HarteA - Dietschp.edsreinforcementoftimberstructures STAR

Reinforcement of Timber Structures
A state-of-the-art report
Editors
Annette M. Harte, Philipp Dietsch
Acknowledgement
The members of COST Action FP1101 would like to thank the COST Office in Brussels for
financially supporting this publication. The financial support of Rotafix UK, Rothoblaas and SPAX is
also gratefully acknowledged.
No permission to reproduce or utilise the contents of this book by any means is necessary, other than
in the case of images, diagrams or other material from other copyright holders.
In such cases, permission of the copyright holders is required. This book may be cited as:
Reinforcement of Timber Structures – a state-of-the-art report.
Neither the COST Office nor any person acting on its behalf is responsible for the use which might be
made of the information contained in this publication. The COST Office is not responsible for the
external websites referred to in this publication.
Copyright Shaker Verlag 2015
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or
otherwise, without the prior permission of the publishers.
Printed in Germany
ISBN: XXX-xxx-xx
ISSN: YYY-yyy
Contents
Preface v
Introduction 1
Annette M. Harte, Philipp Dietsch
Part I: Reinforcement Applications 3

1 Reinforcement of timber beams 5
Steffen Franke, Bettina Franke, Annette M. Harte
2 Reversible timber-to-timber strengthening interventions on wooden floors 25

Alessandra Gubana
3 Reinforcement of timber columns and shear walls 39
Wen-Shao Chang
4 Analysis and strengthening of carpentry joints 55

Jorge M. Branco, Thierry Descamps
5 Reinforcement of dowel type connections 77

Laurent Bléron, Damien Lathuillière, Thierry Descamps, Jean-François Bocquet
6 Seismic Strengthening of Timber Structures 89

Maria Adelaide Parisi, Maurizio Piazza
Part II: Reinforcement Methods 111

7 Adhesives for on-site bonding: characteristics, testing and prospects 113
Benedetto Pizzo, Dave Smedley
8 Glued-in rods 133

René Steiger, Erik Serrano, Mislav Stepinac, Vlatka Rajčić, Caoimhe O’Neill,
Daniel McPolin, Robert Widmann
9 Reinforcement with self-tapping screws 161

Philipp Dietsch, Reinhard Brandner
10 FRP reinforcement of timber structures 183
Kay-Uwe Schober, Annette M. Harte, Robert Kliger, Robert Jockwer, Qingfeng
Xu, Jian-Fei Chen
11 Nanotechnologies for reinforcement and protection of timber structures:

innovative nano-coatings 209
Tanja Marzi
iii
Outlook 231
Philipp Dietsch, Annette M. Harte
List of contributors 232

List of reviewers 234
iv
Preface
This report is a publication developed within the European network COST Action FP1101
“Assessment, reinforcement and monitoring of timber structures”.
COST Action FP1101 is a research network established under the aegis of the COST domain
“Forests, their Products and Services”. The main objective of COST Action FP1101 is to
increase the acceptance of timber in the design of new structures and in the repair of existing
structures by developing and disseminating methods to assess, reinforce and monitor them.
The Action is structured into three working groups: Working Group 1: Assessment of timber
structures, Working Group 2: Reinforcement of timber structures, Working Group 3:
Monitoring of timber structures.
This report has been prepared by members of Working Group 2. It is a distillation of the work
presented and discussed by members at a series of workshops over the course of the Action.
These workshops considered the latest research in reinforcement technologies for both
modern and historical timber structures, current and proposed design methods and also case
studies that illustrate the current state-of-practise for on-site implementation of these
technologies.
Gratitude is expressed to the COST Office for funding these workshops and the publication
of this report. The contribution of Rotafix UK, Rothoblass and SPAX to the technical
activities of COST Action FP1101 and their financial support towards the publication of this
report is also gratefully acknowledged. Finally, special thanks are due to the authors,
reviewers and all working group members who contributed to the preparation of this report.
Annette M. Harte, Philipp Dietsch, Editors
v
vi
Introduction
Harte, A.M., Dietsch, P.
The maintenance and rehabilitation of existing buildings is an area of increasing importance

not only for economical but also for environmental, historical and social reasons. In Europe,
as over 80% of buildings are over 50 years old, currently about 50% of all construction is
connected to the repair, maintenance, and improvement of existing buildings. The need for
structural reinforcement of timber buildings may become necessary for many reasons
including change of use, deterioration due to a lack of monitoring and maintenance,
accidental damage, changes in building regulations, the requirement to increase seismic
resistance or due to the expiration of the planned lifetime. Depending on the situation, the
function of the reinforcement may be to restore a weakened structure to its original load
bearing capacity (repair) or to increase the load bearing capacity of an intact structure
(upgrade).
A large range of reinforcement methods is currently available and further technologies are
under development. Reinforcement approaches include the addition of metallic, wood-based
or FRP reinforcing plates attached to the timber member, the use of mechanical fasteners
such as self-tapping screws and glued-in rods, replacement of decayed material using
prostheses, adhesive/resin repairs, and more recently nanotechnology. Many of these
methodologies are however, not harmonised for their use in-situ nor adapted depending on
whether the structure is part of the regular building stock or belongs to cultural heritage.
Special consideration is required for buildings that are classified as part of the cultural built
heritage. For these buildings, the reinforcement strategy adopted should aim to minimise the
removal of original material, conserve the original function of the structure and be reversible.
In the past, due to a lack of knowledge on how to select and implement appropriate
reinforcement methods, some of the interventions have caused further damage to the
structures. For all building interventions, it is important to minimise the disturbance not only
to the building and but also to its occupants.
The main aim of this report is to summarise the current and emerging methods that are
available to repair or enhance the structural performance of timber structures and to provide
guidance to the use of these methods. The report is organised in two main parts. In Part I, the
different structural elements and subsystems that make up our buildings are considered.
These include beams, floors, columns, shear walls and connections. For each of these, the
types of failure present in the elements are described and the appropriate reinforcement
strategies for each case are presented, including consideration of cultural heritage issues. The
reinforcement of buildings to increase their resistance to seismic actions is also included in
this part. The focus of Part II is on reinforcement materials and methods. These include
adhesive systems, mechanical fasteners such as glued-in rods and self-tapping screws, fibre
reinforced polymer (FRP) laminates and bars, and emerging nano-structured materials.
1
The properties of these materials, their methods of application and relevant design rules are
described.
The report provides details not only of the latest research findings related to the
reinforcement of timber structures, but most importantly how these methods can be best used
in practice. Many examples are given of the implementation of the various reinforcement
methods. Because of this, the report will be of interest not only to the research community,
relevant standardisation bodies, and policy makers but also to practitioners, such as
architects, structural engineers and builders, representatives of the timber construction and
building industry and product developers in the sector of reinforcement technologies.
In conclusion, the reinforcement of existing timber structures is important from economic,

environmental, historical and social perspectives. However, until now knowledge on the
range of reinforcement techniques available, on the criteria for selection of the best approach
for a given situation and on the correct procedures for implementing these methods on site
has been defragmented. It is hoped that this report will address these shortcomings thus
constituting distinct progress in the field of reinforcement of timber structures at a European
level.
2
PART I:
REINFORCEMENT APPLICATIONS
3
4
Reinforcement of timber beams
1
Steffen Franke1, Bettina Franke2, Annette M. Harte3
Summary
High performing, such as highly-loaded, and large span timber beams are often used for sports and
industry halls, public buildings or bridges, and provide an aesthetically pleasing and environmentally
friendly structural solution. Reinforcement of beams may be required to extend the life of the
structure due to deterioration or damage to the material or due to a change of use. The main aim of
this chapter is to summarise the current and emerging methods that are available to repair or enhance
the structural performance of timber beams. An overview of the main materials, cross sections and
geometries used for timber beam structures is presented. Furthermore, their general failure modes are
described and typical reinforcement methods and corresponding retrofitting techniques are given. For
each of the failure modes, the methods and their advantages are summarized. The reinforcement
methods include wood to wood replacements, use of mechanical fasteners such as screws and rods,
and methods which add additional strengthening materials.
1 Typology of timber beams

Timber beams can mainly be classified according to the span, the geometry and the material used, as
summarized in Tab. 1. The focus here is on high-performance, long-span structures. Tab. 2 gives an
overview of typical timber beams in relation to the sizes of the cross section and the span ratio. In
Europe, glulam members or block glued glulam members are the main construction elements used for
large open span spaces, stadiums or bridges in which the primary structure is timber. The typical
layered cross section of glulam reaches from 100 to 250 mm in width and up to 2500 mm in depth but
also in bigger dimensions as block glued glulam. Box or composite beams are alternatives providing a
lower self-weight.
Tab. 1 Classification of timber beams

Material Cross section Geometry
Solid wood Solid cross section Straight Beam
Glulam, Block glued glulam Box-Beam Curved Beam
Laminated veneer lumber I-Beam Tapered Beam
Plywood T-Beam Truss
(OSB, LSL) C-Beam
Cross laminated timber
1)
Professor for Timber Structures, Bern University of Applied Sciences, Biel/Bienne, Switzerland
2)
Research Associate, Bern University of Applied Sciences, Biel/Bienne, Switzerland
3)
Senior Lecturer, College of Engineering and Informatics, National University of Ireland, Galway
5
Tab. 2 Overview of timber beam forms

Timber beam form Cross section Span, Depth ratio
Straight depth h 10 m ≤ l ≤ 40 m
beams span l h ≈ l / 17
Tapered depth h 12 m ≤ l ≤ 25 m
beams h ≈ l / 15
span l
Curved 15 m ≤ l ≤ 35 m
depth h
beams span l h ≈ l / 17
depth h 20 m ≤ l ≤ 85 m
Trusses
h ≈ l / 10
span l
2 Failure modes
2.1 General
Structures have to adopt, and transfer external loads to the ground and also deal with internal loads.
This leads to stresses and deformations in the structure which must not exceed design strength and
deformation limits. In designing new structures, a full cross section with minor damage and correct
material grades are assumed. However, in existing timber structures the cross section and/or the
properties of the material of the members can be reduced due to mechanical and biological damage.
Both types of damage influence the load carrying capacity and serviceability of single members or the
complete construction. Within the assessment of timber structures, damage or failure has to be
detected and assessed for the resistance and serviceability of the timber structure. The net cross
sections observed must be compared to the designed cross sections.
The failure analysis on timber structures in Germany carried out by Blass & Frese [1, 2] gives a good
overview of the distribution of main types of failure classified according to the construction, use and
region. Most assessment reports state that the timber structures have been built using glulam beams of
quality GL28h (see Tab. 3). Their shape, however, is more varied with the most common being, by
order: straight (154/426), tapered straight (124/426), pitched cambered (90/426) and curved (47/426).
80 % of the failure cases could be detected in bending members, followed by 8 % in compression
members. Furthermore in 75 % of the failure cases cracks could be detected. Typical reasons and
types of failure are summarized in Fig. 1 and Fig. 2 .
Tab. 3 Most frequent characteristics of the timber structures assessed, from [1] and [2]
Characteristic Main result Corresponding no. of assessments
Material Glulam 594 80%
Quality (or equivalent) GL28h 68 72%
Load type Bending 470 80%
6
Biological reasons
36% Crack in the
grain direction
Structural- 75%
Physical reasons
38%
Chemical Other
reasons failure
5% 11%
Mechanical Tension
Shear failure
Other reasons reasons failure 6%
12% 9% 8%
Fig. 1 Reasons for damage [1], [2] Fig. 2 Types of failure [1], [2]
For high performing and long span timber members the typical failure modes are described in detail in
the following sections.
2.2 Cracks
The most common type of failure, Fig. 2, was observed as the appearance of cracks in grain direction.
The variation of the surrounding climate at a timber beam changes the moisture content and lead to
shrinkage or swelling of the cross section. Non uniform distributions of the moisture content over the
cross section and/or restraint deformations lead to internal stresses and, if the material strength is
exceeded, to cracks in the cross section which can significantly reduce the capacity, Fig. 3. For the
determination of the influence of cracks in timber beams on the load carrying capacity or stiffness no
comprehensive methods are known. Methods and guidelines for this evaluation are currently under
development at the Bern University of Applied Sciences.
The amount and distribution of cracks depends on several factors, such as timber and glue quality,
defects, loading situation or beam shape. Regarding the distribution of cracks in the timber beams, a
summary of their characteristics can be found in Tab. 4Note: Failure under tension stress
perpendicular to the grain in glulam members has to be distinguished from delamination failure within
the glue lines as shown in Fig. 5. Special techniques can be used for the classification of
delaminations as shown in [3].
a) b) c)
Fig. 3 Glulam cross sections; a) sketch of undamaged cross section with 100% capacity, b) sketch of
cross section with some possible damage and unknown capacity, c) real example of cross section with
internal cracks
7
Fig. 4 Cracks in grain direction at a glulam Fig. 5 Delamination at a glulam member

member
Tab. 4 Characteristics of cracks and their distributions [1], [2]
Location; Length;
Crack cause Quantity Depth ratio* Cases
Stress concentration
At the singularity; 1-10 m;
(Restrained shrinkage, notches, 35%
Single mostly 1.0
transverse forces ...)
Randomly; 0.1-1 m;
Normal climate changes 33%
Numerous 0.1 to 0.4
Element quality At the defect;
17%
(Glue line or finger joints) Depending on its extent
Overloading Various; 1 m;
15%
(Shear or bending stresses) Single to numerous mostly 1.0
* Ratio of depth of crack to width of beam
2.3 Bending failure

Bending results in longitudinal tension and compression stresses distributed over the depth of the
cross section. The tension stress leads to a brittle failure due to the rupture of the wood fibres, as
shown in Fig. 6. Longitudinal compression stress results in elastic and plastic deformations which can
be described as ductile and leads to the so-called kink bands.
Fig. 6 Principal sketch for bending failure Fig. 7 Tension failure under bending
8
Due to natural defects, such as knots, the tension strength can be reduced compared to compression
strength. Therefore, bending failure is mainly described by brittle failure of timber beams within the
tension zone, as shown in Fig. 7. Bending failure is classified as critical and can lead to a single
failure of the structural element or the complete construction.
2.4 Compression failure

Failure under longitudinal compression stress occurs mainly in timber beams or columns. Failure
under compression stress perpendicular to the grain can also be described as a ductile failure with
plastic deformations and occurs mainly at supports or at loading points where high loads have to be
transferred, as shown in Fig. 8 and Fig. 9. In both cases, these plastic deformations can further lead to
eccentricities and load redistributions within the complete structure and therefore overstress parts of
the structure. The overall stability will also be influenced.
Fig. 8 Principal sketch for compression failure at Fig. 9 Compression failure at loading point
support
2.5 Tension failure
Fig. 10 Principal sketch for tension failure Fig. 11 Tension failure perpendicular to the grain
perpendicular to the grain at a notch at a notch
9
a) b)
Fig. 12 Tension failure parallel to the grain of experim. tension tests: a) short-fibred, b) long-fibred,
[4]
Tension stress has to be considered in the parallel to the grain and perpendicular to the grain
directions. When the tensile capacity of the timber is exceeded, brittle failure occurs. Examples of
tensile failure parallel to the grain and perpendicular to the grain are shown in Fig. 10, Fig. 11, and
Fig. 12. However, due to the low tension strength perpendicular to the grain of solid wood and glulam
members, which is almost zero due to natural defects, failure under tension stress perpendicular to the
grain occurs more often. Therefore, wood products are mostly optimized to increase the tension
strength perpendicular to the grain, but it still has carefully to be considered in the design process.
Tension stress perpendicular to the grain occurs in curved, tapered and end-notched glulam members
as well as in members with holes, additional connected structural elements or equipment, and at
connections loaded perpendicular to the grain.
2.6 Shear failure

In most cases, bending stress and deflection limits govern the design of the members. But for short
beams the shear stress can be more important. In general for beams, the shear stress reaches the
maximum value close to the supports (Fig. 13). Additionally, end-notched beams and beams with
holes can lead to shear stress concentrations. Failure due to shear stress is characterised by a sliding of
the fibres and thus cracking parallel to the grain and is considered as a brittle failure. The cracks are
mainly closed and therefore hard to detect if they are not at the end of the beam, as shown in Fig. 14.
Fig. 13 Principal sketch for shear failure Fig. 14 Shear failure at holes
10
2.7 Insects and fungi

Decay due to fungi is possible for timber beams with a moisture content close to/or over the fibre
saturation point, see 5. The fibre saturation point varies from wood to wood species and shows a range
from 26 M% - 32 M% (in mass percentage). The different fungi and their typical appearance and
hazard are summarized in [5].
Generally, decay due to insects can occur within a range of wood moisture content above 6 M%, but
can be neglected in construction of service class 1 or 2, where technical dried wood members like
solid wood, glulam or wood products are used, [6]. The classification and identification of insects is
described in detail in [5].
Fig. 15 Risk of insect and fungal decay in relation of the moisture content
3 Retrofitting techniques
3.1 General
The following sections illustrate possible retrofitting techniques for timber beams. Detailed
descriptions of the different techniques and their design can be found in other chapters of this report.
In the case of damage or decay, the timber beam or parts of the beams have to be replaced as
described in Section 3.3. Retrofit measures to improve the performance of timber beams in bending,
shear, and in tension and compression perpendicular to the grain are described.
Fig. 16 Risk of insect and fungal decay in relation of the moisture content
11
3.2 Repair of shrinkage cracks or delaminations

Repairs of shrinkage cracks or partial delamination of glue lines in glued laminated timber may be
carried out; however, it is not advisable to repair cracks in solid timber members. In many instances,
repair of cracks is carried out in conjunction with other reinforcement interventions. For glued
laminated timber, it is generally considered necessary to repair shrinkage cracks in regions of high
shear stresses and high tensile stresses perpendicular to the grain. The first purpose of the repair is to
restore the load carrying capacity of the glulam member. The visible cracks or delamination in glulam
members have always to be assessed by an expert before planning the repair process regarding the
load carrying capacity of the whole structure.
For crack openings smaller than 10 mm wide and with low to medium fibrosity/splintering the repair
can be done by injection of adhesives, [7]. A number of such products with technical approvals for
these applications are available. The current regulations according to the requirements in the standards
and possible technologies are summarized in [7] for the European market. The methodology of
repairing cracks and delamination is also described in [7].
To ensure adequate carrying capacity, the preparation of the bonding surfaces is important to ensure
the required quality of the applied technology and to avoid defects. In general, the repair procedure
involves cutting out the cracks using a saw, router or grinder to make a clean slot. The slot is then
cleaned, optionally brushed with a primer and preparation of filling and ventilation holes as well as
supporting bracing system before filling with a suitable adhesive. Fig. 16 shows three different
adhesives and various technology for repairing cracks or delamination.
For combination of repair methods, the barrier effect of repaired bonded joints against the ingress of
water and water vapour has to be considered during the maintenance planning.
3.3 Replacement of damaged or decayed parts

Timber that has decayed due to fungal or insect attack is porous, brittle and has very poor strength
properties [8]. This decay often occurs in localised parts of the beam, such as at the ends where the
timber is in direct contact with a masonry supporting wall as seen in Fig. 17. In these cases, the
condition of the rest of the beam is generally good. Other types of accidental damage, such as fire
damage, may cause a reduction in the member cross-section resulting in inadequate strength and
stiffness. Decayed and damaged material should be removed and the member upgraded to restore the
load-bearing capacity of the member.
Repair methods include replacing the damaged section with a timber or engineered wood prosthesis
connected to the original beam by means of a scarf joint with wooden pegs and/or adhesive (Fig. 18),
by means of bonded-in rods or plates (Fig. 19) or replacing the damaged section with a prosthesis
built up from timber boards using screws as shown in Fig. 20 [10], [11].
The most common retrofit method employed involves replacing the damaged timber with a timber
prosthesis which is bonded to the sound timber in the original beam using steel or fibre reinforced
polymer (FRP) rods or plates. Using this approach, the scale of the intervention is limited and the load
bearing function is preserved. The implementation of this type of repair involves a number of
different steps [12]. Initially, the beams are propped. The damaged part of the beam is then removed
by cutting either vertically or at an angle of 45˚ to the vertical, as seen in Fig. 19 and Fig. 21. Holes or
grooves to take the connecting rods/plates are drilled in the beam and the prosthesis and are partially
12
Fig. 17 Beam end decay [9] Fig. 18 Prosthesis with scarf connection
Fig. 19 Beam end repair using bonded in rods or plates
Beam
Screws
Prosthesis
Fig. 20 Beam end repair using bonded in rods or plates
Fig. 21 GFRP rods inserted in ends of propped beams [13]
13
filled with adhesive. The reinforcing elements are inserted into the beam and prosthesis and the
adhesive is topped up in holes/grooves if necessary. Additional props are introduced to support the
prosthesis. When grooves are used, a timber strip is normally inserted to improve the appearance of
the repaired member and to provide fire protection. When the adhesive has fully cured, the supporting
props are removed.
The prosthesis should be of the same species as the timber to be repaired, or be compatible in terms of
its mechanical properties by using, for instance, engineered wood products. The moisture content of
the prosthesis should be the same as that of the beam being repaired. The adhesive used is usually a
thixotropic epoxy resin and the type used should be specially formulated to bond with the timber and
the reinforcement. The design of the repair is based on the requirement that the reinforcement should
provide the same load bearing capacity as the section with sound timber.
3.4 Flexural reinforcement
In order to increase the flexural strength and stiffness of beams, reinforcing elements are added that
act compositely with the existing member. A large variety of reinforcement configurations is
available. The reinforcing elements can be in the form of rods, plates or other structural shapes which
are connected to the beam using mechanical fixings or structural adhesives. These reinforcing
elements can be placed inside or outside of the member and may be slack or prestressed. The
reinforcement material can be a metal, fibre reinforced polymer (FRP) or engineered wood product.
Fig. 22 and Fig. 23 show some possible configurations for external and internal reinforcement. Apart
from the structural requirements, the configuration selected for a particular application may depend on
other factors: the presence of decorative ceilings or painting on beams may require that access for
reinforcement is restricted to the top or sides of the beam; fire protection requirements may exclude
the use of externally bonded plates on exposed surfaces.
As timber beams generally fail in tension in a brittle fashion, positioning of the reinforcement on the
tensile face of the beams is very effective for increasing bending strength. With increasing percentage
of tensile reinforcement, the neutral axis moves towards the bottom of the beam. As a result the
compressive strain in the timber increases relative to the tensile strain and compressive yielding may
occur before the timber eventually fails in tension. The load-deflection response for a timber beam
reinforced with carbon fibre reinforced polymer (CFRP) plates, which were inserted from the top, is
shown in Fig. 24, [14]. The unreinforced beam A has a brittle response. For the reinforced beams, two
of the beams display significant ductility in their response before failure.
Kliger et al. [15] investigated the influence of the distribution of the reinforcement between the tensile
and compressive faces of the beam on the bending strength and stiffness. They concluded that for
maximum strength, 75% of the reinforcement should be on the bottom face and 25% on the top. To
achieve maximum ductility, all of the reinforcement should be placed on the bottom. The maximum
stiffness enhancement was achieved when the reinforcement was equally distributed between the top
and bottom faces. However, for low percentages of reinforcement the stiffness gain by distributing the
reinforcement between the two faces may not justify the additional work involved.
14
Fig. 22 External reinforcement arrangements
Fig. 23 Internal reinforcement arrangements
Fig. 24 Load-deflection response of unreinforced (A) and reinforced (D1, D2, D3) beams, [14]
Steel and other metals have been used for reinforcing timber for many years. Mark [16] bonded
aluminium sheets to the top and bottom faces of timber beams and reported an increase in the flexural
strength and stiffness. Dziuba [17] tested timber beams reinforced with steel rods on the tension face
and noted that compressive yielding occurred prior to failure in tension. DeLuca and Murano [18]
reinforced spruce beams with 0.82% steel bars and recorded mean increases of 48% in peak load, and
26% in stiffness. Nielsen and Ellegaard [19] investigated the use of punched metal plate connectors as
flexural reinforcement for timber but with limited success.
Fibre reinforced polymer in the form of pultruded rods or plates have been the subject of a
considerable amount of research for the reinforcement of timber and have been used in practise to
reinforce solid timber and glulam structures. Several fibre types are available including carbon
(CFRP), aramid (AFRP), glass (GFRP), basalt (BFRP) and steel (SFRP). CFRP [14], [15], [20]-[22]
and GFRP [23]-[26] have been widely used as externally bonded plates or internally as near surface
mounted reinforcement bonded into grooves cut into the beams. CFRP materials have high strength
and stiffness properties and, depending on the properties of the unreinforced beam and the percentage
of reinforcement used, strength and stiffness increases of over 100% can be achieved. For lower grade
timber, less expensive GFRP materials are generally sufficient to provide the required strengthening
but the stiffness increase can be limited. Steel fibre reinforced polymer bars have been found to
provide a significant increase in capacity and ductility but insignificant improvement in stiffness [27].
15
The use of FRP materials has a number of advantages over steel due to their light weight, their
corrosion resistance, and their ease of handling on site. It should be noted that the routing of grooves
to house reinforcement may cause a weakening of the beam as a discontinuity is introduced in the
wood fibres in the vicinity of the grooves.
Prestressed steel or FRP plates bonded on the tension face with epoxy resin [28]-[32] can provide
further increases in strength. A pre-camber is introduced in the beam due to the eccentric prestress,
which can be offset against the deflection to the external loads. However, this technique is currently
not used in practise due to the difficulty in installation and insufficient knowledge of the long-term
performance of the prestressed members.
As the flexural capacity of the beam is enhanced, the shear capacity may be exceeded. In these cases,
a combination of both flexural and shear strengthening may be required.
3.5 Reinforcement in tension perpendicular to the grain

Failure in tension perpendicular to the grain in timber beams can arise in notched beams, around holes
and in curved, tapered or pitched cambered beams. Reinforcement of beams in these situations can be
achieved using internal or external reinforcement. Types of internal reinforcement include self-
tapping screws, bonded-in or drilled-in threaded steel rods or bonded-in FRP rods or tubes. External
reinforcement is achieved by mechanically fixing and/or gluing on sheets of wood-based panels, such
as plywood, or FRP sheets or nail plates.
For the case of notched end beams, the stress concentration at the corner of the notch leads to crack
initiation and rapid crack propagation results in a sudden brittle failure of the beam as shown in Fig.
12. The high tensile stresses perpendicular to the grain are accompanied by high shear stresses.
Different reinforcement methods are illustrated in Fig. 25. The reinforcement can be deployed
perpendicular to or at 45˚ to the beam axis. Due to the presence of high shear stresses, the
performance of notched beams reinforced at 45˚ is superior. This has been validated by a number of
experimental investigations [34]-[37]. Reinforced notches have enhanced load-bearing capacity but
also display less brittle failure modes than is the case for unreinforced notches.
For screws or glued-in rods, the requirements for minimum edge distances and spacing must be
satisfied while keeping the reinforcement as close as possible to the notch corner. Externally bonded
sheets of FRP or plywood are placed on both sides of the beam and extend over the full height.
Screws or nails are normally used to provide the required bonding pressure while the adhesive is
curing.
Irrespective of the type of reinforcement used, the usual design approach is to assume that the tensile
forces perpendicular to the grain are carried entirely by the reinforcement. For beams with a rectangular
notch at the support, the tensile force for which the reinforcement is designed is given by [34]
Ft ,90,d  1,3 Vd 31     2 1    

2 3
(1)
 
where Vd is the design shear force and α is the ratio of the reduced beam height at the notch to the total
beam height.
For the case of beams with a round or rectangular hole,
16
Vd hd  hd2  Md
Ft ,90,d  3  2  0,008  (2)
4h  h  hr
(a) (b)
(c) (d)
(e) (f)
Fig. 25 Typical reinforcement arrangements for notches. (a) & (b) self-tapping screws, (c) & (d)
glued-in rods, (e) & (f) EWP or FRP side plates
Fig. 26 Reinforcement of openings
Fig. 27 Reinforcement of pitched camber beams
where Vd and Md are the design values of the shear force and bending moment at the section,
respectively, h is the beam height, hd is the hole height and hr is the distance from the edge of the hole
to the top or bottom of the beam [34]. Typical reinforcing configurations for beams with holes are
shown in Fig. 26.
For curved, pitched tapered or tapered beams, the tensile stress perpendicular to the grain occurs in
the apex region, which is highlighted in grey in Fig. 27, [39]-[41]. Reinforcement of this region can be
achieved through the use of screws, glued-in rods or side plates, as shown in Fig. 27.
The design tensile stress, σt,90,d, may be calculated as
6M ap ,d
 t ,90,d  k p (3)
b  hap
2
17
where Map,d is the design moment at the apex, hap is the depth of the beam at the apex and kp is a
function of the taper angle, the radius and the depth at the apex [42]. The load to be carried by discrete
connectors, such as screws or glued in rods, is the total of all tensile stresses on an area equal to the
connector spacing by the beam width. The load capacity of the connectors is determined by the
withdrawal capacity and the tensile strength.
The reduced section of the beam due to drilling of holes for the radial reinforcement must be
considered in the design. As the portion of the holes below the neutral axis cannot be considered
effective in tension, the section modulus in bending is reduced [41].
3.6 Shear reinforcement

The methods available to strengthen beams in shear are the same as those described for reinforcement
against tension perpendicular to the grain. These include internal reinforcement in the form of screws
and bonded-in rods of steel or FRP and external reinforcement in the form of side plates.
Akbiyik et al. [43] investigated the shear reinforcement of timber stringers with horizontal splits using
hex bolts, lag screws, and plywood and GFRP side plates. The bolts were epoxy bonded in vertical
holes drilled from the top. The lag screws were installed vertically and at 45˚. The plywood and GFRP
side plates were attached to the sides of the beams using screws. All repair types were effective with
an average increase in the residual shear capacity of 62%. None of the repaired specimens recovered
the original undamaged stiffness. The extent of the existing damage had a big influence of the
effectiveness of the repair. For the more highly damaged stringers, the use of GFRP side plates was
the most efficient method.
Several investigators examined FRP shear reinforcement of beams [37], [43]-[46]. Radford [44]
reported an increase in stiffness of over 270% when using epoxy bonded side plates of GFRP with the
fibres oriented at +/-45˚ to the beam axis. The use of vertical GFRP shear spikes produced a stiffness
enhancement of over 160%. Inserting the shear spikes at a spacing equal to the beam depth was found
to be the most effective. Gentry [46] also used a combination of FRP flexural plate and FRP shear
pins to reinforce glulams. Svecova and Eden [47] used GFRP bars to reinforce beams from a bridge.
This resulted in a significant increase in strength and decrease in variability.
Widmann et al. [37] investigated the shear reinforcement of delaminated glulam beams. Glulam
beams with delaminated middle lamellae were loaded to failure and then reinforced with self-tapping
screws or epoxy bonded CFRP side sheets oriented at 45˚ to the beam axis. Both approaches showed a
significant increase of the shear strength. The ultimate shear strength could not be determined as
different failure modes were found.
Trautz and Koi [48] described a series of tests performed on glued-laminated beams reinforced with
screws using different arrangements to carry tensile and compressive forces. The shear stiffness of
beams reinforced with screws arranged in a nested pattern with screws carrying loads in tension and
compression was superior to that achieved by reinforcing with diagonal tension screws only.
Dietsch et al. [49] describe design approaches for the shear reinforcement of timber beams in the
unfractured and fractured states. The types of reinforcement considered are self-tapping screws or
threaded rods deployed at an angle to the beam axis. The models account for the enhancement in
shear performance resulting from compression induced perpendicular to the grain by the
reinforcement. Comparison with results of experimental tests provided the validation of the shear
stiffness and strength predictions. For unfractured beams, an increase in capacity of 20% is achievable
18
when reinforced with threaded rods. For fractured beams, the maximum increase in bending stress
compared to the intact beam is 33%.
3.7 Reinforcement in compression perpendicular to the grain

Crocetti et al. [50] undertook experimental investigations of the compressive strength perpendicular to
the grain of glulam beams reinforced internally with glued-in steel rods and glued-in hardwood
dowels and externally with screwed-on steel side plates. The beams reinforced with glued-in rods all
failed in buckling and resulted in a significant increase in both strength and stiffness over the
unreinforced beams. The beams reinforced with side plates also produced a significant enhancement
in the compression strength.
Blass and Bejtka [34] proposed a design model for the compressive capacity, R90,d, of a beam support
reinforced using self-tapping screws that accounts for buckling and screw withdrawal. This may be
expressed as
nRd  kc,90  lef ,1  b  f c,90,d 

 
R90,d  min   (4)
b  lef ,2  f c,90,d
 

where n is the number of screws, Rd is the lesser of the withdrawal capacity and the buckling capacity
of the screw, fc,90,d is the design value of the compressive strength perpendicular to the grain, b is the
beam width, lef,1 and lef,2 are the lengths as defined in Fig. 28, and kc,90 is a load distribution coefficient
in the range [1;1.75]
In existing structures, the insertion of screws or glued-in rods at locations of concentrated loading
may be difficult to achieve.
Fig. 28 Reinforcement for compression perpendicular to the grain
3.8 General remarks

In practical terms, the choice of reinforcement method for existing timber beams will be based not
only on the ability of the reinforcement to provide adequate strengthening of the structure but will be
constrained by other factors such as aesthetics, need for reversibility, access for repair, and available
expertise.
As the reinforcing elements generally have different stiffness, thermal expansion and moisture
absorption properties than the timber element, factors such as constrained shrinkage and swelling due
to thermal or moisture changes must be considered and if necessary additional thermal or moisture
induced stresses should be accounted for in design. Where the reinforcement results in a significant
change in beam stiffness, it is important to consider the consequences for the overall behaviour and
load distribution of the entire structure.
19
4 Case studies
4.1 Clyne Castle, Wales – Replacement of decayed parts [13]
Clyne Castle is a Grade II listed building near Swansea in Wales, which was originally built in 1791
but which has had numerous annexes added over the intervening years resulting in a complex roof
structure. Failure of the roof valley drainage system resulted in prolonged exposure of the roof
structure to moisture leading to wet rot in the span beams, hip rafters and ceiling joists.
An upgrade procedure was required that did not interfere with the ornate suspended ceiling. The
solution that was adopted was to replace the decayed timber in-situ with a prosthesis made from
laminated veneer lumber (LVL). Due to the restriction on access, the connection between the
prosthesis and the hip rafters was effected using slots routed in the sides of the prosthesis and the
rafter into which six 16 mm diameter high tensile steel rods were bonded using a two-part epoxy
structural adhesive. Fig. 29 shows the prosthesis in place with the rods inserted in the slots prior to
topping up with adhesive.
The span beams were repaired using an LVL prosthesis that was attached to the beams using six 20
mm diameter steel rods. The rods were bonded into holes drilled into the end grain of the beams and
into matching side slots in the prosthesis, as shown in Fig. 30. Due to the lighter loading carried by
the ceiling joists, it was sufficient to use a C24 softwood timber prosthesis that was bonded to the
joists using two 12 mm GFRP rods. Fig. 21 shows the rods inserted in the joists before adding the
prosthesis.
Fig. 29 Hip rafter repair [13] Fig. 30 Span beam repair [13]
Fig. 31 Long-term monitoring of Sins Bridge, [52]
20
4.2 Sins Bridge, Switzerland – Flexural reinforcement [51], [52]

Sins Bridge is a historic two-span timber arch bridge over Reuss River at Sins in Switzerland. It was
originally built in 1807 and the eastern side was rebuilt after being blown up during the 1852 Civil
War. It comprises two equal spans of 30.8 m and was designed for horse-drawn carriages. In 1992, it
was upgraded to carry 20 tonne trucks. This involved the installation of a new 200 mm thick
transversely prestressed timber deck and the strengthening of two transverse cross-beams with CFRP
laminates.
The cross-beams comprised two solid oak beams placed one on top of the other. The 1 mm thick
CFRP laminates were bonded to the top and bottom surfaces using epoxy resin (Fig. 31). Two types
of CFRP were used: One was a high modulus material (E = 305 GPa, tensile strength = 2300 MPa)
and the other was a high strength material (E = 152 GPa, tensile strength = 2600 MPa).
The reinforced cross-beams and a number of unreinforced crossbeams were instrumented with
electrical resistance strain gauges and Demec gauges in order to monitor their long-term performance
(Fig. 31). The deflection of the reinforced cross-beams was 20 - 50% lower than the unreinforced
beams. Fourteen years after the original installation, the performance of the reinforced beams
continued to be satisfactory [51].
5 Conclusion
Due to the impact of different aspects like moisture changes, fungi and insect attacks, timber elements
can be damaged and result in lower capacity and larger deformations. Furthermore, high stresses
exceeding the strength limits can also lead to different failure cases, like bending, compression,
tension or shear failure, where cracks mostly appear. An analysis of several assessment reports has
shown that most damaged structural timber elements show cracks in the grain direction.
In the case of damage or decay, the timber beam or parts of the beams have to be replaced or
reinforced in bending, shear, and in tension or compression perpendicular to the grain to recover the
performance. The choice of reinforcement method for existing timber beam structures will be based
not only on the ability of the reinforcement to provide adequate strengthening of the structure but will
be constrained by other factors such as aesthetics, need for reversibility, access for repair, and
available expertise.
As the reinforcing elements generally have different stiffness, thermal expansion and moisture
absorption properties than the timber element, factors such as constrained shrinkage and swelling due
to thermal or moisture changes must be considered and if necessary additional thermal or moisture
induced stresses should be accounted for in design. Where the reinforcement results in a significant
change in beam stiffness, it is important to consider the consequences for the overall behaviour and
load distribution of the entire structure.
Many factors have to be considered selecting the best method and sometimes lack of knowledge
exists. Ongoing research needs to be done to further improve the retrofitting and replacing methods
for reliable and efficient repair.
21
6 Acknowledgement
Parts of the report and research work are within the COST Action FP 1101 – Assessment, Monitoring
and Reinforcement of timber structures. The Swiss State Secretariat for Education, Research and
Innovation (SERI) proudly supports the research work.
7 References
[1] Blass, H. J., Frese, M., “Failure analysis on timber structures in Germany”, In: Kohler, J., Fink,
G., Toratti, T., Assessment of failures and malfunctions, Shaker Verlag, Germany, 2011.
[2] Blass, H. J., Frese, M., Failure analysis on timber structures (only in German:
Schadensanalsyse von Hallentragwerken aus Holz), KIT Scientific publishing, Karlsruhe, 2010.
[3] Franke, B., Scharmacher, F. Müller, A. “Assessment of the glue-line quality in glued laminated
timber structures”, Advanced Materials Research. Vol. 778, 2013, pp. 424-431.
[4] Dietsch, P., Gamper, A., Merk, M., Schopbach, H. „Versagensmechanismen im Holzbau“,
DVD, TU Munich, Germany, 2012.
[5] Zabel, R. A.,Morrell, J. J., Wood Microbiology: Decay and Its Prevention, Academic Press
Limited, London, UK, 1992.
[6] Aicher S., Radovic, B., Folland, G. Probability of decay by house longhorn beetle on glued
laminated timber (only in German: Befallswahrscheinlichkeit durch Hausbock bei
Brettschichtholz), IRB-Verlag, Germany, ISBN 978-3-8167-5977-5, 2001.
[7] Lehringer, C., Salzgeber, D., “Repair of cracks and delaminations in glued laminated timber”,
in Franke, B., Franke, S. Cost Workshop – Highly performing timber structures: Reliability,
Assessment, Monitoring and Strengthening, Switzerland, 2014.
[8] Wheeler, A.S., Hutchinson, A.R., “Resin repairs to timber structures”, International Journal of
Adhesion & Adhesives, Vol. 18, 1998, pp. 1-13, 1998.
[9] Tampone, G., “Mechanical failures of the timber structural systems”, In: Proceedings of
ICOMOS IWC - XVI International Symposium, Florence, Venice and Vicenza, 2007.
[10] Pizzo, B., Schober, K.U., “On site interventions on decayed beam ends”, In: Core Document of
COST Action E34, Bonding of timber, University of Natural Resources and Applied Life
Sciences, Vienna, 2008.
[11] Menichelli, C., Adami, A., Balletti, C., Bertolini Cestari, C., Bettiol, G., Biglione, G et al., Le
strutture lignee dell’arsenale di Venezia. Sudi et restauri., In Proceedings of XXV Convegno
Internazionale Scienza et Beni Culturali – Conservare e restaure il legno, Bressanone, Italy,
2009. (In Italian)
[12] Smedley, D., Cruz, H., Paula, R., “Quality control on site”, In: Core Document of COST Action
E34, Bonding of timber, University of Natural Resources and Applied Life Sciences, Vienna,
2008.
[13] Rotafix Ltd., “Clyne castle timber repairs. A case study.” Rotafix Limited, Swansea, UK
[14] Nowak , T.P., Jasienko, J., Czepizak, D., “Experimental tests and numerical analysis of historic
bent timber elements reinforced with CFRP strips”, Construction and Building Materials Vol.
40, 2013, pp. 197–206
[15] Kliger, R., Johansson, M., Crocetti, R., “Strengthening timber with CFRP or steel plates – short
and long-term performance”, In: Proceedings of World Conference on Timber Engineering,
Miyazaki, Japan, 2008.
[16] Mark, R. “Wood-aluminum beams within and beyond the elastic range. Part 1: Rectangular
sections”, Forest Products Journal, Vol. 11, No 10, 1961, pp. 477-484.
[17] Dziuba, T., “The ultimate strength of wooden beams with tension reinforcement”,
Holzforschung and Holzverwertung, Vol. 37, No. 6, 1985, pp. 115-119.
22
[18] DeLuca, V., Murano, C., “A comparison of un-reinforced and reinforced timber glulam with
steel bars”, European Journal of Technology and Advanced Engineering Research, Vol. 2,
2011, pp. 45-54.
[19] Nielsen, J., Ellegaard, P., “Moment capacity of timber reinforced with punched metal plate
fasteners”, In: Proceedings of 1st RILEM Symposium on Timber Engineering, Stockholm,
Sweden, 1999.
[20] Schober K.U., Rautenstrauch K., “Post-strengthening of timber structures with CFRPs”,
Materials and Structures, Vol. 40, 2006, pp. 27–35.
[21] Borri, A, Corradi, M, Grazini, A., “A method for flexural reinforcement of old wood beams
with CFRP materials”, Composites Part B: Engineering, Vol. 36, 2005, pp.143-153.
[22] Li, Y.F., Xie, Y.M., Tsai, M.J., “Enhancement of the flexural performance of retrofitted wood
beams using CFRP composite sheets”, Construction and Building Materials, Vol. 23, 2009, pp.
411-422.
[23] Hernandez, R., Davalos, J.F., Sonti, S.S., Kim, Y., Moody, R.C., “Strength and stiffness of
reinforced yellow-poplar glued-laminated beams”, Research Paper FPL-RP-554, Department
of Agriculture, Forest Service, Forest Products Laboratory, Madison, WI, US, 1997.
[24] Gentile, C., Svecova, D., Rizkalla, S.H.,, “Timber beams strengthened with GFRP bars:
development and applications”. Journal of Composites for Construction, Vol. 6, No. 1, 2002,
pp. 11-20.
[25] Raftery, G.M., Harte, A.M., “Low-grade glued laminated timber reinforced with FRP plate”,
Composites: Part B, Vol. 42, 2011, pp. 724-735.
[26] Raftery, G., Whelan, C., Harte, A., “Bonded-in GFRP rods for the repair of glued laminated
timber”, In: Proceeding of World Conference on Timber Engineering (WCTE12), Auckland,
New Zealand, 2012.
[27] Borri, A., Corradi, M., “Strengthening of timber beams with high strength steel cords”,
Composites: Part B. vol. 42, 2011, pp. 1480-1491.
[28] Peterson, J., “Wood beams prestressed with bonded tension elements”, Journal of the
Structural Division, ASCE, Vol. 91, No. 1, 1965, pp. 103-109.
[29] Triantafillou, T.C., Deskovic, N., “Prestressed FRP sheets as external reinforcement of wood
members”, Journal of Structural Engineering, Vol. 118, No. 5, 1992, pp. 1270-1283.
[30] Lehmann, M., Properzi, M., Pichelin, F., Triboulet, P., “Pre-stressed FRP for the in-situ
strengthening of timber structures”, In: Proceedings of World Conference on Timber
Engineering (WCTE 2006), Portland, OR, USA, 2006.
[31] Brunner, M., Schnueriger, M., “FRP-prestressed timber”, in: Proceedings of 8th International
Symposium on Fiber Reinforced Polymer Reinforcement for Concrete Structures (FRPRCS-8),
Patras, Greece, 2007.
[32] Brady, J.F., Harte, A.M., “Prestressed FRP flexural strengthening of softwood glue-laminated
timber beams”, In: Proceedings of World Conference on Timber Engineering (WCTE 2008),
Miyazaki, 2008.
[33] McConnell, E., McPolin, D, Taylor, S., “Post-tensioning of glulam timber with steel tendons”,
Construction and Building Materials, Vol. 73, 2012, pp.426-433.
[34] Blass, H.J., Bejtka, I., “Reinforcements perpendicular to grain using self-tapping screws”, In:
Proceedings of 8th World Conference on Timber Engineering, Lahti, Finland, 2004.
[35] Amy, K., Svecova, D, “Strengthening of dapped timber beams using glass fibre reinforced
polymer bars”, Canadian Journal of Civil Engineering, Vol. 31, No. 6, 2004, pp. 943-955.
[36] Jockwer, R., Frangi, A., Serrano, E., Steiger, R., “Enhanced Design Approach for Reinforced
Notched Beams”, In: Proceedings CIB – W18 Meeting 46, Vancouver, Canada, 2013.
[37] Widmann, R., Jockwer, R., Frei, R., Haeni, R., “Comparison of different techniques for the
strengthening of glulam members”, In: Proceedings of COST Action FP1004 Early Stage
Researchers Conference, Zagreb, Croatia, 2012.
23
[38] Blumer, H., Spannungsberechnungen an anisotropen Kreisbogenscheiben und Sattelträgern

konstanter Dicke, Lehrstuhl für Ingenieurholzbau und Baukonstruktionen, Universität
Karlsruhe, 1972/1979.
[39] Jonsson, J., “Load carrying capacity of curved glulam beams reinforced with self-tapping
screws”, Holz als Roh- und Werkstoff, Vol.63, 2005, pp. 342-346.
[40] Kasal, B., Heiduschke, A., “Radial reinforcement of curved glue laminated wood beams with
composite materials”, Forest Products Journal, Vol. 54, No. 1, 2004, pp. 74-79.
[41] American Institute of Timber Construction, Standard for radially reinforcing curved glued
laminated timber members to resist radial tension, AITC 404, Centenniel, CO, US, 2005.
[42] CEN, EN 1995-1-1: Eurocode 5: Design of timber structures – Part 1-1: General – Common
rules and rules for buildings, 2004, CEN, Brussels, Belgium.
[43] Akbiyik, A., Lamanna, A.J., Hale, W.M., “Feasibility investigation of the shear repair of timber
stringers with horizontal splits”, Construction and Building Materials, Vol. 21, No. 5, 2007, pp.
991-1000.
[44] Radford, D.W., Van Goethem, D., Gutkowski, R.M., Peterson, M.L., “Composite Repair of
timber structures”, Construction and Building Materials, Vol. 16, No. 7, 2002, pp. 417-425.
[45] Hallstrom, S., Grenestedt, J.L., “Failure analysis of laminated timber beams reinforced with
glass fibre composites”, Wood Science and Technology, Vol. 31, 1997, pp. 17-34.
[46] Gentry, T.R., “Performance of glued-laminated timbers with FRP shear and flexural
reinforcement”, Journal of Composites for Construction, Vol. 51, No. 5, 2011, pp. 861-870.
[47] Svecova, D., Eden, R.J., “Flexural and shear strengthening of timber beams using glass fibre
reinforced polymer bars-an experimental investigation”, Canadian Journal of Civil
Engineering, Vol. 31, No. 1, pp. 45-55.
[48] Trautz, M., Koj, C., “Self-tapping screws as reinforcement for timber structures”, In:
Proceedings of International Association for Shell and Spatial Structures Symposium, Valentia,
Spain, 2009.
[49] Dietsch, P., Kreuzinger, H., Winter, S., “Design of shear reinforcement for timber beams”, In:
Proceedings CIB – W18 Meeting 46, Vancouver, Canada, 2013.
[50] Crocetti, R., Gustafsson, P.J., Ed, D., Hasselqvist, F., “Compression strength perpendicular to
grain – full scale testing of glulam beams with and without reinforcement”, In: Proceedings of
COST Action FP1004 Early Stage Researchers Conference, Zagreb, Croatia, 2012.
[51] Meier, U., “Restoration of a historic wooden bridge with CFRP”, FRP International, 2(1) pp 1-
2 1993.
[52] Czaderski, C., Meier. U., “Long-Term Behaviour of CFRP Strips for Post-Strengthening.” In:
Proceedings of 2nd International fib Congress. Naples, Italy, June 5-8, 2006. Vol. 2. pp. 110-
112.
24
Reversible timber-to-timber strengthening interventions on wooden floors
2
Reversible timber-to-timber strengthening interventions on
wooden floors
Alessandra Gubana1
Summary
The goal of the State-of-the-Art-Report on High Reversible Strengthening Interventions on Timber
Floors is to give an almost complete overview of the new techniques developed and tested to achieve
in-plane and out-of-plane stiffness upgrading by means of less invasive and reversible interventions.
In recent years a growing sensibility towards the preservation and maintenance of heritage buildings
has led researchers to test different dry retrofitting systems. The chapter focuses on strengthening
interventions based on the use of timber or timber based elements: the most adopted or most
promising techniques are so briefly described.
While the problem of upgrading the bending stiffness has a solid analytical background in the
concrete-to-timber composite section theory, the problem of upgrading the in-plane stiffness still
requires attention. The importance of the correct evaluation of the in-plane mechanical properties of
floor timber diaphragms is nowadays clearly assessed in order to determine the building structural
response under lateral seismic loads. In some codes, simplified analytical procedures are proposed to
determine the in-plane stiffness, but generally with regard to new timber building floor typologies.
The tests described in literature are generally referred to different setups, test rigs, boundary
conditions, aspect ratio of the floor samples and also the recorded parameters are not always the same
ones: it is so difficult to compare the experimental data as to achieve a general stiffness evaluation
approach.
1. Introduction
Existing timber floors often require strengthening. The main problem is their low stiffness, which
results in high bending deformations and vibrations under service loads. Permanent deflection due to
creep can also reach critical values. Moreover, in earthquake prone areas, if seismic resistance has to
be assured in existing masonry buildings, floor diaphragm behaviour must be achieved. The
importance of an effective diaphragm action in the floors of multi-storey masonry buildings is well-
known in earthquake engineering. Thanks to the diaphragm action, in fact, the floors can transfer the
forces due to earthquakes to the lateral load resisting systems [1].
One of most widely used and effective techniques for strengthening floors is based on the connection
of a new 40-50 mm height concrete topping on the existing timber joists. Different types of metallic
fasteners, notched shear keys or slotted in perforated plates generally assure the effective
collaboration between the two different materials [2,3,4,5,6,7,8,9].
1)
Associate Professor, University of Udine, Udine, Italy
25
The new composite section ensures a significant floor stiffness upgrade, while the concrete topping
connected with the vertical walls is able to give an effective three dimensional behaviour of the
masonry buildings and therefore it markedly improves the lateral load resistance. The use of concrete
also allows load distribution to take place and provides acoustic and fire insulation. On the other hand,
a thin concrete slab adds undesirable additional weight on the floor, and consequently the seismic
actions and the foundation loads increase.
This technique, while simple and very efficient, is now often considered not sufficiently reversible: in
particular in Italy it is frequently not authorized by the Cultural Heritage Offices, therefore it may be
not allowed in buildings of historical value.
The increased sensitivity towards the conservation of cultural heritage leads to the adoption of restoration
techniques which can guarantee as much as possible the maintenance of the building authenticity and
integrity, the conservation of the materials and of the original function of the structure, the reversibility of
the intervention and its compatibility with the existing parts of the buildings [10,11,12,13].
Various solutions have been studied in recent years with the aim of developing less invasive and
reversible techniques, based for example on the use of steel profiles or mortar topping [14]. The paper
focuses particularly on strengthening interventions based on the use of timber or timber based
elements, dry connected to the existing floors, with the aim of increasing the bending or in-plane
stiffness.
2. Timber-to-Timber Bending Stiffness Upgrade Interventions
Bending stiffness upgrade can be achieved by connecting timber planks or Cross Laminated Timber
panels to existing floor joists. In this way, it is possible to rely on a composite T section beam. In
timber-to-timber composite sections, the use of traditional materials and dry assembly methods are in
agreement with the restoration issues of compatibility, reversibility or recoverability of the
intervention. For all composite sections the mechanical characteristics of the connection are the main
factor which influences the structural response.
The design of timber-to-timber composite sections requires consideration of partial composite
behaviour, due to the deformable shear connection between web and flange. Analysis can follow the
Eurocode 5 [15] approximate ‘gamma method’, where an effective flexural stiffness (EI)ef for the
composite section is calculated as a function of the stiffness of the shear connection, taking into account
the slip between the flange and the joist. Also the ‘shear analogy method’, where the composite beam is
divided into two virtual components coupled with stiff bars can be used to determine internal forces
[16,17,18].
Simplified equations to evaluate the increment, with respect to complete connection, of the camber, of
the central vertical displacement and of the bending stress due to maximum slip at interface can be
found in [14]. For example the increment of the central sag, for ordinary values of the ratio L/H (18÷25),
where L is the length of the beam and H is the height of the cross section, can be estimated with:
w  10 (1)
where w is the increment of the vertical deflection and  is the maximum slip at the interface.
Several connection systems have been proposed for concrete-to-timber systems [19] and some of
them have been tested for timber-to-timber solutions. New connection systems have also been
proposed and different elements for upper flanges have been tested, with good performance in any
case.
26
2.1 Timber flanges connected to main beams by hardwood dowels

The reinforcing technique consists of placing a new plank above each floor joist, connected by means
of dowels (Fig. 1) made of hardwood pins [20]. With proper design of the connection of the beams to
the load bearing walls and additional planks in the upper surface, the system can work also as an
effective diaphragm (Fig. 1b) [21]. A composite timber-to-timber T beam is thus obtained with
deformable connection between the flange and the web. These may or may not be separated by the
existing floor board (Fig. 1d). The presence of the floor boards influences the dowel collapse
mechanisms and consequently the beam failure: the best performance was obtained when boarding
was present.
Fig. 1 Stiffening intervention made with new planks connected to existing beams by means of dry
hardwood pins (from [20])
2.2 Timber flanges connected to main beams by self-tapping screws

Among the different types of possible connectors available, the use of inclined self-tapping screws
seems to give promising results, based on experimental testing of timber-to-timber connections
[22,23,24].
In [25,26] a strengthening intervention on an ancient floor of a castle is described. The existing beams
were coupled with glulam planks strength class GL 24h, 80 mm in thickness. A T-beam compound
section was so obtained, with deformable connections between the flange and web. The web and
flange were separated by a new boarding, with a thickness of 30 mm, which replaced the existing
decayed one (Fig. 2).
The connection system consisted of self-tapping double thread screws, steel grade 10.9, X-crossed at
an angle α = 45˚.
A final load test on a portion of the reconstructed floor showed good performance and good
agreement with the design data.
27
Fig. 2 Diagrams of a) the transversal section, b) longitudinal section of the composite structure, and
c) detail of the geometry of the connection system. (i) Beam interspace; (ls) screw length; (s) spacing
of the fasteners (from[26])
The slip modulus of the connection was determined according to the calculation method proposed
in [22,23,24], on the basis of the results from a series of push-out tests for different screw inclinations
and joint configurations.
2.3 Timber-to-CLT Composite Section
Cross laminated timber panels (XLam or CLT), which has been used to build walls and floor slabs in
new timber buildings for the last 15 years, can be efficiently adopted also in restoration interventions
to obtain a diaphragm effect under seismic actions, by connecting them to the existing timber beams
and to the perimeter walls.
Cross laminated timber panels consist of a sequence of layers made of boards, glued one to the other.
Experimental tests showed they are stiff enough in their plane to assure a diaphragm effect and to
resist the common lateral forces due to earthquakes. In this way the superposition of XLam panels
connected to the existing timber joists can substitute for concrete topping as a less invasive, lighter
and more reversible solution [27].
Structural XLam panels are generally made of a series of layers and can be over 120 mm thick.
Special panels of 60 mm thickness have been specially produced for use in strengthening
interventions, as in restoration it is important not to vary the existing floor levels too much.
A series of experimental tests has been conducted to investigate the behaviour of the composite
section beams. Ten specimens made by XLam panels and glulam beams, six connected together by
means of predrilled steel dowels (T1A÷T6A) [27] and four by self-tapping screws (T1B÷T4B) [28],
were tested under bending loads up to collapse. The screws were inserted with a 90 degree angle with
respect to the beam axis (Fig. 3). All the beams were loaded over 2 points at a distance corresponding
to 1/3 of the beam length. The increasing of the load was checked by displacement control. The data
acquisition was automatically carried out. The instrumentation consisted of 10 potential transducers of
1/1000 mm precision, in order to monitor the vertical deflection in the middle of the beam, the
rotation and the slip at interface near the supports (Fig. 4).
28
Fig.3 Predrilled steel dowel or screw spacing along the beam (from [27,28])
Fig.4 Load configuration and monitoring device set up (mm)(from [27,28])
As for all types of composite section, the behaviour depends on the connection stiffness and it ranges
between the two limit states of composite section without connection (EJ)0 and composite section
with infinite stiffness connection (EJ)∞. The diagrams (Fig. 5) show a notable increase in stiffness.
The use of self-tapping screws as connectors leads to a greater initial stiffness compared with
predrilled steel dowels, as they do not need an initial slip to come in contact with the surrounding
timber [27]. Moreover, there is also a friction effect, due to screws, which gives an initial stiffness
equal to the case of infinite stiffness connection up to 10% of the collapse load.
XLam panels can simultaneously provide diaphragm action as they have significant shear stiffness in
their plane.
Fig. 5 Load-Deflection diagrams of XLam-to-Timber composite section beams
29
2.4 General Remarks

All these interventions have to face the problem of the thickness of the cross section after the
interventions, as in restoration it is important not to vary too much the existing floor levels, otherwise
problems can arise with the window and the door thresholds, with the level of the stairs or with
eventual wall decorations. From this point of view the use of CLT panels of only 6 cm height, or
eventually less, can be an interesting solution.
The efficiency coefficient of the composite section η [7] is a significant parameter to evaluate the
capacity of the connection in order to limit slip between the two parts of the composite section. The
expression of the coupling coefficient η is defined by the following equation:
(2)
where
(EJ) actual bending stiffness, (EJ)0 bending stiffness of the composite section without connection,
(EJ)∞ bending stiffness in case of complete connection.
The results presented in literature show that the typical values of concrete-to-timber cross sections
( [29] or more recently up to 0,9) can be reached also by means of timber-to-timber
techniques: the experimental maximum value obtained by composite CLT to timber beam section was
around =50%, even if numerical analyses show that with different connector spacing it can increase
[28], while the use of inclined self-tapping screws can upgrade the coupling coefficient values
(=0,74 [26]).
If the existing beams have significant cambers due to permanent loads and creep, some problems can
arise during the execution of the intervention, as the bottom of the new element and the upper part of
the existing joist do not match. In these situations compensation board can be needed or forced
shoring can be eventually tried.
3. In-plane stiffness upgrade by using timber or timber based elements.

