of the prefabrication, at the time that the keystones were built, the
works of stacks and stirrups of the bridge were executed. Piles and bridge stirrups were
constructed with poured-in-place reinforced concrete with compressive strength of
50 MPa (Fig. 6).
Fig. 6. View of the construction of piles and stirrups.
4.4 Centring Placement

Until the keystones were inserted and post-tensioned the pedestrian bridge has no
strength and needs the centring to keep the keystones themselves in their correct
relative positions. Figure 7 depicts the porticoed formwork and steel beams used for
this purpose.
Fig. 7. View of the placement of porticoed centring.
4.5 Deliver and Movement of Keystones

At the TITAN SA field shop in Medellin, the 29 keystones were delivered using 29
semi-truck (with a load of one keystone each truck as depicted in Fig. 8 left). At the
pedestrian bridge site, the keystones were moved by crane as showed in Fig. 8 right.
Fig. 8. Delivering of keystones (left). Crane used to move the keystone (right).
4.6 Alignment of Keystones

After the keystones were placed in position, joints were executed with field-cast UHPC.
After that, the post-tensioning process was carried out, creating a continuous girder
(Figs. 9, 10 and 11).
Fig. 9. Alignment of the keystones
Fig. 10. Construction of the joints
Fig. 11. Post-tensioning process
4.7 Bridge Load Test

Finally bridge load test was performed using barrels filled with water as depicted in
Fig. 12.
Fig. 12. Bridge load test
Results of the proof load test performed compiled with the requirements of the
Colombian regulation.
5 Conclusions
From the construction of the first infrastructure in Latin America using UHPFRC, the
following conclusions can be drawn:
• The use of UHPFC instead the previously planned solution in steel meant a saving
of 33% in the material cost of the project.
• The experience gained during the construction of the pedestrian bridge encourages
to continue with the development of infrastructure using UHPFRC.
• Furthermore, this paper shows good practices, large applications of UHPC and
attractive potential market in Colombia.
Acknowledgements. Authors want to acknowledge R&D division of Cementos ARGOS SA,

Escuela Colombiana de Ingeniería Julio Garavito, and EAFIT University for bringing us the
confidence, support and opportunity to launch special concretes like UHPC in Colombia.
References
1. ACI Committe 239: ACI – 239 Committee in Ultra-High Performance Concrete (2018)
2. Abellan, J., Torres, N., Núñez, A., Fernández, J.: Ultra high preformance fiber reinforced
concrete: state of the art, applications and possibilities into the latin american market. In:
XXXVIII Jornadas Sudam. Ingeniería Estructural, Lima, Peru (2018)
3. Soliman, N.A., Tagnit-Hamou, A.: Using particle packing and statistical approach to
optimize eco-efficient ultra-high-performance concrete. ACI Mater. J. 114, 847–858 (2017).
https://doi.org/10.14359/51701001
4. Ghafari, E., Costa, H., Nuno, E., Santos, B.: RSM-based model to predict the performance of
self-compacting UHPC reinforced with hybrid steel micro-fibers. Constr. Build. Mater. 66,
375–383 (2014). https://doi.org/10.1016/j.conbuildmat.2014.05.064
5. Schmidt, C., Schmidt, M.: ‘Whitetopping’ of asphalt and concrete pavements with thin
layers of ultra-high-performance concrete - construction and economic efficiency. In:
Fröhlich, M.S.E.F.C.G.S., Piotrowski, S. (eds.) 3rd International Symposium UHPC
Nanotechnology High Performance Construction Materials, Kassel University, Kassel,
Germany, pp. 921–927 (2012). ISBN online: 978-3-86219-264-9
6. Abbas, S., Nehdi, M.L., Saleem, M.A.: Ultra-High performance concrete: mechanical
performance, durability, sustainability and implementation challenges. Int. J. Conc. Struct.
Mater. 10, 271–295 (2016). https://doi.org/10.1007/s40069-016-0157-4
7. Nehdi, M., Abbas, S., Soliman, A.: Exploratory study of ultra-high performance fiber
reinforced concrete tunnel lining segments with varying steel fiber lengths and dosages. Eng.
Struct. 101, 733–742 (2015). https://doi.org/10.1016/j.engstruct.2015.07.012
8. Filho, R.D.T., Koenders, E.A.B., Formagini, S., Fairbairn, E.M.R.: Performance assessment
of ultra high performance fiber reinforced cementitious composites in view of sustainability.
Mater. Des. 36, 880–888 (2012). https://doi.org/10.1016/j.matdes.2011.09.022
9. Abellan, J., Torres, N., Núñez, A., Fernández, J.: Influencia del exponente de Fuller, la
relación agua conglomerante y el contenido en policarboxilato en concretos de muy altas
prestaciones. In: IV Congress International de Ingenieria Civil, Havana, Cuba (2018)
10. Abellán, J., Fernández, J., Torres, N., Núñez, A.: Statistical optimization of ultra-high-
performance glass concrete. ACI Mater. J. 117, 243–254 (2020). https://doi.org/10.14359/
51720292
11. Abellán-García, J., Núñez-López, A., Torres-Castellanos, N., Fernández-Gómez, J.: Effect of
FC3R on the properties of ultra-high-performance concrete with recycled glass. Efecto del
FC3R en las propiedades del concreto de ultra altas prestaciones con vidrio reciclado. Dyna
86, 84–92 (2019). https://doi.org/10.15446/dyna.v86n211.79596
12. Soliman, N.A., Tagnit-Hamou, A.: Using glass sand as an alternative for quartz sand in
UHPC. Constr. Build. Mater. 145, 243–252 (2017). https://doi.org/10.1016/j.conbuildmat.
2017.03.187
13. Kalny, M., Kvasnicka, V., Komanec, J.: First practical applications of UHPC in the Czech
Republic. In: Fehling, E., Middendorf, B., Thiemicke, J. (eds.) Proceedings of Hipermat
2016 - 4th International Symposium UHPC Nanotechnology Construction Mateials, Kassel,
Germany, pp. 147–148 (2016)
14. Tagnit-Hamou, A., Soliman, N., Omran, A.: Green ultra - high - performance glass concrete.
In: First International Interactive Symposium UHPC – 2016, vol. 3, pp. 1–10 (2016)
15. AFGC and SETRA: Ultra high performance fibre-reinforced concretes - recommendations,
Associatio (2013)
First Experimental Full-Scale Elevated FRSCC
Slab in South America
Luis Segura-Castillo(&), Diego Figueredo, Iliana Rodríguez,

and Nicolás García
Facultad de Ingeniería, Universidad de la República, Montevideo, Uruguay

lsegura@fing.edu.uy
Abstract. A full-scale steel fibre reinforced self-compacting concrete elevated

slab was built and tested up to failure. The experimental slab (without any con-
ventional reinforcement), consisted of four continuous panels of 3.1 3.1 m2
each, supported by columns two meters above ground. The nominal thickness was
130 mm. 90 kg/m3 of steel fibres were used. For the slab testing, two opposite
panels were punctually loaded. Simultaneously, displacements were recorded and
crack pattern registered. Results show a fan type cracking pattern, corresponding
to what is reported by literature. The maximum loads obtained were Fmax,1 =
156.4 kN and Fmax,2 = 211.8 kN, (dcentral 20 mm), and the ultimate loads
FU,1 = 117.3 kN and FU,2 = 187.9 kN (dcentral 60 mm), i.e. for displacements
approximately three times larger than the one registered at peak load, the slab could
carry more than 75% of the peak load, showing good overall ductility. A simple
software, capable of modelling yield line theory, and suitable for routine appli-
cations, was used to model the slab. The software input is the slab geometry,
boundary conditions, and plastic moments. The results showed that using rea-
sonable simplifications, a good agreement between theoretical models and
experimental results can be obtained if characteristic material properties are used in
the model.
Keywords: Steel fibre Self-compacting Concrete Test Failure
1 Introduction
Fiber-reinforced concrete (FRC) has been successfully used as the only reinforcement
of elevated slabs, with dozens of multi-story buildings already completed [1]. The use
of FRC in these elements simplifies the construction with cost and time savings
associated to a reduction in labour and equipment while improving crack control and
durability. These saving are further increased when self-compacting FRC (FRSCC) is
used. For its development, elevated FRC slabs have been studied both experimentally
and numerically by several authors. In general, a good response has been observed in
the experimental slabs already tested, both in terms of load carrying capacity, crack
patterns, and also in terms of ductility [2–6].
Different recommendations and guides exist for the design of elevated FRC slabs,
most notably the ACI 544.6R report [1] and the fib Model-Code 2010 [7]. A notable
difference between them is the procedure to obtain the constitutive equation for the
FRC. While the first one recommends the use of small plates samples according to
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_77
874 L. Segura-Castillo et al.
ASTM standard C1550 [8], the second one prefers the use of small beams, according to
EN 14651 [9], and then applies coefficients to account for fibre distribution and
structural redundancy of the element to be designed.
For structural evaluation both guides accept the use of the Yield Line Method,
developed by Johansen [10] in parallel with Gvozdev [11], for the estimation of the
ultimate load-carrying capacity of slabs. These methods, originally developed to be
performed by hand, have recently been successfully implemented in computer pro-
grams, which allows the analysis of more complex geometries and load cases. E.g. the
software LimitState:SLAB [12], which is based on the discontinuity layout optimization
(DLO) method [13].
One of the main focus of the Structural Concrete Group (GHE – Grupo de Hor-
migón Estructural) of the Universidad de la República, Uruguay, is the introduction of
FRC for structural use in the country. Relevant recent developments by the group are; a
simplified test for the FRC characterization in tension (Montevideo Test) [14], the study
of FRC anisotropy [15], Precast FRC Sandwich Panels [16] or diaphragm walls
strengthened by a layer of FRC [17].
More recently, in a collaboration project with the construction company
ABENGOA-TEYMA, the group built and tested up to failure the first full-scale
experimental elevated FRSCC slab in South America. The objective of this paper is to
describe the main aspects of this experimental program and to show its preliminary
results, mainly, the load–displacement behaviour under the point of loading and the
general crack pattern. Also, the experimental results are compared with those obtained
numerically by a computational program, based in the yield line theory.
2 Methodology
2.1 Experimental
A full-scale fibre reinforced self-compacting concrete (FRSCC) elevated slab was built
and tested up to failure under Short-term loading. The construction site was located in
Montevideo, Uruguay. The slab was built in June 2018 and tested between June and
August 2019.
The experimental slab, which did not have any type of conventional reinforcement,
was 6.2 6.2 m2, dimensions in plant, and with a nominal thickness (t) of 130 mm
(Fig. 1). It consisted of four continuous panels of 3.1 3.1 m2 each, supported on
three lines of three cast in-situ reinforced concrete columns with a square cross-section
of 0.2 0.2 m2. The slab was suspended two meters above ground level. A FRSCC
raft foundation supported the columns.
The materials used for the FRSCC mix were: cement CEM I, water, two natural
river fine aggregates, granitic coarse aggregates with maximum dimension of 19 mm,
superplasticizer and hooked-end steel fibres (Ferrofiber AR65) with a length (l) of
50 mm and a diameter (d) of 0.77 mm, resulting in an aspect ratio (k = l/d) of 65.
A fibre content of 90 kg/m3 was used (equivalent to 1.14% by volume). The mix was
designed to obtain both pumpable and self-compacting characteristics. The self-
compacting requisites of the mix were evaluated by executing the slump-flow [18]
(obtaining a slump flow of 660 mm), and through visual inspection [19].
First Experimental Full-Scale Elevated FRSCC Slab 875
(a) (b)
Fig. 1. Experimental program: a) slab plant; and b) sketch of the slab and foundation.
The FRSCC used for the elevated slab was evaluated under compression at 28 days
by means of cylindrical specimens with a diameter of 150 mm and a nominal height of
300 mm, obtaining a mean compressive strength equal to 64.3 MPa. The post-peak
material tensile behaviour was determined by testing 5 beams using a three-point
bending test, also at 28 days, according to EN 14651 [9]. For all the specimens, the
limit of proportionality fL, the residual flexural tensile strengths fR1, and fR3 for a crack
mouth opening displacement (CMOD) of 0.5 and 2.5 mm, respectively, were calcu-
lated. Average values obtained were (coefficient of variation between brackets): fL =
7.02 MPa (8%); fR1 = 10.23 MPa (10%) and fR3 = 8.63 MPa (12%), with which the
following characteristic residual tensile strengths were computed: fL,k = 6.10 MPa; fR1,
k = 8.58 MPa and fR3,k = 7.00 MPa.
The raft foundation, with a nominal thickness of 130 mm and the same FRSCC mix
than the used in the elevated slab, was cast both to provide support for the structure,
and also to serve as first full scale trial of the mix. It was directly resting on top of a
400 mm compacted gravel layer. Reinforcement Steel was left embedded to connect
both the columns and also the cables used in the load tests. The columns were cast with
conventional reinforced concrete.
The slab was cast by one single pumping step, in approximately 6 min. The cast
sequence is shown in Fig. 2a. The dotted line represents the path followed by the
concreting hose, which was obtained after analyzing a recording of the process. It was
indicated to the operators of the hose to launch the concrete mainly from the centre of
each panel, following a clockwise direction. As can be seen, although the instructions
were somewhat followed, the launch tended to lead a more random course. An image
during the cast is shown in Fig. 2b. Due to the use of the self-compacting concrete, the
slab was appropriately compacted without any poker vibrating. Two void plastic
cylinders were embedded in the centre of the slabs designed for testing, foreseeing the
passage of the testing rod. After casting the slab was water-cured for 5 days.
For the slab testing, two opposite panels were punctually loaded, with
loading/unloading cycles. The loading took place on one panel first (panel 1) and, after
its rupture, the second loading occurred (panel 2). With this scheme, a certain loss of
(a) (b)
Fig. 2. Cast of the slab: a) cast sequence; b) Image during concrete launching.
rigidity is expected for the second panel after testing the first panel. However, in
previous experiences this influence was shown to be reduced, as the cracking take place
in different areas of the slab.
The load test of each panel was carried out in two stages: the first one was load
controlled, recording every measure for constant load increments until the peak load
was registered. After achieving the peak load, in the second stage, the test was con-
trolled by the displacement of the central point of the panel, recording every measure
for constant displacement increments.
To introduce the load in the panel, a hydraulic hollow-jack, controlled by a hand
pump, was used (Fig. 3). The jack was resting over a steel plate that was over the slab.
To induce the load downward, the top of the jack reacted on another steel plate that was
connected to a tie rod which, in turn, passed through a hole towards the bottom of the
slab. Below the slab level, the rod was connected to four suspension cables which were
connected to the bottom of the four columns around the panel under testing. One of
these cables had a manual winch to ensure the tie rod was in a vertical position. The
hydraulic pump was manually controlled to allow for a stepwise increase of the load by
18 kN (200 psi in the jack) and 9 kN (100 psi in the jack) increments for the slab 1 and
2 respectively. The load force was measured from a single load cell placed above the
hydraulic jack, and between the jack and the plate connected to the tie rod.
Simultaneously with the slab loading, seven points (Fig. 1a) of each panel were
selected to record the vertical slab displacements by means of dial indicators. In this
paper only the displacements of the main indicator of each panel, placed in its centre
(Fig. 1a), are used. The dial indicators were installed on 5 steel beams fixed between
columns (Fig. 3), which act as a type of yoke, to record the relative displacement
between the slab and the support columns and minimize the influence of extrinsic
deformations. The load cell and the dial indicators were measured after each load
increment step. Finally, when the load tests were completed, the slab was surveyed for
cracks and the crack pattern registered.
Fig. 3. Experimental FRSCC test set-up.
2.2 Numerical
The structural response of the slab was analyzed by the software LimitState:SLAB [12].
The software is capable of modelling slabs with the yield line theory and is suitable for
routine applications. The ultimate load-carrying capacity and the yield lines associated
can be estimated for a given slab geometry, boundary conditions, plastic moment and
type and position of loads. The program is also able to consider the self-weight of the
element.
The experimental slab geometric parameters were used. Load was simplified to a
point load in the position of the jack. To obtain the plastic moment a cross-section layer
discretization analysis was used, as shown elsewhere (e.g. [20, 21]). Fib Model-Code
2010 [7] design rules were used to obtain the material constitutive law using the
EN14651 test results.
Mean and characteristic plastic moment of resistance per unit length obtained were
mp,m = 26.33 kNm/m and mp,k = 21.67 kNm/m, respectively. In each case, the same
plastic moment was used both for positive and negative yield lines, assuming an
isotropic behaviour.
3 Results

Collapse loads of Pc,m = 198.1 kN and Pc,k = 167.9 kN were obtained with the
LimitState: Slab software for the mean and the characteristic residual tensile strengths,
respectively. The associated collapse mechanism for both cases can be seen in Fig. 4.
The yield lines are shown in blue and red for positive and negative moments,
respectively. It can be seen that the yield-line pattern correspond to the previously
described by other authors for centre point loads, with a fan type characterized by
several positive moment cracking underneath and negative moment cracking in a

circular shape passing by the four footprints of columns at the top of the slab.
Fig. 4. LimitState: Slab collapse mechanism.

The crack pattern observed after the test was finished are schematically represented,
both for positive (Fig. 5a) and negative (Fig. 5b) moments. Thick lines indicate large
(w 0.5 mm) and medium (0.5 mm > w 0.1 mm) cracks. Thin lines indicate
smaller cracks (w < 0.1 mm). Examples of cracking in the slab are shown in Fig. 6.
The positive macro crack at the edge of the slab (see Fig. 6a), was more than 3 mm
wide at the end of the test. The results showed a strong resemblance with both the
patterns reported by literature and with the numerical result previously described.
(a) (b)
Fig. 5. Crack pattern observed in experimental test: a) bottom surface; b) top surface.
(a) (b)
Fig. 6. Examples of cracking in the slab: a) bottom surface; b) top surface.
The load-displacement curves for the tested slab are shown in Fig. 7. The exper-
imental results for the two panels are shown in black lines. Notable points are high-
lighted with circles and also the Plot is amplified in the left of the Figure, showing the
first 10 mm of displacement.
250 250 29,8;

64,4;
211,8
198,1 187,9
200 200
167,9
Load - F [kN]
150 150
19,5; 58,5;
156,4 117,3
100 100
1,94; 1,94; 86,7
86,7 Panel 1 Panel 2
50 50
Num_m Num_k
0 0
0 2 4 6 8 10 0 10 20 30 40 50 60
Mid-span displacement - δ [mm]
Fig. 7. Experimental test results.
General behaviour shows a linear trend up to approximately 80 kN. This value

approximately corresponds to the appearance of the first visible cracks in both panels.
After cracking, a gradual loss of stiffness is registered, which continues until the peak
load is reached. The maximum loads obtained were Fmax,1 = 156.4 kN and Fmax,2 =
211.8 kN, with a corresponding deflection in the centre of the panel of dFmax,1 = 20
mm and dFmax,2 = 29.8 mm. Suffix 1 and 2 indicate results for panel 1 and 2
respectively.
After peak load, a slowly descending branch developed until the test was inter-
rupted. The tests were concluded when a central deflection of dmax,1 = 58.5 mm and
dmax,2 = 64.4 mm was reached in each of the panels, which represents between two to
three times the displacement for the maximum load (dFmax,i). For the maximum dis-
placements, the recorded load were FU,1 = 117.3 kN and FU,2 = 187.9 kN, respec-
tively, which represents approximately 75% and 89% of each maximum load. The
large displacements observed, while still holding a large amount of the maximum load,
indicates a good overall ductility.
The difference between the two slabs may be attributable to several factors, e.g.
possible difference in the fibre distribution and orientation, or the experimental error. In
particular the real slab thickness was slightly different from the nominal values. The
average of the thickness measured in different points of the slab was t1 = 124.8 mm
and t2 = 139.3 mm for each slab.
The numerical estimation of the ultimate load is also plotted in gray lines in Fig. 7,
with the values obtained with the mean and characteristic values of the residual strength
plotted in continuous and dashed lines, respectively. It can be seen that the numerical
estimation, calculated with the mean values of residual strength fall above the exper-
imental values, being 69% and 5% larger for each of the panels. However, the esti-
mation made with the characteristic values fall between the two experimental results
(43% above panel 1 and 11% below panel 2). This may indicate that, if the numerical
method is adequate for the FRC slab analysis, the use of the mean results of residual
strength is not a good predictor of the structural response of the slab. Still, a deeper
analysis should be carried out in order to evaluate other possible factors that may be
influencing the results.
4 Conclusions
A full-scale steel fibre reinforced self-compacting concrete elevated slab was built and
tested up to failure. Results showed:
• A fan type cracking pattern, corresponding to what is reported by literature. The
maximum loads obtained were Fmax,1 = 156.4 kN and Fmax,2 = 211.8 kN and the
ultimate loads FU,1 = 117.3 kN and FU,2 = 187.9 kN. The ultimate load was reg-
istered for displacements (dcentral 60 mm) approximately three times larger than
the one registered at peak load (dcentral 20 mm), still carrying more than 75% of
the peak load, which implies a good overall ductility.
• A simple software, capable of modelling yield line theory, and suitable for routine
applications, was used to model the slab. The results also showed that using rea-
sonable simplifications, a good agreement between theoretical models and experi-
mental results can be obtained if characteristic material properties are used in the
model.
Future work includes the analysis of fibre distribution and structural redundancy of
the slab, in order to improve the analysis of the slab.
Acknowledgements. The authors would like to thank ABENGOA-TEYMA for financial sup-
port, physical resources and the assistance of their staff (especially Mauricio Montaña, Ramiro
Rodriguez and Ignacio Horta without whom this work could never have been completed). The
authors also thank the collaboration of the following companies and institutions: Laboratorio de
Vialidad, MTOP, where testing of small specimens was carried out; Ferrocement, for supplying
the fibres; Atenko, for supplying the scaffolding; and Guardia Republicana – Ministerio del
Interior, on whose premises the slab was built. The authors thankfully acknowledge the students,
assistants and professors that helped during the project. Funding was also made available from
the Agencia Nacional de Investigación e Innovación (ANII), Uruguay, through the Research
Project Herramientas para la innovación - 2017 – “HPI_X_2017_1_141290”.
References
1. ACI Committee 544: ACI 544.6R-15. Report on Design and Construction of Steel Fiber-
Reinforced Concrete Elevated Slabs. American Concrete Institute, Farmington Hills (2018)
2. Mobasher, B., Destrée, X.: Design and construction aspects of steel fiber-reinforced concrete
elevated slabs. In: SP-274 Fiber Reinforced Self-consolidating Concrete: Research and
Applications, C. A. LF, Ed., pp. 95–107 (2010)
3. Salehian, H., Barros, J.A.O.: Prediction of the load carrying capacity of elevated steel fibre
reinforced concrete slabs. Compos. Struct. 170, 169–191 (2017)
4. Colombo, M., Martinelli, P., di Prisco, M.: On the evaluation of the structural redistribution
factor in FRC design: a yield line approach. Mater. Struct. 50(1), 1–18 (2017)
5. Maturana Orellana, A.: Estudio teórico-experimental de la aplicabilidad del hormigón
reforzado con fibras de acero a losas de forjado multidireccionales. Universidad del País
Vasco (2013)
6. di Prisco, M., Martinelli, P., Parmentier, B.: On the reliability of the design approach for
FRC structures according to fib model code 2010: the case of elevated slabs. Struct. Concr.
17(4), 588–603 (2016)
7. FIB: Model Code 2010, Vol 1 & 2. International Federation for Structural Concrete (fib),
8. ASTM: C1550-10a Standard Test Method for Flexural Toughness of Fiber Reinforced
Concrete (Using Centrally Loaded Round Panel). ASTM Stand., pp. 1–14 (2010)
9. EN 14651: Test method for metallic fibre concrete—Measuring the flexural tensile strength
(limit of proportionality (LOP), residual) (2005)
10. Johansen, K.: Yield-line theory. Cement and Concrete Association, London (1962)
11. Gvozdev, A.: The determination of the value of the collapse load for statically indeterminate
systems undergoing plastic deformation. Int. J. Mech. Sci. 1, 322–335 (1960)
12. LimitState Ltd.: LimitState:SLAB, Sheffield, U.K
13. Gilbert, M., He, L., Smith, C.C., Le, C.V.: Automatic yield-line analysis of slabs using
discontinuity layout optimization. Proc. R. Soc. 470(2168), 23 (2014)
14. Segura-Castillo, L., Monte, R., De Figueiredo, A.D.: Characterisation of the tensile
constitutive behaviour of fibre-reinforced concrete: a new configuration for the wedge
splitting test. Constr. Build. Mater. 192, 731–741 (2018)
15. Segura-Castillo, L., Cavalaro, S.H.P., Goodier, C., Aguado, A., Austin, S.: Fibre distribution
and tensile response anisotropy in sprayed fibre reinforced concrete. Mater. Struct. 51(1), 1–
12 (2018). https://doi.org/10.1617/s11527-018-1156-5
16. Segura-castillo, L., García, N., Rodriguez Viacava, I., de Sensale, G.R.: Structural model for
fibre-reinforced precast concrete sandwich panels. Adv. Civ. Eng. 2018, 11 (2018)
17. Segura-Castillo, L., Aguado De Cea, A.: Bi-layer diaphragm walls: evolution of concrete-to-
concrete bond strength at early ages. Constr. Build. Mater. 31, 29–37 (2012)
18. EN 12350-8: Testing fresh concrete - Part 8: Self-compacting concrete - Slump-flow test.
Comité Européen De Normalisation (2010)
19. ASTM C1611: Standard Test Method for Slump Flow of Self-Consolidating Concrete,
pp. 1–6. ASTM International, USA (2005)
20. Segura-Castillo, L., Monte, R., De Figueiredo, A.D.: Analytical correlation between
Montevideo test (MVD) and three-point bending test for fibre reinforced concrete (FRC). In:
4th FIB Congress, p. 8 (2018)
21. de la Fuente, A., Aguado, A., Molins, C., Armengou, J.: Numerical model for the analysis up
to failure of precast concrete sections. Comput. Struct. 106–107, 105–114 (2012)
Masonry Walls Strengthened with Fiber
Reinforced Concrete Subjected to Blast Load
Salah Altoubat1(&), Abdul Saboor Karzad1, Moussa Leblouba1,

Mohamed Maalej1, and Pierre Estephane2
1
Department of Civil and Environmental Engineering, Sustainable Construction
Materials and Structural System Research Group, University of Sharjah, Sharjah,
UAE
saltoubat@sharjah.ac.ae
2
GCP Applied Technologies, Dubai, UAE
Abstract. The aim and objective of this research project was to develop and
investigate a rapid method of strengthening the Unreinforced Masonry walls
(URM) for out-of-plane action. A total of 12 URM walls, built inside a precast
fiber reinforced concrete boundary frame, were strengthened with a cement
based concrete mix with or without steel reinforcement. The concrete mixes
included Fiber Reinforced Concrete (FRC) (two dosage of fibers (4.6 &
6.0 kg/m3)), Engineered Cementitious Composite (ECC), and plain concrete
shotcrete (PLS). Two out of 12 walls were strengthened with steel mesh rein-
forced shotcrete (MRS) such that a steel mesh was attached to the surface of the
walls prior to spraying with concrete. In addition, one wall was plastered only
with conventional cement mortar (RF) to serve as a reference. The strengthening
materials were applied to both sides of URM walls using shotcrete technology.
The walls were exposed to the air pressure of an actual blast from 10 kg of TNT
explosive at a 1 m stand of distance. It was found that strengthening the walls
with the proposed method significantly improved the out-of-plane behaviour of
the URM walls. The results showed that the walls strengthened with ECCS
exhibited the least damage followed by FRS-6, FRS-4.6, MRS, and PLS.
Whereas, the reference wall (RF) did not survive the blast induced pressure and
fully shattered into small pieces.
Keywords: URM walls Out-of-plane Strengthening FRC ECC

Synthetic fibers Steel mesh
1 Introduction
Rising threats of terrorism have boosted the need for engineers to build structures (such
as embassies, airports, and other governmental buildings) capable of withstanding
major impact loads, specifically air pressure generated by explosions. Unreinforced
masonry (URM) walls have become a common building practice in the construction
industry. URM walls compared to other building materials are most vulnerable to
damage in accidental events such as blasts, earthquakes or any similar instances [1].
The destruction of masonry walls in the event of a blast, can lead to severe damage or
full collapse of the building thus resulting in casualties. However, due to the ease of
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_78
884 S. Altoubat et al.
construction, feasibility & availability of these materials, and the simplicity in building,
it is still a preferred building material [2].
Masonry walls are primarily designed for in-plane forces (e.g. combined gravita-
tional and horizontal (wind and/or seismic) loads) [3]. Due to its composite structure
and inherent weak directions (along the head and bed joints) unreinforced masonry
behaves in a highly anisotropic manner. This emphasizes the importance of knowing
the parameters of masonry walls strength. Taking into consideration the action of the
in-plane forces, walls exhibit the behaviour depicted in Fig. 1. As can be seen in figure,
the wall is experiencing an in-plane force which results in a deflection in the wall.
Adding to that, it is interesting to note that due to the deflection, there is a high amount
of energy release, especially in the corners where deflection is greater, which ends up in
diagonal cracks. Coming to the action of the out-of-plane forces, the walls experience
severe damage since it is not designed to withstand these types of loads [4]. The
masonry unit of the structure cannot resist the out-of-plane forces thus shattering the
wall into pieces as illustrated in Fig. 1.
(a) (b)
Fig. 1. Wall behaviour subjected to forces, a) in-plane, b) out-of-plane
To reduce the vulnerability of masonry walls under sever loading is to retrofit such
walls. Many researchers around the globe have looked for methods of strengthening the
walls to resist out-of-plane forces and suggested different methods of strengthening and
their benefits [2, 5–10]. For instance, Sajal verma et al. [8] studied the influence of blast
load on structures. In their study different methods of improving the out of plane
behaviour is suggested such as increasing the mass of walls (as it would decrease the
acceleration when subjected to external pressure), use of polyethylene fibers (which
showed promising results), and using glass fiber reinforcement. In another study by
Carney and Myers [5] the out-of-plane behaviour of URM walls strengthened with FRP
and subjected to blast loading was investigated. Their results showed that the
strengthening method using near surface mounted (NSM) FRP rods with anchorage is
an effective method of strengthening.
As mentioned in the literature review, there have been initiatives that have been
conducted to find feasible solution for strengthening the URM walls against out-of-
plane forces. Similarly, this research was conducted to study out-of-plane behaviour of
the strengthened URM walls subjected to actual blast load which is scarce in the
literature. This research focuses on showing the real effects of the blast on the URM
Masonry Walls Strengthened with Fiber Reinforced Concrete 885
walls with conventional cement mortar and walls strengthened with different types of
concrete using shotcrete technology.
2 Air Blast Background
Air pressure waves also called air blast waves is the resultant of sudden release of
energy from detonation of an explosive material. The generated waves initiate with a
single pulse of increased air pressure that last in milliseconds and travels faster than
sound. The impact of blast waves on the object due to an explosion is known as
overpressure and its magnitude & duration depends on two factors: the stand-off dis-
tance and the amount of charge exploded [11–14].
Figure 2 shows the types of explosive-induced shockwaves which are Friedlander
wave, free-field wave, and complex wave. The idealized form (Friedlander wave) of the
free-field blast wave can be altered considerably along its propagation by the medium
encountered morphology thus can create multiple wave reflections (complex wave).
Even the free-field wave characteristic can change due to the reflection of the blast
waves from ground. For instance, if the blast wave is reflected by rigid boundary or
surroundings, the overpressure can be enlarged by up to 14 times. Particularly, this
situation can occur with more complexity in an urban environment or underground
structures due to presence of multiple boundaries [11, 12, 15].
Fig. 2. Types of blast-induced shock waves; a) idealized form of wave, b) free-field or open-air
blast wave, c) complex or closed environment blast wave (source: [16])
The objective of this research study was to investigate a rapid strengthening method to
enhance the out-of-plane resistance of URM walls against lateral loads. For that a series
of URM walls were built and strengthened with different types of concrete mixes with
and without steel mesh reinforcement. The strengthening mixes were applied using
shotcrete technology and the walls were tested subjected to an actual blast condition.
3.1 Geometry
A total of 12 masonry walls were constructed inside a precast Fiber Reinforced
Concrete (FRC) boundary frame (fiber dosage of 2.3 kg/m3), each measuring 1.27
1.25 0.1 m made of hollow concrete blocks of size 400 200 100 mm. The FRC
boundary frames (1.55 1.92 m) were reinforced using 16 mm diameter steel bars in
both directions. The masonry units (hollow concrete blocks) were made of plain
concrete with an average 28 days compressive strength of 7.5 MPa. The hollow
concrete blocks were built inside the frame using conventional cement mortar between
the layers and the vertical gaps. Figure 3 shows the dimensions of the walls and the
reinforcement details of the frame constructed for the blast test. As depicted in Figure,
the depth of the bottom side has been increased to provide an extra grip when buried
into the ground during blast test.
Fig. 3. Geometry and reinforcement details (units are in cm unless specified in the drawing)
3.2 Strengthening Methodology

The constructed walls in this project consisted of duplicate walls strengthened in the
out-of-plane direction. Each pair of strengthened walls was sprayed with a different
type of concrete mix on both sides of the walls (with a thickness of around 4.5 cm on
each side) using shotcrete technology (see Fig. 4). Four different types of concrete
mixes were sprayed to the walls in addition to a reference wall that was plastered with
cement mortar only. Table 1 shows the type of concrete mix used and the corre-
sponding labels. For ease in identification of each type of wall, the walls were labelled
according to the mix design sprayed on to it.
Table 1. Types of concrete mixes applied on the surface of the walls

Strengthening mixes Short form
Fibre Reinforced Shotcrete with fiber dosage rate of 4.6 kg/m3 FRS-4.6
Fibre Reinforced Shotcrete with fiber dosage rate of 6 kg/m3 FRS-6
Engineered Cementitious Composite with fiber dosage rate of 16 kg/m3 ECCS
Mesh Reinforced Shotcrete (reinforced with steel mesh) MRS
Plain Concrete Shotcrete (Control) PLS
Reference Walls (Plastering mortar only) RF
As mentioned above, four different concrete mix was designed to be used for
strengthening the walls. The mix design proportions of these concrete materials are
presented in Table 2.
Table 2. Mix design proportions of the mixes sprayed to the URM walls from both sides
Type proportions (kg/m3) PLS & MRS FRS-4.6 FRS-6 ECCS
OPC 420 420 420 1449
Silica fume 25 25 25 145
0-5 mm aggregates 1480 1480 1480 0
Dune sand 345 345 345 0
Water (L/m3) 190 190 190 478
ADVA flow 480 (L/m3) 6 11.5 12.5 11.5
Strux 90/40 0 4.6 6 0
Spectra 900 0 0 0 16
Fig. 4. Spraying the walls using shotcrete technology

For FRS mixes the fibers used was synthetic macro-fibers commercially called Strux
90/40 with a length of 40 mm, a rectangular cross Sect. (1.4 0.105 mm), a modulus of
elasticity of 9.5 GPa, and a tensile strength of 620 MPa. The ECCS mix which can be
considered a special case of FRS, incorporates micro fibers commercially named Spectra
900 with a length of 12 mm, diameter of 0.039 mm, modulus of 66 GPa, and a tensile
strength of 2610 MPa. Two walls were strengthened with mesh reinforced shotcrete. To
achieve that, the A98 (Ø5 mm @200 200 mm) steel mesh was attached to the surface
of the masonry walls on both sides before spraying the walls with plain concrete. To
characterize the concrete mixes used, some standard tests were also conducted. Tables 3
and 4 present the test result of the cubes, prisms, and round panels samples tested
according to ASTM C39, ASTM C293, and ASTM C1550.
Table 3. Compressive and flexural strength (MPa)

Mixes properties PLS & MRS FRS-4.6 FRS-6 ECCS
Average Compressive Strength at 28 days 68 59 60 53
Average Flexural Strength at 28 days 4.7 4.6 5.3 4.3
Table 4. Energy absorption of the shotcrete mixes obtained by testing the round panels
Mixes parameters PLS & MRS FRS-4.6 FRS-6 ECCS
Deflection (mm) 39 39 39 39
Energy Absorption (J) 54 168 208 783
3.3 Test Setup and Instrumentation

From the observations in the trial tests, it was decided to place the wall at 1 m stand of
distance from the explosive material and the walls were properly supported. Therefore,
a rigid steel frame with bracing supports (see Fig. 5) was designed and fabricated. The
rigid frame was used to support and hold the walls in its place during the explosion and
to prevent the walls from tilting or flipping after the shock amid blast. The walls were
clamped using steel clamps at four spots to the rigid steel frame to prevent separation of
steel frame and the wall during the shock. For further support, the walls were also
buried in the ground at the base. A total of four tests (first test was trial) were conducted
with a set of 3 walls in each test. For each test the walls were placed around the
explosive such that if looked from above, the walls were staggered in a triangular shape
(see Fig. 5). An amount of 10 kg of TNT explosive was placed in between the three
walls at a height of 60 cm above the ground to match the mid-height of the walls.
In order to measure the acceleration and the air pressure exerted on the walls,
specialized instruments (pressure sensor and accelerometer manufactured by PCB
Piezotronics) were installed on the surface of the walls as shown in Fig. 5. The pressure
sensor was fixed at mid-height facing the explosive material and accelerometer was fixed
at mid-height on the opposite side or in other words at the back of the walls. The sensors
attached to the walls recorded the peak air pressure and the acceleration of the wall with
respect to time. The results obtained from the accelerometer and the pressure sensors
were gathered for each wall and their behaviour was observed against the blast waves.
Fig. 5. Test setup (actual and schematic)

4.1 Damage Identification Through Cracking Pattern and Signal
Processing
4.1.1 Cracking Pattern
After the blast test the walls were visually inspected to determine the cracking pattern
and ultimate damage. Figure 6 shows the photo of all the tested walls along with
diagrams of signal processing implemented on acceleration data recorded during the
tests. It should be noted that all the photos in Fig. 6 are captured from back side of the
walls (opposite side to the side facing blast wave). Referring to the photo in Fig. 6a, the
ECCS walls did not exhibit any noticeable damage except one horizontal crack in the
mid height which is hardly visible. Figure 6b illustrates that the FRS-6 wall exhibited
major diagonal cracks with one or two light horizontal cracks. The crack had small
width and were not very severe to cause any damage, so the wall almost kept its initial
shape. Similarly, the FRS-4.6 walls experienced multiple cracking pattern with a bit
more intensity and wider cracks (see Fig. 6e) compared to the FRS-6 wall. However,
the MRS walls had several much sever and deep cracks that resulted in noticeable
damage shown in Fig. 6c.
The control walls that were PLS (reference for shotcreted walls) and RF (presenting
the conventional method of constructing walls) did not survive the blast waves as
shown in Fig. 6d and 6f. As can be seen in Fig. 6d, the PLS wall exhibited sever
damage associated with spalling of a big chunk of shotcrete from the back side (side
opposite to the blast face). In addition, the side facing the blast force also had a
significant and very deep crack like a punching action that even pushed the surface of
the wall inward. This crack was at the bottom end of the wall only and was running
horizontally. On the other hand, the RF wall completely shattered into pieces and
disappeared from inside the precast boundary frame due to the pressure exerted from
the blast (see Fig. 6f).
4.1.2 Signal Processing

To gain more insight on the damage endured by the walls as a result of the blast, signal
processing was conducted using Fast-Fourier Transform (FFT). Four types of walls
were selected to conduct this analysis: ECC, FRS-6, MRS, and PLS. Given the
acceleration record for each wall, the analysis is carried out in three steps:
1. Select a low-pass cut-off normalized frequency (0\fcut \1) through a visual
analysis of the FFT curve;
2. Apply the Butterworth 2nd order filter to filter out all normalized frequencies higher
than fcut ;
3. Perform FFT on the filtered acceleration record.
The above steps were repeated for acceleration data obtained earlier through drop
weight tests on duplicate walls. This step is important for that it gives a baseline FFT
curves to compare with those obtained from blast tests.
Figure 6 shows the amplitude spectra for each pair of walls after the impact and
blast tests. It is clear from the spectra that frequency content of the blast-tested walls
shifted towards the left indicating a decrease in the fundamental frequency of the walls.
The decrease of the fundamental frequency means that the walls became flexible during
and after the blast tests; this is an important indicator that the walls has experienced a
damage that led to the reduction of its stiffness, thus, lengthened its period.
Excluding the PLS wall, the amount of the frequency shift (i.e., reduction of the
fundamental frequency) is the highest in the MRS wall, followed by FRS-6, then
ECCS. This trend suggests that the among these three wall-types, the ECCS wall
endured the least damage and the MRS wall experienced the highest damage. This has
been also supported by the photos of the corresponding walls included in the same
Figure.
Now for the PLS wall, the impact test that should have served as a baseline resulted
in a damage to the wall, thus, the results depicted in Fig. 6d are for a PLS wall with a
certain degree of damage, hence, the frequency reduction in the duplicate wall after the
blast is seen to be relatively small (the spectra in Fig. 6 are in logscale).
(a) ECCS
(b) FRS-6
(c) MRS
Fig. 6. Amplitude spectra for the selected wall specimens and photo of tested walls.
(a) PLS
(e) FRS-4.6 (f) RF
Fig. 6. (continued)
4.2 Pressure and Acceleration Time Histories

The results collected from blast test were the recorded air pressure, acceleration,
cracking behaviour and the level of damage. The air pressure recorded during the blast
test is plotted against the blast duration in Fig. 7a and shows that the maximum air
pressure was around 130 MPa. The air pressure plot represents a complex wave due to
the rebound of wave after collision on the surface of three surrounded walls.
To confirm the arguments made above concerning the damage level caused by the
blast, the diagrams of acceleration time histories are plotted in Fig. 7b and 7c.
Figure 7b shows that the acceleration response of FRS-4.6 degrades quicker than the
FRS-6 and ECCS walls. This indicates that the FRS-4.6 wall became more flexible due
to the damage caused by the blast. Similarly, The FRS-6 wall that had higher level of
cracking than the ECCS wall, exhibit higher rate of degradation in acceleration
response compared to ECCS wall. However, the ECCS wall shows more stiff beha-
viour suggesting of less damage. Meanwhile, it can be inferred from Fig. 7c that the RF
wall, which was fully shattered after the blast shows significant degradation of
acceleration response compared to PLS wall and the other walls.
Overall it should be noted that all the walls tested in this research had cracks on the
opposite side to the blast side with different level of severity, depth, and width. The side
facing the blast didn’t show any visible cracks except for the control walls. The
behavioural comparison of the tested walls demonstrated that the ECCS and FRS
strengthened walls had the highest level of energy absorption thus mitigated the level of
damage caused by the air blast waves. While the MRS and PLS walls were severely
damaged. In contrast the conventional method of constructing masonry walls had zero
resistance against the air pressure caused by the blast force.
TNT-BLAST INDUCED AIR-PRESSURE

140
120
100
PRESURE (MPA)
80
60
40
20
0
0 2 4 6 8 10 12 14 16 18 20
-20
TIME (MSEC)
(A)
ECCS FRS-6 FRS-4.6 PLS RF
2200 2200
ACCELERATION (G)
1600
ACCELERATION (G)
1600
1000 1000
400 400
-200 -200
-800 -800
-1400 -1400
-2000 -2000
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
TIME (MSEC) TIME (MSEC)
(B) (C)
Fig. 7. Pressure exerted on the walls and acceleration time histories
A total of 12 URM walls were built and then strengthened using cement based fiber
reinforced shotcrete mixes with and without conventional steel reinforcement. Walls
with conventional plaster were also constructed and used as a reference. The walls were
subjected to 10 kg of TNT explosive at a stand-off distance of 1 m. The parameters
measured during the blast test were the air pressure and acceleration.
The acceleration measurement was used to calculate the frequency content of the
walls using FFT and Filtering process. The walls were also visually examined for
damage after the explosion shock. The results of frequency content, degradation in the
acceleration time histories and visual inspection indicated that addition of fibers sig-
nificantly improves the out-of-plane resistance of the URM wall even in the event of
extreme pressure like blast load. The results revealed that the ECCS wall had the least
damage due to blast load followed by FRS-6, FRS-4.6, MRS, and PLS walls. However,
the walls strengthened with plain concrete materials with or without steel mesh rein-
forcement exhibited sever damage compared to the walls strengthened with fiber
reinforced concrete. Moreover, the reference wall with cement mortar plaster
(RF) could not survive the blast waves and was fully damaged.
References
1. Sameer, A., et al.: Out-of-plane Strengthening of masonry walls with reinforced composites.
J. Compos. Constr. 5(August), 139–145 (2001)
2. Ahmad, S., et al.: Experimental study of masonry wall exposed to blast loading. Mater.
Constr. 64(313) (2014)
3. Meisl, C.S.: Out-of-plane seismic performance of unreinforced clay brick masonry walls.
The University of British Columbia (2006)
4. Valluzzi, M.R., da Porto, F., Garbin, E., Panizza, M.: Out-of-plane behaviour of infill masonry
panels strengthened with composite materials. Mater. Struct. 47(12), 2131–2145 (2014)
5. Carney, P., Myers, J.J.: Out-of-plane static and blast resistance of unreinforced masonry wall
connections strengthened with FRP. Aci Spec. Publ. 1(SP-230-14), 229–248 (2005)
6. Kalman, D.: Use of steel fiber reinforced cocrete for blast resistance design. Kansas State
University (2010)
7. Maalej, M., Lin, V.W.J., Nguyen, M.P., Quek, S.T.: Engineered cementitious composites for
effective strengthening of unreinforced masonry walls. Eng. Struct. 32(8), 2432–2439 (2010)
8. Verma, S., Choudhury, M., Saha, P.: Blast resistant design of structure. Int. J. Res. Eng.
Technol. 04(13), 64–69 (2015)
9. Casadei, P., Agneloni, E.: Elastic systems for dynamic retrofitting (ESDR) of structures. In:
Cost Action C26 Urban Habitat Constructions under Catastrophic Events – Proceedings of
Final Conference, pp. 939–948 (2010)
10. Dizhur, D., Griffith, M., Ingham, J.: Out-of-plane strengthening of unreinforced masonry
walls using near surface mounted fibre reinforced polymer strips. Eng. Struct. 59, 330–343
(2014)
11. Chakraborty, T., Larcher, M., Gebbeken, N.: Performance of tunnel lining materials under
internal blast loading. Int. J. Prot. Struct. 5(1), 83–96 (2014)
12. I. of Medicine, Gulf war and health: long-term effects of blast exposure, vol. 9 (2014)
13. Goel, M.D., Matsagar, V.A.: Blast-resistant design of structures. Pract. Period. Struct. Des.
Constr. 19(2) (2014)
14. Tai, Y.S., Chu, T.L., Hu, H.T., Wu, J.Y.: Dynamic response of a reinforced concrete slab
subjected to air blast load. Theor. Appl. Fract. Mech. 56(3), 140–147 (2011)
15. Ning, Y.L., Zhou, Y.G.: Shock tubes and blast injury modeling. Chin. J. Traumatol. – Engl.
Ed. 18(4), 187–193 (2015)
16. Mayorga, M.A.: The pathology of primary blast overpressure injury. Toxicology 121(1), 17–
28 (1997)
Smart FRCs
Interfacial Bond Quality in Functionally
Graded Concretes Incorporating Steel Fibres
and Recycled Aggregates
Ricardo Chan1(&), Isaac Galobardes2, and Charles K. S. Moy1

1
Department of Civil Engineering,
Xi’an Jiaotong-Liverpool University, Suzhou, China
r.chan@xjtlu.edu.cn
2
School of Architecture, Planning and Design, Mohammed VI Polytechnic
University, Ben Guerir, Morocco
Abstract. A functionally graded material (FGM) is a material presenting gra-

dation in composition and structure, designed to attend to specific functions.
FGM produced with concrete, known as functionally graded concrete (FGC),
has been studied for several applications by combining layers of distinct types of
concrete showing technical benefits. Due to the material discontinuity, an
interfacial zone is created between layers, named as layer transition zone (LTZ).
As the weakest link between different concrete layers, the bond quality of LTZ
may influence the mechanical behaviour of FGC. In this paper, the quality of
LTZ in FGC was assessed considering the type of aggregate, content of steel
fibres and casting delay between layers. FGC were produced with a top layer of
plain cement concrete (PCC) and a bottom layer of conventional fibre reinforced
concrete (FRC) or fibre reinforced recycled aggregate concrete (FRRAC).
The FGC were assessed for compressive strength and bond strength between
layers. The results indicate that the bond quality of LTZ is affected by casting
delay and compressive strength of each layer. Moreover, it was noticed that the
impact of adding steel fibres was not significant to alter the bond quality in FGC.
Overall, adequate bond strengths were obtained in FGCs with casting delays of
up to 24 h.
Keywords: Functionally graded concrete Bond strength Fibre reinforced

concrete (FRC) Fibre reinforced recycled aggregate concrete (FRRAC)
1 Introduction
The concept of functionally graded material (FGM) consists of producing an optimised

spatial gradation in material composition and structure, which results in tailored
properties designed to attend specific functional requirements [1]. Most recently, FGM
produced with concrete, known as functionally graded concretes (FGC), have been
studied for specific applications, such as precast shield-tunnel segments [2], impact
resistant panels [3], marine structures [4], and pavements [5, 6]; showing significant
technical benefits. These FGC can be produced by combining two or more layers of
distinct types of concrete, resulting in a material with a discontinuous gradation.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_79
898 R. Chan et al.
Due to this discontinuity, an interfacial zone is created between the layers, named
hereafter as layer transition zone (LTZ). The LTZ can be considered equivalent to the
overlay transition zone (OTZ) defined as the interfacial zone between concretes of
different ages, which is studied similarly as the interfacial transition zone (ITZ) be-
tween the coarse aggregates and cement paste [7, 8]. As the weakest link between
layers [9, 10], the bond quality of LTZ can influence the mechanical behaviour of FGC.
Despite its importance, studies about the bond quality of LTZ in FGC are limited
[11–13]. Therefore, this paper aims to study the quality of LTZ in FGC regarding the
influence of several parameters. An experimental program was conducted to produce
and test FGC specimens with natural and recycled aggregates, different contents of
steel fibre (cf) and different casting delays between layers (Dt). Then, the results
obtained from compressive strength and bond strength tests were presented and dis-
cussed. Finally, conclusions were drawn regarding the impact of those parameters in
the bond quality of FGC.

Cement CEM I-42.5 N [14] and tap water at room temperature (20 °C) were used in
this study. According to the manufacturer, cement initial and final setting times are 172
and 226 min, or 2.87 and 3.77 h, respectively. Natural and recycled aggregates were
adopted for both coarse and fine aggregates. Natural aggregates were composed of
limestone gravel and river sand, while recycled aggregates were produced from crushed
demolition waste. Figure 1 shows the size particle distributions of aggregates,
according to BS 812-103.1:1985 [15]. The main properties of aggregates are presented
in Table 1. These are the oven dry density (qrd), the water absorption (WA) and the
coefficients of uniformity (Cu) and curvature (Cc).
100 100
Cumulative passing (%)
Cumulative passing (%)
80 80
60 60
40 40
20 Natural 20 Natural
Recycled Recycled
0 0
1 10 100 0.1 1 10
(a) Size (mm) (b) Size (mm)
Fig. 1. Particle size distribution and grading limits for (a) coarse and (b) fine aggregates.
Interfacial Bond Quality 899
Table 1. Main characteristics of the aggregates.

Type of aggregate qrd WA Cu Cc
(Mg/m3) (%)
Natural Coarse 2.64 1.11 1.67 0.90
Fine 2.70 2.60 2.70 0.84
Recycled Coarse 2.25 7.51 7.82 3.59
Fine 1.90 15.01 10.6 0.95
Natural aggregates present higher values of qrd and lower values of WA in com-
parison with recycled aggregates. These results are expected, due to the old mortar
attached to recycled aggregates, which impacts qrd and WA [16, 17]. The values of Cu
and Cc suggest that the natural aggregates are more uniformly graded than recycled
aggregates.
As reinforcement, hooked-end steel fibres were used. The fibres present 60 mm of
length, 0.75 mm of diameter and tensile strength of 1150 MPa.
2.2 Mixes and Production

In this study, an FGC configuration optimised for the application in structures subjected
to bending was used: conventional concrete in the top layer and reinforced concrete in
the bottom layer [18]. Thus, to study the effect of type of aggregate, two FGC groups
were considered: one with plain cement concrete (PCC) in the top layer and a bottom
layer of fibre reinforced concrete (FRC), as shown in Fig. 2a; and other with PCC in the
top layer and fibre reinforced recycled aggregate concrete (FRRAC) in the bottom
layer, as shown in Fig. 2b. It should be noted that; for this study, each layer of concrete
corresponded to half of the component’s height (h/2).
h/2 PCC h/2 PCC
h/2 FRC h/2 FRRAC
(a) (b)
Fig. 2. FGC groups: (a) PCC+FRC and (b) PCC+FRRAC.
The mix proportions for PCC/FRC and FRRAC are presented in Table 2. The same
cement content of 475 kg/m3 and free water/cement ratio (w/c) of 0.45 were adopted
for both mixes. Due to the difference between the actual moisture content and water
absorption of aggregates, the quantities of aggregates and water were adjusted
according to [19].
900 R. Chan et al.
Table 2. Mix proportions (kg/m3).

Mix Cement Water Type of aggregate Fine aggregate Coarse aggregate
PCC/FRC 475 215 Natural 790 890
FRRAC Recycled 805 715
The impact of steel fibres was assessed by adopting three values of cf: 0.25, 0.50
and 0.75%, in terms of concrete volume. To assess the influence of casting delay
between layers, Dt of 0.50, 6.00 and 24.0 h were applied to produce FGC, representing
the stages before, after and long after the setting time, respectively. A total of 18 FGC
groups were produced, as listed in Table 3. Each group was identified by the following
codification: Letter + First number + Second number. The letter corresponds to type of
aggregate (N for natural and R for recycled), and the first and second numbers to cf and
Dt, respectively. Furthermore, to simplify the analysis, FGC produced only with natural
aggregates are hereafter referred to as N-concretes, and concretes containing recycled
aggregates, as R-concretes.
Homogenous concretes (PCC, FRC and FRRAC) were used to produce cylinders
with 150 mm of diameter and 300 mm of height, while cubes with 150 mm of nominal
size were produced with the FGC presented in Table 3. The cylindrical specimens were
produced according to BS EN 12390-2:2009 [20]. On the other hand, cubic specimens
were cast as follows: first, the bottom layer was casted, compacted and covered with a
polyethylene sheet until the moment of casting the top layer; then, after waiting for the
Table 3. FGC groups considered in this study.

Mix Group Type of aggregate cf (%) Dt (h)
PCC+FRC N0.25-0.50 Natural 0.25 0.50
N0.25-6.00 6.00
N0.25-24.0 24.0
N0.50-0.50 0.50 0.50
N0.50-6.00 6.00
N0.50-24.0 24.0
N0.75-0.50 0.75 0.50
N0.75-6.00 6.00
N0.75-24.0 24.0
PCC+FRRAC R0.25-0.50 Recycled 0.25 0.50
R0.25-6.00 6.00
R0.25-24.0 24.0
R0.50-0.50 0.50 0.50
R0.50-6.00 6.00
R0.50-24.0 24.0
R0.75-0.50 0.75 0.50
R0.75-6.00 6.00
R0.75-24.0 24.0
respective time Dt, the polyethylene sheet was removed and the top layer was casted on
top of the bottom layer. After casting, the specimens were covered again with poly-
ethylene sheet. After 24 h, the specimens were demoulded and cured in water at a
temperature of approximately 20 °C for 28 days according to BS EN 12390-2:2009
[20].
2.3 Test Methods

The compressive strength of each layer of FGC was used as quality control of concrete.
The compressive strength was taken as the average of three cylindrical specimens
assessed according to BS EN 12390-3:2009 [21], as shown in Fig. 3a. The compressive
strength (fcm) is calculated using Eq. (1), which depends on the maximum compressive
load at failure (F) and the specimen’s cross-section area on which the compressive load
acts (Ac).
(a) (b)
Fig. 3. Setups for (a) compressive test and (b) splitting test.
F
fcm ¼ ð1Þ
Ac
The bond strength in FGC was assessed according to the splitting test described in
BS EN 12390-6:2009 [22] and illustrated in Fig. 3b. For each FGC group, three cubic
specimens were used. The bond strength (fctm) is calculated using Eq. (2), which
depends on the maximum load (F) and the length and width of the specimen (L and d,
respectively). Furthermore, the bond quality of LTZ in FGC was categorized according
to the classification suggested in [23] and presented in Table 4.
2F
fctm ¼ ð2Þ
pLd
902 R. Chan et al.
Table 4. Bond quality classification according to bond strength [23].

Bond quality fctm (MPa)
Excellent 2.1
Very good 1.7–2.1
Good 1.4–1.7
Fair 0.7–1.4
Poor 0.0–0.7
The compressive strength of the concrete used in each layer of FGC are presented in
Fig. 4. The values are referred to the average compressive strength of three specimens
evaluated at the age of 28 days, with the respective standard deviation. As previously
indicated, the specimens from top layer are not reinforced while bottom layer speci-
mens are reinforced with steel fibres.
50.0
Top layer Bottom layer
40.0
fcm (MPa)
30.0
20.0
10.0
0.0
Fig. 4. Compressive test results.
The variation in the results for each layer is below 12.5%, assuring the quality of
the concrete produced. The PCC used in the top layer presents fcm equal to 30.3 MPa.
On the other hand, fcm of FRC and FRRAC is enhanced with the increase of cf, as
observed in previous studies [24, 25]. Also, FRRAC presents, on average, 30% lower
fcm than FRC, as expected [26].
The bond strength of 18 FGC groups are presented in Fig. 5. The values are related
to the average of three results and the respective standard deviation, obtained from
FGC cubes evaluated at the age of 28 days.
4.00
0.50 h 6.00 h 24.0 h
3.00
fctm (MPa)
2.00
1.00
0.00
Fig. 5. Bond strength results.
The results from Fig. 5 indicate that N-concretes present higher fctm than R-
concretes, for Dt = 0.50 h. This is expected, since fctm is affected by fcm of each layer,
as pointed in other studies [13, 27]. However, for Dt > 0.50 h, N-concretes present a
decrease in fctm, while the opposite is verified in R-concretes. This trend difference can
be related to transition in failure mode presented in the splitting test with the increase of
Dt, shown in Fig. 6. The classification adopted for the failure modes observed in this
study is described as follows: (Type A) interfacial failure combined with concrete
fracture; (Type B) interfacial failure combined with partial concrete fracture; and
(C) interfacial failure combined with minimum or non-visible concrete fracture.
For Dt = 0.50 h, N-concretes present type A failure mode, with a main crack in the
LTZ and fracture of one or both of the layers, as indicated in Fig. 6a. This strong bond
may be related to the fact that, at that moment, the concrete did not reach the initial
setting time [8]. However, when Dt is increased to 6.00 h and 24.0 h, the failure mode
changes to type B (see Fig. 6b) and type C (see Fig. 6c) in N-concretes, respectively.
This transition in failure mode suggests that the bond strength of LTZ becomes lower
than the tensile splitting strength of the concrete layers as Dt increases. This is expected,
since the final setting time of 3.76 h was already passed when Dt 6.00 h [8].
On the other hand, type A failure mode is observed in all R-concretes specimens,
independently of Dt, as indicated in Fig. 6d–f. This failure behaviour suggests that the
bond strength of LTZ in R-concretes is not reduced by the increase of Dt. In fact, as
previously pointed out, the bond strength in R-concretes is enhanced with the increase
of Dt. The differential stiffness between layers may contribute to this behaviour, as
shown in other study [28]. Nevertheless, further investigation is required to better
identify other factors affecting the bond strength of LTZ in FGC produced with PCC
and FRRAC.
Furthermore, the results presented in Fig. 5 suggest that the bond strength of LTZ
in FGC is not affected by cf, contrasting with other study [29]. This difference can be
related to the size of steel fibres used in each study. Long fibres (60 mm of length) were
used in this study in contrast to the short fibres (13 mm of length) used in [29], which
are more effective in the control of microcracking than long fibres [30]. In addition, the
fibre orientation could have also played a role on the results difference between studies.
904 R. Chan et al.
The wall effect in moulded cubes induces certain fibre orientation which is different
from the one observed in larger concrete elements, such as slabs and panels [31]. As a
result, the stress distribution provided by fibres could be different between moulded
cubes and extracted cubes from large elements. Thus, further investigation should be
done regarding the effect of fibre orientation in the bond strength of FGC.
(a) Type A (b) Type B (c) Type C
(d) Type A (e) Type A (f) Type A
Fig. 6. Failure modes in N-concrete and R-concrete cubes for Dt equal to: (a, d) 0.50 h; (b, e)
6.00 h; and (c, f) 24.0 h.
Finally, Table 5 presents the classification of bond quality of LTZ in FGC, along
with bond strength results and coefficient of variation, in brackets.
According to Table 5, N-concretes present LTZ with bond quality lowering from
“excellent” to “very good”, as Dt increases. On the other hand, the bond quality of LTZ
in R-concretes enhances from “good” or “very good” to “excellent” with the increase of
Dt. Overall, despite the difference in the test results, the bond quality of LTZ in FGC
assessed in this study can be considered adequate. Thus, in this study, FGC behaves as
a monolithic element, increasing its durability.
Table 5. Classification of bond quality of LTZ in FGC.

Group fctm (MPa) Bond quality Group fctm (MPa) Bond quality
N0.25-0.50 3.07 (9.03%) Excellent R0.25-0.50 1.81 (8.25%) Very good
N0.25-6.00 1.85 (9.63%) Very good R0.25-6.00 1.82 (1.73%) Very good
N0.25-24.0 2.09 (9.56%) Very good R0.25-24.0 2.17 (7.94%) Excellent
N0.50-0.50 2.45 (12.72%) Excellent R0.50-0.50 1.57 (2.82%) Good
N0.75-0.50 2.43 (5.51%) Excellent R0.75-0.50 1.77 (8.17%) Very good
N0.75-6.00 2.00 (10.79%) Very good R0.75-6.00 1.98 (4.07%) Very good
4 Conclusions
The bond quality of LTZ in FGC was assessed by means of an experimental program,
considering the impact of type of aggregate, content of fibre and casting delay. As a
general conclusion, the experimental results suggest that the bond quality is strongly
affected by type of aggregate and casting delay but is not influenced by content of fibre.
Furthermore, the results also indicate that an adequate bond quality can be achieved for
casting delays up to 24 h, expanding the potential of producing FGC in large-scale.
Further conclusions are drawn, as follows:
• FRC presented higher compressive strength than FRRAC, showing that the type of
aggregate is more important than content of fibre;
• N-concretes presented higher bond strength than R-concretes, for casting delay of
0.50 h, indicating the influence of type aggregate. Nevertheless, for longer casting
delays, the bond strength of LTZ decreased in N-concretes but increased in R-
concretes;
• The behaviour of bond strength with the increment in casting delay can be related to
the failure mode observed in the FGC specimens. While N-concretes presented a
transition of failure mode, R-concretes failed according to the same type of failure
mode;
• The long steel fibres used in this study did not significantly affect the bond strength
of LTZ in FGC, possibly due to the low contribution in microcracking control
provided by this type of fibre;
• Adequate bond quality is achieved in all FGC groups studied, regardless of type of
aggregate, content of fibre or casting delay.
Acknowledgements. The authors would like to acknowledge the Xi’an Jiaotong-Liverpool

University (XJTLU) Research Development Fund for the financial support received from the
project with reference RDF-16-02-42.
906 R. Chan et al.
References
1. Kawasaki, A., Watanabe, R.: Concept and P/M fabrication of functionally gradient materials.
Ceram. Int. 23, 73–83 (1997)
2. Zhang, N., et al.: Support performance of functionally graded concrete lining. Constr. Build.
Mater. 147, 35–47 (2017)
3. Mastali, M., Ghasemi Naghibdehi, M., Naghipour, M., Rabiee, S.M.: Experimental
assessment of functionally graded reinforced concrete (FGRC) slabs under drop weight
and projectile impacts. Constr. Build. Mater. 95, 296–311 (2015)
4. Wen, X., Tu, J., Gan, W.: Durability protection of the functionally graded structure concrete
in the splash zone. Constr. Build. Mater. 41, 246–251 (2013)
5. Rao, S., et al.: Composite Pavement Systems, Volume 2: PCC/PCC Composite Pavements,
vol. 2 (2013). https://books.google.com/books?id=o1zn2oqvQmEC&pgis=1
6. Hu, J., Fowler, D.W., Siddiqui, M.S., Whitney, D.: Feasibility study of two-lift concrete
paving: technical report (2014). https://texashistory.unt.edu/ark:/67531/metapth638590/
7. Beushausen, H., Alexander, M.G.: Bond strength development between concretes of
different ages. Mag. Concr. Res. 60, 65–74 (2008)
8. Qian, P., Xu, Q.: Experimental investigation on properties of interface between concrete
layers. Constr. Build. Mater. 174, 120–129 (2018)
9. Rashid, K., Ueda, T., Zhang, D., Miyaguchi, K., Nakai, H.: Experimental and analytical
investigations on the behavior of interface between concrete and polymer cement mortar
under hygrothermal conditions. Constr. Build. Mater. 94, 414–425 (2015)
10. Beushausen, H., Höhlig, B., Talotti, M.: The influence of substrate moisture preparation on
bond strength of concrete overlays and the microstructure of the OTZ. Cem. Concr. Res. 92,
84–91 (2017)
11. Bajaj, K., Shrivastava, Y., Dhoke, P.: Experimental study of functionally graded beam with
fly ash. J. Inst. Eng. Ser. A 94(4), 219–227 (2014)
12. Cao, Y., Li, P., Brouwers, H.J.H., Sluijsmans, M., Yu, Q.: Enhancing flexural performance
of ultra-high performance concrete by an optimized layered-structure concept. Compos.
Part B Eng. 171, 154–165 (2019)
13. Hussein, L., Amleh, L.: Structural behavior of ultra-high performance fiber reinforced
concrete-normal strength concrete or high strength concrete composite members. Constr.
Build. Mater. 93, 1105–1116 (2015)
14. BSI: BS EN 197-1:2011, Cement - Composition, specifications and conformity criteria for
common cements. BSI Standards Limited, London (2011). https://doi.org/10.3403/30205527
15. BSI: BS 812-103.1:1985, Testing aggregates - Method for determination of particle size
distribution - Sieve tests. BSI Standards Limited, London (1998). https://doi.org/10.3403/
00139627
16. Mehta, P.K., Monteiro, P.J.M.: Concrete: Microstructure, Properties, and Materials.
McGraw-Hill Education (2006). https://doi.org/10.1036/0071462899
17. Neville, A.M.: Properties of Concrete. Pearson Education Limited, Harlow (2011)
18. Chan, R., Liu, X., Galobardes, I.: Parametric study of functionally graded concretes
incorporating steel fibres and recycled aggregates. Constr. Build. Mater. 242, 118186 (2020)
19. Teychenné, D.C., Franklin, R.E., Erntroy, H.C.: Design of Normal Concrete Mixes.
Construction Research Communications Ltd., Watford (1997)
20. BSI: BS EN 12390-2:2009, Testing hardened concrete - Making and curing specimens for
strength tests. BSI Standards Limited, London (2009). https://doi.org/10.3403/30164903
21. BSI: BS EN 12390-3:2009, Testing hardened concrete - Compressive strength of test
specimens. BSI Standards Limited, London (2011). https://doi.org/10.3403/30164906
22. BSI: BS EN 12390-6:2009, Testing hardened concrete - Tensile splitting strength of test
specimens. BSI Standards Limited, London (2010). https://doi.org/10.3403/30200045
23. Sprinkel, M.M., Ozyildirim, C.: Evaluation of high performance concrete overlayers placed
on route 60 over lynnhaven inlet in virginia (2000).. http://www.virginiadot.org/vtrc/main/
online_reports/pdf/MicrosoftWord-VTRC01-R1_Sprinkel&Ozyildirim_pdf
24. Senaratne, S., Gerace, D., Mirza, O., Tam, V.W.Y., Kang, W.H.: The costs and benefits of
combining recycled aggregate with steel fibres as a sustainable, structural material. J. Clean.
Prod. 112, 2318–2327 (2016)
25. Carneiro, J.A., Lima, P.R.L., Leite, M.B., Toledo Filho, R.D.: Compressive stress–strain
behavior of steel fiber reinforced-recycled aggregate concrete. Cem. Concr. Compos. 46, 65–
72 (2014)
26. Chan, R., et al.: Analysis of potential use of fibre reinforced recycled aggregate concrete for
sustainable pavements. J. Clean. Prod. 218, 183–191 (2019)
27. Costa, H., Carmo, R.N.F., Júlio, E.: Influence of lightweight aggregates concrete on the bond
strength of concrete-to-concrete interfaces. Constr. Build. Mater. 180, 519–530 (2018)
28. Duarte Santos, P.M., Santos Júlio, E.N.B.: Factors affecting bond between new and old
concrete. ACI Mater. J. 108, 449–456 (2011)
29. Huang, H., Yuan, Y., Zhang, W., Gao, Z.: Bond behavior between lightweight aggregate
concrete and normal weight concrete based on splitting-tensile test. Constr. Build. Mater.
209, 306–314 (2019)
30. Bentur, A., Mindess, S.: Fibre Reinforced Cementitious Composites. Taylor & Francis,
London and New York (2007)
31. Torrents, J.M., et al.: Inductive method for assessing the amount and orientation of steel
fibers in concrete. Mater. Struct. 45, 1577–1592 (2012)
Towards Rebar Substitution by Fibres –
Tailored Supercritical Fibre Contents
Katharina Look1(&), Peter Heek2, and Peter Mark1

1
Ruhr University Bochum, Bochum, Germany
katharina.look@rub.de
2
HOCHTIEF Engineering GmbH, Essen, Germany
Abstract. Structural application of steel fibre reinforced concrete (SFRC)

progressively increases. To create load-bearing components without additional
steel reinforcement, mixtures with supercritical fibre contents tailored to the
structure and application must be devised. An innovative formwork concept is
developed that enables for fibre contents up to 1.0 Vol.-% the steered alignment
of fibres. Based on the observation that fibres orient parallel to formwork edges
it applies internal formworks. The efficiency is verified by measurements of
spatial fibre orientations and flexural tensile strengths. Results on single span
beams indicate that favourable fibre orientations perpendicular to cracks can be
achieved which is even more effectively with increasing fibre content. The
impact does not go along with enhancements of flexural tensile strength. To
investigate the effect of steered fibres on bearing capacities of spatial elements
like slabs, the SFRC’s composition is optimized with regard to fibres’ content
and orientation. The aim is to achieve two-dimensional fibre orientations in
directions of principle tensile stresses by artificially limiting the specimen’s
height, so that fibres align horizontally in concrete’s flow direction without
additional steering. For a supercritical fibre content of 1.8 Vol.-%, it is possible
to substitute conventional (mesh) reinforcements of approximate 6.7 cm2/m
(fyk = 500 N/mm2).
Keywords: Steel fibre reinforced concrete Fibre alignment Rebar

substitution Strain-hardening Formwork concept Steering of fibres
Prefabrication
1 Introduction
Gain of knowledge in the field of SFRC is opening up new areas of application for the
material. The fibres’ effect in e.g. fire [1], fatigue [2] or under creep [3] is subject of
current scientific investigations. Numerous standards (e.g. [4, 5]) and design tools (e.g.
[6–8]) for SFRC are available in the meantime. Although fibres are still mainly used in
non-load-bearing structures or in combination with conventional reinforcement. The
main issue is the ductile load-bearing behaviour, which is - taking into account also the
high scatter in the SFRC’s strength - hard to fulfil with common subcritical fibre
contents. As fibre contents with about 0.5 Vol.-% usually yield to strain-softening after
cracking and thus low calculative residual strength, refined concrete mixtures with high
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_80
Towards Rebar Substitution by Fibres 909
Fig. 1. Concept of substituting crosswise reinforcement with a prefabricated slab of SFRC.
strength fibres are subject of current research to enable reliable strain-hardening

behaviour and ductile tensile failure without additional rebars.
Thereby, effects of composition mixture [9], structural dimensions for optimum
fibre distributions [10], innovative casting methods and sequences [11] to establish
desired fibre orientations as well as scattering in the post-cracking strength [12] are in
the focus of interest. As the individual parameters are strongly interacting, structural
and material interrelationships are itemised subsequently.
Objective is the complete replacement of conventional (bar) reinforcement by
macro steel fibres while ensuring ductile structural behaviour. This should be achieved
by inserting as many fibres as feasible into the concrete for maximum performance. In
addition, the alignment of the fibres is steered in order to achieve the most effective
orientation. This is obtained differently for one- (beams) and two-dimensional struc-
tures (slabs, foundations). In the first case, additional interior formworks apply, as
fibres tend to orient parallel to formwork edges. In two-dimensional members, the
cross-section’s height is artificially limited within the casting sequence to steer fibre
orientations in direction of the fresh concrete’s flow. To assess the accuracy of the
innovative formwork and casting approach, experiments are executed that provide
information on achieved spatial fibre orientations. Resulting load-bearing capacities as
well as scattering of strength values in the post-cracking domain are verified with beam
tests. A tailored fibre mixture for substituting conventional reinforcement is derived
that also takes into account workability of fresh concrete with increasing fibre content.
Starting with beam elements, the findings are transferred in a second step to spatial
structures by optimizing the material and structural parameters. This becomes neces-
sary, as a forced alignment with interior formwork does not work here. To steer fibres’
alignment in the horizontal plane, the specimen’s height is limited. Applications are
seen in hybrid systems, where a prefabricated SFRC-slab with supercritical fibre
content replaces conventional mesh reinforcement, e.g. in foundations or wall-elements
(Fig. 1). The precasted SFRC-slab is transported to the project site and then the
structure is completed with cast-in-place concrete. In order to ensure the transmission
of shear forces in the joint between SFRC and normal concrete, the surface of the
SFRC-slab is roughened.
910 K. Look et al.
2 One-Directional Steering Of Fibres

2.1 General
For beams, which are one-dimensional structures, the fibres should be steered mainly
into the direction of the tensile stress trajectories resulting from bending. Therefore, an
internal formwork is designed consisting of metal segments dividing the beam form-
work into several horizontal sections. The influences of the mechanical alignment in
combination with different fibre quantities on the fibre orientation, workability and
load-bearing capacity are investigated here. For this purpose, typical fibre contents up
to 1.0 Vol.-% apply. Beams are tested in four-point-bending to determine residual
flexural tensile strengths. To investigate the efficiency of the internal formwork con-
cept, fibre orientation of the beams is measured subsequently.