An effective diaphragm action in the floors of multi-storey masonry buildings is particularly
important because the forces due to earthquakes can be transferred to the lateral load resisting walls.
One of the first documents to propose the in-plane stiffness upgrade of timber floors by means of
timber elements was produced by the local authority of Friuli Venezia Giulia Region (Italy) after 1976
earthquake [30], in fact, at the time, Italian national codes did not provide guidance on strengthening
interventions design on existing masonry buildings damaged by earthquakes.
The suggested technique consisted of superposing a second layer of floorboards over the existing one,
but laid in the orthogonal direction (Fig. 6). This multiple layers of floorboards was a technique
commonly diffused in ancient buildings in the earthquake prone areas of the region.
The importance of the connection of the floor with the perimeter walls was underlined by the drawings of
possible details of the joining of the floorboards with walls (Fig. 7) and of the joists with walls (Fig. 8).
30
New layer of floorboards
Existing layer of floorboards
Fig. 6 New layer of floorboards over the existing one (from [30])
New layer of floorboards

Steel plate
Nails
Exisisting layer of floorboards
almost
Masonry wall to stiffening layer connection
Fig. 7 Detail of the connection of the floorboards with the wall (from [30])
Nails spacing
Masonry
wall
Stiffening layer
of floorboards
Existing layer
of floorboards
Masonry
wall
Fig. 8 Detail of the suggested connection of the joist with masonry wall (from [30])
A series of full scale tests to measure experimental in-plane stiffness is described in [31] and [32]. In
these tests, two samples of unstrengthened timber floor elements, having a common typology of
mono-directional floors (FMSB and FM samples), were subjected to five types of strengthening
interventions. The interventions involved the use of diagonal punched metal strips (FMSD), of a
single layer of diagonal planks placed at 45° with respect to the original floorboards with common or
tongue and groove boards (FM SP(A) and FM SP(B)), of a double diagonal layer placed at ± 45°
made with tongue and groove boards (FM± 45°DP(A)), of one diagonal plank equivalent to the
punched steel strip in term of axial stiffness (FMWD(D)) and finally of a double diagonal made with
thick planks (FMWD(E)). As it can be seen in Fig. 9, the monotonic tests on reinforced floors showed
that the best result was achieved by the use of a double board layers. The stiffness increase is
significant with respect to the traditional existing floors.
31
Fig. 9 Force versus top displacement behaviour for the unreinforced and reinforced specimens (from
[32])
Fig. 10 Specimens used in the full-scale tests (from[33])
In [33], five full scale timber floors were tested to analyse the in-plane behaviour: one was made with
timber beams and floorboards only (Fig. 10 S-specimen), the second was strengthened by a second
layer placed orthogonal to the first board (Fig. 10 S-specimen) and the other three by two (Fig. 10
CLT2-specimen) or three CLT panels (Fig. 10 CLT3.1 and CLT3.2). In Fig. 11 the test setup is
described. Test results show that a significant increase in the in-plane stiffness, up to 5-10 times than
that of the unstrengthened specimen, can be achieved, so the efficacy of the intervention was
experimentally demonstrated. Numerical modelling showed that the main influence on the in-plane
32
behaviour was due to the screw connection in the direction perpendicular to the axis of the floor
beams.
Fig. 11 Setup of the floor test (from [33])
In [34] a full-scale experimental program consisting of tests on four as-built diaphragms and four
retrofitted diaphragms in both principal loading directions is presented. As-built configurations were
typical of those found in historic unreinforced masonry buildings in North America and Australasia,
whereas retrofitted diaphragms consisted of plywood panel overlays with stapled sheet metal blocking
systems (SMBS). The nonlinear and low stiffness behaviour of the as-built diaphragms was confirmed
in each principal loading direction. The plywood overlay and SMBS dramatically improved the as-
built diaphragm shear strength and shear stiffness and were shown to perform satisfactorily from a
serviceability perspective.
Fig. 12 Setup of the floor test for parallel to joist loading (a) and perp. to joist loading (from [34])
The tests were performed for loading parallel to joists and perpendicular to joists (Fig.12). Testing in
33
both principal loading directions confirmed the orthotropic nature of timber diaphragms. While the
shear strength remained consistent for as-built diaphragms, the shear stiffness in the direction
perpendicular-to-joists was up to 32% less than the corresponding value in the orthogonal
configuration. For retrofitted diaphragms, the difference in shear stiffness increased to 60%, and the
shear strength in the direction perpendicular-to-joists was almost 50% of the shear strength parallel-
to-joists.
In [35] ten floor specimens were tested, including five floors representing an as-built configuration
and five floors retrofitted through the addition of a plywood layer on the top of the floor boards.
Square-edge straight pine flooring boards (25 × 150 mm) were nailed to the joists with two standard
nails (3.15 mm diameter) at each joist intersection. The flooring board layout is shown in Fig. 13.
Timber floor boards were staggered meaning that the floor was composed of a combination of 2 m
and 1 m long boards. For five specimens, plywood panel overlays were screwed to the original
structure.
Fig. 13 Flooring board layout (from [35])
In Fig. 14 the load-displacement curve related to the specimen R-1 underlines that the effect of the
plywood panel overlays results in a significant increase in both diaphragm strength and stiffness; the
maximum load reached during the test is equal to 150 kN, equal to three times that for the as-built
specimen. The curve shows a strong nonlinearity after the first load cycle, proving that the diaphragm
behaviour is strongly affected by the connection between the timber elements (nails and screws). The
failure occurred principally in the screws between panels and boards (short screws) and between
panels and joists (long screws).
Fig. 14 Load versus displacement curves for as-built (AB-1) and retrofitted specimens R-1 (from[35])
34
3.1 In-plane stiffening of wooden floors by means of nailed plates connecting the
planks.
The in-plane stiffness of wooden floors can be increased by using punched metal plates (gang nails) to
connect the boards longitudinally, so as to prevent the slip due to shear forces (Fig. 15). Different
fasteners may be used as nail plates embedded into the wooden board by using special devices or
knuckle nail plates, applied by hammering in the “knuckle nail” [36]. The technique requires the use
of an appropriate device capable of ensuring that the nail plate achieves complete penetration of the
nails in the wooden boards, such as an hydraulic jack arranged on a special steel frame. The hammer
application technique is simpler, but it may be less effective.
A perimeter L-shaped steel profile is used to increase the floor resistance: the steel profiles
perpendicular to the horizontal action constitute the two chords needed to resist the bending moment,
the stiffened wooden floor supports shear forces and the steel profiles parallel to horizontal action
transfer the in-plane forces to the shear walls.
Fig. 15 Schematic view of the floor (a) and gang nails connecting the boards (b) (from [36])
4. Conclusions
The recent literature on the problem of bending and in-plane floor stiffness upgrading shows that there
is an increasing sensibility towards the problem of reversible and less invasive techniques of
restoration. Several papers on these problems have been presented with an evident increase in number
in the last two to three years.
The design rules for upgrading the bending stiffness by means of timber-to-timber dry techniques are
based on the composite section theories developed for concrete-to-steel and concrete-to-timber
composite sections. As new types of connectors and new types of upper flanges have been introduced,
experimental tests have been performed to check the applicability of the codified formulations or to
validate the new proposed ones. As in all types of composite sections, one of the main parameters
governing the structural answer of the beam is the connection load-slip relationship.
The importance of the correct evaluation of the in-plane mechanical properties of floor timber
diaphragms to determine the building structural response under lateral loads due to earthquakes is
addressed. In some codes, simplified analytical procedures are proposed to determine the in-plane
stiffness, but generally with regard to new timber buildings floor typologies. In [37] design details are
given to obtain a rigid floor, while in [38] formulas are presented to evaluate the different components
35
of the diaphragm horizontal deflection, which are due to nail slip, to flexural deformation and to shear
deformation in the sheathing.
More experimental tests are necessary to check the reliability of the analytical models proposed, to
develop new ones or to complete an exhaustive experimental database. Moreover, the tests described
in literature papers are generally referred to different setups, test rigs, boundary conditions, aspect
ratio of the floor samples and recorded parameters and so it is difficult to compare the experimental
results as to achieve a general stiffness evaluation approach.
5. Acknowledgements
The present paper is a result of the activities of COST Action FP1101 Assessment, Reinforcement and
Monitoring of Timber Structures.
6. References
[1] Tomaževič M., Earthquake-resistant Design of Masonry Buildings, Imperial College Press,
1999.
[2] Turrini G., Piazza M., “Una tecnica di recupero dei solai in legno”, Recuperare, No.5, 1983, pp.
396-407.
[3] Ronca P., Gelfi P., Giuriani E., “Behavior of a wood–concrete composite beam under cyclic and
long term loads”, In: Proceedings of the International Conference on Structural Repair and
Maintenance of Historic Buildings, Seville, Spain, 1991.
[4] Natterer, J., Hamm, J., Favre P., “Composite wood–concrete floors for multi-story buildings”,
In: Proceedings of the 4th International Wood Engineering Conference, New Orleans,
Omnipress, Madison, Wisconsin, USA, 1996.
[5] Gelfi, P., Giuriani, E., Marini A., “Stud shear connection design for composite concrete slab and
wood beams”, Journal of Structural Engineering, 2002, 128(12), pp.1544-1550.
[6] Giuriani, E., “L’organizzazione degli impalcati per gli edifici storici”, L’edilizia 134:30-43,
2004.
[7] Gutkowski, R., Brown, K., Shigidi,A., Natterer J., “Laboratory tests of composite wood–
concrete beams”, Construction and Building Materials, Vol. 22. No.6, 2008, pp.1059-1066.
[8] Bathon, L., Graf, M., “A continuous wood-concrete-composite system”, In: Proceedings of
World Conference of Timber Engineering, Whistler, BC, 2000.
[9] Clouston, P., Bathon, L., Schreyer, A., “Shear and Bending Performance of a Novel Wood–
Concrete Composite System”, Journal of Structural Engineering, Vol. 131, No. 9, 2005, pp.
1404–1412.
[10] Venice Charter, Second International Congress of Architects and Technicians of Historical
Monuments, Venice, May 25-31, 1964.
[11] Krakow Charter, International Conference on Conservation Krakow 2000, Krakow, 2000.
[12] ISCARSAH-ICOMOS, Principles for the analysis, conservation and structural restoration of
architectural heritage, 2003.
[13] UNI Ente Nazionale Italiano di Unificazione, UNI 11138 - Cultural Heritage - Wooden
artefacts - Criteria for the preliminary evaluation, the design and the execution of works,
Milano (Italy), 2004.
[14] Giuriani E., Consolidamento degli edifici storici, UTET, Torino (Italy), 2012.
[15] EN 1995:2004, Eurocode 5 Design of Timber Structures, 2004.
[16] Kreuzinger, H., “Platten, Scheiben und Schalen: Ein Berechnungsmodell für gängige
Statikprogramme“, Bauen mit holz, Vol. 1, 1999, pp. 34-39.
[17] Deutsches Institut für Normung: DIN 1052; Entwurf, Berechnung und Bemessung Deutsches
Institut für Normung: DIN 1052; Entwurf, Berechnung und Bemessung, 2004.
36
[18] Kuhlmann, U., Michelfelder, B., “Optimised design of grooves in timber-concrete composite
slabs”, In: Proceedings of the 10th World Conference on Timber Engineering, Portland, Oregan,
USA, 2006.
[19] Yeoh, D., Fragiacomo, M., De Franceschi, M., Heng Boon K., “State of the art on timber-
concrete composite structures: Literature Review”, Journal of Structural Engineering, Vo. 137,
No. 10, 2011, pp.1085-1095.
[20] Valluzzi, M.R., Garbin, E., Modena, C., “Flexural Strengthening of timber beams by traditional
and innovative techniques”, Journal of Building Appraisal, Vol 3, No. 2, 2007, pp.125-143.
[21] Modena, C., Valluzzi, M.R., Garbin, E., da Porto, F., “A strengthening technique for timber
floors using traditional materials”, In: Proceedings of the Fourth International Conference on
Structural Analysis of Historical Constructions SAHC 04, Padova, Italy, 2004.
[22] Bejtka, I., Blaß, H.J., “Screws with continuous threads in timber connections”, In: International
RILEM Symposium on Joints in Timber Structures, Stuttgart, Germany, ed. S. Aicher and H.W.
Reinhardt, pp.193-201, 2001.
[23] Bejtka, I., Blaß, H.J., “Joints with inclined screws”, In: Proceedings of Meeting 35 of the
International Council for Building Research Studies and Documentation, CIB, Working
Commission W18 – Timber Structures, Kyoto, Japan, CIB Paper No.35-7-4, 2002.
[24] Tomasi, R., Crosatti, A., Piazza, M., “Theoretical and experimental analysis of timber-to-timber
joints connected with inclined screws”, Construction and Building Materials, Vol. 24, 2919, pp.
1560-1571.
[25] Angeli, A., Piazza, M., Riggio, M., Tomasi, R., “Refurbishment of traditional timber floors by
means of wood-wood composite structures assembled with inclined screw connectors”, In:
Proceedings of 11th World Conference on Timber Engineering WCTE 2010, ed. A.Ceccotti and
J.W. Van de Kuilen, Riva del Garda, Italy, 2010.
[26] Riggio, M., Tomasi, R., Piazza, M., “Refurbishment of a traditional timber floor with a
reversible technique: importance of the investigation campaign for design and control of the
intervention”, International Journal of Architectural Heritage, Vol. 8, 2013, pp.74-93.
[27] Gubana, A., “Experimental tests on Timber-to-Cross Lam composite section beams”, In:
Proceedings of 11th World Conference on Timber Engineering WCTE 2010, ed. A.Ceccotti and
J.W. Van de Kuilen, Riva del Garda, Italy, 2010.
[28] De Cillia, L., “Experimental and numerical analysis of Timber to XLam composite section
beam” (in Italian), Master of Civil Engineering Thesis, University of Udine, Gubana A.
Supervisor, 2013.
[29] Piazza, M., Tomasi, R., Modena, R., Strutture in legno, Hoepli, Milano (Italy), 2005.
[30] Regione Autonoma Friuli-Venezia Giulia - Segreteria Generale Straordinaria: Legge Regionale
20 giugno 1977, n. 30 - Recupero statico e funzionale degli edifici. Documento tecnico n.2
DT2: Raccomandazioni per la riparazione strutturale degli edifici in muratura. Gruppo
Disciplinare Centrale, Maggio, 1980.
[31] Valluzzi, M.R., Garbin, E., Dalla Benetta, M., Modena, M., “Experimental Assessment and
modeling of in-plane behaviour of Timber Floors”, In: Proceedings of the VI International
Conference on Structural Analysis of Historic Construction, SAHC 08, Bath, UK, ed. D.
D’Ayala and E. Fodde, 2008.
[32] Valluzzi, M.R., Garbin, E., Dalla Benetta, M., Modena, M., “In-plane Strengthening of Timber
floors for the seismic improvement of masonry buildings”, In: Proceedings of 11th World
Conference on Timber Engineering WCTE 2010, ed. A. Ceccotti and J.W. Van de Kuilen, Riva
del Garda, Italy, 2010.
[33] Branco, J.M., Kekeliak, M., Lourenço, P.B., “In Plane Stiffness of traditional Timber Floors
Strengthened with CLT”, Materials and Joints in Timber Structures, ed. S. Aicher et al., RILEM
Book series 9, pp.725-737, 2014.
[34] Wilson, A., Quenneville, P.J.H., Ingham, J.M., “In-Plane Orthotropic Behavior of Timber Floor
Diaphragms in Unreinforced Masonry Buildings”, Journal of Structural Engineering, Vol. 140,
No. 1, 2013, 04013038:1-11.
[35] Brignola, A., Pampanin, S., Podestà, S., “Experimental Evaluation of the In-Plane Stiffness of
37
Timber Diaphragms”, Earthquake Spectra, Vol. 28, No. 4, 2012, pp. 1687-1709.
[36] Gattesco, N., Macorini, L., “High reversibility technique for in plane stiffening of wooden
floors”, In: Proceedings of the VI International Conference on Structural Analysis of Historic
Construction, SAHC 08, 2-4 July 2008, Bath, UK, ed. D. D’Ayala and E. Fodde, 2008.
[37] EN 1998-1:2005, Eurocode 8 Design of structures for earthquake resistance, 2004.
[38] NZS 3603, New Zealand Timber Structures Standards, 1993.
38
Reinforcement of timber columns and shear walls
3
Wen-Shao Chang1
Summary
This chapter provides an overview of state-of-the-art repairing and reinforcing techniques on timber
columns and shear walls in both research and practice. It covers two levels of intervention, repair and
reinforcement of timber elements. The former focuses on damaged elements and the latter focuses on
enhancing the mechanical properties of the elements. Although it was found that most of the research
foci were on reinforcement of timber connections and flexural members, columns and shear walls
play a crucial role in the prevention of structural collapse. With the future development of taller
timber structures, these issues will become more and more important.
1. The need to reinforce/repair timber columns and walls

A column is a member in a structure that takes vertical load and sometimes bending moment
transferred from a beam via connections. It is crucial to the stability of a structure. A timber shear
wall is an important structural element that provides lateral stability to the structure and resists the
horizontal forces, such as earthquake and wind. They provide substantial in-plane stiffness and only
limited out-of-plane stiffness. The reasons to reinforce timber shear walls are: (1) to enhance stiffness
and strength; (2) to improve ductility; and (3) to increase energy dissipation capacity. Note that in this
chapter; only shear walls made of timber will be discussed. For example, in some half-timber framed
structures, stones, bricks (Fig. 1) and wattle and daub (Fig. 2) are often used as in-fill elements, and
therefore are outside of the scope of this chapter.
There are a number of situations where a column and a shear wall in a building need to be repaired or
reinforced. These include biodeterioration, mechanical failure, cracks, and the need for higher
strength.
1.1 Biodeterioration
Columns, when touching the ground without any measure to isolate them from damp, are prone to
elevated moisture content levels which will lead to bio-deterioration due to insects (such as termites)
and fungal attacks. This is a common form of decay that can be found where the column touches the
ground (Fig. 3). When designing a timber column, one should select the timber carefully as the most
common form of deterioration is from attack of the sapwood by insects, while the heartwood remains
untouched [1]. The rise of moisture content in a column will lead to fungal defects and attract termites
to attack the member, and these are often unseen as shown in Fig. 4. This failure mode in a column
not only reduces the mechanical properties of timber but also reduces the effective section area.
1)
PhD, Department of Architecture and Civil Engineering, University of Bath, UK
39
Fig. 1 Half-timber frame with brick infill Fig. 2 Half-timber frame with wattle and daub
infill
1.1 Biodeterioration
Columns, when touching the ground without any measure to isolate them from damp, are prone to
elevated moisture content levels which will lead to bio-deterioration due to insects (such as termites)
and fungal attacks. This is a common form of decay that can be found where the column touches the
ground (Fig. 3). When designing a timber column, one should select the timber carefully as the most
common form of deterioration is from attack of the sapwood by insects, while the heartwood remains
untouched [1]. The rise of moisture content in a column will lead to fungal defects and attract termites
to attack the member, and these are often unseen as shown in Fig. 4. This failure mode in a column
not only reduces the mechanical properties of timber but also reduces the effective section area.
Fig. 3 Bio-deterioration in a column that has Fig. 4 Timber strut attacked by termites
contact with the ground
1.2 Mechanical failure

Compared with beam members, creep is less onerous in a column member. A column normally takes
only vertical load; in some circumstances it will take combined compression and bending. The former
will lead to buckling of the column, whilst the latter will result in partial yielding or split along the
grain as shown in Fig. 5. Slender compression members are susceptible to buckling. When a
compression member has (1) insufficient section size; (2) vertical cracks so the effective section is
reduced; or (3) low material strength, it is prone to buckle. The buckling of a compression member is
40
a failure that often occurs without warning. It is therefore important to consider whether compression
members within a structure are highly stressed, and if any action needs to be taken to ensure the
prevention of the column from buckling.
1.3 Cracks
Cracks occurring in timber members often result from the differences between the drying speed in
interior layers and outer ones. The drying stresses will be built up if the outer layers are dried to a
level that is much lower than the fibre saturation point while the interior is still saturated [2]. Rupture
in timber occurs and in consequence cracks occur if the drying stress exceeds the strength
perpendicular to the grain as shown in Fig. 6.
1.4 Need for higher strength

Recently there has been a trend all over the world to strive for higher timber constructions. Two mid-
rise timber apartments had been completed in London, UK prior to 2011. Another 10-storey timber
apartment was completed in 2012 in Melbourne, Australia. Further tall timber buildings are at the
planning stage and therefore we will see more and more tall timber buildings in the future. To achieve
taller timber buildings, we need timber products with higher strength, in particular those which will be
used in the lower parts of the buildings. As such, we also need timber columns and shear walls to have
higher strength in order to resist the self-weight built up when the buildings go higher.
2. Repair and reinforcement of timber columns

2.1 Prosthesisation
When dealing with historic buildings, engineers and architects need to balance authenticity of the
structures after renovation/repair and assurance of the strength of the structure to carry the load
needed. To minimise the amount of timber being replaced, prosthesisation has become common
practice when the timber members are bio-deteriorated due to termites or insects. It is a method that
replaces only the decayed or failed part with a new portion. Timber used for prosthesisation, in
particular for the conservation of historic buildings, must be carefully selected so that the nature of the
new timber will match that of the old. The moisture content of the timber being used should be close
to that of the existing members so that moisture movement can be avoided. Fig. 7 shows an example
of a column being prosthesised after it was partially damaged. When a new prosthesis is adopted to
replace the damaged portion in a timber member, two methods exist to connect the old and new
members: (1) local and traditional carpentry as shown in Fig. 8; and (2) glued-in members for the
connection. For both cases, modern adhesives are often used to ensure the continuity of the new
column. Although prosthesisation is common practice nowadays in historic building conservation in
many countries, there is a lack of research work on this method.
2.2 Screw reinforcement
Song et al. [4] carried out a series of tests to study the effect of self-tapping screws to repair timber
columns with vertical cracks and compared that with timber columns repaired by Fibre Reinforced
Polymer (FRP) pads. The vertical cracks were simulated by making slots in the column with a width
of 6 mm and a length of 1500 mm. Vertical load was applied on the columns with pin connections at
both ends. The conditions for the specimens are shown in Tab. 1 and the specimen design is shown in
Fig. 9. The failure modes of each specimen are shown in Fig. 10.
41
Fig. 5 A column damaged due to an earthquake Fig. 6 Vertical cracks occur in a column
Fig. 7 A new timber component was used to Fig. 8 Partial replacement repair in Daibei Temple
partially replace rotten column with traditional (1550), China [3]
carpentry
It was observed from the tests that the

maximum loading capacity of Specimen 2
(cracked and unrepaired) was more than 30%
lower than that of the intact column (Specimen
1), and this shows that the vertical crack will
weaken the column. The experimental results
also showed that the self-tapping screws will
improve the strength of the cracked specimen
to a level close to the intact ones. The
additional work of filling the crack in a column
does not affect the strength of the cracked
column. The strength of the cracked column
repaired by FRP pad showed similar results Fig. 9 Specimen design of repairing cracked timber
those repaired by self-tapping screws. column by using screws and FRP pads [5]
42
Tab. 1 Specifications of column and experimental results by Song et al (data source: [5]).
No. Dimensions (mm) Slotted Filled Retrofit Diameter/width1 Spacing Ultimate

(mm) (mm) strength (kN)
1 200 x 200 x 1800 N ̶ ̶ ̶ ̶ 846
2 200 x 200 x 1800 Y N ̶ ̶ ̶ 571
3 200 x 200 x 1800 Y N STS 6 250 736
4 200 x 200 x 1800 Y N STS 6 250 895
5 200 x 200 x 1800 Y N STS 6 250 675
6 200 x 200 x 1800 Y Y STS 6 250 812
7 200 x 200 x 1800 Y Y FRP 100 200 835
Note: 1 diameter for screws and width for FRP sheets
Fig. 10 Failure mode of columns reinforced by different strategies tested by Song et al. [5]
This study shows self-tapping screws to be a good repairing measure; in particular because it is
reversible, i.e. the self-tapping screws can be removed in the future once more efficient ways of
repairing timber columns have been developed. More work needs to be done on investigating factors,
such as the dimensions of the cracks and the spacing of the screws, before this method can be widely
implemented.
2.3 Steel member reinforcement
In the early stages of reinforcement and repair of timber structures, the focus was mainly on using
metallic reinforcement, such as steel bars and plates. However, the focus was also mainly on beam
43
elements and connections; efforts being devoted to the reinforcement of columns were relatively
scarce. Tanaka et al. compared the effect of a column reinforced by steel plate with that of one
reinforced by carbon fibre sheets [6]. Buckling tests were carried out, and the parameters considered
included (1) slenderness ratio of column, (2) boundary conditions for steel plates in the reinforced
column, and (3) reinforcement methods (steel plates and carbon fibre sheets). The sections of the
specimens of the experiments are shown in Fig. 11 and the reinforcement arrangements are depicted
in Fig. 12.
Fig. 11 Section of the columns reinforced by steel Fig. 12 Different reinforcement arrangements [6]
plates and carbon fibre sheets [6]
The experimental outcomes showed that steel plates reinforced timber columns have load-carrying
capacities at least 2.5 times higher than that of unreinforced timber columns, whilst columns
reinforced by carbon fibre sheets exhibit 1.3 times higher load-carrying capacities than unreinforced.
2.4 Composite material reinforcement
Repair and reinforcement of the damaged timber members by composite material, such as FRP, has
been developed over more than 2 decades. FRP has a remarkable strength-to-weight ratio and leads to
light weight strategies when repairing or reinforcing these damaged members. Substantial amounts of
effort have been devoted to investigating increasing the strength properties of intact timber members
after the application of FRP bonded externally [7-9]. Zhang et al. carried out a series of tests on
repairing cracked columns by using FRP wrapping and developed finite element models to simulate
the behaviour for parametric studies [10]. The factors considered include (1) the column dimensions,
(2) the crack dimensions, (3) whether the crack was filled, (4) FRP properties and (5) FRP spacing. A
total of 17 specimens were tested and six different failure modes were observed. Fig. 13 shows the
specimens tested and factors considered. The experimental results showed that different combinations
of factors, in particular the FRP spacing, will result in different failure modes. It was evidenced in the
series of tests that the load-carrying capacity of a column decreases with increase in the length and
width of the cracks and the influence of the crack width is more significant. It was also observed that
reducing the FRP spacing will increase the recovery of load-carrying capacity of cracked timber
columns.
Oprişan et al. shows different methods of using FRP composite to strengthen a timber column. They
include: (1) FRP fabric with different fibre orientations; (2) FRP strips to provide confinement; (3)
FRP strips to share the load; and (4) using embedded FRP rods and fabric to provide confinement
[11].
44
A series of tests was carried out by

Taheri et al. to investigate the
buckling response of glulam
columns reinforced with FRP sheets
with different lengths and end fixity
[12]. The reinforcement levels
included non-reinforcement
(control), fully reinforced, and
partially reinforced (the FRP sheet
was one-third of the length of the
column and attached in the middle
of the column). The boundary
conditions of the column were
pinned-pinned and fixed-fixed ends.
Fig. 13 The specimens and different factors considered in the
series of tests carried out by Zhang et al. [10]
Fig. 14 Column specimens and carbon Fig. 15 Ultimate strength and elastic modulus of
reinforcement used by Najm et al. [14] columns versus fibre content [14]
Fig. 16 Column to foundation specimen Fig. 17 Experimental setup for post-tensioned strengthening
with post-tensioned reinforcement and of LVL column [15]
external energy dissipater [16]
45
It was found that columns which were fully reinforced had a higher strength compared with the other
conditions. The experimenters concluded that using FRP for partially reinforcing a glulam column
is more effective for the pinned-pinned case as the strength of the column reached almost half of the
increase in strength of those fully reinforced, but only used one third of reinforcing material. Most
FRP composites use organic matrices in manufacturing FRP plates, but since the 90s there has been
significant progress in manufacturing FRP with inorganic matrices that are non-toxic, have good fire
resistance, and are not affected by UV radiation [13]. A series of tests to investigate the confinement
of circular timber columns using inorganic CFRP was carried out by Najm et al. [14]. They tested 40
column specimens in axial compression, two different wrapping methods for the CFRP, spirals and
full wrapping. The specimens and the carbon reinforcement used in the tests are shown in Fig. 18.
The reinforced column specimens exhibited higher strength and stiffness than the unreinforced
specimens. It was also observed that specimens that were fully wrapped had higher strength and
stiffness compared with those that were partially reinforced (spiral reinforcement). With respect to
strength increase and fibre content, it was observed that the average load-carrying capacity of the
column increased with the decrease of the spacing of CFRP, i.e. increase of the amount of CFRP. The
same phenomenon can be found for the axial stiffness of the column specimens. In other words, the
more CFRP used, the better the mechanical properties the columns will have as can be seen in Fig. 19.
2.5 Post-tensioned strengthening
Post-tensioned strengthening is a relatively new development in the seismic field. An extensive
experimental campaign was carried out on beam-to-column, column-to-foundation and wall-to-
foundation subassemblies for the implementation of LVL hybrid solutions [15, 16]. The design was to
use external energy dissipaters together with post-tensioned effect to provide re-centring and energy
dissipation capacity of a timber column. Fig. 20 and Fig. 21 show the specimens and experimental
setup, respectively.
The hysteretic loop (Fig. 22) shows a flag-shape, and it was observed that 4.5% of the storey drift was
achieved in the tests; there was no degradation of stiffness and no structural damage after the tests.
The residual deformation was still negligible as the post-tensioned mechanism helps the column to re-
centre when unloading.
2.6 Enlargement of column cross section
Enlargement of the cross section of a column will help to reduce the stress within the column so as to
reduce the potential for buckling and material yield in compression. In some structures, such as those
found in Japanese temples, large section columns will contribute to resisting lateral load by providing
restoring forces [17]. Suda, Tasiro & Suzuki [18] proposed to enlarge the column base of existing
structures (Fig. 23) to increase the restoring force so as to enhance the aseismic behaviour of
traditional temples.
Shaking table tests were carried out to investigate the effectiveness of the proposed reinforcement
method. The reinforced column shows higher restoring force and larger deformation. This gives the
whole structure better lateral force resistance.
46
Fig. 18 Column specimens and carbon Fig. 19 Ultimate strength and elastic modulus of
reinforcement used by Najm et al. [14] columns versus fibre content [14]
Fig. 20 Column to foundation specimen Fig. 21 Experimental setup for post-tensioned strengthening
with post-tensioned reinforcement and of LVL column [15]
external energy dissipater [16]
Fig. 22 Hysteretic loop of a post-tensioned

strengthening LVL column
47
Fig. 23 Enlargement of column base proposed by Suda, Tasiro & Suzuki [18]
3. Reinforcement of timber shear walls

This section solely discusses the strategy to reinforce timber shear walls; it is worth noting that repair
and strengthening of timber shear walls often are achieved through an intervention on the joints and
beams. There are several solutions to strengthening of timber shear walls including [19]:
 to use additional sheathings
 to reinforce shear walls with steel diagonal elements
 to reinforce existing sheathing of the shear wall with carbon or high-strength synthetic fibre
 to reinforce the beams using hardwood inserts
 to post-tension the walls using prestressing wire.
Fig. 24 The specimen and test setup for

CFRP reinforced timber shear walls [21]
The first solution is the simplest method and is popular. The effectiveness of this method relies
heavily on the stiffness of the fasteners connecting the boards to the frame. The second method is to
attach steel diagonal elements to timber frames so as to share the force with the timber shear walls.
The first two solutions are relatively straightforward and can be designed by calculation, therefore
only limited research efforts have been devoted to these two methods.
The remaining solutions ensure that the reinforced timber shear walls will have higher ductility and
strength. These methods have attracted more attention in research and are discussed below.
48
3.1 Composite material reinforcement

A series of research programmes have been carried out on reinforcing timber shear walls using FRP
strips [20, 21]. A total of nine specimens were tested in three groups and the CFRP strips were glued
on the fibre-plaster board (FPB) attached to the timber frames [20]. Fig. 24 shows the specimen for
the tests. The first group (G1) used two CFRP diagonal strips with width of 300 mm glued on to the
FPB and also onto the timber frame; whereas the second group (G2) used 600 mm wide CFRP
diagonal strips with the other conditions being same as the first group. The third group (G3) has two
300 mm width CFRP strips glued on the FPB but not attached to the timber frame. The experimental
results revealed that the third group had the highest elastic resistance (force forming the first crack)
although it was found to increase in all the CFRP strengthened test samples. The results from this
series of tests showed that the three reinforcement methods do not increase the stiffness but increase
the strength. An analytical model has been further developed to approximate the behaviour of timber
shear walls reinforced by CFRP strips with satisfactory agreement [21].
3.2 Reinforcement by use of timber

Chang, Hsu & Komatsu proposed a new solution to reinforce traditional planked timber shear walls
(Fig. 25) after an earthquake by inserting hardwood strips into grooves in beams that accommodate
these timber planks [22]. Two different species of hardwood were used, Teak (Tectona grandis) and
Padauk (Pterocarpus spp.). The results revealed that the timber shear walls reinforced by Padauk and
Teak show a 100% and 60% increase in strength, respectively, compared with unreinforced and intact
timber shear walls. The reinforced timber shear walls also exhibit better energy dissipation under
cyclic loading.
Fig. 25 Schematic drawing of the reinforcement strategy [22]
3.3 Post-tensioned strengthening

Strengthening of timber shear walls by using the post-tensioned technique provides a very unique
opportunity to achieve better aseismic behaviour for timber walls. In the experimental campaign
described in Section 3.5, two different types of post-tensioned timber shear walls were tested, i.e. the
single (Fig. 26) and coupled timber walls (Fig. 27). In the coupled wall specimens, a U-Shaped
Flexural Plate (UFP) was developed and adopted to connect two smaller units of shear walls. The
hysteretic loop of the coupled walls system shows a promising result as the system has combined
good energy dissipation capacity and recentring effect as shown in Fig. 28.
49
Fig. 26 Schematic illustration of Fig. 27 Schematic illustration of coupled post-

post-tensioned timber shear wall tensioned timber shear walls
This technique shows good potential in the future for seismic-prone areas. However, to achieve a
more robust system, more research should be done to help engineers to deal with long-term creep in
columns caused by post-tensioned and stress relaxation.
Fig. 28 Hysteretic loop of coupled walls [15]
4. Discussion
The previous sections provide an overview on different methods to repair and reinforce damaged and
undamaged timber columns and shear walls, which are tabulated in Tab 2 and Tab 3.
4.1 Reversibility
When dealing with architectural heritage, techniques used to repair or reinforce a structural member
should be reversible whenever possible. The literature has shown there to be a lack of research into
reversible repair techniques for these valuable cultural heritages. Using composite materials, such as
FRP and CFRP, with timber tends to be an irreversible technique due to the adhesive used. The screw
repair technique proposed by Song et al. is reversible [5], but more research should be carried out to
investigate other parameters such as spacing between screws, types of self-tapping screws, the effect
of the size of cracks, etc.
50
Tab. 2 Summary for repair and reinforcement of timber columns

Screw Steel Composite Post-
Prosthesisation
reinforcement member material tensioned
Increase
NA X NA X NA
Cracked strength
members Increase
NA X NA X NA
stiffness
Increase
Bio- X NA NA X NA
strength
decay
members Increase
X NA NA X NA
stiffness
Increase
NA ? O O O
Intact strength
buildings
members
Existing
Increase
NA ? O O O
stiffness
Increase
New buildings
NA ? O O O
strength
members
Increase
NA ? O O O
stiffness
References [3, 23] [5] [6] [7-12, 14] [15, 16]
NA: Not applicable; O: Applicable and will increase properties (such as stiffness and strength)
X: Applicable and will not increase properties (such as stiffness and strength); ?: Need further research
4.2 Long term behaviour of reinforced structural members

Timber is a mechano-absorptive material and therefore creep will need to be considered when long-
term loadings are imposed; it is particularly onerous when the moisture content of the timber members
constantly varies between high and low levels. Post-tensioned reinforcing techniques tend to introduce
high levels of stress into structural members and therefore the long term behaviour of timber columns
and shear walls reinforced by this technique should be investigated. The post-tensioned system
reviewed in this chapter [15, 16] introduces a compressive stress on the timber column perpendicular
to the grain, where beams are connected to the column. This in turn will lead to creep in the material.
How this creep will affect the reinforcement will need to be addressed in the future. Another
important issue to be considered is ageing and weathering of composite material used in
reinforcement and strengthening. Avent tested epoxy reinforced timber connections and pointed out
that the dry condition shear strength of epoxy repaired Southern pine was reduced by one-third when
the repaired member was exposed to natural weathering conditions [24].
51
Tab. 3 Summary for repair and reinforcement of timber shear walls
Composite Timber
Prosthesisation Post-tensioned
material member
Increase
X NA O NA
damaged strength
members Increase
X NA O NA
stiffness
Increase
NA O O NA
Intact strength
buildings
members
Existing
Increase
NA X O NA
stiffness
Increase
New buildings
NA O O O
strength
members
Increase
NA X O O
stiffness
References [19-21] [22] [15, 16]
NA: Not applicable; O: Applicable and will increase properties (such as stiffness and strength)
X: Applicable and will not increase properties (such as stiffness and strength); ?: Need further research
4.3 Compatibility of materials

Pizzo et al. compared the values of the thermal expansion coefficients (TECs) of two wood species
(Spruce and Iroko) and four different epoxy resins. It showed that minor differences in TEC have
been observed between wood in perpendicular-to-the-grain direction and an experimental epoxy
adhesive [25]. Other commercial epoxy adhesives showed greater differences in terms of TEC and a
proportionally decreasing mechanical compatibility. This implies that when a new material and
adhesive are selected to strengthen or reinforce timber columns or shear walls, the TECs of these
material needs to be considered and compatibility should be ensured for long term behaviour. Another
important issue to consider when selecting new material for strengthening is to consider the
compatibility of TEC of timbers so as to ensure the load being transferred from timber to
reinforcement in all conditions.
4.4 Fire performance

Timber has low thermal conductivity; it is measured at approximately 0.8 (measured in
J/h/m2/mm/°C) compared with 12.6 for concrete and 312 for steel [26]. Most epoxies begin to soften
at 90o-120oC and the strength rapidly decreases. Therefore it is good practice to inject epoxy into
timber to repair the interior regions and the timber will protect and slow down the strength reduction
of the epoxy. In the event of fire, the composite material and epoxy are exposed to fire in those
52
composite material reinforcement methods described previously ([5], [7], [8], [9], [20]). This will lead
to the reinforcement and strengthening measurements becoming ineffective. Furthermore, when
selecting the composite material and adhesives for reinforcement and repair, one should ensure that no
toxic emissions occur during the fire.
The strength of steel is halved when exposed at temperature of 600 oC, therefore similar situation can
be found in steel reinforcement of timber members [5, 6]. It is therefore important to develop
appropriate reinforcement and strengthening methods for timber members in the event of fire.
4.5 Effectiveness of prosthesisation
There is no evidence as to how effective the prosthesisation technique is, although this is a widely
accepted practice within architectural heritage conservation programmes. Research effort should be
invested into experiments as well as into developing design guidelines for this practice, such as the
buckling response of timber columns where the lower part of the portion is replaced by new timber
with traditional carpentry.
5. Summary
An extensive overview has been carried out in this chapter on different repair and reinforcement
techniques that should be implemented on timber columns and shear walls under various
circumstances. The existing research has shown that reinforcement of columns by screws and
composite materials such as FRP are effective although there is a need to investigate the long term
performance of these measures. Compared with timber columns, less research has been carried out to
explore the strategies to reinforce and repair timber shear walls, however, reinforcement and repair of
timber shear walls by employing composite materials or hardwood appear to be effective as have been
demonstrated by some authors. This chapter also discusses and analyses the need for future research
on the repair and reinforcement of these structural elements.
Acknowledgement
The author appreciates permissions granted from Prof. Andy Buchanan and University of Canterbury
for Figs. 16-18 and 24, Profs. Suzuki and Suda for Fig. 19, Profs. Weiping Zhang and Xiaobin Song
at Tongji University, China, for Figs. 9 and 10.
References
[1] Yeomans, D.T., The repair of historic timber structures, Thomas Telford, 2003.
[2] Keey, R.B., Langrish, T.A., Walker, J.C., Kiln-drying of lumber, Springer Verlag Berlin
Heidelberg, New York, 2000.
[3] D'Ayala, D., Wang, H., "Conservation Practice of Chinese Timber Structures: ‘No Originality
to be Changed’or ‘Conserve as Found’", Journal of Architectural Conservation, Vol. 12, No. 2,
2006, pp. 7-26.
[4] NAIT5. PixelSkin02. 2008 [cited 2014 December 30]; Available from:
http://nait5.wordpress.com/2008/06/11/pixelskin02/.
[5] Song, X., Jiang, R., Zhang, W., Gu, X., Luo, L., "Compressive behavior of longitudinally
cracked wood columns retrofitted by self-tapping screws", In: Proceedings of World
Conference on Timber Engineering, Auckland, New Zealand, 2012.
53
[6] Tanaka, H., Idota, H., Ono., T., "Evaluation of buckling strength of hybrid timber columns
reinforced with steel plates and carbon fiber sheets", In: Proceedings of 9th World Conference
on Timber Engineering , Oregon, United States, 2006.
[7] Plevris, N., Triantafillou, T.C., "FRP-reinforced wood as structural material", Journal of
Materials in Civil Engineering, Vol. 4, No. 3, 1992, pp. 300-317.
[8] Triantafillou, T.C., Deskovic, N., "Prestressed FRP sheets as external reinforcement of wood
members", Journal of Structural Engineering, Vol. 118, No. 5, 1992, pp. 1270-1284.
[9] Ehsani, M., Se, M.L., "Strengthening of old wood with new technology", STRUCTURE, 2004.
pp. 19.
[10] Zhang, W., Song, X., Gu, X., Tang, H., "Compressive behavior of longitudinally cracked
timber columns retrofitted using FRP sheets", Journal of Structural Engineering, Vol. 138, No.
1, 2011. pp. 90-98.
[11] Oprişan, G., Ţăranu, N., Enţuc, I.-S., "Strengthening of the timber members using fibre
reinforced polymer composites", Buletinul Institului Politehnic DIN IAŞI, Vol. L (LIV)( 1-4),
2004, pp. 67- 75.
[12] Taheri, F., Nagaraj, M., Khosravi, P., "Buckling response of glue-laminated columns reinforced
with fiber-reinforced plastic sheets", Composite Structures, Vol., 88, No. 3, 2009. pp. 481-490.
[13] Davidovits, J., "Geopolymers", Journal of Thermal Analysis and Calorimetry, Vol. 37, No. 8,
1991, pp. 1633-1656.
[14] Najm, H., Secaras, J., Balaguru, P., "Compression tests of circular timber column confined with
carbon fibers using inorganic matrix", Journal of Materials in Civil Engineering, Vol. 19, No.
2, 2007, pp. 198-204.
[15] Palermo, A., Pampanin, S., Fragiacomo, M., Buchanan, A., Deam, B., "Innovative seismic
solutions for multi-storey LVL timber buildings", In: Proceedings of 9th World Conference on
Timber Engineering, Oregon, United States, 2006.
[16] Smith, T., Pampanin, S., Fragiacomo, M., Buchanan, A., "Design and construction of
prestressed timber buildings for seismic areas", NZ Timber Design Journal, Vol. 16, No. 3.
2008, pp. 3-10.
[17] Suzuki, Y., Maeno, M., "Structural mechanism of traditional wooden frames by dynamic and
static tests", Structural Control and Health Monitoring, Vol. 13, No. 1, 2006. pp. 508-522.
[18] Suda, T., Tasiro, Y., Suzuki, Y., "Seismic Reinforcement by Restoring Force due to Column
Rocking for Traditional Wooden Frame (in Japanese)", In: Proceedings of Disaster Prevention
of Historic Cities, 2011.
[19] Dobrila, P., Premrov, M., "Reinforcing methods for composite timber frame–fiberboard wall
panels", Engineering Structures, Vol. 25, No. 11, 2003, pp. 1369-1376.
[20] Premrov, M., Dobrila, P.,Bedenik, B., "Analysis of timber-framed walls coated with CFRP
strips strengthened fibre-plaster boards", International Journal of Solids and Structures, Vol.
41, No. 24, 2004, pp. 7035-7048.
[21] Premrov, M., Dobrila, P., "Mathematical modelling of timber-framed walls strengthened with
CFRP strips", Applied Mathematical Modelling, Vol. 32, No. 5, 2008, pp. 725-737.
[22] Chang, W.-S., Hsu, M.-F., Komatsu, K., "A new proposal to reinforce planked timber shear
walls", Journal of Wood Science, Vol. 57, No. 6, 2011, pp. 493-500.
[23] Pinto, L., Inventory of repair and strengthening methods timber, in Departament d'Enginyeria
de la Construcció, Universitat Politècnica de Catalunya: Barcelona, Spain, 2008.
[24] Avent, R.R., "Decay, Weathering and Epoxy Repair of Timber", Journal of Structural
Engineering, Vol. 111, No. 2, 1985, pp. 328-342.
[25] Pizzo, B., Rizzo, G., Lavisci, P., Megna, B., Berti, S., "Comparison of thermal expansion of
wood and epoxy adhesives", Holz als Roh-und Werkstoff, Vol. 60, No. 4, 2002, pp. 285-290.
[26] Avent, R.R., Issa, C.,A., "Effect of fire on epoxy-repaired timber", Journal of Structural
Engineering, Vol. 110, No. 2, 1984, pp. 2858-2875.
54
Analysis and strengthening of carpentry joints
4
Jorge M. Branco1, Thierry Descamps2
Summary
Joints play a major role in the structural behaviour of old timber frames [1]. Current standards mainly
focus on modern dowel-type joints and usually provide little guidance (with the exception of German
and Swiss NAs) to designers regarding traditional joints. With few exceptions, see e.g. [2], [3], [4],
most of the research undertaken today is focused on the reinforcement of dowel-type connections.
When considering old carpentry joints, it is neither realistic nor useful to try to describe the behaviour
of each and every type of joint. The discussion here is not an extra attempt to classify or compare joint
configurations [5], [6], [7]. Despite the existence of some classification rules which define different
types of carpentry joints, their applicability becomes difficult. This is due to the differences in the way
joints are fashioned depending, on the geographical location and their age. In view of this, it is
mandatory to check the relevance of the calculations as a first step. A limited number of carpentry
joints, along with some calculation rules and possible strengthening techniques are presented here.
1. Timber frameworks and carpentry connections

Timber frameworks are one of the most important and widespread types of timber structures. Their
configurations and joints are usually complex and testify to a high-level of craftsmanship and a good
understanding of the structural behaviour that has resulted from a long evolutionary process of trial
and error. A simplified analysis of (old) timber frameworks, considering only plane parts of the
system, is often hard to realize. Nowadays, a considerable number of timber structures require
structural intervention due to material decay, improper maintenance of the structure, faulty design or
construction, lack of reasonable care in handling of the wood, accidental actions or change of use.
While the assessment of old timber structures is complex, it is an essential precursor to the design of
the reinforcement of the joints. Owing to a lack of knowledge or time, the species and/or grade
assumed are often an overly conservative estimate which can lead to unnecessary replacement, repair
and retrofit decisions along with associated superfluous project costs.
For the design of the reinforcement of old timber structures or joints, the first step is to understand
fully how the structure and the joints work. Old timber structures are usually highly statically
indeterminate structures. This means that loads applied to the structure have different pathways to
reach the supports. Resolving the indeterminate system involves looking for additional equations that
actually express the relative stiffness of all those pathways. To illustrate how the differential stiffness
of elements, joints or supports may influence the behaviour of the structure, a simple collar-braced
roof is presented in Fig. 1.
1)
Assistant Professor ISISE, Dept. Civil Eng., University of Minho, Guimarães, Portugal
2)
Assistant Professor, URBAINE, Dept. Structural Mech. and Civil Eng., University of Mons, Mons, Belgium
55
In the absence of buttressed walls, under vertical loads, the collar (or the tie-beam) is under tension
because it prevents the roof from spreading. If buttressed walls restrain the feet of the rafters, the
collar is in compression. The only difference between these situations is the horizontal stiffness of the
supports (zero or infinite). The mass of the walls to resist the outward thrust is not the only
influencing factor. Most of the time, principal rafters are connected to wall plates that have to be stiff
enough to act as a beam in the horizontal plane spanning between two fixed ends in the walls. If the
rafters are notched, for example, with birdsmouth joints, over the plate at the top, the roof can be hung
from the ridge purlin, depending on the stiffness of the wall plate. The stiffness determines the ability
of the wall plate to act as an additional support. This is valid for most types of carpentry joints as they
usually are statically indeterminate. In conclusion, when working on old carpentry joints, it could be
useful, when possible, to look at the joint as an assembly of equivalent springs. This model allows a
better understanding of how the joints behave and deform and determines where the major stresses
will occur. This could help to avoid incorrect positioning of the reinforcement and thereby circumvent
poor design.
Fig. 1 Collar-braced roof
The main challenges for the structural assessment of carpentry joints are [8]:
• Stiffness and strength of joints depend on the type of loading. As an illustration, the rotational
stiffness of a joint is mostly different under positive and negative bending. Moreover, within most
joints there is an interaction between the different pathways in which the forces are transferred in
terms of stiffness and strength. This interaction should be considered to define the mechanical
behaviour of the connection.
• Despite most current standards not declaring any rules for the assessment of the material strength
under combined stresses, their appearance in carpentry joints is inevitable.
• The design of traditional joints essentially involves a check of the contact pressure between the
assembled elements. It is not easy to calculate the value of contact pressure in the following
situations: unknown contact surfaces and non-uniform stress distributions (because of non-
uniform elastic support due to local defects like knots for example).The values of compressive
strength of timber are different in the direction parallel and perpendicular to the grain. In order to
calculate the strength at any intermediate value of the load angle to the grain Hankinson’s
formula, which has been presented in many standards, may be used. SIA 265:2003 [9] suggests a
different expression that takes into account a reduction because of the difference between the
strength of early wood and latewood. In addition, some standards allow enlarging the real contact
surface by taking into account a so-called effective length [10]. Those slight differences about the
definition of the compressive strength at an angle to the grain highlight a lack of knowledge,
which fortunately, is not of major importance for compression at angles between 30º to 60º
(which represent the most common values).
56
2. Old carpentry joints

Common traditional carpentry joints found in old timber frames can be categorized in four main
types, according to their arrangement and geometry:
• Mortise and Tenon joints: There are countless examples of this type of joint. Tenon joints connect
members that usually form an "L" or "T" type configuration. The joint comprises two
components: the mortise hole and the tenon tongue. The tenon formed on the end of a member is
inserted into a square or rectangular hole cut into the corresponding member. The tenon is cut to
fit the mortise hole exactly and usually has shoulders that sit when the joint fully enters the
mortise hole. The joint may be pinned or locked into place. In the traditional fashion, the pin hole
in the tenon is bored a little closer to the shoulder than in the mortise and the pin pulls the joint
together very tightly. This kind of joint is mainly used when the adjoining pieces connect at an
angle between 45° to 90°. When the angle between the two jointed elements is different from 90º,
the nose of the tenon can be cut off and is called a skewed tenon (see Fig. 3a & 6a).
Fig. 2 (a) Through pinned mortise and tenon (a’) blind pinned mortise and tenon (blind means not
going all the way through). (b) Through tenon with outside wedges (flatwise bending of the tenon (b’)
wedged and pinned dovetail through mortise and tenon
• Notched joints: This kind of joint is linked to the development of king post and king post-like
frames. In order to work successfully, these frames need appropriate joinery at a multitude of
locations. A notch is a "V" shaped groove generally perpendicular to the length of the beam, as
seen in Fig. 3. Examples where notched joints are used include cases where secure footing is
required for the toe of a rafter (or strut) or between the rafter and the king-post. A tenon can be
added to the notched joint to essentially keep all the beams coplanar but the notch is what creates
the strength of the joint (because it is stiffer than the tenon).
Fig. 3 (a) Notched joint between main rafters and tie-beam.(a’) A skewed tenon may be used to help in
keeping all timber pieces co-planar. (b) Peak joint with a notched joint (main rafters and post)
57
• Lap joints: In a full lap joint, no material is removed from either of the members to be joined,
resulting in a joint whose thickness equals the combined thickness of the two members. The
members are held in place by a pin (Fig. 4a). In a half-lap joint, material is removed from each of
the members so that the thickness of the resulting joint is the same as that of the thickest member.
Most commonly, in half-lap joints, the members are of the same thickness and half the thickness
of each is removed. The cogged half-lap joint is a half-lap with additional cogs. The dovetail-lap
joint (named after the shape of the tenon being similar to the tail of a dove) is another way to
fashion the joint in an attempt to reinforce its tensile strength (Fig. 4c).
Fig. 4 Full lap joint (pinned). (b) Half-lap joint. (b’) Cogged half-lap joint. (c)Through dovetailed lap
joint or (c’) wedged dovetailed lap joint if ever the dovetail is embedded in the member
• Scarf joints: Scarf joints (and splice joints), shown in Fig. 5, allow the joining (splicing) of two
members end to end [11], [12]. They are mainly used when the material being joined is not
available in the length required. This technique is recognised as being the strongest form of
unglued member lengthening [13]. The halved-scarf joint is a lap whose surfaces are parallel with
the members. It is similar to a half-lap joint with co-axial members. The scarf joint is simply a
pair of complementary straight sloping cuts secured to each other with pins (also called pegs).
Fig. 5 (a) Common and simplest halved-scarf joint (or half-lap splice joint). (a’)A lapped dovetail
scarf joint is a half-lapped joint in which the lapped portions are shaped like a dovetail joint. (b)Scarf
joint. (c)Scarf joint with under-squinted ends. (d) Trait de Jupiter: particular scarf joint with wedges
(key)
58
Another type of scarf joint is known as the Trait-de-Jupiter or Bolt of lightning, in view of its
resemblance to lightning. It is more efficient in the presence of a key (or several keys, depending
on the number of indentations – see Fig. 23d) made of hardwood to improve contact and to
simplify fabrication. From a mechanical point of view, it is an excellent scarf, since the driving of
its key separates the twin-tables with a primary mechanical force and closes the under-squinted
butts with enormous pressure.
3. Joint stiffness
Numerous examples demonstrate the excellent performance of old timber constructions during
earthquakes or exceptional wind loads. The reason why they are still standing is not only due to their
robustness (highly statically indeterminate structures), but also due to the semi-rigid and ductile
behaviour of their joints, which allow for the dissipation of energy. Also, thanks to its load-
redistribution process, the beams and joints are able to maintain the capacity of the whole structure in
spite of partial damage.
According to common standards, such as Eurocode 5 [10], the rigidity of elements and joints as well
as the eccentricities of the joints have to be taken into account for the computation of the internal
forces. However, in order to simplify the analysis, joints are usually designed by assuming an ideally
pinned (or rigid) behaviour [14]. It is quite obvious that the assumption of pinned joints is
conservative, provided that the joints have enough ductility and are fashioned in a way that their
rotation may develop (deformation capacity is sufficient). Nevertheless, in reality most of the
carpentry joints are not perfect hinges. Though this is not of major importance for the design of the
members, it must be borne in mind that the splitting of timber may occur under low loads (because of
component loads perpendicular to the grain). Therefore, in some cases the joints are assumed to be
rigid. This is conservative for the joints and results in an uneconomic design. Furthermore, carpentry
joints usually have a significant moment-resisting capacity even without any strengthening devices.
Test results on full-scale notched joints show that this capacity is a function of the compression level
in the rafter, the width of the rafter, the friction, the skew angle and the notch depth [3], [15], [16].
Rotational capacity is positively related to the first three parameters.
Undoubtedly, the modelling of the structure taking into account the semi-rigid behaviour of the joints
is the best practice:
 A semi-rigid study of a structure suggests taking into account the stiffness of the joints with
regard to all the components of the loads (normal, shear and bending). In fact, Descamps et al.
[4] have shown that for the computations of the internal forces, the use of the rotational
stiffness alone is not enough. Both axial and rotational stiffness have to be introduced in finite
element models for an accurate study. The shear stiffness is of less importance.
 Uzielli et al. [17] reported on research work in which different assumptions about the joints in
an old timber structure were compared. They found a maximum difference of 20% between
the computed stresses in a semi-rigid model compared to the experimental results, while the
difference increased up to 40% when assuming pinned or rigid joints.
3.1 Component method
Design models are available in all standards for estimating the stiffness of dowel-type joints.
Unfortunately no information is given about how to get the stiffness of carpentry joints in order to
help engineers to gain better results. The component method allows stiffness values to be determined
59
for joints according to their geometrical and mechanical properties. This method has been used
frequently in the field of steel construction and is now applied in research on carpentry joints by
several authors [8], [18], [19], [20], [21], [22]. The problem will be explained by considering a skew
tenon joint under an axial load (Fig.6). Since different loads paths are possible (the joint is statically
indeterminate), the worst case scenario has been chosen (the mortise is longer than the tenon, which is
a common way of fabricating this type of joint).
The component Fh is balanced through a uniform contact pressure on the shoulders of the tenon
(surface (1) in Fig. 6a). The eccentricity of the resulting forces is not considered here. The load
transferred up to the axis of the post (loaded perpendicular to the grain) causes a deformation that can
be defined by the stiffness Kh:
(1)
where E90 is the mean value of the modulus of elasticity perpendicular to the grain, A1 is the area of
the contact surface (1) and L is the height of the section of the post. Assuming a gap between the end
of the tenon and the post, the component Fv is balanced through a uniform contact pressure on the
head of the tenon (surface (2) in Fig. 6a). The load transferred to the tenon causes a deformation that
can be defined by the stiffness Kv:
(2)
where Eα is the mean value of the modulus of elasticity at an angle α to the grain, and A2 is the area of
the contact surface (2). Fig. 6b gives the equivalent spring model of the joint under an axial load. If M
is the moment in the joint and θ is the rotation into the joint (Fig. 6c), the rotational stiffness of the
joint is:
(3)
Fig. 6 (a) & (b) Skewed tenon joint under an axial load and the equivalent beam and spring model. (c)
& (d) Tenon joint under bending and the equivalent beam and spring model
Fig. 7 Definition of a modified modulus of elasticity
60
The applied moment is balanced through contact pressures on surfaces (1), (2) and (3). Those
pressures can be assumed to be uniform or non-uniform. The effect of friction has not been
considered. Each of the contact pressures causes a deformation that can be divided in two
components. One component of deformation is caused by the material (i), for example, loaded in
parallel to the grain and another component of deformation is caused by material (i'), in contact with
(i), for example, loaded perpendicular to the grain. The stiffness Ki and Ki’ can be defined as proposed
by Meisel et al. Ki,i’ is the equivalent stiffness of two springs Ki and Ki’ in series (in material (i) and
(i’), respectively) [8]. Drdácký et al. have proposed another definition of the stiffness Ki based on a
well-known model used to calculate the settlement under a rectangular foundation supported by a
semi-infinite half space [18]:
(4)
If i is the deformation at surface contact (i) and (i’), Fig. 6c:

(5)
For small displacements, such as the surface contact (1):

(6)
Finally, the rotational stiffness is equal to:
(7)
Fig. 6d gives the equivalent spring model under bending. Some enhancements of the method, in
particular for the definition of the stiffness Ki, have been proposed by Descamps et al. [4]:
 Definition of a "modified modulus of elasticity" that takes into account the edge effect that
appears when surface (i) and (i') are next to an edge. C is defined in Fig. 7 where  is the
slenderness of the contact area.
(8)
 New definition of the centre of rotation of the joint. Observations of broken pegs made on
real-size tests suggest that the peg is not the centre of rotation of the connection. The
assumption has been made that the common position of the centre of rotation corresponds to a
minimum of the bending stiffness of the joint (iterative procedure).
Komatsu et al. [20] proposed an enhanced model with a definition of the elastic stiffness Ki that takes
into account an effective length for the definition of Ai (as in Eurocode 5 for the check of compression
perpendicular to the grain). They have also proposed a definition of the stiffness in the post yielding
range. Chang et al. have proposed some enhancements to this model, in particular to take into account
initial gaps and slips [21]. Comparisons with experimentation have demonstrated that the enhanced
model achieves good results not only for the initial stiffness of the joints with gaps, but also for the
initial slip stage which should be regarded as a pin connection in the early stages.
The component method can be applied to other types of carpentry joints. Researchers have
concentrated its use on tenon [4], [8], [18] and lap joints [20], [21], but its application to notched
joints is also possible.
61
4. Evaluation and reinforcement of carpentry joints

Before any intervention, the first step is the assessment of the existing joints in relation to the
material, the strength and the stiffness. Proper assessment of the material (decay) with appropriate
techniques is obviously of major importance and therefore the study of recent state-of-the-art is highly
recommended [23]. This survey may lead to the replacement of a portion or the whole member. On
the other hand, in the case that the member or joint is kept in service and reinforcement is needed, an
accurate assessment of the state of conservation of timber is crucial.
In the past, the first action taken by carpenters to strengthen joints was based on precise observations
of failure modes encountered in real structures and a good understanding of their weakest points. This
led to an improvement in the sketching of the joints and one can say that many carpentry joints are an
evolution or a reinforcement of older primary joints. For example, a notched joint with a tenon (see
Fig. 3a’) can be considered as a “reinforcement” of a tenon joint because the slope of the notch
increases the load bearing capacity of the joint. In the past, joints were realised without any metal
fasteners such as nails, screws or bolts and their ability to carry the loads was achieved through direct
contact and friction. Over the years, various reinforcement techniques such as the use of screws
(including self-tapping-screws), metal plates (strips, stirrup...), glued composites (glass or carbon
fibres, weft knitted textiles) and glued-in rods or even full injection with fluid adhesives among others
have been proposed. It should be noted that special attention has to be paid to any solution that
consists of wrapping the joint in an airtight textile (risk of decay). Furthermore, for the restoration of
historic buildings, all interventions should be reversible; if not completely, they should not limit
further interventions. For this reason, the injection with fluid adhesive directly into the joint is not
recommended anymore [24]. Dowel-type fasteners have been used occasionally in timber joints, for
example, to counteract any out-of-plane displacements which cannot be counteracted by the joint
itself. This practice became common in the 19th century with the development of industrial production
methods and the manufacture of low cost fasteners. Nowadays, the strengthening may aim to locally
reinforce the material in the joint area, for example, to reinforce the timber in shear or tension
perpendicular to grain by means of self-tapping-screws or to avoid the detachment of the connected
elements (joints that could not carry any tension loads for example) or to modify locally the pathway
followed by the loads into the joint. Particularly, in seismic areas, strengthening can prevent loss of
capacity and possible separation of contact surfaces due to the decrease of compression forces, and
may maintain a suitable structural behaviour [2]. The first step of any reinforcement intervention is of
course the definition of a proper model of the joint to assess its strength and stiffness. Models and
reinforcement techniques will be discussed for the most common carpentry joints here after.
4.1 Tenon joints
Tenon joints have a very low stiffness that may cause premature failure of a part or the whole
structure caused by large displacements encountered in the joint [2]. The bearing capacity of skewed
tenon joints is a function of the angle α of the joint, the length of the tenon and the mortise depth [25],
[26], [27]. To check the joint, one may use a simple check on all components of the load that appear
on the surfaces as it has been discussed in Section 3.1.
62
Fig. 8 Configuration and force mechanism in a skewed tenon joint
Each part (i) or (i') of the surfaces in contact is checked in compression at an angle to the grain. Kock
et al. have developed guidelines for design that are suitable for skewed tenon joints, under axial and
shear loading [28].
Surfaces A1 and A2 in contact are presented in Fig 8. μH and μV are the coefficients of friction of A1and
A2, respectively, hs and ls are the height and the length of the strut, respectively, and tx is the distance
between the bottom surface and the loading point of H. The compression loads H and V on A1and A2,
respectively, are defined as follows:
(9)
(10)
where m is the ratio of Q to F:
(11)
4.1.1 Intervention and reinforcement of tenon joints

Feio et al. have tested full-scale notched and skewed tenon joints under compression in order
to assess the local failure in compression and the slipping of the joint [29]. Failure modes
observed in the tested joints are damages due to compression in the brace which are localized
at the tenon end or distributed along the full contact length. An out-of-plane bulging of wood
under the contact length was observed. In some cases, damages in compression associated
with shear failure were observed (Fig. 9). When observed on-site, this type of failure mode
mainly highlights a poor design of the joint (with contact areas that are too small) or of the
structure (unexpected compression forces in one element). No reinforcement can repair
damage in compression perpendicular to the grain and the replacement of the element is
required most of the time.