The test series (M1) covers normal strength concrete of class C30/37 and conventional
fibre contents of 0.5 Vol.-% (M1-1) and 1.0 Vol.-% (M1-2). The utilized fibre type FF-
60/75–1450 (lf = 60 mm, df = 0.75 mm) has a tensile strength of 1450 MPa and a
straight shape with two anchor knots at each end (cf. Fig. 4). Due to this type of
anchorage a better orientation factor and less agglomerations should be achieved [12,
13]. The internal formwork concept consists of a wooden frame in which metal seg-
ments are fixed, so that the formwork of a beam (w h l = 150 150 700 mm3)
is divided into five 30 mm wide sections during casting and compaction (Fig. 2). Three
beams with (notation a) and without internal formwork (notation b) are casted for each
fibre quantity. After casting and compacting, the internal formwork is slowly removed
by hand to prevent delamination of the concrete’s matrix (see grips in Fig. 2). No
compaction energy is added afterwards. Before testing, the specimens are rotated by
90°. Discontinuities among the layers of FRC cannot be observed (Fig. 3).
Fig. 2. Practical realization of the internal Fig. 3. Concept of the internal formwork
formwork. for a beam.
After testing, the beams are cut into cubes. Fibre orientations in all spatial direc-
tions are then measured based on ferromagnetic induction with the measurement device
BSM 100 as described in [14]. The fibre orientation factor ηi for each spatial direction
(i = x, y, z) follows from the measured induction voltage Ui using Eq. (1) [14].
Ui
gi ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð1Þ
Ux2 þ Uy2 þ Uz2
Defining ηi as the ratio of the steel fibre length projected in the respective axis i to
the actual steel fibre length lf [15], the angle a between the horizontal plane (x-y-plane)
and the fibre can be determined with Eq. (2).
a ¼ arcsinðgz Þ ð2Þ
2.3 Results
2.3.1 Flexural Tensile Strengths
Beams are tested in four-point-bending according to the German DAfStb-Guideline
“Steel Fibre Reinforced Concrete” [4]. Figs. 4, 5, 6 and 7 show deflection-stress
diagrams for the individual beams (black lines, primary ordinate) as well as mean
values (black dashed lines, primary ordinate) and deflection-dependent coefficients of
variation (COV) as a measurement of the flexural strength’s scatter (grey lines, sec-
ondary ordinate). Mean values of residual flexural tensile strengths for deflections of
d = 0.5 mm (ffcflm,L1) and d = 3.5 mm (ffcflm,L2) are presented in Table 1. Mean values
of the coefficient of variation (COVm) are also given.
Deflection-hardening behaviour for all specimens with 1.0 Vol.-% of fibres is
achieved (Figs. 5 and 7). For beams with 0.5 Vol.-% (Figs. 4 and 6), the post-cracking
behaviour stabilizes after a stress loss to the pre-cracking initiation state. For test series
M1, the residual flexural tensile strengths can be increased up to 167% by doubling the
fibre content - regardless of additional internal formwork. No essential difference in the
residual flexural tensile strength can be determined with or without internal formwork,
i.e. with steered fibres no enhancement in performance is gained. However, it can be
observed that scattering decreases significantly by using the internal formwork. This
phenomenon is much more pronounced for the lower fibre content (M1-1) than for the
higher fibre content (M1-2). There, the COV for mechanically oriented fibres even
slightly increases. Nevertheless, it remains significantly below the value of M1-1b and
a value of 25%, which is typical for SFRC beam tests [16]. Considering only those
specimens without internal formwork (b), it is noticeable that the scattering in the post-
cracking behaviour decreases with increasing fibre content. This observation can be
traced back to the fact, that at very low intended fibre contents the number of fibres in
the cracked cross-section may even be zero, which exerts a much stronger impact on
the post-cracking strength than scattering fibre numbers at high intended contents. Due
to the high quantity of fibres, they interfere with each other in their alignment. This
effect can be used positively combined with the known correlations between casting
direction and fibre orientation. For example, together with the preferred orientation of
the fibre perpendicular to the casting direction [17], the fibre orientation in two-
dimensional components can be specifically steered and predicted under constant
fabrication conditions.
912 K. Look et al.
Fig. 4. Deflection-stress curves, mean value Fig. 5. Deflection-stress curves, mean value
and COV for M1-1a. and COV for M1-2a.
Fig. 6. Deflection-stress curves, mean value Fig. 7. Deflection-stress curves, mean value
and COV for M1-1b. and COV for M1-2b.
Table 1. Mean values of residual flexural tensile strengths for deflections of 0.5 mm (L1) and
3.5 mm (L2) and mean value of COV for M1.
Mixture f fcflm;L1 [MPa] f fcflm;L2 [MPa] COVm [%]
M1-1a 5.51 4.40 4.7
M1-1b 4.94 4.05 15.8
M1-2a 7.57 6.66 7.6
M1-2b 7.82 6.78 8.1
2.3.2 Fibre Orientation

By means of the aforementioned BSM 100 the percentages of orientation for the x-, y-
and z-direction are measured. Therefore, the tested SFRC beams of series M1 are cut
into five (A–E) cubes (Fig. 8). The results are shown in Fig. 8 for M1-1 and in Fig. 9
for M1-2. In both figures, the dashed lines represent orientation for conventionally
casted beams. Solid lines correspond to beams casted with an internal formwork.
Fig. 8. Percentages of orientation for M1-1. Fig. 9. Percentages of orientation for M1-2.
In SFRCs with high and low fibre con-

tents, the proportions of fibres oriented in the spatial directions do not significantly
differ. Lower percentages for the y-axis (casting direction) in cube B and cube D result
from the geometry of the internal formwork: A cross stud stabilizes the lamellas in the
middle of the beam (cf. Fig. 2), so that direct filling is not possible here and the fibres
cannot orient optimally perpendicular to the casting direction [17]. The fibres in the end
cubes are also predominantly aligned in the y-direction, as the fibres straighten up at the
edge during casting (“wall-effect” [18]), since orientation in z-direction is obstructed by
the internal formwork.
Table 2 lists the mean values of the orientation percentages in each spatial direction
with the corresponding COV in brackets as well as the orientation factor ηz (Eq. (1))
and the angle to the horizontal plane a (Eq. (2)). The mean percentage of fibres that do
not contribute to the load transfer, i.e. those parallel to cracks (z-direction), can be
reduced by 65% compared to conventionally casted beams. Since the orientation is
strongly limited in the height of the beam, the factors scatter in a small range. However,
the scatter for the high fibre content is lower than for the lower fibre content if there is
no internal formwork used - one more indication for correlations between fibre ori-
entation and high fibre contents. In addition, the angle a - limited to 30° by the internal
formwork - is observed with a mean angle of 9° (M1-1a) or 11° (M1-2a), respectively.
This underlines the very good efficiency of the developed formwork concept.
Table 2. Mean values of fibre orientations, orientation factor for z-direction and fibre angle to
the horizontal plane for M1 (COV [%]).
Mixture x [%] y [%] z [%] ηz [-] a [°]
M1-1a 60.6 (7.2) 28.3 (15.9) 11.1 (7.5) 0.16 9°
M1-1b 49.1 (11.6) 18.8 (20.1) 32.1 (8.1) 0.52 31°
M1-2a 61.9 (7.0) 25.0 (15.6) 13.1 (2.9) 0.19 11°
M1-2b 45.6 (9.6) 17.5 (16.0) 36.9 (5.8) 0.60 37°
914 K. Look et al.
Comparing the results of the flexural tensile tests from the previous section with the
measured fibre orientations, the lower scatter for series M1-1b in contrast to M1-2b can
be traced back to a lower percentage of aligned fibres in the z-direction. This confirms
the assumption that larger contents of fibres in fresh concrete impair their free orien-
tation. Most fibres straighten nearly oriented in the direction of principle tensile stress
trajectories which results in lower scattering of flexural tensile strengths, but not in
higher load-bearing capacities (cf. Table 1).
3 Two-Directional Steering of Fibres
3.1 General
The presented findings for the steering of fibres are now transferred to spatial structures.
Two objectives are pursued: First, the performance class is maximized by adding as
many fibres as possible to the concrete. Second, fibre orientations are optimized in term
of two-dimensional alignment in directions of principle tensile stresses. Criteria to
which the concrete must comply with regard to the aspired performance are among
other aspects a very good workability and a homogeneous fibre distribution. To achieve
these requirements, a small maximum grain size of 8 mm, a fine grading curve and a
high amount of cement and fly ash are used. For verification of performance classes,
beams are casted without internal formwork and tested according to Sect. 2. To
investigate the fibre orientation in two-dimensional structures, slabs are casted.
Thereby, the height of the slab is limited to 10 cm in order to minimize the fibre angle
to the horizontal plane [18]. As usual, the width and depth are greater than its height
[19]. For verification of fibre orientations, the slabs are cut into cubes and measured by
means of the aforementioned BSM 100 as well.

In the second series (M2), the fibre content is maximized. Beginning from an initial
content of 1.0 Vol.-% (cf. M1), 0.06 Vol.-% are successively added resulting in a final
content of approx. 1.8 Vol-%. The dosage of the superplasticizer is adjusted after each
fibre addition. The fibre type is also adapted to maximize residual tensile strengths. For
this purpose, the high-performance fibre Dramix 5D-65/60BG (lf = 60 mm, df = 0.9
mm) with a tensile strength of 2300 MPa and a triple deformed hooked end (cf.
Fig. 10) serves. Although straight fibres are recommended to obtain a more pro-
nounced orientation factor, this series focuses on reachable performance classes. In
order to account for the high tensile strength of the fibres -which require sufficient bond
strength with the concrete’s matrix to activate full bearing capacities- a high strength
concrete of strength class C60/75 is applied. Six beams are tested for validation of
performance. Additionally, a slab with dimensions of 50 50 10 cm3 (cf. Fig. 12)
is casted to evaluate the desired two-dimensional fibre orientation for the intended
design of a prefabricated SFRC-slab.
3.3 Results
3.3.1 Flexural Tensile Strengths
The beams of the second series (M2) are also tested in four-point-bending. Corre-
sponding deflection-stress curves are presented in Fig. 10, which show a deflection-
hardening behaviour with the exception of beam 2 with a load-bearing capacity many
times lower compared to the other specimens. After cutting, some irregularities in the
form of matrix imperfections are detected. The beam can bear less force because fibres
are not sufficiently anchored in the concrete and therefore cannot transmit forces.
Additionally, the beam could be eliminated as an outliner with the Grubbs test [20].
However, mean values and COV with and without beam 2 are highlighted in Fig. 10.
A load-bearing capacity with maximum flexural tensile strength of about 9.0 MPa is
reached.
Fig. 10. Deflection-stress curves, mean value and COV (with and without beam 2) for M2-1.
Table 3 presents the mean values of residual flexural tensile strengths for deflec-
tions of d = 0.5 mm (ffcflm,L1) and d = 3.5 mm (ffcflm,L2) and COVs for M2 excluding
beam 2. In contrast to the conventionally casted beams of series M1-2b, scattering is
greater for series M2, although the fibre content is maximized. The reason behind may
be the applied fibre type as the hooked ends are attributed to exert a dual impact. On the
one hand, they improve load-bearing capacities due to their plastic deformation
capacities. On the other hand, they promote adverse fibre distributions and orientations
in fresh concrete due to fibre agglomerations [13].
It is also important to point out that the load-bearing capacity for SFRC with
1.8 Vol.-% of the second fibre type (M2) equals the capacity for SFRC with 1.0 Vol.-
% of the first fibre type (M1–2). Since fibre sedimentation can be excluded, stagnation
of performance is mainly seen as a result of maximum interlocking forces reached in
the fibre-matrix zone. However, other effects, e.g. due to aggregate composition or
concrete strength, have to be part of further investigations.
916 K. Look et al.
Table 3. Mean values of residual flexural tensile strengths for deflections of 0.5 mm (L1) and
3.5 mm (L2) and mean value of COV for M2.
Mixture f fcflm;L1 [MPa] f fcflm;L2 [MPa] COVm [%]
M2 7.98 6.76 12.7
3.3.2 Fibre Orientation

The moulds of the slabs are casted from the centre to the edges while continuous
external vibrations prevail. The aim is to keep the flow direction of the concrete
constant in this direction (tangential). Thereby, the fibres are aligned perpendicular to
the flow direction [21]. In beams, fibres tend to align parallel to the flow direction of
fresh concrete as described by the so called “tunnel-effect” [10]. In contrast, slabs allow
for spatial dispersal of fresh concrete and thus fibres predominately align radially in a
circle around the filling location and perpendicular to the casting direction, i.e. in the
horizontal x-y-plane [21]. The slab is cut into 16 cubes, each with an edge length of
10 cm (Fig. 11). Fig. 11 shows the scheme of dividing the slab and the notation of the
cubes. To eliminate areas with “wall-effect” (cf. Sect. 2.3.2) [18], the first 5 cm on each
side of the specimen are cut off (Fig. 11). Related values for the individual cubes are
shown in Fig. 12.
Mean values of fibre orientation for
Fig. 11. Measuring concept of the slab (di- Fig. 12. Percentages of orientation for M2.
mensions in [cm]).
the whole slab, COV, orientation factor

for the z-direction and calculated fibre
angle to the horizontal plane are presented in Table 4. Mean values of orientation
percentages of the slab are only slightly higher than those of the beams with an internal
formwork concept of M1, although no internal formwork was used here. By reducing
the specimen’s height and applying a very high fibre content, it is possible to achieve a
mean fibre angle a of 15° to the horizontal plane. Nevertheless, it should be noted that
the measured orientation factors for all directions scatter more than those of the beams
do as no additional internal formwork is used here. If the fibre orientation value for the
z-direction (16.2%) is subtracted from the possible 100%, 41.9% of the fibres should be
aligned in both, the x- and y-direction respectively for an ideal two-dimensional fibre
distribution. With an absolute deviation of ±3.2%, an equally, two-dimensional ori-
entation of the fibres is given here.
Table 4. Mean values of fibre orientations, orientation factor for z-direction and fibre angle to
the horizontal plane for M2 (COV [%]).
Mixture x [%] y [%] z [%] ηz [-] a [°]
M2 45.1 (20.3) 38.7 (22.4) 16.2 (19.2) 0.26 15°
4 Practical Application
In order to make reliable predictions on load-bearing capacities, the scatter in post-

cracking behaviour has to be minimized. Since the idea of substituting conventional
reinforcement by a SFRC layer with a bidirectional fibre orientation for two-
dimensional load-bearing behaviour, multiple statically indeterminate systems are
particularly appropriate. Steel fibres transfer tensile forces over crack edges. Plastic
hinges or yield lines arise, which allow the redistributions of stresses. Loads can be
further increased after cracking and load-bearing capacity reserves are activated. In
contrast to statically determined systems, like beams in flexural tensile tests, statically
indeterminate structures with a large cross-sectional area under tension scatter less [22].
Standardization over the whole production process reduces scattering in fibre orien-
tation and load-bearing capacities. Tensile strengths and finally the design can be
statistically predicted and used for dimensioning. Prefabrication with lean construction
methods is a suitable option for continuous quality control and compliance of the
components.
Based on the results derived in the experiments, the replacement of a conventional
(mesh) reinforcement by a prefabricated SFRC-slab is presented. With a maximum
flexural tensile strength of 9.0 MPa (cf. Sect. 2 and 3) and a conservative conversion
factor of 0.37 [4], a calculated tensile stress of 3.33 MPa results. Smearing this stress
over a cross-section with a supposed slab height of 10 cm the slab has a resistance of
333 kN/m in both directions. Assuming a standard reinforcement B500 (fyk = 500
MPa), reinforcement of 6.7 cm2/m is required. This corresponds to a mesh reinforce-
ment of about 2 Q 335 (each with 3.35 cm2/m in both spatial directions), which are
possible to substitute with the presented SFRC mixture. This amount of reinforcement
is used in typical foundations of building construction.
5 Conclusions
In this contribution, an innovative casting concept in terms of location of formwork and

casting sequence is developed to optimize fibre orientations in beams and slabs. The
main conclusions are:
918 K. Look et al.
• For beams, additional interior formworks are suitable since fibres tend to orient
perpendicular to formwork edges. The developed internal formwork concept can
also be adapted to other load distributions, such as the alignment to the inclined
tensile struts under shear load.
• In case of slabs, two-dimensionally alignment of fibres can be achieved by artifi-
cially limiting the cross-sections height. Fibre distributions can be steered by means
of the fresh concrete’s flow direction.
• Residual flexural tensile strengths do not increase by the systematic alignment of
steel fibres for one-dimensional structures. Nevertheless, a mechanical alignment
for low fibre contents can considerably reduce scattering.
• In spatial structures with aligned fibres, a critical fibre content seems to limit
maximum flexural tensile strengths as the further increase of the fibre content do not
significantly enhance load-bearing capacities. The critical fibre content depends on
several conditions and is here for hooked end and macro fibres of high strength steel
about 1.8 Vol.-%. The here achieved maximum post-cracking flexural strength of
9.0 MPa exceeds calculative fibre strength for design purposes according to valid
standards like [4] considerably.
• Tests to determine residual tensile strengths should represent the structural condi-
tions of the actual application. As standard-beam tests just represent one-directional
fibre orientations, uniaxial tensile tests on specimens cut from the component (thin
SFRC-slab) provide the most accurate results of the residual tensile strength and
should therefore be performed.
Acknowledgements. The authors would like to thank BASF Construction Solutions GmbH,
BauMineral GmbH, Bekaert GmbH, Dyckerhoff GmbH, Feel Fiber GmbH and StraTec GmbH
for their friendly provision of the test materials.
References
1. Heek, P., Tkocz, J., Mark, P.: A thermo-mechanical model for SFRC beams or slabs at
elevated temperatures. Mater. Struct. 51, 87 (2018)
2. Heek, P., Ahrens, M.A., Mark, P.: Incremental-iterative model for time-variant analysis of
SFRC subjected to flexural fatigue. Mater. Struct. 50, 62 (2017)
3. Plizzari, G., Serna, P.: Structural effects of FRC creep. Mater. Struct. 51, 167 (2018)
4. DAfStb: DAfStb-Guideline Steel Fibre Reinforced Concrete, Beuth (2015)
5. Fédération Internationale du Béton (fib): Model Code 2010, Ernst & Sohn (2013)
6. Gödde, L., Strack, M., Mark, P.: M-N-Interaktionsdiagramme für stahlfaserverstärkte
Stahlbetonquerschnitte – Anwendung am Beispiel von Tübbingen. Beton- und Stahlbeton-
bau 105(5), 318–323 (2010a)
7. Gödde, L., Strack, M., Mark, P.: ‘Bauteile aus Stahlfaserbeton und stahlfaserverstärktem
Stahlbeton – Hilfsmittel für Bemessung und Verformungsabschätzung nach DAfStb-
Richtlinie Stahlfaserbeton’. Beton- und Stahlbetonbau 105(2), 78–91 (2010b)
8. Look, K., Heek, P., Mark, P.: ‘Stahlfaserbetonbauteile praxisgerecht berechnen, bemessen
und optimieren‘. Beton- und Stahlbetonbau 114(5), 296–306 (2019)
9. Marković, I.: High-Performance hybrid-fibre concrete – development and utilisation.
Dissertation, TU Delft (2006)
10. Stähli, P., Custer, R., van Mier, J.G.M.: On flow properties, fibre distribution, fibre
orientation and flexural behaviour of FRC. Mater. Struct. 41(1), 189–196 (2008)
11. Hadl, P., Tue, N.V.: Einfluss der Faserzugabe auf die Streuung im Zugtragverhalten von
Stahlfaserbeton. Beton- und Stahlbetonbau 111(5), 310–318 (2016)
12. Hadl, P., Gröger, J., Tue, N.V.: ‘Experimentelle Untersuchungen zur Streuung im
Zugtragverhalten von Stahlfaserbeton‘. Bautechnik 92(6), 385–393 (2015)
13. Huß, M.: ‘Neuer Stahlfasertyp eröffnet vielfältige Möglichkeiten für die Betonindustrie‘.
BWI BetonWerk Int. 5, 72–76 (2018)
14. Wichtmann, H.-J., Holst, A., Budelmann, H.: ‘Ein praxisgerechtes Messverfahren zur
Bestimmung der Fasermenge und -orientierung im Stahlfaserbeton‘. Beton- und Stahlbe-
tonbau 108(12), 822–834 (2013)
15. Lin, Y.-z.: Tragverhalten von Stahlfaserbeton, Dissertation, Universität Kaiserslautern
(1999)
16. Heek, P., Lingemann, J., Mark, P., Schnütgen, B., Schulz, M., Zilch, K.: Sicherheitskonzept
der DAfStb-Richtlinie Stahlfaserbeton, Deutscher Ausschuss für Stahlbeton (DAfStb) Heft
614 Erläuterungen zur DAfStb-Richtlinie Stahlfaserbeton, pp. 62–69. Teil B: Allgemeine
Erläuterungen zur Richtlinie Stahlfaserbeton (Autorenbeiträge), Beuth (2015)
17. Empelmann, M., Teutsch, M.: Faserorientierung und Leistungsfähigkeit von Stahlfaser-
sowie Kunststofffaserbeton. Beton 59(6), 254–259 (2009)
18. Soroushian, P., Lee, C.-D.: Distribution and orientation of fibers in steel fiber reinforced
concrete. Mater. J. 87(5), 433–439 (1990)
19. Abrishambaf, A., Barros, J., Cunha, V.: Relation between fibre distribution and post-
cracking behaviour in steel fibre reinforced self-compacting concrete panels. Cem. Concr.
Res. 51, 57–66 (2013)
20. Grubbs, F.E.: Sample criteria for testing outlying observations. Ann. Math. Stat. 21(1), 27–
58 (1950)
21. Barnett, S.J., Lataste, J.-F., Parry, T., Millard, S.G., Soutsos, M.N.: Assessment of fibre
orientation in ultra high performance fibre reinforced concrete and its effects on flexural
strength. Mater. Struct. 43(7), 1009–1023 (2010)
A Constitutive Model for Steel-Fibre-
Reinforced Lightweight Concrete

a.abbas@uel.ac.uk
Abstract. A method is proposed to derive and validate material properties for

lightweight fibrous concrete using experimental and numerical data. The coarse
lightweight material tested (produced by LYTAG) is recycled and offers an
alternative to gravel and quarry resources which are at risk of depletion in future.
Also, this material can lead to reduction in the mass of the structure which
results in economical designs. However, in comparison to normal weight
aggregate concrete (NWAC), lightweight aggregate concrete (LWAC) tends to
be more brittle as it typically shears through the aggregates leading to instan-
taneous drop in peak load in both compression and tension tests. Hence, to
address this brittleness for LWAC, modern hooked-end DRAMIX fibres with
different geometry (hooks), dosages (Vf) and bond strengths (sb) are added to
mixes with different strengths (fck). This paper focuses on both tensile properties
using a direct pullout and indirect notched beam tests, and compressive prop-
erties (fck, E, l) using the conventional compression test for cylinders of plain
and steel-fibre reinforced lightweight concrete (SFRLC). A tensile semi-
empirical multilinear r-x relation was derived besides compressive r-e and
validated against available steel fibre reinforced concrete (SFRC) constitutive
models for the tested fibrous notched beams using ABAQUS.
Keywords: Recycled lightweight concrete Hooked-end fibres Tension

Compression Ductility Pullout Notched beam NLFEA ABAQUS r-x
1 Introduction
The use of structural lightweight aggregate concrete brings several advantages as

compared to the conventional normal weight concrete such as thermal insulation and
fire resistance. Besides, the lightweight aggregate Lytag used in this work is recycled
and offers reduction in CO2 emissions as well as being an alternative to the depleting
gravel and quarry resources (Gerritse 1981). The coarse aggregate Lytag is made from
fly ash which is a by-product of coal-fired power stations, by the process of palleti-
sation. The UK generates about 10% of its electricity from coal. Structural Lytag has
been around since the 1960s, available in the UK and Europe and can be used to
produce concrete strengths of up to 60 N/mm2. Also, the improved strength-to-weight
ratio of LWAC results in smaller cross sections which in turn leads to a decrease in area
of reinforcement and savings in material transport costs due to lower inertial and
gravity loads. The usage of lightweight concrete can therefore be ideal and competitive
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_81
A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete 921
in industry for the growing need for taller and longer span structures, especially in
seismic and dynamic zones (Libre et al. 2011; Campione 2014; Dias-Da-Costa 2014;
Mo et al. 2017). Structural applications of LWAC include bridges, towers and slabs,
with notable modern structures such as the Acton Swing bridge, the Gherkin, the Shard
and the 103 Colmore Row building. A study on Lytag in 2014 (Lytag 2014) showed
that lightweight concrete can bring about 34% savings in CO2 as well as a reduction of
up to 48% of concrete and reinforcement when compared to conventional gravel
concrete allowing a reduction in foundation sizes and increase in spans and building
space. These advantages however come as a trade-off for the increased brittleness of
lightweight concrete. concrete usually being more porous and having a poor aggregate
interlock mechanism as aggregates tend to be as weak as cement, which translates into
lacking a natural toughening mechanism post-crack such as in the case for normal
weight concrete. This causes lower tensile to compressive strength ratio which ulti-
mately lowers shear resistance in structures such as beams and slabs and causes brittle
failure. Also, the porous nature of lightweight concrete leads to a lower modulus of
elasticity (Chen et al. 2010; Badogiannis and Kotsovos 2013) which causes excessive
deflections and cracking (Lim et al. 2006; Wu et al. 2011). In addition, another dis-
advantage highlighted by Lim et al. (2006) is that lightweight aggregate concrete usage
is reduced when compared to normal weight concrete due to a combination of lack of
confidence and guidance for designers (although some exist, they are adapted from
research in the past century on normal weight concrete), and especially the lack of
understanding regarding how this increasingly brittle material behaves on the meso-
and macro scales. The brittle nature of lightweight concrete can be addressed by
incorporating traditional reinforcement. However, the latter solution can become
inherently counterproductive and impractical when reduction in structural elements is
sought by employing lightweight concrete especially at critical zones such as joints.
Therefore, fibre reinforcement which has long proven its effectiveness in controlling
and bridging tensile and shear cracks in the past for both lightweight and normal weight
concretes can become an adequate solution (Gao et al. 1997; Campione and La
Mendola 2004; Abbas et al. 2014a; Di Prisco et al. 2013; Grabois et al. 2016; Mo et al.
2017). For over 40 years, the usage of steel fibres in concrete mixes has been exper-
imented with and used (Ritchie and Kayali 1975), however, comprehensive studies on
fibrous lightweight concrete (especially on the material level which is key in under-
standing concrete behaviour) are still scarce with most work being merely theoretically
linked to SFRC, carried on the structural level only and involve other lightweight
aggregates types uncommon in the UK such as pumice stone and oil palm aggregates
(Swamy et al. 1993; Kang and Kim 2010; Di Prisco et al. 2013; Iqbal et al. 2015;
Grabois et al. 2016; Mo et al. 2017). It should be noted that, at present, there is no
international standards specific for steel fibre reinforced lightweight concrete (SFRLC)
with current guidelines being usually adapted from fibrous normal weight concrete
(SFRC). Hence, it is of environmental, economic and engineering design benefit and
need to study the behaviour of steel fibre reinforced recycled lightweight aggregate
concrete within an adaptable methodology to different types of fibres and aggregates
and derive a generic material and structural relationships capable of guiding designers
to become more confident in using SFRLC for different structural elements, thus
reducing its underutilization and paving the path for future researchers. This work
suggests a methodology to test and derive the complete material constitutive relation in
compression (r-e) including Young’s modulus of Elasticity (E) and Poisson’s ratio (l),
and in tension (r-x) for SFRLC. The compression properties are based on the con-
ventional uniaxial compression test while the tension properties are derived based on
the pullout tests. The parameters used in this work include the type of fibres (3D and
5D), fibre dosage (Vf), interfacial bond strengths (sb) and concrete strengths (flcm).
Fibre embedded length (LE), and fibres with hooks being cut off are also parameters in
pullout tests. Nonlinear finite element analysis using ABAQUS (Habbit et al., 2000) is
employed and discussed to validate notched beams tested according to RILEM TC
162-TDF (2002). The proposed constitutive model derived is also compared to material
models based on RILEM TC 162-TDF (2003), Barros et al. (2005) and fib Model Code
2010 (FIB 2013).
2 Methodology
The methodology of the work done consists of two main parts. The first involves
experimental testing which includes the study of hardened properties of plain Lytag and
fibrous concrete. The focus on this paper will be on compressive tests of cylinders, and
tensile tests on pullout prisms and notched beams. The second part covers nonlinear
numerical testing and finite element analyses of notched beams using ABAQUS which
includes validating the material models derived and comparison with available con-
stitutive SFRC models. Studying the flexural and shear behaviour of reinforced con-
crete beams is shown in an accompanying paper.

As previously noted, the experimental programme includes a variety of specimens. For
uniaxial compression tests, cylinders are tested to study the effect of fibres on short
columns. For tensile tests, direct pullout prisms are tested to evaluate the direct tensile
behaviour of plain and fibrous lightweight concrete while indirect tensile tests on
notched beams are carried out to study the material flexural and tensile behaviour of
SFRLC. For all the specimens tested, fck = 30 MPa being the minimum strength for
structural concrete and hooked-end 3D fibres being the most commonly used in
industry (Abdallah et al. 2018a), are considered to be the control parameters. Other
specimens include different compressive strengths: fck = 35 MPa and fck = 40 MPa,
fibre type 5D and fibre dosages of 0%, 1% and 2%. The choice to skip lower dosages
such as Vf = 0.5% is so since most common structures such as tunnels and pavements
require at least a dosage of Vf = 1% for efficient crack control and thickness reduction
(The concrete Society 2007).
3.1.1 Materials
absorption was estimated to be around 15% per mass of LYTAG. Natural river sand
with a 4.75 mm maximum size was used as the fine aggregate of the concrete. The sand
had a water absorption coefficient of 0.09% and specific gravity of 2.65 complying with
BS EN 12620. The properties of the fibres used are summarized in the table below. It
should be noted that to prevent the possibility of balling, fibres were collated from the
manufacturer.

Fibre Type ru (MPa) lf (mm) df (mm)
3D 65/60 1160 60 0.9
4D 65/60 1500 60 0.9
5D 65/60 2300 60 0.9
3.1.2 Mix Design

The mix design used is summarized in Table 1 below. The mix design of the Lytag
summarised below. These were directly adopted from Lytag (2011) manuals (Table 2).

LC30/33 370 592 668.8 175
LC35/38 420 546 668.8 175
LC40/44 480 485 668.8 175
confirmed by Lytag manual. For this reason, Lytag aggregates were added 24 h after
mixing to obtain a saturated surface dry (SSD) state. The mixing process is shown in
Fig. 1 below.

As mentioned previously, the experiments include a uniaxial tensile pullout, com-
pression cylinder and flexural notched beam tests. A direct uniaxial tensile fibre pull-
out test shown in Fig. 2 was designed and used to investigate the influence of the
embedded fibre onto the tensile behaviour of the concrete. Unlike Robins et al. (2002)
pullout test, this test was designed with the fibre being completely embedded in Lytag
concrete mimicking to a large extent the real behavior of concrete at the crack in a
structure. To ensure breakage of the specimen in tension and to introduce the embedded
length LE as a variable at the monitored middle section, this section of the concrete was
reduced to a diameter corresponding to the area a single DRAMIX fibre of diameter
0.9 mm is predicted to occupy in the concrete (the diameter = 12 mm). This area was
determined based on numerical and statistical models (Krenchel et al. 1975; Romualdi
and Mendel 1964; Soroushian and Lee 1990). Hence, a single fibre pullout test is
equivalent to Vf = 1% while embedding 2 fibres is equivalent to Vf = 2%. The fibre
was placed at the notch of the pullout mould for the chosen LE using fixed suspended
cables mechanically attached to the fibre during casting. These were removed once
concrete was cast to avoid any interference with the results. Thus, this test can be
regarded as a truer and more realistic representation of a crack being bridged by a fibre
than the classical pullout test with the fibre embedded on one side while the other end
of the fibre is clamped and pulled by a tensile machine (Abdallah et al. 2018a). While
one end of the tensile machine is fixed, the other was gradually pulled in tension at a
displacement-controlled loading of 1 mm/min (Fig. 2). The tensile machine was cali-
brated and fitted with a sensitive displacement transducer capable of accurately mea-
suring the slip once crack was initiated. Both concrete blocks embedding the two hooks
of the fibre were assumed to be rigid. This assumption was proven to be correct. The
pullout specimen was designed to only deform between the two gripped carbon steel
bars as shown in Fig. 3.
Fig. 2. A pullout specimen during the uniaxial tensile test

Fig. 3. Pullout mould and dimensions of the specimen
The dimensions of the beam (150 mm, 150 mm, 500 mm) and the notch were
identical to RILEM TC 162 TDF (2002). The LVDTs were glued using high strength
epoxy after the concrete surface in touch with the LVDT was ground in the mid-section
at the front of the beam to enable the LVDTs to fully adhere onto the concrete. The
LVDT’s were connected to a computer software. For the purpose of estimating the
vertical deflection accurately, a steel bar inspired by a technique similar to JSCE-SF4
recommendations was made. The beams were placed onto two frictionless steel sup-
ports. This was deemed adequate as the loading was symmetrical (Fig. 4). A classic
three-point displacement-controlled loading of 0.2 mm/min was adopted using the
hydraulic machine which had a load capacity of 500kN. It should be noted that the
notch was introduced using an incompressible plastic material with width 4 mm (depth
of the notch), length 150 mm and height 25 mm. The plastic material was pre-coated
with a spray to prevent it from binding to concrete in order to remove it before testing.
Fig. 4. Notched beam during testing
Cylinders with 200 mm depth and 100 mm diameter are tested in compression to
generate the complete compressive stress-strain behaviour for lightweight plain and
fibrous concrete. A calibrated compressometer-extensometer steel ring designed
according to ASTM C 469 and fitted with LVDT’s is clamped onto the concrete
cylinders as shown in Fig. 5. A loading rate of 1.2 mm/min was deemed adequate to
study the post-peak behaviour of fibrous concrete in compression.
Fig. 5. Uniaxial compression cylinder test
4 Numerical Study
Three-dimensional nonlinear finite element analyses will be carried out using ABA-
QUS (Zienkiewicz and Taylor 2005). This software has shown to be successful at
predicting the behaviour of SFRC in tension, compression, flexure and shear as well as
the cracking pattern and mode of failure of plain and reinforced elements to a good
level of accuracy (Tlemat et al. 2006; Syed Mohsin 2012; Abbas et al. 2014a, 2014b;
Behinaein et al. 2018). The approach adopted in this work will involve modelling both
plain and fibrous concrete. The fibres will not be modelled discretely, instead they will
be introduced directly into the constitutive tensile and compressive models of the
concrete in ABAQUS. This methodology was opted for since modelling fibres dis-
cretely can be time consuming and will produce a nonflexible FE-based model difficult
to be adopted by designer engineers. Moreover, although modelling fibres discretely
using probabilistic techniques such as Monte Carlo is aimed to account for the random
distribution of fibres (Cunha et al. 2010), its usage can itself be unrealistic as the
prediction might as well be likely to completely miss the actual distribution and
location of fibres in a particular structural element. Hence, modelling fibres as part of
the concrete matrix as shown in a number of design guidelines can offer an easier and
perhaps safer prediction of the behaviour of composite material (Lok and Xiao 1999;
RILEM TC 162-TDF 2002; Barros et al. (2005); Di Prisco et al. 2013). However,
unlike the discrete 3D modelling, it should be noted that the homogenous fibrous
concrete modelling is not aimed to detect local failures on the mesoscale level explicitly
such as fibre rupture and concrete fracturing at the IFZ (Zhang et al. 2018).
Generally, there are two approaches that can be used in FEM to predict the tension
stiffening behaviour of fibrous concrete: the discrete crack approach (r-x) (Ngo and
Scordelis 1967) or smeared crack approach (r-e) (Rashid 1968). Although more
accurate at post-crack, the discrete crack approach can be impractical and numerically
intensive to use (Tlemat et al. 2006), while the more accepted smeared crack approach
that assumes the crack is smeared over an element can be more useful for design. In this
work, a smeared crack approach with (r-x) in tension is adopted using ABAQUS
option to define cracking displacement rather than cracking strain. ABAQUS derives
the strains based on the characteristic length lc which is defined as the mesh size for hex
elements using e = x/lc. The values of crack width were input into ABAQUS directly
from the pullout tests using the constitutive model depicted in Fig. 9. As compared to
the strain, the r-x approach offers a number of advantages since it represents the actual
behaviour of the fibrous material, is member size-effect independent and can be directly
applied to FEM (ABAQUS) from the pullout tests used (De Montaignac et al. 2011).
In ABAQUS, mesh sensitivity is not an issue for r-x relation as compared to r-e.
Also, r-x relationship can provide the necessary information needed to design for
service limit state including fatigue and shrinkage. In compression, the direct results
from the compression r-e tests were adopted using Eq. 1, 2 and 3.
Concrete damaged plasticity (CDP) was calibrated and chosen to model the notched
beam specimens to check the validity of the r-x model. The calibration of CDP on the
material level for both cylinder compression test and pullout test can be found with
more detail elsewhere (Al-Naimi and Abbas 2019). Table 3 below summarises the
CDP parameters used.
Table 3. Parameters usage for CDP

Dilation angle Eccentricity fb0/fc0 K Viscosity
25 0.1 1.16 0.666 0
The explicit solver is adopted. The explicit dynamic analysis was found to be a
more computationally efficient tool at solving the problems used in this work as
compared with the implicit solver which had a tendency to terminate. To ensure a
quasi-static solution the ratio between kinetic energy and internal energy was kept
below 1% throughout the analysis. The analysis was ran using a displacement loading
rate of 0.1 mm/step in a smoothed step with a mass scaling of 50.
The finite element model investigated to validate the notched beams is shown
below. Due to the symmetrical setting only half the beam was modelled with a sym-
metry boundary condition along the Z direction (Fig. 6). The beam was restricted from
moving in the Y direction by applying a displacement rotation boundary condition (in
the initial step) along the middle line of both supports. The displacement-induced load
was applied (in the explicit dynamic step) on the surface of the loading steel plate in a
similar manner to the experiment. To estimate the load, the reaction force was calcu-
lated by summing up the load along the boundary condition line on the supports then
multiplied by 2. The displacement, however, was calculated by taking the average Y
displacement of the surface or line of the structure in middle point of the span of the
beam. It should be noted that hexagonal brick element C3D8R with reduced integration
and hourglass control was used with a size of 20 mm with the exception of notch where
the size was 4 mm by 20 mm. The supports and load blocks were made to be rigid and
a tie coupling to the concrete was applied. The meshing, element size, solver, loading
rate and mass scaling adopted were based on a comprehensive convergence study.
Fig. 6. Finite element model used for the notched beam and boundary conditions (ABAQUS)

5.1.1 Compression
The average compressive stress-strain relationship of cylinders from 6 mixes is shown
in Fig. 7 below.
Fig. 7. Average compressive stress-strain of cylinders from 6 mixes
The figure shows that once the plain cylinders i.e. Vf = 0% with different strengths
reach their compressive peak strength, an instantaneous shear failure takes place.
Following visual inspection of the cylinders, it was evident that the Lytag aggregates
were sheared through which explains the lack of the strain toughening aggregate
interlock mechanism in the lightweight specimens and the occurrence of the brittle
failure. When 3D fibre dosages of 1% and 2% are added to the 30 MPa mix, the
compressive strength remains unaffected, however, it is clear that the failure becomes
ductile with the specimens having Vf = 2% developing a higher compressive ductility
due to the increased lateral confinement. Nonetheless, the specimens reinforced with
Vf = 1% of 5D fibres develop approximately 10% higher peak compressive strength
than that of Vf = 0% although with decreased ductility as compared with the fibrous
specimens having lower compressive strength and similar fibre volume fraction. This
can be explained by the stronger mix naturally being more brittle. It can be concluded
therefore that the more extensive hooks of the 5D fibres render it more difficult for the
fibres to pull out as compared to the 3D fibres causing strength increase prior to peak.
Table 4 below summarises the results of the compression test. It should be noted
that the modulus of elasticity ranged between 19.6 to 22.9GPa depending on the
concrete strength with fibres having generally lower values due to creating air voids in
the lattice. Tests of lightweight concrete in compression with similar fibres agree with
these findings (Li et al. 2018). Poisson’s ratio ranged between 0.16 to 0.19 (which
agrees with Lambert (1982)) with no pattern relating to neither flcm, Vf or fibre type.
Also, the strain for fibrous lightweight concrete at 85% peak varies with the lower
strength specimens developing larger strains. For design calculations, the strain at 85%
should be taken as shown in Eq. (3).
Table 4. Summary of findings from the compression tests

Vf (%) flcm (MPa) E (GPa) l Normal Strain ‰ Ductility (@85% fcm)
eel epeak eult
0 31.84 20.7 0.16 0.51 2.13 2.13 0
1 32.20 19.6 0.16 0.49 2.33 4.43 0.9
2 32.9 19.9 0.16 0.64 3.36 12.1 2.7
0 35.8 21.0 0.16 0.67 2.26 2.26 0
0 41.75 22.9 0.17 0.72 2.53 2.53 0
1 45.41 21.3 0.18 0.85 2.61 3.85 0.48
Based on a more comprehensive study of the compression behaviour of over 100

plain and fibrous lightweight concrete cylinders, the following equations of static
modulus of elasticity and strain at peak were derived:
Estatic ¼ 4:54flcm
0:43
ðGPaÞ with R2 ¼ 0:92 ð1Þ
ð2Þ
ð3Þ
Using all the above it can be deduced that the constitutive r-e model in com-
pression for SFRLC is identical to that of plain concrete with the exception of eult
Eq. (3).
5.1.2 Tension
Figure 8 below illustrates the pullout load-slip of a few plain and fibrous concrete
specimens with different parameters.
Pullout Load-Slip
700
fck=30MPa, Vf=0%
600 fck=30MPa, Vf=1%, 3D, LE=24mm
fck=30MPa, Vf=1%, 3D, LE=19mm
Pullout Load (N)
500 fck=30MPa, Vf=2%, 3D

400 fck=30MPa, Vf=1%, 5D, LE=26mm
fck=30MPa, Vf=1%, 5D, LE=12mm
300 fck=40MPa, Vf=1%, 5D, LE=14mm
fck=30MPa, Vf=1%, 3D hooks cut off
200
100
0
0 5 10 Slip (mm) 15 20 25
Fig. 8. Pullout load-slip of some of the specimens tested
It can be seen that, once plain concrete reaches its maximum strength, it fails
abruptly in a brittle manner. As was mentioned previously, this behaviour illustrates the
absence of any tension stiffening mechanism in the form of the aggregate interlock
present in normal weight concrete. For all fibrous specimens, the behaviour becomes
ductile with specimens with higher embedment length LE developing a more increased
ductility while the peak strength remains unaffected (solid and dashed green lines). The
latter remains correct provided that the embedded length is large enough to fully bond
the hook end length LH, otherwise this can give rise to premature concrete fracture near
the hook preventing the fibre from developing full strength (dashed blue line). For the
hook length to be fully bonded to develop the maximum pullout strength Eq. (1) must
be satisfied.
LE [ LH þ 5df ð4Þ
with df as the fibre diameter in (mm).

This equation also agrees with similar findings in Abdullah et al. (2018) on
DRAMIX hooked-end fibres. Moreover, for the 3D fibres with hooks being cut-off
(dotted black line), it can be seen that the mechanical hook contribution directly
contributes to the peak strength while the remaining length of the fibre only provides
frictional pullout for lightweight concrete. Depending on the fibre type, the pullout
strength varies with 5D resulting in the highest post-peak pullout strength and ductility
(so long as LE is sufficient and identical). The reason behind the differences in strength
is stemmed from the fact that the more extensive mechanical hooks (for 5D) are able to
develop a better bond with the concrete. The ductility differed due to the longer
extensive hooks requiring more work to straighten and eventually pull out. Increasing
Vf from 1% to 2% leads to generating nearly double the pullout strength at Vf = 1%.
Reducing the W/C ratio i.e. increasing the concrete compressive strength which means
having less air voids in the concrete lattice, leads to a higher pullout strength due to
developing a better fibre concrete bond (dashed and long dashed blue lines). Since the
contribution of the concrete is negligible post-crack due to the absence of natural

tension stiffening mechanism, the max residual stress of SFRLC due to fibre rein-
forcement can be written as follows.
Pmax
rSFRLC ¼ rtr1 ¼ rf Vf g0 ¼ ð5Þ
Ac
with Ac (mm2) as area of concrete that pertains to the fibre dosage (Soroushian and Lee,
1990), rf (MPa) fibre stress from the test and Pmax (N) is the maximum pullout. The
orientation factor η0 to account for the randomness of fibres in concrete is defined
differently in literature, and is taken as 0.5 (Hannant 1978) if no size-effects are
involved, or determined using Lee et al. (2011) chart when size-effects are influencial.
Another alternative to calculate the residual strength is by using bond strength sb from
available pullout tests:
sb L E
rtr1 ¼ 4 with df as the fibre diameter in ðmmÞ and sb as the bond strength ðMPaÞ
df
ð6Þ
with rtp (MPa) as the tensile plain concrete strength. Using the tensile tests on plain
concrete of LC30, 35 and 40, the plain concrete strength can be calculated using
Eq. (7).
rtp ¼ 0:065flcm ð7Þ
Based on the above, a semi-empirical r-x is derived in the following figure:
Fig. 9. Constitutive model in tension
etp is calculated by dividing rtp by Young’s modulus of elasticity assuming Ec =

Et. The crack width xtr1 is the slip needed for the fibre to develop maximum tensile
stress and is calculated using Eq. (8) adapted from Lok and Xiao (1999), with Ef as the
modulus of Elasticity of fibre (GPa) and LT as the characteristic length that converts x
to e and is an indication of crack length. LT varies depending on specimen size, Vf,
fibre type, reinforcement, loading level and matrix strength. No agreement has yet been
achieved to determine LT, however for beams, this paper suggests LT as the minimum
of srm (mean distance between cracks or hsp/2 for all phases of SFRLC deformation
with hsp as the unnotched depth (Barros et al. 2005; Tlemat et al. 2006; De Montaignac
et al. 2011; FIB 2013).
rtr1
xtr1 ¼ etr1 LT ¼ LT ð8Þ
E f g0 V f
Also,
rtr2 ¼ 0:6rtr1 at xtr2 ¼ L ð9Þ
rtr3 ¼ 0:3rtr1 at xtr3 ¼ LH ð10Þ
rtu ¼ 0:1rtr1 at xtu ¼ LH þ 5df ð11Þ
It should be noted that, the choice of the residual crack widths also agrees with the
pulley pullout model of Alwan et al. (1999) and its recent revision by Abdallah et al.
(2018b) on hooked-end fibres and is based on the geometry of the mechanical hooks,
while the choice of rtr2 and rtr3 is empirical based on the lower bound results from the
pullout tests. A similar approach can be used to derive constitutive models for any
particular concrete, fibre type and geometry. Figure 10 below compares the constitutive
models based on the notched beam tests up to a x of 2.5 mm (ULS) according to fib
Model Code 2010 (FIB 2013). On ABAQUS, in order to avoid mesh sensitivity for
unreinforced concrete, r-x method was used. To obtain r-x relation from r-e relation
for these constitutive models, the characteristic length LT was chosen as hsp/2 (Barros
et al. 2005). For Model Code 2010, hsp = 125 mm was recommended (Blanco et al.
2013).
Fig. 10. Comparison between the available constitutive models
Figure 11 below shows the Load-CMOD and Load-Deflection from the notched
beam test. As expected, the plain concrete notched beam failed once the flexural load
reached its peak about 7.5 MPa. Both notched beams reinforced with Vf = 1% and
Vf = 2% exhibited increased load capacity and ductility. The first developed a flexural
load capacity almost 3 times higher than that of plain concrete while the second reached
a load capacity about 4 times higher. The fracture energy Gf was calculated using the
area under graph. Gf for fibrous specimen of Vf = 2% was 65% higher than that of
Vf = 1%.
Fig. 11. Notched beams behaviour

Figure 12 shows half the notched beam simulated using the r-x constitutive model on
ABAQUS at displacement of 4 mm. It is clear that the failure pattern was concentrated
at the notch as expected. The simulation remained quasi-static throughout the analysis.
Fig. 12. Notched beam test at displacement 3 mm from ABAQUS
Figure 13 below illustrates the 4 constitutive models discussed in Fig. 10. It can be
seen that the SFRLC proposed constitutive model used predicted the notched beam
behaviour to a good accuracy from both a design and analysis point of view for
Vf = 1% and 2% while other SFRC models overestimated the strength and underes-
timated the ductility of fibrous lightweight concrete.
Fig. 13. Load-deflection predictions of notched beams
6 Conclusions
• The plain concrete fails in compression once it reaches its peak strength due to the
lack of aggregate interlock mechanism.
• The peak compressive strength is unaffected by the addition of conventional 3D
fibres. The more extensive hooks in 5D fibres however enhance flcm slightly.
• The addition of fibres leads to increased confinement and compressive ductility
leading the concrete failing in a ductile manner.
• The r-e behaviour of SFRLC in compression remains identical to that of plain
concrete with the exception of enhanced ductility.
• The pullout test designed was effective at predicting the post-peak behaviour of the
fibrous lightweight concrete.
• Due to the absence of natural tension stiffening, plain concrete fails in a brittle
manner once it reaches its peak tensile strength.
• An increase in Vf, number of hooks and reduction in W/C leads to an increase in
both strength and fracture energy.
• Only the hook length influences the peak strength for fibrous lightweight concrete
while the remaining embedded length of the fibre increases ductility due to fric-
tional pullout.
• A unique semi-empirical r-x constitutive model was derived and compared to
other available SFRC models which appeared to differ in peak load and ductility.
By applying FEA using ABAQUS’ concrete damaged plasticity model in a stress-
cracking displacement method which offers good mesh insensitivity, it was revealed
that the proposed constitutive SFRLC model is successful at predicting the
experimental load-deflection behaviour of the notched beams of different Vf to a
good accuracy and is hence deemed superior than the SFRC based models used in
this work.
References
Gerritse, A.: Design considerations for reinforced lightweight concrete. Int. J. Cem. Compos.
Lightweight Concr. 3(1), 57–69 (1981)
Libre, N., Shekarchi, M., Mahoutian, M., Soroushian, P.: Mechanical properties of hybrid fiber
reinforced lightweight aggregate concrete made with natural pumice. Constr. Build. Mater. 25
(5), 2458–2464 (2011)
Campione, G.: Flexural and shear resistance of steel fiber-reinforced lightweight concrete beams.
J. Struct. Eng. 140(4), 04013103 (2014)
Dias-da-Costa, D., Carmo, R., Graça-e-Costa, R., Valença, J., Alfaiate, J.: Longitudinal
reinforcement ratio in lightweight aggregate concrete beams. Eng. Struct. 81, 219–229 (2014)
durability-related properties of lightweight concrete reinforced with steel fibres. Mater. Struct.
50(1), 46 (2017)
Lytag: Ramboll Frame comparison study (2014). https://www.aggregate.com/our-businesses/
lytag. Accessed 31 Dec 2019
Chen, H., Huang, C., Tang, C.: Dynamic properties of lightweight concrete beams made by
sedimentary lightweight aggregate. J. Mater. Civ. Eng. 22(6), 599–606 (2010)
Badogiannis, E., Kotsovos, M.: Monotonic and cyclic flexural tests on lightweight aggregate
concrete beams. Earthquakes Struct. 6(3), 317–334 (2014)
Lim, H.S., Wee, T.H., Mansur, M.A., Kong, K.H.: Flexural behavior of reinforced lightweight
aggregate concrete beams. In: Asia-Pacific Structural Engineering and Construction Confer-
ence, pp. 68–82. APSEC, Kuala Lumpur (2006)
Wu, C., Kan, Y., Huang, C., Yen, T., Chen, L.: Flexural behavior and size effect of full scale
reinforced lightweight concrete beam. J. Mar. Sci. Technol. 19(2), 132–140 (2011)
lightweight concrete. Cem. Concr. Compos. 19(4), 307–313 (1997)
confined with transverse steel reinforcement. Cem. Concr. Compos. 26(6), 645–656 (2004)
beam–column joints. Eng. Struct. 59, 261–283 (2014a)
Di Prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in fib model code 2010:
Ritchie, A., Kayali, O.: The effects of fiber reinforcement on lightweight aggregate concrete. In:
Neville A (ed.) Proceedings of RILEM Symposium on Fiber Reinforced Cement and
Concrete, pp. 247–256. The Construction Press Ltd. (1975)
Swamy, N., Jones, R., Chiam, A.: Influence of steel fibers on the shear resistance of lightweight
concrete i-beams. ACI Struct. J. 90(1), 103–114 (1993)
Kang, T., Kim, W.: Shear strength of steel fiber-reinforced lightweight concrete beams. Korea
Concrete Institute, Oklahoma, pp. 1386–1392 (2010)
Iqbal, S., Ali, A., Holschemacher, K., Bier, T.: Mechanical properties of steel fiber reinforced
high strength lightweight self-compacting concrete (SHLSCC). Constr. Build. Mater. 98,
325–333 (2015)
RILEM TC 162-TDF: Bending test: final recommendation. Mater. Struct. 35, 579–582 (2002)
RILEM TC 162-TDF: r-e design method: Final Recommendation. Mater. Struct. 36, 560–567
(2003)
Barros, J., Cunha, V., Ribeiro, A., Antunes, J.: Post-cracking behaviour of steel fibre reinforced
concrete. Mater. Struct. 38(275), 47–56 (2005)
Fédération Internationale du béton: fib model code for concrete structures 2010, pp. 147–150.
Ernst and Sohn, Berlin (2013)
Mater. 178, 484–494 (2018a)
The Concrete Society: Guidance for the design of steel-fibre-reinforced concrete. Technical
report No. 63. Cement and Concrete Industry (2007)
Lyag: Technical Manual. Lytag ltd, London (2011)
Robins, P., Austin, S., Jones, P.: Pull-out behaviour of hooked steel fibres. Mater. Struct. RILEM
35, 4343–4442 (2002)
Krenchel, H.: Fiber spacing and specific fiber surface. In: Neville, A.M. (ed.) RILEM
Symposium on Fiber Reinforced Cement and Concrete, pp. 69–79. The Construction Press,
London (1975)
Romualdi, J.P., Mandel, J.A.: Tensile strength of concrete affected by uniformly distributed and
closely spaced short length of wire reinforcement. In: AC1 Journal Proceedings vol. 61, no. 6,
pp. 657–671, June 1964
Soroushian and Lee: Tensile strength of steel fiber reinforced concrete: correlation with some
measures of fiber spacing. ACI Mater. J. 87(6), 542–546 (1990)
Zienkiewicz, O.C., Taylor, R.L.: The Finite Element Method for Solid and Structural Mechanics,
vol. 2, 6th edn. Butterworth-Heinemann, Oxford (2005)
Tlemat, H., Pilakoutas, K., Neocleous, K.: Stress-strain characteristic of SFRC using recycled
fibres. Mater. Struct. 39, 365–377 (2006)
Syed Mohsin, S.M.: Behaviour of fibre-reinforced concrete structures under seismic loading. Ph.
D. thesis, Imperial College London, London, UK (2012)
Abbas, A., Syed Mohsin, S., Cotsovos, D., Ruiz-Teran, A.: Shear behaviour of steel-fibre-
reinforced concrete simply supported beams. Proc. Inst. Civ. Eng. Struct. Build. 167(9), 544–
558 (2014b)
Cotsovos, B.D.M., Abbas, A.A.: Behaviour of steel-fibre-reinforced concrete beams under high-
rate loading. Comput. Concr. 22(3), 337–353 (2018)
Cunha, V., Barros, J., Sena-Cruz, J.: Tensile behavior of steel fiber-reinforced self-compacting
concrete. In: Fiber-Reinforced Self Consolidating Concrete: Research and Applications (ACI
SP-274), pp. 51–68. American Concrete Institute, Detroit (2010)
Lok, T.-S., Xiao, J.R.: Flexural strength assessment of steel fiber reinforced concrete. J. Mater.
Civ. Eng. 11(3), 188–196 (1999)
Zhang, Y.J., Huang, Z.J., Yanga, S.L., Xua, X.W.C.: A discrete-continuum coupled finite
element modelling approach for fibre reinforced concrete. Cem. Concr. Res. 106(2018), 130–
143 (2018)
Ngo, D., Scordelis, A.C.: Finite element analysis of reinforced concrete beams. J. ACI 64(3),
152–163 (1967)
Rashid, Y.R.: Ultimate strength analysis of prestressed concrete pressure vessels. Nucl. Eng. Des.
7(4), 334–344 (1968)
De Montaignac, R., Massicotte, B., Charron, J.-P., Nour, A.: Design of SFRC structural
elements: post-cracking tensile strength measurement. Mater. Struct. 45(4), 609–622 (2012)
Al-Naimi, H., Abbas, A.: Ductility of steel-fibre-reinforced lightweight concrete, Eccomas
Proceedia. In: COMPDYN, pp. 4009–4023, Crete, Greece, 24-26 June 2019 (2019). viewed
31 Dec 2019. www.eccomasproceedia.org
Lambert, G.: Properties and behaviour of structural lightweight (Lytag-sand) concrete. PhD
thesis, University of Sheffield (1982)
Hannant, D.J.: Fibre Cements and Fibre Concretes. Wiley, Hoboken (1978)
Lee, S.-C., Cho, J.-Y., Vecchio, F.J.: Diverse Embedment Model for Steel Fiber-Reinforced
Concrete in Tension: Model Verification. ACI Mater. J. 108(5), 526–535 (2011)
Alwan, J., Naaman, A., Guerrero, P.: Effect of mechanical clamping on the pull-out response of
hooked steel fibers embedded in cementitious matrices. Concr. Sci. Eng. 1, 15–25 (1999)
Sadoon, A., Rees, D.W.A., Ghaffar, S.H., Fan, M.: Predicting pull-out behaviour of 4D/5D
hooked end fibres embedded in normal-high strength concrete. Eng. Struct. 172, 967–980
(2018b)
Blanco, A., Pujadas, P., de la Fuente, A., Cavalaro, S., Aguado, A.: Application of constitutive
models in European codes to RC–FRC. Constr. Build. Mater. 40, 246–259 (2013)
Li, F.-Y., Cao, C.-Y., Cui, Y.-X., Wu, P.-F.: Experimental study of the basic mechanical
properties of directionally distributed steel fibre-reinforced concrete. Adv. Mater. Sci. Eng. 3,
1–11 (2018)
UV-C Treatment to Functionalize the Surfaces
of Pet and PP Fibers for Use in Cementitious
Composites. Adherence Evaluation
María E. Fernández(&), María E. Pereira, Fernando Petrone,

Claudia Chocca, and Gemma Rodríguez
Faculty of Architecture, Design and Urbanism, University of the Republic

(Udelar), Montevideo, Oriental Republic of Uruguay, Uruguay
mefernandez@fadu.edu.uy
Abstract. The objective of this work is to present an evaluation of the adhesion

of synthetic fibers with cementitious matrices when the surface is functionalized
by exposure to UV-C light.
The synthetic fibers used were obtained from post-consumer containers of
polyethylene terephthalate (PET). Their performances were compared with
commercial macro-fibers of polypropylene (PP). Tensile strength and adhesion
in two matrices—one of Portland cement and one with a partial replacement of
Portland cement by ceramic waste—were evaluated at ages of between 7 days
and 6 months. The obtained values were compared with equal samples made
with fibers whose surfaces were not exposed to this radiation.
The results show that surface functionalization by this procedure does not
produce the expected effects in terms of adhesion with either of the two matrices
used. In commercial polypropylene fibers, functionalization is detrimental to its
tensile strength by decreasing it in the environment from 70% to 80%; this loss
of resistance of the polymer causes the start-up test to be interrupted by the
failure of the fiber. This does not occur in PET fibers; functionalization does not
substantially affect their tensile strength.
Keywords: Fibro-reinforced concrete Synthetic fibers UV-C

functionalization
1 Introduction
Waste management of plastics continues to be difficult because plastic production does

not decrease at the same rate as its consumption. This situation generates the need to
look for new ways to value plastics through recycling or reuse, and there are several
investigations that seek solutions to the final destination of synthetic waste; many of
these focus this valorization on their use as reinforcement of cementitious matrix
materials [1–3], particularly fibers obtained from difficult-to-manage industrial and
domestic waste products [4–8].
The adhesion between both materials is affected by the hydrophobic behavior of the
polymer surface and there are possible techniques that can be used to improve this
property [9–12]. One of the techniques investigated is the exposure of the surface of the
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_82
UV-C Treatment to Functionalize the Surfaces 939
polymers to UV radiation [13, 14], particularly to the wavelength fraction between

100–280 nm (UV-C) [15, 16]. The use of polyethylene terephthalate (PET) residues in
cementitious matrices has been shown to produce a superficial degradation of the
polymeric material, which decreases when a mixed matrix with ceramic residue is used
[17]. This hypothesis was used to investigate the performance of the polymer and the
interface through tensile and adhesion tests using PET and PP fibers for both matrices.
The results obtained at an early stage did not allow us to draw conclusive results
regarding the loss of resistance of the polymers, although they demonstrated that the
adhesion of PET fibers is much lower than that of commercial fibers (PP). In trying to
improve that adhesion, the surface functionalization was raised through its exposure to
UV-C radiation. Through the evaluation of the tensile performance of the polymeric
materials exposed to both matrices, and of the fiber-matrix adhesion, this work shows
the results obtained in ages up to 6 months, comparing them with those of the non-
functionalized fibers.
2 Experimental Procedure
For the preparation of the matrices, Portland CPN-40 cement was used. The mixed
matrix was made with a 25% replacement of cement with ground ceramic waste whose
pozzolanic capacity was previously evaluated. The chemical characteristics of the
cement can be seen in Table 1 and the particle size distribution of both materials in
Fig. 1. In both matrices the water/binder ratio was 0.5. For the pull-out samples for
both matrices, a plastic mortar was prepared as indicated in the UNE-EN 196-1
standard [18].
Table 1. Chemical characteristics of the cement used

% %
CaO 63.89 Soluble residue 0.64
Fe2O3 3.30 P.P.C. 1.44
SiO2 22.19
Al2O3 3.83 Compounds
MgO 2.97 C3S 55.70
SO3 1.86 C2S 21.60
K2O 0.22 C3A 4.56
Na2O 0.05 C4AF 10.04
Na2O eq 0.19
The fibers used were obtained from post-consumer containers of polyethylene

terephthalate (PET) and commercial macro-fibers of polypropylene (PP).
Samples were made from both polymers which had surfaces both with and without
functionalization. For the fibers with functionalized surfaces, an exposure of PET films
and PP fibers to UV-C radiation was performed. The exposure to this radiation was
940 M. E. Fernández et al.
Fig. 1. Distribution of particle size of Portland cement and ceramic waste.
carried out in a box made for this purpose (Fig. 2) with two 18 W lamps emitting at a
wavelength of 254 nm. Samples exposed between 1 to 7 days were obtained on each
side, to analyze the variation of the surface tension of the PET by microscope (Fig. 3).
By this method it was determined that at 7 days the surface showed a hydrophilic
behavior, defining the age of exposure for the present work.
Fig. 2. Equipment developed for the exposure of polymers to UV-C radiation.
The conditioning of the PET fibers for the evaluation of tensile strength was carried
out by cutting samples in the form of a halter, as indicated in Standard UNE-EN ISO
527-3 [19]. For the adhesion test (pull-out) the residue was cut using a document
shredder, obtaining fibers 4 mm wide, 0.21 mm thick and 35 mm long. Commercial
fibers were used with their original shapes and sections, 1.07 mm diameter, and their
lengths were conditioned to 35 mm, equal to PET fibers, for pull-out tests.
The samples used to evaluate the evolution of the tensile strength of the fibers were
made in such a way that the degradation produced by the matrix was located in the
central area of the fiber (Fig. 4). They were demolded at 24 h and cured and submerged
in water at laboratory temperature (20 ± 1 °C) until the test date: 7, 28, and 56 days; 3
and 6 months. Prior to the test, the matrix was removed and subsequently washed with
a 1 molar hydrochloric acid solution. Eight samples were tested for each age, 4 without
Fig. 3. Variation of surface tension using optical microscope images.
functionalization (PET with cementitious matrix, CMPET; PET with mixed matrix,
MMPET; PP with cementitious matrix, CMPP; PP with mixed matrix, MMPP) and 4
with functionalization (CMPET-UVC; MMPET-UVC; CMPP-UVC; MMPP-UVC).
The test procedure used was that indicated in standards UNE-EN 14889-2 [20] and
UNE-EN ISO 6892-1 [21].
Fig. 4. Fibers with matrix in central part for subsequent tensile testing. Left: PET. Right: PP.
For the evaluation of the adhesion by means of a double pull-out test, 25 mm

25 mm 100 mm specimens were prepared whose molds were specially designed for
this purpose (Fig. 5). Screws were left in the samples for fastening to the test equip-
ment. After 24 h of filling, they were demolded and kept wrapped in film paper at
laboratory temperature (20 ± 1 °C) until the test dates mentioned above. There were 8
samples for each test age with the same denomination.
The test equipment consisted of a universal ZPM tester equipped with a 1000 N
load cell, 1 45° hardened pyramid jaws, and software that allows the speed to be
regulated continuously and facilitates data acquisition. In both types of trial, the speed
was 1 mm/min, the same as that used in previous studies in order to compare results.
Fig. 5. Molds used to make pull-out samples
3 Results Obtained and Discussion
3.1 Tensile Strength

As mentioned earlier, tensile strength was determined in fiber samples with and without
surface functionalization and with exposure to two matrices in each case. The obtained
values were compared with the control sample, without immersion in any matrix (PET
and PP). Maximum tensile strength and elastic modulus were determined. Figures 6, 7,
and 8 show the tensile strength values obtained.
Fig. 6. Tensile strength results of the fibers studied without functionalization
According to these results, it can be observed that the tensile stress behavior of
these fibers exposed to both matrices until 6 months of age does not show great
variation with respect to the control sample. Although values that differ with the
standard samples were obtained, they do not present a clear pattern of behavior, with
those values included in the dispersion ranges of the results. The tensile strengths of
PET fibers in all cases were lower than those of PP, comprising approximately 50% of
their value. Considering the two matrices studied with both polymers, no statistically
significant differences were found between the results obtained, and the F values were
much lower than the critical value for F (FPET = 0.0164; FPP = 0.079; FCRIT = 4.965)
and the highest p-value than 0.05 for PET and PP (0.901 and 0.785, respectively).
Fig. 7. Results of tensile strength of PET fibers with and without functionalization
Fig. 8. Results of tensile strength of PP fibers with and without functionalization
Analyzing the effect of surface functionalization, we can see that in PET fibers the
exposure to UVC radiation generates a small increase in the tensile strength of the
material before exposing it to the matrices. However, when the polymer is exposed to
the cementitious and mixed matrix, this variation is no longer observed at all ages,
tending to stabilize with age.
Unlike PET fibers, PP fibers with a functionalized surface have tensile strengths
with values of between 20% and 30% of those reached by PP fibers that have not been
exposed to UV-C radiation. Regarding the standard fibers with and without surface
functionalization, they also have a similar behavior, with no observable large differ-
ences in the ages analyzed.
3.2 Elasticity Module

From the results obtained in the tensile tests described above, the modulus of elasticity
was determined for each situation. The results obtained can be seen in Figs. 9, 10, and 11.
In PET fibers embedded in the Portland cement matrix it can be observed that the
modulus of elasticity increases as age progresses, implying greater rigidity of the
material. However, when they are embedded in the mixed matrix, they have a more
variable behavior; their value stabilizes after 3 months of age. PP fibers embedded in
Portland cement matrix have an elastic modulus similar to the standard sample; the one
that is observed when the fibers are embedded in the mixed matrix is more variable.
Fig. 9. Results of the elastic modulus of the fibers studied without functionalization
Fig. 10. Results of the elastic modulus of PET fibers with and without functionalization
Fig. 11. Results of the elastic modulus of PP fibers with and without functionalization
If we analyze the effect of functionalization of the surfaces in the modulus of

elasticity we can observe that in PET fibers, when they are embedded in the Portland
cement matrix for 28 days, the exposure UV-C radiation produces a small increase and
the fibers have a variable behavior from 56 days of age. When PET fibers with a
functionalized surface are exposed to the mixed matrix, they have a variable behavior
that does not allow determination of their pattern.
In commercial fibers, surface functionalization does not produce variation in the
elastic modulus of the standard samples prior to immersion in the different matrices. PP
fibers embedded in the Portland cement matrix, up to 56 days of age, show that the
functionalization of the fibers produces a small decrease in the modulus of elasticity.
From 3 months onwards, the values obtained do not show differences between fibers
with an un-functionalized surface and those that were exposed to UV-C radiation.
However, in commercial PP fibers embedded in the mixed matrix, except for the results
obtained at 7 days, the modulus of elasticity is similar when fibers with and without
functionalized surfaces are used.
3.3 Adherence Resistance

The adhesion strength was determined by the double pull-out test. The samples were
made with fibers of both polymers, PET and PP, embedded in plastic mortars made
with the same matrices used to determine bond stress. In turn, samples were made with
fibers with and without surface functionalization by exposure to UV-C radiation.
The results obtained can be seen in Figs. 12, 13, and 14.
Fig. 12. Bond stress results of the fibers studied without functionalization
The fibers with PET embedded in the Portland cement matrix (CMPET) have an
increasing adhesion resistance over time, except at 3 months of age, at which time the
result is approximately 8% lower than that obtained at 56 days of age. The samples
embedded in mixed matrix (MMPET) have lower adhesion tensions than those
obtained with the Portland cement matrix, except at 3 days of age. However, in PP
fibers, the behavior is variable and the dispersion of results much greater than those
obtained with PET fibers. At all ages, the tear resistance of PET fibers is lower with
values between 20% and 50% of those obtained with PP fibers. These tests ended when
the fibers detached from the matrix across its length.
PET fibers with functionalized surfaces embedded in the cement matrix (CMPET-
UVC) have an adhesion resistance without a clear pattern of behavior. At 6 months of
age, the tear resistance is the same if fibers with surfaces with and without UV-C
exposure are used. The same variable behavior occurs when the fibers used have been
exposed to the mixed matrix. In these fibers, at 6 months, the adhesion obtained with
fibers exposed to UV-C (MMPET-UVC) is 20% lower than that for fibers that did not
undergo functionalization (MMPET). In all samples the test was completed when the
fiber was completely extracted from the matrix.
Fig. 13. Bond stress results of PET fibers with and without functionalization
Fig. 14. Bond stress results of PP fibers with and without functionalization
The results shown in Fig. 14 show that the PP fibers whose surfaces were exposed
to UV-C radiation and were embedded in the Portland cement matrix (CMPP-UVC)
mostly present tear tensions lower than fibers without functionalization. However, the
tests of the samples with ages greater than 7 days, whose values are indicated by an
asterisk, were completed before the detachment of the fiber; in these cases, the test
ended when the fiber broke due to tensile stress. In samples with mixed matrices, this
type of rupture occurred only in samples with a radiation-exposed surface (MMPP-
UVC) at ages 3 and 6 months; for other ages, the adhesion tension presented by the
phases was lower when the fiber was functionalized with light radiation.
4 Conclusions
From the results obtained in this work, the following conclusions can be drawn:
• The functionalization of PET fiber surfaces through exposure to UV-C radiation
does not produce significant changes in tensile strength or the modulus of elasticity.
• Exposure to UV-C radiation of commercial PP fibers produces a decrease in tensile
strength of between 70% and 80%. However, these changes do not affect the
modulus of elasticity.
• In most of the results, the functionalization of the surfaces by means of UV-C

radiation does not show improvements in the adhesion between the fibers and the
different matrices. In the case of PP fibers with ages greater than 28 days, it causes a
significant decrease in joint work, causing the fiber to break due to its degradation,
before its total detachment.
Therefore, it is necessary to continue investigating other techniques for the func-
tionalization of PET surfaces that allow improvement of their adhesion, such as the use
of physical methods that generate a greater roughness on the surface of the material.
References
1. Naik, T.R., Singh, S.S., Huber, C.O., Brodersen, B.S.: Use of post-consumer waste plastics
in cement-based composites. Cem. Concr. Res. 26, 1489–1492 (1996)
2. Sharma, R., Bansal, P.P.: Use of different forms of waste plastic in concrete—a review.
J. Clean. Prod. 112, 473–482 (2015)
3. Gu, L., Ozbakkaloglu, T.: Use of recycled plastics in concrete: a critical review. Waste
Manag 51, 19–42 (2016)
4. Pelisser, F., Montedo, O.R K., Gleize, P.J.P., Roman, H.R.: Mechanical properties of
recycled PET fibers in concrete. Mater. Res. 15, 679–686 (2012)
5. Borg, R.P., Baldacchino, O., Ferrara, L.: Early age performance and mechanical
characteristics of recycled PET fiber reinforced concrete. Constr. Build. Mater. 108, 29–
47 (2016)
6. Vijaya, G.S., Ghorpade, V.G., Sudarsana Rao, H.: The behavior of self-compacting concrete
with waste plastic fibers when subjected to chloride attack. Mater. Today Proc. 5, 1501–1508
(2018)
7. Meza, A., Siddique, S.: Effect of aspect ratio and dosage on the flexural response of FRC
with recycled fiber. Constr. Build. Mater. 213, 286–291 (2019)
8. Mohammed, M.-K., Al-Hadithi, A.I., Mohammed, M.H.: Production and optimization of
eco-efficient self-compacting concrete SCC with limestone and PET. Constr. Build. Mater.
197, 734–746 (2019)
9. Awaja, F., Gilbert, M., Kelly, G., Fox, B., Pigram, P.J.: Adhesion of polymers. Prog. Polym.
Sci. 34, 948–968 (2009)
10. Drobota, M., Persin, Z., Zemljic, L.F., et al.: Chemical modification and characterization of
poly (ethylene terephthalate) surfaces for collagen immobilization. Cent. Eur. J. Chem. 11,
1786–1798 (2013)
11. Cazan, C., Cosnita, M., Duta, A.: Effect of PET functionalization in composites of rubber–
PET–HDPE type. Arab. J. Chem. 10, 300–312 (2017)
12. Michael, F.M., Khalid, M., Walvekar, R., Siddiqui, H., Balaji, A.B.: Surface modification
techniques of biodegradable and biocompatible polymers. Biodegrad. Biocompatible Polym.
Compos. 33–54 (2018). Elsevier
13. Mathieson, I., Bradley, R.H.: Improved adhesion to polymers by UV/ozone surface
oxidation. Int. J. Adhes. Adhes. 16, 29–31 (1996)
14. Moyano, M.A., Martín-Martínez, J.M.: Surface treatment with UV-ozone to improve
adhesion of vulcanized rubber formulated with an excess of processing oil. Int. J. Adhes.
Adhes. 55, 106–113 (2014)
15. Ossola, G., Wojcik, A.: UV modification of tire rubber for use in cementitious composites.
Cem. Concr. Compos. 52, 34–41 (2014)
16. Coopamootoo, K., Masoero, E.: Cement pastes with UV-irradiated polypropylene: fracture
energy and the benefit of adding metakaolin. Constr. Build. Mater. 165, 303–309 (2018)
17. Fernández Iglesias, M.E., Payá, J., Borrachero, M.V., Soriano, L., Mellado, A., Monzó, J.:
Degradation process of postconsumer waste bottle fibers used in portland cement – based
composites. J. Mater. Civ. Eng. 29(10) (2017). Content ID 04017183
18. AENOR: UNE-EN 196-1 Methods of testing cement. Part 1: Determination of strength
(2018)
19. AENOR: UNE-EN ISO 527-3 Plastics. Determination of tensile properties. Part 3: Test
conditions for films and sheets (1996)
20. AENOR: UNE-EN 14889-2 Fibers for concrete. Part 2: Polymeric fibers. Definitions,
specifications and compliance (2008)
21. AENOR: UNE-EN ISO 6892-1 Metal materials. Tensile test Part 1: Test method at room
temperature (2017)
Potential of Using Recycled Carbon Fibers
as Reinforcing Material for Fiber Concrete
Magdalena Kimm1(&), Amna Sabir1,2, Thomas Gries1,

and Piyada Suwanpinij2
1
Institut für Textiltechnik of RWTH Aachen University, Aachen, Germany
magdalena.kimm@ita.rwth-aachen.de
2
The Sirindhorn International Thai-German Graduate School of Engineering,
King Mongkut’s University of Technology North Bangkok, Bangkok, Thailand
Abstract. Carbon fibre reinforced polymers (CFRP) are taking over the aero-
space, automotive, and wind energy sector. The growing demand for CFRP
leads to a growing mass of CFRP waste. Landfilling and incineration are not
effective and environmentally friendly routes for CFRP waste. The primary aim
of this research paper is to determine the potential of using short pyrolyzed
recycled carbon fibres (rCF) from CFRP waste to enhance the mechanical
properties of fibre reinforced concrete (FRC). The secondary objective is to
enhance the mechanical properties of FRC by optimizing fibre-matrix bonding
in the interphase region. Recycled carbon fibre reinforced concrete (rCFRC) 60
specimens were prepared. These specimens consist of different rCF volume
content of 0.00, 0.25, 0.5, 0.75 and 1 vol.-%. 4-point bending test procedure,
along with visual analysis was performed for the characterization of the speci-
mens. Oxygen (O2) plasma treatment has been used to improve the rCF (rein-
forcement) and concrete (matrix) adhesion. A maximum gain of 31% was
achieved in flexural strength at fibre volume content of 0.5 vol.-% as compared
to plain concrete. O2 plasma-treated rCFRC has a much higher value for
elongation at break as compared to untreated rCFRC series.
KEYWORDS: Carbon fibre reinforced polymer Recycled carbon fibres

Fibre reinforced concrete
1 Introduction
Modern industry is interested in materials with high strength-to-weight ratio and

stiffness, e.g. to produce fuel-efficient aircrafts and automobiles, massive wind energy
blades with high-efficiency rates, robust and lightweight pressure vessels or material-
efficient building structures. Carbon fibres (CFs) and CFRP are durable, lightweight,
and possess the ability to mold into complex shapes. These properties make them
potential and primary candidates for manufacturers [1, 2]. Global demand for CFRP in
the year 2018 was approximately 75.5 kt. This demand is forecasted to reach 120.5 kt
in the year 2022 [3]. Increasing demand for CFRP directly leads to an increase in waste
too. By estimation, a sum of 5 kt of recycled CFRP is annually being wasted in Europe.
Incineration is not the solution to CFRP waste because 50% to 70% of slag of
incinerated fibre reinforced polymers are minerals and ashes which still need to be
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_83
950 M. Kimm et al.
landfilled [4]. Industry-specific legislations, such as the End of Life Vehicle Directive
(ELVD), pose high demands to producers to design vehicles towards recyclability [1].
The government’s current primary goal is also to focus on circular economy and
sustainable built environment [6]. Therefore, a sustainable, economic, and nature-
friendly path for CFRP waste recycling is need of the hour.
An attractive approach is the use of rCF from CFRP waste as reinforcement for
concrete. Mechanical properties using conventional fibres made of steel and alkali-
resistant glass can be achieved by rCF. On the other hand, rCF can easily overcome the
disadvantages of both steel and AR-glass fibres. Steel rust easily, which leads to
deterioration of the concrete structure. Additionally, the mechanical properties of steel
FRC are limited. As the workability of concrete becomes lower with increasing fibre
volume content leading to balling effects [7, 8]. In particular, conventional glass fibres
and also alkali-resistant glass fibres become brittle due to alkalinity of cement mortar.
Moreover, alkali-resistant glass fibres show a high price compared to conventional
glass and steel fibres [8].
CFs are high-performance and expensive fibres commonly used in composite
structures for the aerospace and automotive industries. Production process of vCFs is
energy-intensive, which leads to high fibre cost. Due to this reason, vCFs have not been
established in the construction market [9]. Therefore, from an environmental and
economic perspective, it is important to test the potential application of cost-reduced
recycled CFs. There are three main types of recycling, namely mechanical, chemical,
and thermal recycling routes. Mechanical recycling is used in two steps, these steps are
known as primary and secondary processes. Mechanical recycling targets size reduc-
tion by shredding and cutting. Thermal (pyrolysis) and chemical (solvolysis) processes
are identified under the category of tertiary processes. Pyrolysis process is most widely
and commercially used for recovery of fibres from glass and carbon fibre reinforced
composite [2, 5].
Recycling by pyrolysis reduces cost from 57 €/kg (vCFs) to 5-10 €/kg (rCF) and
the required energy from 55–166 kWh/kg to 3–10 kWh/kg [9]. rCF recovered through
pyrolysis retain 98–99% of young’s modulus of vCFs and only lose 4–10% of the
virgin fibre’s tensile strength. The only disadvantage is the loss in fibre length during
the recovery process. In research and current composite material market, rCF from
pyrolysis have not established yet; just reuse of cut-off from vCFs is taking place,
mostly in the companies themselves. Therefore, new markets for the application of rCF
have to be focused. Thus, in this research the influence of short pyrolyzed recycled
carbon fibres as reinforcement for FRC for the construction industry is studied [9, 10].
Some research has been conducted on the integration of recycled carbon fibres in
concrete by Mastali et al. and Saccani et al.:
• In work done by Mastali et al., the fresh and hardened properties of FRC were
assessed, considering rCF at four different fibre volume contents. Maximum fibre
content was up to 2 vol.-% along with three different fibre lengths of 10, 20, and
30 mm [4, 7]. Compressive strength and impact resistance were observed to be
maximum with fibre length of 30 mm and flexural strength was maximum with
20 mm fibre length. CFs were recycled from the remained unusable CFs sheets and
were mechanically shredded into different lengths without any thermal treatment.
Potential of Using Recycled Carbon Fibers as Reinforcing Material for Fiber Concrete 951
Therefore, these results are comparable to the use of chopped vCFs not comparable
to rCF gained by pyrolysis as in this work.
• Saccani et al. investigated the possibility of recycling composites made of carbon
fibres and epoxy resin in different inorganic matrixes without any previous thermal
and chemical treatment in concrete mixture [12]. Short chopped CFRP sticks (5–
8 mm), including resin, were used in different volume fractions up to 5 vol.-%.
The CFRP sticks shift the failure mode of composite from brittle to semi ductile.
• Kimm et al. have investigated the effect of surface modification and volume content
of pyrolyzed rCF in concrete. Fibre volume contents up to 2 vol.-% with an average
length of 14.9 mm were used as reinforcement in concrete [9]. The maximum gain
(+111%) in flexural strength for rCFRC samples in comparison to plain concrete
was observed at 1 vol.-%. Although, with an increase in volume percentage, the
workability of concrete mixer decreases. An enhanced bonding between concrete
and fibres by a silane treatment could not be achieved.
The material properties of FRC are generally influenced by fibre volume content, fibre
arrangement, length, fibre-matrix bonding, and concrete mixture, which predominantly
have been investigated with focus on steel polymer and glass fibres [11]. In general, rCF
show a chemically inert surface, and therefore, it does not react strongly to any matrix
materials unless a sizing is applied. Sizing is a microscopically thin coating of the fibre
surface, usually using polymer-based materials, which intends to improve adhesion to the
surrounding matrix in a composite material. During the thermal recovery process of
carbon fibres (pyrolysis), the sizing layer is completely burned down. Therefore, a
treatment after recycling of fibres is desired to improve the bonding to the concrete matrix.
Only limited work has been done on vCFs, while no significant work is done in case of
rCF. For surface treatment, the main approach is using a plasma, which is a mixture of
particles at the atomic molecular level consisting of partially charged components, ions,
and electrons. Depending on the plasma functional groups, e.g., carboxyl, carbonyl, or
hydroxyl groups, which accumulates on the fibre’s surface and alter fibre properties.
These functional groups modify the chemical or physical structure of fibres, thus tailoring
fibre-matrix bonding strength but without influencing their bulk mechanical properties.
• Schneider et al. investigated the incorporation of plasma-treated CFs yarns along
with mineral coating in concrete matrix [13]. Sized carbon fibre yarns were treated
by plasma. Three different gases (O2, O2 and Ar, O2 and CF4) were used with a
treatment time of 1.6 min and 3.3 min. Pure O2 plasma-treated CFs showed the
most improvement in mechanical properties (increase of pull out energy up to four
times in comparison to the untreated reference specimen).
• In the Ph.D. thesis of Hambach flexural and compressive strength of FRC with
plasma-treated vCFs (2–3 mm) using three different process gases (O2, CO2, NH3)
were investigated [14]. The results show that +6% gain in flexural strength was
achieved using plasma treatment. This gain in flexural strength was achieved at
relatively short exposure (2 min), at low energy (50 W), and using O2 as plasma gas.
Based on Schneider et al. and Hambach’s research work, one can conclude that O2
plasma treatment is a suitable choice for rCF surface treatment. This research work
952 M. Kimm et al.
aims to determine the potential of using recycled carbon fibres as reinforcement in

concrete considering the following aspects:
• Influence of different dosage (0, 0.25, 0.5, 0.75 and 1 vol.-%) of pyrolized rCF on
flexural strength and ductility of rCFRC.
• Effect of O2 plasma-treated rCF on flexural strength and ductility of rCFRC.
2 Experimental
2.1 Materials
Polyacrylonitrile-based rCF obtained by a pyrolysis process was used in the preparation
of FRC and are presented along with their technical data in Table 1 and Fig. 1 (a). The
fibre material was ordered as per company specification in the range of 10–30 mm.
Additionally, length analysis was carried out using a random number of 300 rCF
filaments out of the bulk, according to DIN 53808-1. The fibre length distribution is
shown in Fig. 2.
Fig. 1. a) rCF from pyrolysis b) plasma treated rCF from pyrolysis
Table 1. Properties of rCF

Properties Values
Manufacturer CFK Valley Stade Recycling GmbH & Co. KG, Wischhafen,
Germany
Material specification Carbon (>95%)
Product commercial CarboNXT pure
name
Density 1.77 g/cm3
Length Mean: 27 mm
Standard deviation: 12 mm
(details in Fig. 2(a))
Diameter 7 µm
Tensile strength* 3,620 N/mm2
Young’s modulus* 207,750 kN/mm2
*Values measured at Institut für Textiltechnik of RWTH Aachen University
Relative frequency by number of fibers in [%] 20%
13,9%
11,7%
11,0%
10% 8,9% 8,5%
7,1% 7,1%
6,4% 6,4%
5,7% 3,9%
3,9% 4,3%
2,1% 1,8%
1,1% 1,1% 1,4%
0,0% 0,4% 0,0%
0%
Length interval in [mm]
Fig. 2. Fibre length distribution of rCF measured based on DIN 53808-1
A part of rCF was treated with O2 plasma in the Atto Plasma System of Diener
electronic GmbH & Co KG, Germany. These fibres are shown in Fig. 1 (b) after
treatment. Process parameters and specifications are presented in Table 2.
The concrete matrix is composed of cement CEM I 42.5 R (490 kg/m3), fly ash
(175 kg/m3), micro silica powder (Elkem Microsilica® 940U, 70 kg/m3), quartz flour
(500 kg/m3), sand 0.2–0.6 mm (713 kg/m3), water (245 kg/m3) and Polycarboxylate
Ether-based plasticizer in different contents varying with the fibre volume content
(Master Glenium ACE 460, 7 kg/m3 for 0, 0.25 and 0.5 vol.-%, 7.5 kg/m3 for 0.75 and
8.125 kg/m3 for 1 vol.-%).
Table 2. Process parameters and specifications of plasma treatment

Process parameters Values
rCF quantity of single batch 1.79 kg
Time 5 min/each side of 3–5 cm thick layer of rCF
Pressure 0.3–0.5 mbar
Generate power 50 W
Gas flow rate 5–10 cm3/min
Process gas O2
954 M. Kimm et al.
2.2 Specimen Preparation

The tests were performed at fibre volume contents of 0, 0.25, 0.5, 0.75 and 1 vol.-% on
ten specimens per series with dimensions of 32.5 cm 10 cm 2 cm. In case of
plasma-treated fibres, only one specific fibre volume content 0.5 vol.-% was selected.
During preparation, the solid and liquid components of concrete were first pre-mixed
separately. Liquid components were gradually added into the solid component mixture
and were mixed with the help of a hand mixer. The concrete mixture was placed on
vibrating table for 1 min. rCF were afterwards added in small quantities to the concrete
mixture to avoid fibres agglomeration. The plasma-treated fibres were kept in a vacuum
storage bag after treatment and were used for FRC specimen preparation within 48 h.
A decrease in the workability of concrete mixture with an increase in fibre volume
content was observed. Therefore, at 0.75 and 1 vol.-%, an extra amount of plasticizer
was added (Subsect. 2.1). Molds were filled with the FRC mixture and were manually
leveled with the help of a trowel. FRC specimens in fresh state were placed on a
vibrating table for 20 s. Specimens were covered with plastic sheets and left to dry for
1 day. Then, the specimens were placed in a water bath for 7 more days and 20 more
days to dry in a room climate of 22 ± 3 °C. In total, 60 specimens were prepared.
2.3 Testing Method

For material characterization, EN 1170-5 was followed, which is the standard to
determine the flexural strength of glass FRC using 4-point bending test, see Fig. 3 (a).
A standard climate according to EN ISO 139 was maintained during the tests.
3 Results
3.1 Flexural Strength of RCFRC

In Fig. 3 (b), typical stress elongation curves of rCFRC series are shown. The flexural
strength r and elongation e were calculated based on a formula according to EN 1170-5:
FL 27 Dl
r¼ and e ¼ 2 d ð1Þ
b d2 5 L
with: force F and the geometrical relations L = Span (300 mm), Δl = deformation,
b = width of specimen [mm], and d = Thickness of specimen [mm], see Fig. 3 (a).
Fig. 3. a) Test setup according to EN 1170-5 b) average flexural strength of tested series c)
average elongation at flexural strength d) representative stress elongation curve of tested series
The trend obtained for the tested series resemble the statement of Kimm et al. [9] that
under flexural load, rCF exhibit different behavior in comparison to steel and glass fibre.
Increasing the fibre content up to 0.5 vol.-% increases flexural strength from rMOR =
7.43 to 10.32 MPa and elongation at break from 0.13% to 0.19%. At 0.75 vol.-% (see
Table 3 and Fig. 4 (c)), there is a sudden decrease in LOP, while at 1 vol.-% a broader
plateau with eMOR = 0.2% can be observed (see Fig. 4 (d)). The strength at 1 vol.-% is
higher than at 0.75 vol.-% but lower than both at 0.25 and 0.5 vol.-%. The MOR
decreases with an increase in volume. The broad plateau indicates the shift from brittle
to ductile behavior with increase in fibre volume content. It can be observed in
Fig. 3 (d) that the elongation at break for content of 0.25 vol.-% and 0.5 vol.-% is
increasing in a steady manner, but as the volume contents are increased to 0.75 vol.-%
and 1 vol.-% a substantial gain in elongation at break can be observed.
The result of plasma-treated rCFRC series are being represented by grey triangles
in Fig. 3 (c and d). Comparing plasma-treated rCFRC series to plain concrete, there is
an increase of 13% in flexural strength. But, there is a reduction of 18.5% in rMOR as
compared to rMOR obtained for untreated short rCFRC series at 0.5 vol.-%. Plasma
treatment also has an improving effect on eMOR, which increases by about 43%. The
956 M. Kimm et al.
value obtained for eMOR at 0.5 vol.-% with plasma-treated rCFRC along with standard
deviation is comparable to the content of 1 vol.-% of untreated rCFRC.
Table 3. rCFRC series’ properties under flexural load

Volume content Flexural strength Elongation at
of rCF [vol.-%] rMOR [MPa] break eMOR [%]
0 7.43 ± 0.44 0.13 ± 0.01
0.25 10.12 ± 1.06 0.14 ± 0.01
0.5 10.32 ± 0.38 0.14 ± 0.02
0.5 (plasma treated) 8.42 ± 0.55 0.20 ± 0.01
0.75 9.41 ± 0.6 0.19 ± 0.02
1 9.81 ± 0.29 0.20 ± 0.02
3.2 Visual Analysis of RCFRC

In Fig. 4, the cracked cross-section area of a plain concrete specimen is shown. Air
voids are minimal in number and in range of 0.05 cm to 0.07 cm in diameter.
Air voids
1 cm
Fig. 4. Plain concrete crack cross section
In general, it can be observed that increase in fibre volume content leads to more air
voids and with a greater diameter as compared to plain concrete. As shown in Fig. 5 at
0.25 vol.-%, the cracked cross section area highlights a few prominent air voids while
fibres are not noticeable here. As the volume percentage is increased, rCF can be easily
seen in the crack cross section area with the naked eye. Air voids are prominent at 0.75
vol.-% and 1 vol.-%. As shown in Fig. 1 (a) rCF are presented in bundles form rather
than individual fibres after the pyrolysis recovery process. Under the action of shearing
during mixing in the concrete matrix, some rCF are broken into individual shorter
fibres but bundles of intact rCF can also be seen in Fig. 5 (c, d).
b) 0.5 Vol.-% Air voids

a) 0.25 Vol.-% Air voids
1 cm
1 cm
c) 0.75 Vol.-% Air voids d) 1 Vol.-% Air voids
1 cm 1 cm
Fig. 5. Crack cross section area of rCFRC specimen
4 Discussion
As compared to plain concrete, rCFRC at 0.5 vol.-% shows a maximum gain in flexural
strength of +39%. A slight decrease of flexural strength can be observed with
increasing volume percentage beyond 0.5 vol.-%. According to general composite
mechanics, an increase in aspect ratio (length to diameter ratio) of fibres or volume
percentage of fibres will lead to an increase in mechanical properties. Nevertheless,
there is a practical limit for the fibre volume resulting in a saturation point beyond
which the increase in fibre volume content will deteriorate the mechanical properties of
concrete composite. The decrease in mechanical properties with increasing fibre vol-
ume content can be due to difficulty in mixing, air entrapment, or poor bonding
between fibre and matrix [15].
It can be deduced that plasma treatment dramatically increases the ductility of the
FRC as observed from the stress-elongation behavior of plasma-treated rCFRC series
in Fig. 3(d). Interestingly, the O2 plasma did not contribute to a higher flexural strength
as compared to untreated fibres which is different from the finding of Hambach and
Schneider [13, 14]. An explanation for this behavior is still anonymous.
Considering the uneven pattern of the standard variation of flexural strength, shown
in Table 3, the highest value of standard deviation was observed in case of the lowest
fibre content at 0.25 vol.-% in the contrary, a minimum value was observed in case of
the highest fibre content at 1 vol.-%. Generally, fibre length analysis (see Fig. 2) has
shown a considerable deviation in fibre range from the mean value and an uneven
pattern of fibre length distribution. Hence, fibre length analysis proves that fibre range
differs significantly from the ordered range of 10–30 mm. The maximum fibre length
measured was 64 mm, and the minimum was 4 mm due to the reason that rCF comes
from different origins and have different tensile and young modulus values. This might
have led to the diverging standard deviations of flexural strength between the series.
Especially for low fibre contents, a small amount of fibres are picked from bulk for
testing, which can vary strongly in fibre origin and length. So, the homogenous mixing
of rCF after recovery from CFRP waste is an important factor. In order to avoid
variation of fibre characteristics and to gain isotropic mechanical properties batches of
rCF should already be mixed at recycling companies [9].
In Fig. 6, the relationship between flexural strength and fibre volume content is
illustrated by linear regression analysis for all untreated series. The coefficient of
958 M. Kimm et al.
determination (R2 = 0.3031) is considerably below one, which indicates a strong

divergence of the real values from the regression curve. The value of R2 in the work of
Kimm et al. is 0.8868 for the same interval, but in contrast to this work, Kimm et al.
used a fibre material with shorter mean length and lower length variation [9]. The
coefficient of determination found by Mastali et al. is >0.93 [4, 7].
Fig. 6. Linear regression analysis between flexural strength rMOR and fibre volume content V
An important point in Mastali et al. work is that rCF were originated from a single
source, one fixed type of carbon fibre sheet, and fibres were mechanically shredded to
smaller pieces (no irregularities on fibre surface or deformation on fibre surface from
thermal processing for recovery of fibres) [4, 7]. Hence a good dispersion of fibres can be
easily obtained. So, Mastali obtained a higher coefficient of determination. In the case of
the present work, the fibre’s length had a broad range with twice the value of standard
deviation (Table 1) as compared to Kimm et al. where fibre length mean value is
14.9 mm and the standard deviation is 7.2 mm [9]. On the other hand in current research
work fibres went through a thermal process of recovery and single individual fibres are
not obtained rather than that bundles of fibres are obtained (see Fig. 2). Homogenous
mixing of closely adhered fibres in form of bundles is much difficult to achieve as
compared to mechanically shredded fibres. Considering the difference in origin of fibres,
source of fibres, and the recovery process of fibres, the vast difference in coefficient
of determination values of R2Mastali [ 0:93; R2Kimm ¼ 0:8868, and R2currentresearch ¼
0:3031 can be understood.
5 Conclusion
This work has been conducted to determine the influence of short recycled carbon
fibres (rCF) on the mechanical properties of fibre reinforced concrete (rCFRC). rCF
were added in concrete in different volume percentages of 0, 0.25, 0.75, 1 vol.-% and
analysed in terms of their flexural strength. Visual analysis was also conducted thor-
oughly to connect visual effects with the mechanical properties of the tested specimens.
O2 plasma treatment was carried out for short pyrolyzed rCF to enhance fibre-matrix
bonding. The results concluded based on mechanical and visual analysis are following:
• Concrete mechanical properties are dependent on fibre volume percentage. With the
increase in fibre volume content flexural strength improves, but to a specific volume
percentage limit.
• Maximum flexural strength was achieved at 0.5 vol.-% of rCF. This value is 39%
higher than the flexural strength of plain concrete.
• Maximum elongation at break value was achieved at maximum volume percentage
of conducted series (1 vol.-%). The same value was obtained in case of plasma-
treated rCFRC series at 0.5 vol.-%.
• Based on regression analysis, it can be concluded that origin, dispersion of fibres
and source of rCF effects the coefficient of determination value and further the
predictability of FRC.
• Optimized mechanical method of mixing of rCF in concrete matrix is required and
should be further investigated.
• Plasma treatment influenced the ductility of FRC composite but did not increase the
flexural strength as compared to untreated rCF.
• An increase in the fibre volume leads to decrease in workability of the concrete
paste inducing air voids and surface irregularities on the dried rCFRC test specimen.
Generally, the investigations have shown, that rCF from carbon fibre reinforced
polymers have a very great potential for a high-performance, non-hazardous reuse as
short fibre reinforcement in concrete. In particular, taking proper volume contents, fibre
material characteristics and mixing procedures into account. Also, in comparison to
vCFs, rCF offer exclusive economic benefits due to their lower costs and are therefore
considered to be competitive to conventional fibres in concrete [16].
References
1. Das, S., Warren, J., West, D.: Global Carbon Fiber Composites Supply Chain Competi-
tiveness Analysis, Clean Energy Manufacturing Analysis Center, Oak Ridge National
Laboratory. University of Tennessee, Knoxville (2016)
2. Melendi-Espina, S., Morri, C., Turner, T., Pickering, S.: Recycling of Carbon Fibre
Composites, pp. 1–4 (2016)
3. AKV-Industrial Association Reinforced Plastics eV. Composites Market Report 2018,
Market developments, trends, outlooks and challenges, Frankfurt, Deutschland (2018)
4. Mastali, M., Dalvand, A.: The impact resistance and mechanical properties of self-
compacting concrete reinforced with recycled CFRP pieces. Compos. Part B Eng. 92, 360–
376 (2016)
5. Fernández, A., Lopes, C.S., González, C., López, F.A.: Characterization of Carbon Fibers
Recovered by Pyrolysis of Cured Prepregs and Their Reuse in New Composites. IntechOpen
(2018). http://doi.org/10.5772/intechopen.74281
6. Pimenta, S., Pinho, S.T.: Recycling carbon fibre reinforced polymers for structural
applications: technology review and market outlook. In: Waste Management, New York, N.
Y., vol. 31, pp. 378–392 (2011)
960 M. Kimm et al.
7. Mastali, M., Dalvand, A., Sattarifard, A.: The impact resistance and mechanical properties of
the reinforced self-compacting concrete incorporating recycled CFRP fiber with different
lengths and dosages. Compos. Part B Eng. 112, 74–92 (2016)
8. Rai, A., Joshi, Y.P.: App. Prop. Fibre Reinfor. Concr. 4, 124–127 (2014)
9. Kimm, M.: Investigation of surface modification and volume content of glass and carbon
fibres from fibre reinforced polymer waste for reinforcing concrete. J. Hazard. Mater. (2019).
http://doi.org/10.1016/j.jhazmat.2019.121797
10. Russell, D.: Recycled Carbon Fibre: A New Approach to Cost Effective Lightweighting.
https://www.igcv.fraunhofer.de/content/dam/igcv/de/docs/Travelling_Conference. Accessed
27 Oct 2019
11. Triantafillou, T.C.: Blast Protection of Civil Infrastructures and Vehicles Using Composites.
Woodhead Pub Ltd., Boca Raton (2010)
12. Saccani, A., Manzi, S., Lancellotti, I., Lipparini, L.: Composites obtained by recycling
carbon fibre/epoxy composite wastes in building materials. Constr. Build. Mater. 204, 296–
302 (2019)
13. Schneider, K., Lieboldt, M., Liebscher, M., Fröhlich, M., Hempel, S., Butler, M., Schröfl, C.,
Mechtcherine, V.: Mineral-based coating of plasma-treated carbon fibre rovings for carbon
concrete composites with enhanced mechanical performance. Materials 10, 1–5 (2017)
14. Hambach, M.: High Strength Multifunctional Composites Based on Portland Cement and
Carbon Short Fibres. University of Augsburg, Germany (2016)
15. Naaman, A.E.: Fiber Reinforced Cement and Concrete Composite. Techno press 3000,
Florida (2017)
16. Kimm, M.: Recycling von Carbonbeton - Wie kann eine hochwertige Wiederverwendung
gelingen? In: ‘11. Carbon - und Textilbetontage, Dresden, 24–25 September 2019, pp. 34–
35 (2019)
Textile Reinforced Concrete (TRC)
Development of Textile Reinforced UHPC
with Reduced Steel Fiber Contents
Mengchao Zhai1, Yiming Yao1(&), Jingquan Wang1,

and Barzin Mobasher2
1
School of Civil Engineering, Southeast University, Nanjing, China
yiming.yao@seu.edu.cn
2
School of Sustainable Engineering and the Built Environment,
Arizona State University, Tempe, USA
Abstract. Textile reinforced ultra high performance concrete (TR-UHPC) is

developed to further improve the tensile strength and ductility with reduced steel
fiber contents. Alkali resistant glass textiles are used to partially replace steel
fibers in a hybrid manner. The effects of different steel fiber content on the
tensile and bending properties is investigated for four short fiber content levels.
It is found that the synergistic effects of the hybrid reinforcements can be used to
enhance of strengthening and toughening mechanisms with reduced cost. Pro-
nounced strain hardening and deflection hardening are observed under tension
and bending tests. The highest ductility of the TR-UHPC specimens are
obtained when the steel fiber content is 1.0% vol with enhanced strength.
Keywords: UHPC Textile Low fiber content Tensile behavior Flexural

behavior Strain hardening
1 Introduction
Ultra high performance concrete (UHPC) is a new class of cement-based material with
ultra-high mechanical properties and durability, which has been extensively investi-
gated in recent years [1, 2]. However, the high dosage of steel fibers affects the
workability of UHPC negatively and increases the cost. Textile reinforced concrete
(TRC) is a composite material composing of textiles and fine grained concrete, which
has the characteristics of high strength, large strain capacity and excellent durability
[3]. A combination of high performance textiles and UHPC could be an promising way
to further enhance the tensile properties of UHPC and reduce the use of steel fibers,
especially for thin-walled structures. At present, there are few researches on textile
reinforced ultra high performance concrete (TR-UHPC). Zhou et al. [4] investigated the
effects of different textile types and textile processing methods on the tensile properties
of TR-UHPC; Jiang et al. [5, 6] studied the tensile properties of basalt textile reinforced
engineering cement composites (ECC) and the bonding behavior between fiber and
matrix, the results show that ECC can greatly improve the reinforcement effect of
textile and the crack width can be effectively controlled.
In this study, a TR-UHPC reinforced with alkali-resistant (AR) glass textiles is
developed with different dosages of steel fiber. Mechanical properties of TR-UHPC
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_84
964 M. Zhai et al.
under tension and flexure were investigated and the effects of short steel fiber content
were studied.
2 Material
A proprietary UHPC mix produced by Jiangsu Subote New Materials Co., Ltd was
used. The short steel fibers with length of 6 mm were adopted. Alkali-resistant
(AR) glass textile was used as the reinforcement of the TR-UHPC specimens, and the
specific specification parameters are shown in Table 1.
Table 1. Properties and geometry of yarns made up the textiles.

Tensile Young’s Density Linear Linear density of
Strength Modulus (GPa) (g/cm3) density (tex) monofilament (tex)
(MPa)
1700 72 2.68 1200 19
3 Preparation of Specimens and Test Setup
The 300 mm 300 mm 18 mm square TR-UHPC thin plate was made by using a
laminating technique. The process started with pouring a thin layer of UHPC at the
bottom of the mold. Then a layer of textile was laid on this fresh concrete layer and
pressed properly to ensure the contact between the textile and the matrix. After the first
layer of textile was arranged, the second layer of UHPC was poured. The operations
were repeated until the last layer of textile was laid. All plates were demolded after 1
day, and cured for 28 days in a curing chamber (20 °C, 95% RH). In this experiment,
two layers of textiles with equal spacing were set for all specimens, in other words, the
thicknesses of UHPC layers were all 6 mm.
Six plate specimens with dimension of 50 mm 300 mm 18 mm were cut from
each of the large plates mentioned above for uniaxial tension and four-point bending
tests. The specimens were divided into four groups and labeled as 0.5%-T, 1.0%S-T,
1.5%T and 2.0%-T for tension, and 0.5%-B, 1.0%S-B, 1.5%B and 2.0%-B for bending.
The volume fraction of steel fibers ranged from 0.5% to 2.0%.
Aluminum sheet of 75 mm long were glued on both ends of all the direct tensile
test specimens to strengthen the clamping zone. The gauge length was 150 mm, and the
extensometer was used to measure the tensile deformation (Fig. 1). The uniaxial tensile
test was carried out with a controlled deformation rate of 0.5 mm/min.
Development of Textile Reinforced UHPC 965
Fig. 1. Dimension of TR-UHPC tensile specimen.
In four-point bending test, all specimens were tested with a span of 270 mm (see
Fig. 2). LVDT was placed to measure the mid span deflection of the specimen. The
displacement control was used in the test to ensure that the loading rate of the machine
is 1 mm/min.
Fig. 2. Bending test set up.
4.1 Tensile Behavior of TR-UHPC

The tensile stress-strain curves of TR-UHPC are shown in Fig. 3. The addition of short
fiber has apparent influence on the first-crack stress and tensile strength of the speci-
mens. When 1.5%S-T is compared with 0.5%S-T, in the first crack stress increases for
as much as 50%. Strain hardening stage of the stress-strain curves is observed after
cracking. The post-peak responses are characterized as extended softening behavior as
only two layers of glass textiles were used. While the ductility of was improved
prominently in comparison with normal UHPC. However, as the steel fiber content
increased to 2.0%, tensile strength was decreased. The tensile responses showed that
the hybrid reinforcements can achieve equivalent effects of strengthening and tough-
ening by partially replacing short steel fibers.
966 M. Zhai et al.
14 14
12 0.5%S-T1 12 1.0%S-T1
0.5%S-T2 1.0%S-T2
Tensile Stress, MPa
Tensile Stress, MPa

10 0.5%S-T3 10 1.0%S-T3
8 8
6 6
4 4
2 2
0 0
0 0.01 0.02 0.03 0 0.01 0.02 0.03
Strain, mm/mm Strain, mm/mm
(a) (b)
14 14
12 1.5%S-T1 12 2.0%S-T1
1.5%S-T2 2.0%S-T2
2.0%S-T3
Tensile Stress, MPa
Tensile Stress, MPa
10 1.5%S-T3 10
8 8
6 6
4 4
2 2
0 0
0 0.01 0.02 0.03 0 0.01 0.02 0.03
Strain, mm/mm Strain, mm/mm
(c) (d)
Fig. 3. Tensile stress - strain curves of TR-UHPC with different fiber volume fractions.
The effects of different steel fiber content on the tensile properties of TR-UHPC are
compared in Fig. 4. The strain values corresponding to the ultimate bearing capacity of
all fiber contents had little change. In contrast, when the fiber content was 1.5% vol, the
ultimate bearing capacity of the test specimens was the highest. While when the fiber
content was 2.0% vol, the ultimate bearing capacity was reduced by a certain extent. It
might be explained by higher matrix porosity due to high fiber dosage and interaction
between steel fiber and glass textiles [5]. The laying of textiles under high fiber content
increased the difficulty of the uniform fiber dispersion, which lead to more micro initial
defects.
0.5
14
12 0.4
Tensile Strength, MPa
Strain, mm/mm
10
0.3
8
6 0.2
4
0.1
2
0 0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
Steel Fiber Content, vol% Steel Fiber Content, vol%
(a) (b)
Fig. 4. (a) Average tensile strength and (b) average strain at ultimate bearing capacity at varying
steel fiber content.
4.2 Flexural Behavior of TR-UHPC

Figure 5 shows the experimental curves of bending stress and the midspan deflection,
where the bending stress was the maximum normal stress at the midspan section
calculated according to the plane section assumption. It can be seen that the flexural
strengths are higher than the direct tensile strength of TR-UHPC for same fiber volume
fraction. This can be explained by the differences in the stress distribution profiles of
the two test methods. In the tension test, the entire volume of the specimen is a
potential zone for crack initiation. Comparatively, in the flexural test, only a small
fraction of the tension region is subjected to an equivalent ultimate tensile stress.
With the increase of short fiber content, the bending strength of the specimens
showed a remarkable increase. However, when the fiber content was increased from
1.5% to 2%, the bending strength was slightly decreased, which is consistent with the
trends found in direct tension tests. When the fiber content was 1% vol, and the highest
ductility was obtained and the bending strength was also improved (see Fig. 6).
968 M. Zhai et al.
30 30
1.0%S-B3
25 0.5%S-B4 25 1.0%S-B4
0.5%S-B5
Bending Stress, MPa
Bending Stress, MPa

1.0%S-B5
20 20
15 15
10 10
5 5
0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Deflection, mm Deflection, mm
(a) (b)
30 30
25 1.5%S-B4 25 2.0%S-B4
1.5%S-B5 2.0%S-B5
Bending Stress, MPa
Bending Stress, MPa
1.5%S-B6 2.0%S-B6
20 20
15 15
10 10
5 5
0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Deflection, mm Deflection, mm
(c) (d)
Fig. 5. Bending stress versus mid-span deflection curves of TR-UHPC with different fiber
volume fractions.
30 10
25 8
Flexural Strength, MPa
Deflection, mm
20
6
15
4
10
2
5
0 0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
Steel Fiber Content, vol% Steel Fiber Content, vol%
(a) (b)
Fig. 6. (a) Average flexural strength and (b) average mid-span deflection at ultimate bearing
capacity at varying steel fiber content.
5 Conclusions
TR-UHPC reinforced by AR-glass textile and steel fiber were developed in this study.
The use of textiles in UHPC can effectively reduce the amount of steel fibers and thus
reduce the cost. The synergistic effects of hybrid fibers were ensured the strengthening
and toughening mechanisms of TR-UHPC with reduced steel fibers. Experimental tests
showed an increasing trend of tensile and flexural strength with increasing content of
steel fibers, while the highest strengths were obtained with 1.5% vol. of steel fiber. The
highest ductility was observed at fiber dosage of 1.0% with enhanced strength. This
new type of materials may be ideal for the application in thin-walled structures with
enhanced mechanical properties and pouring workability with reduced fiber contents.
Acknowledgements. This study was supported by the Natural Science Foundation of China
(51908120) and the Natural Science Foundation of Jiangsu Province (BK20180383). The
financial supports are gratefully appreciated.
970 M. Zhai et al.
References
1. de Larrard, F., Sedran, T.: Optimization of ultra-high-performance concrete by the use of a
packing model. Cem. Concr. Res. 24(6), 997–1009 (1994)
2. Chen, B., Ji, T., Huang, Q., et al.: Review of research on ultra-high performance concrete.
J. Architect. Civ. Eng. 31(3), 1–24 (2014)
3. Liu, S., Zhu, D., Li, A.: Research and application progress of textile reinforced concrete.
J. Architect. Civ. Eng. 34(5), 134–146 (2017)
4. Zhou, Z., Zhang, Y., Wang, Y., et al.: Experimental study on tensile mechanical property of
grid reinforced UHPC plates. J. SE Univ. 4, 1 (2019)
5. Jiang, J., Jiang, C., Li, B., et al.: Bond behavior of basalt textile meshes in ultra-high ductility
cementitious composites. Compos. Part B Eng. 174, 107022 (2019)
6. Li, B., Xiong, H., Jiang, J., et al.: Tensile behavior of basalt textile grid reinforced engineering
cementitious composite. Compos. Part B Eng. 156, 185–200 (2019)
Reinforcement of Concrete with Glass
Multifilament Yarns: Effect
of the Impregnation on the Yarn Pull-Out
Behaviour
A.-C. Slama(&), J.-L. Gallias, and B. Fiorio
CY Cergy Paris Université, L2MGC, 95000 Cergy, France

anne-claire.slama@u-cergy.fr
Abstract. The impregnation of the yarn by the cementitious matrix is a key

parameter to predict the mechanical behaviour of textile reinforced concrete.
The mechanism of this impregnation is specific because of the heterogeneous
nature of both materials. Since there is no method to observe and quantify this
impregnation, all the models defined to describe the mechanical failure of this
type of composite are based on statistics and incomplete observations. To adjust
those models to the experimental results and to understand the influence of the
matrix on the impregnation, several pull-out tests were performed on samples
composed of a glass multifilament yarn and several matrices. After pull-out test,
some samples were impregnated with resin in a three steps process that includes
cementitious matrix dissolution and enable to visualize the remaining filaments
and to assess their impregnation degree by the cementitious matrix.
Keywords: Multifilament yarn Mortar Impregnation Pull-out test

Confocal microscopy
1 Introduction
TRC, or textile reinforced concrete, is a promising material for multiple applications such
as prefabricate and structural building [1, 2]. Its advantages, compared to steel rein-
forcements, are no corrosion, lightness and flexibility that enable the manufacture of thin
structure with various innovative shapes. However, to fully use its potential, it is nec-
essary to understand its mechanical behaviour [3, 4]. This mechanical behaviour depends
mainly on the bond between those two assembled materials [5]. On the contrary to the
impregnation of textile reinforcement in most composites with resin polymer matrix, the
impregnation of the textile reinforcement in composite with cement matrix is not com-
plete, due to the specific structure of the mortar and the multifilament yarn [6]. A very
large proportion of the solid particles of the mortar has a bigger size than the spaces in
between the filaments of the yarn [7], leading to a random penetration of the matrix inside
the bundle. Therefore, there is a discontinuous impregnation in the transverse direction of
the yarn but also along it [8]. In the non-impregnated areas, the mechanical behaviour of
the free length of the filaments is also difficult to describe because it depends on the
waviness of those filaments and the inter-filament friction [9, 10]. All those elements were
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_85
972 A.-C. Slama et al.
modelled by several successive models that take into account all those specificities
[11–14]. However, since the visualization of the impregnation is complicated without
destroying the sample, the impregnation degree of the filaments is based on indirect
observation methods and statistical parameters [15, 16]. As a result, the modelling of the
bundle as concentric layer with impregnation level increasing from the outside to the core
is a simplification of the real situation and the impregnation degree is not defined from
experimental data [17]. In order to study the impregnation of a yarn embedded in different
cementitious matrices and with different embedded lengths, several pull-out tests have
been conducted. The matrices were characterized, and then pull-out samples were
manufactured and tested. The results of the pull-out test have been linked to matrix
characteristics and to the impregnation of the yarn by the cementitious matrix thanks to an
innovative visualization method.

2.1 Multifilament Yarns and Cement Matrices
The multifilament yarn used in this study is an Alkali Resistant glass direct roving of
2400 Tex from Owens Corning Cem-l 5325. Its characteristics are given in Table 1
according to [18].
Table 1. Characteristics of the yarn (supplier data)

Linear Diameter of the Sizing Modulus of elasticity Tensile Strength of
weight filaments (µm) (%) of the filaments (GPa) the filaments (MPa)
(Tex)
2400 27 0.8 72 1000–1700
Matrices with various rheological and mechanical properties were used, in order to
understand the effect of the matrix characteristics on the composite properties. They are
all based on standard mortar composition (NF EN 196-1 standard) with normalized
sand, CEM I 52.5 R cement where various quantities of Sika viscocrete Tempo 653
superplasticizer were introduced (expressed by the S/C ratio, where S is the amount of
the superplasticizer’s dry material) in order to obtain matrices with higher workability
and the same W/C ratio or matrices with the same workability and lower W/C ratios
and higher than the standard mortar’s compressive strength. The compositions of the
matrices used in this study are given in Table 2, with their workability (NF P18-452
standard) and compressive strength (NF EN 196-1 standard).
Reinforcement of Concrete with Glass Multifilament Yarns 973
Table 2. Matrices composition and properties

Mortar W/C ratio S/C ratio (%) workability (s) Compressive strength (MPa)
M50-00 0.50 0 10 58 + 2
M50-05 0.50 0.05 4 55 + 3
M50-10 0.50 0.1 2 59 + 3
M47-20 0.47 0.2 2 64 + 2
M45-20 0.45 0.2 10 68 + 2
M45-30 0.45 0.3 4 61 + 4
2.2 Samples Manufacturing and Pull-Out Test

Samples are of prismatic shape (4 4 16 cm), made in steel moulds, with a yarn placed
along the central axis of the prism, as shown in Fig. 1. After placing the yarns, the mortar
is made and cast according to the NF EN 196-1 standard. The samples are removed from
the mould after 24 h and stored in water for 27 days. Then, the edges of the samples are
sawed at the desired embedded length, called Le, to obtain the pull-out samples. After
drying, the pull-out samples are prepared for the pull-out tests by sticking the end of the
free length, Lf (after 10 cm of yarn), in-between two 2.5 5 cm epoxy plates.
After 24 h of drying in room conditions, the pull-out samples are ready for the pull-
out tests. The six matrices are tested with two different embedded lengths: 0.7 and
1 cm. Six pull-out samples are manufactured for each matrix and each embedded
length. However, since the free length of some samples was too damaged during
manufacturing, several samples were excluded from the study.
The pull-out tests are performed on an Instron testing machine at 2 mm/min until
displacement reaches the value of the embedded length. A special device to place the
pull-out sample is designed as shown in Fig. 1. The free length end is hold in the jaw of
the press, with the previously stuck plate. The obtained load/displacement curves are
then smoothed using weighted averages.
Fig. 1. Manufacturing process of the pull-out samples (left) and device for the pull-out test
(right)
2.3 Characterization of the Yarn/Mortar Interface

As described in the literature [13, 19, 20], the pull-out failure of a multifilament yarn
embedded in a cementitious matrix is characterized by the formation of two groups of
filaments with distinct behaviour, according to their impregnation degree: outer fila-
ments of the yarn that are well impregnated in the matrix fail in tension and filaments of
the core of the yarn that are not impregnated enough by the matrix, and that slip and are
extracted from the embedded length.
In order to differentiate the filaments that were well impregnated from those which
were sparsely or not impregnated, a new method of double resin impregnation joined
with confocal microscopy observations was developed.
It is based on two successive impregnations of samples after pull-out test with resin
with two different markers, separated by a step of cement matrix dissolution (presented
Fig. 2). The first step is an impregnation of the yarn inside the cementitious matrix with
a rhodamine (a red coloured marker) marked resin using a syringe to inject it in the hole
left by the extracted filaments. The second step is the cementitious matrix complete
dissolution in acid. The third step is a second resin impregnation with a fluorescein
(green/yellow coloured marker) marked resin of the initially obtained moulding of the
yarn into the cementitious matrix after matrix dissolution.
Fig. 2. Process of double resin impregnation
Thanks to this double marking, it is possible to identify different zones in the yarn
after pull-out test on several considered sections. The first zone is the void left by
filaments extraction on this section, which contains only the rhodamine marked resin
and no filament (zone 1 in Fig. 3). The second zone contains the filaments sparsely
impregnated on this section by the cementitious matrix before its dissolution (zone 2 in
Fig. 3). During the first impregnation, those filaments were not impregnated enough by
the cementitious matrix to prevent the rhodamine resin to reach them. On the contrary,
the third zone contains filaments on this section that were fully impregnated by the
cementitious matrix during the first impregnation by the resin with rhodamine (zone 3
in Fig. 3). After dissolution of the cementitious matrix, they were set free and
impregnated during the second resin impregnation with fluorescein. Despite they were
not reached by the rhodamine marked resin injection due to their complete inclusion in
the cementitious matrix, those filaments are not lost after dissolution of the cementi-
tious matrix because they stay connected to the rhodamine injected part of the yarn
through the excess of rhodamine that remain at the top face of the sample (on the yarn’s
free length side) after the injection. It can therefore be considered that almost no
filaments are lost after the two successive impregnations of the sample.
Fig. 3. Identification of three zones in a section of an impregnated yarn with double marking
observed with confocal microscopy
To study the variability of the filaments/ matrix interaction along the embedded
length of the yarn, different sections of those impregnated samples, made by successive
polishing with a polishing device Buelher, are observed using a confocal inverse
microscope ZEISS LSM 710 with a 10 lens. The areas of zones 1, 2 and, to some
extent, 3 are computed using ImageJ software. The Trainable Weka plugin is used to
count the filaments by machine learning. It is then possible to evaluate the areas of the
different zones and the number of filaments in zone 2 and zone 3. Averages are then
computed on the data obtained for each section.
3 Pull-Out Behaviours
3.1 Two Pull-Out Modes
Two pull-out mechanical behaviours were encountered: failure of the yarn in the
embedded length, followed by the slippage and extraction of some filaments on a
displacement equal to the embedded length or failure of the yarn in the free length of
the sample. In the following, these two failure modes are respectively referred as
mode 1 and mode 2. Examples of the behaviour associated to each of these modes are
presented in Fig. 4. It is also noticeable that some filaments failures along the free
length of the yarn are observed even for mode 1.
Mode 1 is differentiable from mode 2 by the appearance of a hole on the face of the
sawed surface after the pull-out test, which means that a number of filaments has been
extracted from the mortar. Those filaments were not impregnated or not impregnated
enough by the mortar along the embedded length so, during pull-out test, instead of
failure, they experience slippage along the embedded length after failure of some
potential links between the matrix and those filaments [21]. This sliding phenomenon is
visible on the load/displacement curve as a post-peak residual phase, which ends at a
displacement equal to the embedded length.
For mode 2, the mechanical behaviour is similar to the one of the yarn during the
tensile test. However, the pre-peak stiffness, post-peak stiffness and maximum load of
the load/displacement curve of mode 2 pull-out behaviour are lower than the ones of
the load/displacement curve of the yarn tensile test results. It can be due to the yarn
damage during manufacturing of the pull-out samples and the occurrence of some
failures in the embedded length. The load reaches zero for a displacement very inferior
to the embedded length and most maximum load values for this type of pull-out
behaviour are superior to the ones for the first type.
180
Pmax mode 1
160
140 mode 2
120
Load (N)
100
80 Pres
60
40
20
0
0 2 4 6 8
Displacement (mm)
Fig. 4. Two load/displacement curves obtained for pull-out tests of two samples from M50-10
Le = 1 cm and M50-00 Le = 0.7 cm
The results show that the mode 1 pull-out behaviour occurs generally more fre-
quently for lower compressive strength matrices but the correlation between these two
parameters is not strong (R2 = 0.55). On the other hand, the fluidity of the matrix has no
significant influence on the pull-out behaviour.
For all the samples, Pmax, the maximum load, is computed. It is the load value at the
peak.
When a number of filaments slipped and is extracted during the mode 1 pull-out
behaviour, the residual load, Pres, is computed as the load value of the intersection point
of the regression line associated to slope of the post-peak stiffness, and those of the
slope of the residual post-peak stiffness. The post-peak stiffness is the slope of the
regression line computed between 30% and 50% of Pmax in the post-peak phase. The
residual post-peak stiffness is the slope of the regression line computed between 30%
and 50% of the sliding length of the sample.
3.2 Influence of the Matrix on the Pull-Out

Despite a high variability, the results show that matrices with high fluidity and com-
pressive strength lead more frequently to high maximum load values and a failure in
the free length of the pull-out sample (mode 2). On the other hand, low fluidity and
compressive strength lead more frequently to the slippage and the partial extraction of
the filaments (mode 1). It was established also that the embedded length does not have
a significant influence on the pull-out mechanical results. More precisely, in this study,
a relatively strong correlation is found between the compressive strength of the
matrices and the average Pmax of all the tested samples as well as the average Pres
values of samples with mode 1 pull-out test independently of the embedded length
(Fig. 5). It is in accordance with Butler et al. [22] whose study shows that a densifi-
cation of the matrix (so a higher compressive strength) leads to more filaments failure
and less slippage. However, in his case, the maximum load values decrease with this
densification since it is the result of ageing and since other mechanisms can take place
during this ageing. In this present study, all the samples are tested at the same age, so
the maximum load values increase with matrix densification. This correlation does not
exist if the fluidity of the matrix is considered (see Fig. 6), on the contrary to what was
shown in the study of Peled and Mobasher [23], but this contradiction can be explained
by the fact that the fluidity has varied by changing the W/C and S/C ratios in the
present study and by adding silica fume and fly ashes in their case that modified also
the particle size distribution of the matrix.
For samples with mode 1 pull-out behaviour, the residual load value for each
sample is only weakly related to the maximum load value (see Fig. 7) so, even if the
compressive strength has a strong influence on both the maximum and the residual
load, different phenomena might interfere in the peak phase and in the residual phase,
leading to their weak link and high variability for each sample.
250 M50-00
R² = 0.7557
200 M50-05
Average Pmax (N)
150 M50-10
100 M47-20
50 M45-20
M45-30
0
50.00 60.00 70.00
Compressive strength (MPa)
50 M50-00
40
M50-05
Average Pmax (N)
R² = 0.7291
30
M50-10
20
M45-20
10
0 M45-30
50.00 55.00 60.00 65.00 70.00 75.00
-10
Compressive strength (MPa)
Fig. 5. Influence of the compressive strength of the matrix on the pull-out maximum load
(up) and residual load (down)
250 M50-00
Average Pmax (N)
200 M50-05
150 M50-10
100
M47-20
50
M45-20
0
0.00 5.00 10.00 M45-30
workability (s)
Fig. 6. Influence of the rheological properties of the matrix on the maximum load
250
R² = 0.2667
200
M50-00
Pmax (N)
150 M50-05
100 M50-10
M45-20
50
M45-30
0
0 20 40 60
Pres (N)
Fig. 7. Link between the maximum load and the residual load for the samples with mode 1 pull-
out behaviour
4 Influence of the Impregnation on the Pull-Out
To explain the different phenomena influencing the maximum load and the residual
phase, images obtained as described in Sect. 2.3 are analysed. Between the samples
considered previously, only the M50-00 and M50-10 samples, matrices with a majority
of mode 1 pull-out behaviour and the same W/C ratio are observed.
Considering the determined zones by double resin impregnation (Fig. 3), the results
show (Fig. 8) that there is no link between Pres values and the areas of the void left by
filaments extraction (zone 1), on the contrary to what was found by Banholzer [8].
There is no link also between Pmax values and the areas of the zones 1 and 2 or their
sum. However, there is a link between load parameters and the number of filaments
embedded in the cementitious matrix.
It is important to note that the area of the zone 3 is difficult to estimate because the
dissolution of the matrix releases the fully embedded filaments that could be reorga-
nized during the second impregnation, leading to an incorrect estimation of this area.
So, the number of sparsely impregnated filaments in zone 2 and fully impregnated
filaments in zone 3 is then considered to establish the correlation between the level of
impregnation and the pull-out behaviour. The most noticeable links observed are those
between the average number of filaments in zone 3 and the Pmax values (Fig. 9), and to
a lesser extent the Pres values. Therefore, it seems that those values depend essentially
on the number of filaments that are fully impregnated in the matrices. The number of
extracted and sparsely impregnated filaments does not have a direct influence on those
values. However, some other mechanisms that cannot be observed with this post pull-
out test visualisation method (such as friction between the filaments and the matrix
particles during extraction [24], surface roughness of the filaments due to the formation
of hydrate products on them [25] or successive extractions of several groups of fila-
ments at different displacements [8]) might influence the Pres values. Those other
mechanisms need to be investigated in further studies.
30
25
20
Pres (N)
15
10
5
0
0 0.5 1 1.5 2 2.5 3 3.5
Void left by filaments extraction (mm2)
M50-00 L=0,7 M50-10 L=0,7 M50-00 L=1
Fig. 8. Link between the void left by filaments extraction (area of zone 1) and the residual load
200
R² = 0.6366
150
Pmax (N)
100
50
0
-100 100 300 500 700 900
Number of impregnated filaments
M50-00 Le=0.7 M50-10 Le=0.7 M50-00 Le=1
Fig. 9. Link between the filaments fully impregnated and the maximum load
5 Conclusions
By comparison of the relative influence of the different matrices used in the study, the
paper sheds a new light on the impact of the matrix properties and of the yarn
impregnation on the pull-out behaviour of this yarn:
• Two pull-out modes are encountered. The first mode is characterized by the slip-
page followed by the extraction of a number of filaments, it is mostly found for
stronger matrices, and the second mode is characterized by the failure of most of the
filaments, without slippage, mostly found for weaker matrices.
• The matrix compressive strength influences the values obtained for the two
mechanical parameters computed on the pull-out load/displacement curve: the
maximum load for all the samples and the residual load for samples with filaments
slippage. The average maximum load and residual load both increase with the
compressive strength of the matrix. However, the fluidity of the matrix seems to
have no significant influence on the pull-out results.
• For the first mode with filaments slippage and extraction, a new visualization
method is developed, using double resin impregnation and matrix dissolution, and it
enables to conclude that the maximum load values can be explained by the number
of fully impregnated filaments. The residual load, however, is not related to the
extracted filaments.
References
1. Papanicolaou, CG.: 10 - Applications of textile-reinforced concrete in the precast industry.
In: Triantafillou, T., Textile Fibre Composites in Civil Engineering, pp. 227–244. Woodhead
Publishing, Cambridge (2016)
2. Kulas, C.: Actual applications and potential of textile-reinforced concrete in GRCA. In:
Proceedings of an international conference, Dubai, 1–11 April 2016
3. Häußler-Combe, U., Jesse, F., Curbach, M.: Textile reinforced concrete-overview,
experimental and theoretical investigations in Fracture mechanics of concrete structures.
In: Proceedings of an international conference, Vail, 12–16 April 2004
4. Hegger, J., Will, N., Rüberg, K.: Textile reinforced concrete — a new composite material in
advances in construction materials. In: Proceedings of an international conference, Stuttgart,
pp. 147–156, July 2007
5. Xu, S., Krüger, M., Reinhardt, H.-W., Ožbolt, J.: Bond characteristics of carbon, alkali
resistant glass, and aramid textiles in mortar. J. Mater. Civ. Eng. 16(4), 356–364 (2004)
6. Peled, A., Zaguri, E., Marom, G.: Bonding characteristics of multifilament polymer yarns
and cement matrices. Compos. Part Appl. Sci. Manuf. 39(6), 930–939 (2008)
7. Butler, M., Hempel, S., Mechtcherine, V.: Modelling of ageing effects on crack-bridging
behaviour of AR-glass multifilament yarns embedded in cement-based matrix. Cem. Concr.
Res. 41(4), 403–411 (2011)
8. Banholzer, B.: Bond of a strand in a cementitious matrix. Mater. Struct. 39(10), 1015–1028
(2006)
9. Chudoba, R., Vořechovský, M., Konrad, M.: Stochastic modeling of multi-filament yarns. I.
random properties within the cross-section and size effect. Int. J. Solids Struct. 43(3), 413–
434 (2006)
10. Vořechovský, M., Chudoba, R.: Stochastic modeling of multi-filament yarns: II. random
properties over the length and size effect. Int. J. Solids Struct. 43(3), 435–458 (2006)
11. Chudoba, R., Konrad, M., Mombartz, M., Vorechovskỳ, M., Meskouris, K.: ‘Multiscale
modeling of textile reinforced concrete within a consistent modeling framework’ in
‘Computation of Shell and Spatial Structures’. In: Proceedings of an international
conference, Salzburg, June 2005
12. Zastrau, B., Lepenies, I., Richter, M.: On the multi scale modeling of textile reinforced
concrete. Tech. Mech. 28(1), 53–63 (2008)
13. Banholzer, B., Brockmann, T., Brameshuber, W.: Material and bonding characteristics for
dimensioning and modelling of textile reinforced concrete (TRC) elements. Mater. Struct.
39(8), 749–763 (2006)
14. Lepenies, I., Meyer, C., Schorn, H., Zastrau, B.: modeling of load transfer behavior of AR-
glass-rovings in textile reinforced concrete. Spec. Publ. 244, 109–124 (2007)
15. Banholzer, B., Brameshuber, W., Jung, W.: Analytical simulation of pull-out tests—the
direct problem. Cem. Concr. Compos. 27(1), 93–101 (2005)
16. Banholzer, B., Brameshuber, W., Jung, W.: Analytical evaluation of pull-out tests—the
inverse problem. Cem. Concr. Compos. 28(6), 564–571 (2006)
17. Hegger, J., Will, N., Bruckermann, O., Voss, S.: Load–bearing behaviour and simulation of
textile reinforced concrete. Mater. Struct. 39(8), 765–776 (2006)
18. Cem-FIL® Roving 5325 - Owens Corning Composites [Internet]. https://www.
owenscorning.com/composites/product/cem-fil-roving-5325
19. Ohno, S., Hannant, D.J.: Modeling the stress-strain response of continuous fber reinforced
cement composites. Mater. J. 91(3), 306–312 (1994)
20. Homoro, O., Michel, M., Baranger, T.N.: Pull-out response of glass yarn from ettringite
matrix: Effect of pre-impregnation and embedded length. Compos. Sci. Technol. 170, 174–
182 (2019)
21. Badanoiu, A., Holmgren, J.: Cementitious composites reinforced with continuous carbon
fibres for strengthening of concrete structures. Cem. Concr. Compos. 25(3), 387–394 (2003)
22. Butler, M., Mechtcherine, V., Hempel, S.: Durability of textile reinforced concrete made
with AR glass fibre: effect of the matrix composition. Mater. Struct. 43(10), 1351–1368
(2010)
23. Peled, A., Mobasher, B.: Properties of fabric-cement composites made by pultrusion. Mater.
Struct. 39(8), 787–797 (2006)
24. Laws, V., Langley, A.A., West, J.M.: The glass fibre/cement bond. J. Mater. Sci. 21(1), 289–
296 (1986)
25. Vrijdaghs, R., di Prisco, M., Vandewalle, L.: Short-term and creep pull-out behavior of
polypropylene macrofibers at varying embedded lengths and angles from a concrete matrix.
Influence of Fibres Impregnation
on the Tensile Response of Flax Textile
Reinforced Mortar Composite Systems
Giuseppe Ferrara1, Marco Pepe1,2(&), Enzo Martinelli1,2,

and Romildo D. Tolêdo Filho3
1
Department of Civil Engineering, University of Salerno, Salerno, Italy
mapepe@unisa.it
2
TESIS srl, Fisciano, SA, Italy
3
Civil Engineering Department, COPPE, Federal University of Rio de Janeiro,
Rio de Janeiro, Brazil
Abstract. Textile Reinforced Mortar (TRM) composite systems as a technique

to retrofit and reinforce existing structures represents, nowadays, an efficient
application. The use of plant fibres textile as reinforcement, instead of the most
employed industrial ones, resulted a promising solution as response to the
sustainability criteria more and more required in the construction sector. How-
ever, some issues have been manifested as well related to the use of such natural
reinforcements in cement- and lime-based matrices mainly lying in the fibre-to-
matrix interaction and in the durability of both the overall composite and the
reinforcement embedded within the mortar.
In this context, the present study proposes an experimental activity aimed at
investigating the efficiency of an impregnation treatment (by using styrene
butadiene rubber latex) of flax fabrics on the mechanical behaviour of TRMs.
The study confirms that the use of impregnated textile leads to an improve-
ment of the overall behaviour of the composite system and paves the way for
further investigations aimed at verifying the efficiency also in terms of
durability.
Keywords: Plant fibres Composite TRM FRCM
1 Introduction
The use of Textile Reinforced Mortar (TRM) composites as strengthening system for
masonry structures in the last decade became a well-established reinforcement tech-
nique [1]. Such composites are performed by embedding high strength synthetic fibres
in inorganic matrices, typically cementitious or lime-based mortars.
The use of plant fibres, instead of the most conventional high strength synthetic
textiles, as reinforcement in TRMs emerged as a smart solution to reduce the envi-
ronmental impact of the strengthening system. Several studies focused the attention on
this type of application by using jute, sisal, coir, flax, hemp fibres in mortar based
composites [2–4]. These studies emphasised promising mechanical properties of plant
fibres, and a good compatibility with mortars in terms of strength and bond behaviour.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_86
984 G. Ferrara et al.
However, also emerged some drawbacks mainly due to durability issues, a large
deformability of the fibres at low strains, and variability of the properties of the textile
due to non-standardised industrial production processes [5, 6].
Fibre treatments by means of impregnation procedures, came up as an efficient
technique to address these issues. The application of resins on the textile external
surface may lead to fibre stiffness increase and improvement of the adherence with the
mortar [7–10].
In this context, the present study aims at investigating the mechanical behaviour of
TRM composite performed by means of impregnated flax textile. With the purpose of
analyse the influence of the treatment on the composite mechanical behaviour, two
series of specimens, characterised by non-impregnated and impregnated flax textiles,
were performed and tested in tension.
The study, highlighting a much better behaviour of the coupons characterised by
the impregnated textile, shows promising results and paves the way for further
investigations aimed at optimising the adopted treatment and at analysing its durability
performance.
The textile consists of a bidirectional flax fabric (Fig. 1a) whose physical and
mechanical properties are summarized in Table 1. The impregnated configuration
(Fig. 1b) was obtained by coating the textile by means of a carboxylated styrene
butadiene rubber latex, and by drying it for 24 h at a controlled temperature of 38 °C.
The mechanical characterisation of the textile was carried out by means of tensile tests
performed on non-impregnated and impregnated flax threads, constituting the ele-
mentary element of the fabric. The main mechanical properties are showed in Table 1.
Fig. 1. a) Flax textile; b) Impregnated flax textile; c) Mortar bending strength characterisation;
d) Implementation of the TRM composite; e) TRM tensile test.
Influence of Fibres Impregnation on the Tensile Response 985
The matrix consists of a hydraulic lime-based mortar whose mechanical charac-

terisation was carried out according to the EN 196-1 [11] (Fig. 1c). The mortar was
characterised by a compressive strength of 9.81 MPa and a flexural strength of
4.33 MPa.
The mechanical characterisation of the composite was carried out by means of
tensile tests performed on TRM coupons. Two series of specimens were considered:
– Non-impregnated Flax TRM: 5 specimens characterised by a non-impregnated flax
textile;
– Impregnated Flax TRM: 5 specimens characterised by the impregnated flax textile.
The specimens were realized by alternating the first layer of mortar, then the textile
ply, making sure the mesh voids to be fully penetrated by the mortar (Fig. 1d), and the
second layer of mortar. The samples were characterised by a thickness of 7 mm, a
width of 60 mm and a length of 500 mm, of which 300 mm in the middle represents
the gauge length. The two edges of the specimens were clamped by gluing aluminium
plates for a length of 100 mm. The gripping system was the clevis type (Fig. 1e).
Tensile tests were carried out by means of universal testing machine with a load
capacity of 10 kN, in displacement control (0.3 mm/min).
Table 1. Physical and mechanical properties of the flax textile (mean values).
Linear density [Tex] 302
n° threads/cm [−] 4.3
thread cross section [mm2] 0.25
Non-impregnated Young’s modulus [GPa] 9.06
strain to failure [%] 6.09
tensile strength [MPa] 341
Impregnated Young’s modulus [GPa] 9.11
strain to failure [%] 3.63
tensile strength [MPa] 296
Figure 2 shows the typical response of the TRM composite subjected to tensile load.
Specifically, the tensile behaviour can be divided in three phases. The so-called Stage I
represents the elastic response of the composite. Then, when the first crack in the
matrix occurs, the Stage II begins: this stage is characterised by a development of
cracks through the specimen length. Finally, the Stage III is characterised by a linear
behaviour, mainly governed by the textile response, up to the rupture of the textile.
Fig. 2. Load – displacement curve of a representative specimen.
An overview of the whole results obtained herein are summarized, in terms of