63
Fig. 9 Typical experimental failure patterns: (a) joint collapsed in compression, with uniform
distribution of damage, (b) joint collapsed in compression, with out-of-plane bulging, and (c)
combined failure in compression and shear parallel to the grain at the tenon [29]
 To ensure correct strength and stiffness of the joint, it is important to keep all the surfaces of
the joints in contact. In the case of reverse loads or because of high shrinkage of the wood
elements, joints may develop gaps. One traditional reinforcement technique consists in
placing a wooden wedge to ensure perfect contact between the tenon and the mortise (Fig.10).
This wooden wedge should be made of hardwood (for strength and stiffness) and its moisture
content (MC) should be as close as possible to that of the reinforced wooden elements in
order to avoid any shrinkage of the wedge.
Fig. 10 Wooden wedge to ensure the contact between the tenon and the mortise
 Pinned tenon joints also have a very low bearing capacity in tension as only the pin acts. If the
element has to be replaced, a traditional reinforcement technique consists in fashioning the
joint with a dovetail tenon to increase the strength in tension (Fig. 2b’). If the element remains
in service, a binding strip may be used as reinforcement in tension. The strip is screwed onto
the edge of the supporting beam to avoid any crack (Fig. 11). For the design of the fastening
of the strip, Eurocode 5 expressions for double shear in timber-to-steel connections can be
utilized. One may check the tensile stress in steel and the compression perpendicular to the
grain under the strip as follows:
(12)
64
where is the tensile strength of steel and is the shear strength of the fastener. To
avoid any tensile forces in the element, a steel wire may be used as presented in Fig.11b. If
broken pins are observed, replacing the wooden pins by steel ones is not suitable as it.
Fig. 11 Reinforcement of a tenon joint in tension
4.2 Notched joints

In a notched joint, the slope of the notch should minimize the angle between the stresses and the grain
direction for both connected elements, hence increasing the crushing resistance of the joint.
Fig. 12 Force mechanisms in a notched joint (tie beam and rafter for example) with the contact
surface at the front (a) or at the rear (b)
Based on simple geometric considerations, it is possible to demonstrate that the ideal configuration of
the notched joint is the one reported in Fig. 12a, where the angle of the notch is half of the
angle . According to Götz et al [30] and German and Italian standards [31], [32], the depth
of the notch, tv, should not exceed h/4 for  50º and h/6 for >60º (linear interpolation between those
values is proposed). Friction forces and geometric imperfections are not considered. Based on these
assumptions, the axial force is easily resolved into two component forces F1 and F2 perpendicular to
the two surfaces of the notch (Figure 12a):
(13)
65
(14)
If b is the width of the timber elements, the compression at an angle to the grain direction on the notch
and the shear in the frontal shear plane must be checked:
(15)
(16)
For the rear face under compression:

(17)
where d is the length of the rear compressed surface of the notch. This last verification, which is often
neglected, can be of importance because of the risk of high stresses on a surface of limited length.
Parisi et al. proposed the following empirical rule to calculate d [2]:
(18)
It is important to draw attention to the assumptions that have been made. Assuming that there is no
friction is quite far from reality. Friction may increase the stresses in the shear surface. The check of
shear stresses in front of the notch must be done with caution. In consequence of the fragile nature of
the shear failure, for instance, some standards on earthquake resistant structures (e.g.[31]) adopted a
higher partial factor for material properties (equal to 1,3) in the quantification of fv,d.
The notched joint with the contact surface at the rear aims to increase the shear strength of the joints
by increasing the shear surface in front of the notch (Fig.11b). To fashion the joint easily, a notch is
made perpendicular to the direction of one of the two members. In this case the joint strength
decreases because the slope of the notch does not minimize the angle between the stresses and the
grain direction for both connected elements anymore. The compression at an angle to the grain
direction at the notch and the shear in the frontal shear plane must be checked:
(19)
(20)
Owing to the eccentricity between the load F and the contact surface, the joint will turn (open) and a
crack can appear. Even with a gap of 1 or 2mm between the two connected members which prevent
the nose bearing of the rafter, it is almost impossible to avoid the splitting of the rafter.
Double step joints, whose geometry results from the sum of two different single steps joints, increase
the shear surface without a major risk of splitting (Fig. 13). To fashion the joint, the rear notch must
be deeper than the one at the front. To ensure that the joint works well, precision is required so that all
surfaces will be in contact (which is not easy to get especially when it is done manually). According
to Italian standard: and [32]. Whatever the skew angle of the
connection, for a double notched joint, Götz et al. recommends and [30], [32].
66
Fig. 13 Double notched joint.
Strengthening of notched joint also concerns the friction-based behaviour of the joints in its own
plane, preventing the separation of friction surfaces due to the decrease of compression forces (a
notched joint has no tensile strength). A series of monotonic and cyclic tests on unstrengthened and
strengthened joints has been performed by Branco et al. in order to study the initial behaviour of the
connection, as well as its sensitivity to a few parameters [3]. Even without any strengthening devices,
notched joints usually have a significant moment-resisting capacity. This capacity depends on the
axial compression load in the rafter and on the skew angle [2], [3]. Moreover, it is obvious that the
height of the rafter [33], the friction [2], the existence of an additional tenon [30] and the moisture
content [16] are also important.
4.2.1 Intervention and reinforcement for notched joints

 If the wooden elements do not perfectly match in the notched area (lack of precision in the
cutting of the members or because of shrinking), the placement of wooden wedges is
recommended to ensure a perfect contact between connected surfaces with a clear increase in
the load-carrying capacity of the joint [34].
Fig. 14 Wooden wedges used to ensure a perfect contact between elements. Gaps in the front notch (a)
and rear surface (b) and possible use of wedges, (a’) and (b’), respectively’
 The strengthening of existing notched joints mainly aims to avoid shear failure in the front
portion of the notch. Most of the time, an end beam repair is required because of decay and a
wooden prosthesis must be used to replace the degraded material (Fig 15).The prosthesis is
mechanically jointed to sound wood (or resin). One may use the following check:
(21)
where is the effective number of fasteners and is their load-carrying capacity [10].
Using inclined fasteners increases the load-carrying capacity of the prosthesis.
67
Fig. 15 Notched joints reinforced to shear stresses in the frontal part of the notch with a screwed
prosthesis
 In times past, binding strips, stirrups and bolts were used in seismic regions to avoid the
dismantling of the connected members under reverse loads (Fig. 16). When metal elements
were used in the original construction of the joints, or, added later, the intervention usually
included the substitution of the connectors (nails, bolts, etc.) by new ones and the treatment of
the metal.
Fig. 16 Examples of notched joints with metal devices
The strengthening techniques used presently look to reproduce the old techniques even when
using new fasteners like screws and self-tapping screws (see Fig. 17). These kinds of
interventions affect the stiffness of the joint, which should be checked too.
Fig. 17 Contemporary strengthening interventions on notched joints reproducing old techniques
 Strengthened joints with metal devices were tested by Branco et al. under monotonic and
cyclic loading [35]. The purpose was to uncover any advantages and drawbacks in the
behaviour of the joint and of the strengthening as well as to look at different types of
68
strengthening. The four types of strengthened joints tested are modern implementations of
traditional techniques (Fig. 18):
Fig. 18 Traditional strengthening techniques of notched joints: (a) metal stirrups, (b) internal bolt; (c)
binding strip (d) tension ties
o Metal stirrups placed in pairs on two sides of the joint have been a popular reinforcement
in the past. Each stirrup was composed of two steel plates welded in a V-shape and bolted
with seven bolts of 10mm of diameter. Each prong was 50mm wide and 5mm thick.
o The steel rod (12mm of diameter) was fixed by a nut at both ends and secured by using a
special rectangular-shape washer (70×30mm² and 5mm thick). The rod was located at
mid-joint and mid-width and normal to the axis of the tie beam. A notch has been cut in
the rafter to ease the contact between wood and the washer.
o Metal binding strips were also frequently used in the past, particularly to strengthen joints
at an angle of 30° [33]. Two updated versions were considered:
 The joint was bound by a steel ribbon (50mm wide, 5mm thick) located at mid-
connection, normal to the tie-beam.
 The joint was bound with two steel plates located in the bottom surface of the tie
beam and upper surface of the rafter (40×159mm², 10mm thick) tightened
through two rods of 12mm. One may notice that this solution enables a control of
the tightening force during the strengthened lifetime.
All have been analysed for α=30º and α=60º except the rigid binding strip (α=30° only - Fig. 18c).
Force-displacement curves under monotonic loading for unstrengthened and strengthened joints
(whatever the α values) are presented in Fig. 19. It was observed that all the strengthening techniques
used have resulted in improved joint behaviour.
Fig. 19 Force-displacement curves under monotonic loading for unstrengthened and strengthened
connections with α=30º (a) and α=60º (b)
Despite the amount of tests done, results are still insufficient to propose design equations for all the
tested configurations and reinforcements. There is still an evident lack of results and scientific data
69
about this topic, which clearly points out the lack of research in this field. However, some interesting
observations can be drawn. For α= 30°, one may notice:
o All the strengthening techniques used have increased the stiffness, in particular for
the positive loading direction and the maximum load for both directions. One should
check if this modification of the stiffness has any consequence on the whole structure.
o Large improvement of the ductility, especially under a negative loading (the
reinforced joint does not behave in a brittle way anymore).
o Reinforcement with stirrups and binding strips are similar from the point of view of
the maximum load reached. However binding strips have a lower ductility under
negative forces.
o Among all strengthening techniques tested, the least efficient regarding both
maximum force and stiffness is the solution with the tension ties.
For α= 60°:
o The behaviour of the unreinforced joint is ductile. Nonetheless, a significant increase
of the ductility of the reinforced joint has been observed.
o The same conclusion can be given for the increase in maximum load (whatever be the
technique used).
o No significant influence on initial stiffness has been observed.
o The efficiency of the reinforcement is almost the same for all techniques under
positive loading. For a negative loading, the metal stirrup is the most efficient
technique whereas the solutions with tension ties and bolt are similar.
4.3 Lap joints
The pin used in a full lap joint is of major importance as the joint cannot carry any loads without it.
Based only on the strength of the pin, the efficiency of the full lap joint is of course very low. The
half-lap joint is a first improvement of the joint given that it carries the loads by contact in addition to
the pin. In the half-lap joint, both the connected elements are half weakened, however the joints may
be fashioned differently to avoid weakening the members (one-third of the height instead of a half).
As a result the two pieces do not sit flush, which can limit the use of the joint. This configuration is
useful when both members bear a larger load. The lap joint presented in Fig. 4b’ (between the half-lap
joint and the cogging joint) is configured to increase capacity for both members by adding bearing
surfaces. The lower supporting beam has less material removed compared to a half-lap. The side
housings provide better support for the upper girder and lessen shear problems [6].
To check the joint, the total load can be resolved into equivalent loads acting on contact surfaces.
Each component is checked with regards to the strength in compression at an angle to the grain.
4.3.1 Intervention and reinforcement for lap joints
 If the wooden elements do not match perfectly in the notched area (due to lack of precision in
the cutting of the members or because of shrinking), the placement of wooden wedges should
be recommended to ensure perfect contact and, consequently, the attainment of full load-
carrying capacity of the joint.
 If a shear failure is observed in one of the elements, it can be reinforced with fully threaded
70
self-tapping screws (screws are better than bolts in case of shrinkage). Fully threaded screws
are required as they have to suspend the lower part of the beam from the upper part the lower
part of the beam to the upper part. Inclined screws may achieve a better load-carrying
capacity of the reinforced joint than screws screwed perpendicular to the grain direction. To
guarantee an efficient reinforcement, the distance a between the screw and the notch should
be minimized but with .
Fig. 20 Reinforcement in tension perpendicular to grain in lap joints: (1) reinforcement with self-
tapping screws perpendicular to the grain. (1’) reinforcement with inclined screws
Design equations developed for notched beams can be used. The design tensile force,
to be carried by the reinforcement, can be determined according to [36]:
(22)
Where is the design shear force and
 In dovetail-lap joints, the peg keeps all of the wooden pieces together and prevents the
formation of gaps. When a gap occurs, the contact surfaces become smaller and so, the
contact pressure becomes larger. One traditional reinforcement technique consisted in the
placement of a wooden wedge to ensure perfect contact (see notched joints).
 In dovetail-lap joints loaded in tension, the splitting of timber is a common failure mode. The
traditional reinforcement of those joints consists in adding fasteners (bolts, nails, screws, etc.)
restoring the shear mechanism provided by the pin (Fig. 21). The design of this strengthening
technique is based on the calculation of the shear resistance of the new fasteners. This
intervention affects the stiffness of the joint (displacement of the centre of rotation). Binding
strips or steel wire may also be used (see tenon joints).
4.4 Scarf joints
Two timber members connected using scarf jointing techniques cannot match the strength and
stiffness of a single member of the same dimensions. Besides the type of scarf joint, the actual size of
the elements, the strength of the wood and other factors have a substantial effect on the assembled
member’s strength and stiffness. All the different types of existing scarf joints are a proof of all
attempts made by carpenters to strengthen the joints and fit particular requirements. For example, the
scarf joint (Fig. 5b) is an improvement of the half-scarf joint (Fig. 5a) that works better in shear
because less material is removed from each of the members and there is no sharp angle. If the half-lap
is horizontal (and loads are vertical, see Fig. 23a), the maximum moment it can carry is one-quarter of
the moment of a solid beam, because the half-lap has the width of the beam but one-half of the height.
For a vertical half-lap (the scarf is face-halved and loads are vertical too), the width is one half of the
solid beam but the height of the half-lap is equal, so the maximum moment it can carry is one-half of
71
Fig. 21 Traditional reinforcement of dovetail-lap joints under tension loads by adding wooden dowels
Fig. 22 Scarf joints reinforced with metal connectors and plates.
Fig. 23 (a) Scarf joint reinforcement perpendicular to the grain with self-tapping screws. (b)
Reinforcement of bending strength (weak axis) with a cog (half cogged scarf joint). (c)Face-halved
scarf joint. (d) Multiple scarf joint with under-squinted ends.
Fig. 24 Scarf joint reinforced with glued in rods: steel rods are glued in both timber members and
connected with a long nut (Credits: Pascal Lemlyn. Restauration du Moulin de l’abbaye de la Paix
Dieu, Institut du patrimoine Wallon, Belgique).
72
what a solid beam will transfer. A study carried out by TRADA suggested that the limiting moment
capacity of scarf joints (which behave better in bending than the half-lap scarf) is equal to only a third
of the strength of the unjointed beam [37].
For the design, one must check all components of the load that appear on the surfaces in contact.
However, it should be noted that very few research campaigns have been conducted on the
reinforcement of scarf joints or even on their design. Four types of reinforced scarf joints have been
tested by Hirst et al. [12]. The beam sections were 200x150 mm², all pin holes were 19 mm in
diameter and in keeping with traditional practice, the pin holes on one side of the scarf were offset by
3 mm with respect to the other side of the joint (tightening the joint when the pins were driven
through the joint). Once again, the results are still not sufficient to discuss about the advantages and
drawbacks of different techniques or to propose a design equation for reinforced scarf joints.
4.4.1 Intervention and reinforcement for scarf joints
 If the wooden elements do not perfectly match in the notched area (due to a lack of precision
in the cutting of the members or because of shrinking), the placement of wooden wedges
should be recommended to ensure a perfect contact and hence the achievement of the full
load-carrying capacity of the joint.
 The easiest way to reinforce a scarf joint is achieved by adding metal fasteners (screws or
bolts). In ancient times, wooden pegs were used. In case of high loads, lateral metal plates can
be added to improve the load-bearing capacity of the joint and to increase the stiffness. Both
types of reinforcements are used in restoration works (Fig.22).
 Under bending, the rule of thumb that the weak point is the risk of premature splitting of
wood is encountered here too (joints cut with right angles are less suitable). From this point of
view, scarf joints are better than halved-scarf joints. Self-tapping screws can also be used to
strengthen the joint (Fig. 23a). The design equations used for notched beams can be used (see
lap joints). Under tension only, reinforcement screws can be driven only in the overlapping
area (Fig. 23d). This reinforcement can be checked using Johansen's equations assuming that
the tensile load is completely carried by the screws.
 In the case of the Trait-de-Jupiter it is common to add metal connectors passing through the
joint depth to reinforce the joint (Fig. 24). Another solution with glued in rods is presented in
Fig 24.
5. Conclusions
When working on old timber structures, the fact that the structure has stood for decades or centuries
without failure may not be sufficient proof of the load bearing capacity for the future (new imposed
loads etc.). Joints greatly influence the response of the whole structure. Their characterization (the
strength, the stiffness and the ability to be reinforced) still remains a big challenge.
The design of traditional joints essentially involves a check of the contact pressure between the
assembled elements. Even if seemingly trivial, assessment of old carpentry joints still remains a
difficult task. As an illustration, the slight difference in the definition of the compressive strength
(which is of major importance) at an angle to the grain mentioned in different standards, underscore a
basic point that has to be clarified anyway by further research and later on by the revision of current
73
standards. Moreover, not only the strength, but also the stiffness of the joint has to be considered as it
can influence the force distribution within the structure.
If the decay of timber elements is too large, then replacement is clearly the only solution. If repairs are
necessary, specific reliable on-site assessment techniques are required to determine the appropriate
level of intervention needed. This point remains very important to evaluate the replacement, repair
and retrofit solutions along with the associated project costs.
It should be noted that there is still a noticeable lack of scientific results and design rules regarding the
reinforcement of old carpentry joints. This clearly points out the lack of research in this field.
Unfortunately, this lack of information in addition to difficulties in assessment and definition of
grading protocols for old timber elements often lead to unnecessary replacements. Further studies in
the area are deemed necessary in order to establish reliable design models, to set detailing rules and to
provide recommendations for future rehabilitation or strengthening interventions, among others.
Because of the wide variety of carpentry joint geometries in existence, studying them with an
exhaustive approach is neither realistic nor useful. For an accurate study, a good understanding of
how the joint works and how the loads are balanced is the key point. As seen in this review, some
information about strength, stiffness and reinforcement of common joints exists; even though
scientific data is still missing and complementary research is needed. To achieve competency,
engineers need specific tools such as the ones defined for the design of dowelled joints. Hopefully, the
most important outcomes of existing (and ongoing) research will be integrated into the revised version
of Eurocode 5.
6. References
[1] Descamps T., Léoskool L., Laplume D., Van Parys L., Aira J.R., “Sensitivity of timber
hyperstatic frames to the stiffness of step and ridge joints”, In: Proceedings of the 13th World
Conference on Timber Engineering, Quebec, Canada, 2014.
[2] Parisi M., Piazza M., “Mechanics of plain and retrofitted traditional timber connections”,
Journal of Structural Engineering, Vol. 126, No.12, 2000, pp. 1395–1403.
[3] Branco J.M., Piazza M., Cruz P.J.S., “Experimental evaluation of different strengthening
techniques of traditional timber connections”, Engineering Structures, Vol. 33, No. 8, 2011, pp.
2259-2270, http://hdl.handle.net/1822/13592.
[4] Descamps T., Noël J., “Semi-rigid analysis of old timber frames: definition of equivalent
springs for joints modeling. Enhancement of the method, numerical and experimental
validation”, International Review of Mechanical Engineering, Vol. 3, No. 2, 2009, pp. 230-239.
[5] Gerner M., Les assemblages des ossatures et charpentes en bois, Group Eyrolles, Paris, 190
pages, 2012. ISBN: 978 -2 -212-13620 - 3
[6] Sobon J.A., Historic American timber Joinery, a graphic guide, Timber Framers Guild, ,
Becket, MA 01223, 56 pages, 2012.
[7] Seike, k., The Art of Japanese Joinery, Weatherhill/Tankosha Publ., New York, 128 pages, 1977
[8] Meisel A., Moosbrugger T., Schickhofer G., “Survey and Realistic Modelling of Ancient
Austrian Roof Structures”, In: Proceedings of Conservation of Heritage Structure (CSHM-3),
Ottawa, Canada, 2010.
[9] SIA 265: Constructions en Bois. Swiss Society of Engineers and Architects, 2003.
[10] EN 1995-1:2005, Eurocode 5: Design of timber structures - Part 1-1: Common rules and rules
for buildings. Brussels, CEN, European Committee for Standardization, 2005
[11] Hewett C.A., English Historic Carpentry. London & Chichester: Phillimore & Co. Ltd., 1980,
p.270.
74
[12] Hirst E., Brett A., Thomson A. Walker P., Harris R., The Structural Performance of Traditional
Oak Tension & Scarf Joints, In: Proceedings of the 10th World Conference on Timber
Engineering, Miyazaki, Japan, 2008.
[13] Thelandersson, S., Larsen, H. J. Timber Engineering. Chichester: John Wiley & Sons Ltd., 346
pages, 2003.
[14] Larsen H., Jensen J., “Influence of semi-rigidity of joints on the behaviour of timber structures”,
Progress in Structural Engineering and Materials, Vol. 2, No. 3, 2000, pp. 276–277.
[15] Candelpergher L., Sperimentazione, modellazione numerica e caratterizzazione sintetica del
comportamento di collegamenti lignei tradizionali con elementi metallici, Master’s thesis,
Università degli Studi di Trento, 1999.
[16] Palma P., Cruz H., “Mechanical behaviour of traditional timber carpentry joints in service
conditions - results of monotonic tests”, In: Proceedings of “From material to Structure -
Mechanical behaviour and failures of the timber structures,” XVI International Symposium,
ICOMOS IWC, 2007.
[17] Uzielli L., Il manuale del Legno Strutturale, Vol IV - Interventi sulle strutture, Mancosu, Rome,
Italy, in Italian, 2004.
[18] Drdácký M., Wald F., Sokol, Z., “Sensitivity of historic timber structures to their joint
response”. In: Proceedings of the 40th Anniversary Congress of the IASS, Madrid, 1999.
[19] Descamps T., Lambion J., Laplume D. (2006), “Timber Structures: Rotational stiffness of
carpentry joints”, In: Proceedings of the 9th World Conference on Timber Engineering, Portland,
USA, 2006.
[20] Komatsu K., Kitamori A., Jung K. and Mori T., “Estimation of the Mechanical Properties of
Mud Shear Walls Subjecting to Lateral Shear Force”, In: Proceedings of the 11th Int. Conference
on Non-conventional Materials and Technologies, Bath, UK, 2009.
[21] Chang W.-S., Hsu M.-F., Komatsu K., “Rotational performance of traditional Nuki joints with
gap I: theory and verification”. Journal of Wood Sciences, Vol. 52, 2006, pp. 58–62.
[22] Wald F., Mares Z., Sokol M., Drdácký F, “Component Method for Historical Timber Joints”, In:
The Paramount Role of Joints into the Reliable Response of Structures. NATO Science Series
Vol. 4, 2000, pp. 417-424.
[23] Kasal B., Tannert T., "In Situ Assessment of Structural Timber", In: RILEM State-of-the-Art
Reports, Vol. 7, 2011, p. 129.
[24] UNI 11138, Cultural heritage - Wooden artefacts - Building load bearing structures - Criteria
for the preliminary evaluation, the design and the execution of works. UNI Milano, 2004.
[25] Aman R., West H., Cormier D. "An evaluation of loose tenon joint strength", Forest Products
Journal, Vol. 58, No. 3, 2008, pp. 61–64.
[26] Judd J., Fonseca F., Walker C., Thorley P., "Tensile strength of varied-angle mortise and tenon
connections in timber frames", Journal of Structural Engineering, Vol. 137, No. 5, 2012, pp.
636–644.
[27] Likos E., Haviarova E., Eckelman C., Erdil Y., Ozcifci A., "Effect of tenon geometry, grain
orientation, and shoulder on bending moment capacity and moment rotation characteristics of
mortise and tenon joints", Wood Fiber Sciences, Vol. 44, No. 4, 2012, pp. 1–8.
[28] Kock H., Eisenhut L., Seim W., "Multi-mode failure of form-fitting timber connections –
Experimentaland numerical studies on the tapered tenon joint", Engineering Structures ,Vol. 48,
2013, pp, 727–738.
[29] Feio A.O., Lourenço P.B., Machado J.S., "Testing and modeling of a traditional timber mortise
and tenon joint", Materials and Structures, Vol. 47, 2014, pp. 213–225.
[30] Götz K.-H., Hoor D., Möhler K., Natterer J., "Construire en Bois - Choisir, concevoir, realiser".
Presses Polytechniques et Universitaires Romandes, Lausanne, Switzerland, 1993.
[31] DIN 1052, Entwurf, Berechnung und Bemessung von Holzbauwerk. Allgemeine
bemessungsregeln und bemessungsregeln fur den hochbau, 2004.
[32] C.T.E., Documento Básico SEM. Seguridad estructural – Estructuras de madera. A código
técnico de la edificación, ministerio de vivienda, 2006
75
[33] Branco J, Cruz P, Varum H, Piazza M., "Portuguese traditional timber trusses. Static and
dynamic behaviour". Technical Report E-19/05, Guimarães, Portugal. (in Portuguese), 2005.
[34] Derinaldis P.P, Tampone G., "The Failure of the Timber Structures Caused by Incorrect Design-
Execution of the Joints. Two Cases Study", In: ICOMOS IWC, XVI International symposium –
Florence, Venice and Vicenza, 2007.
[35] Branco J.M., Influence of the joints stiffness in the monotonic and cyclic behaviour of
traditional timber trusses. Assessment of the efficacy of different strengthening techniques, PhD
thesis, University of Minho and University of Trento, 2008.
[36] DIN EN 1995-1:2005, NCI NA 6.8.3. National German Annex to Eurocode 5: Design of timber
structures - Part 1-1: Common rules and rules for buildings.Brussels, CEN, European
Committee for Standardization, 2005
[37] Yeomans D., The Repair of Historic Timber Structures, Thomas Telford Publishing, London pp.
147-150, 2003.
[38] EN 1998-1:2004. Eurocode 8: Design of structures for earthquake resistance- Part 1: General
rules. Brussels, CEN, European Committee for Standardization, 2004
[40] Tampone G., "Mechanical Failures of the Timber Structural Systems", In: ICOMOS IWC, XVI
International Symposium – Florence, Venice and Vicenza, 2007.
76
Reinforcement of connections with dowel type fasteners
5
Reinforcement of connections with dowel-type fasteners
Laurent Bléron1, Damien Lathuillière1, Thierry Descamps2, Jean-François Bocquet1
Summary
A good design of connections with dowel type fasteners is essential to ensure a safe design for the
whole structure and a cost effective solution. With reinforced connections, engineers can achieve better
capacity of the connection and safer designs by increasing the ductility of the connection (so-called
"capacity design" with a ductile connection to be the weak point). This paper presents an overview of
the reinforcement of dowel type connections. Most of the failures encountered in dowel type
connections occur because of an excess of shear stresses or tensile stresses perpendicular to the grain in
the connection area. Because of the large amount of possible details and reinforcement techniques, this
report mainly focuses on the reinforcement of dowel-type connections when the dowels are mainly
loaded parallel to the grain. For such connections, modes of failure of un-reinforced and reinforced
connections, updated design models and the effect of the reinforcement on the ductility of the
connection will be presented. Among the various types of reinforcement techniques available,
nowadays, self-tapping screws have found a wide field of application. They will be the main focus of
this chapter.
1. Introduction
In timber structures, the design of dowel type connections may govern the overall design of structural
members. Because of the low load bearing capacity of one fastener (when considered alone) relative
to the total load supported, designers have to design for numerous fasteners and deal with regulations
relating to the minimum spacing requirements. Because of the weakness of timber in shear and in
tension perpendicular to the grain, spacing requirements are mandatory to avoid any failure of
connections at very low load rates (compared to the ultimate load of the global structure). Spacing
requirements may lead to a global over sizing of the timber members and so, a non-effective solution.
Designing large connections (with large spacing between the dowels) may result in uneconomical
solutions and potentially lead to problems caused by shrinkage that is the result of variations in the
timbers moisture content. In that situation, the use of some types of reinforcement is an opportunity to
ensure a better design.
Most of the failures encountered in dowel type connections occur because of an excess of shear
stresses or tensile stresses perpendicular to the grain in the connection area. Because of the
embedment of the fastener in the wood, very complex stress states appear in the wood along the
contact surface (including shear stress and tension perpendicular to the grain). The relative importance
of those stresses depends on the angle between load and grain and different failure modes may occur
as presented in Fig. 1.
1)
University of Lorraine, ENSTIB-LERMAB, Epinal, France
2)
University of Mons, Mons, Belgium
77
Fig. 1 Failed embedment specimens tested at different loads to the grain angles: (a) dowel loaded
parallel to the grain, (d) dowel loaded perpendicular to the grain [1]
Various reinforcement techniques have been proposed among them, truss plates [2], [3], fibreglass
[4], [5], threaded rods [6], glued in rods and self-tapping screws (STS) [7] or glued-on wood-based
panels [8]. All of them mainly aim to avoid splitting of the timber in the connection area. Some
examples are presented in Fig. 2.
Fig. 2 (a) reinforcement using glued-on boards (b) reinforcement by punched metal plate fasteners
and (c) nail plates [8] (d) multiaxial stitch bonded fabric [9]
Haller et al. [9] have used multiaxial stitch bonded and biaxial weft knitted textiles (Fig. d). This
investigation concluded that even light textile reinforcements may significantly increase the strength,
the stiffness and the ductility of doweled connections. It appeared from embedding tests that loop-like
fibre placements achieved the highest embedding strength and stiffness, whereas biaxial knitted
fabrics result in more ductile connection behaviour. However, special attention has to be paid to any
solution involving the wrapping the timber connections in an airtight textile because of the risk of
decay.
Reinforcement helps to overcome timber weaknesses by increasing the shear strength and the tensile
strength perpendicular to the grain and by stopping the propagation of cracks. This report mainly
focuses on the reinforcement of dowel-type connections when the dowels are loaded parallel to the
grain (Fig. 1a). For such connections, modes of failure of unreinforced and reinforced connections,
proposed design models and the effect of the reinforcement on the ductility of the connection will be
discussed.
Situations where dowels are loaded perpendicular to the grain are not developed to a large extent here
after. One may notice that in that situation reinforcements (screws, glued in rods or a surface
reinforcement as a plate) prevent crack propagation (Fig. 3). This kind of reinforcement has been
studied by several authors, among others, Franke et al. [10], Borth et al. [11], Schoenmarkers et al.
78
[12], Jensen et al. [13] & Leijten [14]. Schoenmakers et al. [15] have studied the reinforcement by
screws and tested several configurations (different relative heights h and loaded edge distances he) as
presented in Fig. 3. Blass et al. proposed a design model that assumes that timber has no tensile
strength perpendicular to the grain [16].
Fig. 3 Reinforcement with single plate (a) and one main plate and stitch plates (b) to prevent cracks.
Reinforcement with screws (c)
Initially developed for assembling elements (member or connections), STS have found a new field of
application with the reinforcement of timber beams or timber connections. The benefits of using STS
are generally their efficiency (as reinforcement), their low costs (cheaper than other reinforcement
materials) and the fact that they are easy to use (the setup is less complex than for glued in rods for
example). Moreover, the use of STS does not require any surface preparation or pre-drilling (for the
smaller diameters) and new screw-guns enable fast and effective installation of STS. Reinforced with
STS, connections may behave plastically until failure (because any splitting of wood is prevented).
STS, contrary to traditional wood screws, are fully threaded and made with high tensile strength steel.
They can bear high axial loads (even with a small diameter, reducing the danger of the splitting of
wood) and have high withdrawal strength due to their improved thread geometry. Due to the rapid
growth in the use of STS for reinforcements, they will be the main focus of this chapter.
2. Reinforcement of dowel-type connection with self-tapping screws

2.1 Modes of failure of an unreinforced dowel-type connection
The different failure modes of a connection made with dowel-type fasteners are presented in Fig. 4.
Five failure modes are indicated depending on the connection geometry: edge and end distances of the
fasteners, spacing, the fastener diameter (in relation to the thickness of timber) and the number of
fasteners.
The failure mode (a) is ductile because it is characterized by the embedment of the fasteners into the
wood, the others are not [8], [17], [18]. Timber may show a tendency to split in the connection area
before the embedding strength is reached according to timber thickness, diameter and number of
dowels, load to the grain angle, spacing as well as the end and edge distances of the dowels.
2.2 Self-tapping screws
STS with the thread running along the full length (Fig. 5i) and screwed perpendicular to the grain
increase the local strength in shear and tension perpendicular to the grain so they counteract the
tendency of timber to split. This reinforcement is achieved by ensuring that the tensile stresses
79
perpendicular to the grain (or shear stresses) are mainly supported by the screws and not by the timber
(because the screws embedded into wood are much stiffer than wood) [19]. By protecting against any
premature brittle failure with this technique, the connection failure modes coincide with the ductile
failure modes described by Johansen’s theory (no need for a reduced effective number of fasteners)
[20], [21].
Fig. 4 Different failures modes of unreinforced dowel-type connections
2.3 Self-tapping screws

STS with the thread running along the full length (Fig. 5i) and screwed perpendicular to the grain
increase the local strength in shear and tension perpendicular to the grain so they counteract the
tendency of timber to split. This reinforcement is achieved by ensuring that the tensile stresses
perpendicular to the grain (or shear stresses) are mainly supported by the screws and not by the timber
(because the screws embedded into wood are much stiffer than wood) [19]. By protecting against any
premature brittle failure with this technique, the connection failure modes coincide with the ductile
failure modes described by Johansen’s theory (no need for a reduced effective number of fasteners)
[20], [21].
STS may also increase the load-carrying capacity of one fastener beyond the values defined by
Johansen’s yield theory by increasing the embedding strength. For that purpose, STS screwed
perpendicular into the grain and perpendicular to the fasteners (to be reinforced) are the best. The
tendency of timber to split is minimised by using STS and as a consequence the load carrying capacity
will be increased (Fig. 5ii and 5iii).
Fig. 5 (i) full threaded self-tapping screws (STS), (ii) reinforcement of timber with STS reducing the
wood's tendency to split (iii) reinforcement of timber and fastener embedding strength by placing the
self-tapping screws in contact with the dowel-type fasteners
80
Because of the high tensile strength of steel STS and their special thread, the connection behaviour is
no longer governed by the low tension perpendicular to the grain and shear strength of timber. Bejtka
et al. & Blass et al. first studied this topic [19], [16]. Timber splitting is prevented when the axial
load-carrying capacity of each screw is larger than 30% of the lateral load-carrying capacity per shear
plane of each dowel (calculated according to Johansen´s yield theory). The screws of course prevent
the splitting of timber, but also increase the embedding strength of the fasteners. Some experimental
results are presented in Fig. 6. Two important points can be observed, firstly the increase of the load
bearing capacity and secondly the important ductility of the reinforced connection.
Fig. 6 Typical load-displacement-curves of non-reinforced and reinforced connections [8]

(a) non-reinforced – (b) reinforced with screws without contact with the dowel-type fasteners
– (c) reinforced with screws with contact with the dowel-type fasteners
Fig. 7 (a) example of bolted connections with a slotted in steel plate, (b) example of reinforced
connection layout with reduced spacing, (c) full size sample test [24]
Bejtka et al. [19] studied the influence of the position of the reinforcement in relation to the fasteners
that have to be reinforced. Echavarria [22] has studied the efficiency of STS in stopping the spread of
the cracks (Fig. 5ii) and the related increase in load carrying capacity of the fastener. This study has
been made for a single bolt assembly. The main benefit of this method is the reduced end-distance
without any reduction of the load-carrying capacity of the connection, which leads to a more compact
connection. Tab. 1 shows the ratio between the reinforced and non-reinforced load-carrying capacity
based on tests performed by Echavarria for different geometrical configurations. One can notice that
81
the effect of the reinforcement is maximal for small e/d (small loaded end distance/dowel diameter –
see Fig. 5).
Bocquet et al. [26] have studied block shear failures (Fig. 4d) of dowelled connections in bending and
have proposed design models. Different connections with rectangular and circular patterns have been
studied. Regarding the block shear failure, the tests conducted showed that an optimal number of
reinforcements and an appropriate positioning of STS are required in order to achieve excellent results
(efficient reinforcement). On the one hand, a minimum amount of reinforcement uniformly distributed
over the connection area is necessary in order to prevent splitting. On the other hand, too many
reinforcements can affect the mechanical behaviour of the connection by weakening the timber beams
(reduced cross section). This is especially true if the reinforcements are located where the stresses in
timber are the highest.
Tab. 1 Ratio of reinforced and non-reinforced load-carrying capacity [14]
Bolt diameter Thickness Reinforced load- Non-reinforced Ratio of reinforced and

e/d carrying load-carrying non-reinforced load-
[mm] [mm] capacity [kN] capacity [kN] carrying capacity [%]
15.9 2 38 9.8 4.4 120%
15.9 3 38 12.7 12.1 4.7%
15.9 4 38 17.4 17.3 0.8%
15.9 5 38 18.4 18.2 1.0%
Gehloff et al. [24] enabled the use of STS both to reduce the edge distance of bolts and to increase the
bending capacity of a moment resisting timber connection (Fig. 7). It was observed that STS
prevented early splitting of the members even though small edge distances of bolt were applied (even
for stocky dowels). An important decrease of the edge distance (up to a half) in relation to the
reference connection (Fig. 7a) resulted in an increase in ultimate moment capacity by a factor of about
1.3. A significant increase in connection capacity could not be achieved when placing the STS near
the bolts. One may also notice that based on cyclic loadings, placing the STS close to the bolts reduce
the standard deviation and might have a positive impact on the energy dissipation.
2.4 Proposed design models

Bejtka et al. [19] have developed a calculation model based on Johansen’s yield theory. The STS are
placed in contact with the dowels, perpendicular to the dowel axis and to the grain direction (Fig.iii).
They are loaded just as the dowels themselves perpendicular to their axis. As for Johansen´s yield
theory, all components (STS, timber and fasteners) have ideal elastic-plastic behaviour. The STS
buckle perpendicular to their axes and in the force direction, when the dowel load component FVE
reaches the load-carrying capacity RVE of the STS. In this case, the screw behaves as a “soft” support
(Fig. 8).
82
Fig. 8 (a) one dowel reinforced by one STS; (b) two dowels reinforced by one STS
Alternatively, for FVE < RVE, the screw does not move and represents a rigid support for the dowel. This
consideration leads to four sub-failure modes for each failure mode in timber-to-timber connections and
two sub-failure modes for each failure mode in steel-to-timber connections in Johansen´s yield theory.
As an example, Fig. 9a and Fig. 9b represent, for steel to timber connections, the sub-failure modes of
failure mode 3 (inner steel plate with two plastic hinges). The efficiency of the reinforcement depends
on its position compared to the position of the plastic hinge. According to the distance p of the STS
relative to the dowel and the distance x between the shear plane and the plastic hinge in the
unreinforced connection (see Fig. 9), the STS does not reinforce the connection (if p>x) or on the
contrary the reinforcement increases the resistance of the connection (if p <x). In the latter case,
different sub-failure modes “rigid” or “soft” may occur. Eq. 1 & 2 (moment equilibrium in the shear
plane) give the maximum load Fv for the two sub-failure modes illustrated on Fig. 9a and Fig. 9b. Eq.
3 is for the regular maximum load of the unreinforced connection:
2 M y fh  d  p
FV   if FVE  RVE (1)
p 2
FV  RVE  2  f h  d   2  M y  RVE  p  if FVE  RVE (2)
FV  2  2  M y  f h  d (3)
Fig. 9 Steel to timber connections with “rigid” sub-failure mode (a) and “soft” sub-failure mode (b).
Timber to timber connections with sub-failure modes “soft” (c), “rigid” (d), “soft-rigid” (e) and
“rigid-soft” (f). These additional sub-failure modes (e) & (f) are encountered in timber to timber
connection when the load-carrying capacities of the STS or/and the load components FVE in both
timber members are not equal (different timber density in both timber members)
83
All developments can be found in [19]. Modified Johansen’s equations for timber to timber reinforced
connections are not developed here. The calculation model is a little bit more complicated because
mixed-mode sub-failure modes have to be considered (Fig. 9e and Fig 9f). The modified Johansen’s
equations proposed by Bejtka et al. introduce a new parameter: the load-carrying capacity RVE of the
STS. Of course, for RVE = 0 the load carrying capacity is equal to the load carrying capacity for a non-
reinforced dowel. Just for illustration, the load-carrying capacity for the sub-failure mode “soft” of a
timber to timber connection, derived from the force and moment equilibrium in the shear plane, is
represented in Eq. 4. All others sub-failure modes can be found in [19].
R 2    1
2
  2 
FV ,rein  R1,VE    f h,1  d   2  M y  R1,VE  p  1     1,VE (4)
1  1  2  1   
with R1,VE and R2,VE the load-carrying capacity RVE of the STS in timber 1 and timber 2, respectively,
fh the embedding strength and My the plastic moment of the dowel. In addition:
f h,2 R2,VE
  (5)
f h,1 R1,VE
The load-carrying capacity Ri,VE of the STS is derived and calculated according to Johansen´s yield
theory as for steel-to-timber connections with inner steel plates. For the case of one dowel being
reinforced by one screw (Fig. 8a), RVE can be calculated as follows:
 f h, STS  d STS  lSTS 

 
  16  M y , STS 
RVE  min  f h , STS  d STS  lSTS    2  1  (6)
 f h , STS  d STS  lSTS
2
  
 4  M y , STS  f h , STS  d STS 
 
When two adjacent dowels are reinforced by one screw (Fig. 8b), the load-carrying capacity RVE of the
STS can be calculated as follow:
 0.5  f h ,STS  d STS  lSTS 
   
 f h ,STS  d STS  lSTS  
2
16  M y ,STS  a2  a2 
   2    1  2  1 
 f d l
2
lSTS 
 2
 h ,STS STS STS  lSTS   
 
 f h ,STS  d STS  a2 
 2  2  M y ,STS  f h,STS  d STS
 2  (7)
 
 
RVE  min 4  M y ,STS  f h,STS  d STS 
  2 
 f h ,STS  d STS  lSTS  16  M y ,STS  a  
 2  1 
M y ,STS a
   2   2  1  4 
 d l
2
f h,STS  d STS  lSTS
2 
 2 f
 h,STS STS STS  lSTS  lSTS
 
 
  2  
 f h ,STS  d STS  lSTS  8  M y ,STS  a2  a2  
     1  1 
 f  
2 

2 d l  STS
l  l

 
h , STS STS STS STS
84
3. Effect of the reinforcements on the ductility of the connection

As timber is a brittle material in tension and shear, connections are a good means to introduce ductility
into the structure. The ductility is of major importance in seismic design, for the redistribution of
internal forces in non-statically determinate structures or just to improve the robustness of the structure.
Lam et al. [25] have studied bolted connections reinforced with STS screwed perpendicular to the grain
(Fig. 10) under monotonic and cyclic bending. Experimental results have shown that the reinforcement
has increased the ductility and the moment capacity of the tested connections. Lam et al. also studied
the efficiency of STS to repair damaged connections by testing unreinforced connections repaired with
STS.
Fig. 10 Example of failures under monotonic loading: (a) unreinforced specimen (b) reinforced
specimen, (c) retrofitted connection specimens(reinforced with STS) [25]
Fig. 11 STS: test results of non-reinforced (M1–M4) vs. reinforced (M5–M10) connections [8]
85
More recently, Blass [8] confirmed the increase in the ductility of reinforced connections, by
conducting many tests on reinforced bolted connections. The results presented on Fig. 11 show that
the displacement at the point of failure in the reinforced connection can increase up to twice as much
as in the unreinforced connection. The improvement in the ductility of the connection under
monotonic loading may also be observed to some extent under cyclic loading.
4. Conclusions
For all connections, reinforcement helps to overcome timber weaknesses by increasing the shear
strength and the tensile strength perpendicular to the grain, also helping to reduce the propagation of
cracks. Reinforcement may help to achieve several aims, for example, increasing the load bearing
capacity, reducing the spacing or increasing the ductility.
Failures that occur in dowel type connections are mostly brittle (no embedding failures). Today
various reinforcement techniques exist (fibreglass, textile, truss plates, threaded rods…) but among all
of them self-tapping screws are the most widespread because of their efficiency and their low cost.
The study of reinforced connections with self-tapping screws is new and not yet included in EN 1995-
1-1 [26] but only in the National Annexes of some countries as Germany [27] or Austria [28].
Investigations on that promising topic have helped to figure out how to overcome timber weaknesses
in connections and have resulted in the proposal of design models. The most important and applicable
outcomes will probably be integrated in the revision of Eurocode 5 to help engineers to optimise their
structures during building design [29].
5. References
[1] Pedersen, M.B.U., Dowel type timber connections. Strength modelling, Ph.D. thesis, Technical
University of Denmark, Report No. BYG.DTU R-039, 2002.
[2] Mastschuch, R., Reinforced multiple bolt timber connections, Ph.D thesis, University of British
Columbia, Vancouver, Canada, 2000.
[3] Hockey, B., Lam, F., Prion, H.G., “Truss plate reinforced bolted connections in parallel strand
lumber,” Canadian Journal of Civil Engineering, Vol. 27, No. 6, 2000, pp. 1150–1161.
[4] Chen, C.J., Mechanical behaviour of fiberglass reinforced timber joints, Ph.D thesis, Swiss
Federal Institue of Technology Lausanne, Switzerland, 1999.
[5] L. A. Soltis, L.A., Ross, R.J., Windorski, D.F., “Fiberglass-reinforced bolted wood
connections,” Forest Products Journal, Vol. 48, No. 9, 1998, pp. 63–67.
[6] Quenneville, J.H.P., Mohammad, M., “Anti-check bolts as means of repair for damaged split
ring connections,” In: Proceedings of World Conference on Timber Engineering, Whistler,
Canada, 2000.
[7] Hansen, K.F., “Mechanical properties of self-tapping screws and nail in wood,” Canadian
Journal of Civil Engineering, Vol. 29, 2002, pp. 725–733.
[8] Blass, H.J., P. Schädle, P., “Ductility aspects of reinforced and non-reinforced timber joints,”
Engineering Structures, Vol. 33, No. 11, 2011, pp. 3018–3026.
[9] Haller, P., Birk, T., “Tailor made textile reinforcements for Timber Connection”, In:
Proceedings of World Conference on Timber Engineering, Portland, USA, 2006.
[10] Franke, B., Quenneville, P., “Analyses of the failure behaviour of transversely loaded dowel
type connections in wood,” In: Proceedings of World Conference on Timber Engineering,
Trento, Italy, 2010.
86
[11] Borth, O., Schober, K.U., Rautenstrauch, K.,“Load-carrying capacity of perpendicular to the
grain loaded timber joints with multiple fasteners,” In: Proceedings of International Council for
Building Research and Innovation Working 18, Kyoto, Japan, 2002.
[12] Schoenmakers, I.J.D., Jorissen, A.J.M., Leijten, A.J.M., “Splitting behaviour of timber loaded
perpendicular to the grain by punched metal plates,” In: Proceedings of World Conference on
Timber Engineering, Miyazaki, Japan, 2008.
[13] Jensen, J.L., “Splitting strength of beams loaded by connections,” In: Proceedings of
International Council for Building Research and Innovation CIB-W18, Colorado, USA, 2003.
[14] Leijten, A. J. M. , “Splitting strength of beams loaded by connections, model comparison,” In:
Proceedings of International Council for Building Research and Innovation CIB-W18, Kyoto,
Japan, 2002.
[15] Schoenmakers, J. C. M., Fracture and failure mechanisms in timber loaded perpendicular to the
grain by mechanical connections, Ph.D thesis, Eindhoven University of Technology, 2010.
[16] Blass, H.J., Bejtka, I., "Reinforcements perpendicular to the grain using self-tapping screws", In:
Proceedings of the 8th World Conference on Timber Engineering, Lahti, Finland, 2004.
[17] Fahlbusch, H., Ein Beitrag zur Frage der Tragfähigkeit von Bolzen in Holz bei statischer
Belastung. Hunold, 1949.
[18] Quenneville, J.H.P., Mohammad, M., “On the failure modes and strength of steel-wood-steel
bolted timber connections loaded parallel-to-grain,” Canadian Journal of Civil Engineering,
Vol. 27, No. 4, 2000, pp. 761–773.
[19] Bejtka, I., Blass, H.J., “Self-tapping screws as reinforcements in connections with dowel-type
fasteners,” In: Proceedings of International Council for Building Research and Innovation CIB-
W18, Karlsruhe, Germany, 2005.
[20] Jorissen, A., Double shear timber connections with dowel type fasteners, Ph.D thesis, University
of technology, Delft, Netherlands, 1998.
[21] Johansen, K.W., Theory of timber connections, Publication 9, International Association of
Bridge and Structural Engineering, Bern, pp 249–262, 1949.
[22] Echavarría, C., “Bolted timber joints with self-tapping screws,” Revista EIA, No. 8, pp. 37–47,
2007.
[23] Bocquet, J.B., Barthram, C., Pineur, A., “L block failure of dowelled connections subject to
bending reinforced with threaded rods,” In: Proceedings of World Conference on Timber
[24] Gehloff, M., Closen, M., Lam, F., “Reduced edge distances in bolted timber moment
connections with perpendicular to grain reinforcements,” In: Proceedings of World Conference
on Timber Engineering, Trento, Italy, 2010.
[25] Lam, F., Wrede, M.C., Yao, C.C., Gu, J.J., “Moment resistance of bolted timber connections
with perpendicular to grain reinforcements,” In: Proceedings of World Conference on Timber
[26] EN 1995-1-1, EC5 - Eurocode - Design of timber structures - Part 1-1: General -Common rules
and rules for buildings. 2005.
[27] DIN EN 1995-1-1/NA, EC5 - Eurocode - Design of timber structures - Part 1-1: General -
Common rules and rules for buildings German National Annex. .
[28] AT EN 1995-1-1/NA, EC5 - Eurocode - Design of timber structures - Part 1-1: General -
Common rules and rules for buildings Austria National Annex.
[29] CEN TC 250 SC5, N300 Report from the WG on reinforcement of timber structures, CEN,
Brussels, 2013.
87
88
Seismic strengthening of timber structures
6
Maria Adelaide Parisi1, Maurizio Piazza2
Summary
In European seismic areas timber structures are found as building frames, in combination with masonry
infills, in bridges, but most frequently in roof structures and floor slabs of traditional buildings. In the
seismic design of new timber structures, members are elastic and ductility resources are concentrated in
the joints. Similarly, seismic strengthening of existing structures should provide a well-defined and
simple path to seismic forces, maintain timber members elastic, and develop as much as possible the
post-elastic behaviour of joints. Provisions must be adopted to avoid sudden loss of capacity and brittle
failure, and to foster ductility. Different criteria for seismic strengthening of floor slabs and of
carpentry joints are presented.
1. Introduction
The development or the updating of seismic zonation in a country may require reconsideration of
existing structures that were built for lower levels of seismic action, or even without seismic
provisions, in order to comply with current safety requirements. In Italy, for instance, the building
code [1] requires that buildings undergoing major renovation or a change of use be checked also for
earthquake loading. Reference is made to the seismic action to be adopted in the design of new
buildings. This rule aims at the progressive upgrading of the significant fraction of the building stock
that was originally built without adequate consideration of seismicity. In the specific case of timber
structures, plain timber sections and carpentry joints, which were the basis of traditional construction,
may not satisfy the new requirements; seismic strengthening may become necessary in order to
continue their use.
In Europe, the regions affected by the highest seismicity are in the south, where buildings completely
erected in timber are not frequent. Yet, in traditional construction, including common buildings as
well as historical architecture and monuments, timber has been used extensively to build structures for
supporting floors and roofs. This chapter will focus particularly on such structures and their
components.
Timber has been used since Antiquity in the Mediterranean region to improve the seismic response of
masonry, as witnessed by findings at archaeological sites, e.g. [2]. Different arrangements of the two
materials have been developed in time, corresponding to different views of the way collaboration
should be accomplished. They range, geographically and formally, from the infilled timber frames of
the Pombaline architecture in Portugal, e.g. [3], through the timber framing of southern Italian “case
1)
Associate Professor, Politecnico di Milano, Milano, Italy
2)
Professor, Università degli Studi di Trento, Trento, Italy
89
baraccate”, [4], to the timber frame-and-wall system found in Greece in the island of Leukada where
a secondary frame parallels the stone masonry walls [5], again to Greece [6] and to the many solutions
offered by Turkey, e.g. [7]. The close interaction between timber and masonry that is inherent in all
these systems leads to their classification as composite structures rather than timber structures. Their
seismic capacity may be deemed insufficient, in spite of the original intention of the builders to
withstand strong earthquakes or because of reduced capacity due to damage and decay. Provisions for
improving them require considering features specific of each type, e.g. [8].
Traditional full-timber buildings in many different forms and structural types are found most
frequently in Turkey, both for residential use [9] and for significant religious sites [10]. Some are
outstanding for their constructive boldness and beauty, like for instance the Prinkipo palace [11] at the
Princess island in the Marmara Sea, as shown in Fig. 1. The 6-storey high building is completely built
in wood, in its main structural parts as well as in its secondary walls and divisions, and in its
decorations, presenting an encyclopaedic collection of elements and techniques constructed in timber.
Unfortunately, the cultural value of the Prinkipo palace and of some of these buildings has not been
recognized early enough and they are highly damaged. The main issue is their survival rather than
their seismic resistance, which could be, in any case, attained with the criteria for interventions
described here.
Fig. 1 The Prinkipo palace (left, the façade;right, interior view).
Bridges, which are also traditional timber structures, are present in a small number in European
seismic areas and are generally considered as heritage structures. Although they may require special
attention and treatment, the considerations developed in the following may be extended to their case.
Common criteria for strengthening interventions usually derive from constructional tradition. Yet,
originally strengthening methods did not address the problem of seismic response but would, rather,
concern malfunctions for common vertical loads or to counter deterioration. Their effectiveness
toward seismic actions needs to be confirmed in order to their application.
New intervention technologies, often based on the application of advanced materials, like polymers
reinforced with fibres of different kind, have been proposed also for timber structures. Some of their
mechanical characteristics and easy implementation make them particularly appealing for
strengthening existing structures. The interventions, again, must comply with criteria and
requirements suitable for seismic conditions. It must be noted that in general interventions for
90
rehabilitation or structural upgrading may be similar in the seismic and non-seismic case. In the
former case, however, interventions must also adhere to principles stemming from the basic
philosophy of seismic design, discussed in the following.
Often, existing timber structures belong to buildings listed as part of the cultural heritage of a
Country: as such, they are subjected to conservation requisites. All structural upgrading operations
need to comply with criteria for restoration of heritage as well as safety. In this perspective, according
to a categorization commonly used in earthquake engineering, the interventions to reduce seismic
vulnerability may be distinguished between,
- seismic strengthening, which technically corresponds to putting the structure in the
condition to withstand a design earthquake of exceptional level like a newly designed one
(sometimes the term upgrading is also used), and
- seismic improvement, which is a milder, less invasive, and more local intervention
intended to eliminate possible criticalities without pointing at a general upgrading of the
structural capacity to the same level as in new designs.
The approach of seismic improvement has been deemed suitable for operations on cultural heritage
also by national Authorities in the field [12]. Nowadays, many recent studies and observations are
pointing out that less invasive interventions may yield better results even in proper seismic
strengthening. In the following, the term “strengthening” will be mainly used in the more general and
generic sense of structural upgrading, yet the distinction of strengthening versus improvement in the
sense above will be recalled when necessary.
In the following, general criteria are first presented; timber structures are then examined as a system
and in their components, indicating which situations may be critical in seismic conditions and what
interventions may be most profitably performed.
2. Strengthening Criteria
The properties of wood as a construction material may influence the seismic response of the structure
positively, because the low density generates fairly small inertia forces, and negatively because of
brittleness, which is particularly marked perpendicular to the fibre/grain direction. When designing
new structures, the material behaviour can be exploited optimizing the favourable contributions and
providing for the drawbacks by choosing suitable structural types and details. For new timber
structures, design codes and norms indicate the approach to be followed in order to obtain a suitable
seismic behaviour. Basically, the current philosophy for seismic design entails exploiting post-elastic
resources to an extent reasonable for the structure, depending on material, type, and construction
details. For timber structures, given the very limited ductility offered by the material, wood
components are designed to remain elastic. Modern connections are based on the load-carrying
contribution of steel connectors. Post-elastic yielding may then be developed in the joints, with
contributions from steel ductility and from the embedding of timber interacting with steel.
Little or no indications are given in structural design codes for general strengthening of existing timber
structures, and especially for seismic strengthening. Eurocode 8 for earthquake resistant design [13]
does not refer to timber in its Part 3, which deals with existing buildings, nor does Eurocode 5 [14].
Yet, using as a paradigm the design of new structures, the same approach may be adopted for
upgrading the existing ones. Interventions should,
91
- supply a well-defined and simple path for seismic forces;