Load-Displacement, in Fig. 3 and Fig. 4 for both Non-impregnated-Flax TRM and
Impregnated-Flax TRM series, respectively.
Fig. 3. Load – displacement curves for non-impregnated Flax TRM specimens.
The tensile response of Non-impregnated Flax TRM specimens is characterised by

a linear branch up to the occurrence of the first crack, and, in some cases a second crack
occurred as well (see Fig. 3). The failure occurred due to the slipping of the textile
through the mortar in the gripping edge of the specimen. On the other hand, the
Impregnated-Flax TRM specimens show a response characterised by the development
of several cracks (see Fig. 4), and by a drop of the load, corresponding to each crack
Fig. 4. Load – displacement curves for impregnated Flax TRM specimens.
Fig. 5. Average value of the main mechanical parameter for both non-impregnated and
impregnated TRM series of specimens.
occurrence, much lower than the one observed in the reference series (i.e., Non-im-
pregnated Flax TRM).
A more comprehensive analysis of the results is proposed in Fig. 5 in which the
results obtained from the two series are summarized in terms of stress and strain
registered in correspondence of the key transition points between: the Stage I and the
Stage II (i.e. r1 and e1), the Stage II and the Stage III (i.e., r2 and e2), as well as in
terms of maximum stress and the corresponding strain (i.e., rmax and emax).
Concerning the point of transition 1, it is clear that the average value of the stress
attained in the reference series is higher than the one observed when the impregnated
textile was adopted. Such experimental evidence may be due to a less participation to the
strength of the non-impregnated textile in the elastic phase, causing a more uniform
distribution of the stress within the mortar, letting it to achieve a higher value of the load.
With respect to the point of transition 2 the figure highlights that the impregnated
configuration is characterised by a lower value of the parameter e2. Such behaviour
may be explained by referring to the tensile behaviour of the textile. As a matter of the
fact, the non-impregnated textile is characterised by a less stiff response when subjected
to low values of strain. Such phenomenon, typical of plant fibres, tends to disappear in
the impregnated textiles, that are, in fact, characterised by a lower failure strain (see
also Table 1). As a matter of fact, the use of impregnated fibres within the composite
reduces its deformability in the Stage II.
Moreover, Fig. 6 and Fig. 7 show the cracks development during the tensile test of
a representative specimen of each series. It can be observed that the Impregnated-Flax
TRM specimens are characterised by a much higher number of cracks with respect to
the reference series. Specifically, for the Non-impregnated Flax TRM series, one or two
cracks, characterized by an average width of 13 mm (ranging between 9 mm and
18 mm), were observed. Contrarily, the Impregnated Flax TRM series presented a
significant higher number of cracks (from 7 to 9) with an average width of 2.8 mm
(ranging between 2.6 mm and 3.1 mm).
The crack development is mainly related to the matrix strength, textile stiffness, and
textile-to-matrix bond capacity. Being the two series of specimens characterised by the
same mortar, it can be asserted that the use of impregnated textile resulted in a textile
stiffness, and in textile-to-matrix bond behaviour much satisfying of those occurred in
the reference series. As a matter principle, the textile impregnation conferred to the
material a significant improvement leading to a tensile response characterised by a
Fig. 6. Cracks development of a representative specimen of the series non-impregnated

Flax TRM (spec. 3).
Fig. 7. Cracks development of a representative specimen of the series impregnated Flax TRM
(spec. 1).
more distributed crack pattern through the specimen length, hence allowing it to work
as an actual composite system. Moreover, this response led to a failure mode char-
acterised by the rupture of the fibres, allowing to exploit the entire strength of the
adopted textile.
4 Conclusions
The study deals with the influence of the textile impregnation on the tensile behaviour a
Flax TRM system. The main findings of the research are reported as follow:
– the impregnation of the textile, although resulting in a slight reduction of strength,
conferred to the textile a lower deformability;
– the tensile response of the Non-Impregnated-Flax TRMs was characterised by the
development of as maximum 2 cracks before the failure occurring due to the
slipping of the textile through the mortar;
– unlike the reference series, the Impregnated-Flax TRMs showed a tensile response
characterised by the three stages behaviour, and by a failure due to the rupture of the
textile;
– the series of specimens reinforced by impregnated textile showed a crack pattern
developed through the entire length of the specimen, allowing the material to work
as an actual composite system.
The study confirmed the potential in the use of plant fibres as reinforcement in
TRMs, and showed that it is possible to improve its mechanical behaviour by means of
fibres impregnation treatment. Moreover, it paves the way for further investigations
aimed at defining the optimum amount of textile and at optimising the impregnation
treatments, in order to improve even more the mechanical behaviour. Finally, further
investigations are needed in order to study the influence of the impregnation treatment
on the durability of the system, an aspect, the latter, of fundamental importance in view
of long terms applications of the reinforcement system.
Acknowledgements. The present study is part of the activities carried out by the Authors within
the “SUPERCONCRETE Project (www.superconcrete-h2020.unisa.it) funded by the European
Union’s Horizon 2020 Research and Innovation Programme under Grant Agreement No 645704
(H2020-MSCA-RISE-2014), whose financial support is gratefully acknowledged. The authors
gratefully acknowledge the company INNOVATIONS s.r.l. for providing the materials tested in
the experimental research presented in this paper.
References
1. Giacomin, G.: Innovative strengthening materials for the post-earthquake reconstruction of
L’Aquila masonries. In: Proceedings of the 10th International Conference on Structural
Analysis of Historical Constructions, SAHC 2016, LEUVEN, Belgium, 16–26 September
2016
2. Codispoti, R., Oliveira, D.V., Olivito, R.S., Lourenço, P.B., Fangueiro, R.: Mechanical
performance of natural fiber-reinforced composites for the strengthening of masonry.
Compos. Part B Eng. 77, 74–83 (2015)
3. Olivito, R.S., Codispoti, R., Cevallos, O.A.: Bond behavior of Flax-FRCM and PBO-FRCM
composites applied on clay bricks: experimental and theoretical study. Compos. Struct. 146,
221–231 (2016)
4. Ghiassi, B., Razavizadeh, A., Oliveira, D.V., Marques, V., Lourenço P.B.: Tensile and bond
characterization of natural fibers embedded in inorganic matrices. In: Proceedings of 2nd
International Conference on Natural Fibers, Azores/Portugal, 27–29 April 2015
5. Olivito, R.S., Cevallos, O.A., Carrozzini, A.: Development of durable cementitious
composites using sisal and flax fabrics for reinforcement of masonry structures. Mater.
Des. 57, 258–268 (2014)
6. Cevallos, O.A., Olivito, R.S.: Effects of fabric parameters on the tensile behaviour of
sustainable cementitious composites. Compos. Part B Eng. 69, 256–266 (2015)
7. Menna, C., Asprone, D., Durante, M., Zinno, A., Balsamo, A., Prota, A.: Structural
behaviour of masonry panels strengthened with an innovative hemp fibre composite grid.
8. de Carvalho Bello, C.B., Boem, I., Cecchi, A., Gattesco, N., Oliveira, D.V.: Experimental
tests for the characterization of sisal fiber reinforced cementitious matrix for strengthening
masonry structures. Constr. Build. Mater. 219, 44–55 (2019)
9. Mercedes, L., Gil, L., Bernat-Maso, E.: Mechanical performance of vegetal fabric reinforced
cementitious matrix (FRCM) composites. Constr. Build. Mater. 175, 161–173 (2018)
10. Ferrara, G., Pepe, M., Martinelli, E., Toledo Filho R.D.: Influence of an impregnation
treatment on the morphology and mechanical behaviour of flax yarns embedded in hydraulic
lime mortar. Fibers 7(4), 30 (2019)
11. EN 196-1:1994 “Methods of testing cement – Part 1: Determination of strength” European
committee for standardization
Experimental Investigation of Mechanical
Properties of Smart Textile Reinforced
Concrete Pipes
Gozdem Dittel(&), Michelle Wangler, Bastian Maiworm,

and Thomas Gries
Institut fuer Textiltechnik of RWTH Aachen University, Aachen, Germany

goezdem.dittel@ita.rwth-aachen.de
Abstract. Leakages in pipes results in a 35% loss of the total water supplied
worldwide, which is a critical issue given the impact of climate change and
global warming. Therefore, early leakage warning systems have to be developed
in order to reduce the water losses occurring due to cracks and leakages in pipes.
However, conventional pipes available today do not contain any integrated
leakage detection mechanism. Hence, the proposed solution, of using conduc-
tive carbon fibres in the reinforcement as leakage sensors allowing for the fault
and leakage determination, is being developed. This principle has paved the way
for research into sustainable hybrid textile reinforced concrete (TRC) pipe
systems. With the aim of realizing an industrial production method for TRC
pipes, different grid-shaped textile reinforcement structures, with integrated
sensory rovings, are developed for concrete pipes at the Institut fuer Textil-
technik (ITA) of RWTH Aachen University. This work forms the future basis of
an automated pipe production.
The aim of this study is to characterise the mechanical properties of these new
age TRC pipes. For this purpose, lab scale TRC pipes with a length of
l = 500 mm, an outer diameter of do = 300 mm and a wall thickness of
d = 25 mm are casted using these smart hybrid textile reinforcement structures
made by using alkali-resistant (AR) glass and carbon rovings. Thereafter,
mechanical tests for compressive strength of the TRC pipes are carried out
according to the DIN EN 1916 standards. The results are evaluated and com-
pared with each other.
Keywords: Textile reinforced concrete pipe Smart pipe Peak value of

compressive strength
1 Introduction
Due to the ever increasing global population, the need for building new settlements is
constantly rising every year. Therefore, in order to meet this growing need, additional
infrastructural materials and products are required. One such construction element is a
reinforced concrete pipe, which is primarily used for the transportation of fluids.
Reinforced concrete pipes are used in a variety of ways, including inlet and outlet pipes
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_87
992 G. Dittel et al.
to households and all other demands, including those wherein the transport of industrial
fluids is required.
In Germany, there exist approximately 600,000 km of sewage pipes, 45% of which
are made using concrete material [1, 2]. According to the LAWA guidelines, sewage
pipes are designed for a service life of 50-80 years and were mostly built in the 1950s
[3]. Hence, given that these sewage pipes are nearing their end of life, frequent and
more severe damages are being observed. The main reason of failure can be attributed
to corrosion, root in-growth and deformations resulting from operational loading.
Damage caused by corrosion accounts for 10% of all damage types [4]. The high
moisture content inside and outside of the sewage pipes causes corrosion of the rein-
forcement and thus the corrosion of the concrete. During corrosion, the steel rein-
forcement reacts with the elements in its environment and this chemical reaction causes
an increase in volume of the steel reinforcement and as a result spalling of the concrete
structure occurs. The assembly and maintenance of water pipelines involves high costs
in terms of transport, handling, logistics and monitoring.
There are different methods to control damages of sewage pipes, which depend on
the use and size of the sewage pipes. A common method is the wireless TV-inspection,
in which a camera mounted robot is remotely driven through the pipe and the recorded
images are then evaluated using a software algorithm. However, it should be noted that
only 5% of the sewage pipes are regularly monitored, because continuous monitoring is
cost intensive and time-consuming [5].
The geometry of steel reinforced concrete pipes varies depending on the static load
requirements. The cross-sectional shapes of sewage pipes are defined in DIN 4045 as
circular, egg-shaped or mouth-shaped cross-sections [3]. The diameter of steel rein-
forced concrete pipes ranges from DN 300 to DN 1200 [6]. For protection of the steel
reinforcement from corrosion, a minimum thickness of concrete material must be
deposited over the reinforcing bars. The required wall thickness depends on the
nominal diameter and ranges from 71 to 141 mm. In order to maintain the high wall
thicknesses, a large amount of cement is required for concrete production. During
cement production, lime is burned in kilns to produce hardened cement paste and is
therefore responsible for 4-8% of the global CO2 emission [7]. This in turn is an
additional challenge that presents itself and has adversarial effects on the environment.
In order to find a solution for the problems mentioned above, smart textile rein-
forced concrete pipes are realized in the project “Smart Pipe: Development of a textile
reinforced pipe system with integrated monitoring functions”. The hybrid textile
reinforcement consists of alkali-resistant (AR) glass rovings and carbon rovings. Due to
the non-corrosive fibre-based reinforcement system, a large amount of the concrete
material deposition requirements over the reinforcement structure can be reduced by up
to 70%. Thus, the wall thickness of these TRC pipes is scaled down by 70% and thus a
significant decrease in the CO2 emission per concrete pipe is achieved. At the same
time, the electrically conductive carbon rovings act as a leakage detection system in the
concrete structure. Through electrical measurements, it is possible to detect hairline
cracks in these TRC based concrete pipes. The use of Carbon rovings both as rein-
forcement as well as leakage sensors enables an integrated early warning system which
has never been developed before.
Experimental Investigation of Mechanical Properties 993
At the Institute for Textile Technology (ITA) of RWTH Aachen University, various
prototypes of textile reinforced concrete pipes with a length of 50 cm and an outer
diameter of 30 cm have been realized. Two different production processes are devel-
oped to implement the optimal textile reinforcement for the smart TRC pipe; biaxial
warp knitted textile grid and wound textile grid. The optimal smart textile reinforce-
ment has to ensure a leakage detection over the entire pipe surface and at the same time
the required structural performance. In this study, the mechanical properties of the
smart TRC pipes are determined by measuring the peak value of compressive strength.
2 Properties of the Reinforcement Materals and the Concrete

Mixture
Both types of textile reinforcement consist of alkali-resistant (AR) glass and carbon
rovings coated with a styrene-butadiene rubber (SBR) polymer (50% material con-
centration) through a dispersion process. The type of the coating material is Lefasol VL
90/1 and the added cross linker is Lefasol VP 4-5LF, both produced by Lefatex Chemie
GmbH. The PES 167f48 yarns are used in the biaxial warp knitted structure as knitting
thread and in the wound structure as wrapping yarn to ensure an oval shaped cross
section for the AR-glass and carbon rovings. AR-glass rovings of Cem-FIL® 5325
produced by Owens Corning for cement based matrices are used for this study. The
type of the carbon rovings used is SIGRAFIL® C50 T024 EPY 382 produced by SGL
Group. The physical and mechanical properties of roving materials are listed in
Table 1.
Table 1. Properties of the roving materials.

Property AR-glass roving Carbon roving
Filament Diameter [µm] 19 7
Filament tensile strength [MPa] 1700 5000
Modulus of elasticity [GPa] 72 270
Elongation at break [%] 2.4 1.9
Linear density of roving [tex] 2400 1600
Density [g/cm3] 2.68 1.81
The composition of the fine grain concrete material used as a matrix has a com-
pressive strength of 74.2 N/mm2 and a flexural strength of 7.6 N/mm2 [8]. The
properties of this material are given in Table 2.
Table 2. Composition of the fine grain concrete mixture [8].

Content Quantity [kg/m3]
Cement CEM I 52,5 N 490
Fly ash 175
Silica fume (Elkem Microsilica® 940 U) 35
Quartz flour 500
Sand 0,2–0,6 mm 713
Water 280
Superplasticizer 7
3 Production of the TRC Pipe Specimens
The fibre based reinforcing structures are manufactured using two textile production
methods: the biaxial warp knitting process and the filament winding process. In the
following sections, the production methodology for each textile grid is presented. Both
textile structures are given in Fig. 1.
sensory
carbon
rovings
AR-
glass
rovings
30 mm 100 mm
Fig. 1. Biaxial warp knitted (left) and wound (right) textile reinforcing structures.
3.1 Biaxial Warp Knitted Textile Reinforcing Grid Structure

The optimal geometry, material design and coating procedure of the biaxial warp
knitted reinforcement for the smart concrete pipe is determined in the preliminary
research [9–14]. AR-glass rovings are used as the main reinforcement base and eight of
them were replaced by using four carbon roving pairs in the warp direction. The carbon
fibres are providing the reinforcement and are also functional as leakage sensors. The
mesh opening in warp and weft direction is approximately 8 mm. The knitting type is
counterlaid tricot. As discussed above, to ensure the required load transmission
between the individual fibre filaments of the roving and to prevent the reinforcement
from shifting, the textile grid is coated with styrene-butadiene rubber (SBR) with 50%
solid content. The two dimensional textile grid is shaped into a cylindrical form with a
diameter of d0 = 270 mm before the concrete matrix material is added. The overlap
area of the two edges is 50 mm wide. The four sensory carbon rovings are positioned in
the longitudinal direction at every 90° across the entire pipe surface [15, 16].
3.2 Wound Textile Reinforcing Grid Structure

The wound reinforcing grid structure consists of AR-glass rovings in the longitudinal
direction and two parallel wound sensory carbon rovings in the circular direction. All
rovings are entwined using the same knitting thread as in the biaxial warp knitted
structure to achieve a comparable elliptical cross-sectional roving shape. The AR-glass
rovings are positioned longitudinally on a winding core with a diameter of
d0 = 270 mm and thereafter are coated. A sensory carbon roving couple is wound on
top of the AR glass rovings in the transverse direction by a single filament machine and
is also coated to stabilize the reinforcement cage. The mesh opening in both directions
is approximately 8 mm.
The winding process is developed to eliminate the overlap area in the biaxial warp
knitted structure and to ensure a leakage detection across the full-surface area of the
concrete pipe [17].
3.3 TRC Pipe Production

The cylindrical reinforcement is positioned in a mould with an inner diameter of
di = 250 mm and outer diameter of do = 300 mm using spacers to ensure the spatial
accuracy of the reinforcement. Concrete is then poured into the vertical mould by
applying vibration. The smart TRC pipes are demoulded after a 24-hour cure cycle and
placed in a water bath for 6 days at room temperature. The specimens are stored for
further 21 days at room temperature to complete the hydration process and achieve the
final strength of the concrete. Three smart TRC pipes with a length of l = 500 mm and
a wall thickness of d = 25 mm are produced for each textile configuration. Figure 2
shows the schematic design and actual prototype images of both configurations (warp
knitted and wound) of the smart TRC pipes [15–17].
4 Determination of the Peak Value of Compressive Strength
In order to test the mechanical properties of the textile reinforced concrete pipes, peak
compressive strength tests are carried out at IKT- Institut für Unterirdische Infrastruktur
gGmbH, Gelsenkirchen according to DIN EN 1916 [18]. This particular standard is an
industrial norm for the testing of steel reinforced concrete pipes. In principle, the smart
TRC pipes will be subjected to the same loading conditions.
4.1 Experimental Setup

The loading unit applying the required force is located centrally above the test spec-
imen. It comprises of an elastomeric uniform beam which is gradually lowered to the
pipe at a predefined rate. Thus, applying a steadily increasing load without any impact
sensory
carbon
rovings
AR-
glass
rovings
20 cm
Fig. 2. Smart TRC pipes with biaxial warp knitted (right) and wound (left) textile grid
reinforcement.
or shock. The pipe movement constrainers are positioned at an angle of 30° below the
test piece (see Fig. 3).
The test is carried out as it is proposed in the methodology in Annex C of DIN EN
1916. The concrete pipe must achieve a minimum apex compressive strength Fn, which
corresponds to its nominal diameter and strength class. The concrete pipe should
withstand a crack force of Fc = 0.67 * Fn, whereby in the concrete tensile zone, a
surface crack must not exceed a crack width of 0.3 mm over a length of 300 mm or
more. As soon as the first cracks appear on the textile reinforced concrete pipes, it is
measured and documented. Thereafter, the test load is further applied, and the crack
widths are continuously monitored and measured. The test is carried out until a
deformation of 25 mm is reached.
elastomeric
beam
TRC pipe
pipe
supports 30 cm
Fig. 3. Test setup for compressive loading to failure measurement in accordance to the Annex C
in DIN EN 1916
4.2 Results and Analysis

Due to the low wall thickness of 25 mm the loading speed, specified in DIN EN 1916,
is reduced by 50% to 0.1 kN/s. Usually, steel reinforced concrete pipes are loaded up to
the initial crack and then the load is increased further to a stabilized longitudinal crack
of 0.3 mm. However, the pipes tested in this study already developed a crack width of
0.3 mm at the initial crack. Hence, further loading in force control mode was aborted.
A change was made to the displacement control mode and the load was continued at a
speed of 1.0 mm/s. In this way, the maximum loading case could be achieved.
In the pipes with biaxial reinforcement structure, after a force of 12 kN, the first
crack is observed at an average load of 23.2 kN/m in the area of the impost, which after
a few seconds reaches a crack width of 0.3 mm. The pipes are further loaded up to a
deformation of 25 mm. The average maximum load recorded is 35.2 kN/m.
In the pipes with wound reinforcement structure, the first crack occurs at an average
load of 23.4 kN/m in the impost and shortly thereafter a crack width of 0.3 mm is
observed. Up to the average maximum force of 42.8 kN/m, further cracks occur in the
area of the impost.
In both test series, cracks occur in the apex area and concrete spalling occurs in the
sole area. The results of both test series are shown in Fig. 4.
5 Discussion and Conclusions
In principle, crack widths of 0.3 mm already determined during the initial crack are be
classified as ‘too high’. Here, there are already risks regarding the load-bearing capacity
and also the tightness of the pipes. This means that the textile reinforcement merely
50 50
42,8
Cracking Load [kN/m]
40 40 35,2
M aximum load [kN/m]

30 23,2 23,4 30
20 20
10 10
0 0
Warp knitted structureWound structure Warp knitted structure Wound structure
Fig. 4. Comparison of mechanical properties for TRC pipes with bi-axial warp knitted and
wound reinforcement
prevents the complete collapse of the pipes. With respect to the crack pattern, it should
be noted that the crack distribution in pipes with wound reinforcement appears to be
more favorable as compared to pipes with biaxial warp knitted reinforcement. The
pipes with wound reinforcement demonstrated a significantly increased crack behavior
with smaller crack widths than the pipes with biaxial warp knitted reinforcement (see
Fig. 5).
Fig. 5. Average cracking load and maximum load of the pipes with different textile
reinforcement
The resulting crack pattern of the wound reinforcement structure is comparable to

the crack pattern of steel reinforced concrete pipes. Cracks occur in the apex and
bottom area as well as in the area of the impost. Due to the applied load, high moment
and bending loads occur in the apex. After exceeding the concrete tensile strength, the
first crack occurs. The concrete spalling in the sole is caused due to the high shear
stresses between the two support points. This is also a typical behaviour in peak
compressive strength tests for steel reinforced concrete pipes.
Due to the load transfer from apex and sole into the impost, the tensile strength of
the textile-reinforced concrete is exceeded, and a full joint is formed there. This
becomes visible by a main crack. Until the maximum force is reached, a few secondary
cracks occur in the impost area.
The first cracks occurred during the determination of the peak value of the com-
pressive strength represent the mechanical properties of the fine grained concrete
mixture. For TRC structures, it is unusual that the cracks reach the maximum allowed
crack width of 0.3 mm after only a few seconds of the load application. One reason for
this is the low wall thickness of the tested TRC pipes. The pipes were not sized to
match their steel reinforced counterparts. The wall thicknesses of equivalent steel
reinforced pipes range between 70 to 80 mm which is three times thicker than the
tested TRC pipes. Thus if a smart TRC pipe is designed for operational loads com-
parable to their steel reinforced counterparts, a significantly different cracking pattern
will be observed.
It is very unusual that reinforced concrete pipes only show one crack in the impost
area, as in the test of the biaxial reinforcement structure. This manner can be explained
by a possible inhomogeneous force absorption of the textile reinforcement. Two-
dimensional biaxial warp-knitted grids are shaped to a three dimensional cylindrical
form, which built an overlapping area of 5 mm. This change in the reinforcement cross-
section generates a weak point due to the force transmission. The continuous roving
placement in the wound reinforcement structure ensures a homogeneous force
absorption by the rovings and thus an increase of 21% in the maximum load
absorption.
The goal of this paper was to perform a destructive test on textile reinforced
concrete pipes for determining the maximum loading conditions and to compare two
different reinforcement typologies, biaxial warp knitted structure and wound structure.
Given the results of the current experiment, it can be conclusively stated that wound
reinforcement structures are more suitable to concrete pipes than bi-axial warp knitted
structures. Concurrently, from the manufacturing point of view, the textile winding
process corresponds to the manufacturing process of the steel reinforcement for con-
crete pipes.
Future work will consist of sizing the pipes to defined operational loads. The
wound reinforcement structure will be further developed regarding the textile archi-
tecture such as the mesh opening size, the pre-stress of the rovings and the optimal
position of the different roving materials in the grid structure. A comparison to steel
reinforced pipes will be accurately made.
Acknowledgements. The authors would like to thank the Federal Ministry of Education and
Research (BMBF) - Germany for funding the project “SmartPipe - Development of a textile
reinforced pipe system with integrated monitoring functions” and the Project Management
Agency Karlsruhe (PTKA) for the project coordination. We would like to acknowledge that the
compressive tests were carried out at the IKT - Institut für unterirdische Infrastruktur gGmbH,
Gelsenkirchen.
References
1. Breitkopf, A.: Länge des Kanalnetzes in Deutschland im Jahr 2016. Accessed 24 Jan 2020.
https://de.statista.com/statistik/daten/studie/152743/umfrage/laenge-des-kanalnetzes-in-deut
schland-im-jahr-2007/. (20th December 2018)
2. Stein, D., Stein, R.: Instandhaltung von Kanalisationen, Aufbau und Randbedingungen von
Kanalisationen, Rohrwerkstoffe und Ausbildung der Rohrverbindungen, Übersicht über
Werkstoffe und Rohrverbindungen (1998). Accessed 24 Jan 2020. https://www.unitracc.de/
know-how/fachbuecher/instandhaltung-von-kanalisationen/aufbau-und-randbedingungen-
von-kanalisationen/rohrwerkstoffe-und-ausbildung-der-rohrverbindungen/uebersicht-ueber-
werkstoffe-und-rohrverbindungen
3. Stein, D., Stein, R.: Instandhaltung von Kanalisationen, Aufbau und Randbedingungen von
Kanalisationen, Querschnittsformen und–abmessungen (1998). Accessed 24 Jan 2020.
https://www.unitracc.de/know-how/fachbuecher/instandhaltung-von-kanalisationen/aufbau-
und-randbedingungen-von-kanalisationen/querschnittsformen-und-abmessungen
4. Statista Research Department: Abwasserkanäle - Verteilung der Schäden in Deutschland
2013, 7th November 2019. Accessed 24 Jan 2020. https://de.statista.com/statistik/daten/
studie/456149/umfrage/verteilung-der-festgestellten-schaeden-an-abwasserkanaelen-in-
deutschland/
5. Bütow, E.: Umweltbundesamt, November 2001. Accessed 24 Jan 2020. https://www.
umweltbundesamt.de/sites/default/files/medien/publikation/long/2052.pdf
6. Betonwerk Bieren GmbH, Datenblatt_1.3_Stahlbetonrohre_ohne_Fuss. Accessed 24 Jan
2020. https://betonwerk-bieren.de/produkte/?gclid=EAIaIQobChMIwu3r0dmc5wIVhIxRC
h2lnQNsEAAYASAAEgLuH_D_BwE
7. Kretschmer, A.: Klimabilanz der Zementindustrie, 25th March 2019. Accessed 24 Jan 2020.
https://www.chemietechnik.de/klimabilanz-der-zementindustrie/
8. Brockmann, T.: Mechanical and fracture mechanical properties of fine-grained concrete for
TRC structures. In: Gross, C.U., (ed.) Advances in Construction Materials, pp. 119–129.
Springer, Berlin (2007)
9. Perry, G., Dittel, G., Gries, T., Goldfeld, Y.: Mutual effect of textile binding and coating on
the structural performance of TRC beams. Journal of Materials in Civil Engineering (2020,
in print)
10. Perry, G., Dittel, G., Gries, T., Goldfeld, Y.: The effect of textile configuration on the
monitoring capabilities of smart carbon-based TRC elements to detect water infiltration,
submitted for publication (2020)
11. Quadflieg, T., Goldfeld, Y. Dittel, G., Gries, T.: New age advanced smart water pipe systems
using textile reinforced concrete. In: 15th Global Conference on Sustainable Manufacturing,
Technion-IIT, Haifa, Israel, 25–27 September 2017
12. Goldfeld, Y., Perry, G., Dittel, G., Gries, T.: Development of a textile reinforced pipe system
with integrated monitoring function (SmartPipe). In: German-Israeli Cooperation in Water
Technology Research Status Conference 2019, Dresden, Germany, 24th–25th September
2019
13. Goldfeld, Y., Perry, G.: Electrical characterization of smart sensory system using carbon
based textile reinforced concrete for leakage detection. Mater. Struct. 51(17), 1–17 (2018)
14. Goldfeld, Y., Perry, G.: A-R glass/carbon-based textile reinforced concrete elements for
detection water infiltration within cracked zones. Struct. Health Monitor. (2018). https://doi.
org/10.1177/1475921718808223
15. Dittel, G., Heins, K., Gries, T.: Development and design of smart textile reinforcement for
concrete pipes. In: Proceedings of the ACI Convention, Ohio, USA, 20–24 October 2019
16. Heins, K.: Entwicklung und Realisierung von Konzepten zur Herstellung intelligenter
textilverstärkter Betonrohre, Aachen (2019)
17. Maiworm, B.: Filament wound smart textile reinforcement for leakage detection in concrete
pipes, Aachen (2019)
18. DIN1916: Rohre und Formstücke aus Beton, Stahlfaserbeton und Stahlbeton; Deutsche
Fassung EN 1916:2002, April 2003
UHPFRC, SHCC and ECC
Full-Scale Construction Test for Improvement
of RC Void Slab Bridges Using
UHPFRC – Part 1: Experimental Test Plan
Tohru Makita1(&), Yuji Watanabe2, Shuji Yanai2,

and Hirokazu Kitagawa1
1
Central Nippon Expressway Company Limited, Nagoya, Japan
t.makita.ab@c-nexco.co.jp
2
Kajima Corporation, Tokyo, Japan
Abstract. Kajima Corporation and NEXCO Central started a joint research and
development (R&D) project to develop a method for improving existing RC
void slab bridges using cast-in-situ UHPFRC in 2016 and as the final investi-
gation of the R&D project a full-scale construction test was conducted. This
paper is the first one of the two papers dealing with the construction test.
UHPFRC mix used in the construction test is called AFt-UHPFRC because it is
characterised by its matrix densified by controlled ettringite (AFt) formation.
UHPFRC was produced in a batching plant built in a testing field of that
production capacity is 3.0 m3 per hour. A slab-on-ground (SOG) structure
modelling top part of RC void slab bridge decks was built in the testing field and
UHPFRC was cast on top of the SOG structures where UHPFRC was trans-
ported and placed by a wheel loader and spread/compacted/finished by newly
developed construction equipment.
Keywords: UHPFRC Strengthening Durability enhancement Bridges
1 Introduction
In Japan, the first expressway was put in service in 1963 and since then about 9,500 km
long expressways have been built so far. Central Nippon Expressway Company
Limited (NEXCO Central) operates and manages about 2,000 km long expressways in
the central part of Japan, 40% of which have been in service for more than 40 years and
average service period of the expressways is now approximately 30 years. Although in-
service period of 30 years is not so long considering the fact that expected service life
of civil engineering structures is over 100 years, recent years have seen growing
number of damaged and deteriorated bridges in expressways of Japan and the number
of bridges in such conditions will be getting larger in the coming decades. In order to
address this issue, extensive expressway renewal project was launched in 2015. In the
project, replacement and improvement (strengthening and durability enhancement) of
bridge decks and girders are planned to be undertaken among others and total of
approximately one trillion yen will be spent for the renewal of NEXCO Central’s
expressway structures during 15 years.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_88
1004 T. Makita et al.
Reinforced concrete (RC) void slab bridge (Fig. 1) is the most common type of
concrete bridges in expressways managed by NEXCO Central, constituting approxi-
mately 25% of the NEXCO Central’s bridges. About 95% of the RC void slab bridges
were designed according to old design codes and strengthening is necessary if those
bridges don’t fulfil today’s traffic load requirements. In addition, waterproof mem-
branes had not been applied as standard to the NEXCO Central’s bridges until 1998
and concrete bridge decks built before that time have been seriously exposed to
deleterious environmental influences (e.g. de-icing salts). Top surface of old RC void
slab bridge decks are often found to be chloride contaminated reaching the depth of top
steel rebars. The above-mentioned conditions motivated the improvement of damaged
and deteriorated RC void slab bridges in the expressway renewal project where
chloride contaminated top surface concrete of RC void slab bridge decks is removed
and refilled with cementitious material followed by application of waterproofing
membrane on the top surface.
12,500
mid-span over pier
600
900
2,205 7,740 2,555

[unit: mm]
Fig. 1. Cross sections of an RC void slab
As the refilling material, normal strength concrete (NSC) or fibre reinforced con-
crete (FRC) made of NSC containing 100 kg/m3 (1.3vol.-%) of steel fibres has been
conventionally used. However, in order to improve the load carrying capacity of RC
void slab bridges (Fig. 2a), the deck thickness often has to be increased with additional
rebars (Fig. 2b), resulting in increase of superstructure self-weight. Consequently, it is
usually necessary to strengthen substructures especially for seismic action. Further-
more, expansion joints need to be replaced for increased height of bridge decks.
By using Ultra-High Performance Fibre Reinforced cement-based Composites
(UHPFRC) instead of conventional NSC or FRC, unwanted increase of the deck
thickness can be avoided due to its excellent mechanical properties: replacement of top
surface concrete of bridge decks with UHPFRC is sufficient for improvement of the
load carrying capacity (Fig. 2c). Moreover, it is not necessary to apply waterproofing
membrane on top surface of bridge decks because UHPFRC function as protective
layer due to its very low permeability. In order to develop a method for improving
existing RC void slab bridges using cast-in-situ UHPFRC, Kajima Corporation and
NEXCO Central started a joint research and development (R&D) project in 2016. All
laboratory tests planned in the R&D project were completed at the end of 2018 and as
the final investigation of the R&D project a full-scale construction test was carried out.
Full-Scale Construction Test for Improvement of RC Void 1005
FRC UHPFRC
Void
Void Void
RC RC RC
(a) (b) (c)
Fig. 2. Schematic comparison of cross sections of (a) RC void slab and RC void slab
strengthened with (b) FRC and (c) UHPFRC
This paper is the first one of the two papers dealing with the construction test, pre-
senting the experimental test plan where motivation/objectives and program of the test
are detailed.
2 Motivation and Objectives of Full-Scale Construction Test
In laboratory tests, UHPFRC was manufactured with small mixer (100 litre twin-shaft
forced action mixer or pan type mixer) and poured manually in specimen moulds.
Besides, UHPFRC specimens were cured in perfectly controlled constant temperature
and humidity rooms. However, when UHPFRC is applied to real bridges, in particular
multi-span and large bridges which comprise a majority of expressway bridges in
Japan, UHPFRC is supposed to be manufactured with larger mixer and
placed/spread/compacted/finished with construction equipment for work efficiency
enhancement. Therefore, it is necessary to understand material properties and beha-
viour of UHPFRC that is manufactured and cast in the same condition as real con-
struction work. In order to meet this need, the full-scale construction test was planned.
3 UHPFRC Densified by Controlled Ettringite Formation

(AFt-UHPFRC)
UHPFRC mix used in the construction test

is characterised by its matrix densified by
controlled ettringite (AFt) formation; thus, it
is called AFt-UHPFRC. Microstructure of
the AFt-UHPFRC matrix is basically
formed by decreasing water/binder ratio
using spherical pozzolan particles and
superplasticiser and packing ultrafine parti-
cles optimally. In addition to that, numerous
needle-shaped ettringite crystals of 1 to
2 µm length (Fig. 3) fill micropores of
hydration structure together with inert and
Fig. 3. Needle-shaped ettringite crystals
Table 1. Composition of AFt-UHPFRC

Component Mass Remarks
[kg/m3]
Portland cement 927
Premixed material 360 pozzolanic materials, ettringite formation
additives
Sand 905 crushed sand, dmax 2.5 mm
Steel fibre 235.5 3.0 vol.%, l = 15 mm, d = 0.2 mm
Superplasticiser 36
Shrinkage reducing 12.9
admixture
Defoaming agent 6.4
Water* 195 W/B = 0.152
* including water in superplasticiser
reactive fine fillers. The AFt-UHPFRC mix composition is shown in Table 1. The
premixed material contains silica fume (mean particle diameter of 0.2 lm), fly ash
(mean particles diameter of 3 lm) and additives allowing ettringite to form in a con-
trolled manner. The mix has crushed sand of maximum diameter of 2.5 mm and
3.0vol.-% of steel fibres with length of 15 mm and diameter of 0.2 mm.
20
10-60μm
18 17.0 1-10μm
16 0.5-1μm
14 13.5 100-500nm
12.3 12.2 10-100nm
12
Porosity (%)
10.6 6-10nm
10 9.28 3-6nm
7.94
8
6.07
6
4.38
4
2.66
2
0
18 MPa 44 MPa 64 MPa 73 MPa 99 MPa 115 MPa 133 MPa 165 MPa 193 MPa 215 MPa
15 h 18 h 21 h 24 h 2 days 3 days 7 days 28 days 20 h 5 years
20 °C 85 °C Cast
Steam in-situ
Fig. 4. Evolution of porosity and pore size distribution of AFt-UHPFRC
Figure 4 shows a result of long-term measurement of porosity and pore size dis-
tribution of the AFt-UHPFRC cured under room conditions (20°C and 60% RH). As
favourable effect of ettringite crystal growth, porosity and pore size gradually reduces
as time proceeds: in 28 days the porosity decreases from 17.0% to 6.07% and the mode
of pore size distribution changes from 10–100 nm pores to 3–6 nm pores. As a ref-
erence the porosities and pore size distributions of steam-cured Aft-UHPFRC and Aft-
UHPFRC cured outdoors are indicated in Fig. 3, from which it is understood the
porosity of the Aft-UHPFRC is lowered significantly in five years even without steam
treatment.
4 Test Program
4.1 Slab-on-Ground Structure Modelling Top Part of an RC Void Slab
Bridge Deck
Since building a mock-up RC void
slab bridge is impractical, a slab-on-
ground (SOG) structure modelling
top part of RC void slab bridge decks
was built in a testing field where NSC
of 30 MPa strength was used
(Fig. 5). The width and length of the
SOG structure were 5 m and 62 m,
respectively. The SOG structure was
divided into six zones and construc-
tion conditions were varied for each
zone (Table 2). Top surface of the
SOG structure was roughened with Fig. 5. SOG structure for construction test
water jets for interfacial bonding
between UHPFRC and concrete. Assuming that top 10 cm surface concrete of RC void
slab bridge decks is replaced with UHPFRC, it was planned that 10 cm thick UHPFRC
layer is cast on top of the SOG structure except zone 4 where UHPFRC layer thickness
was increased to 15 cm in order to investigate the influence of layer thickness on the
compactability of thixotropic UHPFRC.
Steel rebars were arranged on top of the SOG structure, which, however, did not
model fully the arrangement of steel rebars of real bridges; yet, densely arranged steel
rebars were partly modelled in order to check if thixotropic UHPFRC properly fill
narrow rebar spacing and gaps between rebars and concrete substrate. On the basis of
the expressway design manual [1] in that combined slope of expressway roads is
prescribed to range from 2% to 9%, top surface of the SOG structure was sloped
transversely either 2% or 9% (Fig. 6). An epoxy bonding adhesive was applied to some
part of top surface of the SOG structure to investigate the performance of the adhesive
for interfacial bonding between UHPFRC and concrete. The amount of the adhesive
was 1.2 kg/m2.
Table 2. Test parameters and points of interest in construction test

Zone Construction conditions Mix Points of interest
UHPFRC Slope Adhesive* Compaction Fibre
thickness (%) (kg/m2) Energy (vol.-%)
(cm)
0 10 2 1.2 (Trial) 3.0 • Compactability at various
1 Minimum compaction energy
2 1.2 Maximum 3.0 • Fillability
2.0 • Interfacial bonding
3 – 3.0 Assume loss of slump flow
(limit of fluidity/fillability)
4 15 Influence of thickness on
compactability
5 10 9 3.0 Slope torelance and workability
* applied to part of top surface
5,500
250 5,000 250 [unit :mm]
74 74 135 165 8×150=1,200 135 135 135 14×150=2,100 74 74
102 100165 135 165 165 165 100 102
φ13 φ13
φ32
570 80
650
φ13
φ13
120 80
200
φ13
φ13 φ13 φ13 φ13
Fig. 6. Cross section of the slab-on-ground structure transversely sloped 9%
4.2 Production and Casting of UHPFRC

A batching plant was built at the testing field (Fig. 7a) and a twin-shaft forced action
mixer with a capacity of 3.0 m3 was used for producing UHPFRC. Except sand and
steel fibre, UHPFRC components needed for each batch were measured and put in the
mixer automatically; sand and steel fibre were measured and put in the mixer manually
(Fig. 7b). 1.5 m3 of UHPFRC was manufactured for each batch, taking 30 min. Steel
fibre content was 3 vol.-% except for UHPFRC cast in part of zone 2 whose fibre
content was reduced to 2 vol.-% in order to see the influence of fibre content on the
consistency of fresh UHPFRC. Thixotropy was conferred on UHPFRC by adding
mineral-based inorganic powder the amount of which was varied depending on the top
surface slope of the SOG structure and the consistency of fresh UHPFRC. UHPFRC
was transported and placed by a wheel loader (Fig. 7c) and spread/compacted/finished
by new construction equipment that was developed by remodelling a concrete paving
(a) (b)
(c) (d)
Fig. 7. Production and casting of UHPFRC: (a) batching plant, (b) steel fibres manually put in
the mixer, (c) wheel loader and construction equipment, (d) application of sheet membrane
spreader (Fig. 7c). Moving speed of the construction equipment was kept constant to
be 0.5 m/min. Vibration frequency of screed mounted on the construction equipment
was changed so as to understand proper vibration for compacting thixotropic UHPFRC.
In order to prevent plastic shrinkage cracking after casting, sheet membrane curing was
applied on top of UHPFRC immediately after the surface finish (Fig. 7d) and the sheet
membrane was removed the next day.
4.3 Evaluation of Construction Test

The construction test was evaluated during and after the construction test by per-
forming visual observation, measuring the behaviour of UHPFRC and the SOG
structure and examining properties of UHPFRC. Table 3 lists UHPFRC property tests
performed as part of the evaluation of the construction test.
Table 3. List of UHPFRC property test

Property Test method Remarks
Fresh state properties Temperature JIS B 7411
Flow JIS R 5201
Air content JIS A 1128
Mechanical Compressive strength JIS A 1108 Modulus of elasticity
properties Tensile cracking JIS A 1113 Splitting tensile test Direct
strength – tension test
Tensile strength – Direct tension test
Behaviour of Deformation – Mould strain gauge
composite structure embedded in UHPFRC
Interfacial bond JIS A 6909 In-situ test
strength ASTM C 1583
Transport properties Air permeability SIA 262/1 Annex Torrent method
E
Chloride diffusivity JSCE Alternate immersion test in
recommendation salt solution
[2]
Porosity and pore size Mercury
distribution intrusion method
During the construction test, vibration of the SOG structure was measured using
piezoelectric accelerometers installed on top surface of the SOG structure in order to
investigate the influence of frequency of screed vibration and UHPFRC layer thickness
on vibration transmission which was considered to index the degree of compaction of
thixotropic UHPFRC. In addition, shrinkage deformation of UHPFRC cast on top of
the SOG structure was measured with mould strain gauges, which was started imme-
diately after casting of UHPFRC. Fresh UHPFRC properties were checked per batch by
measuring temperature, slump flow and air content and optimal quality control of
UHPFRC on site is also investigated. In addition to material-related measurement and
testing, in order to understand the labour cost, the basic production rate of all works
performed for casting UHPFRC, from producing to curing, was examined. Slope tol-
erance of thixotropic UHPFRC was evaluated qualitatively by visual observation.
After the construction test, the compressive strength, tensile cracking strength and
tensile strength of UHPFRC were tested where test specimens were fabricated using
UHPFRC from one batch of each day (the construction test was four days long).
Interfacial bonding strength between UHPFRC and concrete was also tested by per-
forming in-situ pull-off tests at two locations for each zone. Protective function of
UHPFRC was evaluated by investigating air permeability, chloride diffusivity and
porosity/pore size distribution of UHPFRC. Air permeability testing was carried out at
one location per zone using Torrent method on site. Chloride diffusivity was deter-
mined according to Japanese recommendations for design and construction of
UHPFRC [2] where alternate immersion tests in salt solution were conducted on a
UHPFRC core drilled from zone 0. Porosity and pore size distribution were measured
for UHPFRC cores drilled from each zone except zone 0 and 3 by using a mercury
intrusion porosimetry.
5 Conclusions
This paper presents the program of the full-scale construction test conducted as the final
investigation of an R&D project to develop a method for improving existing RC void
slab bridges using cast-in-situ UHPFRC. In the construction test, a batching plant was
built in a testing field and 1.5 m3 of UHPFRC was produced per batch every 30 min.
UHPFRC was cast on top of a slab-on-ground structure modelling top part of RC void
slab bridge decks where UHPFRC was transported and placed by a wheel loader and
spread/compacted/finished by newly developed construction equipment. Construction
conditions were varied and the construction test was evaluated during and after the test
by measurement of specimen behaviour and investigation of material properties.
Evaluation results of the construction test are described in the second one of the two
papers dealing with the construction test [3].
References
1. Central Nippon Expressway Company Limited: Design Manual Volume 4: Road Geometry –
Mainline Geometry, Nagoya, Japan (2017)
2. Japan Society of Civil Engineers: Recommendations for design and construction of ultra high
strength fiber reinforced concrete structures – draft, Tokyo, Japan (2004)
3. Watanabe, Y., Yanai, S., Makita, T., Kitagawa, H.: Full-scale construction test for
improvement of RC void slab bridges using UHPFRC – part 2: test results. In: RILEM-fib
X International Symposium on Fibre Reinforced Concrete, BEFIB2020, Valencia, 21–23
September 2020
Full-Scale Construction Test
for Improvement of RC Void Slab Bridges
Using UHPFRC – Part 2: Test Results
Yuji Watanabe1(&), Shuji Yanai1, Tohru Makita2,

and Hirokazu Kitagawa2
1
Kajima Corporation, Tokyo, Japan
watanyuj@kajima.com
2
Central Nippon Expressway Company Limited, Nagoya, Japan
Abstract. In Japan, a growing number of damaged and deteriorated bridges on

expressways have been seen in recent years, and the number of bridges in such
conditions is set to increase in the coming decades. In order to address this issue,
an extensive expressway renewal project was launched in 2015. Research and
development have begun upgrading bridge decks by utilizing UHPFRC where
either overlaying UHPFRC or replacing top surface concrete with UHPFRC is
conducted, which leads to an increase in bridge deck stiffness. A full-scale
UHPFRC casting test was carried out at the final stage of the research and
development project. This paper is the second of the two papers regarding the
UHPFRC casting test. Several evaluations of the construction test are presented
regarding the compaction of thixotropic UHPFRC, the bonding properties of
UHPFRC with existing slabs, and the transport properties of UHPFRC.
Keywords: UHPFRC Strengthening Durability enhancement Bridges
1 Introduction
Over the last decade, damaged and deteriorated bridges have been increasingly
observed in Japan’s expressways. Typical examples of those damaged and deteriorated
bridges include reinforced concrete (RC) void slab bridges in which top surface con-
crete is deteriorated due to de-icing agents. Bridge deck overlays are most often used
on existing bridges when their decks require rehabilitation. Overlays using Steel Fibre
Reinforced Concrete (SFRC) can also be employed as a preventative measure for
deterioration on existing decks that are in good structural condition; however, there is a
possibility that the top surface of bridge decks may perhaps deteriorate again due to the
repeated loading of heavy vehicles, severe environmental conditions and improper
construction methods. Moreover, the SFRC overlay thickness could reach to more than
100 mm in order to attain the required load bearing capacity. Furthermore, it is nec-
essary to strengthen the other members of bridges in order to carry the increased self-
weight of overlaid bridge decks.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_89
Full-Scale Construction Test for Improvement of RC 1013
Following the above-mentioned situation, the development of a new bridge deck

overlay method has begun for the expressway renewal project, where the strengthening
of bridge decks is carried out.
The new method uses Ultra-High Performance Fibre Reinforced cement-based
Composites (UHPFRC) as overlay material [1]. UHPFRC has very low permeability
against liquid/gas and high resistance to freezing and thawing action. In addition,
UHPFRC is known for its high strength and high elastic modulus. By using UHPFRC
instead of SFRC, the overlay thickness can be made thinner, and an increase of the
concrete bridge deck thickness can even be made unnecessary by replacing a certain
depth of the top surface concrete with UHPFRC (Fig. 1) [2]. Moreover, UHPFRC
functions, such as the protective layer and application of waterproofing material, is not
needed. Thus, the application of UHPFRC to concrete bridge decks for upgrading is an
efficient and effective method because the increase of the load bearing capacity and
enhancement of the durability are achieved simultaneously.
(a) RC void slab and RC void slab strengthened with

(b) SFRC and (c) UHPFRC
Fig. 1. Schematic comparison of cross Fig. 2. Full-scale construction test

sections
Table 1. Test conditions

Zone Construction conditions Mix Points of interest
UHPFRC Slope Adhesive* Compaction Fibre
thickness (%) (kg/m2) energy (vol.
(mm) %)
0 100 2 1.2 (Trial) 3.0 • Compactability at various
1 Minimum compaction energy
2 1.2 Maximum 3.0 • Fillability
2.0 • Interfacial bonding
3 – 3.0 Assume loss of slump flow
(limit of fluidity/fillability)
4 150 Influence of thickness on
compactability
5 100 9 3.0 Slope torelance and
workability
* applied to part of top surface
1014 Y. Watanabe et al.
Table 2. List of UHPFRC tests

Property Test method Remarks This
paper
Fresh state Temperature JIS B 7411 –
properties Flow JIS R 5201 〇
Air content JIS A 1128 〇
Mechanical Compressive JIS A 1108 Modulus of elasticity 〇
properties strength
Tensile JIS A 1113 Splitting tensile test 〇
cracking New method [4] (cylinder) Static tensile tests
strength
Tensile New method [4] Static tensile tests 〇
strength
Behaviour of Shrinkage – Mold strain gauge embedded 〇
composite structure strain in UHPFRC
Interfacial JIS A 6909 On site 〇
bond strength ASTM C 1583
Transport Air SIA 262/1 Torrent method –
properties permeability Annex E
Chloride JSCE Alternate immersion test in –
diffusion recommendation salt solution
coefficient
Porosity Mercury Core specimen 〇
intrusion
method
Table 3. Mix proportion of AFt-UHPFRC

Component Mass Remarks
(kg/m3)
Portland cement 927
Premixed materials 360 Pozzolanic material, ettringite formation
additives
Sand 905 Crushed sand, dmax < 2.5 mm
Steel fibre** 235.5 3.0 vol.%, d = 0.2 mm, l = 15
Superplasticiser 36
Shrinkage reducing 12.9
admixture
Defoaming agent 6.4
Water* 195 W/B = 0.152
* including water in superplasticizer
**not including in the unit volume
The authors have conducted various detailed laboratory studies after confirming the
feasibility of the overlay method and identifying issues of the construction method
through small-scale construction tests. However, when UHPFRC is applied to real
bridges, in particular multi-span and large bridges that constitute most of expressway
bridges in Japan, UHPFRC is supposed to be manufactured with a larger mixer and
placed/spread/compacted/finished with construction equipment for the enhancement of
work efficiency. Therefore, it is necessary to understand material properties and the
behaviour of UHPFRC that is manufactured and cast under the same conditions as real
construction. In order to meet this need, a full-scale construction test was planned and
carried out (Fig. 2) [2]. This paper is the second of the two papers dealing with the
construction test and presents evaluation results of the construction test.
2 Evaluation Results of Construction Test
The outline of the experimental plan and a slab-on-ground (SOG) structure is as shown
in the first paper “PART1: TEST PLAN” [2]. The SOG structure was divided into six
zones, and construction conditions were varied for each zone (Table 1). Table 2 shows
material testing that is mainly performed after the construction test. In this paper,
results of (1) fresh state property tests, (2) mechanical property tests, (3) deformation
measurement, (4) vibration/compaction energy measurement, (5) bond strength and
(6) porosity measurement are presented.
2.1 Fresh State Properties

The UHPFRC mix used in the test is characterized by its matrix, which is densified by
controlled ettringite (AFt) formation; thus, it is called AFt-UHPFRC [3]. Table 3 shows
the mix proportion of AFt-UHPFRC. The fresh state properties of AFt-UHPFRC were
adjusted according to the slope of the SOG slab and the compaction energy of the
paving machine, and a flow value suitable for thixotropic AFt-UHPFRC was deter-
mined. Figure 2 and Fig. 3 show the results of the mortar flow test performed in
compliance with JIS R 5201. It was found that when the flow value of UHPFRC falls
within a range of 200 ± 25 mm, it is slope tolerant. UHPFRC temperature was varied
depending on atmospheric temperature. The lowest was recorded to be 20.5 °C in the
morning, and the highest was recorded to be 31.5 °C at noon. Air content in UHPFRC
was almost constant, being approximately 4 vol.% where the maximum and minimum
value was 4.4 vol.% and 3.0 vol.%, respectively.

Table 4 shows the strength test results of specimens cured at an outdoor environment
as with the SOG structure. At the age of 28 days, the compressive strength reached over
150 MPa, and the cracking strength reached over 8 MPa. It is notable that the com-
pressive strength reached over 120 MPa even at seven days. Rapid hardening property
is conferred to the AFt-UHPFRC by controlled ettringite formation, and the property is
considered favorable for construction with time constraints.
In order to investigate the tensile property of the UHPFRC, direct tension tests were
performed. Figure 4 shows specimen geometry and the test set-up configuration. The
specimen was 400 mm long and 100 mm wide with varying thickness (dog-bone
shaped) to make fractures occur within the 100 mm long tapered central part of the
specimen. The thickness of the central and end parts of the specimen were 40 mm and
100 mm, respectively; moreover, there were 90 mm long transitional zones between
the central and end parts. The test set-up used for the tests was developed in the
previous study [4]. Stress deformation relationships obtained from the static tensile
tests are shown in Fig. 4.
Fig. 3. Thixotropic formulation (Flow) Fig. 4. Fresh state property (Flow)
The averages of elastic limit strength and ultimate tensile strength were 12.1 MPa
and 12.8 MPa, respectively. In the average stress-deformation curve of UHPFRC,
strain-hardening is hardly observed.
This is probably because the UHPFRC specimens didn’t have sufficient amount of
fibres allowing for strain-hardening. This result is different from the result of the direct
tension tests using specimens fabricated from UHPFRC produced by a small mixer and
cured in perfectly controlled constant temperature and humidity rooms. It is also
possible that the method of sampling the specimen affected the fibre orientation. This
discrepancy will be investigated further.
Prestressing bar (φ23)
100 100 Jig 1

60
Jig 2
90
Jig 3
40
100
400
90
60
[unit: mm]
Fig. 5. Static tensile tests (Fibre 3.0 vol. %)

Fig. 6. Strain measurement results (shrinkage strain)
2.3 Deformation
Figure 5 shows UHPFRC strain measurement results. Strain measurement commenced
immediately after casting. The contraction of UHPFRC was caused by autogenous
shrinkage and dry shrinkage. When comparing the strain values of the UHPFRC cast
on top of the SOG structure with the ones of prism specimens, the former is smaller
than the latter. This is thought to be because contraction of the UHPFRC was restrained
more significantly by concrete substrates and steel rebar (higher degree of restraint) and
mitigated by creep of the UHPFRC. The strain values and evolution of the UHPFRC
layers on top of the SOG structure were the same regardless of thicknesses between
100 mm and 150 mm.
Table 4. Strength test results (Fibre 3.0 vol.%)

Property Material age
7 days 28 days 91 days 257 days
Compressive strength (MPa) Ave. 131 164 184 189
Max. 137 177 196 198
Min. 120 144 166 181
Elastic modulus (GPa) Ave. 41 45 47 46
Cracking strength Splitting tensile test (cylinder) Ave. 7.6 8.7 9.3 11.3
(MPa) Max. 9.3 9.9 10.1 11.8
Min. 5.8 6.4 8.5 10.7
Static tensile tests [4] Ave. - - - 12.1
Max. - - - 12.6
Min. - - - 11.6
Tensile strength Static tensile tests [4] Ave. - - - 12.8
(MPa) Max. - - - 14.6
Min. - - - 11.7
2.4 Compaction Energy

Vibration of the SOG structure was measured for the purpose of understanding that the
construction equipment can compact thixotropic UHPFRC so that proper interfacial
bonding between the UHPFRC and concrete is achieved. The accelerometer was
installed at the end of the SOG structure farthest from the screed vibrator so that the
energy transmitted from the vibrator was minimum.
Figure 6 shows the measurement results of acceleration. In the SOG structure,
vibration frequency of vibrating screed was increased by the same percentage as the
UHPFRC layer thickness: a 100 mm thick UHPFRC layer was screeded and consol-
idated by 2,000 (Minimum) vibrations per minute, while a 150 mm thick UHPFRC
layer was screeded and consolidated by 3,000 (Maximum) vibrations per minute.
Acceleration of the specimen with a 150 mm thick UHPFRC layer was larger than the
specimen with a 100 mm thick UHPFRC layer from which it might be said that the
compaction energy produced by screed vibration transmits to UHPFRC-concrete
interface irrespective of UHPFRC layer thickness.
Fig. 7. Acceleration measurement result
Fig. 8. Pull-off test of the UHPFRC cast on top of the SOG structure
2.5 Bond Strength

The interface bonding strength between UHPFRC and concrete was assessed on site using
the pull-off test method specified in ASTM C 1583 (Fig. 7). A partial core was drilled
through the bonded material and into the substrate material; furthermore, a steel pull-off
disc that is approximately 80 mm in diameter was bonded at the desired test location.
Table 5 shows the bond test results. Failure occurred mostly at the substrate con-
crete irrespective of application of adhesive on the UHPFRC-concrete interface. In
addition, the bond strength determined from all tests was almost the same (about
2.8 MPa) regardless of the UHPFRC thickness, tested location in the SOG structure
(center or edge) and imposed compaction energy.
2.6 Porosity of Core Specimen

Figure 8 shows measurement results of porosity and pore size distribution of the top,
medium and bottom part of the UHPFRC layer. The specimens were made from cores
taken from the SOG structure 91 days after UHPFRC casting and cured at 20 °C and
60% RH for about 270 days. Porosity of all specimens was lower than approximately
6% due to the effect of ettringite crystal growth over time (Fig. 9).
Table 5. Results of bond tests
In all zones, the porosity of the bottom part of the UHPFRC layer was smaller than
that of the top part of the UHPFRC layer, which is brought about by the decrease of
pores between 10 to 100 nm in diameter. This is probably explained by the fact that
more extensive hydration occurred in the bottom part of the UHPFRC layer than in the
top part because the SOG structure was subjected to long-term drying conditions after
sheet curing at the early age, and water was lost at the top part of the UHPFRC layer.
Fig. 9. Porosity evolution of UHPFRC (After 270 days)
From the porosity of UHPFRC of Zone 1, 2 and 5 where UHPFRC layer thickness
is 100 mm, it might be said that the higher the flow value of UHPFRC or the frequency
of the vibrating screed, the lower the UHPFRC porosity will be. However, further
investigation is necessary regarding this tendency.
3 Conclusions
This paper is the second of the two papers dealing with the UHPFRC casting test where
test results are presented. Test results are summarised as follows:
• The fresh state properties of UHPFRC were adjusted according to the slope of the
GOG structure and the compaction energy of the construction equipment and a flow
value (200 ± 25 mm) suitable for thixotropic UHPC was determined.
• At the age of 28 days, the compressive strength reached over 150 MPa, and the
cracking strength determined by splitting tensile tests reached over 8 MPa. Aver-
ages of the cracking strength and ultimate tensile strength were determined to be
12.1 MPa and 12.8 MPa, respectively, from the direct tension test.
• The shrinkage strain of UHPFRC cast on top of the SOG structure was smaller than
that of the prism specimen. This is thought to be because contraction of the
UHPFRC was restrained more significantly by concrete substrates and steel rebar
(higher degree of restraint) and mitigated by creep of the UHPFRC.
• Acceleration of the SOG structure with a 150 mm thick UHPFRC layer was larger
than that with a 100 mm thick UHPFRC layer. It might be said that the compaction
energy of vibrating screed transmits to UHPFRC-Concrete interface irrespective of
UHPFRC layer thickness.
• Failure of specimens of UHPFRC-concrete interfacial bonding tests occurred
mostly in the substrate concrete irrespective of application of adhesive on the
interface. All bond strengths were almost the same (about 2.8 MPa) regardless of
the UHPFRC thickness and compaction energy.
• The porosity of the bottom part of the UHPFRC layer was smaller than that of the
top part of the UHPFRC layer. It might be said that the higher the flow value of
UHPFRC or the frequency of the vibrating screed, the lower the UHPFRC porosity;
however, further investigation is necessary for this tendency.
References
1. Brühwiler, E.: Structural UHPFRC: welcome to the post-concrete era. In: Proceedings of the
First International Interactive Symposium on Ultra-High Performance Concrete, Des Moines,
Iowa, 18–20 July 2016
2. Makita, T., Watanabe, Y., Yanai, S., Kitagawa,H.: Full-scale construction test for
improvement of RC void slab bridges using UHPFRC – PART 1: experimental test plan.
In: RILEM-fib X International Symposium on Fibre Reinforced Concrete, BEFIB2020,
Valencia, 21–23 September 2020
3. Watanabe, Y., Ichinomiya, T., Yanai, S., Iriuchi-jima, K., Suhara, K.: Development of cast-in-
place method of ultra high strength fiber rein-forced concrete. In: Proceedings of the Fifth
International Conference on Construction Materials, CONMAT15, Whistle, British
Columbia, 19–21 August 2015
4. Makita, T., Watanabe, Y., Yanai, S., Ichinomiya, T.: Upgrading of existing bridge decks
using UHPFRC densified by ETtringite Formation (AFt-UHPFRC): preliminary investiga-
tion. AFGC-ACI-fib-RILEM International Symposium on Ultra-High Performance Fibre-
Reinforced Concrete, UHPFRC 2017, Montpellier, 2–4 October 2017
Influence of Fiber Type on the Tensile
Behavior of High-Strength Strain-Hardening
Cement-Based Composites (HS-SHCC) During
and After Exposure to Elevated Temperatures
Iurie Curosu(&), Sarah Burk, Marco Liebscher,

and Viktor Mechtcherine
Institute of Construction Materials, Faculty of Civil Engineering,

Technische Universität Dresden, Dresden, Germany
iurie.curosu@tu-dresden.de
Abstract. The paper summarizes selected results of an extensive experimental

investigation, in which high-strength strain-hardening cement-based composites
(HS-SHCC) made with different high-performance polymer fibers were investi-
gated in terms of mechanical behavior under and after exposure to elevated
temperatures of 105 °C, 150 °C and 200 °C. Besides the ultra-high molecular-
weight polyethylene (UHMWPE) fibers, which are commonly used in HS-SHCC,
high-modulus poly(p-phenylene-2,6-benzobisoxazole) (PBO-HM) fibers have
been analyzed, since they exhibit a considerably higher temperature resistance in
comparison to UHMWPE fibers. In contrast to the expectations, the in-situ and
residual tension experiments at temperatures of up to 150 °C showed that the high-
strength SHCC reinforced with UHMWPE fibers yielded considerably superior
performance and less pronounced decrease of the mechanical properties compared
to the composites made with PBO-HM fibers. Furthermore, the SHCC made with
UHMWPE fibers showed a significant recovery after being cooled down, while
the SHCC made with PBO-HM fibers exhibited a limited recovery; the degra-
dation was proportional to the temperature increase. The 200 °C treatment led to
brittle failure of both composites with dramatically reduced tensile strength and
with low recovery after specimen cooling in the residual experiments.
Keywords: SHCC UHMWPE fiber PBO-HM fiber Elevated temperature

Tension tests
1 Introduction
Strain-hardening cement-based composites (SHCC) represent a novel type of fiber

reinforced concrete with a notably high tensile ductility, which results from the for-
mation of multiple fine cracks under increasing tensile load [1]. Despite their numerous
advantageous features, the range of possible applications of SHCC is restricted by the
high sensitivity of their tensile strength and ductility to elevated temperatures. This
limitation is mainly determined by the polymer fibers suitable for SHCC, which are
made of polyvinyl alcohol (PVA) [2] or ultra-high molecular weight polyethylene
(UHMWPE) [3].
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_90
Influence of Fiber Type on the Tensile Behavior of HS-SHCC 1023
Various studies were performed to evaluate the residual mechanical properties of

PVA-SHCC after exposure to elevated ( 300 °C) and high temperatures (>300 °C).
These investigations also aimed to analyze the mitigation effect of the PVA fibers on
the explosive spalling of the composites at temperatures of up to 800 °C [4–11]. It was
shown that the residual tensile ductility of PVA-SHCC reduces already at 100 °C,
while the residual tensile strength reduces considerably at temperature of 200 °C and
higher due to fiber melting. The most pronounced reduction of tensile strength and
ductility occurs between 200 °C and 300 °C, which corresponds to the range in which
the fibers suffer the most pronounced degradation. At temperatures above 400 °C, not
only the complete decomposition of fibers but also the pronounced deterioration of the
matrix by thermally induced cracks, decomposition of the hydrated phases, increase in
porosity, etc. lead to a brittle behavior in tension and compression with dramatically
reduced mechanical strength. Above 600 °C, the thermal deterioration is similar to that
of concrete [4], while at temperatures of 800 °C severe spalling occurs.
As opposed to the studies on residual properties, only few in-situ investigations on
PVA-SHCC have been reported [12–14]. These studies showed that already at tem-
peratures of 60 °C the first crack stress and tensile strength of PVA-SHCC decrease,
but ductility increases due to the thermally reduced stiffness of the fibers, which leads
to wider crack openings. At 100 °C, the ductility reduces dramatically, while above
150 °C the composites show strain-softening behavior only. At temperatures higher
than 150 °C, the crack bridging of the fibers vanishes completely.
No published studies are known so far on the effect of elevated temperatures on
SHCC reinforced with UHMWPE (short: PE) fibers. While PE fibers exhibit excellent
mechanical properties and high chemical stability in cement-based environments, their
low melting point of approximately 150 °C [15, 16] may impose even more stringent
limitations on their applicability for SHCC compared to PVA fiber.
Various researchers proposed using additionally steel fibers to improve the residual
behavior of SHCC under high temperatures [10, 11, 17]. While the melting of the
polymer fibers provides pathways for vapored water to escape, steel fibers may
improve the ductility at failure localization. However, compared to high-performance
polymer fibers, the steel fibers do not ensure a pre-peak tensile ductility in cementitious
matrices typical for SHCC. Moreover, the addition of steel fibers may have an adverse
effect on the fresh-state properties of SHCC and may affect their applicability by
lamination or spraying. In a previous research by the authors it was demonstrated that
such high-performance polymer fibers as PBO (poly(p-phenylen-2,6-benzobisoxazol)
and para-aramid exhibit desirable mechanical and geometrical properties for high-
strength SHCC, yielding composites with an enhanced first crack stress, tensile strength
and considerably reduced crack width compared to those reinforced with PE fibers [3].
Moreover, according to the producers, the decomposition temperature of the PBO and
aramid fibers is as high as 500 °C and 650 °C, respectively, which makes them
promising for SHCC exposed to elevated temperatures [18, 19].
With the purpose of a detailed assessment of the performance of these fibers in high-
strength SHCC subjected to elevated temperatures, an extensive experimental investi-
gation was performed involving analytical and mechanical experiments at the composite
level as well as on plain cementitious matrix and single fibers. The paper at hand aims to
present briefly some representative results from the mentioned study. Emphasis is put on
1024 I. Curosu et al.
the composites’ tensile behavior during and after exposure to temperatures of up to

200 °C. Besides the reference PE fibers, the high-modulus PBO (PBO-HM) fibers are
discussed in this paper, since they are also representative for the other temperature
resistant polymer fibers, i.e. as-spun PBO (PBO-AS) and para-aramid.
2 Materials
2.1 Fibers
The PE fibers Dyneema® SK62 from DSM, the Netherlands, are gel spun, multi-filament
fibers with high tensile strength and low elongation at break [15]; see Table 2. These
polyolefin fibers exhibit excellent resistance to acids, alkalis and most other chemicals,
including water. Due to their nonpolar molecular structure, they are hydrophobic and
yield weak interfacial bonding towards water-based systems like hardened cement paste.
The PE fibers melt at temperatures of approximately 150 °C and, according to Liu and Yu
[16], the critical temperature for their safe use is around 70 °C.
The alternative to PE fiber presented in this work is the high-modulus p-phenylene-
2,6-benzobisoxazole (PBO-HM) fiber, known under the brand name Zylon®, produced
by Toyobo, Japan. The PBO fibers have high resistance to various organic solvents, acids
and bases [18, 20] and a high decomposition temperature [20–23]. Furthermore, they
have a very high tensile strength, more than twice that of the Dyneema PE fibers, and a
very high tensile modulus of elasticity [18]; see Table 1. As opposed to the PE fibers, the
PBO fibers do not melt, but decompose, and they exhibit a weak hydrophilicity, which
leads to a considerably stronger fiber-matrix bond compared to PE [3].
Table 1. Mechanical, geometrical and physical properties of the fibers under investigation
[15, 18].
Producer DSM Toyobo
Brand Dyneema® Zylon®
Fiber type UHMWPE PBO-HM
Average diameter* [µm] 20 13
Length [mm] 6 6
3
Density [g/cm ] 0.97 1.56
Tensile strength [MPa] 2500 5800
Tensile modulus of elasticity [GPa] 80 270
Elongation at break [%] 3.5 2.5
Decomposing temperature [°C] n.a. 650
Coefficient of linear thermal expansion [10–6 1/K] −12 −6
Melting temperature [°C] 150 n.a.
2.2 Cementitious Matrix

The pronounced hydrophobicity and the high tensile strength of PE fibers determine
their optimal crack bridging behavior in high-strength rather than in normal-strength
cementitious matrices [3, 24–28]. Depending on their composition, high-strength
SHCC can yield tensile and compressive strength values comparable to those of steel
fiber reinforced UHPC, but with considerably higher tensile strain capacity prior to
failure localization [3, 27]. The high-strength cementitious matrix presented in the
paper at hand is identical to that analyzed in the previous study by the authors, in which
the reinforcing performance of PE, PBO and aramid fibers was investigated [3]. It has a
high cement content and a high amount of silica fume as partial cement replacement;
see Table 2. Furthermore, it has a relatively small amount of quartz sand as fine
aggregates and no additional binders or fillers. The fine-grained nature of the matrix is
imposed by the small diameters of the reinforcing fibers and should facilitate proper
fiber dispersion on one hand and a low fracture toughness of the matrix on the other.
Table 2. Composition of the high-strength SHCC under

investigation.
Components Content
[kg/m3]
CEM I 52.5 R-SR3/NA 1460
Silica fume 292
Quartz sand 0.06–0.2 mm 145
Superplasticizer Glenium ACE 460 35
Water 315
UHMWPE fiber (2% by vol.) 20 –
PBO-HM fiber (2% by vol.) – 31
To compensate for the low water-to-binder ratio and negative effect of the fibers on
the workability of SHCC, a high dosage of superplasticizer is used. Note that adequate
fresh-state properties of SHCC are essential for an appropriate homogeneity, fiber
dispersion and robustness of their mechanical properties in hardened state [29]. In the
paper at hand, the SHCC under investigation will be named according to the rein-
forcing fiber, i.e. M-PE and M-PBO-HM, in which M stays for matrix.
3.1 Applied Temperature Treatments

The treatment conditions in the presented study were defined in accordance with the
previous in-situ studies [12–14] by taking into consideration the melting temperature of
PE fibers and the higher temperature resistance of the PBO and aramid fibers. To
achieve an effective 1 h treatment of the specimens at maximum temperature, bench-
mark measurements were performed on heated SHCC specimens with embedded
thermoelements (temperature sensors). The specimens with thermoelements were

positioned in the furnace on an isolating pedestal and had no contact with any con-
ductive elements. Figure 1 shows the treatment profile of a 105 °C experiment.
120
100
Temperature [°C]
80
60
40
specimen core gauge portion
20
furnace
0
0 20 40 60 80 100 120
Time [min]
Fig. 1. Temperature profile in the electric furnace and in the specimen core in a 105 °C
experiment.
As shown in Fig. 1 based on a 105 °C control test, the temperature evolution in the
specimen yields a considerable delay compared to the increase of the air temperature in
the electric furnace and the phase prior to reaching maximum temperature in the
specimen core only slowly converges to the plateau of the air temperature. The time
needed to reach the target temperature inside the specimens was approximately 65 min
for 105 °C, 85 min for 150 °C and 105 min for 200 °C. Thus, the total treatment
duration with a heating rate of 5 K/min was chosen to be 120 min for 105 °C, 150 min
for 150 °C, and 180 min for 200 °C.
3.2 Testing Configuration

The uniaxial tension experiments were carried out in an Instron hydraulic testing
machine on dumbbell shaped SHCC specimens with a geometry commonly used by the
authors in previous studies [3]. Note that the in-situ experiments did not imply testing
inside the electric furnace. For ensuring a short duration of the testing process after
specimen extraction from the furnace, the SHCC specimens were gripped mechanically
in specially fabricated elements made of heat resistant stainless steel; see Fig. 2. The
gripping only ensured axial fixation and imposed almost no restriction to specimen
rotation. Given the short testing duration of approximately 3 min for each specimen, it
was assumed that no negative effects of temperature gradients and no significant
specimen cooling would affect the tensile behavior of the composites. The electric
furnace was positioned in the immediate vicinity of the testing machine as shown in
Fig. 2. Upon reaching the target temperature in the specimens and the required treat-
ment duration, the first specimen was extracted from the electric furnace, weighed,
mounted in the gripping elements, and tested. The same procedure was carried out for
four subsequent specimens, which were tested within a time span of approximately
15 min, this being the difference between the treatment duration of the first and last
tested specimens in a series.
a) b)
cameras for DIC

specimen
furnace
Fig. 2. a) Testing machine with the electric furnace and DIC equipment and b) specimen
mounted in the gripping elements.
Upon the extraction of the last (usually fifth) specimen to be tested, the furnace was
shut down, the door was left ajar and the specimens intended for residual experiments
were left to cool down naturally in the furnace. The cooled down specimens were tested
in the next day in the same testing configuration.
The main reason for testing the heated specimens outside of the electric furnace
(which is otherwise suitable for in-situ tests in the Instron machine) was to enable
optical measurements of specimen deformation, crack formation and their evaluation
using Digital Image Correlation (DIC). Moreover, such a configuration presented a
significant advantage of allowing series of multiple experiments for one parameter
variation, which was an important aspect considering the large number of investigated
material parameters and temperatures in the overall study.
For the optical measurements, all the specimens were sprayed with speckle pattern
using heat resistant paint. The specimens were first provided with a black coat and a
cloud of silver dots was subsequently sprayed on top. A stereo-camera system VIC-3D
from Correlated Solutions was used for the optical measurements. At the beginning of
every test series, the system was calibrated and reference experiments were performed
at room temperature. The DIC evaluation of the specimens’ tensile behavior assumed
defining two virtual calipers on the sides of the resolved surfaces in the gauge portion
of the specimens as shown in Fig. 3. The strains were calculated as caliper elongation,
averaged for two calipers, and used for plotting the stress-strain behavior of the
composites under investigation.
Fig. 3. Virtual calipers positioned on the DIC resolved surface for deriving the global
elongation (strain) of the specimen in the gauge length.
Figures 4 and 6 show representative stress-strain curves obtained in in-situ and residual
experiments on M-PE and M-PBO-HM, respectively. The influence of increased
temperatures on the tensile strength and ductility of the composites is illustrated
comparatively in Figs. 5 and 7. Note that the used gripping elements had the disad-
vantage of inducing pronounced stress concentrations and premature failure of some
specimens in the gripping region. The corresponding specimens were not evaluated in
terms of mechanical properties, which can be seen by the varying number of presented
data points in Figs. 5 and 7.
The temperature treatment at 105 °C did not yield a significant effect on the tensile
strength of M-PE, whereas the corresponding strain capacity (strain at peak load)
yielded a marked increase compared to the non-treated specimens; see Fig. 4a and 5b.
This is a result of a reduced fiber stiffness leading to wider crack openings prior to
failure localization. Furthermore, besides the reduction in fiber stiffness, the tempera-
ture of 105 °C seems to cause a moderate decrease in matrix strength, which led to a
lower first crack stress, thus, to a higher strain-hardening modulus and to a more
pronounced multiple cracking. Based on optical analysis of representative specimens,
the number of cracks in the gauge portion of M-PE at 105 °C was approximately 83, at
room temperature it was approximately 56, while at 150 °C it was 40. The average
crack width increased with temperature from approximately 45 µm at 20 °C to 100 µm
at 150 °C. Whereas at 150 °C the composites still showed a ductile strain-hardening
tensile behavior, at 200 °C M-PE exhibited brittle failure with a dramatically reduced
tensile strength, this being the reason why the in-situ curve was not included in Fig. 4a.
M-PE in-situ M-PE residual

8 8
6 6
Stress [MPa]
Stress [MPa]
4 4
2 room temperature 2 105°C residual

105°C in-situ 150°C residual
150°C in-situ 200°C residual
0 0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
(a) Strain [%] (b) Strain [%]
Fig. 4. Representative (a) in-situ and (b) residual stress-strain curves of M-PE at different
temperature treatments.
Comparing the residual and in-situ tensile behavior of M-PE treated at 105 °C, it
seems that the fibers can partly regain their stiffness, as indicated by the equal tensile
strength but reduced elongation capacity.
M-PE tensile strength M-PE tensile ductility

10 10
in-situ in-situ
residual residual
Strain at peak load [%]
8
Tensile strength [MPa]
6 6
4 4
2 2
0 0
0 25 50 75 100 125 150 175 200 0 25 50 75 100 125 150 175 200
(a) Temperature [°C] (b) Temperature [°C]
Fig. 5. (a) In-situ and residual tensile strength values of individual M-PE specimens and
(b) corresponding strains at peak load.
At 150 °C the residual tensile strength recovered considerably, see Figs. 4 and 5a.
Furthermore, even after 200 °C treatment, M-PE showed some recovery in terms of
tensile strength, while failure exhibited a weak softening; see residual curves in Figs. 4
and 5. It seems that the melting fibers are confined inside their channels, which enables
a partial but beneficial recovery after specimen cooling after 150 °C and even after
200 °C treatment. Whereas protruding fibers could be observed on the fracture surfaces
of specimens treated at 150 °C, the specimens treated at 200 °C showed smooth
fracture surfaces with no visible fibers. In addition, the lateral surfaces of the latter
showed a spider web of fine cracks, demonstrating the severe degradation of the matrix.
As opposed to the previous study [3], the tensile strength and ductility of M-PBO-
HM in the current investigation were lower and the composites yielded a lower
robustness of their mechanical properties. This was be traced back to the rotatable
specimen gripping in the current work. The relatively low crack toughness of M-PBO-
HM especially at higher temperatures augmented the negative effect of rotatable
specimen gripping on the composite performance and led to an uncontrolled failure
with a sudden release of the accumulated strain energy. The influence of boundary
conditions on the tensile performance of these composites will be assessed in a future
work, since this is an important aspect with regard to material testing. The increase in
temperature had a considerable negative effect on both the in-situ and residual tensile
strength and ductility of M-PBO-HM; see Figs. 6 and 7. As shown in Fig. 7, the tensile
strength reduction was proportional with the temperature increase and the recovery in
the residual experiments was lower compared to M-PE. Also the reduction in strain
capacity of M-PBO-HM was proportional to the temperature increase.
M-PBO-HM in-situ M-PBO-HM residual

10 10
105°C residual
150°C residual
8 8 200°C residual
Stress [MPa]
Stress [MPa]
6 6
4 4
2 room temperature 2
105°C in-situ
150°C in-situ
0 0
0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0
Strain [%] Strain [%]
Fig. 6. Representative (a) in-situ and (b) residual stress-strain curves of M-PBO-HM at different
temperature treatments.
The optical analysis of the specimens showed that the crack width in the corre-
sponding SHCC was not significantly affected by the increase in temperature, the
average values being 10.6 µm at 20 °C, 8.6 µm at 105 °C and 11.1 µm at 150 °C. At
the same time, the number of cracks in the gauge length decreased from 38 at 20 °C to
14 at 105 °C and to 9 at 150 °C. Same as in the case of M-PE, the fracture surfaces of M-
PBO-HM treated at 200 °C showed no protruding fibers, explaining the brittle failure of
the in-situ tested specimens and the weak softening in the residual experiments.
M-PBO-HM tensile strength M-PBO-HM tensile ductility

10 1.8
in-situ in-situ
residual residual
1.5
Strain at peak load [%]

Tensile strength [MPa]
8
1.2
6
0.9
4
0.6
2
0.3
0 0.0
0 25 50 75 100 125 150 175 200 0 25 50 75 100 125 150 175 200
(a) Temperature [°C] (b) Temperature [°C]
Fig. 7. (a) In-situ and residual tensile strength values of individual M-PBO-HM specimens and
(b) corresponding strains at peak load.
The reason for a poor mechanical performance of M-PBO-HM at elevated temper-

atures is most probably related to the degradation of the high-strength matrix and to the
fact that the cementitious matrix yields a positive thermal expansion, whereas the
polymer fibers yield a negative one; see Table 1. It could be assumed that the high
stiffness of the PBO-HM fibers might lead to a delamination from the matrix already
during the heating process, i.e. prior to crack formation, drastically reducing their crack
bridging effectiveness. In the case of the PE fibers, the thermal incompatibility might be
compensated by their low Young’s modulus and high elongation capacity. These aspects
will be clarified in a systematic future study involving appropriate testing methods.
5 Conclusions
The experimental results of high-strength SHCC tested under and after exposure to
elevated temperatures indicated clearly the importance of fiber type with regard to the
in-situ and residual tensile behavior of the composites at elevated temperatures.
• The high-strength SHCC made with UHMWPE fibers yielded no reduction in
tensile strength at 105 °C, while the tensile strain capacity doubled due to the more
pronounced multiple cracking and increased crack width. The residual experiments
yielded a reduced strain capacity compared to the in-situ 105 °C tests but still
higher than of the non-heated specimens. This was traced back to the partial
recovery of fibers’ stiffness, restraining the crack openings. At 150 °C, the SHCC
made with UHMWPE fibers showed a pronounced reduction in tensile strength and
a moderate reduction of strain capacity compared to the results obtained at 105 °C.
However, despite this, the composites showed a strain capacity higher than 3% and
a tensile strength of approximately 3 MPa.
• In contrast to the expectations, the composites reinforced with PBO-HM fibers

yielded a pronounced reduction in tensile strength and ductility already at 105 °C,
this decline being proportional to the temperature increase up to 200 °C, at which
the composites yielded brittle failure with no multiple cracking. This effect was
partly traced back to the negative thermal expansion of the fibers. However, further
investigations are necessary for a sound clarifications of the responsible phenom-
ena, which will be a matter of interest in an upcoming study.
Acknowledgements. The authors express their gratitude to the German Research Foundation
(Deutsche Forschungsgemeinschaft - DFG) for the financial support within the Research
Training Group GRK 2250 “Mineral-bonded composites for enhanced structural impact safety”.
Furthermore, the authors express their acknowledgement to Mr. Syed Fasih Mohiuddin, Mr. Kai
Uwe Mehlisch, and Mr. Tilo Günzel for their valuable support in preparing and performing the
experimental investigations. Credit is given to the Institute of Timber Structures of the Tech-
nische Universität Dresden for providing the stereo DIC system.
References
1. Li, V.C.: On engineered cementitious composites (ECC): a review of the material and its
2. Drechsler, A., Frenzel, R., Caspari, A., Michel, S., Holzschuh, M., Synytska, A., Curosu, I.,
Liebscher M., Mechtcherine, V.: Surface modification of poly(vinyl alcohol) fibers to control
the fiber-matrix interaction in composites. Colloid Polymer Sci. 297(7), 1079–1093 (2019)
3. Curosu, I., Liebscher, M., Mechtcherine, V., Bellmann, C., Michel, S.: Tensile behavior of
high-strength strain-hardening cement-based composites (HS-SHCC) made with high-
performance polyethylene, aramid and PBO fibers. Cem. Concr. Res. 98, 71–81 (2017)
4. Sahmaran, M., Lachemi, M., Li, V.C.: Assessing mechanical properties and microstructure
of fire-damaged engineered cementitious composites. ACI Mater. J. 107(3), 297–304 (2010)
5. Bhat, P.S., Chang, V., Li, M.: Effect of elevated temperature on strain-hardening engineered
cementitious composites. Constr. Build. Mater. 69, 370–380 (2014)
6. Magalhaes, M.S., Toledo Filho, R.D., Fairbairn, E.M.R.: Thermal stability of PVA fiber
strain hardening cement-based composites. Constr. Build. Mater. 94, 437–447 (2015)
7. Yu, J., Lin, J., Zhang, Z., Li, V.C.: Mechanical performance of ECC with high-volume fly
ash after sub-elevated temperatures. Constr. Build. Mater. 99, 82–89 (2015)
8. Li, X., Wu, L., Yan, Q., Ma, H., Chen, G., Zhang, H.: Thermal and mechanical properties of
high-performance fiber-reinforced cementitious composites after exposure to high temper-
atures. Constr. Build. Mater. 157, 829–838 (2017)
9. Du, Q., Wei, J., Lv, J.: Effects of high temperature on mechanical properties of polyvinyl
alcohol Engineered Cementitious Composites (PVA-ECC). Int. J. Civil Eng. 16(8), 965–972
(2017)
10. Liu, J.-C., Tan, K.H.: Fire resistance of strain hardening cementitious composite with hybrid
PVA and steel fibers. Constr. Build. Mater. 135, 600–611 (2017)
11. Liu, J.-C., Tan, K., Fan, S.: Residual mechanical properties and spalling resistance of strain-
hardening cementitious composite with Class C fly ash. Constr. Build. Mater. 181, 253–265
(2018)
12. Mechtcherine, V., Silva, F.A., Müller, S., Toledo Folho, R.D.: Coupled strain rate and
temperature effects on the tensile behavior of strain-hardening cement-based composites
(SHCC) with PVA fibers. Cem. Concr. Res. 42(11), 1417–1427 (2012)
13. Purfalah, S.: Behavior of engineered cementitious composites and hybrid engineered
cementitious composites at high temperatures. Constr. Build. Mater. 158, 921–937 (2018)
14. de Oliveira, A.M., Silva, F.A., Fairbairn, E.M.R., Filho, R.D.T.: Coupled temperature and
moisture effects on the tensile behavior of strain hardening cementitious composites (SHCC)
reinforced with PVA fibers. Mater. Struct. 51(3), 1–13 (2018)
15. Dyneema Fact Sheet, Ultra high molecular weight polyethylene fiber from Dyneema,
Eurofibers (2010). https://issuu.com/eurofibers/docs/name8f0d44
16. Liu, X., Yu, W.: Evaluating the thermal stability of high performance fibers by TGA.
J. Appl. Polymer Sci. 99, 937–944 (2006)
17. Deshpande, A.A., Kumar, D., Ranade, R.: Influence of high temperatures on the residual
mechanical properties of a hybrid fiber-reinforced strain-hardening cementitious composite.
18. Technical Information, PBO Fiber Zylon, Toyobo CO., LTD. http://www.toyobo-global.
com/seihin/kc/pbo/zylon-p/bussei-p/technical.pdf
19. Technora, ‘High Tenacity Aramid Fiber, Technical Information’, Teijin Techno Products
Limited, Aramid division, Technora section, June 2004
20. Afshari, M., Kotek, R., Chen, P.: High performance fibers. In: Mittal, V. (ed.) High
Performance Polymers and Engineering Plastics’. Wiley, Hoboken (2011)
21. Ghae, H.G., Kumar, S.: Rigid-rod polymeric fibers. J. Appl. Polymer Sci. 100(1), 791–802
(2006)
22. Kuroki, T., Tanaka, Y., Hokudoh, T., Yabuki, K.: Heat resistance properties of poly(p-
phenylene-2,6-benzobisoxazole) fiber. J. Appl. Polymer Sci. 65(5), 1031–1036 (1997)
23. Liu, X., Weidong, Y.: Degradation of PBO fiber by heat and light. Res. J. Text. Appar. 10
(1), 26–32 (2006)
24. Curosu, I., Mechtcherine, V., Millon, O.: Effect of fiber properties and matrix composition
on the tensile behavior of strain-hardening cement-based composites (SHCCs) subjected to
impact loading. Cem. Concr. Res. 82, 23–35 (2016)
25. Kanda, T., Takaine, Y., Tomoe, S., Takahashi, M., Yamamoto, Y., Kawano, K., Kunieda,
M., Mizobuchi, T.: Development of coupling beam elements utilizing UHP-SHCC for high-
rise R/C building. In: Schlangen, E., Sierra Betran, M.G., Lukovik, M., Ye, G. (Eds.)
Proceedings of the 3rd RILEM Conference on Strain-Hardening Cementitious Composites,
pp. 409–416 (2014)
26. Kunieda, M., Denarié, E., Brühwiler, E., Nakamura, H.: Challenges for strain hardening
cementitious composites – deformability versus matrix density. In: Reinhardt, H.W.,
Naaman, A.E. (eds.) Proceedings of the Fifth International RILEM Workshop on HPFRCC,
pp. 31–38 (2007)
27. Ranade, R., Li, V.C., Heard, W.F.: Tensile rate effects in high strength-high ductility
concrete. Cem. Concr. Res. 68, 94–104 (2015)
28. Kamal, A., Kunieda, M., Ueda, N., Nakamura, H.: Evaluation of crack opening performance
of a repair material with strain hardening behavior. Cem. Concr. Compos. 30(10), 863–871
(2008)
29. Li, M., Li, V.C.: Rheology, fiber dispersion, and robust properties of Engineered
Cementitious Composites. Mater. Struct. 46(3), 405–420 (2013)
Tensile and Compressive Performance of High-
Strength Engineered Cementitious Composites
(ECC) with Seawater and Sea-Sand
Jing Yu1(&), Bo-Tao Huang2, Jia-Qi Wu1, Jian-Guo Dai2,

and Christopher K. Y. Leung1
1
Department of Civil and Environmental Engineering, The Hong Kong
University of Science and Technology, Kowloon, Hong Kong, China
jyuad@connect.ust.hk
2
Department of Civil and Environmental Engineering, The Hong Kong
Polytechnic University, Hung Hom, Hong Kong, China
Abstract. Marine infrastructures play an important role in the social-economic

development of coastal cities. However, the shortage of river/manufactured sand
and fresh water is a major challenge for producing concrete on site, as the
transportation of these materials is not only costly but also environmentally
unfriendly, while desalination of sea-sand and seawater is also pricey. Seawater
sea-sand Engineered Cementitious Composites (SS-ECC) have a great potential
for marine/coastal applications; but the present knowledge on SS-ECC is
extremely limited. This study aims to explore the feasibility of producing high-
strength SS-ECC. The effects of key composition parameters including the
length of polyethylene (PE) fibers (6 mm, 12 mm, and 18 mm) and the maxi-
mum size of sea-sand (1.18 mm, 2.36 mm, and 4.75 mm) on the mechanical
performance of SS-ECC were investigated. SS-ECC with compressive strength
over 130 MPa, tensile strength over 8 MPa and ultimate tensile strain about 5%
were achieved. Test results also showed that the tensile strain capacity increased
with increasing fiber length, while sea-sand size had limited effects on the tensile
performance of SS-ECC. The findings provide insights into the future design
and applications of ECC in marine infrastructures for improving safety, sus-
tainability, and reliability.
Keywords: Fiber-reinforced concrete Engineered cementitious composite

Strain-hardening cementitious composite Marine infrastructures Seawater
Sea-sand Tensile performance
1 Introduction
Marine infrastructures play an important role in the social-economic development of

coastal cities. A major challenge in the development of marine infrastructures is the
shortage of fresh water and river/manufactured sand for producing concrete on site, as the
transportation of these materials is not only costly but also environmentally unfriendly,
while desalination of seawater and sea-sand is also pricey. Additionally, direct use of
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_91
Tensile and Compressive Performance of High-Strength ECC 1035
seawater and sea-sand is generally unsuitable for conventional steel-reinforced concrete

structures, as chloride ions can lead to significant steel corrosion [1–3].
Engineered Cementitious Composite (ECC) is a family of advanced fiber-
reinforced concrete material, which exhibits strain-hardening and multiple-cracking
with fine cracks width (typically 100 lm) under tension [4, 5]. Generally, the tensile
strain capacity of ECC materials is over 3%; the compressive strength is about 20–
80 MPa for normal-strength ECC [6–10] and is about 80–160 MPa for high-strength
ECC [11–16]. Compared to conventional concrete, ECC shows much better durability
performance [17] and mechanical performance under static, cyclic, fatigue and impact
loadings [18–26].
From the perspectives of shortage of fresh water and river sand, ECC made of
seawater and sea-sand (i.e., seawater sea-sand ECC, or SS-ECC) has a great potential
for marine/coastal applications; but the present knowledge on SS-ECC is extremely
limited. This study aims to explore the feasibility of producing high-strength SS-ECC.
An experimental program was conducted to investigate the influence of the length of
fiber (6 mm, 12 mm, and 18 mm) and the maximum size of sea-sand (1.18 mm,
2.36 mm, and 4.75 mm) on the compressive strength and tensile performance of high-
strength SS-ECC.
2.1 Mix Proportion and Raw Materials

Table 1 shows the SS-ECC mixes studied. Ultra-high-molecular-weight polyethylene
(PE) fiber (Table 2) was used for achieving high tensile strength in SS-ECC [27]. In the
mix ID, taking PE12-2.0-S1 for example, “PE12” stands for the length of PE fiber
(12 mm), “2.0” stands for the dosage of PE fiber (2.0 vol.%) and “S1” stands for the
maximum particle size of sea-sand (1.18 mm).
The particle size distribution of sea-sand is shown in Table 3. According to the
design theory of ECC, fine sand is preferred [4, 5, 28]. Hence, the sea-sand with the
size of 1.18 mm was used for the fiber-length series in Table 1. However, the raw
sea-sand contained over 77% particles with a size >1.18 mm (Table 3). To improve the
utilization ratio of the sea-sand, another two groups with the maximum particle sizes of
2.36 and 4.75 mm (PE12-2.0-S2 and PE12-2.0-S4 in Table 1) were prepared.
Table 1. Mix proportions (in weight ratio) of high-strength SS-ECC

Mix ID Cement Silica Sea-sand Seawater HRWR PE fiber
fume 1.18 mm 2.36 mm 4.75 mm Length Vol.
(mm) (%)
PE06-2.0-S1 0.8 0.2 0.3 / / 0.18 0.0135 6 2.0
PE12-2.0-S1 0.8 0.2 0.3 / / 0.18 0.0135 12 2.0
PE18-2.0-S1 0.8 0.2 0.3 / / 0.18 0.0135 18 2.0
PE12-2.0-S2 0.8 0.2 / 0.3 / 0.18 0.0135 12 2.0
PE12-2.0-S4 0.8 0.2 / / 0.3 0.18 0.0135 12 2.0
1036 J. Yu et al.
Table 2. Nominal physical properties of PE fiber

Length (mm) Diameter (µm) Tensile strength (MPa) Elastic modulus (GPa)
6/12/18 24 3000 120
Table 3. Particle size distribution of sea-sand

Particle size 2.36– 1.18– 0.6– 0.3– 0.15– 0.075– <0.075
(mm) 4.75 2.36 1.18 0.6 0.3 0.15
Weight ratio 16.13 61.20 13.83 8.42 0.39 0.01 0.01
(%)
2.2 Specimen Preparation and Test Methods

The mixing process of all the mixes in Table 1 was as follows: (1) the cement, silica
fume and sea-sand were dry mixed for 2 min; (2) the seawater and super-plasticizer
were added and mixed for 5 min; and (3) the PE fibers were then added and mixed for
another 5 min. All the samples were cast in stainless steel molds and demolded after
48 h from casting; they were then cured at a temperature of 23 ± 2 °C and relative
humidity of 95 ± 5% until 28 days.
In the compressive test, at least three 50-mm cubes for each group were prepared
and the loading rate was 0.6 MPa/s [29]. In the direct tensile test, four dumbbell
samples (Fig. 1) for each group were prepared according to a recommendation by the
Japan Society of Civil Engineers [30]. The tensile test was performed using a 25-kN
servo-hydraulic MTS 810 testing system and the loading rate was 0.5 mm/min [30].
The tensile deformation was measured along the middle 80-mm part.
Fig. 1. Dumbbell specimen for direct tension test.

The 28-day compressive strength of the high-strength SS-ECC was >130 MPa for all
the SS-ECC mixes (Table 4). The fiber length (from 6 to 18 mm) had a very little effect
on the compressive strength, which is reasonable as these mixes had the same fiber
dosage. For the sea-sand size, the compressive strength of SS-ECC slightly decreased
as the sea-sand size increased (from 1.18 to 4.75 mm), which may be related to the
higher seashell content in coarser sea-sand. The broken seashell is not only in plate
shape but also generally weaker than silica sand; and therefore sea-sand with higher
seashell content can lead to lower compressive strength in the resulting SS-ECC. It
should also be pointed out that the sand/binder ratio for SS-ECC in this study was
relatively low (only 0.3 as shown in Table 1), and hence the influence of sea-sand size
on the compressive strength was limited.
Table 4. Summary of 28-day compressive strength and tensile performance of high-strength

SS-ECC
Mix ID Compressive Tensile strength Ultimate tensile
strength (MPa) (MPa) strain (%)
Average Deviation Average Deviation Average Deviation
PE06-2.0-S1 132.87 6.23 7.45 0.64 2.43 0.70
PE12-2.0-S1 136.82 6.16 7.92 0.67 5.14 1.42
PE18-2.0-S1 134.00 8.22 7.13 0.73 7.05 2.46
PE12-2.0-S2 133.37 3.16 7.35 0.56 5.03 1.19
PE12-2.0-S4 130.48 3.61 7.03 0.54 4.98 0.90
3.2 Tensile Performance

The tensile strain capacity and strength of the high-strength SS-ECC are summarized in
Table 4, while the tensile stress-strain curves are presented in Fig. 2. It can be found
that with increasing fiber length, the strain capacity increased from 2.43% for 6-mm
fibers to 7.05% for 18-mm fibers, while the strength changed little. According to the
design theory of ECC, with the precondition of limited fiber rupture, fibers with a high
aspect ratio is effective to improve the fiber-bridging effectiveness [27]. On the other
hand, the increase in sand size had very little effect on the strain capacity, but slightly
decreased the tensile strength (Table 4). This phenomenon is similar to the trend in the
compressive strength of sand-size group and it may be related to the seashell content in
sea-sand. Additionally, it can be found in Fig. 2 that the tensile strain capacity of SS-
ECC materials showed considerable scatters, which is widely observed in ECC
materials and is due to the random nature of the fiber distribution and flaw size.
Typical crack patterns of the tensile specimens of SS-ECC after testing are shown
in Fig. 3. All the ECC specimens showed multiple-cracking behaviors. As the fiber
length increased, the crack number significantly increased, while the sea-sand size had
a limited effect on the crack number.
1038 J. Yu et al.
Fig. 2. Influence of fiber length and sea-sand size on the tensile stress-strain curves of SS-ECC.
All mixtures show strain-hardening behavior.
Fig. 3. Influence of fiber length and sea-sand size on the crack pattern of SS-ECC. All mixtures
show multiple cracking behavior.
4 Conclusions
This study explores the feasibility of producing high-strength seawater sea-sand

Engineered Cementitious Composites (SS-ECC). According to the materials used and
results obtained, the following conclusions can be drawn:
• SS-ECC achieved compressive strength >130 MPa, tensile strength >8 MPa and
ultimate tensile strain about 5% at 28 days, which fulfil the mechanical require-
ments of many infrastructures.
• The tensile strain capacity increased with increasing fiber length from 6 mm to
18 mm, while the size of sea-sand up to 4.75 mm had limited effects on the tensile
performance.
• The compressive strength of SS-ECC slightly decreased as the sea-sand size
increased (from 1.18 to 4.75 mm), which may be related to the higher seashell
content in coarser sea-sand.
These findings provide insights into the future design and applications of SS-ECC
in marine/coastal infrastructures.
Acknowledgements. This study was financially supported by the Hong Kong Research Grants
Council (No.: T22-502/18-R) and the National Key Research Program of China (No.:
2017YFC0703403). The authors also thank Dr. Yu Xiang, Mr. Ji-Xiang Zhu and Mr. Ke-Fan
Weng for their assistance in the experiment.
1040 J. Yu et al.
References
1. Xiao, J., Qiang, C., Nanni, A., Zhang, K.: Use of sea-sand and seawater in concrete
construction: current status and future opportunities. Constr. Build. Mater. 155, 1101–1111
(2017)
2. Teng, J.-G., Xiang, Y., Yu, T., Fang, Z.: Development and mechanical behaviour of ultra-
high-performance seawater sea-sand concrete. Adv. Struct. Eng. 22(14), 3100–3120 (2019)
3. Ahmed, A., Guo, S., Zhang, Z., Shi, C., Zhu, D.: A review on durability of fiber reinforced
polymer (FRP) bars reinforced seawater sea sand concrete. Constr. Build. Mater. 256,
119484 (2020)
4. Li, V.C., Leung, C.K.Y.: Steady-state and multiple cracking of short random fiber
composites. J. Eng. Mech. 118(11), 2246–2264 (1992)
5. Li, V.C.: Engineered Cementitious Composites (ECC) - Bendable Concrete for Sustainable
and Resilient Infrastructure. Springer, Heidelberg (2019)
6. Yu, J., Leung, C.K.Y.: Strength improvement of strain-hardening cementitious composites
with ultrahigh-volume fly ash. J. Mater. Civil Eng. 29(9), 05017003 (2017)
7. Yu, J., Li, H., Leung, C.K.Y., Lin, X., Lam, J.Y.K., Sham, I.M.L., Shih, K.: Matrix design
for waterproof engineered cementitious composites (ECCs). Constr. Build. Mater. 139, 438–
446 (2017)
8. Huang, B.-T., Li, Q.-H., Xu, S.-L., Zhou, B.: Strengthening of reinforced concrete structure
using sprayable fiber-reinforced cementitious composites with high ductility. Compos.
Struct. 220, 940–952 (2019)
9. Yu, J., Yao, J., Lin, X., Li, H., Lam, J.Y.K., Leung, C.K.Y., Sham, I.M.L., Shih, K.: Tensile
performance of sustainable Strain-Hardening Cementitious Composites with hybrid PVA
and recycled PET fibers. Cem. Concr. Res. 107, 110–123 (2018)
10. Yu, J., Wu, H.-L., Leung, C.K.Y.: Feasibility of using ultrahigh-volume limestone-calcined
clay blend to develop sustainable medium-strength Engineered Cementitious Composites
(ECC). J. Clean. Prod. 262, 121343 (2020)
11. He, S., Qiu, J., Li, J., Yang, E.-H.: Strain hardening ultra-high performance concrete
(SHUHPC) incorporating CNF-coated polyethylene fibers. Cem. Concr. Res. 98, 50–60
(2017)
12. Chen, Y., Yu, J., Leung, C.K.Y.: Use of high strength strain-hardening cementitious
composites for flexural repair of concrete structures with significant steel corrosion. Constr.
Build. Mater. 167, 325–337 (2018)
13. Ranade, R., Li, V.C., Stults, M.D., Heard, W.F., Rushing, T.S.: Composite properties of
high-strength, high-ductility concrete. ACI Mater. J. 110(4), 413–422 (2013)
14. Chen, Y., Yu, J., Younas, H., Leung, C.K.Y.: Experimental and numerical investigation on
bond between steel rebar and high-strength Strain-Hardening Cementitious Composite
(SHCC) under direct tension. Cem. Concr. Comp. 112, 103666 (2020)
of a repair material with strain hardening behavior. Cem. Concr. Comp. 30(10), 863–871
(2008)
16. Curosu, I., Liebscher, M., Mechtcherine, V., Bellmann, C., Michel, S.: Tensile behavior of
high-strength strain-hardening cement-based composites (HS-SHCC) made with high-
performance polyethylene, aramid and PBO fibers. Cem. Concr. Res. 98, 71–81 (2017)
17. van Zijl, G.P.A.G., Slowik, V.: A framework for durability design with strain-hardening
cement-based composites (SHCC): state-of-the-art report of the RILEM technical committee
240-FDS. In: RILEM State-of-the-Art Reports. RILEM, Netherlands (2017)
18. Li, V.C., Horii, H., Kabele, P., Kanda, T., Lim, Y.M.: Repair and retrofit with engineered
cementitious composites. Eng. Fract. Mech. 65(2–3), 317–334 (2000)
19. Huang, B.-T., Li, Q.-H., Xu, S.-L., Liu, W., Wang, H.-T.: Fatigue deformation behavior and
fiber failure mechanism of ultra-high toughness cementitious composites in compression.
Mater. Des. 157, 457–468 (2018)
20. Lu, C., Yu, J., Leung, C.K.Y.: Tensile performance and impact resistance of strain hardening
cementitious composites (SHCC) with recycled fibers. Constr. Build. Mater. 171, 566–576
(2018)
21. Huang, B.-T., Li, Q.-H., Xu, S.-L., Zhou, B.-M.: Tensile fatigue behavior of fiber-reinforced
cementitious material with high ductility: experimental study and novel P-S-N model.
22. Mechtcherine, V.: Novel cement-based composites for the strengthening and repair of
concrete structures. Constr. Build. Mater. 41, 365–373 (2013)
23. Huang, B.-T., Li, Q.-H., Xu, S.-L., Zhou, B.-M.: Frequency effect on the compressive
fatigue behavior of ultrahigh toughness cementitious composites: experimental study and
probabilistic analysis. J. Struct. Eng. 143(8) (2017)
24. Yu, J., Chen, Y., Leung, C.K.Y.: Mechanical performance of Strain-Hardening Cementitious
Composites (SHCC) with hybrid polyvinyl alcohol and steel fibers. Compos. Struct. 226,
111198 (2019)
25. Huang, B.-T., Li, Q.-H., Xu, S.-L., Zhang, L.: Static and fatigue performance of reinforced
concrete beam strengthened with strain-hardening fiber-reinforced cementitious composite.
Eng. Struct. 199, 109576 (2019)
26. Mechtcherine, V., Silva, F.d.A., Butler, M., Zhu, D., Mobasher, B., Gao, S.-L., Mäder, E.:
Behaviour of strain-hardening cement-based composites under high strain rates. J. Adv.
Concr. Technol. 9(1), 51–62 (2011)
27. Zhang, D., Yu, J., Wu, H., Jaworska, B., Ellis, B., Li, V.C.: Discontinuous micro-fibers as
intrinsic ductile reinforcement for Engineered Cementitious Composites (ECC). Compos.
Part B-Eng. 184, 107741 (2020)
28. Wu, H.-L., Yu, J., Zhang, D., Zheng, J.-X., Li, V.C.: Effect of morphological parameters of
natural sand on mechanical properties of engineered cementitious composites. Cem. Concr.
Comp. 100, 108–119 (2019)
29. ASTM. Standard Test Method for Compressive Strength of Hydraulic Cement Mortars. In:
C109/C109M (ASTM International, West Conshohocken) (2013)
30. JSCE. Recommendations for design and construction of high performance fiber reinforced
cement composites with multiple fine cracks (HPFRCC). Japan Society of Civil Engineers,
Tokyo, Japan (2008)
Effect of Fiber Content Variation in Plastic
Hinge Region of Reinforced UHPC Flexural
Members
Mandeep Pokhrel1(&), Yi Shao2, Sarah Billington2,

and Matthew J. Bandelt1
1
Department of Civil and Environmental Engineering,
New Jersey Institute of Technology, Newark, USA
mp595@njit.edu
2
Stanford University, Stanford, USA
Abstract. Ultra-high performance concrete (UHPC) is used for the construc-

tion of resilient structures that can sustain dynamic loadings such as blast,
impact, and earthquake loadings, among others. In structural components sub-
jected to such loading, it is essential to ensure the formation of a ductile plastic
hinge mechanism for suitable load transfer mechanisms and global stability of
the structure. Experimental research is needed to understand the formation of
plastic hinges in UHPC materials and the impact of plastic hinges on the rotation
capacity of reinforced UHPC structural components. The study presented herein
aims to understand the spread of plasticity and formation of plastic hinge regions
in reinforced UHPC flexural members. Two reinforced UHPC beams with
variation in fiber volume fraction (i.e., Vf = 1% and 2%) were subjected to
monotonic loading. The test results demonstrated that the reinforcement plas-
ticity length increased by 26% with a decrease in fiber volume fraction from 2%
to 1%. The plastic hinge region of specimens with 2% fiber content had crack
localization within the maximum moment region, whereas the specimen with
1% fiber content had a more uniformly distributed localized crack pattern.
Further, analytical models and a recently proposed equivalent plastic hinge
length equation were used to predict and compare the flexural strength and
rotation values at various damage states.
Keywords: Plastic hinge length Fiber content Reinforced UHPC Ultimate

rotation capacity
1 Introduction and Background
Ultra-high performance concrete (UHPC) is an advanced cement-based composite

material designed with optimal particle packing density, such that, it possesses extre-
mely high compressive strength (>120 MPa without heat treatment) and enhanced
durability properties [1]. When combined with short discontinuous fibers, UHPC
materials have high tensile strength (>5 MPa), tensile fracture toughness, and ductile
strain-hardening behavior under uniaxial tension tests [1]. The mechanical properties of
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_92
Effect of Fiber Content Variation in Plastic Hinge Region 1043
UHPC have led researchers and engineers to perform a large number of proof-of-
concept investigations under extreme loading conditions such as blast, impact, earth-
quake, and fire [2]. In high seismic zones, researchers are especially interested in the
applicability of UHPC in plastic hinge regions of structural components undergoing
large inelastic deformations [3]. Various classes of high performance fiber reinforced
cementitious composites (HPFRCCs) have already been effectively used in plastic
hinge region of structural components such as coupling beams, columns, and bridge
piers in recent years [4]. The use of UHPC in the plastic hinge regions of structural
components can enhance the load carrying capacity, ductility, and energy absorption
capacity because the mechanical properties of UHPC can prevent premature failure
associated with damage in plastic hinge regions, as is typically observed in structural
components made with conventional concrete (e.g., spalling of cover, buckling of
rebar, shear cracks, etc.).
Structural components can be engineered to improve the damage tolerance of
structures by using UHPC in plastic hinge region, while using conventional concrete in
the remaining portions of the component [3]. Such an approach can minimize the high
cost associated with UHPC while reducing the overall life cycle cost (i.e., maintenance
and repair cost) of the structure. To optimize the initial construction cost, the fiber
volume fraction used in a UHPC material can be reduced; however, the influence of
such a reduction on structural ductility and performance of plastic hinge regions is not
well understood. A previous numerical study conducted using a wide range of
HPFRCC materials indicated that the tensile strength and ductility of the matrix can
significantly alter the amount of damage and the length of the plastic hinge region in
reinforced HPFRCC flexural members [5, 6]. Therefore, the use of low fiber content in
regions undergoing large displacement reversals may result in an undesired failure
mechanism (e.g., shear cracking) without the formation of a ductile plastic hinge
mechanism.
Other experimental and numerical studies have shown that the flexural behaviour of
reinforced HPFRCCs (including UHPC) in terms of crack progression, reinforcement
plasticity, and failure mechanism is significantly different than conventional reinforced
concrete structural components [7–10]. Specifically, the failure mode of flexural
members is found to be predominantly through the fracture of longitudinal reinforce-
ment rather than compression crushing of an HPFRCC matrix. This is due to a crack
localization phenomenon observed in reinforced HPFRCC structural components,
wherein the plastic damage concentrates in the vicinity of a single or few flexural
cracks. Several bond experiments with lap splice beam specimens have shown that
higher bond strength of HPFRCC matrix restraints the formation of splitting cracks
which leads to such a phenomenon [11–13]. Further, tension stiffening experiments of
reinforced HPFRCC prisms have shown that there is localized strain hardening in
longitudinal reinforcement at such localized cracks [14, 15]. Localized hardening of
steel reinforcement can provide a strengthening mechanism at the critical section of
reinforced HPFRCC flexural members, until the member loses its load-carrying
capacity by reinforcement fracture.
Shao and Billington [16] recently conducted an experimental study with reinforced
UHPC beams consisting of two different reinforcement ratios (q = 0.96% and 2.10%).
The study showed that there can be two different failure paths in reinforced UHPC
1044 M. Pokhrel et al.
beams depending on the amount of longitudinal reinforcement used. The use of low
reinforcement ratio led to failure after crack localization failure path in which there
were three major damage states: yielding, crack localization, and rebar fracture. The use
of a high reinforcement ratio led to failure after gradual strain hardening failure path
in which the intermediate damage state changed from crack localization to compression
crushing (or softening). Compared to specimens that fail after crack localization,
specimens that fail after gradual strain hardening show higher ductility and more failure
warnings. The flexural failure paths and various damage states are found to be pre-
dominantly dependent on longitudinal reinforcement and matrix property based on
similar experimental studies carried out on other HPFRCC materials [10, 17].
Although many experimental studies have investigated the flexural behavior of
reinforced UHPC beams, their specimens mostly fail after crack localization [16, 17].
There is limited experimental study in the literature dedicated to investigate the rein-
forcement plasticity distribution, curvature distribution, UHPC surface strain variation,
and damage propagation in plastic hinge region of UHPC beams that fail after gradual
strain hardening. Further, the influence of fiber content variation on various parameters,
as mentioned above, is important to understand the component flexural behaviour,
especially the maximum load carrying capacity, deformation capacity, and structural
stability of components constructed with such material and cross-section property.
Moreover, a mechanics-based analytical approach is adopted in the study, to predict
strength and rotation capacity of the such specimens, which provides a simplified way
for the practicing engineers to model and design such components in large structural
systems.
To that end, an experimental study consisting of two reinforced UHPC beams with
1% and 2% fiber volume fraction (Vf ) were tested under a four-point bending test
setup. Based on recent experimental studies [16, 17], these two beams were designed to
fail after gradual strain hardening by adopting a high reinforcing ratio of 2.10%. The
flexural response, length of reinforcement yielding, inelastic curvature distribution,
crack distribution, and strain variation within the maximum moment region were
investigated. Analytical models and a recently proposed equivalent plastic hinge length
equation for ductile concrete composites were used to predict flexural strengths and
rotation values at various damage states.
2.1 Materials, Mixture Proportions, and Mechanical Tests

Two types of mixture proportions were used in the experiment as listed in Table 1. The
propriety pre-mix blend contained a mixture of cement, quartz, and silica fume. There
were three types of admixtures used to improve the workability of UHPC during the
casting period. Standard smooth steel fibers with a diameter of 0.2 mm and length of
13 mm were used in both mixtures. The naming convention of the beam specimens
were based on the percentage of fiber volume fraction used in each specimen. There-
fore, UHPC-1% denotes the specimen with a fiber volume fraction of 1% and UHPC-
2% indicates the specimen containing a fiber volume fraction of 2%. The materials
were mixed in a horizontal shear mixer and poured from one end of the beam mold
until the mold was filled up to the full height.
Table 1. Mixture proportions (per m3 )

Specimen Premix Water Fibers [% Admix. Admix. Admix.
blend [kg] [kg] Vol.] A [kg] B [kg] C [kg]
UHPC-1% 1939 194 1.0 20 26 28
UHPC-2% 1960 196 2.0 20 26 28
All UHPC specimens were moist cured and tested at 56 ± 3 days of casting. The
representative results of mechanical tests have been tabulated in Table 2. Both com-
pression and flexural tests of the two types of UHPC mixtures were conducted in
accordance with ASTM C1856-17 [18]. Cylindrical specimens of diameter 75 mm and
height 150 mm were prepared and tested to obtain compressive strength and modulus
of elasticity. Four-point bending tests were performed on UHPC prisms with a cross
section dimension of 75 75 mm and a length of 300 mm to obtain the equivalent
bending stress versus displacement response as shown in Fig. 1(a).
Table 2. Mechanical properties

0
Description a
ft Gf a fc E fy fu eu ef
[MPa] [MPa- [MPa] [GPa] [MPa] [MPa] [%] [%]
mm]
UHPC-1% 5.2 6.4 170 41.6 – – – –
UHPC-2% 9.2 6.7 180 42.5 – – – –
Longitudinal – – – 190 470 780 10 18b
Rebar (#6)
Transverse – – – 200c 510c – – –
Rebar (#3)
a
Obtained using inverse analysis; bExtrapolated value; cManufacturer listed value
A series of inverse analyses were conducted using two-dimensional finite element

simulations to estimate the tensile properties of the two UHPC mixtures. The inverse
analysis scheme adapted in this study has been successfully implemented by other
researchers in the past to characterize the tensile stress-strain curve without resorting to
the direct tension test [9, 10, 17]. A total strain-based fixed-crack constitutive model
was used with trial multilinear tensile stress-strain curve as an input to the numerical
model. It can be observed from Fig. 1(a) that the simulated flexural response from
inverse analysis closely approximates the experimental flexural behaviour of unrein-
forced UHPC prisms. The corresponding tensile strength (f t ) and tensile fracture
energy (Gf ) of the mixtures are listed in Table 2, which can be used in lieu of tensile
25 80
UHPC-1% [Experiment]
Equivalent Bending Stress [MPa]
Inverse Analysis [ft: 5.2 MPa]

UHPC-2% [Experiment] 70
[kNm]
20 Inverse Analysis [ft: 9.2 MPa]
Moment[kNm]
60
50
Applied Moment
15
40
10
Applied
30
20 UHPC-1%
5 UHPC-2%
Yield
10 Softening
Fracture
0 0
0 0.5 1 1.5 2 2.5 0 1 2 3 4 5 6 7 8 9 10 11
(a) Mid-span Displacement [mm] (b) Δ∆/L
/Lshear-span [%]
shear-span [%]
Fig. 1. (a) Equivalent bending stress vs. mid-span displacement of unreinforced UHPC beams
(b) Applied moment vs. drift response of two reinforced UHPC beams with different damage
states
parameters obtained from direct tension test to investigate the flexural behavior and
plastic hinge region of reinforced UHPC beams tested in this study. ASTM A615
Grade 60 steel with a diameter of 19 mm was used as longitudinal reinforcing bar in
both UHPC-1% and UHPC-2% specimens. A uniaxial tension test was conducted
using an extensometer of gauge length 50 mm to obtain characteristic tensile properties
of the longitudinal reinforcement as listed in Table 2. Transverse reinforcement of
Grade 60 steel with manufacturer listed yield strength of 510 MPa was used in both
specimens.
2.2 Test Specimens, Setup and Instrumentation

Two reinforced UHPC beam specimens were tested using a four-point bending setup as
shown in Fig. 2(a). A digital image correlation (DIC) system was used to assess
variations in strain in the constant moment region of 200 mm between the two point
loads. Since the DIC system could not be extended further due to laboratory con-
straints, a series of linear variable displacement transducers (LVDTs) were used to
measure the vertical displacement along one side of the beam to a distance of 200 mm
from the center line of the right point load towards the roller support.
Figure 2 (b–c) shows the design of the beam with longitudinal reinforcement
layout, transverse reinforcement layout, location of strain gauges, and cross-section
details. The strain in the bottom longitudinal reinforcement was measured at five
locations by attaching post-yield strain gauges (YEFLA-2-3LJC-F from Tokyo Mea-
suring Instruments Lab) with a maximum measurement capacity of 10%. Two longi-
tudinal reinforcing bars of diameter 19 mm were used in the top and bottom sides of
the cross section resulting into a tensile longitudinal reinforcement ratio of 2.10%.
Transverse reinforcement was provided at a spacing of one-half of the effective depth
(i.e., d=2) with bars of diameter 9.5 mm. The specimens were subjected to a monotonic
Vertical Actuator
(245 kN)
Reinforced UHPC Beam

(thickness: 150 mm)
DIC Surface Preparation
Pin Support Roller

Support
LVDT @ Center
LVDTs @ 50 mm c/c
800 mm 200 mm 800 mm
(a)
-9.5 mm Stirrups @ d/2 -19 mm Rebar 150 mm

A
180
220
mm
mm
Strain Gauge A
at Center 100 mm 40 mm
Strain Gauges
@ 50 mm c/c (c)
1800 mm
(b)
Fig. 2. (a) Test setup of reinforced UHPC beams with location of DIC surface and LVDTs
(b) Specimen design detail with location of strain gauges (c) cross section at A-A
loading at the rate of 0.097 mm/s until they lost their load carrying capacity by fracture
of the tensile longitudinal reinforcement.
3 Result and Discussion

3.1 Moment-Drift Response
The applied moment versus drift response of the two specimens with various damage
states is shown in Fig. 1 (b). Drift is expressed in percentage (%) and is calculated by
normalizing the vertical displacement at mid-span by the shear-span length
(D=Lshearspan ). The initial elastic response of both beams including stiffness, moment at
yield and drift at yield are similar. The beams were assumed to yield when the strain in the
tensile reinforcement at mid-span reached the yield strain (ey ) of 0.2772%. At yield, the
moment and drift capacities are similar because of the use of the same reinforcement ratio
(q = 2.10%) in both beams. The strength and drift capacities at yield are less sensitive to
variations in tensile strength of the matrix (or fiber volume fraction) in specimens with
higher reinforcement ratios compared to specimens with lower reinforcement ratios [5].
After yield, the flexural load carrying capacity increased in both specimens due to the
combination of fiber bridging action and localized hardening of the tensile reinforcement.
The post-yield stiffness of both beams were similar; however, the nominal moment
(Mn ) capacity (i.e., peak moment capacity) of UHPC-2% (72.2 kNm) was found to be
lower than UHPC-1% (75.9 kNm). It was anticipated that the UHPC-2% specimen, which
has twice the fiber content and a higher tensile strength than UHPC-1% beam, would have
a higher nominal moment capacity than UHPC-1% specimen. However, a higher rate of
post-yield strain accumulation in the compression zone of UHPC-2% specimen was
observed (Fig. 5 (a)). This led to earlier softening of the compression matrix and a lower
flexural load carrying capacity in UHPC-2% specimen than in UHPC-1% specimen. The
drift of UHPC-2% specimen at the nominal level is lower than the drift of UHPC-1% by
31% because of the rapid strain concentration in the compression zone of the UHPC-2%
specimen. For example, at 2.2% drift level, the compression zone strain ðec Þ in UHPC-2%
beam was found to be 35% higher than in UHPC-1% beams (i.e., ðec Þ UHPC-2% = 0.42%
whereas ðec Þ UHPC-1% = 0.31% as shown in Fig. 5 (a)) as further discussed in Sect. 3.5.
Both specimens were able to achieve large deformations without significantly losing load
carrying capacity because the hardened bottom longitudinal reinforcement acted as the
tensile component of the flexural couple before reaching the fracture point. This failure
mechanism with gradual strain hardening of tensile reinforcement is mostly found in
reinforced HPFRCC flexural members with high reinforcement ratios [17]. It is inter-
esting to note that the variation of fiber content did not influence the value of ultimate drift
4.0 4.0
Lpy = 290 mm 1% Drift Lpy = 230 mm 1% Drift
2% Drift 2% Drift
Reinforcement Tensile Strain [%]
4% Drift 4% Drift
3.0 8% Drift 3.0
8% Drift
Yield Strain
Yield Strain
2.0 2.0
1.0 1.0
0.0 0.0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
(a) Distance from Mid-span [mm] (b) Distance from Mid-span [mm]
0.001 0.001
Lcl = 150 mm 1% Drift Lcl = 150 mm 1% Drift
2% Drift 2% Drift
4% Drift 4% Drift
0.0008 0.0008
8% Drift 8% Drift
Curvature [1/mm]
Collapse Drift Collapse Drift
0.0006 0.0006
0.0004 0.0004
0.0002 0.0002
0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
(c) Distance from Mid-span [mm] (d) Distance from Mid-span [mm]
Fig. 3. Longitudinal reinforcement tensile strain vs. distance from mid-span at various drift
levels for (a) UHPC-1% specimen and (b) UHPC-2% specimen Curvature vs. distance from mid-
span at various drift levels for (c) UHPC-1% and (d) UHPC-2% specimens

capacity ( D=Lshearspan UHPC-1% = 10.02% and D=Lshearspan UHPC-2% = 10.15%).
The results indicates that the ultimate rotation or drift capacity of reinforced UHPC beams
with high longitudinal reinforcement (i.e., q [ 2%) is not sensitive to variation in fiber
volume fraction compared to the beams with low to moderate reinforcement ratio (i.e.,
1%\q\2%).
3.2 Strain Distribution

Figure 3 (a–b) presents the variation of reinforcement tensile strain from mid-span to
the right side of the specimens (up to 250 mm). At 1% drift, it can be observed that the
UHPC-1% specimen had a reinforcement yielding length, Lpy , of 50 mm, whereas the
longitudinal reinforcement yielded over a length of 90 mm in UHPC-2% specimen.
However, Lpy in the UHPC-1% specimen increased at higher drift level compared to
UHPC-2% specimen because of the opening of flexural cracks along a longer span
length as shown in Fig. 4. The fiber bridging action and matrix tensile strength of
UHPC-1% specimen is lower than the UHPC-2% specimen. The lower strength of
UHPC-1% allowed cracks top easily open at high drift levels and plasticity distributed
uniformly over a longer length of reinforcement. The length of reinforcement yielding
remained constant at higher drifts in both specimens as the inelastic strain concentration
mostly occurred near a dominant crack location as was observed in previous studies
involving reinforced HPFRCC flexural members [9, 19]. At 8% drift level, the length
of reinforcement yielding region in UHPC-1% specimen (Lpy ¼ 290 mm) was longer
than UHPC-2% specimen (Lpy ¼ 230 mm). These results suggest that the plastic hinge
length in beams with higher fiber content (or higher tensile strength and matrix
toughness) are shorter than those with lower fiber content (or lower tensile strength and
matrix toughness). This is also in agreement with the length and type of cracking
pattern shown in Fig. 4 and further discussed in Sect. 3.4.
3.3 Curvature Distribution

Curvature distribution along the span of the specimens were calculated using the
vertical displacements data obtained from the six LVDTs. Mathematically, curvature at
a section can be approximated using the elastic deflection theory as shown in Eq. (1).
dh dh hi þ 1 hi
/¼ ffi ¼ ð1Þ
ds dx distance between LVDTs
Where hi and hi þ 1 are the angles at sections i and i þ 1: These angles can be computed
using the vertical displacements obtained from LVDTs along the span of the beam. Due
to the opening of localized cracks at higher drifts, the recorded vertical displacement
data at some locations were estimated using a trendline. Figure 3 (c–d) shows curvature
distribution from mid-span to the right side of the specimens at incremental drift levels.
At a lower drift level (1% or 2%), the curvature is maximum below the point load, which
is the assumed hinge location under four-point bending setup. However, at larger drift
levels (4% or more) the curvature is larger at mid-span because of the opening of a major
crack at mid-span, such that the section at mid-span deforms more than the section
below the point load as seen from the crack pattern at in Fig. 4. The overall trend of the
curvature is similar to the theoretical curvature under four-point bending test where the
curvature is maximum near the mid-span and sharply decreasing away towards the
support. The curvature localization region (Lcl ) was 150 mm in both the specimens
indicating there was no substantial effect of fiber content variation in curvature distri-
bution of reinforced UHPC flexural members with high reinforcement ratios.
3.4 Crack Pattern in Plastic Hinge Region

Figure 4 shows the crack pattern and location of reinforcement fracture in UHPC
specimens at impending collapse level drift (i.e., D=Lshearspan UHPC-1% = 10.02%

and D=Lshearspan UHPC-2% = 10.15%). Both specimens contained multiple fine dis-
tributed cracks without any flexural crack localization up to 1% drift level. However,
after yielding of tensile reinforcement, flexural cracks slowly began to open as the
fiber-bridging action declined in some of the cracks. Four major flexural cracks
widened in both specimens, but the major cracks were evenly spaced in UHPC-1%
specimen compared to UHPC-2% specimen as shown in Fig. 4. Major cracks widened
in the region away from maximum moment in the UHPC-1% specimen because the
cracks could open at lower flexural load due to a lower tensile strength and fracture
energy of UHPC-1% matrix. In UHPC-2% specimen, major cracks were confined to
the maximum moment region and damage was pre-dominantly localized in a single
crack (i.e., crack number 3). As crack number 3 widened, rebar plastic strain in the
vicinity of that crack concentrated at a higher rate in the UHPC-2% specimen as
compared to the UHPC-1% specimen (Fig. 5 (b)) as further discussed in Sect. 3.5.
The distance between the extreme minor and major cracks were found to be longer in
UHPC-1% specimen (Lminorcracks = 1350 mm and Lmajorcracks = 420 mm) compared to
UHPC-2% (Lminorcracks = 935 mm and Lmajorcracks = 240 mm) which can be attributed
to the comparatively weaker matrix and fiber-bridging action in UHPC-1% specimen
compared to UHPC-2% specimen.
UHPC-1%
200 mm
1 2 3 4
Lmajor cracks= 420 mm

(a)
Lminor cracks= 1350 mm
UHPC-2%
200 mm
1 2 3 4
Lmajor cracks= 240 mm

(b)
Lminor cracks= 935 mm
Fig. 4. Crack pattern in (a) UHPC-1% and (b) UHPC-2% specimens at impending collapse
2.5 5.0
UHPC-2% UHPC-2%
Mid-span Compression Strain [%]
Mid-span Tensile Rebar Strain [%]

UHPC-1% UHPC-1%
2.0 4.0
1.5 3.0 Softening
1.0 Softening 2.0
Yield Yield
0.5 1.0
0.0 0.0
0 1 2 3 4 5 6 0 1 2 3 4 5 6
(a) ∆ /Lshear-span
∆/L [%]
[%] (b) ∆ /Lshear-span
∆/L [%]
shear-span shear-span [%]
Fig. 5. (a) Mid-span compression strain vs. drift and (b) mid-span longitudinal tensile
reinforcement strain vs. drift
3.5 Variation of Strain in Maximum Moment Region

Figure 5 shows the variation of compression and tension strain at mid-span with
incremental drift levels. The slope of the lines indicates the rate of strain accumulation
in the compression zone (Fig. 5 (a)) or in the tensile reinforcement (Fig. 5 (b)) at mid-
span. After yielding of the specimens, the rate of compression strain accumulation in
the UHPC-2% specimen became marginally higher compared to the UHPC-1%
specimen because the high tensile strength matrix attracts larger forces at smaller
deformation level. Further, higher bond strength in UHPC-2% specimen restricted the
formation of splitting cracks, causing early rebar hardening over a small deboned
length. The hardening led to widening of a mid-span crack and rapid strain concen-
tration in the tensile longitudinal reinforcement in UHPC-2% specimen at a lower drift
level as shown in Fig. 5 (b). For example, the strain in the tensile reinforcement at 2%
drift level in UHPC-2% specimen was 2.00% and that in UHPC-1% specimen was
1.16%. The effect of this rapid strain variation caused softening of
UHPC-2% beam at a
lower drift level compared to UHPC-1% beam ( D=Lshearspan UHPC-1% = 3.2% and

D=Lshearspan UHPC-2% = 2.2%). After softening of the specimens, the compression
strain accumulation rate in the UHPC-1% specimen increased compared to the UHPC-
2% specimen. The difference in compression strain at the same drift level decreased
progressively at higher drift levels. The localized hardening strain in tensile rein-
forcement at mid-span of the UHPC-2% specimen was much higher compared to the
UHPC-1% specimen at the same drift levels. For instance, the strain in tensile rein-
forcement at 4% drift level in UHPC-2% specimen was 4.72% and that in UHPC-1%
specimen was 2.91%. It was anticipated that the UHPC-2% would fail earlier by
fracture of reinforcement based on this trend. However, UHPC-2% had a similar
deformation capacity as UHPC-1% as discussed in Sect. 3.1. Further investigation is
required to understand the failure mechanism at higher strain levels using post-yield
strain gauges of a larger strain capacity.
3.6 Prediction of Flexural Strength and Rotation Capacity

The flexural behavior and failure mechanism of reinforced HPFRCC is significantly
different than the conventional reinforced concrete as demonstrated by several exper-
imental studies [7–10]. A recently proposed analytical model (Fig. 6) was used to
predict flexural strength and rotation capacity at different damage states [10, 20]. The
predictability of the analytical model was measured by comparison with the experi-
mentally obtained values. The flexural strengths and curvatures at various damage
states were computed assuming linear strain distribution and using cross section
properties of the specimens. The beams were assumed to reach the yield level when the
tensile reinforcement strain reached the yield value (i.e., ey Þ: The analytical model
shown in Fig. 6(b) considers a tensile stress block contribution which is ignored in the
flexural calculation of conventional concrete components. Elastic deflection theory was
used to compute yield rotation using the yield curvature value as shown in Eq. (2). It
can be observed from Table 3 that the prediction ratio of both the parameters are close
to 1.00, which indicates that the analytical formulation can be successfully used to
compute yield rotation and moment capacity of reinforced UHPC flexural members.
Fig. 6. Analytical model for section analysis (a) cross-section (b) stress distribution at yield
level [20] (c) stress distribution at nominal level using modified Hognestad stress block [10]
(d) simplified stress distribution at nominal and ultimate level using Whitney stress block [20].
Two analytical models were used to estimate the nominal moment capacity: one
with modified Hognestad compression stress block (Fig. 6(c)) and the second with
simplified Whitney compression stress block (Fig. 6(d)). Both models considered
localized hardening of the reinforcement bar as observed in tension stiffening
Table 3. Comparison of experimental and analytical results at yield level

SpecimenYield Rotation [rad] Yield Moment [kNm]
Exp. Ana. Ana./Exp. Exp. Ana. Ana./Exp.
UHPC-1% 0.0098 0.0095 0.97 58 53 0.92
UHPC-2% 0.0095 0.0098 1.03 60 60 1.00
experiments [14, 15]. The beams were assumed to the reach nominal level when the
strain in the compression zone reached 3% [10] or tensile reinforcement strain reached
ultimate value (eu Þ [20]. It can be observed from Table 4 that the nominal moment
predictability using a modified Hognestad stress block is better compared to the use of
simplified rectangular Whitney stress block; however, both give reasonable estimates
of strength.
Table 4. Comparison of experimental and analytical results at nominal and ultimate level.
Specimen Nominal Moment [kNm] Ultimate Rotation [rad]
Exp. Ana.a Ana.b Ana./Exp.a Ana./Exp.b Exp. Ana. Ana./Exp.
UHPC-1% 76 75 80 0.98 1.06 0.1002 0.083 0.83
UHPC-2% 72 75 81 1.03 1.12 0.1015 0.060 0.59
a
Figure 6(c) [10]; bFigure 6(d) [20]
The ultimate rotation capacity was computed using Eq. (2).
1
hu ¼ hy þ hp ¼ /y Ls þ ð/u /y ÞLp ð2Þ
2
Where Ls is the shear span length (mm), /y and /u are the section curvatures of the
structural member at yield level (mm−1) and collapse level (mm−1), respectively. In the
above equation, Lp is the equivalent plastic hinge length (mm), which was computed
using a recently developed expression based on a range of HPFRCC materials as
shown in Eq. (3) [20].
0:24qfy
Lp ¼ 0:02Ls þ ð3Þ
ft
Where q is the longitudinal reinforcement (%), fy is the yield stress (MPa), and ft is
the tensile strength of UHPC mixture (MPa). The analytical framework underestimated
the ultimate rotation capacity in both specimens (Table 4). The reason for this dis-
crepancy is the underestimation of equivalent plastic hinge length values for reinforced
UHPC specimens tested in this experiment. The expression shown in Eq. (3) was
developed using reinforced HPFRCC beams with typical reinforcement ratio (i.e.,
0:70%\q\1:90%) and a maximum tensile strength of 8 MPa. As such, the majority
of the specimens followed the failure after crack localization path with damage
localization in a dominant crack. In the current experiment, both the specimens fol-
lowed failure after gradual strain hardening failure path due to the use of a high
reinforcement ratio. The damage was uniformly distributed over longer length as seen
from the crack patterns. Therefore, there is a need to further improve the plastic hinge
length expression for highly reinforced UHPC beams (i.e., q [ 2:0%) using a more
rigorous parametric investigation.
4 Conclusions
This study provides valuable insight about the formation of plastic hinges with vari-
ation in fiber volume fraction. The following conclusions can be drawn from this study:
• The variation in fiber content does not impact ultimate rotation capacity in rein-
forced UHPC beams with high longitudinal reinforcement ratio (q > 2.0%).
• The length of plasticity in the longitudinal reinforcement increases with a decrease
in fiber volume fraction because of the formation of multiple distributed flexural
cracks along the plastic hinge region.
• Distribution of visible cracks in the specimens indicated that the damage is much
more localized in specimens with higher fiber content (Lmajorcracks = 240 mm) than
those with low fiber content (Lmajorcracks = 420 mm) because the UHPC matrix with
high fiber content had higher tensile strength, bond strength, and fracture energy
which restrained the formation of splitting cracks and prevented opening of addi-
tional flexural cracks.
• A parametric study with a wider variation in fiber content at high reinforcement
ratios is necessary to further improve a recently developed plastic hinge length
expression such that ultimate rotation capacity can be computed with higher
accuracy.
Acknowledgements. The authors gratefully acknowledge the financial support from the John A.
Reif, Jr., Department of Civil and Environmental Engineering at New Jersey Institute of Tech-
nology and John A. Blume Earthquake Engineering Center at Stanford University. The authors
would also like to recognize King Packaged Materials Company, Boisbriand, QC for providing
materials necessary for the tests.
References
1. Graybeal, B.A.: Material property characterization of ultra-high performance concrete.
FHWA, Virginia, USA, Technical report, No. FHWA-HRT-06-103, pp. 1–176 (2006)
2. Yoo, D., Yoon, Y.: A review on structural behavior, design, and application of ultra-high-
performance fiber-reinforced concrete. Int. J. Concrete Struct. Mater. 10(2), 125–142 (2016)
3. Ichikawa, S., Matsuzaki, H., Moustafa, A., Elgawady, M.A., Kawashima, K.: Seismic-
resistant bridge columns with ultra high-performance concrete segments. J. Bridge Eng. 21
(9), 04016049 (2016)
4. Rokugo, K., Kanda, T., Yokota, H., Sakata, N.: Applications and recommendations of high
performance fiber reinforced cement composites with multiple fine cracking (HPFRCC) in
Japan. Mater. Struct. 42(9), 1197–1208 (2009)
5. Pokhrel, M., Bandelt, M.J.: Material properties and structural characteristics influencing
deformation capacity and plasticity in reinforced ductile cement-based composite structural
components. Compos. Struct. 224, 111013 (2019)
6. Pokhrel, M., Bandelt, M.J.: Plastic hinge region and rotation capacity in reinforced HPFRCC
flexural members at collapse level. In: Proceedings of the Seventh International Colloquium
on Performance, Protection & Strengthening of Structures under Extreme Loading & Events,
Whistler, Canada, pp. 1–15 (2019)
7. Yoo, D.Y., Yoon, Y.S.: Structural performance of ultra-high-performance concrete beams

with different steel fibers. Eng. Struct. 102, 409–423 (2015)
8. Bandelt, M.J., Billington, S.L.: Impact of reinforcement ratio and loading type on the
deformation capacity of high-performance fiber-reinforced cementitious composites rein-
forced with mild steel. J. Struct. Eng. 142(10), 04016084 (2016)
9. Bandelt, M.J., Billington, S.L.: Simulation of deformation capacity in reinforced high-
performance fiber-reinforced cementitious composite flexural members. J. Struct. Eng. 144
(10), 04018188 (2018)
10. Shao, Y., Billington, S.L.: Flexural performance of steel-reinforced engineered cementitious
composites with different reinforcing ratios and steel types. Constr. Build. Mater. 231,
117159 (2020)
11. Bandelt, M.J., Billington, S.L.: Bond behavior of steel reinforcement in high-performance
fiber-reinforced cementitious composite flexural members. Mater. Struct. 49(1–2), 71–86
(2014)
12. Lee, J.: Bonding behavior of lap-spliced reinforcing bars embedded in ultra-high strength
concrete with steel fibers. KSCE J. Civil Eng. 20(1), 273–281 (2016)
13. Dagenais, M.A., Massicotte, B.: Cyclic behavior of lap splices strengthened with ultra high
performance fiber-reinforced concrete. J. Struct. Eng. 143(2), 04016163 (2017)
14. Moreno, D.M., Trono, W., Jen, G., Ostertag, C., Billington, S.L.: Cement & Concrete
Composites Tension stiffening in reinforced high performance fiber reinforced cement-based
composites. Cement Concr. Compos. 50, 36–46 (2014)
15. Hung, C., Lee, H., Nga, S.: Tension-stiffening effect in steel-reinforced UHPC composites:
Constitutive model and effects of steel fibers, loading patterns, and rebar sizes. Compos.
B Eng. 158, 269–278 (2019)
16. Shao, Y., Billington, S.L.: Utilizing full UHPC compressive strength in steel reinforced
UHPC beams. In: Proceedings of Second International Interactive Symposium on Ultra-High
Performance Concrete, Albany, New York, USA, pp. 1–9 (2019)
17. Shao, Y., Billington, S.L.: Predicting the two predominant flexural failure paths of
longitudinally reinforced high-performance fiber-reinforced cementitious composite struc-
tural members. Eng. Struct. 199, 109581 (2019)
18. ASTM: Standard practice for fabricating and testing specimens of ultra-high performance
concrete. ASTM C1856/C1586-17, pp. 1–4 (2017)
19. Pokhrel, M., Bandelt, M.J.: Simulation of reinforced HPFRCC deformation capacity under
flexure- and shear-dominated stress states. In: Proceedings of Computational Modelling of
Concrete and Concrete Structures, Bad Hofgastein, Austria, pp. 633–640 (2018)
20. Pokhrel, M., Bandelt, M.J.: Plastic hinge behavior and rotation capacity in reinforced ductile
concrete flexural members. Eng. Struct. 200, 109699 (2019)
An Eco-Friendly UHPC for Structural
Application: Tensile Mechanical Response
Amin Abrishambaf(&), Mário Pimentel, and Sandra Nunes
CONSTRUCT-LABEST, Faculty of Engineering (FEUP), University of Porto,

Porto, Portugal
aminab@fe.up.pt
Abstract. This paper presents and discusses experimental results on the tensile
mechanical performance of a newly developed ultra-high performance cemen-
titious material, UHPC, incorporating spent equilibrium catalyst (ECat), a waste
generated by the oil refinery industry, as a supplementary cementitious material.
The results are compared to a previously developed conventional UHPC. The
influence of ECat on the heat of hydration in UHPC is evaluated by isothermal
calorimeter under a constant temperature of 20 °C. To determine the evolution
of the tensile behaviour with time, a series of uniaxial tensile tests are performed
on the specimens at different ages, i.e. 1, 3, 7, 28 and 91 days after casting.
Afterwards, the fibre to matrix interfacial bond properties were characterized by
executing a series of single fibre pullout tests at the age of 28 days on the steel
fibres embedded in UHPCs with 0°, 30° and 60° orientation angles. The results
confirmed the adequate performance of the new developed UHPC for the
structural application.
Keywords: Ultra-high performance fibre reinforced cementitious material

Tensile behaviour Spent equilibrium catalyst Fibre to matrix bond
properties Fibre orientation
1 Introduction
Ultra-high performance fibre reinforced cementitious composites (UHPFRC) designate

a family of advanced cementitious materials with enhanced matrix packing density,
very low water/binder ratio (w/b < 0.2), and containing a significant amount of high
strength short and discrete steel fibres. This combination provides superior perfor-
mances in terms of compressive strength, energy absorption capacity, ductility and
durability [1–4]. Cementitious matrix of UHPFRC, designated as UHPC, exhibits large
shrinkage values, mainly in the first days after casting, in which, unlike the conven-
tional concrete, a large portion corresponds to the autogenous shrinkage [5].
Spent Equilibrium Catalyst (ECat) is a waste material generated by the oil refinery
industry. ECat was shown to act as an internal curing agent, and its incorporation in the
matrix of UHPFRC allowed reducing the autogenous shrinkage of the composite. ECat
has a high porous microstructure which provides a high specific surface area with a
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_93
An Eco-Friendly UHPC for Structural Application 1057
high water absorption capacity. Therefore, during the first day of hydration, the
absorbed water in the porous ECat particles is released in the UHPC matrix and provide
a beneficial effect to mitigate autogenous shrinkage. As supplementary cementitious
material, ECat, including the one generated in Portugal and used in this research,
exhibits high pozzolanic activity. Chapelle test results revealed that 1540 mg of Ca
(OH)2 is consumed per gram of ECat during the pozzolanic action which is close to SF
pozzolanic activity, i.e. 1577 mg [6].
This paper presents the experimental results on the mechanical performance
(mainly tensile behaviour) of the new developed UHPC, incorporating spent equilib-
rium catalyst (ECat), as a supplementary cementitious material. The results are com-
pared to a conventional UHPC previously developed. The influence of ECat on the heat
of hydration in UHPC is evaluated by isothermal calorimeter under a constant tem-
perature of 20 °C. To determine the evolution of the tensile behaviour with time, a
series of uniaxial tensile tests are performed on the specimens at different ages.
Afterwards, the fibre to matrix interfacial bond properties was characterized by exe-
cuting a series of single fibre pullout test on the fibres with different orientation angles.
2.1 Materials and Mixtures

Cement CEM I 4.25 R, limestone filler, Spent Equilibrium Catalyst (ECat) and silica
fume (SF) with a specific surface area of 3110, 2680, 2660 and 2200 m2/kg were used
in the preparation of the mixtures. ECat is generated by Sines Refinery, Portugal and
shows about 30% water absorption capacity by mass at 24 h and at the saturated
surface-dry basis. Figure 1 shows SEM images of ECat particles. The ECat’s water
absorption was considered and added to the total water used in the UHPC mixture. The
chemical compositions of CEM, ECat and SF used in this research are shown in
Table 1. Siliceous natural sands with the specific surface area of 2630 m2/kg and 0.3%
absorption were adopted in the mixtures. To guarantee the self-compacting property, a
superplasticizer based on polycarboxylate ethers has been used. The mix compositions
are shown in Table 2. In this table, UHPC and ECat_UHPC designate the conventional
and the new developed mixture incorporating ECat [6]. A low fibre volume fraction of
0.5% is used to decrease the brittleness of UHPC mixtures during mechanical tests.
However, this fibre volume fraction is low enough not to affect the mechanical prop-
erties of plain UHPC. The steel fibres are high strength, short and smooth brass-coated
with a length, lf, of 13 mm and a diameter, df, of 0.2 mm. Both mixtures exhibited a
slump flow diameter of 280 and 282 mm, respectively, without any compaction
energy; and close 28-day compressive strength between 140–150 MPa, without
heat/pressure treatment.
1058 A. Abrishambaf et al.
(a) (b)
Fig. 1. SEM image ECat’s particle morphology in secondary electron mode: (a) general view of
particles (250) and (b) internal porosity (10000).
Table 1. Chemical compositions of the mix constituents (% by mass).