- maintain the timber members in an elastic state, especially avoiding local stress
concentrations;
- address special attention to joints, protecting them from sudden loss of capacity and brittle
failure, and providing them as much as possible with post-elastic behaviour capabilities.
In this perspective, it must be considered that interventions suitable for common static loading may be
inappropriate in the case of seismic action if they do not implement the basic principles of seismic
design. For instance, in wood, the local increase of stiffness that may be acceptable in static situations
is most often associated to a reduction of ductility, which is a fundamental property in seismic
conditions.
In the last decades of the 20th century, in Italy, timber roof structures were either strongly reinforced
by means of massive interventions or, more likely, removed and substituted. The new parts were
usually not timber elements but industrial products fabricated with different and more massive
materials. Most frequently, prefabricated concrete elements were used. The increase of mass and
stiffness could be incompatible with the old and weak masonry walls underneath. In a number of
cases where earthquakes arrived after strengthening, damage and even progressive collapse of the
whole building took place, demonstrating the inadequacy of the works that had been performed. In the
case of Fig. 2, timber trusses had been replaced with concrete trusses.
This outcome, together with a growing cultural

trend recognizing the value of older construction
even beyond the case of protected heritage
buildings has been acknowledged in the current
issue of the Italian Building Code [1], for which
timber roof structures should be preferably
preserved, improving their seismic behaviour by
non-massive interventions when necessary. This
regulation implies defining criteria for effective
interventions, covering different possible
Fig. 2 Collapse in Cesi, Umbria-Marche components and situations.
earthquake, 1997, photo courtesy of V. Petrini.
Interventions on timber floor structures have also evolved with time. While timber roofs have
received limited attention and only in fairly recent times, the strengthening of slabs is more
established and documented. Research programs have been carried out internationally for decades and
have yielded results that are now being regularly applied in practice. The strengthening of timber
floors aims at increasing the mechanical properties of the slab, but keeping also in view the
compatibility between the horizontal floor structure and the vertical walls system. Here again, in
many cases, when the quality of masonry was particularly poor, an excessive stiffening of the slab had
proven harmful for the masonry walls and the overall building behaviour. Research is now focusing
particularly on how to increase the mechanical properties and improve the compatibility of the two
systems, without introducing local irregularities and excessive strength and stiffness that may trigger
failure modes; it points to offer a wide range of solutions that may fit different situations.
92
From the point of view of preservation of timber structures, much work has been carried out in order
to guide the design of interventions on timber cultural heritage. In particular, Guidelines for
interventions on wooden artefacts have been issued by UNI, the Italian Organization for
Standardization as UNI 11138 [15], while a document developed as a result of a COST action, [16]
offers guidance for the assessment of such structures. These documents are not specifically prepared
in relation to seismic situations, but may be considered a basis for developing heritage compatible
interventions in the line of seismic improvement as defined above.
3. Problems and Remedies

The development of earthquake engineering, particularly the identification of structural problems and
the proposal of solutions, has always started from the observation of seismic damage. For timber
structures, however, the documentation collected in post-event recognition campaigns is usually
limited to the most outstanding cases of global collapse. This may be due to the difficulty in reaching
and examining the timber structures within a damaged building, but probably also to the secondary
and temporary status that has always been attributed to these structures. Traditionally, timber
structures have been considered to some extent provisional, and worth replacing rather than repairing
whenever refurbishment or extended maintenance was due. Their role, as outlined above, is
nevertheless decisive for the response.
In order to examine systematically the different causes of poor behaviour and to propose remedial or
preventive interventions, timber structures may be examined as a whole system and in their
components, that is, members and joints. This progression is detailed in the following for roofs and
slabs, but it has a more general value and may be applied to other structural types, like timber frames,
scaling down from the problems that may be inherent in the structural concept, to the state and
capacity of members, and to their connections.
3.1 The level of the structure
Timber roof structures and floor slabs were typically built to bear vertical loads, with no provisions
for responding to seismic actions in the horizontal direction. Wind on roofs, especially in southern
Europe where pitch angles are low, acts mainly in suction; although it may be often destructive for
inadequate structures, it widely differs in its nature and effects from the earthquake as a phenomenon.
The first aspect that should be considered is whether the structural scheme can develop equilibrium
for horizontal forces. Static schemes may be insufficient because of construction errors, when
elements required for horizontal equilibrium are missing, or for the presence of unrestrained thrusts
that impinge on building components, like masonry walls, which are unable to supply the necessary
restraint. Many old roof structures have been built by craftsmen without a formal design. In open
couple roofs, depending on the detailing of supports, the rafters may apply a horizontal thrust on the
walls even under vertical loads; the situation worsens and may be particularly critical when horizontal
loads are also present.
In some trusses, joints with low moment transmission, close to a hinge, supply only a low rotational
restraint. The resulting structural scheme is almost unstable and unfit for asymmetric loads. The low
joint stiffness permits to accommodate some asymmetry in the vertical loads or respond to limited
accidental lateral forces but collapse would very likely develop for higher horizontal forces due to an
earthquake. An example of low-stiffness joint may be seen in Fig. 3.
93
Fig. 3 Truss with low rotational rigidity joints (photo courtesy of R. Tomasi); depending on the joint
stiffness, the structure could become unstable under asymmetric loads.
A case of a statically inadequate scheme was the

19th century roof of a church in Vimercate, Italy
(Fig. 4). The structure consisted in a series of trusses
interconnected by purlins and diagonal bracings. In
the original structure, one truss was missing,
creating a discontinuity. Additionally, from the first
truss an almost horizontal bracing element pointed
directly onto the façade wall in the tympanum area,
inducing an extremely high risk of impinging on the
most delicate area of the wall in seismic conditions.
Fig. 4 Truss with diagonal bracing element (photo
courtesy of C. Tardini).
Fig. 5 Pounding of the ridge beam into the tympanum during earthquake motion may result in its
collapse in a pattern similar to the above; here, other damage causes were present (S. Biagio
Amiterno, L’Aquila earthquake, 2009).
This situation, deriving from a conceptual error, not considering horizontal effects, was promptly corrected
by eliminating the impinging element, separating the truss from the tympanum, and creating a different
horizontal retaining system. Also, a truss was added at the proper location, reconstituting regularity in the
layout.
Problems may stem also from structural schemes that are formally equilibrated in the static sense, but
supply an unsatisfactory dynamic response. From the seismic point of view, a robust structure should
exhibit a sound response in different directions of excitation.
94
The case of structures composed of parallel plane trusses, conceived as a bi-dimensional system,
mildly interlinked to form a three-dimensional one, is very common for instance in heritage church
buildings. These structures may not be sufficiently robust. Usually, timber trusses are quite stiff in
their plane, but may be loosely interconnected. As a result, the roof system may be rather deformable
in the direction normal to the trusses and may develop excessive displacements under seismic motion
[17]. Consequently, pounding of the ridge beam into the gable wall or tympanum and collapse of the
roof-wall system have often occurred (Fig. 5). These causes of malfunctioning due to the structure
configuration are to be recognized and may be remediated, e.g. [18, 19].
A further critical point for a good seismic behaviour is the quality of the supports; even if it may not
be strictly considered a problem related to timber. Roof structures falling off supports during
earthquakes are not rare and often entail progressive collapse of the whole building. An effective
restraint from the supports must be sought and provided as a first step in the seismic strengthening of
timber roof structures.
The reinforcing of timber floors is always performed when an old building has to be rehabilitated for
new use or occupancy and takes on special characteristics in the cases of seismic strengthening or
seismic improvement. For vertical loads, in fact, beyond repairing possible decayed elements, the
issue is usually to upgrade the flexural stiffness and capacity, e.g. [20, 21]. Seismic conditions usually
require raising the in-plane mechanical characteristics of the floor and improving collaboration of the
horizontal structure with the vertical walls system. Caution must be exerted, because in some
situations, when the quality of masonry is particularly poor, excessive stiffening and increased mass
may be dangerous for the walls and the overall building behaviour. Additionally, the different
redistribution of horizontal forces among walls needs be controlled, in order to avoid possible new
parasitic effects from eccentricities. The question of how to attain a suitable level of mechanical
properties and collaboration for the floor without introducing irregularities and overstrength that may
trigger failure modes in other elements is the subject of much study and debate at the moment.
Specific conservation requirements for the original slabs in monumental buildings may also exist,
limiting the choices of intervention.
In a normal case, without special limitations, it is worth considering a whole range of possibilities to
achieve different levels of mechanical characteristics. A particularly effective strengthening method is
the formation of a composite system associating new collaborating elements, for instance a new slab, to
the existing timber structure. Substantial increase of the mechanical characteristics may be attained. In
this case, the possibility of composite timber–concrete structures and of timber-timber structures offers
two groups of solutions, each producing a range of mechanical properties for the floor structure
depending on the type of connection adopted, e.g. [22, 23, 24, 25]. A main difference is in the use of dry
or grouted connectors and in the shape and distribution of these elements, usually in steel, that have the
task of consolidating the beams with the newly produced slab. Grouting with resins forms a very
efficient connection, but may have some inconveniences, particularly in terms of restoration issues.
A fully functioning connection should provide full cooperation with the slab augmenting the beam
cross-section into a T-shape. Yet, collaboration is always partial, to a more or less extended degree. In
all cases, an efficient connecting system is the key to obtain a favourable behaviour.
Fig. 6 shows some examples for a timber-concrete composite structure.
95
Fig. 6 Glued-in connectors, with various arrangement, used for assembling timber – concrete
composite structures
Recent research programs have proposed various solutions particularly in view of improving the in-
plane behaviour. Besides the addition of a concrete slab, or of a timber slab in the cross-wise
96
direction, the application of a lattice of FRP tapes, or of metal bands has been examined, in order to
acquire stiffness without excessive material addition, as shown in Fig. 7. Cross laminated timber
(CLT) toppings are not represented in the figure, considering that a CLT slab, not very thin, may not
be able to adhere properly to an existing, usually sagged beam.
Fig. 7. Different in plane shear strengthening techniques for timber floors: (a) existing simple layer
of wood planks on the timber beams; (b) second layer of wood planks arranged crosswise on the
existing one and fixed with steel screws; (c) diagonal bracing of the existing wood planks by light
gauge steel plates; (d) diagonal bracing of the existing wood planks by FRP laminae; (e) three
layers of plywood panels glued on the existing wood planks; (f) a reinforced concrete slab connected
by means of studs (all measures in mm) [26].
Experimental apparatuses have been set up for in-plane testing of floor slabs in a reduced scale as well
as in full scale (Fig. 8), also assessing the small scale effect, by different research groups, e.g. [26, 27,
97
28, 29, 30, 31, 32, 33, 34]; numerical models have been calibrated for slabs with different
reinforcement based on experimental results. The final purpose of these studies is to evaluate the
effects of the in-plane stiffness reached by different strengthening criteria (Fig. 9) on the capacity
curve of a traditional masonry building. This may be evaluated numerically, by a nonlinear static
(pushover) analysis. Case studies are being carried out to this purpose [35, 36].
Fig. 8 Scheme of an experimental apparatus for in-plane testing of slabs, and an example; s.g.
indicates the strain gage positions [26]
Fig. 9 Results from experimentation of in-plane floor behaviour: resultant force versus mid-span
displacement. (a) existing simple layer of wood planks, (b) second layer of wood planks, (c) diagonal
bracing with light gauge steel plates, (d) diagonal bracing with FRP laminae, (e) three layers of
plywood, (f) reinforced concrete slab [26]
On the basis of these studies, some general indications for strengthening, at the Level of the Structure,
may be given, as follows, even if it must be pointed out that each case shall be often regarded as a
specific case study:
- The structural behaviour of the building as a whole must be enhanced, improving the in-plane
behaviour of the slabs (floors and roofs) and their connections with masonry walls. This
98
should be obtained without excessive stiffness concentration, and without any increase of the
mass, thus the solutions b), c) d), e) of Fig. 7 seem to be more advisable;
- The supports of timber members belonging to floors and roofs must be accurately checked
and strengthened if necessary, assuring bi-lateral connections between the elements and the
walls, disregarding any contribution of friction between wood and masonry;
- The presence of unrestrained thrusts from rafters and other timber elements impinging on
walls must be avoided and, if this is not possible, timber elements must be added in order to
reconstitute the regularity of structural layout;
- Each joint must be able to maintain elements in contact, avoiding possible disassembling of
the joint under exceptional actions, as discussed hereafter.
3.2 The level of the elements
In the general strengthening of a building it is often necessary to improve the state of timber elements
that may present section reductions due to decay or even to insufficient original cross-sections in
relation to the higher stress level in seismic conditions.
Damage often occurs at the extremes of the beams, where the shear force is at its maximum and where
the beam, snugly constrained in the wall, suffers limited air circulation that favours wood moisture
contents above 20% and, as a consequence, the development of biotic attack. It is usually necessary to
eliminate the damaged part and replace it with new material, ensuring collaboration between the new
and the original part, in order to restore full continuity of mechanical properties. Different methods
may be adopted. The additional material may be either embedded in the wood and connected to it in
order to reconstitute the original section, or may be added as a new element beside the first one, e.g.
[37], (Fig. 10). The second approach is particularly useful to increase the original section size and
consequently the original mechanical properties, therefore its major use is in the strengthening of floor
systems. It must be considered that often these interventions are required for heritage buildings for
which conservation rules are enforced. In this case the addition of material that remains visible may
not be acceptable.
If following the first option, the two parts may be joined by inserting glued steel bars across the
discontinuity, as done in some connections for new wood-based materials like glued laminated timber,
for which abundant literature is available. The approach offers several advantages when adopted in
existing structures to improve mechanical properties or restore continuity: a high connection stiffness
without significant settling; the possibility of ductile design with yielding of the steel, in spite of the
adherence based on glue; the protection of glue and of the embedded steel elements from chemicals
and fire (a positive feature that is not shared by other techniques using external materials, for example
those with polymeric matrix); on the non-structural side, the unmodified exterior of the reinforced
element, that maintains the original aspect.
The method is still affected by uncertainties and difficulties and is currently receiving attention by
several researchers. Among the possible drawbacks is the elastic-brittle contribution that the
connection of timber-glue-steel may give to the global behaviour of the section. Besides, suitable
glues must be specified for the job according to the expected service conditions, particularly under
steady high service temperatures [38,39]. Timber elements are not considered dissipative, yet an
acceptable level of ductility must be ensured. An additional question to be clarified concerns the
stresses induced by the dimensional changes of wood due to moisture variations.
99
Fig. 10 Fabrication of specimens with glued-in steel rods [37].

The load bearing capacity of the connection is limited by failure modes concerning pull-out: namely,
a) pull-out of the steel element due to cohesive failure in the glue or in the wood, adhesive
failure in the glue-steel or glue-wood interfaces, or cohesive failure at the adhesive-wood
fibres interface;
b) splitting failure of wood caused by excessive tensile stresses perpendicular to the grain;
c) shear failure of wood (plug-type failure).
If the metal surface is rough, as in the case of threaded or ribbed bars, the adhesive failure at the glue-
steel interface will not occur; yet a good glue-steel adhesion is far more difficult to achieve when the
steel surface is large, like in the case of plates. The other modes are not always physically
distinguishable, since the pull-out of steel and the transversal splitting of the timber element most
often occur simultaneously.
The same connections may be constructed substituting steel bars with fibre reinforced polymeric
materials, FRP, for which similar remarks apply. FRP plates are proposed also for the upgrading of
properties in timber beams: plates may be glued externally on the tension side or inserted and glued-in
along the length, a solution that protects the material and is more compatible with requirements
related to visual aspects e.g. [40]. An important research effort is, indeed, devoted to FRP’s and their
use in timber structures.
As a general remark, in spite of the effectiveness offered by the connection with glued-in elements the
presence of glue in the assembly may introduce some amount of brittleness. Considering the general
philosophy in the seismic design or seismic strengthening of timber structures that bans any brittle
contribution from the areas where different timber members are joined, as discussed in a further
section, these glued-in inter-element connections should be kept far from the joint areas, positioning
them where the timber elements remain elastic.
Also at the level of the elements, some general indications may be given, as follows;
- When it is necessary to eliminate the damaged part of an element, it is better to replace it with
timber prosthesis, possibly of the same species, connecting the two parts by means of glued-in
steel rods; as a general remark, this kind of connection may introduce brittleness, so glued-in
inter-element connections should be kept far from the joint areas;
100
- Glued-in steel or CFRP rods may be used for improving the mechanical performance of
timber element; the best results, also in terms of durability of the intervention, can be obtained
with elements inserted into the element, rather than with elements glued externally.
3.3 The level of joints
A peculiarity of timber structures, both new and old, is the essential role of connections in
determining a successful behaviour of the whole structural system. For structures built according to
current design practice, joints are usually fabricated with metal connectors that undertake the
transmission of loads between timber elements; for ultimate conditions they may be dimensioned to
develop a ductile post-elastic behaviour of the joint exploiting proper interaction with the wood.
Considering the very limited post-elastic resources of timber, joints become the only possible location
of post-elastic yielding and of energy dissipation.
In traditional timber structures, carpentry joints are suitably shaped in order to transmit forces by
direct contact and friction. Their many forms encompass a wealth of heuristic knowledge passed on
by generations of carpenters. Traditionally, metal devices, when present, were mainly meant to
prevent accidental and unexpected misalignment of parts, avoiding disassembling of the joint under
exceptional loads without acting in normal working conditions, according to a robust conception of
the joint. Here again, joints were the critical elements for the safety of the whole structural system. It
is worth noting that old metal connectors already present in a joint may not comply with current safety
requirements and must always be checked.
Seismic design codes offer little or no indications for strengthening or improving carpentry joints.
Interventions are mostly performed according to traditional criteria, which are usually more related to
static conditions.
In order to express criteria for seismic strengthening, the two working conditions of a joint, that is, the
normal service conditions and the limit state in response to exceptional actions, may be assumed as
reference. Since codes give no guidance for existing joints, it is convenient, once more, to refer to the
principles expressed for the design of new timber structures, for which the structural capability of
post-elastic cyclic deformations and of energy dissipation concentrates in the connections, while
timber elements must remain elastic. Tab. 1 defines requirements for joints in the two reference states.
Tab. 1 Performance requirements for carpentry joints
State Goals of the connection
Service conditions 1. transmit forces by contact
Ultimate conditions 2. avoid joint disassembly
3. avoid brittle failure modes
4. enhance ductility/ energy dissipation
Experimental research programs have been carried out over several years to assess the elastic and
post-elastic behaviour of the main types of carpentry joints traditionally used in roof trusses. Extended
programs were conducted in Italy [41, 42, 43, 44], in Portugal, with similar results [45, 46, 47], and in
Belgium [48, 49]. One important objective has been to devise reinforcement criteria that would be
101
effective especially in extreme conditions, but would not change the original way of functioning of
the connection for normal loading situations. The main idea has been to account for the conservation
of cultural values and at the same time provide safety and develop in a modern sense the characteristic
of robustness implicit in tradition. Only metal devices, which are the most traditional kind of
reinforcement, have been considered, disregarding more innovative but less common fibre reinforced
materials. The research work has treated particularly the step or cogging joints used in the rafter-to-
chord node and in other connection areas of trusses, characterizing the mechanical behaviour in
monotonic and cyclic conditions and comparing different types of reinforcement. Fig. 11 shows the
testing apparatus with a full-scale specimen of a rafter-to-chord node reinforced with a binding strip.
A constant pressure is applied to the rafter, while a variable transversal force acts at its upper end,
rotating it. The objectives of the program were: 1) to quantify the joints rotational stiffness in the
elastic range, that is, in service conditions, as a function of the main joint parameters, like geometry
and compression level, in order to develop accurate semi-rigid models; 2) for the post-elastic field, to
investigate the effect of the different types of reinforcement and repair/strengthening with particular
attention to the capability to withstand an adequate number of load cycles, and to point out situations
that could trigger brittle failure modes.
Other joints considered in research programs have been the double step joint, used for large rafters
and chords with a double notch in the chord, the scarf joint, which is usually present in truss chords
composed of two segments and having a characteristic zigzag shape, the half-depth lap joint, typically
connecting cross-bracing elements (Fig. 12), the dovetail joint [43, 50, 51, 52, 53, 54, 55], and the
tenon-and-mortise joint [56, 57].
Some interesting works in the literature deal with connections used in the far eastern carpentry
tradition, although the work is usually not focused on seismic response or strengthening issues [58,
59, 60, 61].
On the basis of these studies, some indications for strengthening interventions may be given, as
follows:
- Traditional carpentry joints transmit forces between connected elements by contact and
friction; yet, contact pressure may be reduced in seismic or other exceptional conditions;
contact may be decreased or even eliminated also by unexpected out-of-plane forces; metal
connectors must be applied to prevent disassembly, yet they should not alter the working
modality in normal conditions; in particular, the connection stiffness should not be
significantly increased; the strengthening interventions should not induce parasitic effects like
over-stiffening; in particular the relative rotational capacity between connected members
should not be drastically reduced or eliminated;
- In limit conditions, post-elastic capacity should be enhanced and brittle failure should be
avoided; to this purpose, attention must be exerted in order to avoid introducing brittle modes
of failure in the timber or metal elements;
- Because brittleness may be introduced by glue, gluing is not admitted in the joint area;
connection of prostheses to the repaired timber member by means of glued-in steel elements
should be performed away from the joint area;
- Similarly, the use of metal connectors with welded parts should be avoided as a possible
cause of brittleness;
102
Fig 11 Experimental setup with a specimen of rafter-to-chord node [43]
Fig. 12 Half-depth lap joint [45] Fig. 13 Encapsulated node [62]
Fig. 14 Connection reinforced with two transversal Fig. 15 Distributed reinforcement to prevent brittle
bolts of small diameter (left) gave the most failure, using fully threaded self-tapping screws or
satisfactory cyclic behaviour glued-in steel rods.
103
- Particularly in the rafter-to-chord node of a truss, the use of metal cages or cuffs containing
the node may cause brittle behaviour in the timber elements and induce uncontrolled decay by
forming an adverse microclimate in the node area (Fig. 13);
- Brittle behaviour may also result more subtly from placing a relatively small amount of
reinforcement with an inappropriate layout, preventing relative rotation of the elements; in the
cases presented in Fig. 14, the best results can be obtained reinforcing the carpentry
connection with two transversal bolts of small diameter (Fig. 14, left);
- Brittle failure modes may occur at critical zones like the chord toe of a rafter-to-chord node
that is also particularly prone to biotic attack. The toe may fail in sliding shear due to the
horizontal force from the rafter, especially for low connection angles, short toe length (and
timber with biotic attack). The reduced section of the rafter tip may also fail by compression;
evenly distributing small diameter connectors, like screws, in these areas should eliminate or
control the problem (see Fig. 15); as a dimensioning indication, at the toe, the total amount of
steel should be able to carry the shear force without relying on the timber.
4. Effectiveness of Interventions
In order to assess the effectiveness of strengthening interventions in the seismic field, both
experimentation and numerical analysis may be used. Given the difficulties and often the
impossibility to perform dynamic tests, the limitations that are encountered in developing accurate
models of the structure or element, and the uncertainties intrinsic in the seismic action, none of the
two approaches, supplying a partial view of the situation, may fully confirm the validity of a solution,
but each may point out particular advantages or criticalities.
Numerical analysis is a powerful tool for assessing the level of seismic response of a structure, as long
as modelling has been accurate, which may be at times problematic especially with old structures.
Nevertheless, much information may be obtained from both static and dynamic analysis of the global
configuration, and from fine modelling of details.
Research projects with extended experimental campaigns, like those quoted above on joints, have
permitted the comparison of different strengthening interventions in the static and in the cyclic field,
extended to the nonlinear behaviour, that is, to ultimate conditions.
Experimentation on full-scale trusses, with spans ranging from 12 to 24 m, after their carpentry joints
had been reinforced according to the methods tested, have given information on the global behaviour
in symmetric and anti-symmetric static loading conditions. Numerical analyses of the same trusses
and of other structures have been performed with an accurate modelling of the behaviour of the
strengthened joints calibrated with experimental results, in order to investigate the dynamic and
seismic response of the upgraded structure. Behaviour factors around 3 have resulted from seismic
time-history analyses, confirming the possibility of good post-elastic response after strengthening
interventions.
The behaviour of timber elements before and after a strengthening intervention has been determined
experimentally; in particular, results are available for beams supporting floor slabs e.g. [63, 64, 65].
104
The effects of increased in-plane stiffness of slabs

as a function of the type of reinforcement have
been and are being tested by different groups in
Europe and worldwide. Tests, accompanied by
numerical analyses of the seismic response of the
wall-floor systems, are giving interesting
information on the effects of different intervention
typologies, as mentioned before [e.g. 35, 36].
Lastly, some cases of positive behaviour have been
observed during a damage evaluation activity after
the L’Aquila earthquake of 2009 in roof systems
where minimally invasive and low-mass
interventions had been carried out, as in the
example of Fig. 16, which gives confidence in
using this approach [e.g. 66].
Fig. 16 Roof system reinforced with steel ties [66].
5. Conclusions
In European seismic areas timber structures are mainly present to support roofs and floor slabs. These
elements were originally conceived for vertical loads, therefore interventions to upgrade their seismic
response are often required.
In planning interventions for the seismic strengthening or improvement of timber roof structures, the
structure should be first examined to detect possible construction errors, like rafters not restrained by
chords and applying thrusts on the supporting walls, or missing elements in general; it must be
ensured that the static scheme can respond to horizontal actions, and that stiffness, and therefore
deformability, is comparable in the main directions.
Excessive increase in the mechanical parameters, and especially of stiffness, should be avoided. This
precaution should be applied to the structure, to the repair and strengthening of individual members,
and particularly, at a local level, to carpentry joints, that are a leading element in the system response.
With reference to exceptional earthquake conditions, the goals of interventions on carpentry joints are
to avoid brittle failure modes, which may result from the original layout or may be consequent to an
increase of stiffness, to foster the deformational capability in the post-elastic field, and to prevent
disassembly, which is attained by inserting suitable metal connectors. The joints, however, should
maintain their original working modality in common service conditions, that is, transmit forces by
direct contact, without contribution from the connectors.
Similar indications hold for slab structures. Interventions increasing the in-plane stiffness are
particularly interesting in the seismic field because the resulting structure will interact with masonry
walls affecting the horizontal force distribution and, therefore, the system response. Here again,
beyond the most common case of collaborating concrete slab, different solutions are being studied to
offer a variety of mechanical properties among which to choose the most appropriate for the context.
105
6. References
[1] D.M. 14.01.2008, Nuove norme tecniche per le costruzioni, G.U 4.02.2008, n.29, Ministero
delle Infrastrutture, dell’Interno e Dipartimento Protezione Civile, Roma, 2008.
[2] Tsakanika, E., “The constructional analysis of timber load bearing systems as a tool for
interpreting Aegean Bronze Age Architecture”, In: Proceedings of Bronze Age Architectural
Traditions in the Eastern Mediterranean: Diffusion and Diversity, Munich, Germany, 2008.
[3] Monteiro, M, Lopes, R., Bento, R., “Dynamic behaviour of a Pombalino quarter”, In:
Proceedings of Conference for the 250th Anniversary of the 1755 Lisbon Earthquake, Lisbon,
2005.
[4] Ruggieri, N., Tampone, G., Zinno, R., “Typical Failures, Seismic Behaviour and Safety of the
“bourbon system” with timber framing”, Advanced Materials Research, Vol. 778, 2013, pp. 58-
65.
[5] Vintzileou, E., Zagkotsis, A., Repapis, C., and Zeris, C, “Seismic behaviour of the historical
system of the island of Lefkada, Greece”, Construction and Building Materials, Vol. 21, No.1,
2007, pp. 225–236.
[6] Vintzileou, E., “Effect of Timber Ties on the Behavior of Historic Masonry”, Journal of.
Structural Engineering, Vol. 134, No. 6, 2008, pp. 961–972.
[7] Gülkan, P., Langenbach, R., “The earthquake resistance of Traditional timber and Masonry
Dwellings in Turkey”, In: Proceedings of 13th World Conference on Earthquake Engineering,
Vancouver, Canada, 2004.
[8] Poletti, E., Vasconcelos, G., “Seismic behaviour and retrofitting of timber frame walls”,
Advanced Materials Research, Vol. 778, 2013, pp. 706-713.
[9] Arun, G., “Traditional timber construction in Turkey”, in: Proceedings of Intl. Symposium
“Timber structures from antiquity to the present”, T.C. Haliç University Press, Istanbul, 2009.
[10] Demir, A., “Wooden-columned mosques in Anatolia”, in: Proceedings of Intl. Symposium
“Timber structures from antiquity to the present”, T.C. Haliç University Press, Istanbul, 2009.
[11] Somer, M.E., “Prinkipo Palace”, in: Proceedings of Intl. Symposium “Timber structures from
antiquity to the present”, T.C. Haliç University Press, Istanbul, 2009.
[12] DPCM 26.02.2011, Linee Guida per la valutazione e riduzione del rischio sismico del
patrimonio culturale con riferimento alle norme tecniche per le costruzioni, (Guidelines for
risk evaluation and reduction of the cultural heritage, in Italian, by the Ministry of Cultural
Heritage, Mibact), G.U. 26.02. 2011 n.47, Roma, 2011.
[13] EN 1998-1, Eurocode 8: Design of structures for earthquake resistance. Part 1: General Rules,
Seismic Actions and Rules for Buildings, CEN, European Committee for Standardization, 2004,
Brussels, Belgium.
[14] EN 1995-1, Eurocode 5: Design of timber structures. Part 1-1: General rules and rules for the
buildings, CEN, European Committee for Standardization, 2005, Brussels, Belgium.
[15] UNI 11138, Cultural heritage - Wooden artefacts - Building load bearing structures - Criteria
for the preliminary evaluation, the design and the execution of works, UNI, (English version:
2004), Milano.
[16] Cruz, H., Yeomans, D., Tsakanika, E., Macchioni, N., Jorissen, A., Touza, M., Mannucci, M.,
Lourenço P. B., “Guidelines for the On-Site Assessment of Historic Timber Structures”, Intl.
Journal of Architectural Heritage, 2013, DOI: 10.1080/ 15583058.2013.774070.
[17] Parisi, M.A., Chesi, C., Tardini, C., “Inferring seismic behaviour from morphology in timber
roofs”, International Journal of Architectural Heritage, Vol. 6, 2012, pp. 100-116.
[18] Parisi, M.A., Piazza, M., “Restoration and Strengthening of Timber Structures: Principles,
Criteria, and Examples”, Practice Periodical on Structural Design and Construction, Vol. 12,
No. 4, 2007, pp. 177-185.
[19] Giuriani E., Marini A., “Wooden roof box structure for the anti-seismic strengthening of
historic buildings”, Journal of Architectural Heritage, Vol. 2, No. 3, 2008, pp. 226-246.
[20] Giongo, I., Piazza, M., Tomasi, R., “Out of Plane Refurbishment Techniques of Existing Timber
Floors by means of Timber to Timber Composites Structures”, In: Proceedings of 12th World
106
Conference on Timber Engineering, Session 45, Auckland, New Zealand, 2012, pp. 544-550.
[21] Giongo, I., Piazza, M., Tomasi, R., “Cambering of timber composite beams by means of screw
fasteners”, Wiadomosci Konserwatorskie – Journal of Heritage Conservation, Journal of the
Association of Monument Conservators, Vol. 32, 2012, pp. 133-136.
[22] Piazza, M., “Restoration of timber floors via a composite timber-timber solution”, In:
Proceedings of the Technical Workshop RILEM Timber: a Structural Material from the Past to
the Future, Trento, 1994.
[23] Tomasi, R., Crosatti, A., Piazza, M., “Theoretical and Experimental Analysis of Timber-to-
Timber Joints connected with inclined Screws”, Construction and Building Materials, Vol. 24,
2010, pp. 1560–1571.
[24] Piazza, M., Polastri, A., Tomasi, R., “Ductility of timber joints under static and cyclic loads”,
In: Proceedings of the Institution of Civil Engineers. Structures and Buildings, Vol. 164, No.
SB2, 2011, pp. 79-90.
[25] Giongo I., Piazza M., Tomasi R., “Investigation on the self-tapping screws capability to induce
internal stress in timber elements”, Advanced Materials Research, Vol. 778, 2013, pp 604-611.
[26] Piazza, M., Baldessari, C., Tomasi, R., “The role of in-plane floor stiffness in the seismic
behaviour of traditional buildings”, In: Proceedings of the 14th World Conference on
Earthquake Engineering, Beijing, 2008.
[27] Filiatrault, A., Fischer, D., Folz, D. Uang, C-M., “Experimental parametric study on the in-
plane stiffness of wood diaphragms”, Canadian Journal of Civil Engineering, Vol. 29, 2002,
pp. 554-566.
[28] Paquette, J., Bruneau, M., Brzev, S., “Seismic Testing of Repaired Unreinforced Masonry
Building having Flexible Diaphragm”, Journal of Structural Engineering, Vol. 130, No. 10,
2004, pp. 1487-1496.
[29] Giuriani E., Marini A., Plizzari G., “Experimental behaviour of stud connected wooden floors
undergoing seismic action”, Restoration of Building and Monuments. Vol.11, No.1, 2005, pp. 3-
24.
[30] Corradi, M. Speranzini, E., Borri, A. & Vignoli, A., “In-plane shear reinforcement of wood
beam floors with FRP”, Composites Part B, Vol. 37, 2006, pp. 310–319.
[31] Baldessari, C., Piazza, M., Tomasi, R., “The refurbishment of existing timber floors:
characterization of the in-plane behaviour”, In: Proceedings of PROHITECH 9 Protection of
Historical Buildings, Taylor & Francis, London, 2009.
[32] Brignola, A., Pampanin, S., Podestà, S., “Experimental Evaluation of the In-Plane Stiffness of
Timber Diaphragms”, Earthquake Spectra, Vol. 28, No. 4, 2012, pp. 1687-1709.
[33] Giongo, I,. Dizhur, D., Tomasi, R., Ingham, J., “In-plane assessment of existing timber
diaphragms in URM buildings via quasi-static and dynamic in-situ tests”, Advanced Materials
Research, Vol. 778, 2013, pp 495-502.
[34] Wilson, A., Quenneville, P. J. H., Ingham, J. M., “In-plane orthotropic behavior of timber floor
diaphragms in unreinforced masonry buildings”, Journal of Structural Engineering, Vol. 140,
No. 1, 2014, pp. 04013038-1-11.
[35] Giongo, I., Piazza M., Tomasi R., “Pushover analysis of traditional masonry buildings:
influence of refurbished timber-floors stiffness”, Paper no. 51, In: Proceedings of SHATIS'11
Intl Conference on Structural Health Assessment of Timber Structures, Lisbon, Portugal, 2011.
[36] Giongo, I., Piazza, M., Rizzardi, A., Rodegher, C., Tomasi, R., “Seismic evaluation of URM
buildings with flexible diaphragms. Proposal of a simplified 'ENT' method”, Paper no. 2629, In:
Proceedings of 15th World Conference on Earthquake Engineering, Lisbon, Portugal, 2012.
[37] Piazza, M., Parisi, M.A., “Rehabilitation of timber structures by new materials and connectors”,
in: Proceedings of 10th Conf. “Structural faults & repair 2003 - Extending the Life of Bridges,
Concrete & Composites, Buildings, Masonry & Civil Structures”, Commonwealth Institute,
London, 2003.
[38] Custodio, J., Broughton, J. & Cruz, H., “Rehabilitation of Timber Structures – preparation and
environmental service condition effects on the bulk performance of epoxy adhesives”,
Construction and Building Materials, Vol. 25 (8), 2011, pp.3570-3582.
107
[39] Custodio, J., Broughton, J. & Cruz, H., “Rehabilitation of Timber Structures –Novel test
method to assess the performance and durability of bonded-in rod connections”, Materials and
Structures, Vol. 45, (1-2), 2012, pp. 199-221.
[40] Borri A, Corradi M, Grazini A., “A method for flexural reinforcement of old wood beams with
CFRP materials”, Composites Part B, Vol. 36, No. 2, 2005, pp.143–53.
[41] Parisi, M.A., Piazza, M., “Mechanics of plain and retrofitted traditional timber connections”,
Journal of Structural Engineering, Vol. 126, No. 12, 2000, pp 1395-1403.
[42] Parisi, M.A., Piazza, M., “Seismic Behavior and Retrofitting of Joints in Traditional Timber
Roof Structures”, Soil Dynamics and Earthquake Engineering, Vol. 22, No. 9-12, 2002, pp
1183-1191.
[43] Parisi, M.A., Piazza, M., “Seismic strengthening of traditional carpentry joints”, In:
Proceedings of the 14th World Conference on Earthquake Engineering, Beijing, 2008.
[44] Parisi, M.A., Piazza, M., “Carpentry Joints in Earthquake Conditions”, Ingegneria Sismica-
International Journal of Earthquake Engineering, Vol. 4, 2013, pp. 55-72.
[45] Palma, P., Cruz, H., “Mechanical behaviour of traditional timber carpentry joints in service
conditions - results of monotonic tests”, In: Proceedings of the XVI International ICOMOS
IWC Symposium “From material to Structure - Mechanical behaviour and failures of the timber
structures”, Venice, Italy, 2007.
[46] Branco, J.M., Piazza, M., Cruz. P.J.S., “Experimental evaluation of different strengthening
techniques of traditional timber connections”, Engineering Structures, Vol. 33, 2011, pp. 2259–
2270.
[47] Palma, P., Garcia, H., Ferreira, J., Appleton, J., Cruz, H., “Behaviour and repair of carpentry
connections – Rotational behaviour of the rafter and tie beam connection in timber roof
structures”, Journal of Cultural Heritage, Vol. 13, No. 3, Supplement: S, 2012, pp. S64-S73.
[48] Descamps T., Lambion J., Laplume D. (2006), “Timber Structures: Rotational stiffness of
carpentry joints”, In: Proceedings of World Conference on Timber Engineering, Portland, USA,
2006
[49] Descamps T., Noël J., Semi-rigid analysis of old timber frames: definition of equivalent springs
for joints modeling. Enhancement of the method, numerical and experimental validation.
International Review of Mechanical Engineering. Vol. 3, No. 2, 2009, pp 230-239.
[50] Bulleit, M.W., Bogue Sandberg, L., Drewek, M.W., o’Bryant, T. L., “Behavior and Modeling of
Wood-pegged Timber Frames”, Journal of Structural Engineering, Vol. 125, No. 1, 1999, pp. 3-
9.
[51] Seo, J-M, Choi, I.-K., and Lee, J.-R., “Static and Cyclic Behavior of Wooden Frames with
Tenon Joints under Lateral Load”, Journal of Structural Engineering, Vol. 125, No. 3, 1999, pp.
344-349.
[52] Villar J.R., Guaita M., Vidal P., Arriaga F., “Analysis of the Stress State at the Cogging Joint in
Timber Structures”, Biosystems Engineering, Vol. 96, No. 1, 2007, pp. 79-90.
[53] Sangree R.H., Schafer B.W., “Experimental and numerical analysis of a halved and tabled
traditional timber scarf joint”, Construction and Building Materials, Vol. 23, 2009, pp. 615-624.
[54] Sangree R.H., Schafer B.W., “Experimental and numerical analysis of a stop-splayed traditional
timber scarf joint with key”, Construction and Building Materials, Vol. 23, 2009, pp. 376-385.
[55] Parisi M.A., Cordié C., “Mechanical behavior of double-step timber joints”, Construction and
Building Materials, Vol. 24, 2010, pp. 1364-1371.
[56] Feio A.O., Lourenço P.B., Machado J.S., “Testing and modeling of a traditional timber mortise
and tenon joint”, Materials and Structures, Vol. 47, 2014, pp. 213–225.
[57] Branco J, Descamps T., “Analysis and strengthening of carpentry joints”, In State of the art
report on reinforcement of timber structures, COST Action FP1101, 2014.
[58] King, W. S., Yen, R. J. Y., Yen, A. Y. N, “Joint characteristics of traditional Chinese wooden
frames”, Engineering Structures, Vol. 18, No. 8, 1996, pp. 635-644.
[59] Guam, Z., Kitamori, A., Komatsu, K., “Experimental study and finite element modelling of
Japanese Nuki joints – Part 1: initial stress state subject to different wedge configurations”,
108
Engineering Structures, Vol. 30, 2008, pp. 2032-2040.

[60] Guam, Z., Kitamori, A., Komatsu, K., “Experimental study and finite element modelling of
Japanese Nuki joints – Part 2: racking resistance subject to different wedge configurations”,
Engineering Structures, Vol. 30, 2008, pp. 2041-2049.
[61] Lam, F., He, M., Yao, C., “Example of Traditional Tall Timber buildings in China – the
Yingxian Pagoda”, Structural Engineering International, Vol. 2, 2008, pp. 126-129.
[62] Parisi, M.A., Piazza, M., “Seismic evaluation and strengthening of timber structures in
traditional buildings”, In: Proceedings of Intl. Conference “250th Anniversary of the Lisbon
earthquake”, Lisbon, 2005.
[63] Piazza, M., Riggio M.P., Tomasi, R., Giongo, I., “Comparison of in-situ and laboratory testing
for the characterization of old timber beams before and after intervention”, Advanced Materials
Research, Vols. 133-134, 2010, pp. 1101-1106.
[64] Riggio, M., Sandak, J., Sandak, A., Piazza, M., “Coupling local semi-destructive techniques
and non-destructive imaging for the characterization of traditional timber structures: a case
study”, in: Proceedings of 5th International Congress "Science and Technology for the
Safeguard of Cultural Heritage in the Mediterranean Basin" , Vol. II, Istanbul, Turkey, 2011.
[65] Riggio, M. P., Tomasi, R., Piazza, M., “Refurbishment of a Traditional Timber Floor with a
Reversible Technique: Importance of the Investigation Campaign for Design and Control of the
Intervention”, International Journal of Architectural Heritage, Vol. 8, No. 1, 2014, DOI:
10.1080/15583058.2012.670364, pp. 74-93.
[66] Parisi, M. A., Piazza, M., Chesi, C., “Seismic response of traditional timber elements and roof
structures: learning from the L’Aquila earthquake”, Paper no. 64, In: Proceedings of SHATIS'11
Intl Conference on Structural Health Assessment of Timber Structures, Lisbon, Portugal, 2011.
109
110
PART II:
REINFORCEMENT METHODS
111
112
Adhesives for on-site bonding: characteristics, testing and prospects
7
Adhesives for on-site bonding: characteristics, testing and
prospects
Benedetto Pizzo1, Dave Smedley2
Summary
The present chapter deals with the main characteristics of those adhesive products normally used for
the reinforcement of timber structures, generally achieved through a site-applied bonding procedure.
The characteristics relate to a combination of the chemical composition and the formulation
characteristics of the products. A section dealing with testing procedures specifically intended for
products to be used on site has also been added, as this aspect does not exist in current European
standards. The principal issues associated with the use of those products, mainly relating to long-term
duration aspects, are considered, providing more of the adhesive characteristics and features of the
adhesive rather than the wood substrate. Finally, some challenges are introduced in order to improve
research in this field.
1. Introduction
The adhesive products which may be used on site have specific properties which are not usually
encountered in aminoplastic adhesives for the reasons indicated below:
a) The quality of wood surfaces is not easy to control on site in contrast to that which may be
encountered in a manufacturing plant, in particular where there is effective control of the surfaces to
be bonded. Often the surfaces are freshly machined by sanding or planing, in a clean environment
with the correct temperature and humidity which enables the moisture content of the wood surfaces to
be regulated and facilitates a good bonding surface. This is in contrast to the conditions on site where
all of the above factors may be variable and/or not easily controllable;
b) Bonding pressure of the timber components is not easily achievable on site, in contrast to a factory
situation where bonding is almost always carried out under controlled pressure, and the adhesive
manufacturers often specify the minimum and maximum application temperatures and pressures to be
applied to the assemblies;
c) It follows from the above that factory produced bonded assemblies usually have bond line
thicknesses of 0.5 mm or lower, in contrast to the bond line thicknesses found in on-site bonding
which may be between 1 and 12 mm.
These three conditions indicate that there is a requirement to use classes of adhesives on site which
are characterised by good substrate wetting properties, high internal cohesive strength, plus final
adhesion properties that are not only related to mechanical interlocking (whose efficacy is strongly
1)
Researcher, CNR-IVALSA, Trees and Timber Institute, I-50019, Sesto Fiorentino (FI), Italy
2)
Technical Director, Rotafix Ltd., Abercrave, Swansea, SA9 1UR, United Kingdom
113
associated with the application of pressure during bonding) but also to a variety of other adhesion
mechanisms, such as specific adhesion. As reported previously by Custodio et al. [1], practical
applications on-site generally utilise epoxy, polyurethane adhesives and additionally some polyester
products for structural bonding applications on timber. Of these varieties, epoxy adhesives have been
used for more than forty years and they are currently still the most widely used choice for bonding
structural timber on site. Epoxy types intended to be used for site bonding of timber components
implies that in practice they have to be able to cure at moderately low temperatures. These epoxy
systems are two-part thermosetting adhesives, and include a wide range of formulations and varying
commercial products with distinct characteristics. These are required for practical reasons to include
good gap-filling shrink resistant properties, excellent tensile/shear strength, high dry and wet strength,
and good resistance against moisture and certain chemicals. Although being the optimum choice,
epoxies have some limitations: they have poor peel strength and may often delaminate when subjected
to in-service repeated wetting and drying.
The fact that epoxy adhesives can be effectively used as structural adhesives for applications on
timber has been evidenced in experimental tests: for the development of an epoxy adhesive specific
for glulam consolidation, Radovic and Goth [2] applied all the EN 302 methods [3]–[9] to thin
bondline joints finding full compliance with EN 301 requirements [10]. Although no specific
standards for epoxy adhesives exist to date, it is widely accepted that they are suitable for limited
exterior service environments corresponding to the Service Classes 1 and 2 as defined in EN 1995-1-1
[11].
Another class of adhesive products commonly used for interventions on site are two-component
polyurethanes. In contrast, monocomponent polyurethanes, although largely used in manufacturing
plants to construct glulam components, are not, in general, preferred for applications on site owing to
their tendency to develop gas bubbles within the thick glue line, thus compromising the cohesive
strength of the bonded components. In specific European countries, phenol-resorcinol based (PRF)
adhesives were occasionally used to repair timber on site, for instance to bond glued-in rods or to glue
side plates for beam reinforcement: although these applications have not been widely used they are
mentioned in order to present a complete picture.
Apart from the difference in chemistry which will be briefly considered in Section 2, all classes of
products included do evidence similar characteristics and problems in their use on-site. For instance,
they can be used in a range of rheological states from thixotropic pastes to fairly low viscosity fluids.
Moreover, they are lacking with regard to specific standards for testing and usage when applied on
site. This is the reason why no distinction will be made amongst them with respect to testing in
Section 3. However, previous tests have evidenced certain differences in performance. Several
experimental trials carried out in the past showed that steel rods bonded with an epoxy adhesive
exhibited timber failures close to and along the adhesive/timber interface, with higher pull-out
strengths compared to both polyurethane and PRF adhesives, which instead exhibited cohesive
adhesive or adhesion failures [12]–[15]. Moreover, fatigue tests also indicated better performance for
epoxies than for the other two adhesives [15] and, with polyurethanes tested under load, significant
short term strength losses were measured for glued-in rods at a temperature of just above 40°C.
However, in the case of a particular epoxy product this strength decrease was observed at
temperatures above 50°C [16]. Additionally, due to low adhesion to steel, specimens prepared with
PRF suffered failure in less time than anticipated after long-duration application of load [15], whereas
performances in gluing GFRP seemed more encouraging when evaluated after moisture-induced
114
cyclic delamination tests, although considerable variability was found depending on the FRP-PFR
adhesive combination [17]. However, also in the case of FRP bonding, structural epoxy adhesives are
the generally accepted products in bonded FRP–wood connections. On the other hand, a different
behaviour concerning the adhesion to humid substrates can also be observed: whereas in the case of
epoxies the moisture content of wood at time of bonding must be lower than 20-22%, irrespective of
timber species [18], in the case of polyurethanes even when gluing at higher moisture levels the
obtained resistance is in accordance with the wood resistance at that corresponding moisture content
[19]. Appreciable delay in curing can be expected in PRF resins, due to the slow-down of the water
release process (water molecules come from both the adhesive mixture and the polycondensation
reaction, see below). It must also be noted that there is limited experience on the long-term duration of
the assemblies prepared on-site with PRF resins. In fact, these adhesives do require consistent and
relatively high and even pressure across the joining area, and there are only very rare site situations
where it would remotely possible to use these factory products with any commercial viable clamping
method. This is one reason why they are less useful as site-based materials compared to their excellent
use in humidity and temperature factory controlled conditions.
It is relevant to emphasize at this juncture that all the above-mentioned results strongly depended on
the specific composition of the products, as will be made clear in Section 2. It is highly possible that
the results obtained are formulation specific to a particular adhesive type.
2. Adhesive material composition

In this section epoxy and polyurethane resin chemistry is introduced, including remarks concerning
the generic effects of several parameters affecting the curing behaviour of these products.
2.1 Epoxy resins
Epoxy resins used for the reinforcement/repair of structural timber elements are two-component
systems in which the prepolymer (average molecular weight usually lower than 700 g/mol), usually
called Part A, is a component containing epoxy functional groups, whereas the hardener or curing
agent, usually known as Part B, includes H-containing moieties able to react with the epoxy ring, thus
forming intermolecular bonds constituting a cross-linked three-dimensional structure. Although
several multi-functional molecules can be used, often the Part A is a prepolymer of diglycidyl ether of
bisphenol A, generally referred to as DGEBA, whose monomer can be schematically shown
in Fig. 1:
It is worth noting that the molecular weight

of this monomer is 340 g/mol, which
implies that Part A is usually constituted by
Fig. 1 Schematic of the monomer of DGEBA a prepolymer with a very low degree of
polymerisation, and hence is highly
reactive (in fact, it contains a large number
of reactive groups per unit mass).
In contrast, the hardeners (curing agents) commonly used with DGEBA are polyfunctional amines or
amides, in which more than one reactive group is present. The properties of the cross-linked polymer
strongly depend on the specific type of hardener. This affects the cross-linking density, and hence the
intermolecular spaces and the flexibility of the molecular branches. For the products used in timber
115
bonding, the hardeners must be sufficiently reactive to allow the composite reaction to fully cure at
room temperature, based on a minimum ambient curing temperature of approximately 5-10°C.
Usually a mixture of different aliphatic hardeners is used. The most used include triethylenetetramine
(TETA), (or others obtained from the reaction of ethylene oxide), and isophoronediamine (IPD),
whose schematic formulae are indicated in Fig. 2:
NH2  (CH2 )2  NH  (CH2 ) 2  NH  (CH2 )2  NH2

(a)
Fig. 2 Schematics of TETA (a) and IPD (b), representing the most
commonly utilised hardeners for epoxy resins used on timber (b)
structures
The reactions between a resin containing an epoxy group in its structure and an amine proceed by a
polyaddition mechanism (that is, without the development of a by-product);
1) reaction between the epoxy ring and one hydrogen atom of a primary amine, with the development
of a secondary amine (Fig. 3);
Fig. 3 Opening reaction of the epoxy

ring by action of primary amines (R
and ~ are different residuals)
2) reaction between a pre-existing or just formed secondary amine and an additional epoxy ring, with
formation of a tertiary amine (Fig. 4). Reactions (1) and (2) can develop both in series and in parallel;

ring by action of secondary amines (R’
and ~ are different residuals)
3) consequently, hydroxyl groups can react with epoxy rings, thus accelerating the reaction (Fig. 5).
However, in normal conditions, reactions (1) and (2) are favoured.

ring by action of hydroxyls (~ are
chemical residuals)
As is apparent from the reactions above, each hydrogen atom in the hardener structure is theoretically
able to react with a different epoxy ring. Therefore, the structures of both the prepolymer and the
hardener can reach a relatively elevated cross-linking degree even at moderately low temperatures, as
mentioned above. This undoubtedly has advantages in the application. However, the effect of
temperature on the curing rate is appreciable: Fig. 6 shows how the tensile-stress-at-rupture vs. time
curve moves upwards by increasing the curing temperature from 20°C to 50°C (that is, the
curing reaction is faster) for a commercial epoxy product.
116
Fig. 6 Tensile stress at rupture vs.

time at different curing
temperatures for a commercial
epoxy product: () 20°C; (•)
30°C; (*) 50°C. Tensile values
have been all measured at 20°C
In contrast, if curing occurs at temperatures lower than the Glass Transition Temperature (T g) of the
product, the effect on the final strength values is less pronounced [20]. Due to the very high viscosity
of the prepolymer and to the strong exothermic cure reaction, plus the consequent brittleness of the
cured product, pure undiluted liquid resins are rarely found in commercial products. Commercial
formulations contain several additives such as fillers, thickeners, reactive epoxy diluents and, rarely,
plasticizers which in the long term may affect the dimensional stability of the cured product. Reactive
diluents are commonly characterised by very low molecular weight epoxies (usually aliphatic). They
decrease viscosity and are mainly used in room temperature-curing products. However, they can also
affect the pot life and the curing rate (which both usually increase), whereas their effect on the
mechanical characteristics strongly depends on the specific diluent and amount used. Finally, other
components within the total formulation which usually include thickeners (rheology modifiers), fillers
and pigments all in practice influence the final physical characteristics and properties of the cured
polymer. Viscosity and rheology are strongly affected by both the amount and the nature of the fillers
and, moreover, their presence appreciably reduces the exothermic effects of the two component liquid
curing reaction. The addition of fillers also reduces the quantity of the more expensive liquid epoxy
and curing agent. A good balance between filler amount, granulometry and chemical nature of the
fillers (of which silicas and other size and shape graded inorganic fillers are generally used) have an
appreciable impact on the mechanical characteristics of the cured bonded wood joints: Fig. 7 shows
how the value of the parameter ka,w can appreciably decrease down to 0 by varying the ratio between
two different fillers which were added to a commercial epoxy resin, while their total amount (i.e.,
A+B) was kept constant in the formulation. A value of 0 for this parameter indicates a bonded joint
that has delaminated after the bonded joint specimens had been subjected to a number of ageing
cycles. The parameter ka,w is discussed in detail in Section 3.1. Generally, the components used in the
formulation of adhesives (including all the ancillaries mentioned above) strongly influence their
rheological behaviour (related to both the liquid mixture and the cured product) and their mechanical
and physical properties: Fig. 8 shows how the formulation methodology can appreciably affect the
stress-strain behaviour of two different products, both intended to be used for structural interventions
on timber elements.
117
3
ka,w
1 k
Fig. 7 Value of the parameter a ,w (as defined in
R² = 0.978 Eq. 1 below) vs. the ratio between two different
0 fillers (A and B). Both values reported in the axes
0 2 4 6 8 10 are dimensionless (adapted from [21])
filler A / filler B ratio
Fig. 8 Stress-strain behaviour for two commercial

epoxy products of different rheology, one brittle
and the other ductile (adapted from [22])
2.2 Polyurethane resins

Two-component polyurethanes are polymers that form when a prepolymer reacts with a hardener. The
prepolymer is obtained by the reaction between a terminal hydroxyl of a polyester or polyether (such
as some polyols) with a polyfunctional isocyanate (they contain the extremely reactive group
‒N=C=O in their molecule). When isocyanates are used in excess in that reaction (for instance, in the
case of a foamed product) the terminal moieties of the prepolymer are constituted by ‒N=C=O
groups, otherwise they are rich in ‒OH groups. Among polyfunctional isocyanates, diisocyanates are
those most commonly used to produce polyurethanes. They include toluene diisocyanate (TDI), 4,4’-
diphenylmethane diisocyanate (MDI), hexamethylene diisocyanate (HDI), and isophorone
diisocyanate (IPDI).
The hardeners of polyurethane resins are OH-rich compounds, very often having a similar chemical
nature to the compounds used to produce prepolymers. However, the same environmental humidity,
or the moisture contained in wood, may start the curing reaction, and hence act as a hardener. In fact,
‒N=C=O groups are able to react with hydroxyls to form amines, according to the reaction
represented in the scheme reported in Fig. 9.
Following that, amines can react

R  N  C  O  H 2O  R  NH 2  CO2
with other ‒NCO groups to form
Fig. 9 Scheme of the reaction between an isocyanate and a
urea-type linkages (‒NH-CO-NH‒),
hydroxyl group (in this case represented by a water molecule) to
form an amine (R is a residual) as schematised in Fig. 10. As a
consequence, the molecular weight
of prepolymers increases very
quickly to form a three-dimensional
118
R  NH 2  R' NCO  R  NH  CO  NH  R' network. It is also possible that, in

Fig. 10 Scheme of the reaction between an amine and an addition to the schemes shown in
isocyanate to produce a disubstituted urea-type linkage (R and R’ Fig. 9 and 10, other reactions
are different residuals) may occur, such as those producing
isocyanate trimers, allophanates, and
biurets.
However, these latter reactions are facilitated at high temperatures.

It should be noted that in the reaction represented in Fig. 9 carbon dioxide is formed: this gas is
responsible for the development of gas bubbles within the cured polymer. However, the amount of gas
developed can be regulated by accurately controlling the amount of ‒NCO groups in the prepolymer,
or by adding more amines in the hardener (see the scheme in Fig. 10) in order to increase the
amine/polyols ratio.
In respect of epoxy resins, it is evidenced that additives and fillers can appreciably influence both the
rheology and the mechanical performances of the formulated products. However, in the case of
polyurethane resins catalysts can also be used to control the curing reaction rate. For instance, dibutyl
tin dilaurate can be used for silicone-diisocyanates. However, both the composition and the real
efficacy of these catalysts strongly depend upon the specific chemical components used to produce
the adhesives.
2.3 Phenol-resorcinol-formaldehyde resins
Phenol-resorcinol-formaldehyde (PRF) resins are two-component thermosetting adhesives which are
mostly used in the production of exterior-grade wood lamination and finger jointing. Although they
produce high strength bonds very resistant to both water and weather when exposed to many climatic
conditions, their use in bonding timber on-site is reduced owing to their limited gap-filling quality and
their requirement to be used in a controlled pressure situation.
Counter to pure phenol-formaldehyde adhesives, the presence of the resorcinol moiety, which is very
reactive, makes PRF products able to also set at room temperature at an acceptable rate of cure. As
resorcinol is highly expensive, the phenol counterpart is used in PRFs to reduce the product costs.
Appreciable reduction in the resorcinol/phenol ratio has been attained since the introduction of these
products into the market having reduced from an almost 1:1 ratio to current values of 0.18 or less This
is also due to the introduction of branched resins as opposed to the previous linear types. The
chemistry at the basis of PRF is related to the reactivity of 1,3-dihydroxybenzene (resorcinol), which
reacts with formaldehyde according to the following scheme (further details can be found in [23]):
CH2O  CH2O
resorcinol  linear resorcinol condensate  resorcinol  formaldehyde crosslinked polymer
 H 2O
Commercial PRF products are distinctively brown coloured (some blue preparations can also be found
when urea is used in the mixture), and they are always used as water or hydro-alcoholic solutions,
usually with a 50-60% solid content, which implies that they are characterised by appreciable
volumetric swelling after curing. Solid hardener is seldom used. Commercial products also contain
fillers (of inorganic or organic nature, such as wood flour), which give a consistency to the mix and
are able to control both viscosity and thixotropy.
119
3. Testing timber joints and adhesives for on-site applications

Several standards and practices exist to test wood adhesive for structural uses. Examples of these
standards are EN 301:2013 and the EN 302 series (this latter standard is made up of 7 parts, from EN
302-1:2013 to EN 302-7:2013, each one dealing with a different aspect of the characterisation), which
are devoted to classifying wood adhesives for structural uses1. The Types defined in EN 301 are
expressly recalled in Eurocode 5 for the design of timber structures: Type I adhesives (such as PRF)
are suitable for prolonged exposure to both high temperature and moisture, whereas Type II products
(such as selected emulsion polymerised isocyanates, in brief EPI) are more suited to use in the interior
of well-ventilated buildings. Additionally Type I adhesives are designated to be included in all three
considered Service classes, whereas Type II adhesives can be used only in Service Class 1.
In addition to the characterisation of structural adhesives, other standards exist that deal with the
characterisation of glued products which are usually devoted to the certification and marking of the
same products. Examples of such products are wood panels and glued laminated timber beams. In
those cases, the standard procedures are targeted in order to characterise the bonding quality (that is,
the quality of the adhesion between the adhesive and the specific product to be bonded) rather than the
adhesive itself.
Nevertheless, the specific properties (as indicated in the introduction) that characterise products to be
used on site make it difficult to incorporate them into current standards.
Therefore, in the following sections some tests are described which are specifically intended to verify
whether the current on-site bonding procedures are correct. These tests are based on a series of former
proposals discussed within CEN TC 193/SC1/WG11, which is currently inactive. However, certain
work carried out under the auspices of the above Working Group is considered useful for adhesive
producers, designers and end users in assessing the quality of site based adhesion related procedures.
3.1 Shear tests
As stated in the introductory notes, one of the distinctive features that differentiates bonding on site
from factory production bonding is the appreciable difference in bondline thickness of the assemblies.
This in turn induces a completely different shear area stress distribution, and in addition the cohesive
internal strength of the adhesive layer assumes a significant role in determining the performance of
the joint. It was previously shown that for a predictive method (based on finite element analysis) of
the strength properties of a glued lap joint, within the range 0.05-0.2 mm, a thinner glueline produces
higher shear stress and lower normal stress [24]; moreover, the shear stress at the critical stress
location (close to the bonded edges) is sensitive to glueline thickness. Therefore, it is doubtful if both
the test configuration and the specimen geometries adopted in standards EN 302 can be directly
reflected in any assessment of thick bondlines. Consequently it is questionable whether other
requirements specified in the standard EN 301 are applicable to thick bondlines. In fact, in EN 301 it
is expressly stated that the same standard does not cover the performance of adhesives for on-site
gluing (the same is also indicated by the other main standardisation bodies, such as ISO or ASTM).
Additionally, both the test geometry and the requirements in EN 301/302 have been specifically
developed for adhesives applied to a timber substrate of beech2 rather than the range of timbers
generally used for on-site construction and subsequent bonding as these would be various hardwoods
and softwoods, such as oak, poplar, chestnut, spruce, fir and pine. Beech as a species is not commonly
1
Other standards and classes are provided for non-structural wood adhesives.
2
Beech was used in the code in order to reduce the influence of timber variability on the adhesive behaviour.
120
encountered in timber structures, whereas conifer elements constitute the most generally used species
in wood construction. It is well known that the surface characteristics (wood anatomy, quality and
quantity of extractives, apparent density and porosity, together with the mechanical properties) of
beech (strength, elasticity and deformability) are quite different from conifers (such as spruce).
With regard to the above, new methods have been specifically designed and proposed for using adhesives
on-site for structural timber [25], [26]. These were developed by comparing several factors, including:
- loading type (tensile shear as in EN 302-1, or compressive shear as in ASTM D3931 [27]);
- specimen configuration (the comparison of three different geometries where loadings were applied
to the gluing plane);
- comparison of joint thicknesses of 1, 3 and 4 mm;
- loading rate (comparisons were made between the load speed and the displacement crosshead control type);
- effect of repeated hygro-thermal cycles on the bond strength. A number of procedures were tested,
including milder systems where only the environmental humidity conditions were varied, and more
severe methods such as those provided in both EN 302-3 and ASTM D2559 [28].
Results provided evidence that compressive shear tests were more effective than tensile tests,
particularly for thick assemblies made with spruce. Measured strengths were higher and less affected
by the other considered factors. This was because specimens tested in tension are predominantly
influenced by normal stresses [29], [30]. This occurrence induces a combined (shear and peeling)
state of stress that can affect the load-bearing capacity of the specimens. Moreover, it was previously
reported that frictional forces on the horizontal surfaces of a compressive shear arrangement are
beneficial from a load-bearing capacity point of view, because such forces can prevent tensile peel
stresses from developing [30].
Compressive shear has been used in laboratory tests. Bröker and Kühl [31] tested epoxy adhesive
products filled with cellulose fibres with a thickness up to 10 mm in a 3-ply compressive shear
specimen. They found no significant differences between the various specimen thicknesses when
tested dry. In contrast, differences were appreciable in specimens subjected to a specific hygrothermal
cycle. Moreover, in a comprehensive analysis of factors affecting the compressive shear strength of
solid wood specimens, Okkonen and River [32] found that, coupled with the effect of wood species,
the presence of an offset in the shear plane had the strongest influence on shear strength (stronger than
specimen shape and size, or grain orientation). It has also been shown in [25] that spruce-epoxy
assemblies with the presence of an offset evidenced a higher strength than the solid spruce wood
specimens of the same dimensions. This was attributed to a larger resisting surface, which in turn
allows greater glue penetration into the wood, which forms a strong interphase between the two
materials and to a higher plasticity of the thick adhesive layer with the consequent minimisation of the
load concentration at critical points. It should be noted that all the foregoing experiments show that
the dissipative ability of a thick adhesive layer is highly important in determining the mechanical
behaviour of glued joints and its performance following, for example, temperature or moisture-
induced changes.
Additionally, it is noted that the test configuration influences the results: the asymmetrical typology
tested by Lavisci et al. [25] (load applied on a single, component of a dual wood adhesive interface)
showed the lowest strength, whereas (based on the results of the statistical Tukey test) the other two
possible symmetrical types were mechanically equivalent. Considering also the failure patterns in
121
each typology, the same authors suggested that using the symmetrical configuration shown in Fig. 11
was appropriate in measuring the shear strength of thick-bondline specimens, and it produced clear
fractures developing close to one of the two interfaces. It was also noted that speed of the load
application did not affect the typology, which allowed for certain plasticity in the test configuration.
When ageing cycles are introduced to bonded wood assemblies, it appears that their effect strongly
depends on the severity of the adopted hygro-thermal conditions. Mild conditions induce very limited
strength variations, often lower than 20% of the value in normal conditions, which in re-conditioned
specimens represents the threshold provided in EN 301 for a cycle to be discriminating [33]. Instead,
more severe conditions, like those adopted in either standard EN 302-3 [5] or ASTM D2559 [28],
induce larger dimensional variations into the specimen and consequently more appreciable differences
(both mentioned codes are for wood-to-wood bonding). Lavisci et al. [25] also evidenced how testing
specimens still at their wet state after the hygro-thermal cycles produced values which were very
different from the ones in standard conditions (as a reference, for wet testing, the same EN 301
provides a strength decrease < 40%), and differentiated results according to the epoxy product tested.
In general, structural adhesives which exhibit a low wet strength are considered with suspicion [34],
because this is the evidence of crack formation, resulting in a damaged interphase [35].
Fig. 11 Scheme and dimensions (in

mm) of the test specimen for
compressive shear tests
The comparison of different commercially available products showed that the test method [25] is an
efficient tool for the characterisation and the differentiation of the performances of structural epoxy
adhesives for on-site applications (such as the structural repair of old timber elements). As a
consequence, it is suitable for the purpose of initial evaluation of products and product approval.
Moreover, in a different work [36] it was suggested that comparing the strength values of the glued
joint to those of solid wood was a better criterion to evaluate the joint performances than using pre-
defined values (compared to those suggested in EN 301). This comparison was based on the actual
shear strengths of both kinds of samples (glued and solid wood), which are measured at the time of
testing and with the same equipment. Given the high variability in wood properties (according to
Green [37] the average coefficient of variation for wood shear strength parallel to the grain is 14%,
with appreciable differences existing between batches), this approach limits the influence of
variability, thus considerably improving the reliability of the results. This proposed approach relies on
calculating a unique parameter for the adhesive, defined as [38]:
ka ,w  dry wet (1)
where for each specimen  dry is the ratio of shear strengths in normal conditions (that is, conditioned
at 20°C and 65% r.h.), and  wet is the ratio of shear strengths in the wet conditions (after the
specimens were subjected to hygro-thermal cycles). Such latter ratios can be calculated as:
122
 J ,std
 dry  (2)
 W ,std
 J ,wet
 wet  (3)
 W ,wet
where  J ,std and  W ,std are the shear strengths of glued joint and solid wood, respectively, in standard
conditions (20°C/65% r.h.), and  J ,wet and  W ,wet are the shear strengths (in wet conditions) of glued
joint and solid wood after accelerated ageing. Through the numerical value of the parameter k a ,w , the
adhesive can be eventually classified for its use on timber structures. For instance, in [36] the authors
suggest the use of products with at least k a ,w >0.8 for applications in Service Class 3, as defined in
Eurocode 5.
3.2 Adhesive strength of bonded rods
Pull-pull or pull-compression configurations are commonly used to evaluate the strength of glued-in
rods in timber parallel to grain, whereas pull-beam or pull-pile foundation configurations can be
additionally used with rods inserted perpendicular to grain [39], [40]. Moreover, while some authors
have preferred to use a one-sided test set-up [22], [41], others have chosen the two-sided set-up [42]–
[44]. Jensen et al. [45] have accurately described the mechanics of the two test set-ups. However, to
date there is no agreement among structural engineers about the correct approach for the design of
glued-in connections, as also evidenced by the lack of a standard procedure adopted at European level
for this kind of joint. However, although non-standard, the configurations mentioned above have
provided evidence that the various design factors that affect the bond performance include anchorage
length, rod diameter, bondline thickness, rod-to-grain angle, edge distance, adhesive stiffness, and
wood density. In addition, it is relatively easy to provide a single direct pull out test configuration for
simplified on-site testing and proof loading.
The testing procedure reported here is biased towards the way the load can be applied on site to glued-
in assemblies and it is intended as a method for assessing the overall performance of a joint within a
specific bonding process used in a workshop. Within this context, the expression ‘bonding process’
refers to all the elements of the process, application procedures and designer’s specifications. Hence
the evaluated performances are more related to verifying the correct application of the product,
according to the designer’s or the manufacturer’s instructions, in an on-site bonding situation. This
implies that tests according to the reported procedure can also be carried out directly on site.
Bonded wood specimens with the internally inserted rods are prepared similarly to the test
configurations mentioned above. This results in the process of rods having to be fixed into drilled
holes where the diameter of the hole exceeds the diameter of the rod, and where the annulus is filled
with the adhesive. Rods are centred into the hole by means of ancillary small devices, such as collars
at the top and bottom of the bonded rod length. In consideration of the fact that the described
procedure is also intended for testing specimens directly on site, they can be prepared as separate and
individual samples to be tested later, or as an integral part of the structure. In this way the prepared
samples are truly representative of the work to be carried out. The rods must protrude from the
bonded area by a suitable length (in the order of 240 mm) in order to allow a hollow centre cylindrical
hydraulic ram to fit over the extended anchor and receive a suitable collet and collet rod gripping
holder (Fig. 12).
123
a) b) Fig. 12
(a) General view of the
equipment for on-site tensile
pull out testing;
(b) typical assembly for the
hollow ram rod and collet
gripping device for one
possible rod as described
within the text.
(c) Easily assembled
lightweight equipment for
on-site testing of specimen
anchorages.
c)
Fig. 13 Test
arrangement (in the
small picture at the
right-top corner) and
output of the test as
Temperature vs. time
curve for three
commercial adhesives
and the ambient climate
Following the minimum recommended cure time for the specific adhesive the load is applied using
the hollow ram assembly described above and the use of a suitable hydraulic single action pump with
certificated gauge (Fig. 12). This is a relatively easy-to-perform on-site test, provided that the
following procedures are followed: a) ensure that the hydraulic ram lower face is set at a 90° angle to
124
the bonded rod (normally a levelling bridge is used in order to avoid any distortion of the adhesively
bonded rod during the loading procedure); b) ensure that the load is applied evenly to minimise any
uneven or excessive speed during the incremental loading. In this respect, a valid procedure would be
that of loading by increments of 20% of the expected strength of the joint, and waiting a pre-defined
time between individual increments, in order to assess if there is any loss at each and every individual
increase of load. This is usually observed by checking any movements of the gauge between increases
of load.
Pull-out values below the expected minimum loads may indicate that an error has (or errors have)
occurred during the original bonding process, or that the anticipated pre-calculated load was
excessive.
3.3 Verification of the on-site mixing
Structural adhesives used for timber structures are bi-component products which have to be mixed
immediately prior to their application. Mixing procedures are relatively easy to carry out. In practice,
the total contents of a can, usually Part B, should be poured into the container which holds Part A.
The two components should be thoroughly mixed with a suitable flat bladed spatula ensuring that the
mix is homogeneous and does not entrap excessive amounts of air. This procedure requires care and
good practice to ensure a thorough mix. However, it is well known that mixing errors sometimes
occur, even with the most experienced operatives. It is also known that the mechanical properties of
the mixed epoxy systems, including properties of bonded wood joints, are at a maximum when the
correct (stoichiometric) ratio between the two components is produced. Nevertheless these properties
may decrease appreciably if: a) one of the two components (A or B) is lacking; or b) when the mixing
of the two components is incomplete. The latter is evidenced by incomplete curing in some zones of
the retained sample of the mixed batch.
A very simple and effective method to verify the mixing on site of individual batches of a mixed
adhesive is by filling one or more small cylindrical standard size pots (approximately 20 cc in
volume) (Fig. 13). They are then individually placed in an insulated box to ensure approximate
adiabatic conditions during the initial cure phase. The pots are left to cure in pre-determined
temperature conditions (for instance, the same temperature as that in the workshop and/or work site),
while a thermocouple placed inside the pot measures and records the exothermic temperature rise
during the curing reaction. The curing adhesive should be left in situ during which time the
temperature will dissipate. When ambient temperature is reached the cylinders of cured adhesive may
be removed and will be ready for compressive testing. The cylinders need to dwell for a certain time
until a sufficient compressive strength is reached. They may then be compressively tested in a
laboratory. Both the cure schedule and the compressive strength can be compared to the
corresponding values declared by the adhesive producer, thus verifying the correct mixing of the
product during use.
It is worth noting that while compressive strengths are commonly found in the manufacturer’s
technical data sheets, which provide a time/temperature comparison against various compressive
strengths, they do not normally include the cure schedule. Therefore, this information has to be
specifically sought from the adhesive manufacturer. It is also worth stating here that the cure schedule
of a product depends not only on its specific adhesive composition but also on the ambient
temperature at which the composition is mixed, the initial temperature of the two components of the
mixture prior to mixing, and finally on the volume of an individual mix.
125
4. Effects related to long-term duration of loads

Several investigations have shown that rods glued in timber (in this present section, glued-in rods are
considered as representative of the wood assemblies prepared on site) and subjected to long-term
duration of load may exhibit a decrease of mechanical performance over a long period of time [1],
[15], [16]. In these cases, experimental data usually conform to the damage accumulation modelling
for timber, described by the Madison curve [46]. In fact, if the bonded joint is sustained during the
entire service life and the ambient temperatures do not substantially exceed the usual temperatures in
indoor conditions (normally 20°C) the long-term behaviour of those bonded joints is similar to the
creep behaviour of wood3. This was observed for epoxies in [16] for a constant very humid climate of
85% RH at 20°C, where the strength decrease for the epoxy-bonded assemblies could be described
roughly by the usual damage behaviour of timber (conversely, they also observed a strength reduction
for the polyurethane-bonded joints). In Eurocode 5 this strength reduction is taken into account by
means of the modification factor, kmod, used to derive the design values. The values of k mod
approximately conform to the behaviour of the stress level versus the logarithm of time to failure for
timber, according to the linear Madison curve [47]. These values are defined through consideration of
load duration classes: permanent (e.g. self-weight), long-term, medium-term, short-term (e.g. snow
and wind), and instantaneous (e.g. accidental load). kmod values are not considered dependent on the
adhesive type, and they are the same for the climatic Service Classes 1 and 2, which are, in the case of
timber elements, the usually accepted classes for bonding on site.
However, several tests also showed that the mechanical performance of bonded-in rods may depend
on both the temperature and the environmental humidity in which the same bonded joints are kept.
Assemblies exposed to cyclically varying climates in terms of temperature and relative humidity
revealed that ambient temperatures above approximately 50°C caused significant short- and long-term
strength losses in joints bonded with epoxies, whereas a temperature above 40°C and humid
environments both considerably reduced the strength in joints bonded with polyurethanes [16].
All behaviours described so far are due to a concurrence of effects:
a) Timber is subjected to creep, whose entity and behaviour strongly depends on both the
environmental humidity and the loading history (this aspect is extensively covered in wood-related
literature and requires no further discussion at this juncture) [48].
b) Visco-elastic adhesives such as epoxies and polyurethanes also suffer from creep: its level of
importance is related to how close the service temperature is to the T g of those products (the closer it
is, the larger the creep effect), to the humidity level (humidity can potentially act as a swelling agent),
and to the adhesive cross-link density. The resulting progressive deformation will continue until
rupture or yielding. According to ASTM D2990-01 [49], a creep displacement versus time curve can
be considered as being in three stages: (i) primary creep, (ii) secondary creep and (iii) tertiary creep.
Fig. 14 shows the creep displacement for two commercial products tested: a) well above its Tg
(Product a), and b) just below it (Product b): the three stages of creep are well evident in both
products (a) and (b), although to a much different extent (it is worth noting that the whole duration of
the experiment was limited, as both Products were tested at a temperature close to their Tg).
3
Creep is usually defined as the progressive deformation that visco-elastic solids evidence under constant load.
126
Fig. 14 Creep displacement versus test

duration time curve at a 20% of the
adhesive strength loads for two
commercial products
In the primary creep stage, the creep rate starts at a relatively high value, but decreases rapidly with
time, which may be due to the slippage and reorientation of polymer chains under persistent stress.
After a certain period, the creep rate reaches a steady-state value in the secondary creep stage, in
which the normal duration is relatively long. Finally, the material transcends into the tertiary creep
stage, where the creep rate increases rapidly and final creep failure occurs. However this general
behaviour is only shown when an adhesive is under or close to its glass transition temperature. The
converse is that when substantially above its Tg, a polymer hardly shows any primary and secondary
creep, but the tertiary creep is very obvious.
c) The occurrence of delamination or cracks taking place at the adherend interfaces may appreciably
decrease the bondline duration, mainly under appreciable levels of load. The extent of this reduction
may be very high: as previously shown [1] specimens subjected to accelerated ageing, constituted by
realistic temperature and relative humidity cyclic variations, showed failures after only 9 months of
weathering. The tension load in those specimens was 20% of their average pull-out strength (this
corresponds to two or three times the actual in-practice working load values). However, cracks and
fissures in both the timber and the adhesive were observed after weathering: as also recognised by the
same authors, these occurrences were due to the very small ratio of the adhesive anchorage length to
rod diameter (both anchorage length and diameter measured 10 mm) and the fact that the bond-line
was not concealed inside the timber element.
From the adhesive standpoint, it is possible to increase the T g of room-temperature-curing products by
partially reformulating a suitable and targeted product component formulation. It was shown in [50]
that the addition of liquid rubber as nano-fillers to the basis of a commercial adhesive induced a strain
recovery under constant load, due to the restriction in the slippage, reorientation and motion of
polymer chains, thus increasing the Tg of the modified product (from 32°C to 43°C). As a
consequence, the addition of these nano-rubber particles reduced the initial creep rate in the primary
creep stage, and a much more stable steady-state creep rate within the second creep stage was also
observed in the pure adhesive subjected to creep4.
4
It is of note that the creep behaviour obtained from experiments using adhesive-timber joints is better (more
precautionary) than that obtained from the adhesives alone under the same conditions.
127
5. Future challenges
On-going research is required to a) increase the current performance of existing adhesives, which are
increasingly used at their limits, and b) to fulfil the expectations and requirements of a growing
number of end users. Those issues which seem to be of particular interest and importance are:
- Temperature-related issues: The majority of current commercial 2 part (also referred to as 2K)
epoxy adhesives have a Glass Transition Temperature, Tg, of 50-70°C or less. These values are very
close to those attainable in summer time within roof structures in the countries of Southern Europe. A
class of products able to cure at room temperature but characterised by higher values of Tg would be
an interesting field of research activity. Some attempts to overcome this limitation by using nano-
fillers have already been introduced, but much more effort is required to make this improvement
widespread and easier to manage. In principle the challenge is to improve the glass transition
temperature whilst still controlling the modulus of elasticity of the cured adhesives. However, there
are on-going observed experiences of in-service temperature conditions substantially exceeding the
theoretical and measured Tg of the installed adhesive. To date, ongoing observations on these on-site
moment resisting joints do not indicate any sign of structural demise.
- Plasticity/rigidity of adhesives as related to long-term effects of loading: Products to be used on
site need to have achieved a balance between a plastic and a rigid behaviour. In effect, a flexible
elastic or ductile plastic behaviour is necessary to follow the dimensional variations undergone by
wood following changes in environmental humidity. At the same time, a stable behaviour allowing an
acceptable resistance to loads applied over a long period is also required.
On the other hand, in most of the wood adhesives commercially available to date, appreciable creep
behaviour is usually also associated with ductility. The current solution is to use additional fillers such
as micro/nanofibers and micro/nanoplates, or other additives which are able to substantially alter this
general behaviour. Any of the current ductile products, although more able to follow the wood
deformations, are also responsible for the reduced performances in tests with long-term duration of
load.
While Section 4 above has shown how nanotechnologies can be successfully used to overcome this
inconvenience, unfortunately only a few products fulfilling this requirement are commercially
available. Further research and product development is required to make these characteristic
improvements possible.
- Smartness: Considering that maintenance costs are increasing, an ideal series of products to be used
on site should be ‘active’, in the sense that they should change one of their easy-to-detect properties
(e.g. aesthetic) when they are no longer active, for instance owing to delamination. As an alternative,
or eventually as an additional requirement, they should be self-repairing, e.g., able to self-repair
limited delamination (self-healing effect). Analogously, multifunctional polymers, which are
optimised for performance in a range of application areas (surfactants, coatings, dispersants etc.), can
be considered as an advanced challenge to face with regard to wood adhesives. In effect this could
even include direct combination with integrated sensors to be used for the active monitoring of timber
structures.
- To be monocomponent: In practice, a monocomponent adhesive is easier to manage in workshops
than a bi-component one. However, in the case of thick, or very thick, bondlines it is difficult for the
interface reaction between the timber and the adhesive to continue reacting into the thick portion of
the glue line (similar to that involving the OH groups in wood) to extend towards the bulk of the resin,
128
mainly when inert fillers are used. A potential solution could be the use of a specific type of latent
catalyst or catalysts.
- Acceptance criteria for products to be used on site: Although this aspect does not require a
research effort, acceptance criteria as a standard is nevertheless important for adhesive manufactures,
designers, and end-users. A good basis for establishing the requirements to be fulfilled is described
previously in Section 3. To date, an attempt to establish the requirements of this class of product is
under discussion at CEN meetings. This acceptance criteria process could include an evaluation of
bonding qualities of the products in existence within existing structures.
6. References
[1] Custódio, J. Broughton, J., Cruz, H., “Rehabilitation of timber structures: novel test method to
assess the durability of bonded-in rod connections,” Materials and Structures, Vol. 45, No. 1–2,
2012, pp. 199–221.
[2] Radovic, R., Goth, H., “Entwicklung und Stand eines Verfahrens zur Sanierung von Fugen im
Brettschichtholz,” Bauen mit Holz, Vol. 9, 1992.
[3] EN 302-1, Adhesives for load-bearing timber structures - Test methods - Part 1: Determination
of longitudinal tensile shear strength. Brussels: European Committee for Standardization CEN,
2013.
of resistance to delamination. Brussels: European Committee for Standardization CEN, 2013.
of the effect of acid damage to wood fibres by temperature and humidity cycling on the
transverse tensile strength. Brussels: European Committee for Standardization CEN, 2013.
of the effects of wood shrinkage on the shear strength. Brussels: European Committee for
Standardization CEN, 2013.
[7] EN 302-5, Adhesives for load-bearing structures - Test methods - Part 5: Determination of
maximum assembly time under referenced conditions. Brussels: European Committee for
of the minimum pressing time under referenced conditions. Brussels: European Committee for
of the working life under referenced conditions. Brussels: European Committee for
[10] EN 301, Adhesives, phenolic and aminoplastic, for load-bearing timber structures -
Classification and performance requirements. Brussels: European Committee for
[11] EN 1995-1-1, Eurocode 5: Design of timber structures - Part 1-1: General - Common rules and
rules for buildings. Brussels: European Committee for Standardization CEN, 2008.
[12] Broughton, J.G., Hutchinson, A.R., “Pull-out behaviour of steel rods bonded into timber,”
Materials and Structures, Vol. 34, No. 2, 2001, pp. 100–109.
[13] Kemmsies, M., Comparison of Pull-out Strengths of 12 Adhesives for Glued-in Rods for Timber
Structures, SP Swedish National Testing and Research Institut, Boras, Sweden, SP Report
1999:20, 1999.
[14] E. Serrano, “Glued-in rods for timber structures. An experimental study of softening behaviour”,
Materials and Structures, Vol. 34, No. 4, 2001, pp. 228–234.
129
[15] Bainbridge, R., Mettem, C., Harvey, K., Ansell, M., “Bonded-in rod connections for timber
structures. Development of design methods and test observations,” Int. Journal of Adhesion &
Adhesives, Vol. 22, No. 1, 2002, pp. 47–59.
[16] Aicher, S., Dill-Langer, G., “Influence of moisture, temperature and load duration on
performance of glued-in rods,” In: PRO 22: International RILEM Symposium on Joints in
Timber Structures, Stuttgart: RILEM Publications, 2001, pp. 383–392.
[17] Raftery, G.M., Harte, A.M., Rodd, P.D., “Bond quality at the FRP–wood interface using wood-
laminating adhesives”, Int. Journal of Adhesion and Adhesives, Vol. 29, No. 2, 2009, pp. 101–
110.
[18] Broughton, J., Hutchinson, A., “Effect of timber moisture content on bonded-in rods,”
Construction and Building Materials, Vol. 15, No. 1, 2001, pp. 17–25.
[19] Sterley, M., Gustafsson, P.J., “Shear fracture characterization of green-glued polyurethane wood
adhesive bonds at various moisture and gluing conditions,” Wood Material Science &
Engineering, Vol. 7, No. 2, 2012, pp. 93–100.
[20] Pizzo, B., Compatibilità, durabilità e reversibilità nel restauro delle strutture lignee. Diagnosi
del degrado, tecniche e materiali per il consolidamento, Ph.D. Thesis, University of Palermo,
1999.
[21] B. Pizzo and P. Lavisci, “Un approccio alla valutazione quantitativa della compatibilità nel
restauro strutturale del legno,” in Atti del Convegno “Dalla Reversibilità alla Compatibilità”,
Conegliano, 13-14 giugno, Firenze: Nardini, 2003, pp. 77–87.
[22] Feligioni, L., Lavisci, P., Duchanois, G., De Ciechi, M., Spinelli, P., “Influence of glue rheology
and joint thickness on the strength of bonded-in rods,” Holz Roh Werkstoff, Vol. 61, No. 4,
2003, pp. 281–287.
[23] Pizzi, A., “Resorcinol adhesives” (chap. 29), In: Pizzi, A. and Mittal, K.L. (Eds.), Handbook of
Adhesive Technology (2nd Ed.), New York: Marcel Dekker, 2003.
[24] Chui, Y.H., Ni, C., “Stress distributions in glued wood lap joints subjected to an axial force”, In:
Proceedings 5th World Conference on Timber Engineering, Lausanne, 1998.
[25] Lavisci, P., Berti, S., Pizzo, B., Triboulot, P., Zanuttini, R., “A shear test for structural adhesives
used in the consolidation of old timber,” Holz Roh Werkstoff, Vol. 59, No. 1–2, 2001, pp. 145–
152.
[26] Pizzo, B., Lavisci, P., Misani, C., Triboulot, P., Macchioni, N., “Measuring the shear strength
ratio of glued joints within the same specimen,” Holz Roh Werkstoff, Vol. 61, No. 4, 2003, pp.
273–280.
[27] ASTM D3931-08, Test Method for Determining Strength of Gap-Filling Adhesive Bonds in
Shear by Compression Loading. West Conshohocken, PA: ASTM International, 2008.
[28] ASTM D2559-12a, Specification for Adhesives for Bonded Structural Wood Products for Use
Under Exterior Exposure Conditions. West Conshohocken, PA: ASTM International, 2012.
[29] Strickler, M.D., “Adhesive durability: specimen designs for accelerated tests,” Forest Products
Journal, Vol. 14, No. 1, 1968, pp. 84–90.
[30] Serrano, E., “A numerical study of the shear-strength-predicting capabilities of test specimens
for wood–adhesive bonds,” Int. Journal of Adhesion & Adhesives, Vol. 24, No. 1, 2004, pp. 23–
35.
[31] Broker, F.V., Kuhl, J., “Untersuchungen an zellolosefasergefüllten Epoxidharzen zur Sanierung
breiter Risse in Bauholz,” Bauen mit Holz, Vol. 93, no. 9, 1991.
[32] Okkonen, E.A., River, B.H., “Factors affecting the strength of block-shear specimens,” Forest
Products Journal, Vol. 39, no. 1, 1989, pp. 43–50.
[33] Ceccotti, A., Mannucci, M., Uzielli, L. “Effetti del riassorbimento di umidità sul comportamento
ad estrazione di barre di acciaio ancorate nel legno mediante resina epossidica,” In: Proc. 2nd
Congresso Nazionale sul Restauro del Legno, Firenze, Firenze, 1990.
[34] Raknes, E., “Durability of structural wood adhesives after 30 years ageing,” Holz Roh Werkstoff,
Vol. 55, No. 2–4, 197, pp. 83–90.
130
[35] River, B.H., Ebewele, R.O., Myers, G.E., “Failure mechanisms in wood joints bonded with
urea-formaldehyde adhesives,” Holz Roh Werkstoff, Vol. 52, No. 3, 1994, pp. 179–184.
[36] Pizzo, B., Lavisci, P., Misani, C., Triboulot, P., “The compatibility of structural adhesives with
wood,” Holz Roh Werkstoff, Vol. 61, No. 4, 2003, pp. 288–290.
[37] Green, D.W., Winandy, J.E., Kretschmann, D.E., “Mechanical properties of wood,” in Wood
Handbook. Wood as an Engineering Material, Madison, WI: Department of Agriculture, Forest
Service, Forest Products Laboratory, 1999.
[38] Pizzo, B., Macchioni, N., Lavisci, P., De Ciechi, M., “On-site consolidation systems of old
timber structures,” In: Interaction between science, technology and architecture in timber
construction, C. Bertolini Cestari, T. Marzi, E. Seip, and P. Touliatos, Eds. Paris: Elsevier,
Heritage series, 2004, pp. 323–352.
[39] Tlustochowicz, G., Serrano, E., Steiger, R., “State-of-the-art review on timber connections with
glued-in steel rods,” Materials and Structures, Vol. 44, No. 5, 2011, pp. 997–1020.
[40] Serrano, E., Steiger, R., Lavisci, P., “Glued-in rods,” in Core Document of the COST Action
E34. Bonding of Timber, M. Dunky, B. Källander, M. Properzi, K. Richter, and M. Van
Leemput, Eds. Vienna: Univerität für Bodenkultur, 2008, pp. 31–39.
[41] Del Senno, M., Piazza, M., Tomasi, R., “Axial glued-in steel timber joints. Experimental and
numerical analysis,” Holz Roh Werkstoff, Vol. 62, No. 2, 2004, pp. 137–146.
[42] Steiger, R., Gehri, E., Widmann, R., “Pull-out strength of axially loaded steel rods bonded in
glulam parallel to the grain,” Mater. Struct., vol. 40, no. 1, pp. 69–78, 2007.
[43] Chans, D.O., Cimadevila, J.E., Gutiérrez, E.M., “Influence of the geometric and material
characteristics on the strength of glued joints made in chestnut timber,” Materials & Design,
Vol. 30, No. 4, 2009, pp. 1325–1332.
[44] Serrano, E., “Glued-in rods for timber structures. A 3D model and finite element parameter
studies,” Int. Journal of Adhesion & Adhesives, Vol. 21, No. 2, 2001, pp. 115–127.
[45] Jensen, J.L., Koizumi, A., Sasaki, T., Tamura, Y., Iijima, Y., “Axially loaded glued-in
hardwood dowels,” Wood Science & Technology, Vol. 35, No. 1–2, 2001, pp. 73–83.
[46] Hoffmeyer P., Sørensen, J.D., “Duration of load revisited,” Wood Science and Technology, Vol.
41, No. 8, 2007, pp. 687–711.
[47] Morlier, P., Ranta-Maunus, A., “DOL effect of different sized timber beams,” Holz Roh
Werkstoff, Vol. 56, No. 5, 1998, pp. 279–284.
[48] Morlier, P., Creep in Timber Structures. London: Chapman & Hall (Spon), 1994.
[49] ASTM D2990-09, Test Methods for Tensile, Compressive, and Flexural Creep and Creep-
Rupture of Plastics. West Conshohocken, PA: ASTM International, 2009.
[50] Ahmad, Z., Ansell, M.P., Smedley, D., Md Tahir, P., “The Effect of Long Term Loading on
Epoxy-Based Adhesive Reinforced with Nano-Particles for In Situ Timber Bonding,” Advanced
Materials Research, Vol. 545, 2012, pp. 111–118.
131
132
Reinforcement with glued-in rods
8
René Steiger1, Erik Serrano2, Mislav Stepinac3, Vlatka Rajčić4, Caoimhe O’Neill5, Daniel
McPolin6, Robert Widmann7
Summary
Glued-in rods (GiR) have been successfully used for both constructing new and strengthening existing
timber structures. The research and development of connecting and strengthening timber structural
elements with GiR has been going on since the 1980s. However, agreement regarding design criteria
for these applications has not been reached. Today, some few technical approvals for specific
adhesives suitable to GiR exist, but an approach for the design of connections or reinforcement with
GiR has not been included in the European design code EN 1995 so far. Therefore, it is desired to
gather the current state of knowledge to enable application in practice of the existing and documented
knowledge and experience. This state-of-the-art review (STAR) summarises results from research
done regarding connections and reinforcement with GiR. The review considers manufacturing
methods, mechanisms and parameters governing the performance and strength of GiR, theoretical
approaches and existing design recommendations. For GiR applied as reinforcement similar rules and
requirements apply as for GiR being used as connectors.
1. Introduction
Glued-in rods (GiR) are an effective way of producing stiff, high-capacity connections in timber
structures. In addition GiR have been successfully used for almost 30 years for in-situ repair and
strengthening of structures, as well as for new construction works. GiR are used for column
foundations, moment-resisting connections in beams and frame corners, as shear connectors and for
strengthening structural elements when extensively loaded perpendicular to grain and in shear. Early
examples of their use also include the connection of windmill blades made from glued laminated
timber (glulam) [1, 2].
1)
PhD, Senior Scientist, Empa, Dübendorf, Switzerland
2)
Professor, PhD, Linnæus University, Växjö, Sweden
3)
PhD Candidate, University of Zagreb, Croatia
4)
Professor, PhD, University of Zagreb, Croatia
5)
PhD Candidate, Queen’s University, Belfast, UK
6)
Lecturer, Queen’s University, Belfast, UK
7)
Researcher, Empa, Dübendorf, Switzerland
133
Most applications have used the GiR connections/reinforcement with metal bars glued into softwood. In
practice, glulam made from softwood in combination with rods with metric threads is the most
commonly used combination. Great experience has been gathered in the repair and strengthening of
beams made of solid timber, both softwood and hardwood, and in connecting concrete slabs to floor
beams. For applications where corrosion or weight could be of concern, the use of pultruded FRP rods is
quite common. Some investigations have also aimed at the use of reinforcement bars (rebars), e.g. [3, 4].
Basically all types of adhesives useful for wood bonding have also been tried for GiR, but one and
two-component epoxies, PUR and resorcinol types are those most frequently used in practice. Specific
adhesive products have been formulated to fulfil the needs of GiR connections/reinforcement with
timber, which offer much better performance with respect to strength and durability. A large number
of parameters impact the strength of GiR. Hence, the challenge is to adequately account for these in
design and to provide quality control measures to guarantee a reliable load bearing behaviour of GiR,
which are usually assigned high loads by the designer.
2. Application – Gluing-in the rods

2.1 Variants
There are several possibilities on how to glue rods into the wood [5]. Most often, a hole is drilled into
the timber member with a diameter that exceeds the nominal diameter of the rod by 1 mm to 4 mm.
This will result in glue line thicknesses from less than 1 mm to 2 mm. Thin glue lines are usually
preferred over thick glue lines as many adhesives perform better the thinner the glue line is made and,
in addition the necessary quantity of the expensive adhesive is reduced. In general the holes can be
drilled in any direction relative to the grain. An important step after drilling is to clean the hole
thoroughly. If pressurized air is used for this purpose it has to be verified that the air is free of oil-dust.
If the rods can be set into holes with openings situated at the top of an element an easy variant is to first
pour a defined quantity of adhesive into the hole and then to set the rod (Fig. 1(a)). Depending on the
viscosity and the open time of the adhesive the rods may sink into the adhesive-filled hole due to their
own weight or the rods have to be pushed into the adhesive filled hole. A disadvantage of this method
is that there is no adequate control of the glue line quality in terms of assuring that the adhesive fills all
cavities completely and no voids are present in the glue line.
Another often used technique for setting the rod is to drill a second hole, this second hole being drilled
perpendicular to the hole drilled for the rod. This hole should lead to the lower end of the rod and thus
the adhesive can be injected under pressure from the bottom (Fig. 1(b)). For every rod the injection of
adhesive will be continued until it can be observed that the adhesive pours out at the top of the hole that
contains the rod or at another hole positioned at the desired position. The rod has to be fixed while the
adhesive is injected. If the opening between rod and hole is sealed (for example by means of a moulded
part or super glue), it is also possible to set the rods in a horizontal or overhead configuration as shown
in Fig. 1(c) and (d).
134
(a) (b) (c) (d)

aa a a bb b b cc c c d d d d
Fig.1 Variants for the application of GiR
(a) (b) (c) (d) (e)
Fig. 2 For optimum performance avoid: unwanted inclination of drilled hole (a), inclined setting of
rod in hole (b), eccentric position of rod in hole (c), incomplete insertion of rod in hole (d), voids in
glue line (e)
In the literature other variants of the application of GiR can be found. One of these uses a
concentric continuous hole in the rod for the injection of the adhesive [6]. In another variant
the rod is drilled into an adhesive filled hole with a diameter equal to or smaller than the
nominal diameter of the rod. This procedure can be regarded as a combination of glued-in and
drilled-in rods technology. However, today these two latter methods are not of significant
importance for practical applications of GiR.
2.2 Quality control
As for all glued connections quality control of the manufacturing process is of great
importance. The following parameters have to be checked when GiR connections or
reinforcements are applied:
Material
 Timber: strength class, moisture content (MC)
135
 Adhesive: proven to be suitable for gluing in rods, technical specifications, climatic

conditions, open time, curing time
 Rod: correct geometry, correct type/strength according to design (an overstrength rod
can have adverse effects in this application), corrosion, condition of surface (free of oil
and/or lubricants)
Application
 Hole: position (including edge and rod distances), diameter, depth, inclination,
straightness, cleanliness (Fig. 2a)
 Complete and centered positioning of rod in the hole (Fig. 2b-d). Depending on glue
line thickness the use of spacers and/or centering devices like e.g. plastic or metal rings
or a countersink at the bottom of the hole might be required.
 Adhesive: application according to manufacturer specifications, control of filling level,
presence of voids (Fig. 2e)
3. Key parameters
The following parameters impact the load bearing capacity of GiR connections/reinforcement [5]:
Geometry
 Ratios of area of wood, adhesive area and rod area
 Absolute size of the anchoring zone (represented by hole diameter d h and anchorage length  )
 Slenderness ratio, which is defined as    / d h
 Number of rods, edge distances and rod-to-rod distances

 Rod-to-grain angle (including unintentional deviations from planned angle due to production
process, definition of a tolerance-range)
Material stiffness
 Moduli of elasticity (MOE) and shear moduli of rod, adhesive and wood
 Ratios of MOE to shear modulus for each material (especially important for the wood
material, this being strongly orthotropic)
Material strength
 Strength of the wood (especially shear strength and tensile or compression strength
perpendicular to the grain). Note that the strength of wood is influenced by the density and
that solid timber and glulam are usually assigned to strength classes according to EN 338 [7,
8] or EN 14080 [9] respectively. (Also applies to engineered wood products!)
 Cohesive and adhesive strength of the adhesive
 Ultimate strength of the rod material (for steel rods the yield strength is also important).
136
Fracture mechanical properties of wood and adhesive

 Fracture energy and fracture softening characteristics
Variability of all properties

 Irregularities, i.e. deviation from nominal properties
 Variations in mechanical properties of wood, rod and adhesive
Loading conditions
 Direction of external load on the rod in relation to its axis (pull-out, shearing) and reaction
forces on the specimen that counteract the external load in the tests (Fig. 3)
 Load duration (static)
 Number of load cycles, frequency and amplitude (dynamic)
Other parameters
 Wood species
 Special features to reduce stress peaks and/or to guarantee for a ductile failure mode
 Manufacturing practice (curing time and pressure, surface characteristics etc.)
 Quality control.
F
F/2 F/2
F F
(a) pull - pull (b) pull - compression
F F
e
F/2 F/2 F/2 F/2
e
Pile foundations
(c) pull - beam (d) pull - pile foundation

Fig.3 Different types of loading GiR specimens may be subjected to in tests of axially loaded rods[5,
10]
4. Adhesives
A variety of adhesives have been tested to glue in rods. In early years, traditional wood adhesives
based on phenol-resorcinol (PRF) or epoxies (EPX) were used, while later work has included also the
use of polyurethanes (PUR) for instance. In 1999, Kemmsies made an investigation regarding the
suitability of 12 different adhesives and adhesive/sealants [11]. In the experiments conducted within a
large European research project in the late 1990s, (GIROD), three types of adhesives were used and
compared [12]: PRF, EPX and PUR. From the tests and analyses it was concluded that the adhesives
137
revealed increasing strength in pull-out tests in the following order: fibre reinforced PRF, PUR and
EPX. EPX adhesives develop a strong bond with both steel and the wood, resulting in the wood
becoming the weakest link of the connection. Thus the fracture properties of the wood or the
wood/adhesive interface are decisive for the pull-out strength.
Characterising an adhesive only by terms like EPX or PUR is not sufficient. There are many
adhesives available of each type and they ‘‘can show all types of constitutive behaviour’’ (regarding
EPX: [13]). The pull-out strength of the GiR is obviously related to the adhesive type, but also to the
used wood species, since different adherents may develop different bonding strength with different
adhesives [14]. Generally speaking, and to a varying degree depending on the specific adhesive used,
the bonding strength can be affected by shrinkage during initial hardening, by the adhesive’s
sensitivity to elevated temperatures, by its limited gap-filling qualities and by the sensitivity to
moisture content changes [14]. All these effects have to be taken into account in design [5, 15].
Adhesives for GiR connections must, in addition to good strength and durability, have acceptable
creep and creep-rupture properties. In order to assess these properties tests based on existing methods
(e.g. longitudinal shear strength according to EN 302-1) were developed [16]. However, up to now
there are no specific standards or guidelines available on how to test adhesives for GiR.
The choice of adhesive is not independent of the method used to produce the connections. The main
parameters of concern are the adhesion to the wood, the mechanical link to the rod (interlocking), the
thickness of the glue line and the properties (e.g. viscosity) of the bonding agent [5]. The adhesive
should demonstrate good gap-filling properties.
For the connections with GiR there are many failure locations and modes which can be decisive for
the load bearing capacity (see 5.3). The adhesive might be chosen during the design of the connection
taking into account geometrical properties, requests of application methods and with the aim of
avoiding a brittle failure mode finally making sure that the adhesive bond will not be the weakest link
of the connection [17] in order to profit from the full capacity in shear strength that wood offers. In
countries like Sweden, UK, Switzerland, Germany [18] and New Zealand [19] the most commonly
used adhesives for connections and reinforcement with GiR are 2-component PUR and EPX.
5. Mechanics, failure modes, design philosophy

5.1 Mechanical behaviour of GiR connections
Mostly, our knowledge about the mechanical performance of GiR connections is based on practical
experience and design formulas developed by curve-fitting of empirical data [5]. The majority of the
studies performed have focused on axial pull-out strength of a single GiR and its dependency on
various material and/or geometrical parameters.
During axial pulling, the load transfer between timber and rod is governed by shear of the adhesive.
Depending on the strength of the adhesive and the surface characteristics of the rod and its surface
treatment, the anchorage between the threaded rod and the adhesive may act as a mechanical
connection [20, 21] similar to screws [22, 23]. Some design codes (e.g. [24, 25]) even do not allow to
use rods lacking a threaded surface, since a pure adhesive bond is suspected not to be able to
guarantee a reliable and durable force transfer. The force transferring mechanism is also influenced by
the ratio of the diameter of the hole to the diameter of the rod, i.e. the bond line thickness. In some
sources it is claimed that GiR connections act like a combination of glued and mechanical connections
138
[13, 26, 27]. For rods inserted in undersized holes, it can be expected that the connection strength to a
large extent results from the mechanical interaction between the wood and the thread of the rod [28].
One major advantage with the GiR connections is the transfer of forces directly into the inner part of
the members’ cross-sections [29]. The connection is actually a hybrid one, made up of typically three
different materials (wood, adhesive, rod) with different stiffness and strength properties [14] which
have to work under loading simultaneously. This severely complicates the analysis of the connections
and is one of the reasons for today’s lack of full understanding of the behaviour of this connection
type as well as for agreeing on a design model.
5.2 Theoretical approaches to describe the behaviour of the adhesive bond
The adhesive bond line (by bond line or glue line is meant the adhesive layer plus the interface
between adhesive and adherends) plays a major role in the overall behaviour of the GiR. Different
approaches to describe the laws governing the behaviour of adhesive connections can be found in
literature: (a) traditional strength analyses, (b) analyses based on linear elastic fracture mechanics
(LEFM) and, finally, (c) so called non-linear fracture mechanics (NLFM) analyses [5].
In a traditional strength analysis, one tries to predict the stress (and strain) distribution in the GiR for a
given loading situation, and then applies some failure criterion for this distribution. The failure
criterion can be based on stress or strain, involving also multi-dimensional criteria. The approach will
give a prediction of the load bearing capacity of the GiR, and also give a prediction of the stiffness.
The stress (and strain) distribution can be determined with analytical or numerical methods, the
former e.g. according to the Volkersen theory [30-33].
When using the framework of LEFM, the situation of loading a connection with a pre-existing crack
is considered. The crack introduces a stress (and strain) singularity, and thus a traditional single point
maximum stress criterion is not useful. Instead the crack driving force, also known as the energy
release rate is calculated. The energy release rate is defined as the amount of (elastic) energy released
during crack propagation. The critical energy release rate of the connection, Gc, is the amount of
energy needed to increase the crack area. By assuming that failure of the connection takes place when
the strain energy released is equal to the critical energy release rate of the connection, the load bearing
capacity can be calculated [34]. NLFM provides a framework that takes into account not only the
strength of the bond line (like in a strength analysis) neither only the fracture energy of the connection
(like in the LEFM approach), but in fact accounts for both [34].
NLFM provides a framework that can be said to include both the framework of traditional strength
analysis and LEFM. In traditional strength analysis it is assumed that the strength of the material is
limited and that the fracture energy is either zero or infinite, the latter for the case of perfect plasticity.
If a crack exists, such traditional strength analyses methods will fail since infinite stress (or strain)
will be predicted. The framework of LEFM will, as mentioned above, only be applicable to the cases
with a pre-existing crack. LEFM assumes finite fracture energy but an infinite strength of the material.
NLFM takes into account not only the limited strength of the bond line but assumes also that fracture
energy is limited. Thus NLFM can be used for any situation, independent of whether a crack exists or
not. In NLFM the stress-strain relation used in conventional approaches is exchanged by a non-linear
stress-displacement relation, such that the bond line, after stress has reached the strength of the
material can still transfer load. This post peak-stress load transferring capacity diminishes with
increasing displacement (normal opening or shear slip across the bond line) and will eventually reach
139
zero. Thus, a typical stress versus displacement relation involves both an ascending part (typically the
linear elastic response) and a post peak-stress descending part, known as strain softening [34].
The choice of theory to be applied basically depends on the predicted failure characteristics (brittle or
ductile) of the adhesive bond, related to the properties of the bonding agent, to the size and shape of
the connection and to the stiffness of the adherents [34]. For ductile adhesive bonds, stress based
approaches can be useful. For very brittle adhesive bonds, an approach based on LEFM can be
appropriate, and in theory, a NLFM-approach can be used for both these cases and any in-between
situation. It must be emphasized that the failure characteristic of the bond line (brittle or ductile)
depends on material (strength and stiffness of timber, type and strength of adhesive), geometry
(surface and thickness of bond line) and loading conditions.
The main tendency in improving the theoretical basis for analysis of connections with GiR follows
Gustafsson’s further development of the Volkersen theory taking into consideration the NLFM. Since
the GiR connection is a type of adhesive connection, usually those theories are the base for the stress
analysis. The broad description of available theories and the historical development of them are
available in many sources, for instance in [13].
5.3 Failure modes
The GiR connection acts like a chain consisting of the serial links “rod”, “adhesive” and “wood” [8],
the load bearing capacity and failure mode being influenced by the parameters listed in chapter 3. The
following failure modes are relevant for a single rod. Although such connections are of little interest
in practice, they form the basis for research and the design of group of rods.
1. Failure of the rod due to

a. material failure
b. buckling of the rod in case of compression loading
2. Pull-out of the rod due to

a. adhesive failure at the steel-adhesive interface (in case of lack of rods without profiled surface)
b. cohesive failure in the adhesive
c. adhesive failure at the wood-adhesive interface
d. cohesive failure in the wood close to the bond line
3. Pull-out of wood-plug
4. Splitting failure of the wood due to

a. short edge distances
b. the rod being not set perfectly parallel to the grain
c. excessive perpendicular to the grain loading
5. Tensile failure in the net or gross wood cross-section

In addition to these failure modes for single-rod connections, the following are of interest for multiple
rod connections:
140
6. Splitting failure due to short rod-to-rod distance
7. Group pull-out
Splitting due to shrinkage or excessive shear stresses, and especially due to the stress peaks that are
typically formed at the end of the rod [5, 22, 31] can be prevented by transversely reinforcing the
connection, e.g. by means of self-tapping screws or threaded steel bars glued into drilled holes [35],
crossing potential crack lines, approximately 50 mm from the end of the member [36]. Other
possibilities to overcome the peaks in the shear stress distribution are to countersink the drill hole or
to widen its diameter at the face end [10]. In references [4, 37] it is suggested to shift the anchorage
zone to the inner part of the member (i.e. away from the surface) by either applying no adhesive at the
face end of the drill hole or by turning off the thread of the bar over a certain length in order to
prevent indentation and shear force transfer there. Successful experiments with widened bottom parts
of the drill hole which allow the adhesive to spread in bulbs are reported in [38].
Since moisture induced stresses increase the risk of splitting, the application of GiR is usually
restricted to service classes 1 and 2 [39].
5.4 Design philosophy
Depending on the design philosophy basically each of the aforementioned links can be assigned to be
the weakest one. Whilst it is straightforward to calculate the tensile strength of the rod, at least in
cases where the material quality is clearly defined and is not influenced by too high variations, the
load bearing capacity in the wood and in the adhesive as well as in the interfaces is more difficult to
estimate. In practice, the failure load for each of the failure modes must be assessed and the design
philosophy set, in order that a chosen failure mode can be ensured or prevented respectively. It has to
be clearly differentiated between experimental investigations and design for practice. In the first case
the connections are designed such that usually the wood should be the weakest link (in order to
identify the maximum load bearing capacity of the connection being subject of investigation). In the
second case assigning the rod to be the weakest link allows for ductility and robustness.
Several design approaches have been suggested [5]. One approach could be to ensure that a
connection fails in a ductile failure mode, such as by failure in the steel, which of course must allow
large plastic strains to develop with constant or monotonically increasing load capacity until final
collapse [37, 40]. Some design codes (e.g. the Swiss design code SIA 265:2012 [21]) prescribe this
type of ductile failure, which is favourable for any design case, regardless of materials in use and
regardless of the possibility of seismic actions. In case of multiple rod connections it is even of greater
importance to aim for a ductile failure mode. Only if the steel rods are the weakest link a uniform
distribution of the load among all rods of the connection is possible [37]. Plastic deformations in the
steel rod can develop only if there is sufficient free length for elongation. To achieve this, a part of the
rod near the surface of the timber should be left unbonded [2, 4, 20, 41, 42] and if possible also
necked down to a slightly smaller diameter by turning off the thread [4, 41]. This helps to prevent
mechanical interlocking in this particular part of the anchorage zone and to “force” plastic
deformations to develop in this zone [4, 37, 43]. With respect to ductility there is certainly an
advantage in using mild steel with large yield capacity. For GiR connections in high strength timber
like beech or ash rods of quality 8.8 may be indicated. This is also the case when (in experimental
investigations) pull-out failures are to be achieved in order to derive the optimal anchorage length, to
check performance of a specific adhesive or to study the influence of parameters like wood density or
shear strength of the wood.
141
It is worthwhile mentioning that no matter what failure mode is intended the engineer has to be able to
assess all of the above failure modes, in order to perform the design [5]. The adhesive used, in any
case, shall not be the weakest link, because this would not allow for making use of the full capacity
the glued-in rod connection provides. There is thus no contradiction in performing large test series
intended to assess the pull-out strength of GiR, even if the practising engineer would rather choose a
failure mode based on plastic failure taking place in the rod.
In order to optimize performance of GiR connections (1) the transfer of stresses should be steady, (2)
deviations between force and grain direction should be small, (3) both the rod(s) and the timber
should have similar stiffness (i.e. ETimber  ATimber  ERod  ARod , which in case of steel rods results in
ATimber  16 to 20  ASteel ) and (4) the deformation in rod and timber should be in similar range and
not exceed the ultimate deformation capacity (2 to 3 ‰ for Norway spruce) [37, 41].
6. Design of GiR connections
6.1 Background
Over the past 25 years, despite many national research projects, European projects, COST Actions
(e.g. E13, E34) and constant practical application of GiR there is still no universal standard for the
design thereof [44, 45]. This mainly seems to originate from the many different design approaches
available in the literature for defining the behaviour of the adhesive connections and due to the fact
that a large number of parameters impact the design.
An early design approach was published in 1988 by Riberholt [46], who proposed an equation for the
estimation of the pull-out strength of an axially loaded single GiR. In the 1990’s a considerable
amount of experimental work was done and different design methods were presented (see below).
Certain design methods were introduced into national design standards and in 1997 a proposal was
implemented in a pre-version of the Eurocode, the pre-standard prEN 1995-2 [47]. Although not
being exclusively related to the design of timber bridges, the design rules for GiR had to be included
in part 2 of EN 1995 since, at that time work on prEN 1995-1-1 had already been finalized and it was
not possible anymore to amend this part of prEN 1995. In 1998, the European GIROD project was
launched. The main objective of this project was to establish design rules and the project result was a
new calculation model based on the generalized Volkersen theory (GIROD Project Report 2002,
[48]). This resulted in a proposal to be implemented in the pre-standard prEN 1995-2, Annex C [49].
During the CEN/TC 250/SC 5 meeting in 2003 it was decided to discard the Annex C. Delegates
argued that the proposed code text did not meet the actual status of research (e.g. [50-52]). Recently
both past and actual research has been re-visited with the purpose to propose a design approach that
could replace several national design rules. Proposals and design rules developed during the years are
shown in Fig. 4.
A calculation model has to take into account all relevant parameters that impact the load bearing
capacity of glued-in rods (see chapter 3). Although there are numerous studies and calculation
methods, and although in an earlier version of EN 1995 design methods exists, the basic problem is
still which method to accept and to implement in EN 1995. It is clear that a lack of a common
European design approach is a serious obstacle to the exploitation of the GiR connection [44].
For more than ten years many research efforts and research programs have contributed to the
knowledge about GiR and attempted to provide the information required to prepare design rules
which would allow an increased, more advanced and more reliable use of GiR in timber structures
142
[53]. Stepinac et al. [54] obtained, via online survey, information about the practical use of GiR and
about problems the designer faces when designing this connection. Results were as expected:
Available design rules were characterized as unreliable and unsatisfying. The most commonly applied
design approaches turned out to be the ones in prEN 1995-2, Annex C [49] and in DIN 1052 [24].
Beside the fact that a unified design approach still does not exist, the key reservations with the
available design rules are [54]:
 Definition of rod spacing and edge distances are not reliable for rods under tension and shear
load.
 Design rules (and requirements in rod spacing and edge distances) often are too conservative.
 Ductility should be treated as a key issue.
 There are no reliable rules for multiple rod connections.
 The duration of load (DOL) effect is not accounted for.
 There are no design rules for the case of interacting axial load and transverse load.
 The influence of load-to-grain angle is not addressed.
 Some of the available design approaches contain non user-friendly formulae and/or
parameters which are difficult to assess.
Fig. 4 Standards and proposals containing design rules to estimate the pull-out strength of GiR and
researchers involved in the development in the last 25 years.
6.2 Comparison of design rules

Most of the available design equations are focused only on the pull-out strength of single axially
loaded GiR, since a lot of research has been carried out exclusively dealing with pulling-out of single
rods. In Sections 6.3 and 6.4 calculation models for rods set perpendicular to the grain and rules for
multiple rods are introduced briefly. In the present Section rules commonly applied for the design of
GiR are compared. Diagrams in this Section in general show graphs on characteristic level, except
when stated in the caption of the respective figure.
143
6.2.1 Axially loaded single GiR parallel to the grain