SiO2 Al2O3 Fe2O3 CaO MgO SO3 Na2O K2O LoI
CEM 19.80 5.08 3.16 61.31 1.82 2.90 0.15 0.58 2.54
ECat 40.30 54.45 0.45 0.06 0.15 0.00 0.43 0.02 1.05
SF >90 <3.0
Table 2. Mix compositions (kg/m3).

Mix Cement SF ECat Filler Water SP Sand Fibres
UHPC 794.90 39.74 – 311.43 185.20 30.03 1005.21 39.25
ECat_UHPC 690.19 33.56 155.45 250.58 207.50(*) 19.48 839.26 39.25
(*) the absorption water of ECat is included.
2.2 Specimens and Test Setups

2.2.1 Heat of Hydration
To determine the influence of ECat particles on the heat of hydration, JAF60 isothermal
calorimeter [7] was used. The calorimeter tests were performed at a constant temper-
ature of 20 °C. In each mixture, Table 2, the sand particles were excluded. After
mixing, pastes were poured inside a plastic bag, which was thereafter carefully sealed
by thermoplastic welding. The weight of each sample was determined in order to
guarantee that it contains 30 gr of cement. The measurement was started 30 min after
adding water to the dry powders and kept reading for a period of 3 days. For each
mixture, 4 samples were tested.
JAF60 isothermal calorimeter details are shown in Fig. 2. Each calorimetric unit
contains a polystyrene element to isothermally insulate the sample container. The
calorimetric units are placed inside the principal bath under a controlled temperature.
More details of this setup can be found in [7].
Principal bath
Secondary bath
Cooling unit
Calorimetric units
Fig. 2. JAF60 isothermal calorimeter.
2.2.2 Uniaxial Tensile Test

To evaluate the tensile behaviour of UHPCs at different ages, dog-bone shaped
specimens were cast with the dimensions presented in Fig. 3(a). The specimens have a
cross-section of 30 mm 40 mm in the constant length area. The specimens were cast
by moving the material container throughout the mould to prevent any preferential
orientation of fibres. After 24 h, the specimens were demoulded and cured underwater
until the day of performing the test. To investigate the evolution of the tensile beha-
viour of UHPCs with time, the following ages were studied: 1, 3, 7, 28 and 91 days
after casting. For each age and UHPC mixture, five specimens were cast which leads to
the production of 50 specimens in total.
Fig. 3. Specimen’s geometry and dimensions of the uniaxial tensile test.
The uniaxial tensile tests were performed in an Instron universal testing machine
with a 300 kN capacity load cell using hinge-hinge supporting conditions in both
extremities. Figure 4(a) shows the uniaxial tensile test setup. Special care was taken to
centre and align the specimen in the setup before running the test. The tests were
carried out under closed-loop displacement control with a displacement rate of
0.003 mm/s and were kept constant during the test. An LVDT was installed on each
side of the specimen to record elongation in a base length of 92 mm during the test.
LVDT’s connection detail is depicted in Fig. 4(b).
(a) (b)
Fig. 4. Uniaxial tensile test setup: (a) general view and (b) LVDTs’ connection.
2.2.3 Fibre Pullout Test

To characterize the interfacial bond properties between fibre and matrix, a series of
single fibre pullout tests were performed on smooth brass-coated steel fibres (as
commercially used in the production of UHPFRC, see Sect. 2.1 for the fibre charac-
teristics) embedded in UHPC and ECat_UHPC. The fibre pullout test specimens
included fibres with an embedded length of le = lf /2 and fibre orientation angles 0°, 30°
and 60°. To prepare the specimens, firstly, the fibres were embedded with a desired
length and orientation in a thick polystyrene foam layer, see Fig. 5(a). Afterwards, a
wood frame was bolted to the polystyrene foam, and the matrix was then cast from the
centre of the moulds, see Fig. 5(b). The specimens were demoulded 24 h after casting
and cured inside a chamber room with 20 °C and 95% RH. For each series, 7 speci-
mens were prepared. The tests were performed after 28 days of curing.
The test was performed in an Instron universal testing machine under a closed-loop
displacement control with a constant displacement increment of 0.01 mm/s. To achieve
the necessary accuracy in the load measurements, an external load cell with a maximum
load-bearing capacity of 500 N was used. To record the displacement, an LVDT was
installed on the backside of the clamping system. The test setup is depicted in Fig. 5(d)
and more details are discussed in [4].
(a) (b)
Load cell
LVDT
Clamping system
Holding plate
Specimen
(c) (d)
Fig. 5. Fibre pullout test: (a) embedding fibres in polystyrene foam, (b) casting specimen,
(c) final view of a specimen and (d) test setup.
3.1 Heat of Hydration
6 300
UHPC
Heat evolution rate (W/kg)
Cumulative heat (kJ/kg)
5 ECat_UHPC 250
4 200
3 150
2 100
1 50 UHPC
ECat_UHPC
0 0
0 12 24 36 48 60 72 0 12 24 36 48 60 72
Time (h) Time (h)
(a) (b)
Fig. 6. Heat evolution rate of UHPCs.
Figure 6(a) shows the rate of heat evolution of two UHPCs. As it can be seen, by
incorporating ECat in UHPC, the duration of the dormant period decreases from 5.7 to
2.6 h. As hydration proceeded, the hydration peak of UHPC reduces from 4.61 to 3.83
(W/kg) and the time to reach the hydration peak becomes shorter from 11.28 to 8.08 (h),
for UHPC and ECat_UHPC, correspondingly. Besides, from Fig. 6(b), it is observed
that up to 12 h, cumulative heat increases roughly in ECat_UHPC with a higher rate and
it seems that after 72 h it still shows an upward trend compared to UHPC which is
approximately stabilized. This means that, in ECat_UHPC, the hydration reaction was
not entirely promoted up to this time. This is possible because the ECat water retention
capacity can release the absorbed water inside UHPC for longer than 3 days, or the
already released water was not fully used for the hydration reaction until 3 days.
3.2 Uniaxial Tensile Behaviour

Figure 8 shows tensile stress-strain relationships (up to the tensile strength) obtained
from uniaxial tensile behaviour of each UHPC mixture at different ages. The strain
value was determined by dividing the average elongation recorded in LVDTs by the
measuring length (92 mm). In Fig. 8, only the results of the valid tested specimens
were included, that is, those in which the crack was localized in the measuring length
covered by the LVDTs. Strong linear stress-strain relationship was observed for all the
specimens. For the specimens tested at early ages, i.e. 1 and 3 days, a higher scattering
in the tensile behaviour between the same specimens is evidenced. This scattering can
be expected since the hydration of the binder is undergoing. Once a crack is formed, an
abrupt reduction in load was observed. This part was excluded from the figures due to
the lack of interest. However, the post-cracking tensile behaviour of the new developed
ECat_UHPC incorporating 3% fibre volume factions is shown in Fig. 7(f). This
12 12
UHPC UHPC
10 10
ECat_UHPC ECat_UHPC
8 8
6 6
4 4
2 2
0 0
0.0000 0.0002 0.0004 0.0006 0.0008 0.0000 0.0002 0.0004 0.0006 0.0008
Strain (-) Strain (-)
(a) (b)
12 12
UHPC UHPC
10
10
ECat_UHPC ECat_UHPC
8 8
6 6
4 4
2 2
0 0
0.0000 0.0002 0.0004 0.0006 0.0008 0.0000 0.0002 0.0004 0.0006 0.0008
(c) (d)
Fig. 7. Evolution of the tensile stress-strain behaviour of UHPCs with time: (a) 1 day, (b) 3
days, (c) 7 days and (d) 28 days.
12 20
UHPC 18
10 ECat_UHPC 16

14
8
12
6 10
8
4
6
2 4 Well oriented
Randomly oriented
2
Average curve
0 0
0.0000 0.0002 0.0004 0.0006 0.0008 0.000 0.001 0.002 0.003 0.004 0.005
(e) (f)
Fig. 8. (continiue) Evolution of the tensile stress-strain behaviour of UHPCs with time: (e) 91
days and (f) tensile response of ECat_UHPFRC at the age of 28 days with Vf= 3.0%.
confirms an adequate tensile performance of the new developed mixture for structural
application. More details about this discussion can be found elsewhere [8].
Figure 9 depicts the development of the average tensile strength and Young’s
modulus of elasticity for UHPC and ECat_UHPC with time. The modulus of elasticity
was determined as the slope of the best line fitted to the data from origin up to a stress
value corresponds to 60% of the tensile strength. Up to 3 days, both mixtures show a
very close tensile strength; however, afterwards the development of the tensile strength
was faster in UHPC. On the other hand, at 7 days, 83.8% and 72.5% of the corre-
sponding tensile strength at 90 days was achieved for UHPC and ECat_UHPC,
respectively. This is reasonable since UHPC mixture contains a slightly higher SF
content, which is highly reactive at early ages. Therefore, a high hydration degree was
already reached and further hydration becomes minimal, see Fig. 6. Even though, after
90 days, ECat_UHPC shows only 0.86 MPa lower tensile strength than UHPC. Similar
to tensile strength, the evolution of modulus of elasticity was faster in UHPC. In the
early age, ECat_UHPC shows a slightly lower modulus of elasticity. However, it is
worth noting that after 91 days of water curing, ECat_UHPC shows a very close
modulus of elasticity to UHPC. This confirms that, unlike UHPC, in the ECat_UHPC
the hydration reaction was continuously promoted until later ages, see Fig. 6.
12 45
Modulus of elasticity (GPa)
Tensile strength (MPa)
10 40
8 35
6 30
4 25
2 UHPC 20 UHPC
ECat_UHPC ECat_UHPC
0 15
0 20 40 60 80 100 0 20 40 60 80 100
Time after casting (days) Time after casting (days)
(a) (b)
Fig. 9. Development of the average: (a) tensile strength and (b) modulus of elasticity with time.
3.3 Fibre to Matrix Interfacial Bond Properties

Figure 10 shows the interfacial bond properties determined from fibre pullout test. To
characterize the interfacial bond properties, the following parameters were assessed: 1)
debonding strength, sdeb: , where the chemical bond between the fibre and matrix was
destroyed and a small reduction in the load is observable; 2) average bond strength,
savg: , which corresponds to the maximum load recorded during the test. In the case of
UHPC, this parameter can be named as frictional bond strength as well, since after
chemical debonding a slip hardening behaviour can be achieved, particularly in the
case of inclined fibres; 3) pullout work, Wp , determined as the area under force – slip
curve. These parameters can be determined by the following equations:
Pdeb:
sdeb: ¼ ð1Þ
pdf le
Pmax:
savg: ¼ ð2Þ
pdf le
Z
Wp ¼ s¼le
s¼0
PðsÞds ð3Þ
where, Pdeb: is the load at the fibre debonding stage and Pmax: is the maximum load
recorded during the test. Two failure modes were observed during the tests: the fibres
were mainly pullout, however, in the case of the inclined fibres, few fibres were ruptured
as well. It should be mentioned that the fibre pullout tests were performed on the fibres
with a maximum embedded length which can increase the risk of fibre rupturing.
However, in the composite level, this failure mode is expected to rarely occur.
fibre rupture 600

20 UHPC Avg.UHPC 20 UHPC Avg.UHPC
Pullout work (N.mm)
ECat_UHPC Avg.ECat_UHPC 500 ECat_UHPC Avg.ECat_UHPC

16 16
τdeb. (MPa)
τavg. (MPa)
400
12 12
300
8 8
200
fibre rupture
4 4 UHPC Avg.UHPC
100
ECat_UHPC Avg.ECat_UHPC
0 0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60 0 10 20 30 40 50 60
Fibre orientation angle (degree) Fibre orientation angle (degree) Fibre orientation angle (degree)
(a) (b) (c)
Fig. 10. Interfacial bond properties: (a) debonding load, (b) maximum load and (c) pullout work.
A significant improvement in the interfacial bond properties of the inclined fibres

(30° and 60°) compare to the aligned fibres (0°) was observed in both mixtures. This
improvement is due to the snubbing effect that improves the frictional bond between
the fibre and matrix and imposes severe damage on the brass coated surface of the
inclined steel fibres. The fibre surface morphology analysis confirms a high level of
damage on the surface of the inclined fibres compared to the aligned fibres, see Fig. 11.
The fibre rupturing mainly occurred in the inclined fibres embedded in UHPC mixture.
In some specimens of this series, the matrix spalling was not formed in the fibre exit
point to overwhelms the snubbing effect, therefore fibres were ruptured. This is
expected since the UHPC mixture shows a slightly higher tensile strength at 28 days
compared to ECat_UHPC, see Fig. 9(a).
A similar sdeb: is obtained for all the fibre inclination angles embedded in both
mixtures. However, a slightly higher savg: and pullout work are achieved for the fibres
embedded in ECat_UHPC mixture. It is worth noting that the inclined fibres embedded
in ECat_UHPC mixture show savg: close to the fibre rupturing, however, unlike the
UHPC mixture, only a few fibres ruptured in this series (only in 60° inclination), since
the matrix spalling occurred in the fibre exit point to reduce the snubbing effect. SEM
analysis on the fracture surface of ECat_UHPC matrix confirms that the ECat particles
are either broken or debonded from the matrix, see Fig. 12. These extra abraded fine
particles are squeezed against the surface of the fibre when the fibre moves in its matrix
channel, therefore, increasing scratching and peeling off the fibre brass coating. This
contributes to higher interfacial shear resistance. Fibre surface morphology analysis
also confirmed this fact and demonstrates slightly higher friction between the fibre and
ECat_UHPC, see Fig. 11.
(a) (b) (c)
(d) (e)
Fig. 11. SEM image fibre surface morphology in secondary electron mode (500): (a) virgin
fibre, (b) UHPC mixture and aligne fibre, (c) UHPC mixture and inclined fibre, (d) ECat_UHPC
mixture and aligned fibre, and (e) ECat_UHPC mixture and inclined fibre.
Broken ECat
Dedonded ECat
(a) (b)
Fig. 12. SEM analyses on the fracture surface of ECat_UHPC mixture in secondary electron
mode.
4 Conclusions
This paper presents the experimental results on a new developed ultra-high perfor-
mance fibre reinforced cementitious material, UHPC, incorporating spent equilibrium
catalyst (ECat). The results are compared to a conventional UHPC previously
developed.
The calorimetry results show that by incorporating ECat in UHPC the duration of
dormant period decreases from 5.7 to 2.6 h. As hydration proceeded, the hydration
peak of UHPC reduces from 4.61 to 3.83 (W/kg) and the time to reach the hydration
peak becomes shorter from 11.28 to 8.08 (h), for UHPC and ECat_UHPC,
correspondingly.
Up to 3 days, ECat_UHPC shows a very close tensile strength to UHPC; however,
the development of the tensile strength was faster in the earlier ages in UHPC. On the
other hand, at 7 days, 83.8% and 72.5% of the corresponding tensile strength at 90 days
was achieved for UHPC and ECat_UHPC, respectively. Similarly, the evolution of the
modulus of elasticity was faster in UHPC. In the early ages, ECat_UHPC shows a
slightly lower modulus of elasticity. However, it is worth noting that after 91 days of
water curing, ECat_UHPC shows a very close modulus of elasticity to UHPC.
Concerning the fibre to matrix bond properties, similar debonding strength is
obtained for all the fibre inclination angles embedded in both mixtures. However, a
slightly higher average bond strength and pullout work are achieved for the fibres
embedded in ECat_UHPC mixture. SEM analysis on the fracture surface of ECa-
t_UHPC matrix confirms that the ECat particles are either broken or debonded from the
matrix. These extra abraded fine particles are squeezed against the surface of the fibre
when the fibre moves in its matrix channel, therefore, increasing scratching and peeling
off the fibre brass coating. This contributes to higher interfacial shear resistance.
In general, marginal differences between the tensile and fibre to matrix bond
properties of ECat_UHPC and UHPC were observed. This seems to confirm the
adequate performance of the new ECat_UHPC mixture for structural applications
which can lead to 13.2%, 15.6%, 19.5% and 17.6% reduction in the content of cement,
silica fume, filler and sand, respectively.
Acknowledgements. This work was financially supported by: Base Funding -

UIDB/04708/2020 and Programmatic Funding - UIDP/04708/2020 of CONSTRUCT -Instituto
de I&D em Estruturas e Construções funded by national funds through the FCT/MCTES
(PIDDAC); and by the project PTDC/ECI-EST/31777/2017 – “UHPGRADE - Next generation
of ultra-high performance fibre-reinforced cement based composites for rehabilitation and
strengthening of the existing infrastructure” funded by FEDER funds through COMPETE2020-
Programa Operacional Competitividade e Internacionalização (POCI) and by national funds
(PIDDAC) through FCT/MCTES. Collaboration and materials supply by Concremat, Secil,
Omya Comital, Sika and Bekaert is gratefully acknowledged.
References
1. Yoo, D.Y., Lee, J.H., Yoon, Y.S.: Effect of fibre content on mechanical and fracture properties
of ultra high performance fibre reinforced cementitious composites. Compos. Struct. 106,
742–753 (2013)
2. Charron, J.P., Denarié, E., Brühwiler, E.: Permeability of ultra high performance fiber
reinforced concretes (UHPFRC) under high stresses. Mater. Struct. 40, 269–277 (2007)
3. Abrishambaf, A., Pimentel, M., Nunes, S.: Influence of fibre orientation on the tensile
behaviour of ultra-high performance fibre reinforced cementitious composites. Cem. Concr.
Res. 97, 28–40 (2017)
4. Abrishambaf, A., Pimentel, M., Nunes, S.: A meso-mechanical model to simulate the tensile
behaviour of ultra-high performance fibre reinforced cementitious composites. Compos.
Struct. 222, 110911 (2019)
5. Fehling, E., Schmidt, M., Walraven, J., Leutbecher, T., Frohlich, S.: Ultra-High Performance
Concrete UHPC: Fundamentals, Design, Examples. Wiley, Berlin (2014)
6. Matos, A.M., Nunes, S., Costa, C., Barroso-Aguiar, J.L.: Spent equilibrium catalyst as
internal curing agent in UHPFRC. Cem. Concrete Compos. 104, 103362 (2019)
7. Wexham Developments: JAF Calorimeter - Operating Manual, 7th edn. United Kingdoom
(2005)
8. Abrishambaf, A., Pimentel, M., Nunes, S.: Tensile behaviour of an ultra-high performance
fibre reinforced cementitious composites incorporating Spent Equilibrium Catalyst. In: 5th
International Symposium on Ultra-High Performance Concrete and High Performance
Construction Materials, Proceedings of an International Conference, Germany, March 2020
Characterization of Ultra High Performance
Fiber Reinforced Concrete (UHPFRC) Tensile
Behaviour
Nicola Generosi(&), Jacopo Donnini, and Valeria Corinaldesi
Università Politecnica delle Marche, Ancona, Italy

n.generosi@pm.univpm.it
Abstract. Ultra-high performance fiber reinforced concrete (UHPFRC) is

considered a promising material for many applications where high compressive
and tensile strength, small thickness and high energy absorption capacity are
required. However, although these materials were introduced in the mid-1990s,
a comprehensive investigation regarding its tensile characterization is still par-
ticularly challenging. Different tensile test setups have been used by many
researchers, in order to obtain reliable results, but today there are currently no
testing standards available that define test conditions, specimen geometry and
analytical procedures to fully characterize the tensile properties of UHPFRC.
In this study, the tensile behaviour of UHPFRC has been investigated with
direct tensile tests on dog-bone specimens, using a gripping system with rotating
boundary conditions. Tests have been performed in displacement control.
Digital Image Correlation (DIC) has been used to measure displacements,
deformations, number and width of cracks in experimental testing.
The effect of hooked steel fibers with diameter of 0.38 mm and length of
30 mm on the tensile behaviour of UHPFRC has been investigated, varying the
amount of fibers from 0% up to 2.6% by volume. The fiber volume fraction
greatly influences the tensile strength of the material, strain at peak strength,
energy absorption capacity and post-cracking behaviour.
Keywords: Digital image correlation Fiber reinforced concrete Steel fibers

UHPFRC
1 Introduction
The interest of the scientific community on the development of Ultra-High-Performance

Fiber Reinforced Concrete (UHPFRC) has been significantly increased in recent years.
UHPFRCs are a new generation of cementitious composites with excellent mechanical
properties. To achieve the ultra-high compressive strength, fine powders, such as quartz
flour and silica fume are normally used in mixture to ensure a highly dense matrix [1–4].
The benefits of fiber reinforcement in improving the fracture toughness, impact
resistance, fatigue endurance and energy absorption capacity of concrete are well
known. Fibers resist the nucleation of cracks by acting as stress-transfer bridges, and
once cracks nucleate, fibers abate their propagation by providing crack tip plasticity
and increased fracture toughness [5].
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_94
Characterization of Ultra High Performance Fiber Reinforced Concrete (UHPFRC) 1069
UHPFRC can be classified in tension as “strain-softening” or “strain-hardening”. In

the case of higher fiber volume fraction (for example Vf > 2%), UHPFRC specimens
subjected to uniaxial tensile loads show a strain hardening behaviour after the formation
of the first crack. As a result, the tensile strength of the composite is higher than that at
first cracking [6]. Figure 1 illustrates a simplified response of strain-hardening Fiber
Reinforced Concrete (FRC). This idealized modelling approach distinguishes the tensile
behaviour into three parts: Part I (Elastic behaviour up to cracking strength cc); Part II
(Strain hardening behaviour with multiple cracking); Part III (Softening behaviour).
Fig. 1. Idealized response of strain-hardening FRC in tension
There are already few national recommendations on the Design and Construction
with UHPFRC: the French AFGC-SETRA [7] and the Japanese JSCE [8] both provide
recommendations on how to perform uniaxial tensile tests on UHPFRC materials. The
Fib Model Code 2010 [9], which serves as a basis for the future code for concrete
structures, also provides important information on the mechanical properties of Fiber
Reinforced Concrete (FRC) elements. The Italian CSLP (Consiglio Superiore dei Lavori
Pubblici) recently published a guideline about qualification, technical assessment cer-
tification and acceptance control of FRC. It also specifies how to test the flexural and
direct tensile strength of FRC. Further details on test methods can be found in [10].
UHPFRC material is under detailed exploration: many experts from all over the
world presented their research results regarding its tensile characterization. Wille et al.
[11] experimentally investigated the direct tensile behaviour of nine dog-bone different
UHPFRC specimens (by varying fiber type and fiber volume fraction). The analysed
results showed, as expected, a strong dependency on fiber volume fraction. In Yoo D.
et al. [12] the effects of steel fiber type on the tensile performance of UHPFRC were
investigated. Four different steel fiber types were used: S-straight, T-twisted, H-hooked,
and HH-half-hooked. The order of effectiveness in enhancing the tensile performance
of dog-bone specimens was S-fibers > T-fibers > HH-fiber > H-fiber.
Direct tensile tests are challenging to perform, since it is difficult to achieve
homogeneously distributed stress throughout the specimen cross section and to control
a stable load versus displacement/crack opening response [13].
Furthermore, specimen’s geometry and dimension, together with the gripping
methods used during the test, can significantly influence the test results. Different
1070 N. Generosi et al.
tensile test setups have been proposed by researcher to evaluate the direct tensile
behaviour of cementitious materials [11, 14, 15].
The objective of this research work is to investigate the mechanical properties of
UHPFRC reinforced with different amount of hooked steel fibers, from 0.6% up to
2.6% by volume. First, the material has been characterized by means of compression
and flexural tests. Then, direct tensile tests have been performed on dog-bone shape
specimens to fully characterize their tensile properties and to evaluate the effect of
different amounts of hooked steel fibers both on the tensile strength, deformation
capacity and softening or hardening behaviour.
A Commercial Portland-limestone blended cement (CEM I 52.5 R), in compliance with

EN-197/1 [16] was used. The Blaine fineness of cement was 0.48 m2/g and its specific
gravity was 3.15 g/cm3. As aggregate two different types of quartz sand with particle size
0–0.6 mm and 0.6–1.0 mm were suitably combined. Silica fume was added at a dosage of
125 kg/m3. In addition, an acrylic-based water-reducing admixture (WRA) was added in
powder at dosage of 1.1% by weight of cement with a water to cement ratio (w/c) of 0.20.
Hooked steel fibers (Fig. 2), 30-mm long with aspect ratio equal to 80, were added at
increasing dosages: 50, 100, 150 and 200 kg/m3. The corresponding fibers volume frac-
tions are 0.6%, 1.25%, 1.9% and 2.6%. The mixture proportions are reported in Table 1.
Fig. 2. Hooked steel fibers (a) and dog-bone specimens manufacturing (b)
Table 1. UHPFRC mixtures (kg/m3)

Specimen CEM I 52.5R Water Sand 0/0.6 Sand 0.6/1 Silica fume WRA Fibers
UHPC 1000 200 400 600 125 11 –
UHPFRC-0.6 50
UHPFRC-1.25 100
UHPFRC-1.9 150
UHPFRC-2.6 200
3.1 Compression and Flexural Tests (4 4 16 cm)
Three prismatic specimens were prepared for each concrete mixture, 4 4 16 cm in
size, by casting them in steel forms. After 2 days they were demoulded and kept up to 28
days in climatic chamber at 20 °C and 50% R.H. to test their behaviour in the worst
curing conditions. They were tested in bending and then in compression, according to the
procedure described in EN 1015-11 [17]. Even if this type of test method is specific for
mortars and it is not suitable for concrete samples, it was adopted in this study to obtain
preliminary information on the mechanical properties of the specimens by varying the
fibers dosage. Results obtained for the different UHPFRC mixtures are listed in Table 2,
while load-midspan displacement curves of flexural tests are reported in Fig. 3.
Table 2. Experimental results of compression and flexural tests on specimens with dimensions
of 40 40 160 mm3
Specimen Compression Increase of Flexural Increase of Flexural
strength compression strength flexural toughness Uf
(MPa) strength (MPa) strength (kNmm)
UHPC Average 98.5 – 13.8 – 1.14
CoV(%) 5 5 11
UHPFRC- Average 121.4 +23.3% 17.2 +24.8% 9.34
0.6 CoV(%) 5 7 3
UHPFRC- Average 124.0 +25.8% 27.2 +97.9% 16.23
1.25 CoV(%) 5 9 8
UHPFRC- Average 138.8 +40.9% 31.7 +130.4% 20.15
1.9 CoV(%) 4 5 5
UHPFRC- Average 143.4 +45.6% 43.6 +217.1% 28.63
2.6 CoV(%) 4 4 3
It can be observed that the reference UHPC cured in dry environment showed an
average compression strength of 98.5 MPa (a quite low value for a mixture with a w/c
of 0.20, due to the dry curing conditions).
The addition of hooked steel fibers was able to gradually increase both the com-
pression and the flexural strength of the concrete mixtures. Concerning the compressive
strength, the reason for this result can be attributed to the confinement effect of the fiber
reinforcement. The maximum increase was found for the mixture with 200 kg/m3 of
steel fibers (equal to about 2.6% by volume): compressive strength raised of about 45%
while flexural strength showed an increase higher than 200%.
Flexural toughness has been calculated by integrating the area under the load-
displacement curves of flexural tests, up to a midspan displacement equal to 2 mm. The
addition of steel fibers allows to considerably increase the flexural toughness of
UHPFRC specimens. This value raises by increasing the amount of fibers, up to a value
of about 28 kNmm for specimens reinforced with 2.6% of fibers. With respect to the
reference UHPC mixture without fibers, the UHPFRCs were able to absorb up to 25-
time higher energy amount with the highest dosage of fibers.
UHPC - Flexural test UHPFRC-0.6 - Flexural test

20 20
18 18
A A
16 16
14 B 14 B
Load (kN)
Load (kN)
12 12
C
10 10 C
8 8
6 6
4 4
2 2
0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
Midspan displacement (mm) Midspan displacement (mm)
UHPFRC-1.25 - Flexural test UHPFRC-1.9 - Flexural test

20 20
18 A 18 A
16 16
B
14 B 14
Load (kN)
Load (kN)
12 12 C
C
10 10
8 8
6 6
4 4
2 2
0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
Midspan displacement (mm) Midspan displacement (mm)
UHPFRC-2.6 - Flexural test

20
18
16
14
Load (kN)
12
10
8 A
6 B
4
2 C
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Midspan displacement (mm)
Fig. 3. Load-displacement curves of flexural tests on UHPFRC specimens

3.2 Direct Tensile Tests (Dog Bone Specimens)

Dog-bone specimens with a representative cross section of about 30 mm 45 mm
and with a total length of 330 mm were used to experimentally determine the tensile
stress-strain responses of UHPFRC (Fig. 4). Specimens were cast in wood frameworks
(Fig. 2), after 2 days they were demoulded and kept up to 28 days in climatic chamber
at 20 °C and 50% R.H. to test their behaviour in the worst curing conditions. A steel
frame realized with welded steel plates and four steel cylinders (Fig. 4) was used to
grab the specimen and to transfer the tensile force without applying compression at the
specimen ends. In this configuration rotations at the ends of the specimen are allowed.
Tensile tests have been performed on a tensile testing machine with a load capacity of
50 kN in displacement control, with a loading rate of 0.5 mm/min. Digital Image
Correlation (DIC) was used to measure displacements and deformations in the speci-
mens. During the tests, pictures of the frontal surface of the specimens have been
acquired by two digital cameras (model Pixelink® B371F) at 2 frame per second,
collecting about 600 images. The cameras were equipped with a lens having a focal
length of 16 mm and placed about 400 mm away from the specimen, in order to reduce
the perspective errors due to eventual out-plane motions. The specimen was illuminated
using an LED spotlight. A second camera, placed on the right side, was used to monitor
possible motions out of the plane. Tensile strain, e, was measured on a free length,
LDIC, equal to 100 mm as marked in red in Fig. 4b. Results of the tensile tests have
been reported in Table 3 (the average value of 3 specimens for each mixture and
relative Coefficient of Variation (%)), specimens after testing have been shown in
Fig. 6 and the stress-strain curves have been shown in Fig. 5. Tensile strength was
calculated by dividing the tensile load recorded by the testing machine by the speci-
men’s cross section, which has been calculated as the mean value of three sections at
both ends and the centre of the specimen (kept in the central part of the specimen with
total length of 80 mm). The first cracking stress rt was defined as the tensile stress in
the specimen corresponding to the first crack formation. The formation of the first crack
has been determined by visual inspection (with the help of DIC frames) or looking at
the stress-strain curves, in correspondence with a stress decrease in the elastic phase.
The average maximum tensile strength rmax and the corresponding deformation emax,
for each group of specimens were also reported in Table 3.
(a) (b)
Fig. 4. Specimens dimension in mm (a), and direct tensile test setup (b)
Table 3. Experimental results of direct tensile tests on dog-bone specimens

Specimen rt (MPa) et (%) rmax (MPa) emax (%)
UHPC Average 5.65 0.037 5.65 0.037
CoV(%) 17 28 17 28
UHPFRC-0.6 Average 3.74 0.029 4.49 0.07
CoV(%) 17 31 14 19
UHPFRC-1.25 Average 4.90 0.035 5.40 0.49
CoV(%) 5 17 6 45
UHPFRC-1.9 Average 3.93 0.036 5.88 0.56
CoV(%) 14 25 17 5
UHPFRC-2.6 Average 6.14 0.062 8.11 0.87
CoV(%) 14 34 10 40
In uniaxial tensile tests, specimens without fibers (UHPC) showed a linear elastic
behaviour up to failure, with the formation of a single pass-through crack (Fig. 5 and
6). Failure was always brittle after the formation of one unique crack. The addition of
hooked steel fibers, even at the lowest dosage (0.6% by volume), is able to avoid brittle
failure and it allows the specimen to undergo plastic deformation after the formation of
the first crack. However, with fibers dosages of 50 and 100 kg/m3, the formation of the
first crack is reached at lower tensile stresses with respect to the plain UHPC specimen.
This is probably due to a bad dispersion and misalignment of the fibers within the
cross-section, thus reducing the resistant section: it might have constituted a discon-
tinuity, thus anticipating the formation of the first crack. Only by exceeding a critical
threshold on the fibers dosage (about 200 kg/m3) the addition of steel fibers is able to
increase the tensile strength at first cracking.
UHPC UHPFRC-0.6
10 10
A A
8 8
B B
6 6
σ (MPa)
σ (MPa)
C C
4 4
2 2
0 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
ε (%) ε (%)
UHPFRC-1.25 UHPFRC-1.9
10 10 A
A
8 8 B (failed)
B
C
6 C 6
σ (MPa)
σ (MPa)
4 4
2 2
0 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
ε (%) ε (%)
UHPFRC-2.6
10
6
σ (MPa)
4 A
B
2
C
0
0 0.5 1 1.5 2 2.5 3
ε (%)
Fig. 5. Stress-strain curves of direct tensile tests on dog-bone UHPFRC specimens

It can be noticed that the post-cracking behaviour changes by varying the amount of
steel fibers. For lower dosages (50 and 100 kg/m3) a softening behaviour is observed
while with higher dosages (150 and 200 kg/m3) the post-cracking behaviour changes
and significant stress-hardening branches can be noticed in the stress-strain curves
(Fig. 5). This change from softening to hardening behaviour in direct tensile tests,
when a critical threshold in fibers dosage is exceeded, is confirmed by some studies
present in literature [11, 18].
Failure modes observed in UHPFRC specimens are different depending on the
amount of steel fibers, as shown in Fig. 6. UHPFRC specimens remain intact due to the
presence of the steel fibers while those without fibers failed immediately after the
formation of a unique pass-through crack. Specimens with lower fibers dosages
(UHPFRC-0.6 and UHPFRC-1.25) showed the formation of a big crack which grows
at increasing load. Only specimens with higher fibers dosages (UHPFRC-1.9 and
UHPFRC-2.6) showed the formation of multiple cracks. However, once the maximum
load is reached, only one of them continues to grow in the softening branch.
Fig. 6. Failure modes of dog-bone UHPFRC specimens under tensile loads
4 Conclusions
The objective of this study was to evaluate the tensile properties of UHPFRC mixtures,
reinforced with different amount of 30-mm long hooked steel fibers, varying from 0.6%
up to 2.6% by volume. Direct tensile tests have been carried out on dog-bone shaped
specimens, with total length of 330 mm and cross section of 45 30 mm2. Com-
pression and flexural tests have been performed on 40 40 160 mm3 specimens.
Experimental results allowed to compare bending and tensile properties for UHPFRC
specimens with different fiber dosages.
The flowing concluding remarks can be drawn:

• The addition of hooked steel fibers allowed to significantly increase the compres-
sion and flexural strength of UHPFRC mixtures (up to 45% and more than 200%,
respectively, at the highest dosage of 2.6%).
• Flexural toughness has been greatly raised by increasing the steel fiber volume: with
respect to the reference UHPC mixture without fibers, UHPFRCs were able to
absorb up to 25-time higher energy amount with the highest dosage of fibers.
• The addition of steel fibers significantly modified the behaviour of UHPFRC
specimens subject to uniaxial tension, by avoiding their brittle failure even at low
dosages (0.6% by volume).
• The post-cracking behaviour in direct tensile tests was strongly influenced by the
amount of steel fibers, passing from softening to hardening if high amount of fibers
was used (higher than 1.9% by volume).
• This study confirms that the post-cracking behaviour in uniaxial tensile tests is
different from the post-cracking behaviour in bending tests. Once cracked, the
UHPFRC is still able to increase its flexural strength even at low fibers dosage while
its tensile strength can be increased only when high fibers volume is used.
References
1. Yu, R., Spiesz, P., Brouwers, H.J.H.: Mix design and properties assessment of Ultra-High
Performance Fibre Reinforced Concrete (UHPFRC). Cem. Concr. Res. 56, 29–39 (2014).
https://doi.org/10.1016/j.cemconres.2013.11.002
2. Yoo, D.-Y., Yoon, Y.-S.: A review on structural behavior, design, and application of ultra-
high-performance fiber-reinforced concrete. Int. J. Concr. Struct. Mater. 10(2), 125–142
(2016)
3. Song, Q., Yu, R., Shui, Z., Wang, X., Rao, S., Lin, Z.: Optimization of fibre orientation and
distribution for a sustainable Ultra-High Performance Fibre Reinforced Concrete
(UHPFRC): Experiments and mechanism analysis. Constr. Build. Mater. 169, 8–19
(2018). https://doi.org/10.1016/j.conbuildmat.2018.02.130
4. Corinaldesi, V., Donnini, J., Nardinocchi, A.: The influence of calcium oxide addition on
properties of fiber reinforced cement-based composites. J. Build. Eng. 4, 14–20 (2015)
5. Banthia, N., Nandakumar, N.: Crack growth resistance of hybrid fiber reinforced cement
composites. Cem. Concr. Compos. 25, 3–9 (2003). https://doi.org/10.1016/S0958-9465(01)
00043-9
6. Fantilli, A.P., Mihashi, H., Vallini, P.: Multiple cracking and strain hardening in fiber-
reinforced concrete under uniaxial tension. Cem. Concr. Res. 39, 1217–1229 (2009)
7. Bffup, A.: Ultra High Performance Fiber-Reinforced Concretes: Interim Recommendations:
Scientific and Technical Committee. Assoc Française Genie Civ (2002)
8. Japan Society of Civil Engineers. Recommendations for Design and Construction of High
Performance Fiber Reinforced Cement Composites with Multiple Fine Cracks (HPFRCC).
Concr. Eng. Ser. 82: Testing Method 6–10 (2008)
9. fib Model Code for Concrete Structures 2010 (2013). https://doi.org/10.1002/
9783433604090
10. Delle, C., In, S., Armato, C.: Consiglio Superiore dei Lavori Pubblici Servizio Tecnico
Centrale. Cemento 93, 62–00146 (2003)
11. Wille, K., El-Tawil, S., Naaman, A.E.: Properties of strain hardening ultra high performance
fiber reinforced concrete (UHP-FRC) under direct tensile loading. Cem. Concr. Compos. 48,
53–66 (2014). https://doi.org/10.1016/j.cemconcomp.2013.12.015
12. Yoo, D.Y., Kim, S., Kim, J.J., Chun, B.: An experimental study on pullout and tensile
behavior of ultra-high-performance concrete reinforced with various steel fibers. Constr.
Build. Mater. 206, 46–61 (2019). https://doi.org/10.1016/j.conbuildmat.2019.02.058
13. Wille, K., Kim, D.J., Naaman, A.E.: Strain-hardening UHP-FRC with low fiber contents.
Mater. Struct. Constr. 44, 583–598 (2011). https://doi.org/10.1617/s11527-010-9650-4
of a repair material with strain hardening behavior. Cem. Concr. Compos. 30, 863–871
(2008). https://doi.org/10.1016/j.cemconcomp.2008.08.003
15. Jun, P., Mechtcherine, V.: Behaviour of strain-hardening cement-based composites (SHCC)
under monotonic and cyclic tensile loading: Part 1 - Experimental investigations. Cem.
Concr. Compos. 32, 801–809 (2010). https://doi.org/10.1016/j.cemconcomp.2010.07.019
16. EN 197-1. Cement - Part 1: Composition, specifications and conformity criteria for common
cements (2011)
17. EN 1015-11. Metodi di prova per malte per opere murarie Parte 11 : Determinazione della
resistenza a flessione e a compressione della malta indurita (2007)
18. Naaman, A.E., Reinhardt, H.W.: Proposed classification of HPFRC composites based on
their tensile response. Mater. Struct. Constr. 39, 547–555 (2006). https://doi.org/10.1617/
s11527-006-9103-2
Slip-Hardening Bond: A Key to the Success
of Ultra High Performance FRC Composites
Antoine E. Naaman(&)
Department of Civil and Environmental Engineering, University of Michgan,

Ann Arbor, USA
antoinenaaman@yahoo.com
Abstract. Bond is recognized as a fundamental causal parameter in the

mechanics of fiber reinforced composites. Its paramount function is to transmit
forces between fibers and matrix and vice-versa. This paper focuses on “slip-
hardening bond” which can be achieved in cement composites with some fibers,
as compared to commonly observed slip-softening bond or constant bond. In
particular: 1) it describes how bond is characterized from a pull-out test or a
pull-through test on a single fiber; 2) it gives examples of fibers with bond stress
versus slip exhibiting slip-hardening behavior; and 3) it clarifies how slip-
hardening response can be likely achieved. Slip-hardening bond is an extremely
important characteristic which should be evaluated at the onset of design; it
implies that, at the composite level, once a crack is formed in the matrix, the
fibers bridging the crack provide increasing resistance to crack opening. This is
likely to encourage multiple cracking in the composite and helps lead to com-
posites with strain-hardening behavior in tension, as well as large composite
strains at failure. Slip-hardening bond is considered critical for the further
development and success of high performance and ultra-high performance fiber
reinforced cement composites [1, 6].
Keywords: Concrete Fiber Fiber concrete Bond Pull-out Pull-through

Slip-softening Slip-hardening UHPC composites
1 Introduction
Bond is recognized as a fundamental causal parameter in the mechanics of fiber

reinforced composites. Its paramount function is to transmit forces between fibers and
matrix and vice-versa. The success of composite action, as described by the compos-
ite’s strength, modulus, fracture energy, and ductility, depends on bond. Bond is not a
physical variable; unlike the fiber or the matrix it does not have a volume, but it is as
important. Of the three key components of a composite (fiber, matrix, and bond), bond
is the most difficult to define, understand, characterize, control, optimize, … and, in
short, the most elusive [Chapter 13 in 1]. Bond is also the most critical differential
property among fibers of same material (such as in the case of steel fibers).
In fiber reinforced cementitious composites, bond is mostly evaluated through
single fiber pull-out tests during which the applied load and the fiber slip are contin-
uously measured leading to a pull-out load versus slip relationship, from which a bond
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_95
1080 A. E. Naaman
stress versus slip curve is derived [2–5]. Such a curve can be considered a fundamental
constitutive relation for bond and can vary in nature from brittle to ductile and similarly
from slip-softening to slip-hardening behavior. Moreover, precise pull-out tests allow
us to uncover various bond mechanisms, and then characterize, model separately, and
integrate them in analytical models to optimize composite properties.
This paper focuses on “slip-hardening bond” which can be achieved in fiber
reinforced cement composites, as compared to commonly observed slip-softening bond
or constant bond. In particular: 1) it describes how bond is identified from a pull-out
test or a pull-through test on a single fiber; 2) it gives examples of fibers with bond
stress versus slip exhibiting slip-hardening behavior; and 3) it clarifies how slip-
hardening can be possibly achieved. So far, the most effective way to achieve slip-
hardening wiht steel fibers is through mechanical deformations along the fibers. Slip-
hardening bond is an extremely important characteristics of some fibers; it implies that,
at the composite level, once a crack is formed in the matrix, the fibers bridging the
crack could be designed to provide increasing resistance to crack opening. This is likely
to encourage multiple cracking in the composite and helps lead to composites with
strain-hardening behavior in tension, as well as large composite strains and high energy
absorption at failure. This author considers that slip-hardening bond is essential for
furthering the development and success of strain-hardening high performance and ultra-
high performance (UHPC) fiber reinforced cement composites.
2 Basic Pull-Out and Pull-Through Tests

for Characterization of Bond
In order to better understand bond, it is helpful to understand typical experimental bond

tests. Two are described in Fig. 1, namely the pull-out test and the pull-through test.
Both lead to a pull-out load versus slip relationship from which various bond param-
eters can be extracted. In the pull-out test, the fiber embedded length (that is the length
of contact between fiber and matrix) decreases with an increase in slip; while with the
pull-through test, the embeded length remains constant. The pull-out load versus slip
response recorded from a test is used to determine the maximum and average bond
stress (over a given slip) to use in various mechanical models and in design. Generally,
the maximum load obtained in a pull-out or pull-through test is used to determine the
maximum bond strength. In the mechanics of materials, bond stress is essentially
defined as a shear stress and thus the terms are interchangeably used in this paper, often
referring to bond-shear stress.
Figure 1 (right end) also shows a “push-through” test generally used with micro-
fibers and fibers that are very delicate to be gripped effectively. The specimen for such
test is obtained by slicing a prism containing one longitudinal fiber; the slice is gen-
erally very thin and the fiber is pushed through by an indenter leading to a load versus
slip response from which a bond measure is derived. This test is not addressed here but
mentioned for completeness. Figure 2 shows photos of pull-out and pull-through test
set-ups.
Slip-Hardening Bond 1081
Pull-Out Pull-Through Push-Through

P P
Penetration Fiber S End Fiber Indenter

Slip
surface
Matrix Matrix P
Variable Le
Le contact Constant
length contact Fiber
during length
test during
Matrix
test
Fig. 1. Typical illustrations of tests to determine the bond or shear stress versus slip response of
a fiber embedded in a cement matrix [Ref. 1].
(a) (b ) (c)
Fig. 2. (a) Pull-out test set-up. (b) Photo of ongoing pull-out test. (c) Photo of ongoing pull-
through test [Refs. 1–4, 8].
3 Basic Calculations of Bond Stress or Shear Stress
Given the results of a pull-out test which measures load, P, and slip, S (Figs. 1 and 2),
the following relation is typically used to estimate a bond shear stress versus end slip
response [1]:
( PðSÞ
sðSÞ ¼ wðL e SÞ
for a fiber of any section
PðSÞ ð1Þ
sðSÞ ¼ pdðL e SÞ
for a circular fiber
1082 A. E. Naaman
where S is the end slip, sðSÞ is the bond or shear stress at slip S, PðSÞ is the pull-out
load at slip S, w is the fiber perimeter (p in other notation), d is the fiber diameter, and
Le is the embedded length of the fiber.
Similarly, for a pull-through test where the area of contact between fiber and matrix
remains constant, the bond shear stress at any slip can be calculated from:
( PðSÞ
sðSÞ ¼ wL for a fiber of any section
e
PðSÞ ð2Þ
sðSÞ ¼ pdL e
for a circular fiber
In the pull-through test, the slip can be defined at the midsection or an average
between the two ends; the embedded length can be small but is generally taken to
represent the average surface condition of the fiber; for example, for an indented fiber, a
slice with at least one or two indentations is needed.
The bond-shear stress versus slip relationship provides an excellent means to
compare different types of fibers. Another measure that also helps evaluate the effi-
ciency of a typical fiber under pull-out in comparison to other fibers, is the tensile stress
generated in the fiber by the pull-out load recorded; it is simply given as the pull-out
load at a given slip divided by the cross-sectional area of a fiber. Thus:
PðSÞ
rðSÞ ¼ ð3Þ
Af
The higher the tensile stress generated the more efficient is the fiber; an optimum
design would seek to maximize the stress without leading to failure of the fiber.
Often in practice, the bond-shear strength (or maximum stress), smax , is calculated
for the maximum pull-out load Pmax but assuming S ¼ 0 in Eq. (1), that is, assuming
the initial embedment length, Le , even though the test is a pull-out test. This is because
generally the maximum load occurs at a relatively small slip. Thus for all practical
purposes one can write:
Pmax
smax ð4Þ
wLe
Many parameters affect the shape of the bond stress versus slip response of a fiber
and include chemical surface treatment as well as mechanical surface indentation or
deformation. Numerous examples are given in Chapter 13 of Ref. [1].
4 Typical Bond-Shear Stress Versus Slip Curves

4.1 Typical Idealized Curves
For convenience, the pull-through test is easier to understand at first (Figs. 1 and 2).
Indeed, in a pull-through test, the bond contact surface between fiber and matrix
remains constant during the test, making the pull-out load directly proportional to the
bond-shear stress (Eq. 2). Thus the load axis, can be visualized at the bond-shear stress
axis. In Fig. 3, the contact surface has a width Le, but could generally be considered an
increment dx in an integration for a pull-out test. Theoretically, if the bond shear stress
versus slip is to be considered a true constitutive property, then the slip should be the
observed displacement of the fiber with respect to the matrix over a distance dx con-
sidered representative of the fiber surface.
Fig. 3. Typical bond shear stress versus slip curves illustrating the definition of some variables
and terminology [Ref. 1].
Figure 3 illustrates the types of bond-shear stress versus slip curves observed with
steel fibers. Generally the curves start with an ascending portion (almost vetical linear
at the scale plotted) up to a point beyond wich a significant decrease in stiffness (slope
of the curve) occurs followed by one of four cases: 1) brittle failure or sudden drop of
bond stress to zero); 2) decaying bond (also called slip-softning) or slow gradual
decrease of bond stress with increasing slip; 3) constant bond, a mostly hypothetical
case where bond remains constant with slip (this is also a theoretical boundary con-
dition between slip-softening and slip-hardening); and 4) slip-hardening bond, where
the bond stress increases with slip up to a reasonable and useful level of slip. Most
smooth steel fibers as well as carbon, glass, and kevlar fibers and many polymeric (PP,
PE, Nylon, polyesther) generally show a decaying bond-shear stress versus slip
response. Some steel fibers with mechanical deformation may show slip-hardening
bond up to a different extensts of slip; for instance hooked-end fibers embedded in
1084 A. E. Naaman
normal strength concretes, show a slip hardening bond behavior up to a slip of about
1.5 mm (about equal to the hook length) while, everything esle being equal, an
equivalent diameter twisted steel fiber can show a slip-hardening bond up to 10 mm.
The difference is critical in the development of strain-hardening FRC composites
(defined as having a post-cracking strength higher than their cracking strength) and in
drastically improving their energy absorbtion capacity.
Although the curves in Fig. 3 assume a pull-through test to illustrate the direct
proportionality between the pull-through load and the bond-shear stress, similar curves
can be observed when derived from pull-out tests results using Eq. (1) instead of
Eq. (2).
4.2 Example of Pull-Through and Pull-Out Test Results for Bond

with Steel Fibers
Two examples of pull-through load versus slip response curves for a smooth steel fiber
are shown in Fig. 4 [2]. The test was carried out as per Fig. 2c. Since according to
Eq. (2), the bond-shear stress is directily proportional to the load, the y axis could
represent the bond-shear stress. It can be observed that in this example, after the initial
peak point, the bond slightly deteriorates with increasing slip, thus leading to a slip-
softening or decaying bond up to a slip representing about 20% of the embedded
length. Note that for a pull-out load of 10 lb, the corresponding bond-shear stress is
about 318 psi (2.2 MPa); a much smaller value of bond-shear stress would be used in
design to reflect an average response related to the expected slip.
Fig. 4. Example of pull-through test illustrating decaying bond with a smooth steel fiber and
influence of having the fiber embedded in a plain matrix versus a SIFCON matrix [Ref. 2].
Given any experimental pull-out load versus end slip curve for a fiber (as per the
test of Figs. 2a and 2b), a bond shear stress versus end slip curve can be obtained from
Eq. (1), point by point, and may take on various forms similar to those described in
Fig. 3.
Figure 5a shows the actual pull-out load versus slip response obtained from a pull-
out test such as shown in Fig. 2a. Here, after an initial steep ascending response, the
load drops suddenly and then decreases at a slow rate with increasing slip. Using
Eq. (1), the bond-shear stress is first computed from the pull-out load, and the corre-
sponding bond-shear stress versus slip is plotted in Fig. 5b [5]. It can be observed that,
after the maximum point, the stress decreases suddenly then stabilizes up to relatively
large slips about 50% of embedded length. The value of bond at the plateau level
(1.2 MPa) is only 38% of the maximum bond stress (3.2 MPa). Such behavior justifies
the often used assumption in design that bond is about constant (for example taken as
1.2 MPa), in which case the initial portion of the curve is ignored.
Fig. 5. (a). Typical pull-out load verus slip response from a test such as in Fig. 2a. (b). Derived
bond-shear stress versus slip response for the curve of Fig. 5a [Refs. 2, 5].
Figures 4 describes two examples of typical decaying or slip-softening bond-shear

stress versus slip curves (the bond-shear stress is proportional to the load). If we ignore
the initial portion which occurs over a very small slip, Fig. 5b can be used as a typical
example of a constant (1.2 MPa) bond-shear stress versus slip response. Examples of
slip-hardening response are described next.
5 Examples of Experimental Slip-Hardening Bond Shear

Stress Versus Slip Response
Most generally, the bond shear stress versus slip relationship of smooth fibers is
observed to be with decaying frictional response; however, in some cases, slip-
hardening can be observed (Figs. 6 and 7). Slip hardening is attributed to some
mechanical action which is not intuitive with smooth fibers. For instance, in the case of
smooth brass-coated steel fibers embedded in UHPC, the brass-coating, if scraped from
the surface during pull-out, creates a mechanical blocking that may lead to a slip
1086 A. E. Naaman
hardening response; this is illustrated in Fig. 6 taken from [5]. Note the very high
values of bond stress achieved with an UHPC of compressive strength exceeding
200 MPa. In a similar manner monofilament PP fibers can fibrillate along their surface
during pull-out significantly improving their frictional resistance leading in some cases
to a slip-hardening bond behavior, although the numerical values of bond are much
much smaller.
Fig. 6. Slip-hardening bond-shear stress versus slip curves for brass coated steel fibers
embedded in UHPC matrices with various amounts of fine sand; embedded length = 6.5 mm
[Ref. 5].
Figure 7 illustrates several bond stress versus end slip curves obtained from pull-
out load versus slip tests of high strength twisted steel fibers [5–8]. The x axis repre-
sents the normalized value of slip divided by the embedded length. It can be observed
that, following the initial ascending portion, the bond stress for the twisted fibers
increases continuously with increasing slip, leading to very high bond stresses at large
slips (up to 80% of embedded length), prior to complete pull-out. The curves can be
described as “slip-hardening”. Slip-hardening characteristics have been also observed
with some other deformed fibers such as hooked-ends steel fibers, but hardening was
limited along a slip smaller than the length of the hook, prior to decaying, suggesting
the important contribution of the mechanical component of bond.
Note generally that while slip-hardening is not very common with smooth fibers, it
is more likely to develop with mechanically deformed fibers at least over the range of
slip over which the mechanical deformation provides the most anchorage and resis-
tance. There is need to explore fully this type of behavior and carry out in depth
research to uncover if other treatments, such as chemical treatment of the fiber surface
or a change in fiber material, lead to a slip-hardening bond-shear stress versus slip
response.
Fig. 7. Slip-hardening bond shear stress versus end slip curves obtained from experimental pull-
out load versus slip curves with twisted triangular steel fibers Eq. (1) [Refs. 1, 6–9].
6 Concluding Remarks
As evident from the above discussion, bond is a very complex subject. Some com-
mendable successes have been achieved so far in understanding and modeling bond in
cementitious composites, but more can be done. An important simplification used in
design, is to assume a bond shear stress that is constant and independent of slip; this
can be a very optimistic and unsafe guess-estimate. Indeed, it is not realistic to consider
bond stress without associating a maximum slip value to it; both should be quantified
for a particular application. Thus, the author strongly recommends for standard design,
to select an average value of bond-shear stress (such as from Eq. (1 or 2)) over a
predetermined slip based on expected or prescribed crack opening in the composite,
and for a given level of performance (service or ultimate limit state, or in-between).
A very safe lower bound value can be taken as the equivalent bond obtained from the
pull-out work such as estimated from Eq. (13.8) in Ref. [1] and in Ref. [10].
Among the numerous methods to improve bond in steel fibers, the use of suitably
designed twisted triangular or square shaped fibers offer so far the most advantages [6–
9]. Proper design implies that the fibers un-twist during pull-out or crack opening. Note
that round fibers do not offer the benefits of mechanical ribs when twisted; and very flat
rectangular fibers, if twisted, will form tunnel like sections that may not be penetrated
by the cement matrix and are undesirable sites of stress concentration. Crimping, which
induces a sinusoidal wave form, while effective in improving mechanical bond, leads to
a significant reduction in the equivalent elastic modulus of the fiber. Hooked and
paddled ends fibers offer an effective anchorage concentrated at their ends, but not
distributed along the fiber, thus unable to help for bond at large slips; thir corre-
sponding pull-work [11], while better than that of smooth fibers, can be significantly
smaller than that of equivalent twisted steel fibers.
To achieve strain-hardening behavior in tension, the post-cracking tensile strength
of the composite should exceed its strength at first precolation cracking, and one can
1088 A. E. Naaman
set, for a given composite and fiber, a critical value of fiber volume fraction to insure
such behavior [1]. The post-cracking tensile strength is tied to the product s *Vf *L/d
[1]; practical limitations on both Vf and L/d (such as for instance for steel fibers,
L/d < 100 and Vf < 3% in the premix process), leave the bond as the next most critical
causal variable. In UHPC where the practical limits on Vf and L/d are about attained,
slip-hardening bond becomes crucial not only performance-wise (post-cracking tensile
strength) but also ductility-wise (particularly in improving the strain at peak tensile
stress), and therefore cost-wise as for any other FRC composite.
Future developments in new fibers for concrete should focus on improving fiber
efficiency and achieving better bond characteristics, preferably with bond stress-slip
hardening behavior. Fibers with bond properties characterized by a bond shear stress
versus slip response that is either elastic-plastic or better (slip hardening) up to rela-
tively large slips, enhance crack bridging resistance of a composite under larger crack
openings and thus enhance structural performance and damage tolerance in all respects.
Such fibers allow the practical development of high performance and ultra-high per-
formance (UHPC) FRC composites capable of exhibiting strain-hardening behavior in
tension at relatively low volume fractions of fibers. They are suitable in structural
applications in combination with reinforcing bars or prestressing strands such as in
blast and seismic resistant structures, as well as in stand-alone applications (with no
other reinforcement) such as in thin sheet products for housing, claddings for buildings,
precast products, shells, pipes, and the like.
Acknowledgements. The author has carried out research on bond in fibers for concrete since
1970, as part of his Ph.D. thesis. Since then, the following former colleagues and students have
helped in the evolution of ideas briefly described in this paper: S.P. Shah, U. Gokoz, K.
Visalvanich, G. Nammur, K. Kosa, H. Najm, J. Alwan, R.E. Robertson, P. Guerrero, C.
Sujivorakul, A. Waas, D.J. Kim, and K. Wille. Their collaboration is sincerely acknowledged.
The author offers his apologies to all those who have specifically addressed in their research
“slip-hardening” bond but are not cited here due to space limitations.
References
1. Naaman, A.E.: Fiber Reinforced Cement and Concrete Composites. Techno Press 3000, 765
p. (2018). http://www.technopress3000.com/books. ISBN 978-0-9674939-3-0; LCCN
2017916342
2. Naaman, A.E., Najm, H.: Bond-slip mechanisms of steel fibers in concrete. ACI Mater. J. 88
(2), 135–145 (1991)
3. Guerrero, P., Naaman, A.E.: Effect of mortar fineness and adhesive agents on pullout
response of steel fibers. ACI Mater. J. 97(1), 12–20 (2000)
4. Sujivorakul, C., Naaman, A.E.: Evaluation of bond-slip behavior of twisted wire strand steel
fibers embedded in cement matrix. In: Balaguru, P., Naaman, A.E., Weiss, W. (eds.)
Proceedings of ACI Symposium on Concrete: Material Science to Applications, a Tribute to
S.P. Shah, ACI SP-206, pp. 271–292, April 2002
5. Wille, K., Naaman, A.E.: Bond-slip behavior of steel fibers embedded in ultra high
performance concrete. In: Dresden, V.M., Kaliske, M. (eds.) Proceedings of 18 European
Conference on Fracture and Damage of Advanced Fiber-Reinforced Cement-Based
Materials, Contribution to ECF 18, pp. 99–111. Aedificatio Publishers, Freiburg, September
2010
6. Naaman, A.E.: Fibers with slip-hardening bond. In: Reinhardt, H.W., Naaman, A.E. (eds.)
High Performance Fiber Reinforced Cement Composites - HPFRCC 3, RILEM Pro 6,
pp. 371–385. RILEM Publisations S.A.R.L., Cachan, France, May 1999
7. Naaman, A.E.: Engineered steel fibers with optimal properties for reinforcement of cement
composites. J. Adv. Con. Technol. 1(3), 241–252 (2003)
8. Naaman, A.E., Sujivorakul, C.: Pull-out mechanisms of twisted steel fibers embedded in
concrete. In: Proceedings of International Conference on Applications of Shotcrete,
Tasmania, Australia, 7 p., April 2001
9. Sujivorakul, C., Naaman, A.E.: Modeling bond components of deformed steel fibers in FRC
composites. In: Naaman, A.E., Reinhardt, H.W. (eds.) High Performance Fiber Reinforced
Cement Composites (HPFRCC-4), , pp. 35–48. RILEM Pub., Pro. 30, June 2003
10. Kim, D.J., El-Tawil, S., Naaman, A.E.: Loading rate effect on pull-out behavior of deformed
steel fiber. ACI Mater. J. 105(6), 576–584 (2008)
11. Alwan, J., Naaman, A.E., Hansen, W.: Pull-out work of steel fibers from cementitious
matrices - analytical investigation. J. Cem. Concr. Compos. 13(4), 247–255 (1991)
Testing of Thin UHPFRC Cantilever Stairs
with Bolted Connections
Ioan Sosa(&), Camelia Negrutiu, Bogdan Heghes, and Adel Todor
Technical University of Cluj-Napoca, Cluj-Napoca, Romania

ioan.sosa@dst.utcluj.ro
Abstract. The purpose of this research is to analyze the suitability of UHPFRC

for applications to precast cantilever stairs that can be easily connected to the
main structure. Several cantilever stair elements were tested under a concen-
trated static load applied at the free end. The anchored end was connected to the
testing frame with four bolts that provided a partial fixed end. Each stair had a
tread, a riser and an end plate casted with four holes to accommodate the bolts.
The riser and the tread had a thickness of 20 mm whereas the end plates were
20 mm and 30 mm thick. The length of the stairs was 600 mm. Only short steel
fibers were provided as reinforcement (2.35 vol.%) resulting a 180 MPa com-
pression strength and 25 MPa flexural strength. The failure of the elements
occurred at concentrated load values of 4 to 7 kN depending on the thickness of
the end plate. The equivalent static force was above the one resulting from
standardized static loads for staircases.
Keywords: Fibers Ultra-high performance concrete Cantilever Stairs

Precast Bolts
1 Introduction
Ultra High Performance Fiber Reinforced Concrete (UHPFRC) is one of most promising
concretes for the construction industry. Its strong points such as ultra-high compressive
strength (over 150 MPa), very good durability, ductility provided by fiber reinforcement
lead to slender cross-sections and a significantly decrease of the concrete elements’
weight [1, 2]. The reduced maintenance costs and improved performances recommend it
as a sustainable construction material [3]. However, uncertainties about UHPFCR
behavior and the lack of standardized design rules limited the expected boost of
UHPFRC use. The most significant and complete publications related to UHPFRC are
referring to the few commercially available compositions (Ductal, CERACEM,
CEMTECmultiscale): Setra&AFGC French recommendations (2002 and updated 2012)
[2], JSCE recommendation (2006) [3], U.S. Federal Highway report in 2006- updated
state-of-the art report (2013) [4]. Some large-scale structures are available worldwide but
are limited mainly to footbridge applications. Some of the most impressive are in: France
(Pont du Diable Footbridge, 2008), Germany (Gärtnerplatzbridge Kassel, 2007), South
Korea (Seonyu Footbridge, 2002), Japan (Sakata Mirai Footbridge, 2002), Spain (Ovejas
ravine in Alicante, 2013). One of the uncertainties regarding UHPFRC is the fiber
orientation and its dependability of the casting procedure. As most of the researches use
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_96
Testing of Thin UHPFRC Cantilever Stairs with Bolted Connections 1091
self-compacting fiber reinforced concrete, the fibers movement and rotation possibilities
are far more pronounced that those in a less workable concretes. Two trends are observed
for fibers orientation depending elements shape: beam like and slab like elements. In the
beam-like elements the concrete can unrestrictedly flow in the longitudinal direction
conducting to a funnel type flow dominated by shear and wall effects. In the slab-like
elements the concrete can flow unrestricted in any horizontal direction namely a radial
flow [5, 6]. However, efforts are made to use the preferential orientation of the fibers as a
way to improve the flexural behavior of UHPFRC elements [7, 8].
2 Research Scope
The scope of the research is to develop a series of easy to assemble precast UHPFRC
cantilever elements, with different connections types, that would appeal to the con-
struction industry. The first stage was the experimental testing and analysis of the
UHPFRC stairs with no other reinforcement but fibers. The possible advantages of
UPFRC stairs are the thin walls and the cantilever system with a great impact on the
esthetics of a building. Moreover, using bolts to connect the stairs to the main structure
greatly reduces the time involved in the construction. The advantage of the bolts is that
they can be either fixed in concrete during casting or more conveniently post-installed
using chemical anchors. This way, the support elements for the stairs in the form of
concrete walls or beams can be cast without interruption. The research findings will be
used to extend the research to other types of cantilever elements that require a small
self-weight and fast assembling such as canopies and balconies. Moreover, other types
of connections are planned to be tested including by post tensioning which will also
increase the capacity.
3 Experimental Research
3.1 Cantilever Stairs Design

The cantilever stairs were cast without reinforcement except the short steel fibers that
were included in the concrete composition. Each stair consisted of a tread, riser and an
end plate with four holes (25 mm diameter) to accommodate the bolts for the con-
nection. Two sets of stairs were researched: Set A and Set B (Fig. 1). Set A had an end
plate with a thickness of 20 mm and Set B had an end plate of 30 mm. Both Sets, A
and B, had 20 mm thick treads and the rise. Supplementary, Set B was casted with a
30 mm strengthening shoulder at the intersection of the tread with the end plate. The
variation was designed in order to increase the bending capacity of the end plate. For
each set, two stairs elements were casted.
1092 I. Sosa et al.
Fig. 1. Design of the stairs: Set A (left) and Set B (right)
3.2 Concrete Composition, Specimens, Casting and Curing

The UHPFRC composition was developed within the Technical University of Cluj
research group on concretes and the concrete was extensively tested for mechanical
behavior before [9]. The cement and the sands were locally produced, whereas the
silica fume and steel fibers were imported from the EU. The components (Table 1)
were: early strength CEM I 52,5 R produced by Holcim, dry silica fume powder with
min 90% SiO2 (Microsilica Grade 940) produced by Elkem, 0.175 mm diameter/6 mm
long straight fibers produced by Baumbach Metall, modified acrylic polymer super-
plasticizer Dynamon SR3 by Mapei. The locally used aggregates consisted of quartz
sands divided into fine sand (0 to 0.3 mm) medium fine sand (0.3 to 0.63 mm), and
coarse sand (0.63 to 1.2 mm). No quartz powder was added. The fiber volume was 2.35
vol.%. The concrete resulting workability was high as the concrete freely flowed within
the formworks without the need for vibrations.
Table 1. UHPFRC composition [kg/m3].