Tlustochowicz et al. [5] and Stepinac et al. [54] explained in detail proposals and design rules
published in the last 25 years. In this state-of-the-art review six design rules and methods which are
most commonly applied are analysed and explained in detail:
Riberholt equation, 1998 [46]: Rax, k  f w1  c  d  lg (1)
prEN 1995-2, 2003 [49]: Rax, k    dequ  la  f ax, k  tan   /  (2)
GIROD equation, 2003 [48]: Pf   f    d  l  tan  /   (3)
Proposal by Gehri, Steiger, Widmann, 2007 [43]: Fax, mean  fv,0, mean    dh  l (4)
New Zealand Design Guide, 2007 [36]:

Qk  6,73  kb  ke  km  l / d 0,86  d / 201,62  h / d 0,5  e / d 0,5 (5)
DIN 1052:2008 [24] and CNR DT 206/2007 [55]: Rax, d    d  lad  f k1, d (6)
where:
Rax,k / Pf / Qk characteristic value of axial resistance [N], [kN]
Rax,d design value of axial strength [N], [kN]
Fax,mean mean value axial resistance [N], [kN]
l / la / lg / lad glued-in length / effective anchorage length [mm]
d nominal diameter of the rod [mm]
dh / h diameter of the drill hole [mm]
dequ equivalent diameter [mm]
e edge distance [mm]
kb / km / ke bar type factor / moisture factor / epoxy factor
ω stiffness ratio of the connection
τf local shear strength of the bond line [N/mm2]
fw1 / fv,α,k / fv,k / fax,k / fk1,d strength parameter / characteristic value of the shear strength of the wood
/ design value of the shear strength of wood across the grain /
characteristic value of the shear strength of the wood at the angle
between the rod and grain direction / design value of the bond line
strength [N/mm2]
fv,0,mean nominal shear strength parallel to the grain of a single axially loaded rod
[N/mm2].
d = diameter of rod
dh d
l = anchorage length
l
dh = strength
Fig. 5 Geometric parameters influencing the pull-out diameterof
ofaxially
hole loaded single GiR.
e = glue line thickness
144
The pull-out strength depends primarily on the interfacial layer and shear strength parameter which is
influenced by mechanical and geometrical properties of three different materials. Hence, a simplified
calculation model for axial loading could be similar to the one of screws:
Rax,k    d  l  f v,k (7)
where:
Rax,k characteristic value of pull-out strength
l anchorage length
d diameter
fv,k shear strength parameter.
However, the mechanics of GiR are complex, so any attempted simplification from the designer’s
point of view would be helpful since making the design of GiR straight forward but might result in
uneconomic connection design. A closer look at the simplified equation reveals several unanswered
questions such as: Which diameter (diameter of rod, diameter of hole or equivalent diameter) and
anchorage length (length of bonded rod or equivalent anchorage length) to use? Can the geometry of
the hole be described by the slenderness ratio    / d ? Which parameters must be included in the
shear strength parameter (timber density, MC of timber, MOE of timber, rod and adhesive, rod
surface, rod material, type of adhesive, slenderness ratio, geometrical factors, etc.)? This (among other
things) is the reason for present standards and proposals differing significantly (Fig. 6 and Fig. 7).
From experts discussions it can be concluded that the most common design rules like the ones in
prEN 1995-2 [49], the former DIN 1052 [24] are on the “safe side” while equations proposed in
various scientific papers in most cases relying on experimental data derived from tests on specific
connection systems deliver much higher values for the pull-out strength. The glue line thickness e is
considered only in some formulas. Some standards propose a maximum value of 2 mm [24], [57],
[21] but do not provide answers for thinner glue-lines. Differences and the influence on the calculated
load bearing capacity are shown in Fig. 7.
Fig. 8 and Fig. 9 show the characteristic value of the pull-out strength of one single axially loaded rod
estimated on basis of different design rules whereby the diameter of the rod and the anchorage length
were varied. Problems occur when defining these two parameters in the equation. The diameter d is
sometimes the diameter of the rod [46], [24], the diameter of the drill hole [43] or an equivalent
diameter [59], [56]. A similar problem applies for the definition of the anchorage length. The former
prEN 1995-2 equation [49], which was based on the GIROD project findings, included several
different parameters. Some of these parameters e.g. fracture mechanics parameters, cannot be
determined easily by engineers in practice.
The influence of wood density has been subject of several studies (e.g. [46], [56], [43], [59]) (Fig. 10).
The opinions in the matter of influence of density on the pull-out strength of glued-in rods differ. The
recommendations given in [47] for the design of GiR connections indicate that the axial strength of
glued-in rods depends on the density of the wooden element. It could be expected that such a relation
exists, having in mind that it has been demonstrated that the pull-out strength of nailed and screwed
connections is dependent on the density of the wooden member [23, 60-62]. On the other hand, the
correlation between density and strength of wood in general is poor [63].
145
Fig. 6 Comparison of the pull-out strength [kN] derived with different design approaches ([36], [43],
[49], [53], [24], [56], [48], [47], [46]), (EPX, l=200 mm, ρk=370 kg/m3 (MC<14%), d=20 mm,
e=2 mm). Blue bars represent characteristic values; red bars represent mean values. References are
given in top down sequence as listed in the graph.
Fig. 7 Influence of glue line thickness on the pull-out strength [kN] (EPX, l=200 mm, ρk=370 kg/m3
(MC<14%), d=20 mm) ([36], [43], [27], [57], [49], [53], [24], [48], [58],). References are given in
top down sequence as listed in the graph.
146
A recent study on the influence of density [43], based on performed pull-out tests on low and high
density specimens of Norway spruce glulam demonstrated that the influence of density on pull-out
strength of the rods bonded in parallel to grain direction can be quantified by a power function of
density  c with an exponent of c0  0,55 . The adhesive used in this case was EPX. The further
testing of rods glued-in perpendicular to grain [59] revealed less consistent results and therefore the
influence of the density of the timber was recommended not to be taken into account or to account for
by using an exponent of c90  0,25 . Bernasconi [64] also reported having found such a relation.
However, other studies [65, 66] showed that if such a correlation exists, it is hard to identify.
Fig. 8 Comparison of the pull-out strength [kN] derived with different design rules ([46], [49], [48],
[24], [56], [48]) when varying the diameter of the rod (EPX, l=200 mm, ρk=370 kg/m3, e=2 mm)
References are given in top down sequence as listed in the graph.
Fig. 9 Comparison of the pull-out strength [kN] derived with different design rules when varying the
anchorage length ([46], [49], [43], [36], [53], [24], [48]), (EPX, d=12 mm, e=2 mm d=20 mm).
147
References are given in top down sequence as listed in the graph.
Fig. 10 Comparison of the pull-out strength [kN] parallel to the grain derived with different design
rules when varying the timber density (EPX, l=200 mm, e=2 mm, d=20 mm) [46], [49], [48], [24],
[56], [48].
From a theoretical point of view, the influence of the density is often regarded as a secondary effect,
meaning that changing the density changes the value of the parameters in the theoretical expressions
for pull-out strength. Thus, an increased density of the wood can influence the load bearing capacity
by increased shear strength of the wood, reduced adhesion to the wood, increased stiffness of the
wood, etc. Consequently, a number of factors can in part counteract each other. It should be noted,
that a possible influence of density on the load-bearing capacity of GiR can only be derived from test
series where the failure occurred in the wood or in the wood/adhesive interface.
6.3 Axially loaded single GiR set in timber perpendicular to the grain
Although in most design rules and proposals equations for pull-out strength of single GiR set parallel
and perpendicular to grain do not differ, it is known that the rod-to-grain angle markedly impacts the
pull-out strength of GiR. In applications with rods set perpendicular to the grain one of the main
parameters is the tensile strength of the timber perpendicular to the grain. Widmann et al. [59], [43]
tested and compared specimens set perpendicular and parallel to grain. The rods set perpendicular to
the grain achieved higher pull-out strengths than those set parallel to the grain, therefore rod-to-grain
angle is regarded to be a parameter which cannot be neglected [43]. Blass & Laskewitz [67] proposed
a mechanical model which (as a simplified version) has been implemented in German standards [24].
From their online survey Stepinac et al. [54] concluded that designers are using the same equations for
rods set perpendicular and parallel to the grain, or are referring to [59] where the pull-out strength is
estimated as follows:
Fax,mean  k1  Agk2 with Ag  l    d h ; k1 = 0,045 and k2 = 0,8 (8)
l anchorage length [mm]

dh diameter of drill hole [mm]
148
6.4 Multiple rod connections

In multiple rod connections non-uniform distribution of forces and interference between the rods
occur [5]. In prEN 1995-2 [49] there was an equation to estimate the pull-out strength of a group of
rods inserted parallel to the grain. This design approach however, was based on failure in the timber
element. The characteristic load bearing capacity of one rod Rax,k was taken as:
Rax,k  ft ,0,k  Aef (9)
where ft,0,k is the characteristic tensile strength of the wood and Aef is the effective timber failure area.
This formulation caused huge dissatisfaction and it was characterized as not reliable (e.g. brittleness
could lead to progressive failure in multiple rod connections). An easy way to reach an uniform
distribution of forces among all rods is to use steel rods and to design the connection such that the
steel rods are the weakest link [37]. In fact, very little data on the behaviour of multiple GiR
connections are available. In a recent study Parida et al. [40] concluded that the use of mild steel as
well as more rods of smaller diameter are effective measures to increase the ductility of the
connection.
For multiple rod connections the spacing between the rods and the edge distances are key issues
governing the load bearing capacity of the connection [5]. Blass et al. [68] studied the influence of
these parameters for axially GiR. The load bearing capacity decreased if the edge distance was less
than 2.5 times the rod diameter. The results of a study by Broughton et al. [10] also confirmed this,
demonstrating how multiple rods spaced too closely do not act individually but pull-out as one plug.
Edge distances are a crucial factor taking pronounced impact on the load bearing capacity because too
small edge distances may cause splitting of the wood [69]. However, there are some differences in the
proposals; more than 2 d [46], more than 2.3 d [43] but values for minimum edge distances of 2.5 d
are present in most design equations (Tab. 1).
According to the provisions in [36] the pull-out strength of a group of GiR must be reduced by a
factor kg for groups of bars (0,8 for 5 or 6 bars in a group, 0,9 for 3 or 4 bars in a group and 1,0 for 1
or 2 bars in a group).
Tab. 1 Edge distances and distances between rods as proposed in different design rules for
connections with rods set parallel to the grain.
Design rule Rods set parallel to the grain: Minimum distances
a1 - between the rods a2 – edge distances
Riberholt [46], Deng [20] 1,5d 2d
prEN 1995-2 [49], CNR DT [55] 4d 2,5d
GIROD [48], DIN 1052:2008 [24] 5d 2,5d
French rules [57] 3d 2,5d
Steiger et al. [43] 4d > 2,3d
New Zealand Timber 2d 1,5d (no shear force)

Design Guide [36]
2,5d
149
Rod spacing and edge distances are key parameters regarding not only the prevention of early splitting
of the connection or of plug failure in case of multiple rod connections but also the overall
performance of a GiR connection in terms of percentage of the load bearing capacity of the timber
gross cross-section transferred by the connection. This means e.g. that distances between rods as well
as edge distances should be fixed taking into account that timber and rod should have similar stiffness
(i.e. ETimber  ATimber  ERod  ARod , which in case of steel rods results in ATimber  16 to 20  ASteel
(see 5.4).
6.5 Technical approvals
Neither an EC design approach nor a product standard (EN) for GiR connections is available so far. In
order to account for the specific features incorporated within different systems of GiR, companies
offering such systems or adhesives for gluing in rods enabled the practical application of their
products/systems by means of technical approvals (TA). Examples include e. g. the GSA ® system
[70], the Purbond PUR adhesive CR 421 [71] and WEVO-Spezialharz EP 32 S /B 22 TS [72]. In
Germany the Studiengemeinschaft Holzleimbau e.V. holds a technical approval [73] containing
general specifications and design rules (referring to the former DIN 1052 standard [24]) for the
application of GiR in practice.
Amongst others, the above mentioned product related TA’s provide detailed information and relevant
data regarding application (service classes, temperatures, type of load), system components (timber,
adhesive, rods) and system design (design loads, rod to rod and rod to edge distances). In general the
determination of the design loads according to the mentioned TA's is based on the German National
Annex to EN 1995 [25] or the preceding standard DIN 1052 [24] respectively (both standards contain
identical design approaches). Hence, the basic design equation is similar to equation (7). As a
consequence, the design can lead to different results compared to the experimentally derived
performance of a connection or reinforcement realised with a particular product or system. The main
reason for this is that basic parameters like characteristic values of pull-out strength and/or required
rod to rod and rod to edge distances can differ from product to product. Therefore it makes sense for
the engineer/user to compare the different products based on the approvals in order to achieve a
maximum performance of a GiR connection or reinforcement.
7. Reinforcement of structural elements with GiR

Key deficiencies of timber in terms of comparably low tensile and compression strength perpendicular
to the grain as well as moderate shear strength can be overcome by strengthening the timber with GiR
in zones subjected to excessive stress. Examples are notched beams or beams with holes, curved or
tapered beams and contact zones / supports with high compression stresses perpendicular to the grain
(Fig. 11). Due to their availability in different length and their high stiffness, GiR are an efficient tool
in strengthening of timber structures. Since, however, their application in practice is quite demanding
(see Section 2), self-tapping screws are often preferred by the designers (see Chapter 9). This in
particular applies for existing structures.
Reinforcing of timber structures is considered an important topic. Hence, in the course of the currently
launched further development of EN 1995, one working group is exclusively dealing with this topic.
Work thereby is based on document CEN/TC 250/ SC 5 N 300 [74] which describes the state-of the
art related to reinforcement of timber structures.
150
(a) (b) (c)
(d) (e) (f)
Fig. 11 Application of GiR to strengthen timber structural elements: zones of high tensile stresses
perpendicular to the grain in curved (a) and tapered beams (b), notched beams (c), beams with holes
(d); zones of excessive shear stresses (d) / compression stresses perpendicular to the grain at supports
(e, f).
7.1 Strengthening in tension perpendicular to the grain

Amongst the earliest applications of GiR to strengthen timber structures were members with excessive
tension stresses perpendicular to the grain (curved and tapered beams, notched beams, beams with
holes) [22, 75, 76]. The GiR reinforcement in these cases will prevent the members from early
cracking (design of new structures) or stop crack propagation and restore initial load bearing capacity
in/of members in existing structures suffering from damages caused by severe cracks [77]. The GiR
reinforcement acts like rebars in concrete. Design rules for GiR applied to strengthen members
perpendicular to the grain can be found in chapter 6.8 of the German National Annex to EN 1995-1-1
[25]. According to these rules, glued-in rods with metric thread as well as glued-in profiled rebars can
be applied. When designing the reinforcement of notches or holes, the perpendicular to the grain
tensile strength is not taken into account, i.e. cracking of the structural member is assumed to have
taken place already [78].
7.2 Strengthening in shear
Beside a general need in higher load bearing capacity, due to their big impact on shear resistance,
already existing cracks or the intent to prevent them may provoke strengthening of beams. From
numerical and experimental studies on shear reinforcement by means of GiR or self-tapping screws
[79-84] it can be concluded that GiR (and self-tapping screws) set under an angle of 45° with respect
to the beam axis provide an efficient mean to increase the shear strength of beams. Beams
strengthened in shear will reach higher load bearing capacities in bending since early shear failures
are prevented. The reinforcing elements in addition contribute to a considerable extent to an increased
flexural stiffness of the beams. So far no design rules are available. Whereas self-tapping screws
provide more ductility and allow for an easy self-setting into the beams, GiR provide high stiffness
but require a higher effort in their application (drilling of holes, centring of rod, gluing).
151
7.3 Zones of concentrated compression forces perpendicular to the grain

If the designer faces the problem of high compression forces to be transferred to the timber element or
from the element to the support, he has to either provide an adequate area of contact (in order to
reduce the compression stresses perpendicular to the grain) or to strengthen the timber locally. This
can be achieved by means of self-tapping screws or GiR both of them acting similar to pile
foundations by transferring the concentrated force via contact pressure and shear stresses along the
rod [22, 85].
7.4 Reinforcement in bending
Some researchers successfully applied rods made from steel or from Fibre Reinforced Polymers
(FRP) to strengthen beams in zones of excessive bending stresses (e.g. [86-92]). Application of this
reinforcement technique in practice may be indicated in the case of decayed tension face of beams or
increased load.
7.5 Moisture induced stresses
When designing reinforcement of timber structures not only the stresses from external loads but also
moisture induced stresses (MIS) should be taken into account [93]. MIS can result from changing
climatic conditions or from drying of beams with MC higher to the one to be expected on site [94,
95].
8. Rods made from Fibre Reinforced Polymers (FRP)

8.1 Background
Fibre Reinforced Polymers (FRPs) are composite materials consisting of load bearing fibres held in a
polymer matrix that protects the fibres and enables load to be transferred between them. Hence, the
strength of an FRP is determined by the strength of the fibrous matrix used. Carbon, glass, aramid or
basalt fibres and a thermosetting or thermoplastic polymer such as epoxy or PFA [96, 97] can be used.
FRPs come in two forms; unidirectional parallel fibres or layered fabrics. Rods are the former, and are
created through a pultrusion process. This is where the fibres are pulled through a resin bath in which
they are impregnated with the polymer; they then enter a heated die with a constant cross-section to
create the required diameter of rod [98].
Fibre Reinforced Polymers have been used in concrete and masonry structures for many years. The
use of FRP in timber dates back to the 1960s where a number of laminated timber structures were
reinforced with Glass Fibre Reinforced Polymer (GFRP). The introduction of Carbon Fibre
Reinforced Polymer (CFRP) and Aramid Fibre Reinforced Polymer (AFRP) in timber construction
[99] first occurred in the 1990’s. In the past two decades much work has been done investigating the
potential of bonded-in FRPs in timber as an alternative to steel rods [15, 99-102].
8.2 Material properties
As Tab. 2 demonstrates, even the weakest FRP is stronger in tension than steel and they are all of
much lower density. Both Basalt Fibre Reinforced Polymer (BFRP) and Glass Fibre Reinforced
Polymer (GFRP) have a much lower modulus of elasticity than steel. Therefore when used in timber
these FRPs should be more compatible with most timbers.
152
Tab. 2 Material properties of bar materials [103-108].

Tensile
Density Yield strength Elastic modulus Cost*
Material strength
(kg/m3) (MPa) (GPa) (Euro/m3)
(MPa)
Steel 7’800 400 – 700 275 - 500 200 6’700
Aramid
1’450 3’000 - 77 - 135 82’000
FRP
Basalt FRP 2’700 1’000 - 90 14’000
Carbon
1’500 1’600 - 120 - 300 90’000
FRP
Glass FRP 1’800 850 - 46 11’500
* Costs are based on 2008 figures and will vary depending on the bar diameter [104, 108].
The higher strength compared with steel rods allows a lesser equivalent volume to be used to achieve
the desired performance. From a cost perspective, both BFRP and GFRP are cost-effective options but
BFRP has a higher tensile strength and slightly better corrosion resistance than equivalent GFRP [14,
104, 106].
8.3 Application and design
In reinforcement or a connection using a bonded-in FRP rod, failure will occur in the timber, close to
the glue-timber interface, as this is the weakest part in the bond, provided a good bond was achieved
in the first place. Adhesives which have good viscosity and gap-filling properties, such as an epoxy or
PFA, should be used to bond FRP rods to timber. The timber should be freshly drilled and cleaned out
and the FRP abraded and wiped down with a solvent or a peel-ply method used to guarantee a good
quality bond.
When designing FRP connections the orientation of fibres in the FRP should be considered. FRPs are
anisotropic materials; they are strong parallel to the direction of their fibres but are weaker
perpendicular to them. Therefore load-carrying components should be designed using FRP orientated
parallel to the load, and connections that may require some flexibility should use fibres perpendicular
to loading.
At present there is no guidance for design using FRP in Eurocode 5 however, the Italian design guides
[109] have information on using FRPs for retrofit and include strengthening in bending, simultaneous
bending and axial force, in-plane actions and connections.
8.4 Advantages and disadvantages
FRP rods have a much higher strength-to-weight ratio than steel rods of equivalent diameter; therefore
they can be used to produce lightweight structures with equal strength. This also makes them easier to
handle and install and reduces transportation costs. FRPs are corrosion resistant and so can be used in
harsh environments such as chloride-rich splash zones where steel would be at risk from corrosion. As
a result of this corrosion resistance, structures using FRPs have a longer service life than when steel is
used, with less monitoring and maintenance required and thus, reduced expenditure where this is
concerned.
153
The cost of using FRPs is higher than steel and this can be a major barrier to their use. As FRPs are
not as readily available as steel their manufacturing process is more costly, leading to an overall
increase in cost of use. The level of expertise and availability of personnel with such experience and
skill is also an issue to be considered. Disposal of waste FRP is another end stage component related
to increased costs; as they cannot be separated in to their original components they are very difficult
to recycle [110]. However, with time and as more experience is gained about using FRPs the cost of
using them should decrease and come in to line with those associated with steel. Tab also
demonstrates that FRPs behave in a brittle fashion whereas steel exhibits ductile behaviour, hence
FRPs not having a yield strength value. However, in cases where a bonded-in rod connection is
designed in such a way that failure occurs due to timber shear, the brittle failure mode of the rod is not
a critical issue.
9. Further development
Rules for the design of GiR connections are desperately needed; be it in the form of an EN 1995
design equation, a product standard or additional technical approvals for established GiR connection
systems. If it should not be possible to develop a widely acceptable design rule for EN 1995 covering
all issues and parameters described in the preceding chapters of this state-of-the-art review, at least an
approach for pre-designing connections/reinforcement with GiR or some general rules (e.g. similar to
SIA 265 [21]) should be given in EN 1995.
One way to untie the “Gordian knot” could be to start not from available experimental data and design
proposals but rather from answering the question: “What are the key advantages and what is the
potential the GiR offers compared to other types of connections/reinforcement and what requirements
have to be fulfilled in order to profit best from these advantages/this potential?”
In the course of the CEN/TC 250/SC5 work programme for the next five years (“towards a 2nd
generation of EN Eurocodes”), glued-in rods are pointed out as an important work item [111] because
they are widely used all over the world. COST Actions FP1004 and FP1101, among others, are
dealing with GiR and hopefully, by the end of the Actions, technical guidelines will be accessible to
designers, industry and scientists, which provide a sound technical background for the preparation of
design rules which will be implemented in the 2nd generation of EN 1995. As minimum expectation it
could be formulated that at least agreement should be reached on how to proceed in general with the
codification of GiR.
10. References
[1] Heymann M., Die Geschichte der Windenergienutzung 1890-1990, Campus Verlag,
Frankfurt/Main, 1995, p. 518.
[2] Riberholt H., Spoer P., Glued-in bolts for the root to hub connection, Nibe-B windmill, Rapport
167 (in Danish), Department of Structural Engineering,Technical University of Denmark,
Lyngby, 1983.
[3] Bernasconi A., Tragverhalten von Holz senkrecht zur Faserrichtung mit unterschiedlicher
Anordnung der Schub- und Biegearmierung, PhD Thesis, ETH Zürich, Professur für
Holztechnologie, Zürich, 1996, p. 155.
[4] Fabris A., Verbesserung der Zugeigenschaften von Bauholz parallel zur Faser mittels Verbund
mit profilierten Stahlstangen, PhD Thesis, ETH Zürich, Professur für Holztechnologie, Zürich,
2001, p. 265.
154
[5] Tlustochowicz G., Serrano E., Steiger R., "State-of-the-art review on timber connections with
glued-in steel rods", Materials and Structures, Vol. 44, No. 5. 2010, pp. 997-1020.
[6] Bainbridge R. J., Mettem C. J., "A review of moment resistant structural timber connections",
Structural Building Engineering, Vol. 128, No. 4. 1998, pp. 323-331.
[7] EN 338, Structural timber - Strength classes, European Committee for Standardization, Brussels,
Belgium, 2009.
[8] Steiger R., "In Brettschichtholz eingeklebte Gewindestangen − Stand des Wissens zu einer
leistungsfähigen Verbindungstechnik", in: Proceedings of the 18. Internationales Holzbauforum,
Garmisch, Deutschland, 2012.
[9] EN 14080, Timber structures - Glued laminated timber and glued solid timber - Requirements,
European Committee for Standardization, Brussels, Belgium, 2013.
[10] Broughton J. G., Hutchinson A. R., "Pull-out behaviour of steel rods bonded into timber",
Materials and Structures, Vol. 34, No. 2. 2001, pp. 100-109.
[11] Kemmsies M., "Comparison of pull-out strengths of 12 adhesives for glued-in rods for timber
structures", SP Report 1999:20, SP Swedish National Testing and Research Institute, Boras,
Sweden, 1999.
[12] Gustafsson P. J., Serrano E., Glued-in rods for timber structures, Report TVSM-3056, Lund
Universtity, Division of Structural Mechanics, Lund, Sweden, 2001.
[13] Aicher S., "Structural adhesive joints including glued-in bolts", In: Thelandersson, S.: Timber
Engineering, Wiley, Chichester, 2003, pp. 333-363.
[14] Harvey K., Ansell M. P., "Improved timber connections using bonded-in GFRP rods", In:
Proceedings of the 6th World Conference on Timber Engineering, Whistler, Canada, 2000.
[15] Ansell M. P., Harvey K., "Improved timber connections using bonded-in GFRP rods", In:
Proceedings of the 6th World Conference on Timber Engineering, Whistler, Canada, 2000.
[16] Bengtsson C., Johansson C.-J., "Glued-in rods - Development of test methods for adhesives", In:
International RILEM Symposium on Joints in Timber Structures, RILEM Publications s.a.r.l.,
Stuttgart, Germany, 2001, pp. 393-402.
[17] Steiger R., Gehri E., Widmann R., "CIB-W18/37-7-8: Glued-in steel rods: A design approach for
axially loaded single rods set parallel to the grain", In: Proceedings of the CIB-W18 Meeting
Thirty-Seven, Edinburgh, United Kingdom, 2004.
[18] Serrano E., Steiger R., Lavisci P., "Glued-in rods", in: Core Document of the COST Action E34
“Bonding of Timber”, University of Natural Resources and Applied Life Sciences, Vienna,
Austria, 2008, pp. 31-39.
[19] Batchelar M., McIntosh K. A., "Structural joints in glulam", In: Proceedings of the 5th World
Conference on Timber Engineering WCTE, Montreux, Switzerland, 1998.
[20] Deng J. X., Strength of epoxy bonded steel connections in glued-laminated timber, Civil
Engineering Research Report 97/4, University of Canterbury, Christchurch, New Zealand, 1997.
[21] SIA 265, Timber Structures, Swiss Society of Engineers and Architects, Zurich, Switzerland,
2012.
[22] Gerold M., "Verbund von Holz und Gewindestangen aus Stahl", Bautechnik, Vol. 69, No. 4.
1992, pp. 167-178.
[23] Meierhofer U., "Schraubenauszugfestigkeit und Tragfähigkeit von Fichten- und Tannenholz",
Holz als Roh- und Werkstoff, Vol. 46, No. 1. 1988, pp. 15-17.
[24] DIN 1052, Entwurf, Berechnung und Bemessung von Holzbauwerken - Allgemeine
Bemessungsregeln und Bemessungsregeln für den Hochbau, Deutsches Institut für Normung
e.V., Berlin, Germany, 2008.
[25] DIN EN 1995-1-1/NA, National Annex, Nationally determined parameters - Eurocode 5: Design
of timber structures - Part 1-1: General - Common rules and rules for buildings,
Normenausschuss Bauwesen (NABau) im DIN, Berlin, Germany, 2010.
[26] Broughton J. G., Hutchinson A. R., "Adhesive systems for structural connections in timber",
International Journal of Adhesion and Adhesives, Vol. 21, No. 3. 2001, pp. 177-186.
155
[27] Carling O., Dimensionering av träkonstruktioner, AB Svensk Byggtjänst och Trätek, Stockholm,
Sweden, 1992, p. 372.
[28] Johansson C.-J., "Lecture C14: Glued-in bolts", In: Timber Engineering STEP 1,
STEP/Eurofortech, Centrum Hout, Almere, The Netherlands, 1995, pp. C14/11-C14/17.
[29] Pedersen M. U., Clorius C. O., Damkilde L., Hoffmeyer P., "Strength of glued-in bolts after full
scale loading", Journal of Performance of Constructed Facilities, Vol. 13, No. 3. 1999, pp. 107-
113.
[30] Volkersen O., "Die Nietkraftverteilung in zugbeanspruchten Nietverbindungen mit konstanten
Laschenquerschnitten", Luftfahrtforschung, Vol. 15, No. Lfg. 1/2. 1938, pp. 41-47.
[31] Volkersen O., "Die Schubkraftverteilung in Leim-, Niet- und Bolzenverbindungen, Teil 1",
Energie und Technik, Vol. 5, No. 3. 1953, pp. 68-71.
[34] Serrano E., Gustafsson P. J., "Fracture mechanics in timber engineering - Strength analyses of
components and joints", Materials and Structures, Vol. 40, No. 1. 2006, pp. 87-96.
[35] Gaunt D., "Joints in glulam using groups of epoxy-grouted steel bars", New Zealand Timber
Design Journal, Vol. 26, No. 1. 1999, pp. 34-38.
[36] NZW 14085 SC, New Zealand Timber Design Guide, Timber Industry Federation Inc.,
Wellington, New Zealand, 2007.
[37] Gehri E., "Ductile behaviour and group effect of glued-in steel rods", In: International RILEM
Symposium on Joints in Timber Structures, RILEM Publications s.a.r.l., Stuttgart, Germany,
2001, pp. 333-342.
[38] Estévez Cimadevila J., Otero Chans D., Martín Gutiérrez E., Vázquez Rodríguez J., "New
anchoring system with adhesive bulbs for steel rod joints in wood", Construction and Building
Materials, Vol. 30, No. 2012, pp. 583-589.
[39] EN 1995-1-1, Design of timber structures, Part 1-1: General - Common rules and rules for
buildings, European Committee for Standardization, Brussels, Belgium, 2004.
[40] Parida G., Johnsson H., Fragiacomo M., "Provisions for ductile behavior of timber-to-steel
connections with multiple glued-in rods", Journal of Structural Engineering, Vol. 139, No. 9.
2013, pp. 1468-1477.
[41] Gehri E., "Klassische Verbindungen neu betrachtet", In: Proceedings of the 17. Dreiländer
Holztagung "Holz: Architecture - Research - Technology", Luzern, 2000.
[42] Pedersen M. B. U., Clorius C. O., Damkilde L., Hoffmeyer P., "The strength of glued-in bolts
after 9 years in situ loading", Journal of Performance of Constructed Facilities, Vol. 13, No. 3.
1999, pp. 102-113.
[43] Steiger R., Gehri E., Widmann R., "Pull-out strength of axially loaded steel rods bonded in
glulam parallel to the grain", Materials and Structures, Vol. 40, No. 1. 2007, pp. 57-68.
[44] Källander B., "CIB-W18/37-7-9: Glued-in rods in load bearing timber structures - Status
regarding European Standards for test procedures", In: Proceedings of the CIB-W18 Meeting
Thirty-Seven, Edinburgh, United Kingdom, 2004.
[45] Larsen H. J., "Essay 4.3: The sad story about glued-in bolts in Eurocode 5", cib-w18.com, 2011.
[46] Riberholt H., "CIB-W18/21-7-2: Glued bolts in glulam - Proposal for CIB code", In:
Proceedings of the CIB-W18 Meeting Twenty-One, Parksville, Vancouver Island, Canada, 1988.
[47] ENV 1995-2, Design of timber structures, Part 2: Bridges, European Committee for
Standardization, Brussels, Belgium, 1997.
[48] Johansson C.-J., Aicher S., Bainbridge R. J., Bengtsson C., Blass H.-J., Görlacher R., Gustafsson
P.-J., Laskewitz B., Mettem C. J., Serrano E., GIROD - Glued in Rods for Timber Structures, SP
Swedish National Testing and Research Institute, Boras, Sweden, 2002.
[49] prEN 1995-2, Design of timber structures, Part 2: Bridges. Final Project Team draft. Stage 34,
European Committee for Standardization, Brussels, Belgium, 2003.
156
[50] CEN/TC 250/SC5 Document N 204, Swiss comments on prEN 1995-2, Final Project Team Draft
(Stage 34), European Committee for Standardization, Brussels, Belgium, 2003.
[51] CEN/TC 250/SC5 Document N 202, UK comments on prEN 1995-2, Final Project Team Draft
(Stage 34), European Committee for Standardization, Brussels, Belgium, 2003.
[52] CEN/TC 250/SC5 Document N 201, Finish comments on prEN 1995-2, Final Project Team
Draft (Stage 34), European Committee for Standardization, Brussels, Belgium, 2003.
[53] Rossignon A., Espion B., "Experimental assessment of the pull-out strength of single rods
bonded in glulam parallel to the grain", Holz als Roh- und Werkstoff, Vol. 66, No. 6. 2008, pp.
419-432.
[54] Stepinac M., Rajcic V., Hunger F., van de Kuilen J.-W., Tomasi R., Serrano E., "CIB-W18/46-7-
10: Comparison of design rules for glued-in rods and design rule proposal for implementation in
European standards", In: Proceedings of the CIB-W18 Meeting Forty-six, Vancouver, Canada,
2013.
[55] CNR-DT 206/2007, Istruzioni per la progettazione, l’esecuzione ed il controllo delle strutture di
legno, Italian National Research Council, Roma, Italia, 2007.
[56] Feligioni L., Lavisci P., Duchanois G., De Ciechi M., Spineli P., "Influence of glue rheology and
joint thickness on the strength of bonded-in rods", Holz als Roh- und Werkstoff, Vol. 61, No. 4.
2003, pp. 281-287.
[57] Faye C., Le Magorou L., Morlier P., Surleau J., "CIB-W18/37-7-10: French data concerning
glued-in rods", In: Proceedings of the CIB-W18 Meeting Thirty-Seven, Edinburgh, United
Kingdom, 2004.
[58] Townsend P. K., Steel dowels epoxy bonded in glue laminated timber, Research Report 90-11,
Department of Civil Engineering, University of Canterbury, New Zealand, Christchurch, New
Zealand, 1990.
[59] Widmann R., Steiger R., Gehri E., "Pull-out strength of axially loaded steel rods bonded in
glulam perpendicular to the grain", Materials and Structures, Vol. 40, No. 8. 2007, pp. 827-839.
[60] Ehlbeck J., "Versuche mit Sondernägeln für den Holzbau", Holz als Roh- und Werkstoff, Vol.
34, No. 7. 1976, pp. 205-211.
[61] Görlacher R., "Untersuchungen an altem Konstruktionsholz: Die Ausziehwiderstandsmessung",
Bauen mit Holz, Vol. 92, No. 12. 1990, pp. 904-908.
[62] Werner H., Siebert W., "Neue Untersuchungen mit Nägeln für den Holzbau", Holz als Roh- und
Werkstoff, Vol. 49, No. 5. 1991, pp. 191-198.
[63] Bengtsson C., Johansson C.-J., "CIB-W18/33-7-8: Test methods for glued-in rods for timber
structures", In: Proceedings of the CIB-W18 Meeting Thirty-Three, Delft, The Netherlands,
2000.
[64] Bernasconi A., "CIB-W18/34-7-6: Behaviour of axially loaded glued-in rods - Requirements and
resistance, especially for spruce timber perpendicular to the grain direction", In: Proceedings of
the CIB-W18 Meeting Thirty-Four, Venice, Italy, 2001.
[65] Otero Chans D., Cimadevila J. E., Gutiérrez E. M., "Glued joints in hardwood timber",
International Journal of Adhesion and Adhesives, Vol. 28, No. 8. 2008, pp. 457-463.
[66] Serrano E., "Glued-in rods for timber structures - An experimental study of softening
behaviour", Materials and Structures, Vol. 34, No. 4. 2001, pp. 228-234.
[67] Blass H. J., Laskewitz B., "Load-carrying capacity of axially loaded rods glued-in perpendicular
to the grain", In: International RILEM Symposium on Joints in Timber Structures, RILEM
Publications s.a.r.l., Stuttgart, Germany, 2001, pp. 363-371.
[68] Blass H. J., Laskewitz B., "CIB-W18/32-7-12: Effect of spacing and edge distance on the axial
strength of glued-in rods", In: Proceedings of the CIB-W18 Meeting Thirty-Two, Graz, Austria,
1999.
[69] Serrano E., "Glued-in rods for timber structures - A 3D model and finite element parameters
studies", International Journal of Adhesion & Adhesives, Vol. 21, No. 2. 2000, pp. 115-127.
[70] Zulassung Z-9.1-778, 2K-EP-Klebstoff GSA-Harz und GSA-Härter für das Einkleben von
Stahlstäben in Holzbaustoffe, Deutsches Institut für Bautechnik, Berlin, Deutschland, 2012.
157
[71] Zulassung Z-9.1-707, 2K-PUR-Klebstoff PURBOND® CR 421 zum Einkleben von Stahlstäben
in Holzbaustoff, Deutsches Institut für Bautechnik, Berlin, Deutschland, 2010.
[72] Zulassung Z-9.1-705, 2K-EP-Klebstoff WEVO-Spezialharz EP 32 S mit WEVO-Härter B 22 TS
zum Einkleben von Stahlstäben in Holzbaustoffen, Deutsches Institut für Bautechnik, Berlin,
Deutschland, 2009.
[73] Zulassung Z-9.1-791, Verbindungen mit faserparallel in Brettschichtholz eingeklebten
Stahlstäben, Deutsches Institut für Bautechnik, Berlin, Deutschland, 2012.
[74] CEN/TC 250/SC5 Document N 300, Report from the working (evolution) group on
Reinforcement of timber structures, European Committee for Standardization, Brussels,
Belgium, 2013.
[75] Gerold M., "Anwendung von in Holz eingebrachten, in Schaftrichtung beanspruchten
Gewindestangen aus Stahl", Bautechnik, Vol. 70, No. 10. 1993, pp. 603-613.
[76] Möhler K., Hemmer K., "Eingeleimte Gewindestangen", Bauen mit Holz, Vol. 83, No. 5. 1981,
pp. 296-298.
[77] Gehri E., "Krafteinleitung mittels Stahlanker", In: Proceedings of the 28. SAH-Fortbildungskurs
"Brettschichtholz", Weinfelden, Schweiz, 1996.
[78] Blass H. J., Ehlbeck J., Kreuzinger H., Steck G., Erläuterungen zu DIN 1052: 2004-08 Entwurf,
Berechnung und Bemessung von Holzbauwerken, Bruderverlag, Karlsruhe, 2004, p. 217.
[79] Blass H. J., Krüger O., "Schubverstärkung von Holz mit Holzschrauben und Gewindestangen",
Karlsruher Berichte zum Ingenieur-Holzbau, Universitätsverlag Karlsruhe, 2010.
[80] Dietsch P., Kreuzinger H., Winter S., "CIB-W18/46-7-9: Design of shear reinforcement for
timber beams", In: Proceedings of the CIB-W18 Meeting Forty-six, Vancouver, Canada, 2013.
[81] Dietsch P., Mestek P., Winter S., "Analytischer Ansatz zur Erfassung von
Tragfahigkeitssteigerungen infolge von Schubverstarkungen in Bauteilen aus Brettschichtholz
und Brettsperrholz", Bautechnik, Vol. 89, No. 6. 2012, pp. 402-414.
[82] Koj C., Trautz M., "Self-tapping screws as connectors and reinforcements - long-term tests on
rigid frame corners in exterior climatic conditions", Bautechnik, Vol. 91, No. 1. 2014, pp. 40-47.
[83] Trautz M., Koj C., "Self-tapping screws as reinforcement", Bautechnik, Vol. 85, No. 3. 2008, pp.
190-196.
[84] Trautz M., Koj C., "Self-tapping screws as reinforcement - new results.", Bautechnik, Vol. 86,
No. 4. 2009, pp. 228-238.
[85] Gehri E., "Eingeklebte Anker – Anforderungen und Umsetzungen", In: Proceedings of the 15.
Internationales Holzbauforum 09, Garmisch Partenkirchen, 2009.
[86] Alhayek H., Svecova D., "Flexural Stiffness and Strength of GFRP-Reinforced Timber Beams",
Journal of Composites for Construction, Vol. 16, No. 3. 2012, pp. 245-252.
[87] Gentile C., Svecova D., Saltzberg W., Rizkalla S. H., "Flexural strengthening of timber beams
using GFRP", Advanced Composite Materials in Bridges and Structures, Vol. No. 2000, pp.
637-644.
[88] Gentile C., Svecova D., Rizkalla S. H., "Timber beams strengthened with GFRP bars:
Development and applications", Journal of Composites for Construction, Vol. 6, No. 1. 2002, pp.
11-20.
[89] Svecova D., Eden R. J., "Flexural and shear strengthening of timber beams using glass fibre
reinforced polymer bars - an experimental investigation", Canadian Journal of Civil Engineering,
Vol. 31, No. 1. 2004, pp. 45-55.
[90] Jobin J., Garzon Barragan O. L., Flexural strengthening of glued laminated timber beams with
steel and carbon fiber reinforced polymers, Masters Thesis, Chalmers University, Gothenburg,
Sweden, 2007, p. 164.
[91] Alshurafa S. A., Alhayek H., Alshorafa M., "Strength evaluation of long douglas fir stringers
reinforced with GFRP rods", International Journal of Civil and Structural Engineering, Vol. 3,
No. 3. 2013, pp. 613-620.
158
[92] Raftery G., Whelan C., Harte A., "Bonded-in GFRP rods for the repair of glued laminated
timber", In: Proceedings of the 12th World Conference on Timber Engineering, Auckland, New
Zealand, 2012.
[93] Ranta-Maunus A., Gowda S. S., Curved and cambered glulam beams - Part 2: Long term tests
under cyclically varying humidity, Report 177, VTT, Espoo, Finland, 1994.
[94] Dietsch P., Kreuzinger H., Winter S., "Effects of changes in moisture content in reinforced
glulam beams", In: Proceedings of the 13th World Conference on Timber Engineering, Quebec,
Canada, 2014.
[95] Wallner B., Versuchstechnische Evaluierung feuchteinduzierter Kräfte in Brettschichtholz
verursacht durch das Einbringen von Schraubstangen, PhD Thesis, Graz University of
Technology, Graz, 2012, p. 154.
[96] Patinak A., "Applications of Basalt Fibre Reinforced (BFRP) reinforcement for transportaion
infrastructure", In: Proceedings of the Transportation Research Board Conference "Developing a
research agenda for transportation infrastructure preservation and renewal", Washington D.C.,
2009.
[97] Van de Velde V. K., Kiekens P., Van Langenhove L., "Basalt fibres as reinforcement for
composites", In: Proceedings of the 10th International Conference on Composites / Nano
Engineering, New Orleans, LA, 2003.
[98] Astrom B. T., Manufacturing of polymer composites, CRC Press, London, 2002, p. 469.
[99] Micelli F., Scialpi V., La Tegola A., "Flexural reinforcement of glulam timber beams and joints
with carbon fiber-reinforced polymer rods", Journal of Composites for Construction, Vol. 9, No.
4. 2005, pp. 337-347.
[100] Borri A., Corradi M., Grazini A., "A method for flexural reinforcement of old wood beams with
CFRP materials", Composites Part B-Engineering Vol. 36, No. 2. 2005, pp. 143-153.
[101] Madhoushi M., Ansell M. P., "Flexural fatigue of beam to beam connections using glued-in
GFRP rods", In: Proceedings of the 8th World Conference on Timber Engineering, Lahti,
Finland, 2004.
[102] Yeboah D., "Rigid connections in structural timber assemblies", PhD Thesis, Queen's University,
Belfast, UK, 2012, p.
[103] Aslan FRP, Aslan 100 Glass Fiber Reinforced Polymer (GFRP) Rebar: Product Data Sheet,
2011.
[104] Balafas I., Burgoyne C. J., "Economic design of beams with FRP rebar or prestress", Magazine
of Concrete Research Vol. 64, No. 10. 2012, pp. 885-898.
[105] Linear Composites, Parafil: The Ultimate Synthetic Rope, 2013.
[106] Magmatech, RockBar corrosion resistant basalt fibre reinforcing bars, 2013.
[107] Mettem C. J., Bainbridge R. J., Harvey K., Ansell M. P., Broughton J. G., Hutchinson A. R.,
"CIB-W18/32-7-13: Evaluation of material combinations for bonded in rods to achieve improved
timber connections", In: Proceedings of the CIB-W18 Meeting Thirty-Two, Graz, Austria, 1999.
[108] Williams B. M., Personal Communication, 2014.
[109] CNR-DT 201/2005, Guidelines for the design and construction of externally bonded FRP
systems for strengthening existing timber structures, Italian National Research Council, Rome,
Italy, 2005.
[110] Smallman R., Bishop R. J., Modern physical metallurgy and materials engineering science,
process, applications, Butterworth-Heinemann, Oxford, 1999, p. 438.
[111] Dietsch P., Winter S., "Eurocode 5 - Future Developments towards a more comprehensive code
on timber structures", Structural Engineering International, Vol. 22, No. 2. 2012, pp. 223-231.
159
160
Reinforcement with self-tapping screws and threaded rods
9
Philipp Dietsch1, Reinhard Brandner2
Summary
In timber engineering, self-tapping screws, optimized primarily for axial loading, represent the state-of-
the-art in fastener and reinforcement technology. Their economic advantages and comparatively easy
handling make them one of the first choices for application in both domains. This paper focuses on
self-tapping screws and threaded rods applied as reinforcement, illustrating the state-of-the-art in
application and design approaches in Europe, in conjunction with numerous references for background
information. With regard to medium to large span timber structures which are predominately erected
by using linear timber members, from e.g. glued laminated timber, the focus of this paper is on their
reinforcement against stresses perpendicular to grain as well as shear. However, latest findings with
respect to cross laminated timber are included as well.
1. Introduction
Wood is a highly anisotropic material, featuring low capacities in tension and compression
perpendicular to the grain as well as shear. When designing structural elements from timber, it should
be aimed at minimising e.g. tensile and compressive stresses perpendicular to the grain. If this cannot
be fully achieved, the timber element can be reinforced to compensate for these low strength
properties. The list of applicable internal or external reinforcement is – amongst other factors – based
on the necessity of a continuous interconnection between the timber and the reinforcement as well as
sufficient stiffness of this connection (to prevent cracking). Fully threaded, self-tapping screws are a
very economical alternative to the traditional reinforcement by glued-in steel rods or wood-based
panels which are glued onto the timber member. After a brief introduction to wood screws in general,
we focus on the state-of-the-art of fully-threaded self-tapping screws and threaded rods and associated
design approaches common in Europe. In principle, the approaches presented in the following also
apply to glued-in rods. For an in-depth overview on glued-in rods, the interested reader is kindly
referred to the corresponding Chapter in this book. [1]
1)
Dr.-Ing., Team Leader Timber Structures, Chair of Timber Structures and Building Construction Technische
Universität München, Germany
2)
Ass.-Prof. Dr. techn.,Institute of Timber Engineering and Wood Technology Graz University of Technology,
Austria
161
Traditional wood screws feature a shank with a threaded and a smooth part, the outer diameter of the
thread generally equaling the smooth shank diameter. Such screws are described in national standards,
like DIN 7998 [2]. Typically, the length of the threaded part is only 60 % of the total length. For shank
diameters d > 6 mm, pre-drilling is required (depending on e.g. density of the timber). Their axial load-
carrying capacity is mostly limited by the pull-through capacity of the screw head. Thus the additional
use of adequately designed washers is commonly meaningful. The combination of threaded and
smooth shank and the use of mild- or low carbon steel allow using them for loading in shear and
tension or a combination of both.
Constraints in the load-carrying capacity of primarily axially-loaded screws were overcome with the
development of self-tapping screws. In contrast to traditional screws which have their threaded part
turned down from the original rod diameter, the thread of self-tapping screws is produced by rolling or
forging a wire rod around the shank, which consequently features a smaller diameter when compared
to the outer cross sectional thread diameter (see also Fig. 1). Self-tapping screws mostly feature a
continuous thread over the whole length. This leads to a more uniform load transfer between the screw
and the wood material as well as a considerably enhanced axial load-carrying capacity, the type of
loading for which they are optimized.
During manufacture, their thread is hardened, leading to an increased bending and torsion capacity,
but also to a more brittle failure mechanism. In combination with the development of optimized drill
tips and threads, self-tapping screws, featuring diameters up to 14 mm and lengths up to 1000 mm, are
produced and applied today, see Fig. 1 and Fig. 2.
Fig. 1 Different forms of drill tips and threads (with and without shank cutter) as well as variations of
screw heads; comparison with DIN-screw (third from right), from [5]
Fig. 2 Development of screw length (and implicitly load-carrying capacity) from traditional wood
screws to self-tapping screws, from [6]
162
An extension of these geometric limits is possible by the application of threaded rods. These are a
modification of self-tapping, fully threaded screws featuring screw threads over the full length. These
can reach diameters of up to 20 mm and lengths of up to 3000 mm. Threaded rods with screw threads
need pre-drilling with the core diameter and a coating and/or lubricant to reduce friction stresses when
driving them in. In dependency of the length of the screw or threaded rod, their axial load-carrying
capacity may be limited by the tensile capacity of the steel or, if loaded in compression, by buckling.
Following EN 1995-1-1 [3] and investigations made in e.g. [4], a group of axially loaded screws may
also fail in block shear.
Requirements for self-tapping screws are for example given in EN 14592 [7]. Here, the nominal
diameter equals the thread diameter d which has to be 2.4 mm ≤ d ≤ 24 mm (practical range: 8 mm,
10 mm and 12 mm for screws, 16 mm and 20 mm for threaded rods). The core diameter d1 has to be
0.6 ∙ d ≤ d1 ≤ 0.9 ∙ d (0.6 ∙ d ≤ d1 ≤ 0.75 ∙ d according to EN 1995-1-1 [3]). The minimum thread
length is restricted to lg ≥ 4 ∙ d (minimum embedment depth or effective penetration (anchoring)
length, lef, according to EN 1995-1-1 [3] is: for axially loaded screws 6 ∙ d and for laterally loaded
screws 4 ∙ d,), see also Fig. 3.
l
lg d … nominal (thread) diameter
d1 … core diameter
dh … head diameter
d1
d
l … length
dh
lg … thread length
Fig. 3 Geometry parameters for screws
The mechanical properties are defined by the characteristic values of (i) the yield moment, (ii) the
withdrawal capacity of the threaded part, (iii) the tear-off capacity of the screw head, (iv) the pull-
through capacity of the screw head, (v) the tensile strength of the screw, and (vi) the torsional strength
of the screw,. The mechanical properties can be derived from tests or from equations given in e.g.
EN 1995-1-1 [3]. The use of other screws as specified in EN 14592 [7] is allowed, provided their
applicability is proven by a technical approval. These used to be issued by national building authorities;
currently most national approvals are being converted into European Technical Approvals (ETAs).
Between the products available on the market, there are a variety of head and thread forms and
differences in shank-, tip- and thread-diameter as well as different ratios between thread to core
diameter, see Fig. 1. However, a comprehensive comparative study on self-tapping screws of five
different manufactures, tested at angles between screw axis and fibre direction of 90°, 45° and 0°,
showed only minor differences in the withdrawal capacities. The range of ± 10 % is in-line with the
technical approvals. More relevant are the differences in the practical application of the screws, see e.g.
[5].
Meanwhile, many regulations in technical approvals have been adapted between different approvals,
facilitating design and comparability. Nevertheless, there are a number of approvals in which special
rules are given that must be followed. For example, there are different rules in the following areas:
 minimum spacings and distances, minimum member thickness requirements;
 axial withdrawal capacities (as a function of characteristic (5 %-quantile) density);
 permissible angles between screw axis and grain direction;
163
 wood species (mostly softwoods);

 tensile capacity;
 stability of the screws, i.e. buckling failure when loaded in compression;
 stiffness values (Kser, Kax,ser, Ku).
Because of these differences, specified screws shall not be substituted by other screws.
2. General rules on the application of self-tapping screws

The European basis for the design of self-tapping screws is the design concept given in Eurocode 5
(EN 1995-1-1 [3]) in combination with the provisions given in the European Technical Approvals
(ETAs). For applications which are not covered by Eurocode 5, design approaches can be given in the
National Annexes to Eurocode 5 (as non-contradictory, complementary information (NCCI)). Some
ETAs, e.g. [8], [9] and [10], also feature annexes containing design provisions for certain
applications.
One advantage of self-tapping screws is that they do not require pre-drilling, given the density of the
timber is not too high (e.g. ρ < 500 kg/m3, given for most softwoods). However, research results in
[11] indicate, that for the application of self-tapping screws in timber with temperature below zero,
pre-drilling may be required to prevent splitting of the timber which becomes more brittle at these
temperatures. In more recent technical approvals, pre-drilling is allowed. Here, the borehole diameter
shall not be greater than the core diameter of the screw. Pre-drilling can have a positive effect on the
precision of the screw positioning. This is of special importance if the screws are positioned at an
angle to the grain, which is challenging if carried out free-handed, especially if screws of small
diameter (e.g. d = 6 mm) and high slenderness are used. For this, placement devices or CNC
machinery can be used. Some ETAs allow to reduce the minimum spacing if pre-drilling is applied.
Most self-tapping screws feature a special drill-tip which allows the use of reduced spacings and
distances even without pre-drilling. For such screws, most technical approvals (e.g. [8], [9] and [10])
contain the spacing and distance requirements for axially loaded screws given in Tab. 1 (see also Fig.
4). In the case of screws positioned at an angle between grain and screw axis, the centroid of the
threaded part of the screw in the respective timber member shall be used as reference for determining
spacings and distances (see Fig. 4 and e.g. [8], [9], [10] and [12]). In the case of reinforcement against
tensile stresses perpendicular to the grain, the distance between the reinforcement and the (mostly
localized) area of stress peaks should be minimized, e.g. by using the minimum of a3,c (see e.g. [12]).
Tab. 1 Typical minimum spacings and minimum distance

requirements given in technical approvals for axially
loaded self-tapping screws
a1 a2 a3,c a4,c
5∙d a 5∙d 4∙d b
5∙d a
a
may be reduced to 2.5∙d, if the condition a1·a2 ≥ 25·d² is fulfilled.
b
in some cases, 3∙d is allowed.
164
Fig. 4 Definition of spacings, end and edge distances for axially loaded screws
According to Eurocode 5 (EN 1995-1-1 [3]), a reduction in cross-section, caused by screws do not
need to be considered in cases of d ≤ 6 mm (without pre-drilling) and for screws placed in areas under
compressive stresses. For nominal diameters d > 6 mm, a reduction of the cross-section shall be taken
into account. This is typically realized by considering only the cross-sectional area of the screws, cut
at an angle α at the timber cross section featuring the highest number of penetrating screws (e.g. [12]).
However, tension parallel to grain tests on butt-joints with inclined fully threaded self-tapping screws
showed, that the net cross section is better approximated by the timber area between the projected
screws, see Fig. 5, right. Thus, the sole consideration of the screw holes as loss in net cross section
was shown to be not sufficient [13]. Recent research results [14] indicate that a reduction of the cross-
section should also be considered in members under compression, even if the holes are filled with a
material of higher stiffness than the wood. This is explained by stress peaks developing in the vicinity
of the fasteners. The exception is glued-in rods since the glue-line shall lead to a reduction in stress
peaks.
Fig. 5 Butt-joint with self-tapping screws and proposal for definition of net cross section, from [13]
The characteristic axial load-carrying capacity of the screw, Fax,Rk, is determined as the minimum of
the characteristic tensile capacity, Ft,Rk, (or characteristic buckling capacity, Fki,Rk) and the
characteristic withdrawal capacity, Fax,α,Rk, of the screws, see Eq. (1). The characteristic withdrawal
capacity, Fax,α,Rk, is calculated as the product of the characteristic withdrawal parameter, fax,k, (given in
the technical approval; conforming with the characteristic withrawal strength multiplied by π = 3.14),
the nominal diameter, d, and the effective anchorage length, ℓef. In the case that screws are positioned
at an angle α between grain and screw axis, most approvals contain an equation with which the
withdrawal capacity can be determined according to the values for α = 90°. Some technical approvals
allow to use the values determined for α = 90° in case of angles α ≥ 45°.
165
 Ft , Rk or Fki , Rk 
Fax, Rk  min   (1)
 f ax,k  d  lef 
The slip moduli, Kser (slip moduli in shear) and Kax,ser (axial slip moduli), are needed e.g. in the case of
mechanically jointed (e.g. doubled) beams which are designed with the γ-method or shear analogy
(see chapter 6). Experiments to determine the axial slip modulus, Kax,ser, show large variations in
results (e.g. [15] and [16] contain values approximately twice as high as in [17]). This variety of
results is also observed in different technical approvals, which report different values of the slip
modulus. Eq. (2), which is based on tests on screws from different manufacturers featuring maximum
penetration lengths of 120 mm [17]), is generally reported in many technical approvals. Other
technical approvals report different equations delivering higher results.
K ax, ser  780  d 0.2  lef

0.4
(2)
Test results reported in [15] and [18] show that the slip moduli are also significantly dependent on the
angle α between screw axis and grain direction. In contrast to the withdrawal capacity the slip moduli
increase with decreasing angle α. In [18], tests on screwed-in rods featuring penetration lengths of
200 mm and 400 mm are reported that indicate a disproportionate (above-average) increase of the
axial stiffness when doubling the penetration length.
3. Self-tapping screws as reinforcement to carry tensile stresses

perpendicular to the grain
Screws under axial loading are very stiff, favoring their use as reinforcing elements. In the following
subsections, typical applications of self-tapping screws as reinforcement to carry tensile stresses
perpendicular to the grain in timber structures are presented. Within the approaches presented, the
tensile capacity perpendicular to the grain of the timber is neglected, i.e. a cracked tension zone is
assumed. This is different from to the approach taken by [19] in which only the force components,
exceeding the tensile strength perpendicular to the grain of the timber, are considered for the design of
the reinforcement.
3.1 Reinforcement of connections with a tensile force component perpendicular to the
grain
The approach to design the reinforcement of connections with a tensile force component
perpendicular to the grain explained in the following is standardized [20], previous sources include
[21] and [22]. The approach is explained in [23] as well as [24], previous works include [25].
The design tensile force perpendicular to the grain, Ft,90,d, is the resultant of the tensile stresses
perpendicular to the grain on the plane defined by the loaded edge distance to the center of the most
distant fastener, he, (see e.g. [26]). According to beam theory, the connection force component
perpendicular to the grain results in a step in the shear force distribution. The tensile force
perpendicular to the grain, Ft,90,d, see Eq. (3), is determined from this change in shear stress by
integration of the shear stress in the area between the row of fasteners considered and the unloaded
edge, as indicated by the shaded area in Fig. 6 right. A derivation of this approach can be found in e.g.
[27].
166
 
Ft ,90,d  1  3   2  2   3  Fv ,Ed (3)
with :   he h
Fv ,Ed design value of the shear force component perpendicular to the grain
The design approach for reinforcement of connections with a tensile force component perpendicular
to the grain can also be translated to the reinforcement of girder hangers or the mortise part of a
dovetail connection.
Fig. 6 Reinforced cross-connection: reinforcement (left) and distribution of shear stresses and shear
flow (right, see also [27])
The highest tensile stresses perpendicular to the grain occur in direct vicinity of the fasteners.
Therefore, the distance between the screw and the fasteners, a1, should be minimal. It is ideal to place
the reinforcement (also) in between the fasteners, as shown in Fig. 6 left. With respect to further
geometric considerations for the reinforcement, the specifications given in 0 apply. The load-carrying
capacity of the screw is determined in dependence of the length of the screw between the row of
fasteners considered and the unloaded edge, ℓad,t. The thread of the reinforcement should at least cover
75 % of the beam height, see e.g. [28]. In all other cases, the tensile stresses perpendicular to the grain
at the screw tip have to be verified as well.
3.2 Reinforcement of notched members
The approach to design the reinforcement of notches in members with rectangular cross-section
explained in the following is standardized [20], previous sources include [21] and [22]. The approach
is explained in [23], [24] as well as [12]. Preceding works include e.g. [25]. The tensile force
perpendicular to the grain, Ft,90,d, can be approximated by integration of the shear stress in the area
between the inner and outer corner of the notch (see shaded area in Fig. 7 right). A derivation of this
approach can be found in [27].
A more detailed analysis of the magnitude of the tensile stresses perpendicular to the grain around the
notch, using plate theory, has shown that these stresses are increased due to the eccentricity between
the support and the inner edge of the notch ([29], [30]). For relationships x ≤ hef / 3 (see Fig. 7), the
tensile force perpendicular to the grain, Ft,90,d, can be sufficiently estimated by applying an increase
factor of 1.3.
The design tensile force, Ft,90,d, to be carried by the reinforcement, can be determined according to Eq. (4).
167
Ft ,90,d  1.3 Vd  [3  1     2  1    ], for x  hef 3

2 3
(4)
with:   hef h
Vd design value of the shear force
The design approach for notched members can also be used for the reinforcement of the tenon part of
a dovetail connection, as e.g. applied by [31].
Due to the limited distribution length of the tensile stresses perpendicular to the grain outside the
corner of the notch, the distance between the screw and the notch, a3,c, should be minimized, see Fig.
7 left. Only one row of screws at a distance a3,c should be considered, hence it should be aimed at
placing all necessary screws in the first row, utilizing – if necessary – the minimum possible distance
a2 between the screws (see Fig. 4).
The load-carrying capacity of the screw is determined in dependence of the smaller of both anchorage
lengths, with ℓef = ℓad.
Fig. 7 Notched beam: reinforcement (left) and distribution of shear stresses (right)
Since end-grain is exposed and commonly not sealed at a notch, the superposition of moisture induced
stresses and load-dependent tensile stresses perpendicular to the grain around holes can be significant
[32]. Therefore, some authors recommend that notched members should always be reinforced, see e.g.
[12], [33].
Experiments on notched members reinforced with screws have shown the potential of a crack
developing from the corner of the notch. The crack development ends just after it has crossed the
screw. The reason for this can be seen in the very small deformation capacity of timber before
exceeding the tensile strength perpendicular to grain. Although it could be considered a visual
deficiency, such a crack is not a sign of reduced load-carrying capacity of the notched beam.
Recent research [34] indicates that Eq. (4) might not be conservative for all cases since it does not
account for shear failure of the reinforced notch. The authors state that a combination of
reinforcement to cover both shear and tensile stresses perpendicular to the grain, has a positive effect
on the structural behavior of reinforced notched beams.
Recent developments also include self-tapping screws with changing angle of the thread along the
screw. Such screws can be used to induce compressive stresses perpendicular to the grain, thereby
reducing the effect of load-dependent or moisture induced tensile stresses perpendicular to the grain
[35]. The maximum acceptable displacement between the thread and the wood material is limited,
implying that this approach could be most adequate for details with localized areas of high tensile
stresses perpendicular to the grain. So far, the influence of relaxation on the long-term occurrence of
the imposed compressive stresses perpendicular to the grain could not be clarified.
168
3.3 Reinforcement of members with holes

The approach to design the reinforcement of members with holes explained in the following is
standardized [20], previous sources include [21] and [22]. The approach is explained in [23] as well as
[24], research on this approach is presented in [36]. Modifications to the approach are explained in
[37] and [38]. Modifications constitute the limitation of the permissible relative dimensions of the
hole in dependency of the type of reinforcement (e.g. hd ≤ 0.3 ∙ h in case of reinforcement with
screws). Information on the behavior of holes in timber beams in general can be found in e.g. [39],
[40], [41], [42] and [43].
The tensile force perpendicular to the grain, Ft,90,V,d, see Eq. (5), can be approximated by integration of
the shear stress between the axis of the member and the corner of the hole under tensile stresses
perpendicular to the grain as indicated by the shaded area in Fig. 8 right. A derivation of this approach
can be found in [27]. Experimental results given in [40] also show a tensile failure perpendicular to
the grain at holes in areas without shear. From the experimental results, a tensile force component due
to bending moment, Ft,90,M,d, was determined which has to be added to the force component due to
shear. In the case of circular holes, the parameter hr may be increased by 0.15 ∙ hd to account for the
fact that potential fracture will occur at a position described by an angle of 45° from the center line of
the hole. The applicable effective anchorage length, ℓef, is equal to the applicable distance hr from the
edge of hole to the upper/lower edge of the member, given the length of the screw after the potential
failure plane exceeds hr.
Vd  hd  h 2 M
Ft ,90, d  Ft ,90,V , d  Ft ,90, M , d   3  d2   0.008  d
4h  h  hr
with: Vd design value of the shear force at the edge of the hole
M d design value of the bending moment at the edge of the hole
a hole length
hd hole depth (5)
h rl distance from lower edge of hole to bottom of member
h ru distance from upper edge of hole to top of member
 min hro ; hru  for rectangular holes
hr = 
 min hro  0,15  hd ; hru  0,15  hd  for round holes
lA support distance of a hole
Fig. 8 Beam with hole: reinforcement (left) and distribution of shear stresses (right)
The superposition of moisture induced stresses and load-dependent tensile stresses perpendicular to
the grain around holes can be significant [32], in particular if the end-grain in the holes is not sealed.
Therefore, some authors recommend that members with holes should always be reinforced, see e.g.
[33]. In most cases, the necessity to reinforce holes will be given by the geometrical boundary
169
conditions for unreinforced holes (hd ≤ 0.15 ∙ h). With respect to geometric considerations for the
reinforcement, e.g. distances, the specifications given in 0 apply.
It shall be emphasized that besides the reinforcement with screws also a verification of the shear
strength of the timber in the vicinity of the hole is required. The distribution of the shear stresses in
the vicinity of the hole deviates considerably (polynomial stress distribution instead of parabolic stress
distribution), its maxima can reach significantly higher values compared to the values determined
according to beam theory. A description as well as an associated design equation is given in [24], see
Eq. (6). In [44] it is recommended to apply the same verification for round holes as well. In the same
publication, a method is described to verify the bending stresses above or below rectangular holes,
including the additional longitudinal stresses from the frame action (lever of the shear force) around
the hole (see also [40]).
0.2
1,5 Vd  a  h  1,5 Vd
max   max   1,84  1     d  
b   h  hd   h   h b   h  hd 
with:  max factor to take account for the increased shear stresses in the area of the edge of the hole (6)
description of terms see Eq. (5)
h d may be replaced by 0.7  h d in case of round holes
3.4 Reinforcement of double tapered, curved and pitched cambered beams

Double tapered, curved and pitched cambered beams mostly feature beam depths which exceed the
maximum length of self-tapping screws. An alternative reinforcement is given by threaded rods which
are produced in lengths of up to 3000 mm and installed by aid of pre-drilling with the core diameter.
Threaded rods featuring screw threads constitute a modification of self-tapping, fully threaded screws,
i.e. the main specifications given above apply.
A standardized approach to design the reinforcement of curved and pitched cambered beams is given
in [20], previous sources include [21] and [22]. The approach is explained in [23]. It is differentiated
between reinforcement to carry the full tensile stresses perpendicular to the grain and reinforcement to
only carry the tensile stresses perpendicular to the grain from climatic conditions, i.e. the moisture
induced tensile stresses perpendicular to the grain.
The first approach is based on an integration of the sum of tensile stresses perpendicular to the grain
in the plane of zero longitudinal stresses. The equations given in [3] (based on [45]) only provide the
maximum tensile stresses perpendicular to the grain in the apex. Depending on the form and loading
of the beam, the tensile stresses perpendicular to the grain decrease with increasing distance from the
apex. They even spread to some extent into the straight parts of the beam, see Fig 9 and [46]. In [47]
and [48], these results were verified and extended to beams with mechanically fixed apex, see Fig 9
below right. For simplification, reinforcement in the inner quarters of the area exposed to tensile
stresses perpendicular to the grain is designed for the full tensile stresses perpendicular to the grain,
see Eq. (7).
t ,90,d  b  a1
Ft ,90,d  (7)
n
170
with: Ft ,90,d design tensile force perpendicular to grain in the reinforcement

t ,90,d design tensile stress perpendicular to grain
b beam width, in [mm]
a1 distance between fasteners in grain direction
n number of reinforcing fasteners within a1
Fig. 9 Distribution of tensile stresses perpendicular to the grain over the beam height and length in
double tapered (above left), curved (above right), pitched cambered (below left) and cuved beam with
mechanically fixed apex, i.e. secondary apexes (below right)
In the outer quarters, the tensile stresses perpendicular to the grain are assumed to reach 2/3 of the
maximum tensile stresses perpendicular to the grain determined with Eq. (7), see Fig 9.. The spacing
between the reinforcement is limited to a1 ≤ 0.75 ∙ hap (hap as height of the apex) to ensure that the
whole area exposed to tensile stresses perpendicular to the grain is covered by reinforcement.
Even if the requirements regarding systematic, load-dependent tensile stresses perpendicular to the
grain can be met, see [3], it is state of the art to reinforce double tapered, curved and pitched
cambered beams. Reason is the superposition of the load-dependent stresses with moisture induced
stresses perpendicular to the grain due to e.g. changing climatic conditions or a drying of the beam
after the opening of the building, see e.g. [49], [50]. In the lack of a method to reliably predict the
magnitude of tensile stresses perpendicular to the grain, it is custom to apply reinforcement if the
maximum load-dependent tensile stresses perpendicular to the grain exceeded 60 % of the design
tensile strength of the timber member perpendicular to the grain. Some authors recommend that
double tapered, curved and pitched cambered beams are always reinforced, independent from the
magnitude of tensile stresses perpendicular to the grain, see e.g. [33]. Eq. (8) represents one approach
to design reinforcements to carry moisture induced tensile stresses perpendicular to the grain ([20],
[21] and [44]). It is based on the assumption that 1/4 of the tensile stresses perpendicular to the grain
from external loads are carried by the reinforcement. It is also based on the assumption that the
potential magnitude of moisture induced stresses increases with increasing member width, i.e.
increases due to decelerated adaption of timber moisture content in the interior of the cross-section.
For a member width b = 160 mm, Eq. (8) lends ¼ of the full reinforcement determined by Eq. (7), less
reinforcement for smaller member widths and more reinforcement for larger member widths.
171
t ,90,d  b2  a1
Ft ,90,d 
640  n
with: Ft ,90,d design tensile force perpendicular to grain in the reinforcement
(8)
t ,90,d design tensile stress perpendicular to grain
a1 distance between fasteners in grain direction
n number of reinforcing fasteners within a1
In the case of reinforcement to carry the tensile stresses perpendicular to the grain from climatic
conditions, the spacing between the reinforcement should be kept constant. It is recommended that the
spacing is limited to the member depth [44].
Recent research has begun to examine the question of the influence of reinforcement on the
magnitude of moisture induced stresses since reinforcement restricts the free shrinkage or swelling of
the timber beam. Experimental studies (short-term tests) and analytical considerations, presented in
[51] and [52], indicate that a reduction of timber moisture content of 3 % to 4 % around threaded rods,
positioned perpendicular to the grain, can lead to critical stresses with respect to moisture induced
cracks. In addition, a substantial mutual influence of adjacent reinforcing elements could be
identified. Experiments on drying glulam members reinforced by screwed-in threaded rods placed
perpendicular to the grain are presented in [53]. The results indicate that moisture induced stresses in
the timber can lead to forces in the reinforcement in the order of the steel capacity of the rods.
3.5 Reinforcement of connections with a tensile force component parallel to the grain
The load-carrying capacity of connections with multiple fasteners in one row parallel to the grain can
be lower than the sum of the load-carrying capacities of the single fasteners (depending on the
fastener distances parallel to the grain). This is due to the splitting forces (tensile stresses
perpendicular to the grain) induced by the fasteners. The tendency to splitting increases with
decreasing spacing of the fasteners parallel to the grain, a1 (Fig. 10). In the codes, this is accounted for
by an effective number of fasteners, nef, [3] (based on [54]). Placing self-tapping screws with
continuous thread perpendicular to the fastener axis and to the grain direction may prevent splitting,
i.e. nef = n may be used, see Fig. 10. The closer the screw is placed to the dowel-type fastener, the
better the effect. Since the splitting force is highest close to the joint between two connected
members, the screws should be positioned with a minimum edge distance, a4,c, see Fig. 10 right. In
[44] it is recommended to design the self-tapping screws for an axial force of 30 % of the load
transferred by each dowel-type fastener and shear plane.
Fig. 10 Reinforcement of dowel-type connections

According to [55], [56] as well as [57], the load-carrying capacity of dowel-type connections,
reinforced with self-tapping screws with continuous thread can be further increased, if the screw is in
direct contact with the dowel-type fastener. This necessitates very exact positioning which is not
always feasible. Pre-drilling can have a positive effect on the precision of screw positioning.
172
4. Self-tapping screws as reinforcement to carry compressive stresses

perpendicular to the grain
Structural details in which the timber is loaded in compression perpendicular to the grain are very
common, e.g. beam supports or sills / sole plates. The combination of high loads to be transferred over
localized areas and low capacities in compression perpendicular to the grain can make it difficult to
meet the associated verifications. Fully threaded, self-tapping screws are a means to improve the
stress dispersion into the timber, see Fig. 11. Research on this type of reinforcement is presented in
[56], background information as well as a design approach is presented in [58].
Fig. 11 Reinforcement of support areas

The load-carrying capacity of a reinforced support can be determined under the assumption of an
interaction between the timber under compressive stresses perpendicular to the grain and the screws
under compression. This assumption is valid if certain deformations of the loaded edge are accepted.
In addition it should be verified that the compression capacity perpendicular to the grain of the timber
is not exceeded at the screw tips (transition between reinforced and unreinforced section). For this, an
angle of stress distribution of 45° may be applied. On the safe side, this angle should be measured
from the screw heads, see Fig. 9. Here, the factor kc,90 shall be taken as kc,90 = 1.0.
 Fc,90, Rd  nS  Fax , Rd

F90, Rd  min 

 b  lef  f c,90,d
with: F90, Rd design force perpendicular to grain with reinforcement

Fc,90, Rd design compression perp. to grain capacity without reinforcement,
with Fc,90, Rd  kc,90  b  l  f c,90, d (see [2]) (8)
Fax , Rd design axial load-carrying capacity according to Eq. (1) (without Rt , Rk )
f c,90,d design compression perp. to grain strength
kc,90 factor adjusting compression perp. to grain strength to real design situations (see [3])
nS total number of screws
b beam width (support width  b)
l length
l ef effective length
The compression force must be evenly distributed to all screws and the compression stresses at the
screw heads have to be absorbed by the bearing material. These two requirements can only be met by a
hard bearing material. This can be realized in form of a hard intermediate layer from e.g. steel, designed
in adequate thickness and thus capable to transfer the load uniformly. The screws shall be equally
distributed over the bearing area and the screw heads shall be on one line with the surface of the timber
member. The distance requirements are the same as for screws in tension, see Tab. 1 and Fig. 4.
173
5. Self-tapping screws as reinforcement to carry shear stresses in glued-

laminated timber and cross-laminated timber
5.1 Reinforcement of glued-laminated timber against shear stresses
The shear strength of timber is in the range of five times the magnitude of tension perpendicular to the
grain strength. However, there can be applications in which the shear stresses exceed the shear
capacity of a timber beam. Examples are double tapered or pitched cambered beams, where the
changing depth leads to high shear stresses in the area of the supports, see Fig. 12. With respect to an
economic use of reinforcing elements it is of interest, whether a proportionate distribution of shear
stresses between the timber beam and the shear reinforcement can be achieved in the unfractured
state. In [51] (see also [59] and [60]), an analytical approach is proposed to determine the load-
carrying capacity of timber beams in the uncracked state, featuring shear reinforcement like screws or
threaded rods. This approach is based on common theoretical concepts and constitutive equations for
material properties and enables the incorporation of the semi-rigid composite action between the
reinforcement and wood material, as well as the interaction of shear stresses and stresses
perpendicular to the grain. The applicability and accuracy of the approach is verified by laboratory
tests, also taking into account research carried out and presented in [18]. It appears that the
redistribution of load from the timber to the shear reinforcement is comparatively low. Considering
the uncracked state, comparative calculations indicate that, under realistic construction conditions, an
increase in shear capacity of up to 20 % is feasible.
Fig. 12 Pitched cambered beam – potential reinforcement zones (from [59])

Since wood is characterized by very brittle failure mechanisms both in shear and tension
perpendicular to the grain, it is beneficial to design the corresponding reinforcements so that they are
able to also carry the corresponding stresses in the fractured state, see e.g. [61]. The shear analogy
([62], [63], [64]) represents an applicable approach to calculate the semi-rigid composite action
between both sections in the fractured state. A numerical study on highly stressed shapes of glulam
beams (see [51] and [59]), featuring the minimum required reinforcement to carry the stresses that are
released in the case of cracking shows, that the maximum increase in bending stresses between the
intact state and the fractured state is in the range of one third.
5.2 Reinforcement of cross-laminated timber against shear stresses
The load-carrying capacity of cross-laminated timber (CLT) elements can be limited by the rolling
shear strength, which is in the range of only 1/3 of the shear strength of softwood parallel to the grain,
see Fig. 13. Possible examples are CLT elements loaded out of plane under high concentrated loads or
point supports. Fully threaded, self-tapping screws used as shear reinforcement can have a distinct
positive influence on the shear capacity of CLT elements. The reason is that CLT elements feature
significantly larger shear deformations than GLT elements. This can mainly be explained with the low
rolling shear modulus of the transverse layers (GR ≈ 0.1 ∙ G), see Fig. 13. This results in a larger share
of the screws in the proportionate distribution of shear stresses between the timber beam and the shear
reinforcement.
174
Fig. 13 Shear reinforcement of CLT and GLT – deformation and strength capacities, from [60]
In [65] and [66] a design concept is proposed which is validated by means of experiments, see also
[67]. The concept is based on a strut-and-tie model and takes into account the positive influence of the
interaction of compression perpendicular to the grain and rolling shear stresses, by a factor kR,90, as
well as the load-carrying capacity of the screws in direction of the potential fracture plane, see Eq.
(10) and Fig. 14. In [67] it could also be shown that in the case of biaxial load transfer, additional
effects are activated, leading to an increase in the rolling shear capacity compared to that of beam
elements.
Rax,k 2
f R,k  k R,90  f R,k  (10)
2  a1  a2,ef
1  0,35  c,90 Rax ,k 2

with: k R,90  min  and c,90 
 1,20 2  a1  a2,ef
f R,k characteristic load-carrying capacity of the reinforced CLT under stress [N/mm 2 ]
f R,k characteristic rolling shear capacity (according to technical approvals) [N/mm 2 ]
Rax ,k characteristic load-carrying capacity of a screw parallel to its axis [N]
a1 distance of the screws parallel to the load-bearing direction [mm]
a2,ef effective distance of the screws perpendicular to the load-bearing direction [mm]
l ef effective embedment length of the screws for calculation of R ax,k [mm]
k R,90 parameter for the consideration of the stress interaction [-]
Fig. 12 Design concept for shear reinforcement in CLT on the basis of a strut and tie model, adapted
from [65]
175
6. Self-tapping screws to realize mechanically jointed beams

Fully threaded, self-tapping screws can also be used to strengthen existing beams by mechanically
joining an additional cross-section to the existing beam. Screws feature significantly higher axial slip
moduli, Kax,ser, compared to slip moduli in shear, Kser, see Fig. 15. Therefore they should be positioned
at an angle to the joint (e.g. 45°), see Fig. 16. Models for calculation of these slip moduli can be found
in e.g. [17], see Eq. (2) and [68] and [69].
Fig. 15 Comparsion of axial slip moduli Kax,ser (see Eq. (2)) and slip moduli in shear Kser (see [3]) for
different screw diameters and lengths
Fig. 16 Example of two cross-sections, mechanically jointed by screws incl. schematic illustration of
composition of shear stresses and longitudinal stresses
A common, standardized method [3] to design such beams is the “gamma-method” [70]. The
coefficient γ symbolizes the efficiency of the connection (γ = 1 → rigid connection, γ = 0 → no
connection). It is used to determine an effective moment of inertia of the beam with which the stresses
in the beam as well as the shear flow in the joint can be determined. For cases, in which the basic
assumptions for the “gamma-method” are not met (e.g. continuous joint stiffness and cross-section,
single-span or symmetric continuous beams), an alternative is given by the “shear analogy method”
[62], [63] and [64].
The timber members to be connected by screws should be fixed in their position so that no gap occurs
in the joint when the screws are driven in. If the screws are positioned so that they are loaded in axial
tension, they could close a potential gap due to a rope-effect. Another option is to place the screws
crosswise, i.e. one screw is mainly loaded in axial tension, and the other screw is mainly loaded in
axial compression. The shear flow in the joint can be distributed to the screws using a triangle of
176
forces. When determining the joint stiffness of a mechanically jointed beam, the slip moduli of the
threaded parts of the screws in both beams have to be accounted for. For common types of
mechanically jointed beams with self-tapping screws set at an angle between screw axis and fibre
direction of 45°, values in the range of γ = 0.8 can be obtained [8]-[10].
7. Conclusions and remarks

Fully-threaded, self-tapping screws represent the latest developments in screwing technology for
timber engineering, providing a significant load-carrying capacity if loaded in axial direction. The
screws enable stiff connections but limited plastic potential. Apart from their obvious application as
fasteners in timber connections, fully-threaded, self-tapping screws feature a high potential for
numerous reinforcement applications. An overview of possibilities and related design procedures is
given, including relevant literature for background information.
For numerous applications, design procedures exist, which have already been clarified to an extent
satisfying engineering needs, e.g. reinforcement taking the full tensile stresses perpendicular to the
grain. Other applications are well developed but require some additional research, e.g. reinforcement
approaches which consider proportional load sharing between the timber and the reinforcement,
including the modelling of stiffness properties (Kser, Kax,ser, Ku), as well as reinforcement against
compressive stresses perpendicular to the grain. However, there is a need for further efforts in
research, e.g. considering the potentially harmful effect of reinforcement restricting the free shrinkage
or swelling of the timber.
Although fully threaded, self-tapping screws have undoubtedly a great potential for use in timber
structures, should their potential application as reinforcement always be considered with care. For
centuries, excellent timber structures have been designed without the need for reinforcement. Wood
features a multitude of positive characteristics in view of its application as building material, i.e. a
general reinforcement of timber elements – as known from concrete structures – is not necessary.
Although recent developments have largely increased the range and forms of structures to be realized
with timber products, good design of timber structures should still aim at minimizing stresses for
which timber only features small capacities and brittle failure mechanisms (e.g. tensile stresses
perpendicular to the grain and shear), thereby avoiding or at least minimizing the necessity for
reinforcement.
8. References
[1] Steiger, R., Serrano, E. et al., “Glued-in rods”, in: Harte, A., Dietsch, P. eds. Reinforcement of
Timber Structures, Shaker, Aachen, Germany, 2015 (this publication)
[2] DIN 7998:1975-02, Threads and Thread Ends for Wood Screws, DIN, Berlin, 1975.
[3] EN 1995-1-1:2004-11, Eurocode 5: Design of timber structures – Part 1-1: General – Common
rules and rules for buildings, CEN, Brussels, 2004.
[4] Mahlknecht, U., Brandner, R., Ringhofer, A., Schickhofer G., „Resistance and Failure Modes of
Axially Loaded Groups of Screws“, in: Proceedings of the RILEM Timber Structures
Conference „Materials and Joints in Timber Structures – Recent Advancement of Technology“,
Stuttgart, Germany, 2013.
[5] Pirnbacher, G., Schickhofer G., „Schrauben im Vergleich – eine empirische Betrachtung“, in:
177
Proceedings of the 6. Grazer Holzbau-Fachtagung (6. GraHFT’07) „Verbindungstechnik im

Ingenieurholzbau“, F 1–22, Graz, Austria, 2007.
[6] Dietsch, P., Winter, S., “Eurocode 5 – Future Developments towards a More Comprehensive
Code on Timber Structures“, Structural Engineering International, Vol. 21, No. 2, 2012, pp.
223-231.
[7] EN 14592:2012-05, Timber structures – Dowel type fasteners – Requirements, CEN, Brussels,
2012.
[8] ETA-12/0062, SFS intec AG: Self-tapping screws for use in timber constructions, European
Technical Approval, OIB, p. 17, 18.06.2012.
[9] ETA-11/0112, SPAX International GmbH & Co. KG: Self-tapping screws for use in timber
constructions, European Technical Approval, ETA-Danmark, p. 84, 05.09.2012.
[10] ETA-11/0190, Adolf Würth GmbH & Co. KG: Self-tapping screws for use in timber
constructions, European Technical Approval, DIBt, p. 99, 27.06.2013.
[11] Pirnbacher, G., Brandner, R., Schickhofer, G., „Base Parameters of Self-Tapping Screws“, CIB-
W18 / 42-7-1, Proceedings of the international council for research and innovation in building
and construction, Working commission W18 – timber structures, Meeting 42, Duebendorf,
Switzerland, 2009.
[12] Colling, F., Holzbau – Grundlagen und Bemessung nach EC5, 3. Auflage, Springer Vieweg,
ISBN 978-3-8348-1789-1, 2012.
[13] Ringhofer, A., Brandner, R., Schickhofer, G., „Entwicklung einer optimierten
Schraubengeometrie für hochbeanspruchte Stahl-Holz-Verbindungen“, Bautechnik, Vol. 91, No.
1, 2014, pp. 31-37
[14] Enders-Comberg, M., Blaß, H.J., “Influence of fasteners in the compression area of timber
members”, CIB-W18 / 46-7-8, Proceedings of the international council for research and
innovation in building and construction, Working commission W18 – timber structures,
Meeting 46, Vancouver, Canada, 2013.
[15] Ringhofer, A., Grabner, M., Brandner, R., Schickhofer, G., „SGSC 3.2.1_1 – Prüftechnische
Ermittlung des Tragverhaltens der Einzelschraube in der BSP-Schmalfläche“, Internal
Research Report, holz.bau forschungs gmbh, Institute of Timber Engineering and Wood
Technology, Graz University of Technology, 2013.
[16] Krenn, H., Schickhofer, G., „Joints with inclined screws and steel plates as outer members“,
CIB-W18 / 42-7-2, Proceedings of the international council for research and innovation in
building and construction, Working commission W18 – timber structures, Meeting 42,
Duebendorf, Switzerland, 2009.
[17] Blaß, H. J., Bejtka, I., Uibel, T., „Tragfähigkeit von Verbindungen mit selbstbohrenden
Holzschrauben mit Vollgewinde“, Forschungsbericht der Versuchsanstalt für Stahl, Holz und
Steine, Abt. Ingenieurholzbau, Universität Karlsruhe (TH), 2006.
[18] Blaß, H.-J., Krüger, O., Schubverstärkung von Holz mit Holzschrauben und Gewindestangen,
Karlsruher Berichte zum Ingenieurholzbau, Band 15, Universitätsverlag Karlsruhe, 2010.
[19] Brüninghoff, H., Schmidt, C., Wiegand, T., „Praxisnahe Empfehlungen zur Reduzierung von
Querzugrissen bei geleimten Satteldachbindern aus Brettschichtholz“, Bauen mit Holz, Vol. 95,
No. 11, 1993, pp. 928-937.
[20] DIN EN 1995-1-1/NA:2010-12, National Annex – Nationally determined parameters –
Eurocode 5: Design of timber structures – Part 1-1: General – Common rules and rules for
buildings, DIN, Berlin, 2010.
[21] DIN 1052:2008-12, Entwurf, Berechnung und Bemessung von Holzbauwerken – Allgemeine
Bemessungsregeln und Bemessungsregeln für den Hochbau, DIN, Berlin, 2008.
178
[22] DIN 1052:2004-08, Entwurf, Berechnung und Bemessung von Holzbauwerken – Allgemeine
Bemessungsregeln und Bemessungsregeln für den Hochbau, DIN, Berlin, 2004.
[23] Blaß, H.J., Steck, G., „Querzugverstärkungen von Holzbauteilen: Teil 1 – Teil 3“, Bauen mit
Holz, Vol. 101, No. 3, pp. 42-46, No. 4, pp. 44-49, No. 5, pp. 46-50, 1999.
[24] Blaß, H.J., Bejtka, J., “Reinforcements perpendicular to the grain using self-tapping screws”,
in: Proceedings of the 8th World Conference on Timber Engineering, Lahti, Finland, 2004.
[25] Möhler, K., Siebert, W., „Untersuchungen zur Erhöhung der Querzugfestigkeit in gefährdeten
Bereichen“, Bauen mit Holz, Vol. 86, No. 6, 1984, pp. 388-393.
[26] Ehlbeck, J., Görlacher, R., Werner, H., “Determination of Perpendicular-to-Grain Tensile
Stresses in Joints with Dowel Type-Fasteners”, CIB-W18 / 22-7-2, Proceedings of the
international council for research and innovation in building and construction, Working
commission W18 – timber structures, Meeting 22, Berlin, West Germany, 1989.
[27] Kreuzinger, H., „Holzbau“, in: Zilch, K., Diederichs, C.J., Katzenbach, R. (eds.), Handbuch für
Bauingenieure, Springer, Berlin, 2002.
[28] Mahlknecht, U., Brandner, R., „focus_sts 3.1.2_1: Untersuchungen des mechanischen
Verhaltens von Schrauben – Verbindungsmittelgruppen in VH, BSH und BSP“, Internal
Research Report, holz.bau forschungs gmbh, Institute of Timber Engineering and Wood
[29] Henrici, D., Beitrag zur Spannungsermittlung in ausgeklinkten Biegeträgern aus Holz,
Dissertation, Technische Universität München, 1984.
[30] Henrici, D., „Beitrag zur Bemessung ausgeklinkter Brettschichtholzträger“, Bauen mit Holz,
Vol. 92, No. 11, 1990, pp. 806-811.
[31] Tannert, T., Lam, F.; “Self-tapping screws as reinforcement for rounded dovetail connections”,
Structural Control and Health Monitoring, Vol. 16, No. 3, 2009, pp. 374-384.
[32] Gustafsson, P.J., “Notched beams and holes in glulam beams”, in: Blaß, H.J., Aune, P., Choo,
B.S., et al. (eds). Timber Engineering STEP 1 – Basis of design, material properties, structural
components and joints, Centrum Hout, Almere, 1995
[33] Neuhaus, H.; Ingenieurholzbau – Grundlagen, Bemessung, Nachweise, Beispiele, 3. Auflage,
Vieweg&Teubner, ISBN 978-3-8348-1286-5, 2010.
[34] Jockwer, R., Frangi, A., Steiger, R., Serrano, E., “Enhanced design approach for reinforced
notched beams”, CIB-W18/ 46-6-1, Proceedings of the international council for research and
Meeting 46, Vancouver, Canada, 2013.
[35] Steilner, M., Blaß, H.J., „Vorspannen von Holz mit Vollgewindeschrauben“, Proceedings 17.
Internationales Holzbau-Forum, Garmisch-Partenkirchen, 2011.
[36] Blaß, H.J., Bejtka, I., Querzugverstärkungen in gefährdeten Bereichen mit selbstbohrenden
Holzschrauben, Forschungsbericht der Versuchsanstalt für Stahl, Holz und Steine, Abt.
Ingenieurholzbau, Universität Karlsruhe (TH), 2003.
[37] Aicher, S., Höfflin, L., “Glulam Beams with Holes Reinforced by Steel Bars”, CIB-W18 / 42-
12-1, Proceedings of the international council for research and innovation in building and
construction, Working commission W18 – timber structures, Meeting 42, Duebendorf,
Switzerland, 2009.
[38] Aicher, S., “Glulam Beams with Internally and Externally Reinforced Holes – Test, Detailing
and Design”, CIB-W18 / 44-12-4, Proceedings of the international council for research and
Meeting 44, Alghero, Italy, 2011.
179
[39] Kolb, H., Frech, P., „Untersuchungen an durchbrochenen Bindern aus Brettschichtholz“, Holz
als Roh- und Werkstoff, Vol. 35, No. 4, pp. 125-134
[40] Kolb, H., Epple, A., Verstärkungen von durchbrochenen Brettschichtholzbindern,
Schlussbericht zum Forschungsvorhaben I.4 – 34810, Forschungs- und Materialprüfungs-
anstalt Baden-Württemberg, Stuttgart, 1985.
[41] Aicher, S., Höfflin, L., “Glulam Beams with Round Holes – a Comparison of Different Design
Approaches vs. Test Data”, CIB-W18 / 35-12-1, Proceedings of the international council for
research and innovation in building and construction, Working commission W18 – timber
structures, Meeting 35, Kyoto, Japan, 2002.
[42] Aicher, S., Höfflin, L., “Design of rectangular holes in glulam beams”, Otto-Graf-Journal, Vol.
14, 2003, pp. 211-229.
[43] Aicher, S., Höfflin, L., Reinhardt, H.-W., „Runde Durchbrüche in Biegeträgern aus
Brettschichtholz, Teil 2: Tragfähigkeit und Bemessung“, Bautechnik, Vol. 84, No. 12, 2007, pp.
867-880.
[44] Blaß, H.J., Ehlbeck, J., Kreuzinger, H., Steck, G., Erläuterungen zu DIN 1052:2004-08,
Bruderverlag, Karlsruhe, 2004.
[45] Blumer, H., Spannungsberechnungen an anisotropen Kreisbogenscheiben und Sattelträgern
konstanter Dicke, Lehrstuhl für Ingenieurholzbau und Baukonstruktionen, Universität
Karlsruhe, 1972/1979.
[46] Ehlbeck, J, Kürth, J., „Influence of perpendicular-to-grain stressed volume on the load-carrying
capacity of curved and tapered glulam beams“, CIB-W18 / 24-12-2, Proceedings of the
commission W18 – timber structures, Meeting 24, Oxford, United Kingdom, 1991.
[47] Moser, M., Theoretische Betrachtung der Berechnungsgrundlagen für auf Querzug
beanspruchte BSH-Bauteile, Master Thesis, Institute of Timber Engineering and Wood
[48] Dietsch, P., Winter, S.; Untersuchung von nicht in DIN EN 1995-1-1 geregelten Formen von
Satteldachträgern im Hinblick auf den Nachweis der Querzugspannungen, Research Report,
Lehrstuhl für Holzbau und Baukonstruktion, Technische Universität München, 2014.
[49] Gowda, S, Ranta-Maunus, A., Curved and cambered glulam beams - Part 1: Short term load
Tests, Research Notes 1500, Technical Research Centre of Finland, Espoo, 1993.
[50] Dietsch, P., Gamper, A., Merk, M., Winter, S.; "Building Climate – long-term measurements to
determine the effect on the moisture gradient in large-span timber structures", CIB-W18 / 45-
construction, Working commission W18 – timber structures, Meeting 45, Växjö, Sweden, 2012.
[51] Dietsch, P., Einsatz und Berechnung von Schubverstärkungen für Brettschichtholzbauteile,
Dissertation, Technische Universität München, 2012.
[52] Dietsch, P., Kreuzinger, H., Winter, S.; “Effects of changes in moisture content in reinforced
glulam beams”, in: Proceedings of the 13th World Conference on Timber Engineering, Quebec,
Canada, 2014.
[53] Wallner, B., Versuchstechnische Evaluierung feuchteinduzierter Kräfte in Brettschichtholz
verursacht durch das Einbringen von Schraubstangen, Master Thesis, Institute of Timber
Engineering and Wood Technology, Graz University of Technology, 2012.
[54] Jorissen, A.J.M., Double shear timber connections with dowel type fasteners, Dissertation,
Delft University of Technology, Netherlands, 1998.
[55] Bejtka, I., Blaß, H.J., “Self-tapping screws as reinforcements in connections with dowel-type
fasteners”, CIB-W18 / 38-7-4, Proceedings of the international council for research and
180

Meeting 38, Karlsruhe, Germany, 2005.
[56] Bejtka, I., Verstärkung von Bauteilen aus Holz mit Vollgewindeschrauben, Dissertation, Band 2
der Reihe Karlsruher Berichte zum Ingenieurholzbau, KIT Scientific Publishing, Karlsruhe,
2006.
[57] Kobel, P., Modelling of strengthened connections for large span truss structures, Master Thesis,
Institute of Structural Engineering, ETH Zürich, Switzerland & Division of Structural
Engineering at Lund University/LTH, Sweden2011.
[58] Bejtka, I., Blaß, H.J., “Self-tapping screws as reinforcements in beam supports”, CIB-W18/ 39-
construction, Working commission W18 – timber structures, Meeting 39, Florence, Italy, 2006.
[59] Dietsch, P., Kreuzinger, H., Winter, S., “Design of shear reinforcement for timber beams”, CIB-
W18/ 46-7-9, Proceedings of the international council for research and innovation in building
and construction, Working commission W18 – timber structures, Meeting 46, Vancouver,
Canada, 2013.
[60] Dietsch, P., Mestek, P., Winter, S., "Analytischer Ansatz zur Erfassung von
Tragfähigkeitssteigerungen infolge von Schubverstärkungen in Bauteilen aus Brettschichtholz
und Brettsperrholz", Bautechnik, Vol. 89, No. 6, 2012, pp. 402-414.
[61] Dietsch, P., "Robustness of large-span timber roof structures – Structural aspects", Engineering
Structures, Vol. 33, No. 11, 2011, pp. 3106-3112.
[62] Kreuzinger, H., „Platten, Scheiben und Schalen – ein Berechnungsmodell für gängige
Statikprogramme“, Bauen mit Holz, Vol. 101, No. 1, 1999, pp. 34-39.
[63] Kreuzinger, H., „Verbundkonstruktionen“, in: Holzbaukalender 2002, Bruderverlag, Karlsruhe,
2001, pp. 598-621.
[64] Kreuzinger, H., “Mechanically Jointed Beams: Possibilities of Analysis and some special
Problems”, CIB-W18 / 34-12-7, Proceedings of the international council for research and
Meeting 34, Venice, Italy, 2001.
[65] Mestek, P., Kreuzinger, H., Winter, S., “Design Concept for CLT – Reinforced with Self-
Tapping Screws”, CIB-W18/ 44-7-2, Proceedings of the international council for research and
Meeting 44, Alghero, Italy, 2011.
[66] Mestek, P., Winter, S., „Punktstützung von Brettsperrholzkonstruktionen – Schubverstärkungen
mit Vollgewindeschrauben“, Bauingenieur, Vol. 86, No. 12, 2011, pp. 529-540.
[67] Mestek, P., Punktgestützte Flächentragwerke aus Brettsperrholz (BSP) – Schubbemessung
unter Berücksichtigung von Schubverstärkungen, Dissertation, Technische Universität
München, 2011.
[68] Kevarinmäki, A., “Joints with inclined screws”, CIB-W18 / 35-7-4, Proceedings of the
commission W18 – timber structures, Meeting 35, Kyoto, Japan, 2002.
[69] Tomasi, R., Crosatti, A., Piazza, M. "Theoretical and experimental analysis of timber-to-timber
joints connected with inclined screws", Construction and Building Materials, Vol. 24, 2010, pp.
1560-1571.
[70] Möhler, K., Über das Tragverhalten von Biegeträgern und Druckstäben mit zusammengesetzten
Querschnitten und nachgiebigen Verbindungsmitteln, Habilitation, Universität Karlsruhe, 1956.
181
182
FRP reinforcement of timber structures
10
Kay-Uwe Schober1, Annette M. Harte2, Robert Kliger3, Robert Jockwer4, Qingfeng Xu5,
Jian-Fei Chen6
Summary
Timber engineering has advanced over recent decades to offer an alternative to traditional materials
and methods. The bonding of fibre reinforced plastics (FRP) with adhesives to timber structures for
repair and strengthening has many advantages. However, the lack of established design rules has
strongly restrained the use of FRP strengthening in many situations, where these could be a preferable
option to most traditional techniques. A significant body of research has been carried out in recent
years on the performance of FRP reinforced timber and engineered wood products. This chapter gives
a State of the Art summary of material formulations, application areas, design approaches and quality
control issues for practical engineers to introduce on-site bonding of FRP to timber as a new way in
design for structural repair and rehabilitation.
1. Introduction
Fibre reinforced polymer materials combining high strength fibres and a resin matrix have a wide
variety of industrial applications due to their high strength-to-weight ratio and ease of handling. Their
versatility is reflected in the construction industry where they have been widely used for many years,
especially for strengthening of concrete structures [1],[2]. More recently, the techniques have also
been extended to timber structures [3]-[6].
FRP reinforcement in the form of pultruded rods or plates and woven fabrics has an expanding and
particularly effective application in structural repair of timber structures. They may be inserted at
critical locations for enhanced loading capacity, or for load transmission when damaged beam-ends
are cut off and replaced with new timber or epoxy mortar. The reinforcement techniques are usually
based on the use of adhesives on site, and use procedures that are common for the repair or the
1)
Professor of Timber Engineering and Structural Design, Mainz University of Applied Sciences, Mainz,
Germany
2)
Senior Lecturer in Engineering, National University of Ireland, Galway, Ireland
3)
Professor of Structural Engineering, Chalmers University of Technology, Gothenburg, Sweden
4)
Postdoctoral Researcher in Timber Engineering, ETH Swiss Federal Institute of Technology, Zurich,
Switzerland
5)
Deputy Chief Engineer, Shanghai Research Institute of Building Sciences, Shanghai, China
6)
Professor of Civil and Structural Engineering, Queens University Belfast, Belfast, Northern Ireland, UK
183
upgrading of concrete and metallic structures. Such techniques minimize the disturbance to the
building and its occupants during the intervention. However, some concerns have prevented the wider
use of adhesives, particularly in historical timber structures, where sufficient reliability cannot yet be
guaranteed. One reason is that a long service life has not yet been fully proven for synthetic adhesives,
since the oldest bonded joints are only around sixty years and greater ages cannot be simulated by
existing accelerated ageing tests.
Fibre reinforced polymers and structural adhesives have been used to repair or strengthen structural
members for many years. This approach is nevertheless difficult to extend to less common wood
species, less favourable environmental conditions, new adhesive formulations and other variables.
The lack of established design rules available to engineers and other decision makers has significantly
restrained the use of FRP strengthening techniques in many situations where these could otherwise be
a preferable option to most traditional techniques or to the total replacement of timber members.
In this chapter, the materials used in FRP reinforcement of timber structures are discussed, a design
approach for bondline delamination is presented, current and potential applications of FRP for
reinforcement of timber structures are described, design rules are outlined and finally relevant quality
control procedures for on-site bonding are summarised.
2. Materials
2.1 FRP reinforcement materials
FRP materials are composites comprising fibres that provide the load-bearing capacity and stiffness,
embedded in a polymeric resin that transfers loads between fibres and provides protection for the
fibres. They are available in a wide variety of forms, and have properties that vary considerably
depending on the fibre material, volume fraction and orientation. Typical properties of the common
fibres and polymers are given in Tab. 1
Tab. 1 Fibre and polymer properties [1]
Material Modulus of Elasticity Tensile Strength Failure Strain CTE Density

[GPa] [MPa] [%] [10-6 °C-1] [g/cm3]
E-glass 70-80 2000-4800 3.5-4.5 5.0-5.4 2.5-2.6
Carbon (HM) 390-760 2400-3400 0.5-0.8 -1.45 1.85-1.90
Carbon (HS) 240-280 4100-5100 1.60-1.73 -0.6 - -0.9 1.75
Aramid 62-180 3600-3800 1.9-5.5 -2.0 1.44-1.47
Basalt 82-110 860-3450 5.5 3.15 1.52-2.7
Polymer 2.7-3.6 40-82 1.4-5.2 30-54 1.10-1.25
CTE: Coefficient of thermal expansion; HM: High modulus; HS: High strength
For structural reinforcement, two main forms of FRP are generally used, namely, pultruded rods or
plates and fabrics. For internal reinforcement, pultruded rods and plates are bonded into slots or
grooves formed in the timber element. For external reinforcement, FRP plates or fabric materials are
used.
184
2.2 Adhesives
The reinforcement of timber with FRP is normally implemented by adhesive bonding. For bonding of
pultruded FRP plates, the adhesive is applied to the timber substrate and the FRP is then applied to the
adhesive under pressure. FRP fabric reinforcement is generally applied using a wet lay-up method,
whereby the fabric is impregnated with the adhesive first and then this is applied to the timber.
Epoxy based adhesives have been used in most cases for on-site repair jobs, most formulations were
developed for other materials. These adhesives are generally too rigid for bonding timber and there is
no chemical bonding or suitable mechanical anchorage in wood. The bond surface is prone to failure
because of dimensional changes in the wood induced by moisture content variations, even for indoor
applications such as those defined under Eurocode 5 Service Class 2. More recently, epoxy
formulations have been developed specifically for use with wood. On-site application of adhesives
may be difficult and the quality of adhesive bond is not easy to evaluate. Since properties of
reinforced elements very much depend on construction quality, relevant procedures for applying and
controlling adhesive quality[9],[10],[10] have to be followed.
The selection of the adhesive for bonding of FRP to timber has to be undertaken with great care. The
adhesive must be capable of bonding with both the FRP and timber and should have adequate
strength. Four categories of adhesive are available: epoxies, polyurethanes, polyesters, phenolics and
aminoplastics [11]. While many of these adhesive types have been shown to provide satisfactory
bonding performance when used in a controlled environment [12],[16], two-part cold-cure epoxy
adhesives have generally been found to be most suitable for on-site bonding as they have good gap-
filling properties, are thixotropic and have low curing shrinkage. Additional information on adhesives
for on-site bonding: characteristics, testing, applications can be found in [17].
A wide variety of epoxy formulations is available and compatibility with the adherents must be
verified by the manufacturer. On site, the moisture content and the surface quality of the timber are
difficult to control and it may not be possible to achieve the desired bonding pressure. Thick bond
lines in the order 1-3 mm are common. Careful surface preparation is essential in order to achieve
good bond strength and durability. Surfaces to be bonded should be dry, free from contaminants, such
as release agents and dust, and have sufficient surface roughness. For the FRP this involves abrasion
followed by solvent cleaning or else removal of the peel ply if one is provided. Surface preparation
should be carried out immediately prior to bonding while preparation and application of the adhesive
should be in accordance with the manufacturer’s instruction. The importance of using experienced
operatives cannot be underestimated. Quality control measures should be implemented at each stage
of the process [18].
2.3 Bond behaviour of FRP-timber interface
Central to successful reinforcement is the integrity of the bond between the FRP and timber substrate
[1],[19]. Over the last two decades, a number of studies were undertaken by different researchers on
bond behaviour of FRP-to-wood and many useful results were acquired, including failure modes,
stress distribution and local bond-slip relationship.
In the early studies on the behaviour of FRP to wood bonds, the test configuration, shown in Fig. 1(a),
was a modified form of the ISO 6238 and ASTM D905-03 block shear test [12],[13],[20]-[23]. In
those tests, the shear strength of the bond was derived as an average stress over the bonded plate
length. As the surface of the FRP was sandwiched between two pieces of timber, FRP plate surface
strains were difficult to monitor, not to mention the shear stress distribution and bond-slip responses.
185
SILVA et al. [24] conducted four-point bending tests on timber-FRP joints with the strengthening
techniques of near-surface mounted (NSM) and externally bonded reinforcement (EBR) which was
convenient to get the FRP strain distribution, shear stress distribution and bond-slip responses. On the
basis of SILVA’s tests, Juvandes and Barbos [25] analyzed the maximum anchor strength of the
composite and the maximum composite strain, and proposed the effective bonding length for the EBR
and NSM reinforcements
Wan et al. [26] introduced single-lap FRP-to-timber joint shear tests, shown in Fig. 1(b), in which the
strength of the bond between the FRP and timber was examined. The shear tests were conducted on
softwood (Pine) which was strengthened with carbon FRP with the main test variables being the FRP
bond length and the growth characteristics of the timber. Extensive strain gauging of the FRP has
enabled the onset and propagation of debonding cracks to be monitored. Failure modes of the joints
and effective bond length were identified based on the test results. Wan et al. [19] also conducted a
series of tests on 86 single-shear FRP-to-timber joints to study on FRP-to-timber bonded interfaces.
The test parameters included adhesive type, FRP plate type and timber species. Test results showed
that all softwood joints failed predominantly in the timber while the hardwood joints exhibited failure
at different interfacial positions. Load-slip, strain and bond stress distribution, and bond stress-slip
responses were consistent with those of FRP-to-concrete bonds. A test-based theoretical bond stress-
slip law utilizing the J-integral method was proposed and could be implemented in analytical and
numerical models.
wood
FRP
b)
a)
Fig.1 (a) Shear block test configuration; (b) single lap joint shear test
The FRP composites mentioned above were all FRP sheets or strips. There was much less research on
bond behaviour of wood and FRP bars. Lorenzis et al. [27] conducted pull-out tests to study the bond
performance of FRP rods epoxied into glulam timber. The test variables were bonded length, surface
configuration of the rod and direction of the wood fibres with respect to the longitudinal axis of the
joint. The observed failure modes were cohesive failure in timber for rods glued-in parallel to the
grain and adhesion failure for rods glued-in perpendicular to the grain.
In addition to the experimental investigations, finite element simulation on bond behaviour of FRP-to-
timber was developed in recent years. Valipour and Crews [28] proposed a novel force-based element
in the framework of the total secant approach for nonlinear analysis of timber beams strengthened
with FRP sheet (bar), including bond-slip effects. The formulation takes account of material
nonlinearities and preserves the continuity of slip shear. Further, a composite Simpson integration
scheme with a finite difference scheme was employed to calculate the bond shear forces along the
element. It was concluded that for the considered cases, bond shear-slip between the FRP sheet (bar)
and timber has a minor effect on the ultimate loading capacity as well as the load-deflection response
186
of timber beams strengthened with FRP sheets (bars). The assumption of perfect bond between the
FRP and timber beam was observed to be acceptable in most cases.
In summary, the research on the bond behaviour of FRP-to-timber is still in its infancy. Further
investigations should be carried out to determine the influence of the various factors affecting the
bond between FRP and wood including timber properties and specimens, thicknesses (diameters) and
specimens of FRP sheets (bars), and thickness of adhesive. The local bond-slip relationship for FRP-
to-timber which can be directly used for finite element analysis is definitely the research focus in the
future.
2.4 FRP damage and delamination by loads
2.4.1 Global fracture criteria for composite design
A realistic design approach to account for the fracture and delamination in FRP strengthened timber
structures needs appropriate material models describing physically-based failure criteria of the
anisotropic and non-linear properties of these composite structures. Global fracture criteria with
complete stress interaction were the first failure criteria developed [29]. The main advantage is the
ease of use for analytical, numerical and design approaches due to a single scalar equation of failure
for unidirectional laminates. These hypotheses contain no information about the fracture mode of
composites in three-dimensional stress states; the fibre-parallel fracture plane is unknown. Stress
combinations like (21, 2) and (31, 3) are set to equal and do not consider the importance of loads
perpendicular to the fibre within the fibre plane which are critical due to the dimensions of most
structural elements.
The most common criterion used in FE-codes has been developed by TSAI and WU, shown in Eq. (1).
The symmetry of material properties is considered by strength coefficients which can be easily
transformed in their invariances.
D0 D D 2 2 2
F  1  90  2  90  3  1  2  3
F11 F22 F22 F11 F22 F22
(1)
 122  132  232
 F12 1 2  F12 1 3  2 F23 2 3    1
f v2,90,0 f v2,90,0 f v2,90,90
with
F11  1  f t ,0 f c ,0  F1  1 f t ,0  1 f c ,0
F22  1  f t ,90 f c ,90  F2  1 f t ,90  1 f c ,90
2 1
F12  0 2 F23  
f c ,90 f t ,90 f v2,90,90
D0  f c ,0  f t ,0 D90  f c ,90  f t ,90
ft,0 tensile strength in fibre direction
fc,0 compressive strength in fibre direction
ft,90 tensile strength perpendicular to fibre direction
fc,90 compressive strength perpendicular to fibre direction
fv,90,0 shear strength in laminate area
fv,90,90 shear strength perpendicular to laminate area
187
1, 2, 3 are normal stresses and 23 = 32, 13 = 31, 12 = 21 shear stresses related to the (xl, x2, x3)
coordinate system of the unidirectional laminate area, where (xl, x2)-area describes the laminate area
and x3 the thickness direction (Fig. 2).
Fig. 2 Stresses in unidirectional laminates

With this hypothesis most of fracture modes can be described but fibre fracture (FF) and interlaminar
fibre fracture (IFF) cannot be identified. Furthermore, the strength increase cannot be represented by a
maximum criterion but only by an interactive criterion. Therefore, for FRPs the combination of the
general Tsai-Wu criterion with the hypothesis of maximum stresses is recommended, otherwise stress
states greater than the single value of the strength will be accepted.
2.4.2 Failure-mode based strength criteria for composite design

Each failure mechanism is governed by one mode-associated strength. Failure conditions are required
to be simply formulated, numerically robust, and physically-based and to allow a simple
determination of the highest mode stress effort (design driving). The most important investigations
have been made by Puck in a further development of the classical Mohr theory and Hashin’s failure
criteria [30]. The key statement in his proposal is that failure in a fibre parallel fracture plane under
tension perpendicular to the plane is caused by the tension stresses and also by the shear stresses nt
and n1 (see 2), while compression forces perpendicular to the fibre plane increase the shear resistance.
A further enhancement of this criterion was the introduction of an additional term of the internal
Coulomb friction to account for stress stiffening below the fracture limit due to compression. For the
case of tension reinforcement, the general fracture criterion of Puck is given by an easy to modify 7-
parameter model identifying FF and IFF for design issues. A more sophisticated version has been
developed in the form of the mathematically simpler Puck-Knaust-IFF-failure criteria [30].
 22   32   2 3   232  1 1   212   312

F     2   3   2
f t ,90 f c ,90  f t ,90 f c ,90  f v ,90,0
 1
(2)
1 1 
  12    1
f t ,0, IFF f c ,0, IFF  f t ,0, IFF f c ,0, IFF 
F 1 
with F() = fracture plane
2.4.3 Practical application for engineering problems

The definition of failure is an issue during failure analysis. In particular, initial failure like
delamination has to be defined in the right way. The classification of failure criteria regarding their
original aim has to be assessed before their use:
188
 Analysis range (first ply or post failure)

 Physical model (micro, macro or component level)
 Mathematical approach (limit, polynomial tensor or physically based)
One of the biggest problems for circulation of knowledge on failure of FRP is the non-availability of
respective post-processing software codes. The Tsai-Wu criterion is implemented in most software
e.g. ANSYS® or ABAQUS® being able to analyze the composite structures. Due to the same
mathematical layout of the parabolic criterion of the Puck-Knaust and the Tsai-Wu criterion in form
of tensor polynomials it is possible to implement the Puck-Knaust criterion in commercial FE-codes
for an appropriate design of composite structures by enhancing the included failure criteria. The
complete mathematical expression of the modification and implementation for engineering design
problems is given in Tab. 2.
Tab. 2 Strength indices for Puck-Knaust modified Tsai-Wu criteria for FRPs
IFF Strength FF IFF

Strength index FF modification
modification index modification modification
ft,x ft,0 1,3 ft,0 fv,xy > 104 MPa fv,0,90
fc,x fc,0 1,3 fc,0 fv,yz > 104 MPa 1/ 3 f t ,90 f c ,90
ft,y > 104 MPa ft,90 fv,zx > 104 MPa fv,90,90
fc,y > 104 MPa fc,90 Cxy 0 0
ft,z > 104 MPa ft,90 Cyz 0 –1
fc,z > 104 MPa fc,90 Czx 0 0
Due to the mathematically similar definition of the Tsai-Wu criteria and the Pick-Knaust criteria by
tensor polynomials, the implemented algorithm in FE-codes can be modified using Tab. 2 for the
specific of fiber reinforced polymers and own stress and failure analyses.
2.4.4 Numerical aspects of bond-line delamination

Decohesion along interfaces plays an important role in a wide variety of failure processes in structures
when using chemical bonding as the optimal form of combining two surfaces with each other. The
analysis of delamination initiation is based on stresses and interaction criteria of the interlaminar
stresses in conjunction with a characteristic distance as a function of geometry and material
properties. Crack propagation is predicted by a fracture mechanics (FM) approach. This avoids the
difficulties associated with a stress singularity at the crack front but requires the presence of a pre-
existing delamination. When used in isolation, neither the strength-based approach nor the FM
approach is adequate for a progressive delamination failure analysis.
For the combination of anisotropic materials like wood with FRP reinforcement in an interface
damage law, the interaction between shear and normal stresses has to be considered. This problem can
be solved by using a delamination analysis and an exponential interface damage law [32]. The
definition of traction and separation depend on the finite element and the applied material model.
189
Decohesion response is specified in terms of a surface potential () relating the interface tractions
and the relative normal and tangential displacements n and t across the interface. In most of the
computations all cohesive surfaces are taken to have identical cohesive properties which simplifies the
surface potential [33] to
    2 
( )  e  c  n 1  exp   n  t2  (3)
   n  t 
e Euler constant
c maximum normal traction at the interface
n normal separation where the maximum normal traction is attained with t = 0
t shear separation where the maximum shear traction is attained
The potential leads by derivation on the displacements to the stresses if  < max where the traction
components T are coupled to both normal and tangential crack opening displacements. The main
advantage of the cohesive zone modelling is that, when it is known where fracture may occur a priori,
a cohesive zone may be placed anywhere along element interfaces in that area, to take these effects
into account. Furthermore, using decohesion elements, both onset and propagation of delamination
can be simulated without previous knowledge of crack location and propagation direction and is
therefore suitable for structural design and evaluation of composite beams.
2.5 FRP damage by fire
When FRP is subjected to elevated temperatures, the strength and stiffness are reduced due to the loss
of mechanical integrity of the polymer matrix [3]. The two-part epoxy adhesives normally used in
bonding have glass transition temperatures less than 100°C. For externally bonded FRP
reinforcement, the exposure to high temperatures in a fire situation can therefore lead to both loss in
the reinforcing effect of the FRP and degradation of the bond between the FRP and timber. In order to
prevent this situation, the innate insulating properties of timber can be used to protect the
reinforcement by providing adequate timber covering to the FRP.
Martin and Tingley [34] undertook a test program in which FRP reinforced and unreinforced control
glulam beams were subjected to standard fire exposure and mechanical loads equal to or in excess of
the design loads. The beams were reinforced with three different types of FRP plates, bonded either
externally to the soffit or between the bottom two laminations. The results showed that internally
reinforced beams had a 44% higher fire endurance compared to the externally reinforced specimens.
No difference in fire performance was found between the different types of FRP. Williamson [35]
showed that glulam beams with externally bonded FRP can be designed to achieve a 1-hour rating.
Fire exposure testing on glulam beams with beam end repairs comprising short epoxy resin
replacement beams attached to the main beams by means of steel reinforcing bars [36] showed that
these repairs could give one hour fire resistance. In this case, the steel bars had a minimum cover of
60 mm to the surface.
190
3. Applications
3.1 Beam end reinforcement
In order to make the interventions minimally intrusive, the connecting elements are confined to the
corners of the beam and their length is reduced to the minimum necessary. The operating procedures
depend on the specific requirements of the site. This starts with the propping of the beams followed
by removal of the decayed portion of the timber, usually terminated in an inclined cut (3). The new
section of wood is shaped to match the external dimensions of the original decayed section. After the
preparation of internal holes or external grooves between the original and newly introduced wood for
the positioning of the connecting elements (4), the holes or groove are partially filled with adhesive
and the reinforcing elements are inserted. The last stage is the insertion of a final wood fillet, to hide
the grooves (5) followed by removal of the beam supports after the complete curing of the adhesive.
Fig. 3 Inclined cut-decayed end Fig. 4 Linking grooves Fig. 5 Insertion of final wedge
Epoxy resins have been used in specific cases to repair timber that has deteriorated from decay or
insect attack and certain structural deficiencies in existing construction. The compressive strength and
filling capabilities of epoxies can aid in repairing timber structures. The tensile, shear, and bond
strength of epoxies as structural wood adhesives are, in certain cases, limited and are further subject to
variability due to conditions of use.
Mechanical reinforcement should be used in conjunction
with epoxies for repairs intended to develop shear
capacity. For structural rehabilitation and restoration the
decayed ends are cut off and replaced by timber or
polymer concrete where the shear and tension forces
resulting from loads are transmitted by GFRP
reinforcement with a PU or EP resin filling compound Fig. 6 Shear reinforcement with GFRP
between the new and the existing part (Fig. 6).
In some cases the decayed part of the structure will be replaced by a polymer concrete (PC) supplement
using rapid-setting organic polymers as binders and small grade aggregates such as sand or gravel. The
application of PC in structural rehabilitation and restoration has several advantages. Historically
significant structures can be protected by minimum disturbance of the construction and minimum
replacement of decayed parts. The appearance of the timber structure will not be changed. The section
design can be easily executed by timber formwork on the level of the necessary construction height. All
work can be undertaken from the top side so suspended ceilings will remain unaffected. The floor below
the reconstruction work can be used with some restrictions. The method is suitable for inhabited floors
and restoration of complicated timber joints. The full load-carrying capacity is achieved after one day
with an increase of the structural performance by 140% for timber on-timber supplements and 30% for
timber-on-PC supplements as reported by Schober et al. [32],[33].
191
3.2 Tension reinforcement perpendicular to the grain