Binder
Cement Silica fume Water Super. W + S/B Fibers Sands
740 156 150 60 0.23 185 1242
The formwork of the stairs was crafted out of wood which was brushed with release
agent to minimize the water absorption by the wood. The wood formwork is not that
best option for UHPFRC as any water loss can affect the properties but was chosen as it
is easy to work with and has low costs. Two identical formworks were produced in
order to cast the stairs from the same concrete batch.
As the direction of casting can induce preferential orientation of the fibers, the
decision was to pour from the riser part of the stairs (Fig. 2). Due to the concrete flow
and formwork walls, a favorable, longitudinal direction orientation of the fibers was
presumed for the riser and the tread, meaning a potential better flexural capacity.
However, the fiber orientation in the end plate would be unfavorable in terms of flexure
as the fibers would align parallel with the presumed flexural cracks.
Fig. 2. Stair elements positioned for casting (left), flexural test specimens (right)
In order evaluate the fiber orientation and flexural behavior, two series of thin plate
elements were casted, both with a thickness of 20 mm, identical with the stairs tread
and riser. One of the plate series was poured in a similar way with the stair tread
(horizontal) and one in horizontal position similar with the riser (vertical). Furthermore,
two series of 40 mm x 40 mm x 160 mm prisms were cast horizontally and vertically
to further assess the influence of fibers orientation on flexural behavior (Fig. 2). The
vertical cast prims would emulate the end plate in terms of fiber orientation.
To evaluate the concrete compression strength, 50 mm x 50 mm x 50 mm cubes
were casted. The relatively small size of the specimens was chosen because of the small
particle size of the aggregates (sands) and the thin cross section of the stair elements.
After pouring the concrete settled in the formworks for 24 h, followed immediately
after demolding by thermal treatment for 24 h. The parameters of the thermal treatment
were: temperature of 90°C and relative humidity of 80%.
3.3 Testing Methodology

All the specimens and the stair element were tested approximately one week after the
finish of the thermal treatment. The compressive strength was determined on
50x50x50mm cubes using an automated testing machine.
The flexural behavior was tested using the three-point flexural test setup for the
prisms and the thin plates (Fig. 3). The span was 140 mm for all flexural test speci-
mens. The thin plates were cut out of the casted 800 mm long plates. Displacement and
force were recorded by a computer via a data acquisition system. Due to the test setup,
the displacement transducer was placed on the machine frame, thus the initial readings
for mid-span displacement included the machine movement. Subsequently, those initial
movement due to preload were removed from the diagrams.
The stair element was connected to a rigid steel frame using four M20 bolts and
50 mm washer. The stair element was positioned upside down to facilitate the appli-
cation of the load (Fig. 4). One concentrated load was applied at the free end of the stair
1094 I. Sosa et al.
Fig. 3. Flexural test setup of prims (left) and thin plates (right)
element using 100 mm x 100 mm steel plate, thus resulting a span of 550 mm for the
cantilever. The displacement transducer was connected at the very end of the cantilever.
Both the force sensor and displacement transducer were connected to the computer.
Fig. 4. Test setup of the cantilever stairs.
The obtained compression and flexural strengths results are listed in Table 2. The
average compressive strength was 186 N/mm2 and is more than the minimum required
for the ultra-high-performance concretes (150 N/mm2).
The flexural strength and behavior were tested on two different kinds of specimens,
each casted in two positions in order to evaluate the preferential fiber orientation effect
due to casting. Regarding the thin plates the effect of casting direction was minimal,
Table 2. Compression and flexural strength of specimens

Strength Specimen Specimen Average value
test type size
[mm] [N/mm2]
Compression Cubes 50 50 50 186
Flexure Plate vertical 20 95 140 23
Flexure Plate horiz. 20 95 140 25
Flexure Prism vertical 40 40 140 19
Flexure Prims horiz. 40 40 140 24
both the flexural strengths and displacements being in the same range (Figs. 5 and 6).
Thus, is presumed that the riser and the tread of the stair elements have similar dis-
tribution of the fibers and consequently will have comparable response in flexure.
A more pronounced fiber orientation effect was observed for the prims. The vertically
cast prims had an approximately 20% lower flexural strength compared with the
horizontally cast prims or either of the thin plates. Therefore, the end plate, casted in a
similar position with the vertically cast prims is presumed to have lower strengths than
the riser or the tread.
35
Vertical cast thin plates
30 V1
25 V3
V4
fct,fl [N/mm2]
20 V5
15
10
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Middle span displacement [μm]
Fig. 5. Flexural stress – displacement of vertical cast thin plates
The cantilever stairs elements test results are listed in Figs. 7, 8, 9 in terms of force
vs. displacement. One of the elements from Set B was damaged during demolding and
was excluded from testing. Set B stairs with an end plate of 30 mm and a strengthening
shoulder have an almost 85% larger maximum loading capacity compared with the set
A with an end plate of 20 mm. Both sets displayed a semi-ductile behavior, with slow
development of the crack openings. The prescriptions of EN 1991-1-1:2004 regarding
the imposed loads for stairs require a minimum concentrated load of 2 kN for stairs
applied in the most unfavorable position. In the case of the tested cantilever stairs the
1096 I. Sosa et al.
35
Horizontal cast thin plates
30
H1
25 H2
20
H3
fct,fl [N/mm2]
H4
15 H5
10
5
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Middle span displacement [μm]
Fig. 6. Flexural stress – displacement of horizontal cast thin plates
most unfavorable position is the free end. Considering a safety factor of 1.5 the
required imposed load is 3 kN. The tested cantilever stairs exceeded the loading
requirements of the EN 1991-1-1:2004: Set A with 30% and Set B with 145%.
5
Set A, Element 1
4
Force [kN]
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Free end displacement [mm]
Fig. 7. Force – Displacement of Set A, Element 1 (20 mm thick end plate)
The thickening of the end place from 20 mm to 30 mm also changed the pattern of
cracks (Fig. 10). Whereas for the Set A all the cracks developed within the end plate,
the Set B element displayed failure cracks on the riser as well. Nevertheless, for both
types of elements the concrete cracks initiated on the end plate, at the first row of bolts,
indicating a semi-rigid connection.
5
Set A, Element 2
4
Force [kN]
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
Fig. 8. Force – Displacement of Set A, Element 2 (20 mm thick end plate)
8
Set B, Element 1
7
4
Force [kN]
0
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
Fig. 9. Force – Displacement of Set B, Element 1 (30 mm thick end plate + shoulder)
1098 I. Sosa et al.
Fig. 10. Crack patterns after failure: Set A with 20 mm end plate (left) and Set B with 30 mm
end plate (right)
5 Conclusions
The present research focused on the potential use of UHPFRC cantilever stair elements,
with fiber reinforcement only and a simple bolted connection. The experimental flex-
ural tests on cantilever elements and concrete specimens led to the following
conclusions:
• The cantilever stairs exceeded with up to 145% the required imposed loads from EN
1991-1-1:2004
• The bolted steel connection offers only a fixed end to the stairs, the crack initiation
and failure was observed in all cases in the end plate
• Increasing the end plate thickness by 10 mm increased the loading capacity by 85%
• Preferential fiber orientation due to casting and formwork walls can decrease the
flexural strength by 20%
• The research was limited to a relatively small span for the stairs and a single type of
connection: larger openings and moment resting connections will be tested and
evaluated together with a finite element analysis.
References
1. Naaman, A., Wille, K.: The path to UHP-FRC: Five Decades of Progress. In: Proceedings of
Hipermat 2012 3rd International Symposium on UHPC and Nanotechnology for High
Performance Construction Materials, Kassel, 7–9 March 2012, pp. 3-15 (2012).
2. Respledino, J., Toutlemonde, F., et al.: Ultra-high performance fiber-reinforced concretes.
Interim Recommend., AFGC -Setra (2002)
3. Japan Society of Civil Engineers, ‘Recommendations for design and constructions of ultra
high strength concrete structures’, Japan (2006)
4. Russell, H., Graybeal, B.: UHPC ‘A State-of-the-Art Report for the Bridge Community’, US
Highway Admin.,Report No. FHWA-HRT-13–060, USA (2013)
5. Lataste, B., et al.: Assessment of fiber orientation in ultra high performance fiber reinforced
concrete and its effect on flexural strength. Mater. Struct. 43(7), 1009–1023 (2010)
6. Bartoli, L., et al.: The production effect on the performance of panels cast with self-
compacting fiber reinforced concrete. In: 3rd International RILEM Conference on Strain
Hardening Cementitious Composites, 03 – 05 November, RILEM, Dordrecht 2014.
7. Yu, R. et al.: ‚Sustainable development of ultra-high performance fiber reinforced concrete
(UHPFRC): towards to an optimized concrete matrix and efficient fiber application. J. Clean.
Prod. 162, 220–233 (2017)
8. Huang, H., et al.: Improvement effect of steel fiber orientation control on mechanical
performance of UHPC. Constr. Build. Mater. 188, 709–72 (2018)
9. Magureanu, C., et al.: – ‘Mechanical properties and durability of ultra high performance
concrete’. ACI Mater. J. 109(2), 177–184 (2012)
Studying of Processing-Structure-Properties
Relation of Strain Hardening Cementitious
Composites (SHCC)
Zhenghao Li1(&), Jiajia Zhou1, Cong Lu2,

and Christopher K. Y. Leung1
1
Clear Water Bay, Hong Kong SAR, People’s Republic of China
zlifb@connect.ust.hk
2
Department of Civil Engineering, Southeast University, Nanjing, China
Abstract. Strain hardening cementitious composites (SHCC) are a class of fiber

reinforced materials exhibiting tensile strain hardening behavior up to strain of
several percent, accompanied by the formation of fine multiple cracks with
openings below 50 lm. To model the full stress-strain relation of SHCC (which
governs ductility and energy absorption) and the crack width versus strain
relation (which governs durability), the sequential formation of cracks needs to
be analysed. The cracking process is related to the internal structure of SHCC,
such as the fiber and flaw size distributions, which varies with material rheology
and mixing sequence. As a first study of the processing-structure-property
relation of SHCC, tensile specimens are prepared with mixes exhibiting different
viscosities. To account for sequential cracking, the variation in SHCC ‘structure’
is represented by the size variation of equivalent spherical flaws inside the
member according to the normal distribution, while fibres are assumed to be
uniformly distributed. By fitting the measured tensile stress-strain curves for
various mixes with a micromechanical model developed at HKUST, the effect of
matrix viscosity on the flaw size distribution is determined. The results will
provide insight on the micromechanics-based design of SHCC for various
requirements on ductility and crack control.
Keywords: Strain hardening composites Tensile performance Crack width

control Durability
1 Introduction
Strain hardening cementitious composites (SHCC), also known as engineered

cementitious composites (ECC), exhibit tensile strain hardening behavior up to several
percent strain, accompanied by the formation of fine multiple cracks with opening
below about 50 lm. In structural applications, these properties will translate into high
deformation capacity and energy absorption as well as high durability, as fine cracks
will not facilitate the penetration of water or other chemicals to induce concrete
deterioration or corrosion of steel rebars. Based on understanding of micromechanics
and fracture mechanics, the criteria for achieving SHCC behavior with short random
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_97
Studying of Processing-Structure-Properties Relation of SHCC 1101
fibers was first proposed in [1] and further refined in [2–4]. Based on these criteria,
SHCCs with different compositions have been successfully made by various
researchers groups.
In engineering applications, besides knowing the failure mode of the material
(hardening or softening), it is very important to have the full stress-strain relation up to
maximum stress, as this will reflect the tensile strain that can be reached and the total
energy that can be absorbed. Moreover, this constitutive relation can be employed in
finite element analysis to predict the behavior of structural components. The crack
width versus strain is another relation with practical significance. Penetration of water
and chemicals can be accelerated by the opening of cracks, but the effect will be
minimal if the crack width is below about 50 lm. At the serviceability state, the
maximum strain in a member can be calculated. If the corresponding crack width is
sufficiently small, long-term durability can be assured.
While the stress versus strain and crack width versus strain relations can be
experimentally measured, it is highly desirable to have micromechanical models that
can predict them based on properties of the fiber, matrix and fiber/matrix interface as
well as fiber dimensions and volume fraction. SHCC can then be designed to achieve
the performance requirements for various applications [5]. Such models have been
developed recently at HKUST [6, 7]. The strain corresponding to a certain applied
stress and the crack width at a particular strain are dependent on the crack spacing,
which is governed by two factors. The first factor is the effectiveness of stress transfer
from the fiber at the crack back to the matrix, which has been analysed in [6]. The
second factor is the variation of matrix cracking strength in the material, which
determines if the transferred stress at a certain section is sufficient for cracking to occur.
In [6] and [7], different assumptions have been made on how the cracking strength
varies, and fitting of test data was conducted to find the parameters governing the
variation.
From a scientific point of view, the strength in various sections should depend on
the variation of fiber and flaw size distributions within the material, which are in turn
governed by the rheological properties and the mixing sequence. The effect of Marsh
cone flow time (which reflects the viscosity [8]), on SHCC behavior was revealed in [9]
while the effect of mixing sequence was shown in [10]. To be able to predict SHCC
behavior (so the proper trial mix can be designed), a link between processing param-
eters, internal structure (i.e., fiber and flaw distributions) and properties of SHCC need
to be established. The goal of our research program is to study the effect of processing
parameters on the fiber and flaw size distributions, and the effect of such distributions
on the tensile stress-strain relation and crack width versus strain relation. As the first
paper reporting our finding, only the effect of viscosity (i.e., flow cone time) is con-
sidered, with the mixing sequence kept unchanged. Also, the variation in SHCC
‘structure’ is represented by the size variation of equivalent spherical flaws within the
member according to the normal distribution, while the fiber content is taken to be
uniform for all sections. Based on the fitting of measured tensile behavior with the
model in [7], the effect of viscosity on the flaw size distribution can be determined. This
information should be useful for the design of SHCC. In the following sections, the
experimental program will be presented with major test results. After a brief description
1102 Z. Li et al.
of the micromechanical modelling, fitting of the tensile curves will be described and the
relevance of the results will be discussed.
2.1 Material
Three mixtures with the same mix proportion (Table 1) but different superplasticizer
content were designed in this experimental investigation. For the SHCC specimens, the
main constituents included Ordinary Portland cement, Class I Type F fly ash, silica sand
with the size distribution of 125–180 lm, polycarboxylate-type superplasticizer and
water. PVA fibers were added into the matrix at a volume fraction of 2%. All mixtures
are listed in Table 1. The PVA fibers are 12 mm in length and 39 lm in diameter, and
the Young’s modulus and tensile strength are 40 GPa and 1600 MPa respectively.
Table 1. Mixture Proportions (by Weight) of the tested SHCCs

Cement Fly ash Silica sand Water Super-plasticizer PVA fiber (Vf)
SHCC_I 0.2 0.8 0.2 0.22 0.50% 2%
SHCC_II 0.2 0.8 0.2 0.22 0.40% 2%
SHCC_III 0.2 0.8 0.2 0.22 0.36% 2%
2.2 Testing of Fresh Properties

In this study, the Marsh cone test was carried out to evaluate the fluidity of the fresh
ECC mortar. A metallic Marsh cone with the dimensions recommened in [9] was
employed (Fig. 1). The cone was fully filled with ECC mortar, and the time taken for
all of the mortar to flow through the cone was measured and denoted as the flow time.
The higher the flow time, the lower is the fluidity of the mortar.
2.3 Mixing, Casting and Curing

All mixtures were prepared using a 12 HL mixer, following the same mixing sequence,
speed, and time. Solid ingredients, including cement, fly ash, and silica sand, were first
mixed at low speed for 2 min. Water and super-plasticizer were then added into the dry
mixture and mixed at high speed for 2 min. PVA fibers were added and mixed at high
speed for 2 min.
The mixtures were then cast into steel molds, and external vibration was applied to
improve the compaction. After casting and curing the SHCC specimens in steel molds
for 24 h, all the specimens were demolded and cured in the standard curing environ-
ment at 95% relative humidity and 23 ± 2 °C. When the curing period reached 28
days, the specimens were removed from the curing room until testing. Five samples
were prepared for each mix.
2.4 Tensile Test

The uniaxial tensile test was carried out with the dumbbell SHCC specimens at the age
of 28 days. Figure 2 illustrates the dimension of the specimen. Before testing, four
aluminum plates were glued on the both ends of each specimen to facilitate gripping.
Tests were carried out using a 250 kN capacity MTS machine under a displacement
control rate of 0.005 mm/s. Two external linear variable displacement transducers
(LVDTs) were attached to the middle part of the specimen with a gage length of
80 mm to measure the deformation of the specimen during the test. The tensile stress–
strain curve of each specimen was recorded.
120 mm
330
13
300 mm
60
30
25 mm
20 mm
85 40 80 40 85
Fig. 1. Marsh cone flow Fig. 2. Dumbbell specimen for tensile testing.
time test [9].
The Marsh cone flow times of the fresh mortar and the tensile properties of the
hardened SHCC samples are summarized in Table 2. The flow time of the fresh mortar
decreased with increasing dosage of the superplasticizer.
Table 2. Flow times of the fresh mortars and tensile properties of the hardened SHCC samples
SHCC_I Flow time 19 s
First cracking strength (MPa) 3.37 3.33 3.36 2.12 4.02
Ultimate strength (MPa) 4.73 5.30 4.68 4.19 4.92
Ultimate strain (%) 3.34 4.79 4.96 2.12 3.76
SHCC_II Flow time 30 s
Ultimate strain (%) 4.64 4.73 5.55 6.22 6.03
SHCC_III Flow time 48 s
Ultimate strain (%) 2.10 4.37 3.87 2.09 3.15
1104 Z. Li et al.
Figure 3 shows the tensile stress–strain curves of the three groups of SHCC
samples. All specimens show tension-hardening and multiple cracking behavior. When
the flow time was 19 s, large variation existed in both the ultimate strain and the
ultimate strength. The tensile strain could vary from 2.12% to 4.96%. Large variation in
tensile strain capacity was also found when the flow time was 48 s. However, when the
flow time was 30 s, the SHCC samples exhibited more consistent tensile behavior as
well as greatly improved tensile strain capacity, with an average value of 5.45%.
Flow time = 19s Flow time = 30s Flow time = 48s
5 4
5
4 4 3
Stress(MPa)
Stress(MPa)
Stress(MPa)
3 3
2
2 2
1
1 1
0 0 0
0 1 2 3 4 5 0 1 2 3 4 5 6 0 1 2 3 4
Strain(%) Strain(%) Strain(%)
Fig. 3. Tensile stress–strain curves of the SHCC samples.
4 Simulation of Tensile Behavior
To interpret the variation of tensile behavior from the perspective of microstructure,

curve fitting method was adopted. In this section, the micromechanical model applied
will be briefly introduced first, then the fitting parameter will be obtained by inversion
method. The correlation between viscosity, microstructure parameter, and tensile
behavior will be discussed.
4.1 Micromechanical Model

The model applied in this article for predicting the tensile behavior of SHCC was
proposed by Lu [7], which considered the effect of fiber distribution, flaw distribution
and stress transfer distance under different load stages. Key features of the model are
described below.
Fibres were assumed to be 3D randomly distributed in this model, which means the
V
number of fibers across each section is equal to the theoretical value N ¼ prf2 , the
f
L
embedment length of short side is randomly distributed from 0 to 2f , and the probability
density of fiber inclination angle h is sinh in the range of 0 to p/2 [11]. Fibers are
divided into 5 different categories according to different statuses or stress levels,
namely Two-way debonding, Pullout-debonding, Two-way Pullout, Ruptured, and
Pullout, which have different contributions to the crack bridging force. By summing up
the contribution of each fiber, the stress-cracking opening relation is calculated.
The stress transfer to matrix is calculated in two parts. The first part is the pulley
force, which is caused by snubbing effect as shown in Fig. 4. The pulley force can be
derived by force equilibrium and occurs right at the crack plane. The second part is
frictional force. Frictional force is caused by chemical bond and frictional bond of
interface, so the contribution of fibers are different at different statuses. Similarly, by
summing up all contributions from fibers with different embedded lengths and incli-
nation angles, the stress field near a single crack can be obtained.
Fig. 4. Schematic diagram for snubbing effect at exit point [7].
To model the distribution of matrix strength, a set of initial flaws are introduced to
the matrix. The flaws are assumed to be spherical, with their size following the normal
distribution. The flaw location is randomly assigned within the member as suggested by
Kabele [12]. Then the equivalent penny-shaped cracks are calculated according to the
largest flaw on each section as shown in Fig. 5 [1], and the normalized cracking
strength can be derived from
pffiffiffi
e
r fc pK 4 pffiffiffi 1
¼ þ c c ð1Þ
g 2 c 3 2
where
2 pf =2

g 1 þ e ð2Þ
4þf2
is governed by snubbing coefficient f. In Eq. (1), r ~fc refers to cracking strength at the

section normalized by r0 Vf s Lf =df =2, in which Vf is the fiber volume fraction, Lf
and df is the length and diameter of fiber. K and c are normalized fracture toughness and
flaw size, defined by
pffiffiffiffiffi
K Ktip =r0 c0 =g~ d ð3Þ
1~1
c ¼ ðc=c0 Þ2d ð4Þ
1106 Z. Li et al.
Where
2
Lf EC p
c0 ð5Þ
16ð1 v2 Þ2
2Ktip

~d 2s=ð1 þ gÞEf Lf =df ð6Þ
With proper modeling of fiber and flaw distributions as well as stress transfer
analysis, simulation of multiple cracking process can be conducted. In the process of
the simulation, stress field for each load step is calculated and compared with the
cracking strength distribution. If the stress reaches the cracking strength, the matrix will
crack at the section and stress field will be calculated again. The process ends when the
load reaches peak bridging force. Knowing the cracking strength and post-cracking
bridging stress vs crack opening relation of different sections, the cracking sequence
and width of each crack can be obtained. The stress-strain curve can then be
established.
Fig. 5. Schematic diagram for calculating flaw size [7].
4.2 Parameters of Flaw Size Distribution

To study the effect of microstructure on the macro tensile behavior of SHCC, inversion
method was used to get the distribution parameter of flaws, specifically the mean size
and its standard deviation. Based on previous research [13–15], a typical set of material
parameter was adopted for the simulation, as listed in Table 3.
Table 3. A typical set of material parameter.

Fiber Fiber volume fraction 2%
Fiber length lf (mm) 12
Fiber diameter df (mm) 0.039
Fiber elastic modulus Ef (GPa) 18
Nominal fiber strength 1060
(continued)
Matrix Elastic modulus Em (GPa) 20
Fracture toughness Ktip (MPam0.5) 0.15
Poisson’s ratio m 0.2
Interface Frictional bond s0 (MPa) 1.31
Debonding fracture strength Gd (J/m2) 1.08
Snubbing coefficient f 0.5
Fiber strength reduction factor f’ 0.33
Slip hardening parameter b 0.58
With the parameter in Table 3, the bridging stress – crack opening relation curve
obtained by the model is shown in Fig. 6 and the peak stress is about 5.1 MPa at 93 µm
crack opening. To simplify the model and improve calculation efficiency, the same
bridging stress – crack opening curve is used for each section. In other words, the
variation of number of fibers among different sections is not considered, and the average
bridging stress – crack opening relation is employed. With this approach, the section
with the smallest value of maximum crack bridging stress, which governs the ultimate
strength of SHCC, cannot be properly simulated. To deal with this problem, the ultimate
stress of ‘the weakest section’ is set according to the test data, i.e. the average ultimate
stress of all the samples in a batch. When the stress reaches the ‘failure stress’, the
simulation program will stop and the simulated strain-stress curve will be plotted.
Three main characteristic parameters are most concerned in the simulation, namely
the first cracking strength, the ultimate stress and the ultimate strain. The sample is cut
into 200 sections and a total of 60 individual flaws are incorporated. To fit the different
tensile curves of the SHCC specimens with different viscosity, the mean and standard
deviation of the flaw size distribution were varied. Note that the distributions are
truncated at zero as the crack size cannot be negative. The attained flaw size distri-
bution parameters are listed in Table 4, and probability density curves of flaw sizes are
displayed in Fig. 7. With these parameters, the theoretical simulation of tensile
behavior coincides well with the experimental curve as plotted in Fig. 8.
5 Stress-crack opening 19s

Bridging Stress (MPa)
relation for a crack 30s

0.6 48s
4
Density
3
0.4
2
0.2
1
0 0
0 100 200 300 400 500 600 2 3 4 5 6
Crack Opening ( m) Diameter of flaws(mm)
Fig. 6. Stress-cracking opening relation Fig. 7. Probability density curve of simu-

for a single crack. lated flaw size.
1108 Z. Li et al.
Table 4. Parameters of flaw size distribution.

Marsh flow
cone time
19 s 30 s 48 s
Distribution of flaw size Mean value (mm) 2.6 2.9 3.5
Standard derivation (mm) 0.5 0.6 1.0
Other parameters Number of flaws 60 60 60
Number of sections 200 200 200
It should be mentioned that the fitting parameters, namely the mean value and
standard derivation of flaw size, are not the same as the physical flaw size and its
distribution. As mentioned in a previous section, many assumptions have been made in
the model. For example, flaws are assumed to be spherical and are treated as penny-
shaped flaws in analysing their effect on the cracking strength of individual sections.
The cracking strength of a section is assumed to be governed by the largest flaw only so
interaction among individual flaws is neglected. Moreover, many material parameters
employed in the model (shown in Table 3) are not directly measured but taken from the
literature. These assumptions may cause some inconsistencies with the real material.
Nevertheless, there should still be close correlation between the simulated parameters
and the real flaw distribution. Also, from the point of view of engineering application,
if these ‘effective’ flaw distribution parameters can be obtained as a function of flow
properties (e.g., a certain flow cone time), the effect of processing on SHCC behavior
can be predicted with the existing model. Although not a precise description of the
actual microstructure, the use of the ‘effective’ flaw size parameters can be a useful first
step in studying the processing-structure-properties relation of SHCC. In the following,
the effects of viscosity on flaw size parameters and SHCC behavior will be discussed.
4.3 Correlation of Flowability, Distribution Parameter and Tensile

Behavior
Flowability of mortars affects the flaw size distribution within the specimens and their
relation is reflected by the fitting parameters. From a general perspective, mortar with
better flowability is more likely to be better compacted. It is because it is easier for
entrapped air (especially large bubble) to escape, and hence densifying the matrix. This
point is consistent with the simulating parameters in Table 4 where the mean value and
standard deviation of flaw size both increase with the Marsh cone flow time.
From Table 2, the first cracking strength decreases with increasing flow time. In
general, first cracking strength of a specimen is dependent on the largest flaw size. As
shown in Fig. 7, the mean value and variation of flaws size of SHCC_III (48 s) is the
largest, thus there will be more large flaws so the first cracking strength is expected to
be the lowest. Similarly, the first cracking strength of SHCC_I (19 s) is the highest and
that of SHCC_II (30 s) is medium.
Fig. 8. Stress-strain curve with simulating parameter.
The ultimate strength of SHCC_II (30 s) is the highest among the three matrix, and
also shows less variation. During the tensile test, when the stress of a certain section
reaches its cracking strength, a crack is formed and the stress will be carried by the
fibers crossing the section. Thus the ultimate strength is governed by the weakest
section, which has the least number of fibers. For SHCC_I (19 s), a better flowability
means lower viscosity and lower shear force generated during mixing. In this case,
some fibers may stay as clusters, leading to poorer fiber dispersion and reduction of the
ultimate strength. If the viscosity is very high, as in SHCC-III (48 s), the fibers can
disperse well during mixing but the ultimate strength is found to be reduced (as shown
in Table 2). There are two plausible explanations. Firstly, as shown in Fig. 7, there are
many large flaws in the mix with low flowability, together with a high variation of the
flaw size. As a result, there will exist some sections with high porosity (associated with
large flaws) and others with low porosity. If the fibers are uniformly distributed within
the matrix, sections with high porosity will exhibit lower fiber content (as the matrix
fraction is reduced) [9], so the ultimate strength will also be reduced. Secondly, low
flowability may be associated with poorer wettability of the fibers, causing reduction in
the interfacial fracture energy and friction, which governs the fiber bridging stress.
These aspects will be experimentally investigated in our future work.
The ultimate strain of SHCC_II (30 s) is significantly higher than the other cases.
The total strain of SHCC is directly influenced by the number of cracks and the average
crack width at failure. As the same bridging stress – crack opening relation was used for
the three SHCCs, the average crack width is the same at the same stress level. The
number of cracks is affected by the ultimate stress and the matrix strength. In the tensile
test, the sections with the cracking strength lower than ultimate strength may crack. With
more sections undergoing cracking, the higher is the ultimate strain. For SHCC_III
(48 s), the large amount of large flaws weakened the matrix strength and reduced the
ultimate strength to only about 4 MPa. Therefore, only a small number of sections can
crack before failure, resulting in a lower ultimate strain. For SHCC_I (19 s), the ultimate
strength is about 5 MPa which is similar to SHCC_II (30 s). However, the relative small
flaw size makes the cracking strength of most sections higher than 5 MPa. With a
smaller number of cracked sections, the ultimate strain is reduced.
From the analysis above, SHCC_II (30 s) represents a mixing process closer to the
optimal. The resulting flaw sizes are not too big to influence the fiber distribution or
weaken the section too much, and the proper viscosity can disperse the fibers well so
1110 Z. Li et al.
that the ultimate strength among different sections are similar. Therefore, the ultimate
stress and strain are less variable, showing the best robustness.
5 Conclusion
This study investigated the correlation between the flowability of SHCC mortar, flaw
distribution parameters and macroscopic tensile behavior. Based on the numerical
model, flaw distribution parameters are attained by inversion method, which is con-
sistent with the rheology analysis and tensile results.
• The first cracking strength shows a negative correlation with Marsh cone flow time.
Low flowability of SHCC mortar could lead to large flaws and reduce the first
cracking strength.
• High flowability mortars could not provide enough shear force when mixing and
lead to poorer fiber dispersion. Low flowability mortars may introduce too many
large flaws and induce lower content of fibers in the weakest section. In both cases,
the ultimate strength is reduced.
• Ultimate strain is governed by the region where the cracking strength is lower than
the ultimate strength of the SHCC, which is governed by the weakest section. The
number of cracks that can form in the member is limited by both high flowability
(which results in dense matrix with high cracking strength for most of the sections)
and low flowability (which results in low ultimate strength).
• Based on the above, the highest ultimate strength and ultimate strain are obtained
for an intermediate flowability, which corresponds to Marsh cone flow time of 30 s
in our study.
In this study, the size distribution of effective spherical flaws is investigated by
inversion method. In future work, the variation of both the physical flaw size and fiber
distribution among various cross sections will be experimentally studied to provide
insights for the development of an improved model for the comprehensive tensile
behavior of SHCC.
Acknowledgements. The support of this research by the Hong Kong Research Grant Council
through the General Research Fund (Project Number 16215018) is gratefully acknowledged.
References
1. Li, V.C., Leung, C.K.Y.: Steady state and multiple cracking of short random fiber
composites. ASCE J. Eng. Mech. 188(11), 2246–2264 (1992)
2. Li, V.C.: From micromechanics to structural engineering–the design of cementitious
composites for civil engineering applications. JSCE J. Struct. Mech. Earth. Eng. 10(2), 37–
48 (1993)
3. Leung, C.K.Y.: Design criteria for pseudo-ductile fiber composites. ASCE J. Eng. Mech.
122(1), 10–18 (1996)
4. Kanda, T., Li, V.C.: A new micromechanics design theory for pseudo strain hardening
cementitious composite. ASCE J. Eng. Mech. 125(4), 373–381 (1999)
5. Leung, C.K.Y.: Performance-based design of SHCC components – research and challenges.

In: ‘Strain Hardening Cement-based Composites’, Proceedings of an International Confer-
ence, Dresden, pp. 429–440. Springer, Cham (2017)
6. Lu, C., Leung, C.K.Y.: A new model for the cracking process and tensile ductility of strain
hardening cementitious composites (SHCC). Cem. Concr. Res. 79, 353–365 (2016)
7. Lu, C., Leung, C.K.Y., Li, V.C.: Numerical model on the stress field and multiple cracking
behavior of engineered cementitious composites. Constr. Build. Mat. 133, 118–127 (2017)
8. Roussel, N., Le Roy, R.: The marsh cone: a test or a rheological apparatus? Cem. Concr.
Res. 35(5), 823–830 (2005)
9. Li, M., Li, V.C.: Rheology, fiber dispersion and robust properties of engineered cementitious
10. Zhou, J., Qian, S., Ye, G., Copuroglu, O., van Breugel, K., Li, V.C.: Improved fiber
distribution and mechanical properties of engineered cementitious composites by adjusting
the mixing sequence. Cem. Conc. Comp. 34, 342–348 (2012)
11. Aveston, J., Kelly, A.: Theory of multiple fracture of fibrous composites. J. Mater. Sci. 8,
352–362 (1973)
12. Kabele, P., Stemberk, M.: Stochastic model of multiple cracking process in fiber reinforced
cementitious composites. In: Proceedings of the 11th International Conference on Fracture.
Turin: CCI Centro Congressi Internazionale srl, Citeseer (2005)
13. Yang, E., Wang, S., Yang, Y., Li, V.C.: Fiber-bridging constitutive law of engineered
cementitious composites. J. Adv. Concr. Technol. 6, 181–193 (2008)
14. Kanda, T., Li, V.C.: Effect of fiber strength and fiber-matrix interface on crack bridging in
cement composites. J. Eng. Mech. 125, 290–299 (1999)
15. Lin, Z., Li, V.C.: Crack bridging in fiber reinforced cementitious composites with slip-
hardening interfaces. J. Mech. Phys. Solids 45, 763–787 (1997)
The Effect of Fiber Content
on the Post-cracking Tensile Stiffness
Capacity of R-UHPFRC
M. Khorami1,2, Juan Navarro-Gregori1(&), and Pedro Serna1

1
Institute of Science and Concrete Technology, ICITECH,
Universitat Politècnica de València, 46022 València, Spain
juanagre@cst.upv.es
2
Universidad UTE, Facultad de Arquitectura y Urbanismo,
Calle Rumipamba s/n y Bourgeois, Quito, Ecuador
Abstract. Concrete cracking can be controlled by adding fibers to concrete,

with the expected desirable behavior under serviceability conditions due to
narrower close space cracks compared to similar concrete without fibers. Using
fibers to produce Ultra-High Performance Fibre Reinforced Concrete
(UHPFRC) has enhanced the post-cracking tensile capacity of composite
material and increased the related energy absorption capacity for the cracked
member. Accordingly, the amount and type of fiber in the matrix affect post-
cracking behavior. In this experimental study, specimens reinforced by con-
ventional steel rebars with a constant cross-sectional dimension and reinforced
steel ratio were tested. The tested variables were: 1) type and length of fibers; 2)
fiber content. The Direct Tensile Test was conducted, and the tensile behaviour
of specimens was obtained. The results showed that the increment in fiber
content (80 kg/m3 to 160 kg/m3 in this research) or the combination of micro
and macro steel fibers with the same content (80 kg/m3 for each fiber type) had
no significant effect on the post-cracking stiffness capacity. Moreover, all the R-
UHPFRC specimens provided the full tension stiffening with the quasi same
post-cracking stiffness capacity close to the bare bar axial stiffness.
Keywords: Tension stiffening Post-cracking tensile stiffness Serviceability

UHPFRC
1 Introduction
In the structural concrete design, it is commonly assumed that reinforcement carries all
tensile force at the crack face [1]. Away from the crack face, due to the bond between
the steel rebar and the surrounding concrete, tensile stresses are shared between the
concrete and steel rebar. This contribution of concrete between cracks in tension is
commonly termed tension stiffening. This phenomenon has an effect on member
stiffness and is essential for determining serviceability deflection [2] and crack widths
[3]. The tension stiffening response should be included in the analysis to predict
member behaviour, under serviceability conditions.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_98
The Effect of Fiber Content 1113
The effects of cracking and reinforcement on member stiffness can be taken into
account with effective axial member stiffness (EAÞeff . The effective modulus method is
adopted by the fib Model Code 2010. It suggests an effective modulus for reinforce-
ment (E b ) and then the axial member response can be predicted using P ¼ E
b :As :em : (As
and em are the reinforcing bar area and the average member strain, respectively). The
load-strain relation and the effective modulus method to predict the member response
are shown in Fig. 1.
Fig. 1. Load-strain relationship and effective modulus method (MC2010)
The presence of the fiber in the concrete leads to enhanced post-cracking behaviour
and, as a result, improves the overall tensile response of the tensile elements. This
improvement is due to the combination of tension stiffening and strain hardening
mechanism. Strain hardening refers to the bridging effect and transmission of tensile
stresses by fibers across crack faces, and tension stiffening refers to the bond behaviour
of reinforcement and concrete [4].
Many theoretical and experimental studies have been performed to evaluate the
fiber content effect on post-cracking behaviour at the material consideration level for
FRC or UHPFRC [5–9], while real structural elements are combined by the rein-
forcement steel bar. Thus studying post-cracking behaviour for these reinforcement
elements is essential. In line with this, the present work focuses on the axial tensile
stiffness of cracked tensile elements (herein called post-cracking tensile stiffness). The
effect of fiber content was studied by employing three different doses and two types of
fibers for UHPFRC and comparing to Ultra-High Performance Concrete (UHPC). The
average tension stress-strain of tensile elements was obtained. Post-cracking tensile
stiffness was calculated by considering the tensile behaviour curve slope in the
elastoplastic region of behavior. Finally based on the experimental results, the cracking
behaviour and influence of fibers are also reported.
1114 M. Khorami et al.
2.1 Materials
Four different types of UHPFRC were used in this study given the difference in fiber
content and the type of steel fibers. Two types of steel fibers were considered in this
study. The first type was high-strength micro smooth steel fibers with a fine diameter
(df = 0.2 mm) and relatively short fiber length (Lf = 13 ± 0.1 mm). The second type
was macro hooked end fibers (length Lf = 30 ± 1 mm and diameter df = 0.375 mm).
The geometry and parameters of the steel fibers are presented in Table 1.
Table 1. Properties of steel fibers
The four types of UHPFRC were named as follows:

• (C160): the UHPFRC with fiber type (SF1) and fiber content of Vf = 2% or
160 kg/m3.
• (C80): the UHPFRC with fiber type (SF2) and fiber content of Vf = 1% or
80 kg/m3.
• (C8080): the UHPFRC with fiber type (SF1&SF2) and fiber content of Vf = 1% for
SF1, and Vf = 1% for SF2.
• (C0): the UHPC without fibers.
The average compressive strength of UHPFRC and UHPC was determined at the
tensile elements’ testing age with four cube specimens (100 100 100 mm). The
resulting values are found in Table 2.
Table 2. UHPFRC and UHPC average compressive strength

Concrete type fc ðMPaÞ
C160 158.55
C80 151.42
C8080 142.59
C0 136.69
The matrix composition of UHPFRC herein used in all the concretes was based on
our research group’s previous experiences [10–14]. Table 3 provides the matrix
composition of the employed UHPFRC and the UHPC matrix.
Table 3. Composition of the matrix mixture by the weight ratio for UHPFRC and UHPC
Cement Medium Fine Silica Silica fume Superplasticizer Water
(type I) sand sand flour (Elkem Microsilica,
0.6–1.2 mm 0.5 mm U-S500 grade 940)
1.00 0.70 0.37 0.28 0.22 0.037 0.20
2.2 Specimen Geometry and Test Set-Up

The uniaxial tensile tie test was carried out using prismatic concrete specimens with a
100x100 mm square cross-section and 1000 ± 3 mm length. The tensile element was
reinforced with a 12-mm central rebar (Es ¼ 200 GPa, fy ’ 550 MPa). Ten specimens
were cast for this study, six of them made with UHPFRC C160 and C80 (three for each
type), and two specimens for each C8080 and C0. The uniaxial tensile test was con-
ducted using a hydraulic jack machine with a loading capacity of 200 kN under
displacement control at a loading rate of 0.5 mm/min. The complete details and testing
procedure are available in M. Khorami et al. [15]. Figures 2a and 2b show the test set-
up and the R-UHPFRC tensile element.
a)
b)
Fig. 2. Uniaxial tensile test: (a) test set-up, (b) R-UHPFRC specimen
Eight displacement transducers (DTs) were installed on the surfaces of specimens

(four of them on the right side and four on the left side of the specimen) to record
element elongation during the test and to capture any undesired bending applied to the
specimen due to unforeseen load eccentricities. Each DT measured the length variation
between the fixing points placed with a 350 mm length from the center of specimens
toward the ends (see Fig. 2b).
3 Test Results and General Discussion

3.1 Tensile Behavior
The average stress-strain curves obtained for each tensile element type are shown in
Figs. 3a–c. The tensile stress in reinforcement (rs ) was calculated by dividing the
tensile load (N) by the reinforcement area (As ). The tensile elongation was measured
when the test started and after curing and the storage time. The authors have studied the
shrinkage effect on R-UHPFRC. UHPFRC shrinkage would shorten the member
without reinforcement, while embedded reinforcement would restrain concrete short-
ening. This leads to a negative pre-strain (es;sh ) with compressive stress in the rebar, and
initial tensile strain in the UHPFRC (ec;sh ). Hence the real origin of the bare steel rebar
response was modified by moving the compression strain value caused by shrinkage.
The experimental value obtained for the reinforcement strain due to UHPFRC
shrinkage was approximately (es;sh ¼ 0:40%).
The cracking stress level at the interaction point between the fit line over the
uncracked and cracked responses was defined. The slope of the elastoplastic region of
the tensile behavior represents the post-cracking tensile stiffness for the R-UHPFRC
tensile elements and was herein evaluated. The values calculated over the average
response of two or three specimens (red-colored curve) for each concrete are presented
in Table 4 according to the criteria shown in Fig. 4.
Fig. 3. Average tensile response of tensile elements for four concrete types.
According to the obtained results (Table 4), the post-cracking tensile stiffness
increased by a low value when increasing the fiber content when using the hybrid fibers
for TC8080. Moreover, these values came very close to the axial stiffness of the bare
bar with a value of c ¼ Es ¼ 200 GPa. To better understand the influence of fiber
content on post-cracking tensile stiffness and the difference between them, the stress-
strain relations for four concrete types are presented together, as shown in Fig. 5.
Table 4. Elastic tensile stiffness and post-cracking tensile stiffness

Specimen Stress in reinforcement at Stress in UHPFRC at Post-cracking tensile
ID cracking rs;cr ðMPaÞ cracking rc;cr ðMPaÞ stiffness c ðGPaÞ
TC0 112.00 1.28 N.A
TC80 455.00 5.20 205.83
TC160 563.00 6.44 225.43
TC8080 591.00 6.76 237.59
Fig. 4. Criteria for calculating R-UHPFRC tensile response.
Fig. 5. Tensile response of tensile elements with four concrete types.

As can be seen, R-UHPFRC tensile elements TC160, and TC8080 almost have the
same tensile cracking stress. Moreover, the post-cracking tensile stiffness for all three
types of R-UHPFRC tensile elements is in parallel to the bare bar response, while TC0
exhibits a very similar behavior to the tensile elements with conventional concrete (this
finding is discussed in more detail in Sect. 4). By comparing the tensile response of
TC80 with TC160, and TC8080, it can be concluded that if using a double fiber content
is necessary (e.g. for the durability aspect of UHPFRC), employing the hybrid fiber for
R-UHPFRC (80 kg/m3 for the micro smooth steel fibers, plus 80 kg/m3 for the macro
hooked end steel fibers) will present better tensile behavior on the one hand and, on the
other hand, the employed steel fiber weight will be the same for both cases (TC8080
and TC160). The cracking behavior aspect and the influence of fiber on cracking
propagation and crack width are discussed in Sect. 5.
4 UHPFRC Contribution in Tension
The tension stiffening response refers to the tension carried by the concrete between
cracks due to the reinforcing bar’s bond behavior. This ability increases the element’s
stiffness before reinforcement yields and can be used to predict tensile behaviour,
multiple crack spacing and crack widths.
Figure 6 is a qualitative representation of the tensile behavior of both R-UHPFRC
and R-UHPC. The cracked R-UHPFRC tensile element exhibits constant contribution
in tension, which refers to the parallel region of the behavior with the bare steel bar
(called herein full tension stiffening), while the tension stiffening response for the
UHPC tensile element after crack stabilization gradually decreases as the applied load
increases, and the member response curve moves closer to the bare bar response. The
bond factor parameter (b) accounts for the variation in the concrete average tensile
stress between cracks, and is generally expressed as an average tensile stress/cracking
stress ratio. The bond factor value vary from zero to one for no bonded reinforcement
and fully bonded, respectively (0\b\1).
Fig. 6. Qualitative representation of the tensile behavior of the R-UHPFRC and R-UHPC
elements, full tension stiffening concept.
The R-UHPFRC tensile elements (TC80, TC160, and TC8080) exhibit the full
tension stiffening with Beta values equaling one (see Fig. 5). This R-UHPFRC property
is essential when the serviceability state is evaluated, especially for deflection control.
Consequently, the high tension stiffening capacity of R-UHPFRC leads to good per-
formance under serviceability conditions and affirms the use of this material for special
structures in which durability or permeability of concrete elements is essential.
Moreover, it is evidenced that the cracking stress of the specimen without fibers (TC0)
is much lower compared to all the other specimens with fibers (TC80, TC160, and
TC8080).
For modeling the post-cracking tensile response of RC or FRC members, an
empirical relation is needed to represent the gradual transition from the uncracked axial
response to the fully cracked member (bare steel bar). An alternative approach to
predict the post-cracking tensile response consists in using an effective axial stiffness
(EAÞeff for the cracked member, which depends on the member’s strain level.
According to the fib Model Code 2010, the average strain of the RC or FRC member
(em ) is calculated by taking into account the tension stiffening effect, and can be
calculated by the average reinforcement strain (eb ) minus the average concrete strain
(Dec ):
em ¼ eb Dec ð1Þ
As the (Dec ) is variable for the RC and FRC tensile members, the R-UHPFRC
tensile elements provide a constant tension stiffening effect. Hence post-cracking ten-
sile modeling and the deflection calculation may involve less complexity.
5 Cracking Behavior
5.1 Fiber Content Effect
The crack distribution along the entire length element was determined to evaluate the
crack behavior of the tensile elements. The number of cracks was obtained at the end of
the test when the average tensile strain reached 2%. Water was used to wet the surface
so that micro cracks would be visible given the narrow width of cracks, which could
not be seen by the naked eye, and the crack pattern was painted on the specimen
surface. The number of cracks was recorded on each lateral surface at two surface
edges over the red line (see Fig. 7) and the average number of cracks was calculated.
The cracking measurement approach is shown in Fig. 7.
Fig. 7. Cracking measurement approach of the R-UHPFRC ties elements
It is worth mentioning that the R-UHPC tensile element exhibited a completely

different cracking behavior with the localized macro cracks and wide crack spacing,
while the R-UHPFRC ties provided a distribution crack propagation along the entire tie
length.
The average crack width (wm ) was calculated by dividing the total tension elon-
gation (et l) to the total number of cracks (n) as follows:
wm ¼ ðet lÞ=n ð2Þ
where (et ) is the average tensile strain at the time the number of cracks was measured
(2%), and (l) is tie length, which is 1000 mm. The parameter (n) is the average number
of the cracks. The obtained results are presented in Table 5, which reveal a clear
influence of crack width on the fiber content.
Employing a high-dose fiber content for concrete causes higher bond strength and
shorter transfer length [16]. Thus crack spacing will narrow, and the number of cracks
will increase. This phenomenon was observed when comparing the results for similar
tensile elements with different UHPFRC types in fiber content terms.
In serviceability behavior terms, the multiple-cracking with the distributed crack
propagation of R-UHPFRC led to very thin cracks (micro cracks) at the high tension
strain (em ¼ 2% in this study). The serviceability limit state can be controlled by
applying limitations for the tensile stresses in reinforcement. These limitations are
made to avoid inelastic strain, unacceptable cracking or deformation. Eurocode 2
indicates that the tensile stress in reinforcement cannot exceed 0.8 fyk , where fyk is the
characteristic yield strength of reinforcement.
Table 5. Average crack width at the tensile strain of 2%

Specimen Total number of Average crack width Eq. (2) Mean value
ID cracks (mm) (mm)
TC80-1 57 0.035 0.042
TC80-2 46 0.043
TC80-3 43 0.047
TC8080-1 63 0.032 0.028
TC8080-2 79 0.025
TC160-1 137 0.015 0.014
TC160-2 113 0.018
TC160-3 114 0.018
TC0-1 9 0.220 0.200
TC0-2 11 0.181
By comparing the mean crack width of TC160 to TC8080, it can be concluded that
using micro fibers led to a more efficient behavior compared to hybrid concrete
(TC8080), and both cases were better than TC80. In addition, the micro cracking
process was better controlled in UHPFRC with 160 kg/m3 fiber content. From the
serviceability perspective, the tensile elements with C160 exhibited better behavior
than those with C8080 despite the steel fiber content being the same.
6 Conclusions
In the present study, the impact of fiber content on the axial post-cracking tensile
stiffness capacity was studied by employing three UHPFRCs with fiber content and
fiber type variation, and one type of non fiber UHPC. The uniaxial tensile test was
conducted for tensile elements. Based on the experimental results, the following
conclusive remarks are drawn:
• The reduction of the axial tensile stiffness for the R-UHPFRC specimens between
the elastic region and micro crack stabilization region (the parallel zone of the
behaviour with the bare bar response) rapidly happens. Moreover, during the
applied tensile load, most micro cracks appear at the same time in this region of
behaviour.
• The tensile strength capacity in the cracking region depends on the fiber content of
R-UHPFRC. However, the slope of the overall tensile behaviour of all the R-
UHPFRC tensile elements (referring to the curve slope: c) almost parallels the bare
bar response. Consequently, the R-UHPFRC tensile elements provide the full
tension stiffening effect.
• Providing full tension stiffening for the R-UHPFRC members with a constant value
for concrete contribution for the cracked member leads to a facility of the com-
puting deflection of reinforced members with UHPFRC, while for reinforced RC or
FRC members an empirical model with descending branch after cracking is needed.
• Multiple cracking behaviour, and thus a large number of R-UHPFRC micro cracks,
emphasizes the high potential of UHPFRC composite material for structures in
which durability and permeability aspects are essential. From the serviceability
point of view, R-UHPFRC with a 2% micro fibers content provides better behavior
(crack width at a high rate of tensile strain) compared to the R-UHPFRC with
hybrid fibers (1% of micro fibers, plus 1% of macro fibers), while both contain the
same fiber content.
Acknowledgments. This study forms a part of Project BIA2016-78460-C3-1-R supported by

the State Research Agency of Spain.
References
1. Bischoff, P.H.: Reevaluation of deflection prediction for concrete beams reinforced with
steel and fiber reinforced polymer bars. J. Struct. Eng. 131, 752–767 (2005)
2. Visintin, P., Oehlers, D., Muhamad, R., Wu, C.: Partial-interaction short term serviceability
deflection of RC beams. Eng. Struct. 56, 993–1006 (2013)
3. Bischoff, P.H.: Tension stiffening and cracking of steel fiber-reinforced concrete. J. Mater.
Civ. Eng. 15, 174–182 (2003)
4. Bernardi, P., Cerioni, R., Michelini, E.: Analysis of post-cracking stage in SFRC elements
through a non-linear numerical approach. Eng. Fract. Mech. 108, 238–250 (2013)
5. Kooiman, A., Walraven, C.: Modelling the post-cracking behaviour of steel fibre reinforced
concrete for structural design purposes. HERON 45(4), 2000 (2000)
fibre-reinforced concretes. Constr. Build. Mater. 25, 2713–2722 (2011)
7. Abrishambaf, A., Barros, J.A., Cunha, V.M.: Relation between fibre distribution and post-
cracking behaviour in steel fibre reinforced self-compacting concrete panels. Cem. Concr.
Res. 51, 57–66 (2013)
8. Pereira, E., Barros, J.A., Ribeiro, A.F., Camões, A.: Post-cracking behaviour of selfcom-
pacting steel fibre reinforced concrete. In: 6th International RILEM Symposium on Fibre-
Reinforced Concretes (2004)
9. Zhou, B., Uchida, Y.: Relationship between fiber orientation/distribution and post-cracking
behaviour in ultra-high-performance fiber-reinforced concrete (UHPFRC). Cement Concr.
Compos. 83, 66–75 (2017)
10. López, J., Serna, P., Navarro-Gregori, J., Coll, H.: Comparison between inverse analysis
procedure results and experimental measurements obtained from UHPFRC Four-Point
Bending Tests. In: Proceedings of the 7th RILEM Workshop on High Performance Fiber
Reinforced Cement Composites (HPFRCC7), pp. 185–192 (2015)
Mater. Struct. 48(11), 3703–3718 (2014)
12. López Martínez, J.A.: Characterisation of the tensile behaviour of UHPFRC by means of
four-point bending tests. Ph.D. thesis (2017)
13. Mezquida-Alcaraz, E.J., Navarro-Gregori, J., Lopez, J.A., Serna-Ros, P.: Validation of a
non-linear hinge model for tensile behavior of UHPFRC using a Finite Element Model.
Comput. Concr. 23, 11–23 (2019)
14. Mezquida-Alcaraz, E., Navarro-Gregori, J., Serna-Ros, P.: Numerical validation of a

simplified inverse analysis method to characterize the tensile properties in strain-softening
UHPFRC. In: IOP Conference Series: Materials Science and Engineering, p. 012006. IOP
Publishing (2019)
15. Khorami, M., Navarro-Gregori, J., Serna, P., Navarro-Laguarda, M.: A testing method for
studying the serviceability behavior of reinforced UHPFRC tensile ties. In: IOP Conference
Series: Materials Science and Engineering, p. 012022. IOP Publishing (2019)
16. Harajli, M., Hamad, B., Karam, K.: Bond-slip response of reinforcing bars embedded in
plain and fiber concrete. J. Mater. Civ. Eng. 14, 503–511 (2002)
Controlling Strength and Ductility
of Strain-Hardening Cementitious
Composites by Nano-Engineering
Ousmane A. Hisseine(&) and Arezki T. Hamou
Cement and Concrete Research Group, University of Sherbrooke,

Sherbrooke, QC, Canada
O.Hisseine@USherbrooke.ca
Abstract. Considering the hierarchical nature of cracking in cement compos-

ites, multi-scale reinforcement bears the potential to enhance the fracture per-
formance of fibre-reinforced cementitious composites. This study shows how
nanoscale cellulose filaments (CF) can be used as a novel tool for tailoring the
properties of strain-hardening cementitious composites (SHCC) towards
improved strength and ductility. SHCC with fly ash-to-cement ratio of 1.2 and
incorporating CF at rates 0.03, 0.05 and 0.10% of cement mass were developed
following the micromechanical principles for pseudo-ductile cement compos-
ites. Results indicate that the incorporation of CF in SHCC allows nanoengi-
neering matrix and interface properties by increasing matrix elastic modulus and
imparting a significant slip-hardening effect. Consequently, higher comple-
mentary energy and lower crack tip toughness were obtained, thereby leading to
enhanced ductility as also validated by tensile and flexural tests. As such, the
incorporation of CF enhanced composite tensile strength by up to 23% and
increased the ultimate strain capacity in tension by up to 26% and the deflection
capacity in flexure by up to 36%. Therefore, nano-engineering SHCC with CF
yields multi-scale composites with higher ductility without necessarily
increasing the volume fraction of PVA fibres while exhibiting higher strength
without necessarily increasing the binder content.
Keywords: Cellulose filaments (CF) Engineered cementitious composites

(ECC) Nanocellulose Nanoengineered concrete Recycled glass powder
(RGP) Ground-glass pozzolans (GP) Strain-hardening cementitious
composites (SHCC)
1 Introduction
Bearing in mind the increasing socioeconomic burden associated with the rehabilitation
and reconstruction of aging concrete infrastructure, today’s concrete technology practi-
tioners are challenged to develop concrete recipes demonstrating the highest perfor-
mance, the least ecological footprint, and the maximum performance-to-investment ratio.
Among concrete types bearing the potential to meet such stringent requirements is strain-
hardening cementitious composite (SHCC) also known as engineered cementitious
composite (ECC). SHCC belongs to high-performance fibre-reinforced cementitious
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_99
Controlling Strength and Ductility of Strain-Hardening Cementitious Composites 1125
composites (HPFRCC). The salient features of SHCC include its significant tensile strain
capacity reaching up to 11%, the tight multiple cracking as well as the strain-hardening
behaviour at low fibre content, commonly 2% [1–4]. While extended multiple
cracking is a fundamental requirement in SHCC, crack localization often leads to a
softening behaviour, thereby jeopardizing composite ductility. One way to foster SHCC
ductility is to consider nano-engineering matrix and interface properties. Nano-
reinforcement can favour a multi-scale crack resistance such that sequential multiple
cracking and strain-hardening can be obtained. The survey of literature indicates that the
incorporation of 0.08 wt% graphene oxide (GO) in SHCC enhanced the tensile strength
by up to 38% and the flexural capacity by up to 81% [5]. Nanocellulose fibres were proven
to enhance the fracture energy of UHPC by more than 50% [6]. SHCC with nano-
carbonate whiskers incorporated at 0.5 vol. % led to 30% enhancement in compressive
strength, 53% in ultimate tensile strength, 114% in the tensile strain capacity [7]. Con-
sidering the effect of nano-reinforcement on enhancing the ductility of SHCC, the
hypothesis of this study is that the use nanoscale cellulose filaments (CF) can also enhance
the ductility of SHCC.
CF are nanoscale rod-like cellulosic particles with 30–400 nm diameter and 100–
2000 µm length and belong to nanocellulose materials (NCM) such as cellulose
nanocrystals (CNC); microfibrillated cellulose (MFC; and nanofibrillated cellulose
(NFC) [8, 9]. Owing to their intrinsic mechanical strength imparting inherent strength
to plants, NCM are attracting substantial attention in versatile applications including
cement and concrete composites [8–10]. Our former investigations on CF demonstrate
enhancement in elastic modulus (18%) [8], flexural capacity (25%), and toughness
(96%) [9] attributable to higher microstructure properties (increased degree of hydra-
tion of 12% and enhanced micromechanical properties of C-S-H gel matrix of
12–25%) [8]. The current study aims at leveraging the advantages of CF to enhance the
ductility of SHCC. Research outcomes are expected to contribute towards the devel-
opment of high-performance cement composites while contributing towards enhancing
concrete ecoefficiency.
2.1 Materials Properties
SHCC ingredients include type HS cement, type F- fly ash (FA), and quartz sand
(QS) with a maximum particle size of 600 µm. The cement has a specific gravity
(SG) of 3.18, Blaine fineness of 438 m2/kg, and mean particle diameter (d50) of 12 µm.
The FA used in the study fulfils the requirements of CAN/CSA A3000 specifications
and has an SG of 2.55, Blaine surface area of 363 m2/kg, and d50 of 17. The GP used
herein is of high alkali content and has an SG of 2.51 and d50 of 27 µm. As for the QS,
it has an SG of 2.70, maximum particle size (dmax) of 600 µm, and a d50 of 250 µm.
Figure 1 provides the particle size distribution of granular materials used in this study.
Polyvinyl-alcohol (PVA) fibres (with 38 µm diameter, 8 mm length, 40 GPa elastic
modulus and 1400 MPa tensile strength were added at 2% per volume). The cellulose
filaments (CF) shown in Fig. 4 have 30–400 nm diameter, 100–2,000 µm length, and a
1126 O. A. Hisseine and A. T. Hamou
surface area of more than 80 m2/g. Further information about CF characteristics and
sustainability features can be found in our former works [8]. Figure 2 depicts a
scanning electron microscope (SEM) image of a CF diluted aqueous suspension at a
concentration of 0.10%.
Fig. 1. Particle-size distribution of SHCC ingredients.
2.2 Mixture Proportions

Four SHCC were considered in this investigation. This includes a reference SHCC
mixture as well as three SHCC incorporating cellulose filaments (CF). The reference
SHCC has FA/cement ratio of 1.2, QS to binder ratio of 0.35 and a water-to-binder
(w/c) ratio of 0.28. CF was incorporated at dosages of 0, 0.03, 0.05, and 0.10% per
mass of cement. Table 2 shows mixture proportioning. The amount of high-range water
reducing admixture was adjusted in function of CF in order to achieve the target slump-
flow of 300 mm in the suspended mortar and 250 mm in final SHCC. Resulting
mixtures were tailored via the micromechanics design approach of SHCC proposed by
Li et al. [11–13].
2.3 Mixing Procedures and Specimen Preparations

A pan mixer of type Mortarman 360 was used to prepare the different batches using the
following sequence:
• All granular materials were first dry–mixed for 7 min prior to adding water and
HRWRA.
• A 90% of the HRWRA diluted into 95% of the mixing water was added to the
mixer slowly during 0.5 min then mixing continued for 2.5 min. For mixtures
incorporating CF, a CF-water suspension was first prepared from readily dispersed
CF (colloidal suspensions with 1.2% CF solid content). Thereafter, 90% of
HRWRA was diluted into 95% of CF-water suspension. The remaining 10% of
HRWRA is used for final adjustment of mixture flowability.
• The mixer was then stopped for 0.5 min to scrape its blades and edges then mixing
continued for another 1 min. The consistency of the suspended mortar was then
checked such that when a mini-slump flow diameter of 300 mm was obtained [14],
Fig. 2. SEM micrograph of nanoscale cellulose filaments (CF). Figure adapted from [9] with
permission from ASCE Journal of Materials in Civil Engineering.
only the remaining water is added (after adjustment of the amount of water con-
tained in the unused 10% of HRWRA). Otherwise, a gradual amount of HRWRA is
added (along with an adjusted amount of the remaining 5% of water) until the
desired 300 mm mini-slump flow diameter is achieved.
• PVA fibres were added slowly during 1 min. Mixing continued for another 3 min to
allow PVA fibres to be evenly dispersed.
• Finally, following 2 min of rest, concrete is remixed for 1 min then sampled for the
different tests. Specimens for mechanical properties were covered with plastic
sheets and kept in a room with relative humidity and temperature of approximately
50% and 23 °C, respectively then demoulded or 24 ± 1 h later. The specimens
were then sealed inside plastic bags then transferred for storage in a fog room at
100% RH and 22 °C temperature until the age of testing.
2.4 Testing
Testing included two distinct series: the first was conducted on non-fibrous (suspended)
mortar to collect the parameters necessary for micromechanical tailoring of SHCC
mixtures. Those tests are the fracture toughness, the direct tension, and the single-fibre
pull-out tests.
• The elastic modulus (E) of the matrix was evaluated on 100 200 mm cylinders
at 28 days as per ASTM C469 [15].
• The fracture toughness (Km) of the matrix was determined following a procedure
adapted from ASTM E399 [16]. The test was conducted at 28 days on
(100 100 400 mm) plain prisms of 25 mm deep, 3 mm central notch tested in
three-point bending configuration in a displacement-controlled mode using a dis-
placement rate of 0.05 mm/min.
• The direct tensile strength was conducted on dog-bone shaped coupons (Fig. 3) in
order to determine the tensile strength of the plain matrix (rfc). The test was conducted
under a displacement-controlled mode at a rate of 0.2 mm/min. The test set-up is

equipped with a 24.4 mm wire extensometer attached at the middle of the sample.
Fig. 3. Uniaxial tension test: Sample dimensions (left) and test configuration (right).
• Single fibre pull-out test (SFPT) was conducted to evaluate fibre/matrix interface
properties necessary for micromechanical tailoring. A 5-N capacity of load cell was
used for this purpose. The test was conducted under a displacement-controlled
mode at a rate of 0.2 mm/min (See Fig. 4 (a) for the test set-up and Fig. 4 (b) for the
actual test configuration). Further details about SFPT as well as the determination of
interface parameters [frictional bond (s0), chemical bond (Gd), and slip hardening
coefficient (b)] are provided elsewhere [17].
Fig. 4. Single-fibre pull-out test: Schematic of test set-up (a), actual test configuration with a
focus on a single fibre being pulled from the matrix (b). Figure adapted from Hisseine et al. [17]
with permission from Elsevier.
The second test series was conducted on SHCC to assess the behaviour at the
composite level as well as to validate the outcome of the micromechanical tailoring.
Those tests include the compressive strength, the uniaxial tensile strength and the
flexural capacity. The compressive strength was assessed on 50 50 50 mm cubes

(ASTM C109-16) [18]. The uniaxial tensile strength was evaluated at 28 days fol-
lowing the same procedures and specimen dimensions described earlier for the case of
the plain matrix. The flexural capacity was evaluated on 100 100 400 mm prisms
(ASTM C78-18) [19] at 28 days. Flexural tests were performed under four–point
bending configuration in a displacement-controlled mode using a displacement rate of
0.05 mm/min. Elaborated experimental details can be found elsewhere [17, 20, 21]
3.1 Effect of Nano-Modification with CF on Micromechanical Properties

Table 1 presents the results of the micromechanical investigation conducted in this
study. Results show that the incorporation of nanoscale cellulose filaments
(CF) influence the properties of SHCC matrix as well as fibre/matrix interface prop-
erties. For matrix properties, CF appears to moderately increases the elastic modulus
(Em), the fracture toughness (Km), and the first cracking strength (rfc). The enhance-
ment in the Em (around 15%) can affect the fracture behaviour of SHCC. Given the
definition of crack tip toughness (Jtip) as a function of the elastic modulus and the
fracture toughness (Jtip ¼ Km2 =Em ), it is perceivable that that the increase in Em can
lower Jtip and allow fulfilling the energy criterion (Jtip Jb0 ) for SHCC [24]. In terms
of interface properties, the use of CF slightly attenuated the frictional bond s0, sig-
nificantly reduced the chemical bond Gd, and imparted a characteristic slip-hardening
effect b. CF influences interface properties by altering the fibre/matrix interfacial
transition zone (ITZ). The observed reduction in frictional bond s0 in the presence of
CF is attributable to the interference of the omnipresent CF between the matrix and
PVA fibres (Fig. 5), thereby reducing the contact sites between the matrix and the PVA
fibres. While high frictional bond s0 favours higher bridging stress, too high s0 can
cause fibre premature rupture, thereby jeopardizing composite ductility. Likewise, the
observed reduction in the chemical bond Gd in the presence of CF is ascribable to the
attenuation in contact sites between the PVA fibres and active cations from the matrix
such as Ca2+ known to influence the affinity of PVA fibres to adhering to the matrix
[22, 23]. The characteristic slip-hardening behaviour observed in systems with CF, on
the other hand, can sprout from the propensity of the flexible nanoscale CF to causing a
jamming effect during PVA fibre pull-out, thereby leading to resisting higher pull-out
load.
3.2 Effect of Nano-Modification with CF on Fibre Bridging Capacity

Figure 6 presents the fibre bridging stress-crack opening response (r–d) for the dif-
ferent SHCC developed in this study. Using the micromechanical principles of SHCC,
the strain hardening performance indicators for the different SHCC were also obtained
and are presented in Table 2. The results of Fig. 6 indicate that the incorporation of CF
has no observable influence on the maximum bridging capacity (r0) in spite of the
aforementioned positive effect of matric strength [elastic modulus (Em), fracture
Table 1. Mixture proportions

Mixture Mixture Composition (kg/m3)
name Cement Fly ash Quartz sand Water HRWRA (solid CF
(FA) (QS) extract) (solid
extract)
M 0.00CF 597 717 460 366 3.251 –
M 0.03CF 597 717 460 366 4.014 0.175
M 0.05CF 597 717 460 366 4.125 0.292
M 0.10CF 597 717 460 366 4.589 0.584
Fig. 5. Effect of nanomodification with CF on matric and interface properties of SHCC:

(a) omnipresence of CF on PVA fibre surface, (b) interference of CF on fibre/matrix interface,
and (c) nan-reinforcing of matrix by CF. Figure adapted from Hisseine et al. [17] with permission
from Elsevier.
toughness (Km), and first cracking strength (rfc)]. This can be a consequence of the
slight reduction in frictional bond s0 in the presence of CF. Nonetheless, all systems
with CF recorded higher complementary energy Jb0 . Consequently, the assessment of
strain-hardening indicators presented in Table 2 demonstrates that the index Jb0 /Jtip is
higher in all SHCC incorporating CF than in the reference. Likewise, the ratio r0/rfc is
maintained above 1.45 as required for substantial strain-hardening behaviour. Kanda
and Li [24] have suggested that additional to satisfying the strength and energy criteria
required, SHCC should also meet two more conditions: Jb0 /Jtip > 3 and r0/rfc> 1.45.
With both ratios enhanced in all SHCC incorporating CF, it is perceivable that nano-
modifying SHCC with CF fosters composite ductility.
3.3 Effect of Nano-Modification with CF on Composite Performance

To validate the outcome of the micromechanical investigation, the performance of
resulting SHCC was evaluated at the composite level in terms of compressive strength,
uniaxial tensile strength and flexural capacity.
Fig. 6. Computed fibre bridging stress versus crack opening relationship (r–d) for varying GP
content.
3.3.1. Effect of CF on compressive strength

Figure 7 presents the effect of CF on the evolution of compressive strength up to 91
days. Overall, the incorporation of CF leads to strength enhancement particularly from
7 days and onward. While it is not evident to make a clear distinction between the effect
of different CF dosages, it is observable that all CF concentrations led to higher
compressive strength compared to that of the reference SHCC (without CF). With
better effects observed at later curing ages, strength enhancement of 15-20% were
recorded. This can be ascribed to the mechanism underlying the effect of CF on cement
composites whereby the hydrophilic and hygroscopic CF can contribute to enhancing
composite strength by an internal curing effect [9] as well as by a nanoreinforcing effect
allowing to bridge cracks at the scale of hydrates [8]. Assessment of autogenous
shrinkage in low water-to-cement ratio systems in our former work indicate that CF can
contribute by internal curing effect to control volumetric instability whereby a reduc-
tion in autogenous shrinkage of up to 31% were obtained [9]. On the other hand,
microstructure assessment as well as nanoindentation studies revealed that CF
strengthen cement composites by a twofold effect: (i) increased degree of hydration
(15%) and (ii) higher micromechanical properties of C-S-H matrix ( 12–25%) [8].
3.3.2. Effect of CF on uniaxial tensile strength and on flexural capacity

Figure 8 depicts the uniaxial tensile behaviour for the different SHCC considered
herein. The figure indicates that nano-modification by CF has influenced the first
cracking strength (rfc), the post-cracking strength (rpc), the multiple cracking and
strain-hardening response as well as the ultimate tensile strain capacity (eu). The highest
enhancement in the first cracking strength rfc was obtained at 0.10% CF where an
increase of 20% was observed. On the other had, the three respective CF dosages
enhanced the average post-cracking strength rpc from 3.34 MPa in the reference SHCC
to 3.48, 3.65, and 4.10 MPa, corresponding to enhancements of 5, 10, and 23%,
respectively. As for the ultimate tensile strain capacity (eu), the incorporating CF
enhanced eu from 3.01% in the reference SHCC [Fig. 8 (a)] to 3.33, 3.60, and 3.78% in
the systems with 0.03, 0.05, and 0.10% CF [(Fig. 8 (b), (c), (d)]. This corresponds to
enhancements of 11, 20, and 26%, respectively. The enhancement in eu is a conse-
quence of a more prominent multiple cracking response reflected by the intensity and
1132
Table 2. Results of micromechanical investigation and strain-hardening indicators for SHCC

Mixture Matrix parameters Interface parameters SHCC pseudo-ductility performance
name Jb; r0
Elastic Fracture First crack Frictional Chemical Slip Maximum Comple- Crack rfc
Jtip
modulus, toughness, uniaxial bond, bond, hardening bridging mentary tip
O. A. Hisseine and A. T. Hamou
Em Km strength, rfc s0 Gd coefficient, stress, energy, toughness,

pffiffiffiffi
(GPa) (MPa. m) (MPa) (MPa) (J/m2) b r0 Jb; Jtip
(MPa) (J/m2) (J/m2)
M0.00CF 22.50 0.48 0.68 0.05 2.59 0.11 2.95 0.52 2.65 0.14 0.003 4.82 64 20.5 3.12 1.86
M0.03CF 25.74 0.21 0.70 0.08 2.90 0.12 2.94 0.56 0.27 0.10 1.24 0.39 5.03 67 19.0 3.50 1.73
M0.05CF 25.50 0.19 0.71 0.04 2.90 0.08 2.82 0.66 0.25 0.12 1.51 0.20 5.15 66 19.8 3.34 1.78
M0.10CF 24.69 0.32 0.71 0.07 2.92 0.04 2.80 0.78 0.21 0.06 1.31 0.34 5.18 67 20.4 3.28 1.77
90
Compressive strength, fc (MPa) M 0.0%CF
75 M 0.03%CF
M 0.05%CF
60 M 0.10%CF
45
30
15
0
1 7 28 56 91
Age (days)
Fig. 7. Compressive strength development
frequency of peaks in the uniaxial tensile behaviour. This is more pronounced in the
SHCC with 0.10% CF [Fig. 8 (d)].
The results of uniaxial tensile response are further confirmed by the flexural
response shown in Fig. 9. The figure shows that the incorporation of CF influenced
particularly the deflection capacity du (identified by the mid-span deflection corre-
sponding to the maximum flexural load). As such, compared to the reference SHCC
[Fig. 9 (a)] recording an average du of 3.91 mm, the incorporation of CF at 0.03, 0.05,
and 0.10% [Fig. 9 (b), (c), (d), respectively] resulted in average du of 4.51, 4.92, and
5.31 mm. This corresponds to 15, 26 and 36% higher du. The improved deflection
capacity demonstrates the effectiveness of CF in imparting higher ductility. These
results can be linked to the micromechanical investigation where nano-modified SHCC
showed higher complementary energy (Jb; ). The enhanced strain-hardening response
and extended deflection capability in systems with CF can also be linked to the
characteristic slip-hardening behaviour (b) imparted by CF as discussed earlier. The
slip-hardening response imparted by CF was demonstrated elsewhere to allow the
matrix to withstand pull-out loads even higher than the load at which the initial
fibre/matrix bond deteriorated. As such, further energy is consumed to complete fibre
pull-out. This can be reflected by the observed enhancement in the multiple cracking
response as well as in the deflection capacity and in the overall ductility [17, 21].
Fig. 8. Uniaxial tensile strength
Fig. 9. Flexural capacity

Nano-engineered strain-hardening cementitious composites (SHCC) have been devel-

oped in this study following micromechanical tailoring. Cellulose filaments (CF) were
incorporated at dosages of 0, 0.03, 0.05, and 0.10% in SHCC with fly ash to cement
ratio of 1.2 and a water-to-binder ratio of 0.28. Results demonstrate that the incorpo-
ration of CF enhances both strength and ductility characteristics of SHCC. Specific
findings are as follows:
• The incorporation of CF influenced the mechanical properties of the plain matrix as
well as the micromechanical properties at fibre/matrix interface. SHCC with CF
recorded higher elastic modulus, a reduced chemical bond energy, and a charac-
teristic slip-hardening effect.
• The effect of CF on elastic modulus was positively reflected by an improved
complementary energy and a reduced crack tip toughness. Consequently, higher
strain-hardening performance was obtained.
• Experimental validation by tensile and flexural tests proved that the incorporation of
CF enhanced composite tensile strength by up to 23% and increased the ultimate strain
capacity in tension by up to 26% and the deflection capacity in flexure by up to 36%.
Finally, results support the effectiveness of nano-engineering SHCC with CF to foster
composite strength and ductility without necessarily increasing the volume fraction of
PVA fibres or the binder content. This provides a twofold benefit in terms of utilization of
the most abundant and renewable natural polymer on the planet to enhance the perfor-
mance of cement and concrete composites as well as to foster their ecoefficiency.
Acknowledgements. This project is jointly supported by a Cooperative Research and Devel-

opment (CRD) grant from the Natural Sciences and Engineering Research Council of Canada
(NSERC), Canada Vanier Graduate Scholarship (CGS) program award no: 360284, Kruger
Biomaterials Inc. (QC, Canada), and Euclid Chemicals. The authors are grateful to the financial
support from all these partners.
References
1. Li, V.C., Leung, C.K.Y.: Steady state and multiple cracking of short random fibre
composites. ASCE J. Eng. Mech. 188(11), 2246–2264 (1992)
2. Guan, X., Li, Y., Liu, T., Zhang, C., Li, H., Ou, J.: An economical ultra-high ductile
engineered cementitious composite with large amount of coarse river sand. Constr. Build.
Mater. 201, 461–472 (2019)
3. Yu, K.Q., Yu, J.T., Dai, J.G., Lu, Z.D., Shah, S.P.: Development of ultra-high performance
engineered cementitious composites using polyethylene (PE) fibres. Constr. Build. Mater.
158, 217–227 (2018)
4. Ding, Y., Yu, J., Yu, K.Q., Xu, S.: Basic mechanical properties of ultra-high ductility
cementitious composites: From 40 MPa to 120 MPa. Compos. Struct. 185, 634–645 (2018)
5. Meng, W., Khayat, K.H.: Mechanical properties of ultra-high-performance concrete
enhanced with graphite nanoplatelets and carbon nanofibers. Compos. B Eng. 107, 113–
122 (2016)
6. Peters, S.J., Rushing, T.S., Landis, E.N., Cummins, T.K.: Nanocellulose and microcellulose
fibres for concrete. Transport. Res. Rec. 2142, 25–28 (2010)
7. Ma, H., Cai, J., Lin, Z., Qian, S., Li, V.C.: CaCO3 whisker modified Engineered
Cementitious Composite with local ingredients. Constr. Build. Mater. 151, 1–8 (2017)
8. Hisseine, O.A., Wilson, W., Sorelli, L., Tolnai, B., Tagnit-Hamou, A.: Nanocellulose for
improved concrete performance: a macro-to-micro investigation for disclosing the effects of
cellulose filaments on strength of cement systems. Constr. Build. Mater. 206, 84–96 (2019)
9. Hisseine, O.A., Omran, A.F., Tagnit-Hamou, A.: Influence of cellulose filaments on cement
pastes and concrete. J. Mater. Civ. Eng. 30(6), 04018109 (2018)
10. Hisseine, O.A., Basic, N., Omran, A.F., Tagnit-Hamou, A.: Feasibility of using cellulose
filaments as a viscosity modifying agent in self-consolidating concrete. Cem. Concr.
Compos. 94, 327–340 (2018)
11. Li, V.C.: Engineered Cementitious Composites (ECC) – Tailored Composites Through
Micromechanical Modeling, Fiber Reinforced Concrete: Present and the Future, pp. 64–97.
Canadian Soc. Civil Eng., Montreal (1998)
12. Li, V.C., Wang, S., Wu, C.: Tensile strain-hardening behaviour of polyvinyl alcohol
engineered cementitious composite. ACI Mater. J. 98(6), 483–492 (2001)
13. Li, V.C.: On engineered cementitious composites (ECC) – a review of the material and its
application. J. Adv. Concr. Technol. 1, 215–230 (2003)
14. Kong, H., Bike, S.G., Li, V.C.: Development of a self-consolidating engineered cementitious
composite employing electrosteric dispersion/stabilization. Cem. Concr. Compos. 25(3),
301–309 (2003)
15. ASTM C469/ C469M-14, Standard Test Method for Static Modulus of Elasticity and
Poisson’s Ratio of Concrete in Compression, ASTM International, West Conshohocken, PA,
2014 www.astm.org
16. ASTM E399-12, Standard test method for linear-elastic plane-strain fracture toughness KIc
of metallic materials, ASTM International, West Conshohocken, PA, 2012. www.astm.org
17. Hisseine, O.A., Tagnit-Hamou, A.: Characterization and nano-engineering the interface
properties of PVA fibres in strain-hardening cementitious composites incorporating high-
volume ground-glass pozzolans, Constr. Build. Mater. 234, 117213 (2020)
18. ASTM C109/ C109M-16a, Standard Test Method for Compressive Strength of Hydraulic
Cement Mortars (Using 2-in. or [50-mm] Cube Specimens), ASTM International, West
Conshohocken, PA, 2016 www.astm.org
19. ASTM C78/ C78M-18, Standard Test Method for Flexural Strength of Concrete (Using
Simple Beam with Third-Point Loading), ASTM International, West Conshohocken, PA,
2018 www.astm.org
20. Hisseine, O.A., Tagnit-Hamou, A.: Development of ecological strain-hardening cementitious
composites incorporating high-volume ground-glass pozzolans. Constr. Build. Mater. 238,
117740 (2020)
21. Hisseine, O.A., Tagnit-Hamou, A.: Nanocellulose for the development of ecological
nanoengineered strain-hardening cementitious composites incorporating high-volume
ground-glass pozzolans, article under review by Cement and Concrete Composites (2020)
22. Rodger, S.A., Brooks, S.A., Sinclair, W., Groves, G.W., Double, D.D.: High strength
cement pastes. J. Mater. Sci. 20, 2853–2860 (1985)
23. Gulgun, M.A., Kriven, W.M., Tan, L.S., McHugh, A.J.: Evolution of mechano-chemistry
and microstructure of a calcium aluminate-polymer composite: Part I—Mixing Time Effects.
J. Mater. Res. 10(7), 1746–1755 (1995)
24. Kanda, T., Li, V.C.: Multiple cracking sequence and saturation in fibre reinforced
cementitious composite, Concr. Res. Technol., JCI 9 (2) (1998) 19–33
Comprehensive Characterization of UHPFRC
Mixes for Seismic and Durability
Rehabilitation of Bridge Piers
C. Sevigny-Vallières1(&), P. Marchand2, B. Terrade2, N. Roy1,

F. Toutlemonde2, and A. Tagnit-Hamou1
1
Université de Sherbrooke, Sherbrooke, Québec, Canada
claudine.sevigny.vallieres@usherbrooke.ca
2
Materials and Structures Department, Université Gustave Eiffel,
Marne-la-Vallée, France
Abstract. Towards design of UHPFRC jacketing for rehabilitation of bridge

piers, one UHPFRC mix has been extensively characterized in order to provide
the necessary material characteristics for such a design verification process. This
mix is an innovative environmental-friendly UHPFRC mix, comprising recycled
glass powder as cement and quartz powder replacement. The mechanical
characterization of the mix comprises compressive strength and its evolution,
Young’s modulus at early age and mature state, development of autogenous and
total shrinkage, and tensile behaviour identified by flexural testing both on
standard prismatic specimens as well as on thin plates representative of the
UHPFRC jacketing layer cast around the existing concrete pier. Development of
this effort has been undertaken in a joint project within ECOMAT international
laboratory with Université de Sherbrooke (Canada) and the Materials and
Structures Department of Université Gustave Eiffel (France).
Keywords: UHPFRC Glass powder Early age properties Mechanical

properties Tensile characterization Bridge piers
1 Introduction
Both in Canada and France, the bottom parts of a large number of concrete bridge piers
along motorways, especially the central piers of common overpasses, suffer degrada-
tion due to chloride ingress, because this part is generally directly exposed to splash of
deicing salts and salty polluted water caused by the traffic. Moreover, a significant part
of these piers date back to more than 30 years ago, and the level of seismicity con-
sidered for their design as well as the corresponding detailing provisions do not meet
present requirements for structures that should not endanger the operation of major
corridors in case of an important seismic event. UHPFRC jacketing, possibly combined
with additional transverse reinforcement, in substitution to existing damaged regular
concrete cover, has appeared as a promising solution, which may prove technically and
economically efficient. This option can offer a better durability to the existing column,
the UHPFRC having a low porosity, and it can ensure confinement by the presence of
the fibers in the matrix.
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_100
1138 C. Sevigny-Vallières et al.
The obtained structural ductility of UHPFRC jacketed piers may, however, highly
depend on the pier-axial reinforcement ratio, on axial loading rate as compared to the
global capacity, and on detailing of the UHPFRC jacket implementation with respect to
the existing pier and footing critical cross-sections (variation of the concrete area, end
of lap bars, etc.). In addition, the structural behaviour of the jacket is affected by the
quantity and orientation of the fibers in the matrix. The influence of casting conditions
and fiber orientation on the tensile strength is of great importance for the design of the
reinforcing jacket. Indeed, fibers oriented parallel to the longitudinal axis of the column
allow a gain in the bending strength of the column and fibers oriented perpendicularly
allow a gain in confinement. For design purposes, it is also important to know the
shrinkage of the concrete in the first hours after setting. Restrained deformations, due to
the presence of the existing column, could induce an initial confinement to the section
and possibly lead to cracking of the matrix. Therefore, optimization of the combined
seismic and durability rehabilitation must be investigated, with the calibration of a
rational repair design methodology.
Development of this effort has been undertaken in a joint project within ECOMAT
international laboratory with Université de Sherbrooke (Canada) and the Materials and
Structures Department of Université Gustave Eiffel (France). As a first step, one
UHPFRC mix has been extensively characterized in order to provide the necessary
material characteristics for such a design verification process. This mix is an innovative
environmental-friendly UHPFRC mix, comprising recycled glass powder as cement
and quartz powder replacement (UHPFRC-GP). This new concrete have been devel-
oped at the University of Sherbrooke for different applications [1] and for different
replacement ratios [2]. The mechanical characterization of the mix comprises com-
pressive strength and its evolution, Young’s modulus at early age and mature state,
development of autogenous and total shrinkage, and tensile behavior identified by
flexural testing both on standard prismatic specimens as well as on thin plates repre-
sentative of the UHPFRC jacketing layer cast around the existing concrete pier. This
research project has extended the knowledge of the structural properties as well as the
early age properties of UHPFRC-GP. Indeed, in France and in Europe, different testing
standards, applying specifically to UHPFRC, are available, which is not the case in
Canada. It is therefore of great interest to combine North American and European
expertise to perform a complete characterization of UHPFRC-GP. The range of
answers of UHPFRC to these standards is also given in this study.
The paper will focus on the results of this comprehensive characterization protocol,
which will enable validating the repair design methodology based on the structural tests
to be detailed and realized in a next step of the research program.
2 Test Description
For the compression tests, the french standard NF-EN 12390-3 [3] was used. For the
evaluation of the Young’s modulus and the Poisson’s ratio in the long term, a test
protocol close to the standard NF EN 12390-13 [4] was followed, to which are added
the specifications of the standard NF P18-470 [5] and NF P18-710 [6] for the
UHPFRC. The strain measurement device J2P [7] was used to measure the Young’s
Comprehensive Characterization of UHPFRC Mixes 1139
modulus and Poisson’s ratio in a compression test. In order to determine the properties
of UHPFRC at early age, the BT JADE [8] test and the FreshCon [9, 10] test were
performed. The BTJASPE test [9, 11], developed at IFSTTAR, allows to follow the
evolution of the Young’s modulus of concrete at early age, from the first hours after the
concrete is poured. The temperature variations of the specimen are limited by a con-
stant water circulation around the mould, regulated by a thermocouple installed in the
center of the concrete mass. The dynamic Young’s modulus can be measured using the
FreshCon test [9, 10] developed at the University of Stuttgart. The velocity of the
compressional (P) and shear (S) waves is measured in order to quantify the stiffening of
the cementitious matrix in the different directions. The BT JASPE and FreshCon tests
were performed on a single sample and the J2P test was performed on 3 samples per
test period.
According to NF P18-427 [12], shrinkage must be measured at 2, 7, 28, 91 and 365
days. However, this technique does not provide information on the first hours of setting
of the concrete. The BT JADE test [8], developed at LCPC (IFSTTAR), allows a
continuous measurement of autogenous shrinkage as soon as the test specimens are
poured. This measurement is particularly important for concrete with a low water to
binder ratio, where autogenous shrinkage is particularly important and can cause
cracking when deformation is restrained [8]. Sensors are used to measure and control
the temperature with the specimen immersed in a thermostatically controlled bath. The
deformations due to autogenous shrinkage only can be isolated in this way. Thus, the
BT JADE test was carried out in conjunction with conventional shrinkage measure-
ments, allowing information on the shrinkage at different phases of concrete maturation
to be obtained. The autogenous shrinkage was measured by the BT JADE test [8],
during the first two weeks of concrete curing. Total shrinkage was measured on
prismatic specimens stored in an climate chamber at 20°C and at a relative humidity of
50%. Shrinkage measurements by the BT JADE device were performed on a single
specimen and conventional shrinkage measurements were performed on 3 specimens.
In order to characterize the tensile behaviour of UHPFRC, bending tests were
performed as recommended by NF P18-470 [5]. Compared to the direct tensile tests,
the bending tests present a simpler execution and allow a better accounting of the real
fabrication conditions [13, 15]. Indeed, the uniaxial tensile test on a “dogbone” shaped
specimen requires the use of adapted press jaws, which must be compatible with the
shape of the specimen [16]. Furthermore, the reduced width of the center of the
specimen causes a preferential orientation of the fibers due to a wall effect, which
overestimates the tensile capacity of the UHPFRC [17]. For these reasons, the bending
test was used. In order to obtain the stress-strain law, or stress-crack-opening law, a
point-by-point inverse analysis, based on the equilibrium of moments and forces of a
section in bending, was performed on each of the results obtained [5, 13–15]. In order
to obtain the identity card of the UHPFRC under study, standard NF P 18-470 [5]
suggests to perform 3-point bending tests on notched prisms to obtain the law of post-
peak tensile behaviour. The notch allows the localization of the crack and an exact
measurement of the resistance to crack opening. The 4-point bending tests give the
elastic and plastic behaviour of prisms and thin plates. The prisms tested had a cross-
section of 70 mm by 70 mm and a length of 280 mm. They were cast in inclined open
metal moulds to favour a preferential orientation of the fibers. The study of the
behaviour of the thin plates makes it possible to represent the projected application, i.e.
the rehabilitation of bridge piers with UHPFRC-GP. For the fabrication of the rein-
forcement jacket, the concrete will be poured vertically. It is therefore interesting to
study the strength of the UHPFRC-GP horizontally (providing a confinement) and
vertically (providing bending strength). Two plate orientations were studied, vertical
and horizontal, as shown in Fig. 1. The thin plates had a cross-section of 104 mm by
30 mm and a length of 600 mm. They were cast in closed moulds with a small
inclination to allow flow along the panels.
Fig. 1. Casting direction of the plates
Since the stress distribution is different in notched (3PBT) and un-notched (4PBT)
prisms, the inverse analyses performed are also different [5, 15]. As shown in Fig. 2,
the stress distribution of the notched prism is linear in the uncracked area and depends
on the elastic curvature. The stress distribution is non-linear in the cracked area and
depends on the crack opening. Thus, by computing the forces equilibrium on the
bending section and using the kinematic relationship developed by Casanova [18] to
connect the curvature of the uncracked and the cracked part, it is possible to obtain the
relationship between the equivalent tensile stress and the crack opening using the
inverse analysis [5, 19].
Fig. 2. Strain and stress distribution in a notched prism

4-point bending tests on an unnoticed prism are used to determine the yield strength
under tension and, if the tensile behaviour is strain-hardening, an inverse analysis can
be performed to obtain a post-cracking behaviour law (stress-strain). The stress varies
linearly with the elastic curvature until the yield strength is reached. The relationship
between stress and strain beyond the yield point is obtained by considering the overall
moment-curvature relationship. Thus, by computing the forces equilibrium on the
bending section, the equivalent tensile stress vs strain can be obtained by using the
inverse analysis.
3 Material
The ultra-high performance fiber-reinforced concrete mix studied in this test program
has a water-to-binder ratio of 0.18 and contains 2% metal fibers. The quantity of
cement is reduced by 25% and replaced by glass powder and 100% of the quartz
powder is replaced by glass powder. The glass powder is considered as a supple-
mentary cementitious material in Canadian standard (CSA, A3000).
4.1 Compressive Strength, Young’s Modulus, Shear Modulus

and Poisson’s Ratio
Table 1 shows the average compressive strength (fcm), Young’s modulus (Ecm) and
Poisson’s ratio (m) obtained from the J2P measuring device [7] in pure compression
tests. In comparison, the UHPFRCs commonly used in France have an average 28-day
compressive strength between 150 and 200 MPa, a Young’s modulus between 45 and
65 GPa and a Poisson’s ratio of 0.2 [5, 6]. However, these values are in the same range
at 90 days.
Table 1. Evolution of long term properties of UHPFRC-GP

Time (days) Fcm (MPa) Ecm (GPa) m
3 72 38.5 0.184
7 85 42.3 0.196
28 140 46.4 0.192
90 169 47.6 0.183
The evolution of the modulus of elasticity (E) of UHPFRC-GP, shown in Fig. 3,

was measured with FreshCon [9, 10], BT JASPE [9, 11] and J2P [7] tests. The red
curve represents the dynamic modulus of elasticity obtained for a reference UHPFRC
mix. The dynamic Young’s modulus at early age, measured with the FreshCon test,
reaches a value of 32.9 GPa at 7 days. The static Young’s modulus at early age,
measured with the BT JASPE test and the J2P test, reaches a value of 37.6 GPa and
42.3 GPa respectively at 7 days. In comparison, the UHPFRCs have a dynamic

Young’s modulus between 40 and 45 GPa (17.7% higher) and a static Young’s
modulus between 45 and 65 GPa (6 to 16% higher) [5, 6].
Fig. 3. Evolution of elastic modulus of UHPFRC-GP
The dynamic shear modulus (Gdyn) of the UHPFRC-GP, obtained in the FreshCon
test [9, 10], is 15.9 GPa, compared to a value of 20 to 25 GPa for the UHPFRCs [5, 6],
which is 20% lower in the short term. The expected value at 7 days, considering the
Young’s modulus and Poisson’s ratio obtained with the J2P test is 20.4 GPa. The
FreshCon test [9, 10] also determined that the initial setting of UHPFRC with glass
powder begins approximately 16 h after the addition of the liquids and that the stiffness
of the material stabilizes after 30 h. The initial setting is slightly delayed compared to
the reference concrete, beginning 10 h after the addition of the liquids.
4.2 Shrinkage
The autogenous shrinkage of UHPFRC-GP and the total measured shrinkage are shown
in Fig. 4. Following the initiation of concrete setting, the shrinkage increases rapidly
and reaches 350 lm/m in 20 h. The measured shrinkage stabilizes after 50 days,
reaching a maximum value of 700 lm/m. In comparison, UHPFRC commonly used in
France have a shrinkage between 550 and 800 lm/m [15].
Fig. 4. Autogenous and total shrinkage of UHPFRC-GP
4.3 Tensile Behaviour

The stress-strain relationship resulting from the inverse analysis on the 4-point bending
tests on six un-notched prisms is shown in Fig. 5 and the mean curve is shown in
Fig. 6. The post-cracking behaviour versus crack opening, derived from the inverse
analysis of the 3-point bending tests on notched prisms, is shown in Fig. 7 and the
mean distribution is shown in Fig. 8. Finally, thin plates were also tested in 4-point
bending tests to obtain the stress-strain response. The results of the horizontal thin
plates are shown in Fig. 9 and the mean curve is shown in Fig. 10. The results of the
vertical thin plates are shown in Fig. 11 and the mean curve is shown in Fig. 12. Those
stress-strain relationships were obtained between 29 days and 39 days. It should be
noted that in Fig. 9 and Fig. 11, the H5, V1, V2 and V3 plates showed a crack location
outside the instrumented area. The average strength obtained for UHPFRCs commonly
used in France is between 7 and 12 MPa [6, 20].
Fig. 5. Stress-strain relationship obtained Fig. 6. Average stress-strain relationship

by inverse analysis (4PBT) obtained by inverse analysis (4PBT)
Fig. 7. Post-pic behaviour obtained by Fig. 8. Average post-pic behaviour obtained

inverse analysis (3PBT) by inverse analysis (3PBT)
Fig. 9. Stress-strain relationship obtained by Fig. 10. Average stress-strain relation-

inverse analysis (4PBT-Horizontal thin plates) ship obtained by inverse analysis
(4PBT-Horizontal thin plates)
Fig. 11. Stress-strain relationship obtained Fig. 12. Average stress-strain relationship
by inverse analysis (4PBT-Vertical thin obtained by inverse analysis (4PBT-
plates) Vertical thin plates)
The waves observed in the curves resulting from inverse analysis come from the
successive cracks opening which create successive decrease and increase in the tensile
stress. It is also resulting from the inverse analysis which tends to amplify these waves
(inverse analysis is a kind of derivation of raw curves). These waves should not be
considered to determine design curves.
5 Discussion
The 28-day compressive strength of UHPFRC-GP is 6% lower than that of the ref-
erence UHPFRCs. However, at 90 days, the strength continues to increase and is in the
same range as the common UHPFRC [6]. Young’s modulus is 5 to 20% lower than the
short-term (7 days) reference values. In the long term, the modulus is in the same range
as the reference values. It can thus be concluded that the resistance gain occurs later. It
is well known that cement replacement by pozzolanic material such as glass powder
results in delay in hydration which reduce the mechanical performances at early age but
not at long term (56 days and more). Indeed, the strength of UHPFRC-GP increases
considerably between 28 and 90 days.
The shrinkage measured for UHPFRC-GP is in the same range of values as the
UHPFRC.
The tensile strength measured on prisms and thin plates is in the same range of
values as the UHPFRC. In Figs. 5 to 8, it can be observed that the UHPFRC-GP prisms
tested have a slightly strain-hardening behaviour. Figure 13, adapted from Mousa et al.
[21], shows the idealized behaviour expected for a strain-hardening UHPFRC. The
curves used for the design need to be smoothed, as presented in Fig. 14. In this figure,
the different phases of the constitutive law are easily identifiable, i.e. the linear zone,
corresponding to the elastic behaviour, and the constant zone, corresponding to the fine
multi-cracking. The increase in stress corresponds to the opening of the cracks. The
maximum stress reached corresponds to the localization of a crack and the pull-out of
the fibers [13, 14]. Figure 7 and Fig. 8 show the last phases of the constitutive law as a
function of crack opening. Figures 9 to 12 show that the tensile strength of horizontal
plates is higher than that of vertical plates and that a softening behaviour is observed,
due to the rapid localization of the crack. Thus, it can be concluded that the fibers have
an orientation following the flow of the fluid to the base of the mould, a preferential
horizontal orientation being obtained, as shown in Fig. 15. For the projected applica-
tion, a preferential orientation of the fibers perpendicular to the column axis is therefore
to be expected, which is favourable for a column confinement application. Neverthe-
less, a post-cracking tensile stress higher in the case of thin plates would have been
expected compared to prisms. This is not the case here. The way of introducing the
material into mould would certainly need to be improved.
Fig. 13. Idealized expected tensile behav- Fig. 14. Data smoothing of 4PBT on
ior of UHPFRC prism for design
Fig. 15. Direction of flow and fiber orientation
6 Conclusion
Based on the experimental results presented above, the following conclusions can be
drawn:
• Replacing a proportion of the cement and all of the quartz powder with glass
powder makes it possible to obtain properties that are comparable (compressive and
tensile strength, Young’s modulus, shear modulus, etc.) to those of the UHPFRCs
commonly used in France, with better environmental impact
• The gain in strength of UHPFRC-GP is delayed compared to the UHPFRCs
commonly used in France at 28 days, with comparable strength being reached at 90
days.
• The prisms tested in bending showed a hardening behaviour, whereas the thin plates
rather showed a softening behaviour due to the rapid localization of the crack.
• For the planned application, a preferential orientation of the fibers perpendicular to
the column axis is therefore to be expected.
The durability tests are in progress.
Acknowledgement. This research project was financially supported by Chaire de recherche

SAQ, Université de Sherbrooke (grant), LIA-Écomat and Université Gustave Eiffel. The authors
would like to express their gratitude to Aghiles Begriche, PhD student at Université de Sher-
brooke and to the technical personnel of Université Gustave Eiffel for their help and support.
References
1. Tagnit-Hamou, A., Soliman, N., Omran, A.: «Green Ultra-High Performance Glass
Concrete». In: Proceedings of the First International Interactive Symposium on UHPC,
Des Moines, Iowa, USA, (2016)https://doi.org/10.21838/uhpc.2016.35
2. Soliman, N.A., Tagnit-Hamou, A.: Development of ultra-high-performance concrete using
glass powder – towards ecofriendly concrete. Constr. Build. Mater. 125, 600–612 (2016).
https://doi.org/10.1016/j.conbuildmat.2016.08.073
3. AFNOR, «NF EN 12390–3 - Essais pour le béton durci- Partie 3- Résistance à la
compression des éprouvettes» (2019)
4. AFNOR, «NF EN 12390–13- Essais pour le béton durci- Partie 13: Détermination du
module sécant d’élasticitié et du module de compression» (2014)
5. AFNOR, «NF P18–470- Bétons fibrés à Ultra Hautes Performances- Spécifications,
performance, production et conformité» (2016)
6. AFNOR, «NF P18–710-Complément national à l’Eurocode 2- Calcul des structures en
béton: règles spécifiques pour les Bétons Fibrés à Ultra-Hautes Performances
(BFUP)» (2016)
7. Boulay, C., et al.: «Un extensomètre à béton éliminant l’influence des déformations
transversales sur la mesure des déformations longitudinales». Matér. Constr. 14, 35–38
(1981)
8. Boulay, C.: «Développement d’un dispositif de mesure du retrait endogène d’un béton au
jeune âge», présenté à Huitième édition des journées scientifiques du Regroupement
francophone pour la recherche et la formation sur le béton (RF)2B, Montréal, Canada, 2007,
[En ligne]. Disponible sur: https://scholar.google.com/scholar?hl=fr&as_sdt=0%2C5&q=
Boulay+C.+2007.+D%C3%A9veloppement+d%27un+dispositif+de+mesure+du+retrait+
endog%C3%A8ne+d%27un+b%C3%A9ton+au+jeune+%C3%A2ge%2C+Huiti%C3%
A8me+%C3%A9dition+des+Journ%C3%A9es+scientifiques+du+Regroupement+
francophone+pour+la+recherche+et+la+formation+sur+le+b%C3%A9ton%2C+Montr%
C3%A9al%2C+Canada.&btnG=
9. Boulay, C., et al.: «Monitoring elastic properties of concrete since very early age by means
of cyclic loadings, ultrasonic measurements, natural resonant frequency of composite beam
(emm-arm) and with smart aggregates» (2013)
10. Carette, J., Staquet, S.: Monitoring the setting process of mortars by ultrasonic P and S-wave
transmission velocity measurement. Constr. Build. Mater. 94, 196–208 (2015). https://doi.
org/10.1016/j.conbuildmat.2015.06.054
11. Boulay, C., et al.: How to monitor the modulus of elasticity of concrete, automatically since
the earliest age? Mater. Struct. 47(1–2), 141–155 (2014). https://doi.org/10.1617/s11527-
013-0051-3
12. AFNOR, «NF P18–427- Détermination des variations dimensionnelles entre deux faces
opposées d’éprouvettes de béton durci» (1996)
characterization: inverse analysis of four-point bending test results. Mater. Struct. 46(8),
1337 (2013)
14. Baby, F., Graybeal, B., Marchand, P., Toutlemonde, F.: Proposed flexural test method and
associated inverse analysis for ultra-high-performance fiber-reinforced concrete. ACI Mater.
J. 109(5), 545–555 (2012)
15. Resplendino, J., Marchand, P.: Bétons fibrés à ultra-hautes performances. Recommanda-
tions/ Ultra-high performance fiber-reinforced concretes. Recommendations. AFGC, Paris,
France, Revised edition (2013). https://doi.org/10.1002/9781118557839.ch47
16. di Prisco, M., Ferrara, L., Lamperti, M.: Double edge wedge splitting (DEWS): an indirect
tension test to identify post-cracking behaviour of fibre reinforced cementitious composites.
Mater. Struct. 46(11), 1893 (2013)
17. Delsol, S., Charron, J.-P.: Numerical modeling of UHPFRC mechanical behavior based on
fibre orientation. In: Proceedings of the RILEM-fib-AFGC International Symposium on
Ultra-High Performance Fibre-Reinforced Concrete, UHPFRC. p. 10 (2013)
18. Casanova, P.: «Bétons renforcés de fibres métalliques: du matériau à la structure. Etude
expérimentale et analyse du comportement de poutres soumises à la flexion et à l’effort
tranchant», École nationale des ponts et chaussées, Paris, France (1996)
19. Marchand, P.: «Caractérisation et utilisation des BFUP pour des applications structurelles» ,
PhD Thesis, Université Paris-Est, (2019)
20. Herrera, A.: «Fonctionnement des jonctions âmes-membrures en Béton Fibrés à Ultra-
Hautes Performances (BFUP)», PhDThesis, Université Paris-Est (2017)
21. Mousa, M., Cuenca, E., Ferrara, L., Roy, N., Tagnit-Hamou, A.: Tensile characterization of
an “eco-friendly” uhpfrc with waste glass powder and glass sand. Strain-Hardening Cem.-
Based Compos. 15, 238–248 (2018). https://doi.org/10.1007/978-94-024-1194-2_28
Evaluation of the Splitting Tensile Strength
of Ultra-High Performance Concrete
An Hoang Le(&)
NTT Hi-Tech Institute, Nguyen Tat Thanh University, Ho Chi Minh City,
Vietnam
lhan@ntt.edu.vn
Abstract. The splitting tensile strength of ultra-high performance concrete

(UHPC) is much larger than that of normal concrete. It was found that the
studies on UHPC has mainly focused on the direct tensile strength or flexural
strength, while there has been insufficient work to evaluate the splitting strength
characteristics of UHPC. Therefore, this study is aimed at presenting experi-
mental and statistical evaluation of the splitting tensile strength of UHPC. The
splitting tests were conducted on cylindrical specimens of 100 200 mm size.
UHPC was designed to achieve a nominal compressive strength of 200 MPa at
the age of 28 days. Macro steel fibers were used to reinforce the UHPC by
volumetric percentages of 0, 1, and 2%. The effect of fiber volume on the
splitting tensile strength was investigated by the test results. The values of the
splitting tensile strength of UHPC with and without fibers in some previous
studies were collected together with this study and subsequently verified with
the predictions of the splitting tensile strength obtained from the existing
models. The appropriateness of these existing models was clarified. Finally,
based on the regression analysis on the collected test results, a simplified
equation was proposed to estimate the splitting tensile strength of UHPC having
compressive strength varying between 120 and 200 MPa.
Keywords: Ultra-high performance concrete Splitting tensile strength

Fibers Compressive strength Splitting tests
1 Introduction
Ultra-high performance concrete (UHPC) is considered as a composite material that

comprises fine powder such as Portland cement, silica fume, fine sand. In addition to a
very low water-to-binder ratio, a super plasticizer is required to be used in the UHPC
mixture to ensure the very high workability [2, 17]. Due to a very dense matrix, UHPC
is characterized by a very high compressive strength exceeding 150 MPa, possibly
attaining 250 MPa and an outstanding durability [2–4]. The inclusion of internal fiber
reinforcement ensures the ductile behavior of UHPC under tension and significantly
increases the tensile, flexural and shear strength under different types of structural
actions. The superior properties of UHPC can generate many structural advantages,
thereby leading to a great attention to many researchers throughout the world. Along
with the compressive strength, the tensile strength is one of the most important aspects
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5_101
1150 A. H. Le
in the design of structures made of UHPC, especially when UHPC reinforced by fibers
(UHPFRC) is applied.
Generally, there are three test methods to determine the tensile strength of concrete
either directly or indirectly: (1) Direct tensile strength test; (2) Splitting strength test;
(3) Flexural tensile strength test [7]. The direct tensile test (DTT) has some impedi-
ments with its performance such as the slippage or the alignment between gripping
apparatus and concrete specimens, the stress concentration at the gripping devices [15].
The results from the flexural tensile test (FTT) are usually sensitive to the geometry and
size of concrete specimen, loading application and strain rate. Both DTT and FTT give
the tensile behavior of concrete, however, these two test methods require special test
equipment and comparatively high cost to conduct [7]. On the other hand, the splitting
test (ST), which is an indirect tensile strength test of concrete, was commonly adopted
by many standards (ASTM C496, AS 1012.10 – Standards Australia 2000) and easier
to carry out than DTT and FTT [7, 15, 16].
In terms of UHPC and UHPFRC, most previous studies have focused on DTT and
FTT in order to comprehensively provide the effect of fibers on the pre-peak and post-
peak response of tensile stress - strain behavior. Few studies have been conducted in
the literature to investigate the splitting strength of UHPC and UHPFRC. Graybeal
(2006) [17] suggested some modifications of the splitting test in ASTM C496 to
capture the post-cracking behaviors of UHPC cylinders. El-Helou et al. (2014) [13]
reported the splitting test results of UHPC cylinders using 0%, 2% and 4% steel fibers,
and compared with the direct tensile test. El-Din et al. (2016) [12] investigated the
influence of fiber volume and aspect ratio on the splitting strength of UHPC and
developed an equation to predict the splitting strength of UHPC. Shafieifar et al.
(2017) [23] compared the splitting strength between UHPC and normal strength con-
crete (NSC) cylinders. Bae et al. (2017) [7] studied the indirect tensile strength tests
including splitting test and flexural test of ultra high strength concrete (UHSC) with and
without steel fibers to show the effect of compressive strength and fiber content on the
splitting strength. The authors also proposed equation for calculating the splitting
strength considering the concrete compressive strength and fiber content simultane-
ously. Goaiz et al. (2018) [15] presented a comparison among the splitting test, the
double punch test and the direct tensile test of notched prisms for determining the
tensile strength of reactive powder concrete (RPC – a type of UHPC). Guler et al.
(2018) [26] studied on splitting strength of concrete using steel, synthetic and hybrid
fibers. Moreover, these authors evaluated the existing strength models for predicting
the splitting tensile strength.
It was found that the current information in the literature about the splitting tensile
strength of UHPC and UHPFRC is limited. Therefore, the main aim of this paper is to
examine the splitting tensile strength of UHPC and UHPFRC by splitting tests on
concrete cylinders of 100 200 mm. The effect of steel fiber volumes of 0%, 1% and
2% on the splitting tensile strength was investigated. Based on the database of splitting
test results obtained from this study and previous studies, this paper provide a statistical
evaluation on some existing models for estimating the splitting strength of UHPC and
UHPFRC. Finally, a simplified equation was derived from the regression analysis to
predict the splitting tensile strength of UHPC and UHPFRC having compressive
strength varying between 120 and 200 MPa.
Evaluation of the Splitting Tensile Strength 1151
2 General Specifications
A total of nine UHPC and UHPFRC mixtures were prepared for this study. The recipe
of M3Q, which was developed at University of Kassel - Germany, was adopted for
producing UHPC and UHPFRC [2]. The details of the mix proportions are given in
Table 1. It should be noted that the M3Q mix was designed to provide a very high self-
compacting characteristic and a compressive strength of concrete cylinder at about
200 MPa. Therefore, the necessity for compacting of concrete using external vibration
was eliminated. This is favorable for the preparation of test specimens. The steel fibers
(smooth and brass coated surface) with a diameter df of 0.175 mm and a length lf of
13 mm were added to the UHPC mix in volume fraction of 0%, 1% and 2% (UHPC,
UHPFRC-SF1% and UHPFRC-SF2%). The mechanical properties of steel fibers are
illustrated in Table 2.
A mixer (Zyklos Gleichlaufmischer ZZ 150 HE) with maximum capacity of 170 L
was used for mixing each concrete batch. The sequence for mixing UHPC and
UHPFRC is presented in Table 2. When UHPC and UHPFRC mixtures were ready, for
each batch of concrete, they were poured into six cylindrical moulds having a diameter
of 100 mm and a height of 200 mm without any vibrations. For six concrete cylinders
of each concrete batch, three cylinders were used for uniaxial compression test and
three remaining cylinders were used for splitting tests. A total of fifty-four concrete
cylinders were fabricated for testing. All test specimens were covered by plastic sheets
immediately after casting in order to prevent moisture loss. The test specimens were
demolded 48 h after casting and cured at ambient temperature until testing day. It
should be noted that all concrete cylinders were cast and tested for 28-day compressive
strength and splitting tensile strength. Figure 1 shows the slump flow of UHPFRC-
SF2% and the cylinders for casting and concrete cylinders after demolding.
a) Slump flow b) Cylinders for casting c) Concrete cylinders of

(UHPFRC-SF2%) concrere 100x200 mm for testing
Fig. 1. Specimen preparation for splitting tests

1152 A. H. Le
Table 1. Composition of UHPC and UHPFRC mixes

Mix composition Unit UHPC UHPFRC-SF1% UHPFRC-SF2%
Water kg/m3 188.0 186.0 184.0
CEM I 52.5R HS-NA kg/m3 795.4 787.5 779.5
Silica fume kg/m3 168.6 166.9 165.2
Super plasticizer Sika Viscorete 2810 kg/m3 24.1 23.7 23.6
Ground Quartz W12 kg/m3 198.4 196.4 194.4
Quartz sand 0.125/0.5 kg/m3 971.0 961.3 951.6
Steel fibers kg/m3 – 78.5 157.0
2.2 Testing Procedure

The flowability of the fresh concrete for each mix was checked immediately after
mixing, using a mini-slump cone on a flow table in accordance with DIN EN 12350-8:
2010-12 [9]. The slump-flow test includes two governing parameters: the slump flow
and the flow time t500. The minimum value of slump flow is higher than 800 mm,
indicating that the concrete mixtures in this study were fully self-compacting even with
the use of steel fibers 2% by volume (see Fig. 1a). The compressive strengths fc and
elastic modulus Ec were determined from the compression tests on 3 cylindrical
specimens of 100 200 mm for each batch of concrete in accordance with DIN EN
12390-3:2009-07 [10] and DIN 1045-1 [8], respectively. Prior to the compression tests,
the two ends of each concrete cylinder were ground using a grinding wheel so that the
two ends were parallel and the load was transferred uniformly to the cross section. All
cylinders were tested uniformly force-controlled using a 4000 kN capacity compression
machine. In addition, elastic modulus was measured using a compress meter installed at
the mid-height of concrete cylinder measuring the average compressive strain.
Splitting tensile strength tests were executed on three concrete cylinders of 100
200 mm for each batch of concrete according to DIN EN 12390-6:2009 [11] to
determine the average splitting strength, as shown in Figure. The splitting test was
performed by applying compressive line loads along two opposing lengths on the side
of the concrete cylinder (see Fig. 2a). Two timber strips having the dimensions of
300 mm in length, 10 mm in width, and 5 mm in thickness were located between the
loading plates and the specimen surfaces along the full length of the specimen as
bearing strips (see Fig. 2a). A conventional compression testing machine at a loading
rate of 0.01 mm/s was employed for the splitting test. The splitting tensile strength was
calculated using the following equation:
2P
fspl ¼ ð1Þ
pLD
where fspl is the splitting strength (MPa), P is the maximum applied load (kN), L is the
length of the concrete cylinder (mm), and D is the diameter of the concrete cylinder
(mm). Figure 2 shows the splitting test setup.
Table 2. UHPC and UHPFRC mixing plan

Process Time
Material filling –
Dry mixing (2 min) 0 min – 2 min
Water and superplasticizer (3 min) 2 min – 5 min
Break (2 min) 5 min – 7 min
Mixing 7 min – 15 min
Steel fiber filling 13 min
Break (10 min) 15 min – 25 min
Final mixing 25 min – 25.5 min
UHPC
UHPFRC-SF1%
a) Splitting test setup
UHPFRC-SF2%
b) Failure modes
Fig. 2. Splitting test on concrete cylinders 100 200 mm and failure modes
3 Test Results and Discussions
Table 3 presents the average compressive strength (fc) and the average elastic modulus
(Ec) derived from the compression tests on three concrete cylinders for each concrete
cast. Also, the average splitting tensile strength fspl measured from the splitting test on
three concrete cylinders for each concrete cast is tabulated in Table 3.
1154 A. H. Le
Table 3. Results of compression tests and splitting tests

Series Mix Steel fiber volume Average Average elastic Average splitting
Vf (%) compressive strength modulus Ec (MPa) tensile strength fspl
fc (MPa) (MPa)
1 Cast 1 0 178.9 48370 3.4
Cast 2 1 195.5 49645 11.2
Cast 3 2 188.2 48421 19.5
2 Cast 4 0 198.0 46937 4.4
Cast 5 1 195.5 47881 14.2
Cast 6 2 187.8 48580 19.5
3 Cast 7 0 190.4 46186 5.0
Cast 8 1 195.6 48689 10.8
Cast 9 2 192.4 48557 14.0
a) The average splitting tensile strength b) The splitting tensile strength of all concrete
corresponding with fiber volume batches corresponding with fiber volume
Fig. 3. The effect of steel fiber volume on the splitting tensile strength
The typical failure modes of concrete cylinders in splitting tests were observed after
testing and shown in Fig. 2(b). The concrete cylinders without steel fibers (UHPC)
experienced one failure surface along the line of the loading strip. The concrete
specimens without steel fibers failed in brittle manner with a sudden failure when all
specimens were totally spitted into two parts. However, the concrete cylinders with
steel fibers (UHPFRC-SF1% and UHPFRC-SF2%) showed a ductile failure mode. All
specimens using steel fibers remained intact with a small crack which appeared along
the middle line of the concrete section. The crack of specimens using higher volume of
steel fibers (2%) was observed to be shorter than that of specimens using lower volume
of steel fibers (1%). The ductile failure mode is attributed to the influence of steel
fibers. Steel fibers provide the distributed stresses along the failure surface, thus pre-
venting the complete splitting failure.
Table 3 shows that presence of steel fibers significantly increased the splitting
tensile strength as compared to that in the case of no fibers. The higher volume of steel
fibers also resulted in higher splitting tensile strength. In this study, UHPC had the
splitting tensile strength varying between 3.4 and 5.0 MPa, while the spitting tensile
strength was found to be between 10.8 and 14.2 MPa for UHPFRC-SF1%, and
between 14.0 and 19.5 MPa for UHPFRC-SF2%.
Figure 3 shows the effect of steel fiber volume on the splitting tensile strength. The
average splitting tensile strength of UHPFRC-SF1% and UHPFRC-SF2% were
increased by 205.74% and 348.20% as compared to UHPC. Similarly, the average
splitting tensile strength of UHPFRC-SF2% was increased by 46.59% as compared to
UHPFRC-SF1%. The highest splitting tensile strength was achieved by UHPFRC-SF2%
(cast 3 and cast 6) which had the highest steel fiber volume of 2%. This observation
indicates that the splitting tensile strength increases significantly with increasing steel
fiber volume from 0% to 1% and from 0% to 2%, while the increase rate in the splitting
tensile strength is lesser with steel fiber volume between 1% and 2%.
It was observed from Fig. 3(b) that there was a large scatter of the splitting strength
from batch to batch for each fiber volume. This is due to the different distribution and
dispersion of fibers when casting specimens in vertical direction without vibration.
4 Evaluation of Existing Models
Most of the international design codes introduce equations based on the compressive
strength (fc) to predict the splitting tensile strength (fspl). However, these codes do not
take into account the effect of fibers. Previous researchers attempted to propose models
considering the effect of fibers including fiber volume (Vf), aspect ratio (lf/df), fiber type
to estimate the splitting tensile strength.
Table 4. Equations of existing models

Authors Equations Explanations
Wafa and Ashour (1992) [29] fspl ¼ 0:58ðfc Þ0:5 þ 3:02Vf2 fc 100 MPa
Song and Hwang (2004) [27] fspl ¼ 0:63ðfc Þ0:5 þ 3:01Vf 0:02Vf2 fc 100 MPa

Musmar (2013) [31] l pffiffiffiffi 4 MPa fc 120 MPa
fspl ¼ 0:6 þ 0:4Vf dff fc
Thomas and Ramaswamy (2007) [28] fspl ¼ 0:63ðfc Þ0:5 þ 0:288Vf lf f 0:5 0:02V 2 fc 85 MPa
df c f
h pffiffiffiffi i
El-Din et al. (2016) [12] fspl ¼ 0:076 fc þ 10 F 3 fc 20 F ¼ dlf Vf bf , where bf is the bond
f
factor depending on the type of

fibers
125 MPa fc 155 MPa
qffiffiffiffiffiffiffiffiffi
Narayanan and Darwish (1987) [19] fspl ¼ q
fc ffiffiffiffiffiffi
þ 0:7 þ Vf dff
l N/A
l
20 Vf df
f
Arioglu et al. (2006) [5] fspl ¼ 0:321fc0:66 4 MPa fc 120 MPa

ACI 318-14 (ACI, 2014) [1] fspl ¼ 0:321fc0:66 21 MPa fc 69 MPa
SNZ 3101 (SNZ, 2006) [25] fspl ¼ 0:36fc0:5 25 MPa fc 100 MPa

fspl ¼ 2:12ln 1 þ fc 67 MPa
10 =0:9
DIN 1045-1 (DIN, 2008) [8] fc

Mode Code 2010 (MC2010, fspl ¼ 2:12ln 1 þ fc 50 MPa fc 120 MPa
10
FIB 2013) [14]
JSCE (JSCE, 2007) [18] fspl ¼ 0:23fc2=3 20 MPa fc 80 MPa
1156 A. H. Le
It was found that the compressive strength in all existing models and design codes is
limited up to around 120 MPa and usually applicable to normal or high strength con-
crete (NSC and HSC). Therefore, there is a need to evaluate the suitability of the existing
models and design codes for UHPC and UHPFRC. In this study, UHPC and UHPFRC
are defined as concrete having compressive strength of cylinders (100 200 mm)
higher than 120 MPa. Table 4 shows the equations of seven existing models and five
codes collected in the literature, which are used for the evaluation.
The results of splitting tests on the 165 cylinders of 100 200 mm obtained from
previous works and this study were used. Table 5 summarizes the content of database
of collected splitting tests. The compressive strengths were in the range 124.0-
216.52 MPa, and the steel fiber volumes varied between 0 and 6%. Table 6 shows the
results of statistical analysis of the predictions from existing models and design codes.
The statistical measures include mean, standard deviation (SD), coefficient of variation
(COV), and integral absolute error (IAE), which were calculated by the ratio of pre-
dicted splitting tensile strength (fspl,pre) to test result (fspl,test). It should be noted that the
value of IAE is determined using the following equation in Bae et al. (2017):
2
0:5
X fspl;testi fspl;prei
IAE ¼ P ð2Þ
fspl;prei 100
It was revealed from Table 6 that the models proposed by Wafa and Ashour (1992),
Thomas and Ramaswamy (2007), Narayanan and Darwish (1987), Arioglu et al.
(2006) underestimated the splitting tensile strength, however the mean value of fspl,pre/
fspl,test obtained from each model was close to unity. Furthermore, all design codes gave
an remarkable underestimation with low mean value of fspl,pre/ fspl,test. The models
suggested by Musmar (2013), and Song and Hwang (2004) slightly overestimated the
splitting tensile strength, while the model of El-Din et al. (2016) gave an overesti-
mation with high mean value of 1.19. In terms of statistical variables, all models and
Table 5. Collected database of previous splitting tests

Authors Fc (MPa) Fspl (MPa) Vf (%) Lf/df (mm/mm) Steel fiber Number of
type specimens
Wang et al. 129–185 7.4–17.1 0–3 13/0.2 Straight 12
(2018) [30]
Shafieifar et al. 138 20.7 2 12.5/0.2 Straight 3
(2017) [23]
Pansuk et al. 136.81–147.69 10.38–16.27 0–1.6 13/0.2 Straight 18
(2017) [22]
Nehdi et al. 151–173 9.4–39.8 1–6 8/0.2;12/0.2; 16/0.2 Straight 30
(2015) [20]
El-Din et al. 124–154 11.5–19.1 0–3 30/1; 50/1 Hooked 21
(2016) [12] ended
Othman (2016) [21] 151.9–174.1 6.8–15.3 1–3 13//0.2 Straight 27
Shin (2017) [24] 184.8 12.9 2 13/0.2 Straight 3
Bae et al. (2017) [7] 149.4–216.52 9.01–11.96 0.5–2 13/0.2 Straight 24
Current study 178.9–198.0 3.38–19.53 0–2 13/0.2 Straight 27
Table 6. Results of statistical analysis

Authors Mean Standard deviation Coefficient of variation IAE
(fspl,pre/fspl,test) (SD) (COV)
Wafa and Ashour (1992) [29] 0.98 0.41 0.43 16.25
Song and Hwang (2004) [27] 1.02 0.44 0.43 15.97
Musmar (2013) [31] 1.01 0.44 0.44 17.38
Thomas and Ramaswamy 0.84 0.45 0.54 21.02
(2007) [28]
El-Din et al. (2016) [12] 1.19 0.70 0.59 23.00
Narayanan and Darwish 0.87 0.49 0.57 20.48
(1987) [19]
Arioglu et al. (2006) [5] 0.81 0.50 0.62 21.88
ACI 318-14 (ACI, 2014) [1] 0.62 0.38 0.60 26.65
SNZ 3101 (SNZ, 2006) [25] 0.40 0.24 0.60 34.57
DIN 1045-1 (DIN, 2008) [8] 0.58 0.34 0.59 27.78
Mode Code 2010 (MC2010, FIB 0.52 0.31 0.59 29.90
2013) [14]
JSCE (JSCE, 2007) [18] 0.60 0.37 0.62 27.61
design codes exhibited relatively high values of SD and COV, especially in the case of
design codes. It is mentioned that the design codes do not consider the effect of fibers
on the splitting tensile strength, thereby resulting in the low accuracy of the prediction
as compared with the test results. Likewise, in terms of IAE, an increasing tendency in
IAE appeared with design codes, while the predictions from existing models had lower
values of IAE. Among existing models, the model of Wafa and Ashour (1992) was the
most suitable approach to predict the splitting tensile strength because this models gave
a safe prediction with low values of SD, COV and IAE as compared with the remaining
models. It is also recommended that the effect of fibers should be considered in the
design codes to obtain more accurate prediction of the splitting strength for UHPFRC.
5 Proposed Equation for Predicting the Splitting Tensile

Strength of UHPC and UHPFRC
The collected database of splitting tests in Table 5 was used for the regression analysis.
The approach of El-Din et al. (2016) [12] was adopted for the proposed equation of the
splitting tensile strength. In El-Din et al. (2016) [12], the splitting tensile strength is a
function of the compressive strength fc and a variable (F) which can reflect the effect of
steel fiber comprehensively. This variable (F) is given as follows:
lf
F ¼ Vf bf ð3Þ
df
where Vf is the fiber volume, lf/df is the fiber aspect ratio, and bf is the bond factor
(bf = 0.5 for fibers having circular section, bf = 0.75 for hooked or crimped fibers).
1158 A. H. Le
Fig. 4. Regression analysis on the collected database
ffi
Deriving from the regression analysis of the relation between the ratio fspl pffiffiffi
fc and F
in Fig. 4, the equation for predicting the splitting tensile strength of UHPC and
UHPFRC can be proposed as below:

lf pffiffiffiffi
fspl ¼ 0:94Vf bf þ 0:67 fc ð4Þ
df
It is mentioned that the above proposed equation is applicable for UHPC and
UHPFRC having compressive strength from 120 to 200 MPa
6 Conclusions
Some conclusions and recommendations can be drawn from this study as follows:
• In this experimental tests, UHPC had average splitting tensile strength between 3.4
and 5.0 MPa, while UHPFRC had average splitting tensile strength varying
between 10.8 and 14.2 for 1% steel fibers by volume and between 14.0 and
19.5 MPa for 2% steel fibers by volume;
• The inclusion of steel fibers significantly increases the splitting tensile strength.
Higher volume of steel fibers results in higher splitting tensile strength;
• The failure mode of concrete cylinders under splitting load becomes more ductile
with the use of steel fibers;
• All selected design codes underestimated the splitting tensile strength with low
accuracy. The effect of steel fibers should be considered in these codes;
• Among existing models, the model of Wafa and Ashour (1992) was found to be the
best choice for predicting the splitting tensile strength;
• A simplified equation for estimating the splitting tensile strength of UHPC and
UHPFRC having compressive strength between 120 and 200 MPa was proposed
with considering the effect of steel fibers. More test results of splitting strengths
should be collected get better regression analysis.
• Further studies on splitting tensile strength of UHPC and UHPFRC using various
fiber types, fiber aspect ratios, fiber volumes should be conducted.
Acknowledgements. This research is funded by Vietnam National Foundation for Science and
Technology Development (NAFOSTED) under grant number 107.01-2019.325.
References
1. ACI, Building code requirements for structural concrete (ACI 318–14) and commentary. In:
ACI, 318-14 Farmington Hills, MI, USA (2014)
2. An, L.H., Fehling, E.: Influence of steel fiber content and type on the uniaxial tensile and
compressive behavior of UHPC. Constr. Build. Mater. 153, 790–806 (2017)
3. An, L.H., Fehling, E., Binglin, L., Kien, T.D., Chau, V.N.: Experimental study on structural
performance of UHPC and UHPFRC columns confined with steel tube. Eng. Struct. 187,
457–477 (2019)
4. An, L.H., Fehling, E., Kien, T.D., Chau, V.N.: Evaluation of axial strength in circular STCC
columns using UHPC and UHPFRC. J. Constr. Steel Res. 153, 533–549 (2019)
5. Arioglu, N., Girgin, Z.C., Arioglu, E.: Evaluation of ratio between splitting tensile strength
and compressive strength for concretes up to 120 MPa and its application in strength
criterion. ACI Mater. J. 103(1), 18–24 (2006)
6. ASTM, C.: 496–90 Test method for splitting tensile strength of cylindrical concrete
specimens. In: Annual Book of ASTM Standards ASTM International, West Conshohocken
(1990)
7. Bae, B.L., Lee, M.S., Choi, H.K., Choi, C.S.: Indirect tensile strength of UHSC reinforced
with steel fibres and its correlation with compressive strength. Mag. Concrete Res. 69(15),
772–786 (2017)
8. DIN (Deutsches Institut für Normung) DIN 1045–1, Plain, Reinforced and Prestressed
Concrete Structures – Part 1: Design and Construction. Beuth Verlag, Berlin, Germany
(2008)
9. DIN EN 12350–8:2010-12, Testing fresh concrete-Part 8: Self-compacting concrete-Slump-
flow test. German version EN 12350-8, Beuth Verlag, Berlin (2010)
10. DIN EN 12390–3:2009-7, Testing hardened concrete-Part 3: Compressive strength of test
specimens, German version EN 12390-3, Beuth Verlag, Berlin (2009)
11. DIN EN 12390–6, Testing Hardened concrete – Part 6: Tensile splitting strength of test
specimens. Beuth Verlag, Berlin (2009)
12. El-Din, H.K.S., Heba, A.M., Khater, M.A.E., Sayed, A.: Effect of steel fibers on behavior of
ultra high performance concrete. In: First International Interactive Symposium on UHPC,
USA (2016)
13. El-Helou, R.G., Moen, C.D., Cusatis G.: Ultra-High performance fiber-reinforced concrete:
extensive material characterization, model validation, and structural simulations. Presenta-
tion at ACI Fall 2014 Convention Washington, DC (2014)
14. Fib (Fédération Internationale du Béton), Model Code for Concrete Structures 2010. Fib.
15. Goaiz, H.A., Farhan, N.A., Sheikh, M.N., Yu, T., Hadi, M.N.S.: Experimental evaluation of
tensile strength test methods for steel fibre reinforced concrete. Mag. Concrete Res. 71(8), 1–
42 (2018)
16. Goaiz, H.A., Tao, Y., Hadi, M.S.N.: Quality evaluation tests for tensile strength of reactive
powder concrete. J. Mater. Civ. Eng. 30(5), 04018070 (2018)
17. Graybeal, B.A.: Practical means for determination of the tensile behavior of ultra-high
performance concrete. J. ASTM Int. 3(8), 1–9 (2006)
1160 A. H. Le
18. JSCE (Japan Society of Civil Engineers), Standard Specifications for Concrete Structures –
Design, JSCE Guidelines for Concrete No. 15 Tokyo, Japan, (2007)
19. Narayanan, R., Darwish, I.Y.S.: Use of steel fibers as shear reinforcement. ACI Struct. J.
84(3), 216–227 (1987)
20. Nehdi, M.L., Abbas, S., Soliman, A.M.: Exploratory study of ultra-high performance fiber
reinforced concrete tunnel lining segments with varying steel fiber lengths and dosages. Eng.
Struct. 101, 733–742 (2015)
21. Othman, H.A.B.: Performance of Ultra-High Performance Fibre Reinforced Concrete Plates
under Impact Loads. PhD dissertation (Ryerson University, 2016)
22. Pansuk, W., Thuc, N.N., Yasushiko, S., Uijl, J.A.D., Walraven, J.C.: Shear capacity of high
performance fiber reinforced concrete I-beams. Constr. Build. Mater. 157, 182–193 (2017)
23. Shafieifar, M., Farzad, M., Azizinamini, A.: Experimental and numerical study on
mechanical properties of Ultra High Performance Concrete (UHPC). Constr. Build. Mater.
156, 402–411 (2017)
24. Shin, J.: Ultra-High Performance Concrete (UHPC) Precast Segmental Bridges: Flexural
Behaviour and Joint Design. PhD dissertation University of Hamburg (TUHH) (2017)
25. SNZ (Standards New Zealand) NZS 3101:2006, Concrete structures standard’, SNZ
Wellington, New Zealand, (2006)
26. Soner, G., Demet, Y., Fuat, K., Ashraf, A.: Strength prediction models for steel, synthetic,
and hybrid fiber reinforced concretes. Struct. Concrete 20(1), 1–18 (2018)
27. Song, P.S., Hwang, S.: Mechanical properties of high-strength steel fiber-reinforced
concrete. Constr. Build. Mater. 18(9), 669–673 (2004)
J. Mater. Civ. Eng. 19(5), 385–392 (2007)
29. Wafa, F.F., Ashour, S.A.: Mechanical properties of high-strength fiber reinforced concrete.
ACI Mater. J. 89(4), 449–455 (1992)
30. Wang, R., Gao, X., Li, Q., Yang, Y.: Influence of splitting load on transport properties of
ultra-high performance concrete. Constr. Build. Mater. 171, 708–718 (2018)
31. Musmar, M.: Tensile strength of steel fiber reinforced concrete. Contemp. Eng. Sci. 6(5),
225–237 (2013)
Author Index
A Breitenbücher, Rolf, 176

Abbas, Ali A., 359, 596, 730, 920 Burk, Sarah, 1022
Abellán-García, Joaquín, 570, 864
Abrishambaf, Amin, 1056 C
Agra, Ronney Rodrigues, 233 Caballero-Jorna, Marta, 322
Al Marahla, Razan H., 301, 392 Camille, Christophe, 548, 717
Alberti, Marcos G., 693 Capuzzo, V. M. S., 99
Al-Naimi, Hasanain K., 359, 730, 920 Cardoso, M. G., 99
Alsabbagh, Ahmed, 111 Carmona, Sergio, 253
Al-Tabbaa, Abir, 466 Cavalaro, S. H. P., 610
Altoubat, Salah, 883 Chan, Ricardo, 897
Ambrosini, D., 536 Charron, J.-P., 453
Amin, Ali, 368 Chen, E., 477
Andrade, Carmen, 417 Cheung, Andrés B., 75, 745
Arangjelovski, Toni, 380 Chocca, Claudia, 938
Arango-Campo, Samuel E., 864 Clarke, Todd, 140, 548, 717
Ayoubi, Mazen, 221 Cleven, Simon, 24, 402
Codina, R., 536
Combrinck, Riaan, 199
B Conforti, Antonio, 423
Bakhshi, Mehdi, 621, 815 Constantinescu, Horia, 433
Bakhshi, Mohammad, 703 Corinaldesi, Valeria, 1068
Baloch, Hassan, 3 Curosu, Iurie, 1022
Bandelt, Matthew J., 1042
Barros, Joaquim A. O., 49, 703 D
Bauwens, Thomas, 64 da Silva Oliveira, Kaio Cézar, 270
Bernard, E. Stefan, 37 da Silva Ramos Barboza, Aline, 270
Berrocal, Carlos G., 477 Dai, Jian-Guo, 1034
Biermann, Dirk, 801 de Carvalho, Isadora Queiroz Freire, 270
Billington, Sarah, 1042 de Figueiredo, Antonio Domingues, 233
Blanco, A., 610 de la Fuente, A., 610
Blazy, Julia, 209 de Melo Lameiras, Rodrigo, 279
Bogart, Kurt, 140 de Morais Silva, Wandersson Bruno Alcides,
Bokern, Jürgen, 402 270
Boshoff, William P., 199 De Schutter, Geert, 64
© RILEM 2021
https://doi.org/10.1007/978-3-030-58482-5
1162 Author Index
De Smedt, Maure, 151 H

De Wilder, Kristof, 151 Hafezolghorani, Milad, 527
Dehn, Frank, 779 Hamou, Arezki T., 1124
Desmettre, C., 453 Heek, Peter, 908
Destrée, Xavier, 841 Heghes, Bogdan, 433, 1090
di Prisco, Marco, 347 Hewage, Dayani Kahagala, 548, 717
Dias, Gabriela Silva, 270 Hisseine, Ousmane A., 1124
Dias, Salvador, 49 Huang, Bo-Tao, 1034
Dittel, Gozdem, 991 Hunger, Martin, 402
Dlouhý, L., 681
Donnini, Jacopo, 1068 I
Doostkami, Hesam, 489 Ian Gilbert, R., 368
dos Santos, Ana Carolina Parapinski, 279
Dreier, Julia, 801 Isla, F., 536
E J
Enfedaque, Alejandro, 693 Jain, Kranti, 791
Ercegovič, Rok, 290
Estephane, Pierre, 883 K
Karzad, Abdul Saboor, 883
F Kaufmann, Walter, 87, 163, 368
Fataar, Humaira, 199 Khorami, M., 1112
Fernández, María E., 938 Kim, Sungwook, 123
Fernández-Gómez, Jaime A., 570 Kimm, Magdalena, 949
Ferrara, Giuseppe, 983 Kirkland, Brendan, 140, 548, 717
Ferrara, Liberato, 779 Kitagawa, Hirokazu, 1003, 1012
Fiengo, F., 536 Kopálová, M., 681
Figueredo, Diego, 873 Krasnikovs, Andrejs, 841
Filho, Christiano Augusto Ferrario Várady, 270 Kunieda, Minoru, 445
Fiorio, B., 971
Formagini, Sidiclei, 75, 745
Freitas, Danilo José Pereira, 270 L
Lameiras, R. M., 99
G Lauch, K.-S., 453
Gallias, J. -L., 971 Le, An Hoang, 1149
Galobardes, Isaac, 897 Leblouba, Moussa, 883
Gálvez, Jaime C., 693 Lee, Namkon, 123
Garbeth, Michael R., 757 Lehký, David, 527
García, Nicolás, 873 Leite, João, 49
Garcia-Taengua, Emilio, 301, 392, 827 Lesage, Karel, 3
Gebhard, Lukas, 87 Leung, Christopher K. Y., 1034, 1100
Generosi, Nicola, 1068 Li, Jiabin, 333
Genikomsou, Aikaterini S., 651 Li, Jianchun, 661
Gettu, Ravindra, 770 Li, Zhenghao, 1100
Gherman, Oana, 433 Liebscher, Marco, 1022
Giaccio, Graciela M., 189, 423, 536 Lipowczan, Martin, 527
Giraldo Soto, Alejandro, 163 Litina, Chrysoula, 466
Girardello, P., 852 Löfgren, Ingemar, 477
Glynn, Brian, 757 Look, Katharina, 908
Gries, Thomas, 949, 991 Lu, Cong, 1100
Grünewald, Steffen, 3, 64, 779 Luccioni, B., 536
Guzlena, S., 262 Lundgren, Karin, 477
Author Index 1163
M Pleesudjai, Chidchanok, 815

Maalej, Mohamed, 883 Plückelmann, Sven, 176
Magalhães, Margareth S., 313 Pokhrel, Mandeep, 1042
Mahrenholtz, Christoph, 221 Polanec, David, 290
Maiworm, Bastian, 991 Poletanovic, Bojan, 245
Makita, Tohru, 1003, 1012 Polvere, Rafael R., 75, 745
Marchand, P., 1137 Příbramský, V., 681
Mark, Peter, 176, 380, 908 Pukl, Radomír, 527
Markić, Tomislav, 87
Markovski, Goran, 380 R
Martinelli, Enzo, 983 Ramezansefat, Honeyeh, 703
Mashiri, Fidelis, 548, 717 Ramos Barboza, Aline S., 584
Mata-Falcón, Jaime, 87 Rezazadeh, Mohammadali, 703
Matthys, Stijn, 3 Riva, P., 852
Mechtcherine, Viktor, 1022 Rodríguez, Gemma, 938
Medina, Néstor Fabián Acosta, 279 Rodríguez, Iliana, 873
Menna, Demewoz W., 651 Roig-Flores, Marta, 322, 489
Merta, Ildiko, 245 Rossi, Pierre, 515
Mezquida-Alcaraz, Eduardo J., 489, 639 Roy, N., 1137
Mirza, Olivia, 140, 548, 717
Mobasher, Barzin, 815, 963 S
Molins, Climent, 253 Sabir, Amna, 949
Moy, Charles K. S., 897 Saha, Suman, 49
Sakale, G., 262
N Sanjuán, Miguel A., 417
Naaman, Antoine E., 1079 Santana, F. B., 99
Najari, A., 610 Sedran, Thierry, 515
Nakov, Darko, 380 Segura-Castillo, Luis, 873
Nasri, Verya, 621, 815 Serafini, Ramoel, 233
Navarro-Gregori, Juan, 639, 670, 1112 Serna, Pedro, 322, 489, 639, 670, 1112
Negi, Bichitra S., 791 Sevigny-Vallières, C., 1137
Negrini, Alberto, 489 Shahrbijari, Kamyar Bagherinejad, 49
Negrutiu, Camelia, 433, 1090 Shao, Yi, 1042
Nell, Wilhelm, 221 Sirtoli, D., 852
Nemeth, Gergely, 245 Slama, A. -C., 971
Novák, Drahomír, 527 Smarslik, Mario, 176
Nunes, Sandra, 209, 1056 Sosa, Ioan, 433, 1090
Núñez-López, Andrés M., 570, 864 Sousa, Carlos, 209
Soyemi, Olugbenga B., 596
O Spyridis, Panagiotis, 801
O’Flaherty, Tomas, 558 Stephen, Stefie J., 770
Oliveira, T. T, 99 Suksawang, Nakin, 12, 111
Ortiz-Navas, Francisco, 670 Šušteršič, Jakob, 290
Suwanpinij, Piyada, 949
P
Park, Gijoon, 123 T
Patel, Devansh, 815 Tagnit-Hamou, A., 1137
Pepe, Marco, 983 Tailhan, Jean-Louis, 515
Pereira Lima, Iva E., 584 Talavera-Sánchez, Santiago, 670
Pereira, María E., 938 Tang, Zixuan, 466
Petrone, Fernando, 938 Tavares, Maria Elizabeth N., 313
Picazo, Álvaro, 693 Teixeira, Paulo José B., 313
Pimentel, Mário, 209, 1056 Terrade, B., 1137
Pires, Ana R. L., 75, 745 Todor, Adel, 1090
1164 Author Index
Tolêdo Filho, Romildo D., 983 Wangler, Michelle, 991

Torkaman, Javad, 133 Watanabe, Yuji, 1003, 1012
Torres-Castellanos, Nancy, 570 Watts, Murray, 368
Torrijos, María C., 189, 423, 536 Willrich, Fábio Luiz, 279
Toutlemonde, F., 1137 Wilson, William., 558
Tran Thanh, Hai, 661 Winterberg, Ralf, 757
Tri, Le V., 445 Wolf, Sébastien, 24, 841
Tsutsui, Masaki, 445 Wtaife, Salam, 111
Wu, Jia-Qi, 1034
V
Valente, Isabel B., 49, 703
van den Bos, A. A., 503 Y
van der Aa, P. J., 503 Yanai, Shuji, 1003, 1012
Vandevyvere, Brecht, 333 Yao, Yiming, 963
Vandewalle, Lucie, 151, 333, 347 Yohannes, Daniel, 12
Verstrynge, Els, 151, 333 Yu, Jing, 1034
Vivas, Juan C., 189, 536
Vlasák, Oldrich, 24 Z
Vrijdaghs, Rutger, 151, 347, 402 Zajc, Andrej, 290
Zerbino, Raúl L., 189, 423, 536
W Zhai, Mengchao, 963
Walter, Lars, 801 Zhang, Y. X., 661
Wang, Jingquan, 963 Zhou, Jiajia, 1100

2021 Book FibreReinforcedConcreteImprove

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2021 Book FibreReinforcedConcreteImprove

Uploaded by

Copyright:

Available Formats

RILEM Bookseries

RILEM, The International Union of Laboratories and Experts in Construction

More information about this series at http://www.springer.com/series/8781

José R. Martí-Vargas Juan Navarro-Gregori

Fibre Reinforced Concrete:

José R. Martí-Vargas Juan Navarro-Gregori

ISSN 2211-0844 ISSN 2211-0852 (electronic)

2020, a weird year.

Pedro Serna (Chairman), Spain

M. A. Aiello, Italy E. Cuenca, Italy

K. Kobayashi, Japan K. A. Rieder, Germany

RILEM fib ACI

Rheology and Early-Age Properties

Aligned Interlayer Fibre Reinforcement for Digital Fabrication

Development of an HPFRC for Use in Flat Slabs . . . . . . . . . . . . . . . . . . 209

An Experimental Study of the Inﬂuence of Moderate Temperatures

Self-healing of Fibre Reinforced Concrete Containing an Expansive

Analytical and Numerical Models

Machine Learning Prediction of Flexural Behavior of UHPFRC . . . . . . 570

Modeling the Compressive Behavior of Steel Fiber Reinforced

Codes and Standards

Using Decades of Data to Rethink Proportioning and Optimisation of

Case Studies: Structural and Industrial Applications

Potential of Using Recycled Carbon Fibers as Reinforcing Material for

Textile Reinforced Concrete (TRC)

UHPFRC, SHCC and ECC

Characterization of Ultra High Performance Fiber Reinforced

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1161

RILEM Proceedings (PRO)

PRO 1: Durability of High Performance Concrete (ISBN: 2-912143-03-9; e-ISBN:

PRO 9: 2nd International RILEM Symposium on Adhesion between Polymers and

PRO 24: 2nd International RILEM Workshop on Frost Resistance of Concrete

PRO 51: 2nd International Symposium on Advances in Concrete through

PRO 64: Long Term Performance of Cementitious Barriers and Reinforced

PRO 77: International Conference on ‘Material Science - AdIPoC - Additions

PRO 89: RILEM International workshop on performance-based speciﬁcation and

RILEM Reports (REP)

Report 24: Workability and Rheology of Fresh Concrete: Compendium of Tests

Report 39: Final Report of RILEM Technical Committee TC 187-SOC

Hassan Baloch1(&), Steffen Grünewald1,2, Karel Lesage1,

Abstract. Strain-hardening cementitious composites (SHCC) have a high

Keywords: Fibres Strain-hardening cementitious composites Rheology

Strain-hardening cementitious composites (SHCC) are a special class of material which

2 Materials and Methods

Table 1. Properties of ﬁbres used

2.2 Mix Formulations and Mixing Regime

Table 2. Mixture formulations

2.3 Test Methods

2.3.2 Rheological Investigation

3 Results and Discussion

Table 3. Flow spread test results.

3.1.2 Effect of the Fibre Factor

Fig. 1. Effect of HDPE ﬁbre addition on the flow spread

Flow Spread (cm)

Fig. 2. a. Traditional fibre factor b. Modified fibre influence factor

3.2 Rheological Investigation

3.2.1 Relation Between Yield Stress and Flow Spread

3.2.2 Relation Between Viscosity and Fibre Influence Factor

Shear Stress (Pa)

She ar rate 1/s She ar rate 1/s

Slump Flow (mm)

y = 2E-05x2 - 0,0572x + 21,976 8 y = 0,0001x2 - 0,0982x + 27,109 8 y = 0,0001x2 - 0,1033x + 31,088

y = 5E-07x2 + 0,0045x + 11,885 5 5

Nakin Suksawang1(&) and Daniel Yohannes2