The tensile strength of the timber perpendicular to the grain is considerably lower compared to the
tensile strength parallel to the grain. High tension stresses perpendicular to the grain stresses occur at
notches, holes or curved beams and beams with variable height. Efficient reinforcement techniques
are required for increasing or maintaining the load-carrying capacity of these structures. The
reinforcement should offer high strength and stiffness and should create a more ductile failure of the
structure. Reinforcement made of FRP offers high strength and stiffness along its fibre direction;
however, it commonly does not lead to more ductility of the structure. There is relatively little
information on research and applications regarding FRP reinforcement of regions with tension
perpendicular to the grain reported in literature due to the existence of other reinforcing techniques
offering more ductility, like e.g. self-tapping screws.
In general, the reinforcement should be applied close to the region of high tensile stresses
perpendicular to the grain. As discussed by Hallström and Grenestedt [37], the use of internal
reinforcement may be necessary for internal members that are not accessible at their external sides.
For wrapping of FRP fabrics around the timber member good accessibility of the member at all sides
is crucial.
Studies on glued laminated timber (glulam) beams with holes reinforced by means of glass fibre
reinforced polymers (GFRP) are reported by in Hallström and Grenestedt [37],[38]. It was aimed at
reducing the stress singularity in the corner of the hole by changing the anisotropy by means of the
reinforcement. However, the theoretical decrease of stresses at the corner of the hole was considerably
lower than the level of increase of load-carrying capacity observed in the tests. The effect of load
redistribution from the timber to the GFRP was confirmed in comparative FE simulations.
Results of tests on notched glulam beams reinforced by means of GFRP plates are reported by
Coureau et al. [39]. The reinforced notched beams showed higher load-carrying capacity compared to
reference tests on unreinforced beams. Higher load-carrying capacities were achieved in the tests with
an increase in the width of the FRP plates. Failure of the reinforced notched beams was accompanied
by delamination of the FRP plates. Debonding of the FRP laminates was also observed in tests
performed by Jockwer [40]. In the case of perpendicular to the grain reinforcement, debonding
occurred in the lower beam section due to insufficient bond line area, whereas in the tests reinforced
with an angle of 45° to the grain debonding occurred in the upper beam section near the support due
to the high relative stress difference between the timber in compression perpendicular to the grain and
the CFRP in tension in fibre direction. In both types of configuration of the CFRP, it was not possible
to prevent crack initiation in the timber at the notch corner by the reinforcement. The two
configurations are illustrated in Fig. 7.
Fig. 7 FRP reinforcement of notched beams
192
Tests on curved and pitched cambered beams reinforced by means of GFRP are reported by Enquist et
al. [41]. The reinforcement perpendicular to the grain led to a considerable increase in load-carrying
capacity compared to unreinforced beams. Tensile failure in the GFRP and debonding occurred in the
tests. Radial reinforcement repairs for curved bending members can also be accomplished by placing
GFRP rods in oversized holes filled with epoxy PC. Here, the radial stresses are transmitted through
the epoxy PC in shear.
The need for further research on the application of GFRP as reinforcement perp. to grain includes [41]:
 Long-term tests to cover the effects from variation in moisture and ambient conditions
 Tests on full-size beams in order to cover scale effects.
3.3 Bending reinforcement
The reinforcement of timber members in bending using pultruded FRP rods, strips, plates or other
structural shapes and fabric wraps has been the subject of a large number of research programmes
during the last 50 years. The reinforcement can be deployed internally by bonding rods or strips in
grooves cut into the tension and compression faces of the member or externally by bonding FRP plates
to the tension face but not on the compression face due to the risk of buckling failure in the FRP.
Experimental campaigns to investigate the reinforcement of solid timber beams and glulam with CFRP
reinforcement by several researchers [42]-[47] have demonstrated that the use of a small percentage of
reinforcement in the order of 1.5-2.5% can result increases in the bending strength and stiffness of up to
90% and 100%, respectively. In addition, reinforced beams were shown to have less variability in their
properties than unreinforced beams. With increasing amounts of tensile reinforcement, the load-
deflection response becomes ductile due to compression yielding at high strain levels.
Due to the high costs associated with CFRP materials, several investigators [47]-[50] have studied the
flexural reinforcement of timber members with GFRP. As for the case of CFRP, large increases in
strength were reported when using small percentages of reinforcement. However, the increases in
stiffness were less significant due to the lower stiffness of the GFRP material. Decreased variability in
the performance of the reinforced members was found. Limited ductility was achieved, which
increased with increasing percentage of reinforcement. More recently, flexural reinforcement of
timber using basalt fibre reinforced polymer (BFRP) has been undertaken with promising results [52].
3.4 Shear reinforcement
Various studies on the reinforcement and repair of regions with high shear stresses in glulam beams
are reported in the literature. The studies were often performed more in order to compare different
reinforcement techniques and to evaluate the reinforcing effect rather than to validate or verify design
procedures. In general the tests can be roughly separated whether the reinforcing was applied
externally or internally. External reinforcement was commonly made by means of FRP plates, fabrics
or roving. Internal reinforcement was commonly made by FRP rods. The different techniques and
relevant parameters are shown below.
V V V V
β β a1
Fig. 8 Externally applied FRP plates Fig.9 Internal applied FRP rods
193
A good summary of the literature regarding shear strengthening of timber beams by means FRP is
given by Andre [4]. The main focus in this work is put on the properties of different fibre materials.
The behaviour of mechanically connected GFRP plates as shear reinforcement of timber beams was
evaluated by Akbiyik et al. [53]. However, it is not possible to restore the stiffness of the undamaged
beams by this reinforcement method.
An approach for the calculation of the stresses in the timber of a beam reinforced by means of FRP
panels is proposed and evaluated by Triantafillou [5],[54]. The approach is verified in a series of
experiments on small size specimens. The main focus of the study was the required area fraction of
the FRP panels and the relative height of the reinforcement in order to reduce the shear stresses in the
timber. The approach is valid only for undamaged timber beams with no checks in the cross-section
and, hence, not adequate for the repair of structures. As discussed in [5],[54] the reinforcing effect of
the FRP depends on its stiffness in the grain direction of the timber. Hence, the use of multidirectional
FRP fabrics or unidirectional FRP roving with small angle β relative to the grain direction of the
timber are beneficial both for strength and stiffness of the reinforced beam.
The advantages of internal reinforcement by means of FRP rods are the good aesthetics of the wooden
surface of the beam and the possibility of reinforcing beams with reduced accessibility from the sides.
Hence, FRP rods are a good method for repairing beams in situ [55]. A large test series on damaged
beams of railway bridges reinforced by means of GFRP rods installed with β = 90° is reported by
Radford et al. [6] and Burgers et al. [56]. A considerable increase of the shear stiffness of damaged
beams was observed in the tests. Compared to external FRP wraps internal FRP rods show a lower
stiffness but also less material use and an easier installation. The latter point is of special importance
if beams are only accessible from the bottom surface. GFRP rods with β = 90° were tested in small
and medium scale glulam beams by Gentry [57]. A general good reinforcing effect and an increase of
approximately 50% of the load-carrying capacity in shear were observed. The method of reinforcing
beams by means of FRP rods with β = 90° is also referred to as the Z-spiking method.
Tests on glulam beams reinforced with GFRP dowels in shear and bending were performed by
Svecova and Eden [58]. The internal shear reinforcement was applied with different distance a1
between the bars partly along the whole length of the beam and partly only in the region of high shear
force near the support. The rods were installed perpendicular to the grain at β = 90°. It was possible to
increase the load-carrying capacity, to reduce the variability of the beams, to increase the deflection
before failure and hence, introduce some kind of ductility to the beams. A dowel distance equal to the
beam depth is recommended by the authors based on tests.
An example of an application of CFRP wraps in practice was given by Lauber [58]. A rafter was
reinforced in shear by wrapping it with CFRP fabric in a U-shape. The advantage of fabric compared
to plates is better flexibility during installation and its possibility to adapt to other connecting
elements. An exact surface preparation of the existing beam is necessary in order to smooth the
surface and to create good bond between timber and CFRP. It is emphasized that constant ambient
conditions and constant moisture content of the structure are required in order to prevent delamination
and to guarantee the long term strength.
Sonti and Gangarao [60] found FRP wraps to be adequate for increasing the strength and stiffness of
timber beams in infrastructural application. Widmann et al. [61],[62] performed a test series on full
size beam specimen with shear cracks reinforced by means of CFRP roving with β = 45°. A
considerable increase of the stiffness of the damaged beam was observed at that fibre angle.
194
3.5 Pre-stressed FRP

The reinforcing efficiency of FRP materials can be improved by prestressing. In passive or slack
reinforcement, the load-carrying capacity of the FRP is often not reached as failure commonly occurs
in the timber element. In the case of tensile reinforcement of beams, a tensile force is usually applied
to the FRP sheet or rod by means of hydraulic jacks before pressure bonding to the timber element.
The eccentric prestress induces significant compressive stresses in the bottom of the beam, which
oppose the tensile stresses due to the external loads. In this way, the bending strength of the beam is
increased. In addition, higher compressive stresses in the timber under the action of the external loads
give rise to a plastic response. Prestressing of the beam also influences the deformation of the
member. When the jacking force is released, the beam cambers in the upward direction. This
deflection can be offset against the deflection due to the external loads thereby giving an apparent
increase in flexural stiffness.
Several researchers have investigated the feasibility of prestressing timber beams with FRP plates. As far
back as 1992, Triantafillou and Deskovic [63] undertook a small scale testing programme, in which one
timber beam was reinforced with 2.5% prestressed CFRP plate bonded to the beam soffit. A strength gain
of 40% compared to 16% for passive reinforcement was achieved. More recently, Dagher et al. [64] tested
6.7 m long glulam beams with 1% GFRP plate reinforcement bonded to the soffit using a PRF adhesive. A
prestressing force of 30% of the ultimate tensile strength was applied to the FRP plate using hydraulic
jacks. The strength of the prestressed beams was found to be 95% higher than unreinforced beams and
38% higher than passively reinforced beams. The average precamber of the beams was 10.9 mm, which
can be offset against the allowable deflection. Rodd and Pope [65] investigated the behaviour of a GFRP
prestressed glulam beam with a bumper lamination glued to the bottom face of the FRP plate. Lehmann et
al. [65] reported increases in bending resistance of 30% in solid timber beams when prestressed with 0.3%
CFRP. The precamber was about 30% of the allowable deflection.
A key limitation is the fact that the prestressing force needs to be anchored at the ends of the beams.
Load transfer takes place over a short length at the end of the bond line and delamination may occur
due to the local stress concentration. Brunner and Schnueriger [67] described a gradiented prestressing
device developed by EMPA as a possible solution to this problem. In this method, the curing of the
epoxy adhesive is carried out in a controlled fashion using heat to accelerate curing. Starting at mid-
span and gradually moving towards the support while reducing the prestressing force, ensures that the
prestressing force is anchored over a longer length and the force at the end is reduced. An alternative
approach [64] is to release the jacking force immediately after applying the bonding pressure.
Prestressing of FRP using hydraulic jacks may not be
suitable for on-site applications. An alternative method
proposed by Negrão et al. [65] involves the pre-cambering
of the timber before installing the FRP reinforcement as
shown in 9. This is achieved using an adjustable prop
located at the centre of the beam. This approach has the
additional advantage of inducing a triangular bending
moment distribution in the beam due to the prop force
resulting in a low, constant shear stress in the glue line.
This method may not be suitable for large beams as the
Fig. 10 On-site procedure for FRP
prop force required may be too great. prestressing of timber [65]
195
Another method, most recently developed at Chalmers University of Technology, is the stepwise pre-
stressing of the laminate, which greatly reduces stress concentrations at the ends of the laminate,
making mechanical anchorage unnecessary, [67]. This method differs from the gradiented pre-
stressing device developed by EMPA in that it is based on the stepwise introduction of the pre-
stressing force in the laminate, rather than the gradual release of the pre-stressing force. Special
equipment was also developed to pull the laminate in such a way that the entire pre-stressing force is
gradually introduced into the laminate in a discrete or continuous manner by distributing the total
force over ten steps. Numerical and experimental studies show that it is possible to reduce the shear
and peeling stresses in the bond line to levels below 1 and 0.2 MPa respectively, for a pre-stressing
force of 100 kN. These values are well below the shear and tensile strength perpendicular to the grain
of wood and adhesives.
Design models for FRP prestressed timber beams have been presented by a number of researchers.
Brunner and Schnueriger [67] describe a calculation model for the case of FRP prestressed plate
bonded to the bottom of the beam using an iterative approach for the moment capacity and a linear
elastic perfectly plastic model for timber in compression. Triantafillou and Deskovic [63] considered
the same situation but used a bilinear model with a falling branch post-yield for compressive
response. Brady and Harte [69] presented a closed form expression for the moment capacity of
prestressed glued-laminated beams incorporating a bumper lamination covering the FRP, where the
timber in compression was modelled using a bilinear model with a falling branch post-yield.
McConnell et al. [70] presents a theoretical model for glulam members post-tensioned with BFRP
tendons.
The long-term performance of prestressed timber requires further examination before this method can
be recommended for use.
4. Design
4.1 Flexural strengthening
Flexural reinforcement can be placed on the tension and compression faces of the member and can be
in the form of externally bonded plates (EBP) or near surface mounted (NSM) rods, plates or strips.
Externally bonded plates are not recommended for compression reinforcement due to the likelihood of
buckling.
Analysis of reinforced timber flexural members is based on a classical strength of materials approach.
The analysis is based on the following assumptions:
 The member cross-section is symmetric in the plane of bending
 Plane sections remain plane
 De-bonding or slippage does not occur between the FRP and wood
 The FRP material is linear elastic to failure in tension and compression
 The timber is linear elastic to failure in tension and nonlinear in compression
Various constitutive models have been used to model the non-linear behaviour of timber in
compression. The bilinear Bazan-Buchanan model [71] assumes linear elastic behaviour up to the
yield point followed by a falling branch with a negative slope. This model has been found by a
196
number of authors [45],[48],[69] to match well with experimental results. In many cases, data on the
slope of the falling branch is not available and, in that case, a simplified linear elastic, perfectly plastic
model has been used [43],[49]. A quadratic approximation has also been used successfully [46].
In order to determine the ultimate moment capacity, all possible failure modes must be considered. In
a large number of test programmes over the last twenty years, it has been found that failure of the FRP
reinforcement is unlikely to occur and, in practise, only two failure modes need to be considered.
These are:
 Mode 1: Failure of the timber in tension while in compression the response is linear elastic
 Mode 2: Failure of the timber in tension after the onset of compressive yielding
These two scenarios are illustrated below for beams with NSM reinforcement on both faces.
For failure mode 1, the maximum tensile strain in the timber, ε2, reaches the ultimate value while the
maximum compressive strain, ε1, is less than the yield strain.
 2   tu ; 1   cy (4)
where εtu is the ultimate tensile strain for the timber and εcy is the timber compressive yield strain
Fig. 11 Failure mode I Fig. 12 Failure mode 2
The analysis is based on strain compatibility and force equilibrium. Equilibrium of the axial forces
acting on the reinforced section requires that
Ff 1  Ff 2  Fcw  Ftw  0 (5)
where Ff1 and Ff2 are the forces in the compressive and tensile reinforcement, respectively, and Fcw
and Ftw are the total compressive and tensile forces in the wood, respectively.
As all of the strain terms are linearly related, Eq. (5) reduces to
h  hNA  h f 1 hNA  h f 2 h2
  E f  Ew  Af 2 
1 1
E f Af 1    Ewb   h  hNA    Ewb  NA  0 (6)
h  hNA h  hNA 2 2 h  hNA
where the dimensions are defined in Fig. 11, Ef and Ew are the elastic moduli of the FRP and timber,
respectively, Af1 and Af2 are the areas of tensile and compressive reinforcement, respectively, and εf1
and εf2 are the corresponding strain terms. Solving this quadratic equation gives the location of the
neutral axis, hNA. Increasing the axial stiffness EA of the FRP in tension causes the neutral axis to
move down, while increasing that of the FRP in compression causes the neutral axis to move up.
Knowing the location of the neutral axis, the ultimate moment capacity of the reinforced section is
determined by taking moments of the forces about the neutral axis
197
M u  Ff 1   h  hNA  h f 1   Ff 2   hNA  h f 2   Fcw   h  hNA   Ftw  hNA

2 2
(7)
3 3
For failure Mode 2, the maximum tensile strain in the timber, ε2, reaches the ultimate value while the
maximum compressive strain, ε1, is greater than the yield strain (Fig. 12).
Ff 1  Ff 2  Fcw1  Fcw2  Ftw  0 (8)
This again gives a quadratic expression which can be solved to give the location of the neutral axis.
The ultimate moment capacity is then determined by taking moments of the normal forces about the
neutral axis
M u  Ff 1   h  hNA  h f 1   Ff 2   hNA  h f 2   Fcw1  hcy  Fcw2  hcy 2  Ftw  hNA

2 2 2
(9)
3 3 3
The extent of compression yielding in the timber depends on the relative magnitudes of εtu and εcy and
on the amount of tension and compression reinforcement. Increasing the amount of reinforcement on
the tension side leads to more ductile behaviour and increased strength. Ideally, the timber beam
should be unloaded prior to application of the FRP reinforcement. If this is not possible, the existing
strain in the structure before FRP strengthening takes place must be taken into account.
Fundamental to the design process is the selection of appropriate design values for the material
properties. While some countries have developed national guidelines, there is currently no
European Standard specifying the design of FRP structures or FRP reinforcement for timber, steel or
concrete structures. The Italian National Research Council has developed guidelines [2],[3] for FRP
reinforcement of existing structures. Using a limit state approach, the design value of a property is
expressed as
Xd   Xk /  m (10)
where η is a conversion factor to account for environmental and creep effects, Xk is the characteristic
value and γm is the partial material factor. Some of the recommended values for the conversion factor
η from CNR-200 [2] are summarised in Tab. 3. CNR-DT200 provides different partial material
factors for FRP depending on the failure modes and the types of FRP system used. For FRP rupture,
γm is given as 1.1 for certified systems and 1.25 for uncertified systems.
Tab. 3 FRP design conversion factors  [2]
Exposure Loading mode

FRP
InternalExternalAggressive Creep Fatigue
Glass / Epoxy 0.75 0.85 0.95 0.30 0.50
Aramid / Epoxy 0.65 0.75 0.85 0.50 0.50
Carbon / Epoxy 0.50 0.70 0.85 0.80 0.50
198
A major challenge in the design of reinforcement for existing timber structures is the determination of
the properties of the timber for use in design calculations. This is particularly difficult for older
structures. Non-destructive testing carried out during the assessment of the structure prior to
reinforcement is necessary in order to get characteristic values for design. For timber, the design value
of a property is expressed as
X d  kmod  X k n  /  m (11)
where kmod is the modification factor for service classes and load-duration and Xk(n) is the characteristic
value of the property from on-site tests. It has been found by several researchers [47],[48],[50] that
reinforced timber beams fail in tension at a higher stress than unreinforced beams. This is because the
reinforcement in the tension zone bridges defects such as knots and constrains crack opening so that
the timber can carry a higher load before failure. Gentile et al. [48] found that an enhancement in
tensile strength of between 18 and 46% could be achieved depending on the unreinforced timber
strength.
For the serviceability limit state, the materials are assumed to behave elastically. The flexural stiffness
EI, of the reinforced beam can be determined using the transformed section method. Depending on the
situation, a gravity load test may be carried out on the unreinforced beam to measure unreinforced
flexural stiffness. From this the mean modulus of elasticity Ew for the timber parallel to grain and the
modular ratio n can be calculated, where
n  E f / Ew (12)
In the transformed section method, an equivalent section in which the FRP reinforcement is
transformed to an equivalent area of timber is considered. The bending stiffness for the reinforced
section is then found by multiplying the second moment of the transformed section by the modulus of
elasticity of the timber. The maximum increase in flexural stiffness can be achieved by placing half of
the reinforcement on the compressive face and half on the tensile face [43], however, the increases
compared to reinforcing only on the tensile face may not be significant enough to justify the
additional material and labour costs. Increasing the modular ratio will result in greater increase in the
flexural stiffness. An efficient method for investigating different reinforcing configurations is finite
element modelling incorporating nonlinear material behaviour for timber [42],[72].
4.2 Shear strengthening
4.2.1 Shear strengthening by means of external reinforcement
The design of shear reinforcement is different for intact cross-sections without cracks and for
damaged members. In the former case, both the timber and the FRP elements carry the forces whereas
in the latter case only the FRP element carries the load in the regions of the cracked member. As a
consequence the design procedures are different for both cases.
Triantafillou [54] summarized the design procedure for the case where both timber and FRP elements
carry loads proportionally to their stiffness in direction of the beams axis. In addition to Eq. (12), the
area fraction of the FRP αfrp can be used for simplification.
Afrp 2  t  h frp
 frp   (13)
Atimber bh
where the geometric quantities are defined in Fig. 13 and Fig. 14.
199
h/2
t t
hfrp h hfrp
h/2
b b
Fig. 13 Intact timber cross-section Fig. 14 Cracked timber cross-section
The maximum shear stress in the timber, due to the shear force V, can be calculated by choosing the
timber as the reference material:
V Scomp ,max
 timber ,max  (14)
I comp bcomp
where the effective geometrical properties of the composite cross-section are denoted by the subscript
comp. The effective width of the composite cross-section is:
 h 
bcomp  b  2 tn  btimber  1  n

 frp  (15)
 h frp 
The effective moment of inertia of the composite cross-section is:
3
b h 3 2 n t h frp   h frp 
2

I comp    I timber 1  n    frp  (16)
12 12   h  
and the effective static moment of area in the centre of the composite cross-section is:
hh h h  h 
Scomp ,max  b  2nt frp frp  Stimber ,max 1  n frp  frp  (17)
24 2 4  h 
Hence, the stresses in the timber according to Eq. (14) can be calculated as follows:
h frp
V Stimber ,max 1  n h  frp 1
 timber ,max    2
(18)
h
btimber I timber 1  n  frp 1  n  h frp  
h frp  h  frp
 
And consequently the maximum shear stresses in the FRP are:
 frp,max  n  timber ,max (19)
Using Eq. (12) the amount (area fraction αfrp) and properties (n) of the FRP can be adjusted in order to
reduce the shear stresses in the timber and satisfy the shear strength of the FRP. For a good
reinforcing effect FRP roving should be applied with an angle smaller than 90° with respect to the
grain direction of the timber. Widmann et al. [61] reported good performance with an angle of 45°
between FRP fibre and grain of the timber. In addition, the bondline stresses have to be checked.
200
In the situation of a damaged beam by shear cracking in the cross-section centre (Fig. 14), shear is
transferred only by the FRP. For a perfect reinforcement the full shear stress has to be carried by the
FRP. However, in reality the reinforcing effect connecting the upper and lower beam parts will not be
perfectly stiff and, hence, a reduction of the transferred shear forces will be the result. Reinforcement
with low stiffness, like FRP roving perpendicular to the grain, will not be able to transfer loads
between the timber members and will not contribute to the stiffness of the beam. The two beam parts
will act separately as individual members. Hence, an inclined application as described by Widmann et
al. [61] should be preferred. An estimation of the maximum stresses acting in the FRP can be made
assuming perfectly stiff reinforcement, where shear stresses at the crack location are carried in full by
the FRP.
Stress peaks occur in the bond line between FRP and timber in vicinity of the shear crack. Securing an
intact bond line is crucial for achieving the optimal reinforcing effect. Special consideration has to be
paid to the long term behaviour of the bond line under the influence of moisture variations and
changes in ambient conditions. The slenderness of the FRP sheets together with the high stiffness can
cause high tension stresses in the laminate in the case of swelling of the timber or buckling due to
compression in the case of shrinkage of the timber.
4.2.2 Shear strengthening by means of internal reinforcement
Internal reinforcement with FRP rods act locally along the beam axis. In the undamaged cross-section
the distribution of shear stresses is affected only in the reinforced area. For design of undamaged
cross-sections, the internal FRP reinforcement should be neglected. Similar considerations as for
reinforcement by means of self-tapping screws [73] can be made also for FRP rods. Due to the low
relative portion of FRP compared to the timber cross-section the impact of the reinforcement on
stiffness can be considered to be low. Nevertheless, FRP rods can be an adequate measure for
reinforcing damaged beams with shear cracks. The FRP rods should be designed to carry the forces
acting between the upper and lower timber parts. In the literature, most tests have been made with
FRP rods installed perpendicular to the grain, however, an inclination should be chosen in order to
benefit from the high tensile strength of the reinforcing elements. Svecova and Eden [58] observed a
good reinforcing effect with a dowel distance equal to the beam depth in their tests.
4.3 Long-term deformations in timber and composites
4.3.1 Overview
Timber structures are subject to long-term deformations due to creep, which are exacerbated in the
presence of moisture and especially moisture variation, known as mechano-sorptive creep. By
reinforcing timber with material that has superior properties when it comes to short- and long-term
stiffness, the long-term behaviour could be improved and deflection could be reduced. The first
recordings of accelerated creep in wood due to varying humidity conditions were described in the
1960s [74]. These recordings were later verified by performing bending tests on small wood
specimens in both cyclic and constant humidity [75]. Specimens subjected to constant humidity show
an almost constant creep rate. For cyclic humidity the deflections varies with the drying and wetting
cycles, but the total deflection increases for every cycle. For higher stress rates, the rate of deflection
increases significantly in the presence of cyclic humidity.
Generally, the fibres in composite materials are not expected to creep significantly. It is widely known
that the creep for specimens loaded in the fibre direction is negligible [76]. On the other hand, the
resins and adhesives exhibit marked rheological properties (viscous properties) experiencing
201
significant creep, which are strongly influenced by temperature. To improve the long-term properties
of two-component epoxy type adhesives, a study was made where dog-bone shape epoxy specimens
were cast [77]. Half of the specimens were reinforced with 0.5% carbon fibres. These specimens had a
creep rate 38% lower than the unreinforced specimens.
4.3.2 Experimental creep test on solid wood strengthened with composites
Creep testing of unreinforced and reinforced beam specimens was carried out at Chalmers University
of Technology [78]. In the study, 24 specimens were used with the cross-section 45x70 mm and
length of 1.1 m. In order to minimise the variability of the material properties, the specimens were
well defined in terms of the origin of the raw material and sawing pattern [79]. The specimens were
assigned to four different groups with approximately the same mean value and standard deviation in
modulus of elasticity. The different reinforcement schemes are shown in Tab. 4. Two types of CFRP
were used, namely, CFK 150/2000 (CFRP 165) by S&P and Carbodur H514 (CFRP 300) by Sika. No
creep was expected to occur in the CFRP reinforcement or in the steel as it was loaded in the direction
of the fibres. The adhesive used was in all cases S&P Resin 220 Epoxy. In order to allow moisture
exchange on three parts of the cross-section, the tension surface of unstrengthened timber specimens
were sealed with epoxy.
Tab. 4 Creep test specimens and reference beam, sealed with epoxy on the tension side.
Beam types
Timber CFRP 165 CFRP 300 Steel
Strengthening schemes Sealed with epoxy CFK 150/2000* CarboDur H514** Steel
Reinforcement area [mm²] - 70 70 100
E-Modulus [GPa] - 165 300 210
Tensile strength [MPa] - 2310 1350 235
* Sto S&P flexural strengthening product (plates)
** Sika Chemicals flexural strengthening product (plates)
The creep tests were conducted in four-point bending at a stress level of 8 MPa. Directly after
loading of the specimens the climate was set to 90% RH thereafter the climate was changed
between 30% RH and 90% RH in two week cycles, while the room temperature was held
constant at 23°C. The mid-point curvature was measured using a LVDT gauge. The results
showed that the reinforced specimens had a lower initial deflection than the un-reinforced
specimens. This result was expected due to the higher stiffness of the reinforced specimens.
The creep results were presented as relative creep, i.e. the deflection divided with initial deflection
(after 60 seconds). The mean values of the relative creep from all series are shown in Fig. 15.
The variation in relative creep between the wet and dry periods is very high, much higher for the
reinforced specimens than for the unreinforced specimens. For the reinforced specimens the variation
between 90% and 30% RH is of the same magnitude as the initial deflection. This is caused by the
longitudinal shrinkage of the timber material while the reinforcement material does not shrink. This
effect is much larger when one-sided reinforcement was used. The conclusion is that the
reinforcement material should be distributed on both the tension and compressive side of the beams.
The trend lines in Fig. 16 were calculated by taking the sum of the initial value and the final value for
202
each 2 week period. The trend showed clearly that the reinforcement prevents the mechano-sorptive
creep rather well. After the first 4 weeks the increase of the relative creep for the steel and CFRP-300
reinforced specimens is only a fourth of the increase for timber with epoxy.
Fig. 15 Mean relative creep for the four different Fig. 16 Mean relative creep - trend lines for the
reinforcement schemes four different reinforcement schemes
4.3.3 Design aspects
In order to minimise the mechano-sorptive creep, it is beneficial to strengthen both the tension and
compression sides of a timber beam to avoid large moisture-induced movements. EC5 does not provide
kdef factors for FRP strengthened beams. However, the main benefit is due to increased short-term
stiffness of the strengthened beams, but the long-term creep deformations are the same as for
unstrengthened beams. More research is needed to investigate various climate classes and load durations
for certain standardised strengthening system. Calculation of deflection for a reinforced beam (with
about 2% reinforcement) in a residential house showed that it is possible to increase the span length by
as much as 20% or reduce the size of the cross-section compared to an un-reinforced beam.
5. Quality Control for Bonding on Site
The success of the reinforcement intervention depends on the implementation of adequate quality
control measures at each stage of the process. All work must be carried by appropriately experienced
personnel, using suitable materials, procedures and equipment in accordance with the Quality Plan. A
procedure for quality control on site was developed as part of the EU CRAFT Project LICONS [80].
This was subsequently adopted by COST Action E34 [18]. Draft standards for on-site acceptance
testing of mixing and application of adhesives have been prepared by CEN TC 193/SC1/WG11 [7]-
[10].
These groups recommended that the Quality Plan should include the following checks and should
specify the frequency of each check / test.
1. Reception of Materials: A record should be kept of all materials delivered to site. The materials
should be checked to ensure that they match the specification and the expiry date should be
checked.
2. Inspection and tests
a) Elements to be repaired: the timber moisture content should be checked and recorded, the
condition of the gluing surface should be checked and the dimensions and locations of holes,
slots etc. should be checked and recorded.
203
b) Mixing and application of adhesives: the following tests should be carried out in accordance
with the draft standards [7]-[10]
 On-site sampling and measurement of the cure schedule for adhesives
 On-site sampling and subsequent laboratory measurement of the shear strength of
adhesive joints
 On-site sampling and proof loading of the strength of adhesive joints
c) Visual inspection of final repair including dimensional checks.
Based on these tests, the actual properties obtained on site can be compared to the values used in the
design.
6. Outlook and recommendations
Using bonded fibre-reinforced polymer laminates for the strengthening and repair of wooden
structural members has been shown to be an effective and economical method. The high strength and
stiffness, light weight and good durability properties of FRP composites, together with advantages
offered by adhesive bonding, have made it a suitable alternative for traditional strengthening and
repair techniques. FRP materials and adhesives and their properties suitable for various methods are
discussed. It is pointed out that careful surface preparation is essential in order to achieve good bond
strength and durability.
Two-part, cold-cure epoxy adhesives have generally been found to be most suitable for on-site
bonding, as they have good gap-filling properties and low curing shrinkage. It is pointed out that
careful surface preparation is essential in order to achieve good bond strength and durability. Research
on the bond behaviour of FRP to timber is still in its infancy. More research is needed, focusing in
particular on various factors affecting the bond between FRP and wood, including combined
properties and the local bond-slip relationship for FRP to timber, which can be used directly for finite
element analysis.
A design approach, which accounts for the fracture and delamination in FRP strengthening, including
appropriate material models, is presented. Failure-mode-based strength criteria are discussed and the
development made by Puck is shown in terms of the mathematical formulation and strength indices
for the Puck-Knaust failure criterion. The applications presented are related to beam-end
reinforcements, improving tension strength perpendicular to the grain, shear reinforcements and the
pre-stressing of FRP on the tension side of a beam in flexure.
Beam-end reinforcement is a very important repair method for the decayed part of a structure and
specifically the restoration of historically significant structures. An increase in structural performance
of 140% for timber-on-timber supplements has been reported. Studies of the slack reinforcements of
beams in flexure and shear reinforcement are briefly summarised, as studies of this kind have been
conducted for many years and a large amount of research is available. The pre-stressing of FRP
materials on the tension side of a beam in flexure offers the most effective utilisation of these
materials; increasing the load-bearing capacity and pre-cambering of existing beams and, by doing so,
improving the serviceability limit state which often governs the design.
It is recognised that one major challenge when it comes to the design of reinforced timber members in
the ultimate limit state is the lack of appropriate properties of timber and of older timber structures in
particular. The maximum increase in flexural stiffness (in the serviceability limit state) can be
achieved by placing half the reinforcement on the compressive face and half on the tensile face. There
204
are no creep factors for long-term deflection as defined in design codes for timber beams strengthened
in flexure. It is therefore not possible to take advantage of lower mechano-sorptive creep than that in
unstrengthened beams. In general, more research is needed on long-term tests and durability
performance to cover the effects of variations in moisture and ambient conditions. The need for
appropriate quality control for bonding on site is recognised and a quality plan based on various
standards is presented.
References
[1] Hollaway, L., and Teng, J., Strengthening and Rehabilitation of Civil Infrastructures Using Fibre-
Reinforced Polymer (FRP) Composites, Woodhead Publishing Limited, Cambridge, UK, 2008.
[2] CNR-DT 200/2004, Guide for design and construction of externally bonded FRP systems for
strengthening existing structures. Materials, RC and PC structures, masonry structures, National
Research Council. Advisory Committee on Technical Recommendations for Construction,
Rome, 2004.
[3] CNR-DT 201/2005, Guidelines for the design and construction of externally bonded FRP
systems for strengthening existing structures. Timber Structures, National Research Council.
Advisory Committee on Technical Recommendations for Construction, Rome, 2007.
[4] André, A., Fibres for strengthening of timber structures, Master Thesis, Luleå tekniska
universitet, Civil and Environmental Engineering, Structural Engineering, 2006.
[5] Triantafillou, T., “Composites: a new possibility for the shear strengthening of concrete,
masonry and wood”, Composites Science and Technology, Vol. 58, No.8, 1998, pp. 1285-1295.
[6] Radford, D. W., Van Goethem, D., Gutkowski, R.M., Peterson, M. L., “Composite repair of
timber structures”, Construction and Building Materials, Vol. 16, No. 7, 2002, pp. 417–425.
[7] Greco, A., Maffezzoli, A., Casciaro, G., Caretto, F., “Mechanical properties of basalt fibers and
their adhesion to polypropylene matrices”, Composites Part B: Engineering, Vol. 67,
2014, pp. 233-238.
[8] CEN TC 193/SC1/WG11, Adhesives for on site assembling or restoration of timber structures.
On site acceptance testing: Part 1: Sampling and measurement of the adhesives cure schedule,
Doc N20, CEN TC 193, 2007.
On site acceptance testing: Part 2: Verification of the shear strength of an adhesive joint, Doc
N21, CEN TC 193, 2007.
On site acceptance testing: Part 3: Verification of the adhesive bond strength using tensile proof
loading, Doc N22, CEN TC 193, 2007.
[11] Broughton, J.G., Hutchinson, A.R., Review of relevant materials and their requirements for
timber repair and restoration, LICONS CRAF_1999-71216, 2003.
[12] Raftery G.M., Harte A.M., Rodd P.D., “Bond quality at the FRP–wood interface using wood
laminating adhesives”, International Journal of Adhesion and Adhesives, Vol. 29, No. 2, 2009,
pp. 101–110.
[13] Raftery, G.M., Harte, A.M., Rodd, P.D., “Bonding FRP materials to wood using thin epoxy glue
lines”, International Journal of Adhesion and Adhesives, Vol. 29, No. 5, 2009, pp. 580-588.
[14] Davalos, J.F., Qiao, P.Z., Trimble, B.S., “Fiber-reinforced composite and wood bonded
interfaces: Part1. Durability and shear strength”, Journal of Composite Technology and
Research, Vol. 22, No. 4, 2000, pp. 224-231.
[15] Harvey K., Ansell, M.P., “Improved timber connections using bonded-in GFRP rods”, In:
Proceedings of 6th World Conference on Timber Engineering, Whistler, Canada, 2000.
[16] Custodio, J., Broughton, J.G., “Factors influencing bond performance”, In Core Document of
COST Action E34, Bonding of timber, University of Natural Resources and Applied Life
Sciences, Vienna, 2008.
205
[17] Pizzo, B., Smedley, D., “Adhesives for on-site bonding: characteristics, testing, applications”, In
State-of-the-art report on the Reinforcement of Timber Structures, 2015, COST Action FP1101.
[18] Smedley, D., Cruz, H., Paula, R., “Quality control on site”, In: Core Document of COST Action
E34, Bonding of timber, University of Natural Resources and Applied Life Sciences, Vienna,
2008.
[19] Wan J., Smith S., Qiao P., and Chen F., “Experimental Investigation on FRP-to-Timber Bonded
Interfaces”, Journal of Composites for Construction, Vol. 18, No. 3, 2013, pp. A4013006:1-9.
[20] Barbero E., Davalos J., and Munipalle U., “Bond strength of FRP-wood interface”, Journal of
Reinforced Plastics and Composites, Vol. 13, No. 9, 1994, pp. 835-854.
[21] Davalos J., Qiao P., and Trimble B., “Fiber-reinforced composite and wood bonded interfaces:
Part 1. durability and shear strength”, Journal of Composites Technology and Research, Vol. 22,
No. 4, 2000, pp. 224-231.
[22] Lopez-Anido R., Gardner D., and Hensley J., “Adhesive bonding of eastern hemlock glulam panels
with E-glass/vinyl ester reinforcement”, Forest Products Journal, Vol. 50, No. 11/12, 2000, pp. 43-47.
[23] Crews K., and Smith S., “Tests on FRP-strengthened timber joints”, In: Proceedings, Third
International Conference on FRP Composites in Civil Engineering, CICE 2006, Miami, USA,
pp. 677-680.
[24] Silva P., Cachim P., and Juvandes L., “Técnicas avançadas de reforço de estruturas de Madeira
com compósitos reforçados com fibras (FRP)”, In: Proceeding 1st Iberic Congress Cimad’04-
Wood in Construction, Guimarāes, 2004, pp. 613-622.
[25] Juvandes L., and Barbos R., “Bond analysis of timber structures strengthened with FRP
systems”, Strain, Vol. 48, No. 2, 2012, pp. 124-135.
[26] Wan J., Smith S., and Qiao P., “FRP-to-softwood joints: experimental investigation”, In:
Proceedings of Fifth International Conference on FRP Composites in Civil Engineering, CICE
2010, Beijing, China, pp. 951-954.
[27] Lorenzis L., Scialpi V., and Tegola A., “Analytical and experimental study on bonded-in CFRP
bars in glulam timber”, Composites Part B: Engineering, Vol.36, No. 4, 2005, pp. 279-289.
[28] Valipour H., and Crews K., “Efficient finite element modeling of timber beams strengthened
with bonded fibre reinforced polymers”, Construction and Building Materials, Vol. 25, No. 8,
2011, pp. 3291-3300.
[29] Talreja, R., Singh, C.V, Damage and Failure of Composite Materials. Cambridge University
Press, 2012.
[30] Puck, A., “A Failure Criterion shows the Direction – Further Thoughts of the Design of
Laminates”, Kunststoffe German Plastics, Vol. 82, 1992, pp. 29-32.
[31] Puck, A., “Praxisgerechte Bruchkriterien für hochbeanspruchte Faser-Kunststoffverbunde”,
Kunststoffe, Vol. 82, No. 2, 1992, pp. 149-155.
[32] Schober, K.U., Rautenstrauch, K., “On the application of cohesive zone modeling in timber
composite structures”, In: Proceedings of the 10th World Conference on Timber Engineering
(WCTE 2008), Miyazaki, Japan, 2008.
[33] Schober, K.U., Rautenstrauch, K., “Structural Behaviour of Hybrid Timber-Composite Beams”,
In: Proceedings Composites & Polycon 2009, American Composites Manufacturers Association,
Tampa, FL, USA, 2009.
[34] Martin ZA, Tingley DA., “Fire resistance of FRP reinforced glulam beams”, In: Proceedings of
World Conference on Timber Engineering, Whistler, Canada, 2000.
[35] Williamson T.G., “Fire performance of fiber reinforced polymer glued laminated timber”, In:
Proceedings of 9th World Conference on Timber Engineering, Portland, Oregon, 2006.
[36] Metten C.J., Page A.V., Robinson G.C., Repair of structural timbers. Part 2 Fire resistant
repairs, TRADA, High Wycombe, UK, 1993.
[37] Hallström S., Grenestedt, J. L., “Failure analysis of laminated timber beams reinforced with glass
fibre composites”, Wood Science and Technology, Vol. 31, No. 1, 1997, pp. 17-34.
[38] Hallström S., “Glass fibre reinforced Holes in Laminated timber”, Wood Science and
Technology, Vol. 30, 1996, pp. 323-337.
206
[39] Coureau J.L., Cuvillier E., Lavergne C., “Strength of locally PGF reinforced end-notched
beams”, In: Joints in Timber Structures, Aicher S., Reinhardt H.W. (eds) International RILEM
Symposium, 2001.
[40] Jockwer, R., Structural behaviour of glued laminated timber beams with unreinforced and
reinforced notches, PhD Thesis, ETH Zurich, Institute of Structural Engineering, Zurich, 2014.
[41] Enquist, B., Gustafsson, P. J., Larsen, H. J., “Glass-fibre reinforcement perpendicular to the grain”,
In: Proceeding of International Timber Engineering Conference, London, 1991, pp. 3.242-3.250.
[42] Nowak, T.P., Jasienko, J., Czepizak, D., “Experimental tests and numerical analysis of historic
bent timber elements reinforced with CFRP strips”, Construction and Building Materials Vol.
40, 2013, pp. 197–206.
[43] Kliger, R., Johansson, M., Crocetti, R., “Strengthening timber with CFRP or steel plates – short
and long-term performance”, In: Proceedings of World Conference on Timber Engineering,
Miyazaki, Japan, 2008.
[44] Schober K.U., Rautenstrauch K., “Post-strengthening of timber structures with CFRPs”,
Materials and Structures, Vol. 40, 2006, pp.27–35.
[45] Borri, A., Corradi, M., Grazini, A., “A method for flexural reinforcement of old wood beams
with CFRP materials”, Composites Part B: Engineering, Vol. 36, 2005, pp.143-153.
[46] Li, Y.F., Xie, Y.M., Tsai, M.J., “Enhancement of the flexural performance of retrofitted wood
beams using CFRP composite sheets”, Construction and Building Materials, Vol. 23, 2009,
pp.411-422.
[47] Johns K.C., Lacroix, S., “Composite reinforcement of timber in bending”, Canadian Journal of
Civil Engineering, Vol. 27, 2000, pp. 899-906.
[48] Gentile, C., Svecova, D., Rizkalla, S.H., “Timber beams strengthened with GFRP bars:
development and applications”, Journal of Composites for Construction, Vol. 6, No. 1, 2002, pp.
11-20.
[49] Hernandez, R., Davalos, J.F., Sonti, S.S., Kim, Y., Moody, R.C., Strength and stiffness of
reinforced yellow-poplar glued-laminated beams, Research Paper FPL-RP-554, Department of
Agriculture, Forest Service, Forest Products Laboratory, Madison, WI, US, 1997.
[50] Raftery, G.M., Harte, A.M (2011): Low-grade glued laminated timber reinforced with FRP
plate, Composites Part B: Engineering 42(4):724-735.
[51] Alam, P., Ansell, M.P. Smedley, D., “Mechanical repair of timber beams fractured in flexure
using bonded-in reinforcements”, Composites Part B: Engineering, Vol. 40, No. 2, 2009, pp.95-
106.
[52] Pengyi, Z., Shijie, S., Chunmei, M., “Strengthening Mechanical Properties of Glulam with
Basalt Fiber”, Advances in Natural Science, Vol. 4, No. 2, pp. 130-133.
[53] Akbiyik, A., Lamanna, A. J., Hale, W. M., “Feasibility investigation of the shear repair of timber
stringers with horizontal splits”, Construction and Building Materials, Vol. 21, No. 5, 2007, pp.
991-1000.
[54] Triantafillou, T., “Shear Reinforcement of Wood Using FRP Materials”, Journal of Materials in
Civil Engineering, Vol. 9, No. 2, 1997, pp. 65-69.
[55] Alam, P., Ansell, M. P., Smedley, D. “Mechanical repair of timber beams fractured in flexure
using bonded-in reinforcements”, Composites PartB: Engineering, 40(2), pp. 95-106.
[56] Burgers, T., Gutkowski, R., Balogh, J., and Radford, D., “Repair of Full-Scale Timber Bridge
Chord Members by Shear Spiking,” Journal of Bridge Engineering, Vol. 13, No. 4, 2008, pp.
310-318.
[57] Gentry, T. R., "Performance of glued-laminated timbers with FRP shear and flexural
reinforcement.", Journal of Composites for Construction, Vol. 15, No. 5, 2011, pp. 861-870.
[58] Svecova, D., Eden, R.J., ”Flexural and shear strengthening of timber beams using glass fibre
reinforced polymer bars -an experimental investigation”, Canadian Journal of Civil Engineering,
Vol. 31, 2004, pp. 45-55.
[59] Lauber, B., “Ertüchtigung von Holztragwerken - Bespiele aus der Praxis”, In: Proceedings of the
Holzbautag Biel, Biel, 2008.
207
[60] Sonti, S.S., GangaRao, H.V.S., “Strength and Stiffness Evaluations of Wood Laminates with
Composite Wraps”, In: Proceedings of the 50th Annual Conference, Composites Institute, The
Society of Plastics Industry Inc., 1995.
[61] Widmann, R., Jockwer, R., Frei, R., Haeni, R., “Comparison of different techniques for the
strengthening of glulam members”, In: Enhance mechanical properties of timber, engineered
wood products and timber structures: COST Action FP1004 Early Stage Researchers
Conference, Zagreb, Croatia, 2012, pp. 57-62.
[62] Widmann,R., Tannert,T., Frei,R., “Comparison of different techniques for the shear
strengthening of glulam members”, In: Proc. of Second Conference on Smart Monitoring,
Assessment and Rehabilitation of Civil Structures (SMAR 2013), Empa, Istanbul, Turkey, 2013,
8 pp.
[63] Triantafillou, T.C., Deskovic, N., “Prestressed FRP sheets as external reinforcement of wood
members”, Journal of Structural Engineering, Vol.118, No. 5, 1992, pp. 1270-1283.
[64] Dagher, H., Gray, H., Davids, W., Silva-Hernandez, R., Nader, J., “Variable prestressing of
FRP-reinforced glulam beams: methodology and behaviour”, In: Proceedings of World
Conference on Timber Engineering (WCTE 2010), Riva del Garda, Italy, 2010.
[65] Rodd, P.D., Pope, D.J., “Prestressing as a means of better utilising low quality wood in glued
laminated beams”, In: Proceedings of the International Conference on Forest Products,
Daejoen, Korea, 2003.
[66] Negrão, J., Brunner, M., Lehmann, M., “Pre-stressing of timber”, In: COST E34 - Bonding of
Wood - WG1: Bonding on site - Core Document, K. Richter and H. Cruz eds., 2008.
[67] Brunner, M., Schnueriger, M., “FRP-prestressed timber”, In: Proceedings of 8th International
Symposium on Fiber Reinforced Polymer Reinforcement for Concrete Structures (FRPRCS-8),
Patras, Greece, 2007
[68] Haghani, R., Al-Emrani, M. A new method and device for application of bonded prestressed
FRP laminates. Proceedings of the second international conference on advances in civil and
structural engineering, ISBN 978-1-63248-035-4, 2014, Kuala Lampur, Malaysia.
[69] Brady, J.F., Harte, A.M., “Flexural reinforcement of glue-laminated timber beams using
prestressed FRP plates”, In: Proceedings of 4th International Conference on Advanced
Composites in Construction (ACIC), Edinburgh, UK, 2008.
[70] McConnell, E., McPolin, D., Taylor, S., “Post-tensioning of glulam timber with BFRP tendons”,
Proceedings of ICE- Construction Materials, 2014, p. 1-9.
[71] Buchanan, A., “Bending strength of lumber”, ASCE Journal of Structural Engineering, Vol. 116,
No. 5, 1990, pp. 1213-1229.
[72] Raftery, G.M., Harte A.M., “Nonlinear numerical modeling of FRP reinforced glued laminated
timber”, Composites: Part B, Vol. 52, 2013, pp. 40-50.
[73] Dietsch, P., “Einsatz und Berechnung von Schubverstärkungen für Brettschichtholzbau-teile”,
PhD Thesis, Technical University of Munich, Munich, 2012.
[74] Armstrong, L.D., Kingston R.S.T., “Effect of moisture changes on creep in wood”, Nature, Vol.
185, 1960, pp.862-863.
[75] Hearmon R.F.S., Paton J.M., “Moisture content changes and creep of wood”, Forest Products
Journal, Vol. 14, 1964, pp. 357-359.
[76] Lou, Y., Schapery, R.A., “Viscoelastic characterization of a nonlinear fiber-reinforced plastic”,
Journal of Composite Materials, Vol. 5, No. 2, 1971, pp. 208-234.
[77] Miravalles, M., Dharmawan, I., The creep behaviour of adhesives. A numerical and experimental
investigation. Chalmers Uni. of Technology. Master’s thesis, 2007.
[78] Hansson, S., Karlsson, K., Moisture-related creep of reinforced timber. Theoretical studies and
laboratory tests, Chalmers Uni. of Technology. Master’s thesis, 2007.
[79] Perstorper, M., Pellicane, P.J., Kliger, I.R., Johansson, G., “Quality of timber products from
Norway spruce”, Wood Science and Technology, Vol. 29, 1995, pp. 157-170.
[80] Anon., Low intrusion conservation systems for timber structures (LICONS),
http://www.licons.org/, 2006.
208
Nanotechnologies for reinforcement and protection of timber structures: innovative nano-coatings
11
Nanotechnologies for reinforcement and protection of timber
structures: innovative nano-coatings
Tanja Marzi1
Summary
In this chapter the latest technological innovations in nano-structured materials for the reinforcement
and protection of timber structures are presented. Starting from the definition of nanotechnologies
applied in the construction field, the chapter briefly describes nano-materials already existing on the
market, nano-coatings and their wood surface protection functions, classification and compatibility
with the different wood species, focusing on their potential usage and their application in the fields of
architecture, civil engineering and cultural heritage.
The second part of the chapter describes an experimental research on the application of a polymeric
resin reinforced with carbon nanotubes on historic timber structures belonging to cultural heritage sites.
1. Introduction
Nanosciences and nanotechnologies represent a new scientific and technological approach for
manipulating material structure and behaviour at the atomic and molecular scales, where properties
differ significantly from those observed at a larger scale. Descending from the normal scale to the
infinitely small (the prefix "nano" means 10-9), we enter the domain of quantum physics: the use of
nano-particles makes it possible to obtain materials with new chemical, physical and mechanical
properties and to increase the original performance of conventional materials (e.g. carbon atoms
connected to form nanotubes, can produce materials that are stronger than steel) [1]. Nanotechnology
involves multidisciplinary areas of investigation, which in recent years have affected all sectors of
industry with significant economic implications.
For the European Union, nanotechnologies represent one of the major fields of scientific development
for the near future and have been identified as priority areas for European economic and industrial
development. The growing interest in the potential applications of nanotechnology is confirmed at
international level in the strategic documents for the planning of scientific research and technological
innovation funding. In 2008, the European Commission also adopted the "Code of Conduct for
responsible research in the field of nanotechnology." In the absence of a comprehensive structured
legislation, this voluntary code aims at promoting integrated, safe, ethical and responsible research,
recalling some principles (i.e. "sustainability" and "precaution") on which the EU States are invited to
take concrete action.
As nanotechnology is an emerging research field, there is a great debate regarding to what extent
nanotechnology will benefit or pose risks for human health and environment and many research
projects at international level are specifically focusing on these issues.
1)
Arch. Ph.D., Post-doc Research Fellow, Politecnico di Torino, Department of Architecture and
Design, Torino, Italy
209
Hundreds of products containing nano-materials are already in use in many sectors of life (including
health, information society, industry, energy, transport and space). In particular, in the field of
construction and cultural heritage, nanotechnologies are providing a significant boost to innovation in
traditional processes and products. Some applications in relevant international architecture are
contributing to their diffusion. At present, the most promising applications relate to nano-structured
coatings. Together with the common performance requirements (long term stability, durability and
weather resistance, good adhesion to the substrate, transparency, sustainability of the production
process, etc..), they introduce additional functionality such as self-cleaning, photocatalysis, water
resistance, fire resistance, scratch resistance, graffiti resistance, antibacterial coatings [2,3].
Currently, the nanotechnology industry is rapidly developing in the field of construction and cultural
heritage [4]. However, it should be highlighted that most research has been directed at materials such
as concrete and metal. To date, several ready-to-use nano-products for wood protection are available
on the market but very little research has focused on the reinforcement of timber structures. The last
part of this chapter focuses on this specific topic, describing also experimental research carried out at
Politecnico di Torino on the application of a polymeric resin reinforced with carbon nanotubes on
historic timber structures. The research addressed the definition and assessment of a methodology for
preparation and application of the nano-composite. The aim was to verify whether there was an
increased mechanical resistance in comparison with traditional reinforcement methods.
2. Nanotechnologies for wood in Cultural Heritage

Due to its anatomical features, wood is considered a natural nano-structured composite material,
anatomically similar to strong piping bonded with a thermoplastic matrix, the lignin, and equipped
with strong dissipative capacity with regards to fracturing energy (Fig. 1). It can be seen as a
polymeric composite of cellulose, hemicellulose, protein and lignin, as well as, at the nanoscale level,
a cellulosic fibrillar composite [5]. This natural nano-composite can potentially offer important
applications in the field of nanotechnologies mainly as nano-materials derived from forest products
(i.e. nano-cellulose, cellulose nano-composites) or as nanotechnology incorporated into traditional forest-
based products (i.e. nano-coating to enhance wood durability) [6].
Wooden cultural heritage is widespread, with different formal, colour and structural features that make
it unique. Architectural heritage built in wood is an important sector of our cultural heritage, including
different typologies (such as floors, roof trusses, bridges, etc.). On the topic of their conservation it
has become widely accepted that such structures should be maintained and preserved, with
interventions respectful of their original conception, but also of their material: wood [7, 8].
Wood frequently discolours as a result of exposure to ultraviolet (UV) light, moisture and bio-
organisms. Treatments introduced by nanotechnologies can improve its durability or stability, which
are negative factors associated with timber exposure to the environment.
Despite the highly innovative character of these techniques (some still experimental), recent
international case studies of applications of nano-structured materials on historic timber structures
belonging to cultural heritage sites are beginning to be known (Fig. 2).
210
a)
b)
Fig. 1 Different levels of observation of wood (a) (Artwork by Mark Harrington, copyright University
of Canterbury, 1996) and variability of wood structure at different scales of observation (b) (re-
elaborated from R.Zanuttini)
a) b)
Fig. 2 Examples of nano-coatings application on relevant architectures: pavilion at Hakone Open-air
Museum (a) and timber structures of traditional Japanese homes (b) (from [13])
3. Nano-coatings for wood protection

Nano-coatings can improve the performance and functionality of wood, extending its stability, which
is often a factor that limits its use. On the other hand, it is noted that traditional products and
211
treatments available on the market, aimed at maintaining wood durability, are often highly toxic for
humans and for the environment. There is therefore the need for new non-toxic products.
Nano-materials are usually incorporated into coatings, in an aqueous, organic or polymeric medium
and different nano-impregnation of timber is possible. An essential aspect is that nano-particles are
well dispersed in a suitable medium to avoid aggregation. Nano-particles are less than 100 nm in size
and they protect or enhance the properties of the substrate. Their high ratio of surface area to mass
ensures that a loading of only a few percent by weight in coatings can significantly enhance chemical,
thermal and physical properties [9].
A survey was carried out on nano-based coatings specifically addressed to wood that are already
present on the international market [10]. Among the main fields of application of these products there
is water repellency, followed by UV rays protection (Fig. 3). It should be highlighted that there is a
wide choice of nano-particles which offer different functionality or multi-functionality. Some nano-
materials, such as silica (SiO2) and titania (TiO2), are multi-functional and there is frequently a choice
of nano-materials for a given application.
There are also health and safety issues associated with the abrasion and wear of nano-materials,
releasing potentially hazardous particles into the atmosphere [11]. In-life and end-of-life health and
safety factors should be considered in developing nano-materials for timber [9].
Fig. 3 Main fields of application of nano-based coatings addressed to wood already present on the
market: waterproofing (Clay, SiO2, CeO2, TiO2); UV protection (TiO2, ZnO, SiO2, CeO2, Fe2O3,
clays); biotic decay protection (Ag, ZnO, Cu); fire resistance (SiO2, Clay, TiO2); self-
cleaning/photocatalitic (TiO2, ZnO); anti-scratching, hardness (TiO2, Al2O3, SiO2, clays, lime) (from
[10])
3.1. Waterproofing
Hydrophobic (water-repellent) nano-coatings are among the promising developments for facade
treatments and weather protection of timber. These products mainly consist of silica nanoparticles, by
exploiting the so-called "lotus effect", properties of water repellence and at the same time
breathability of the surface can be obtained (Fig. 4). It is important to highlight that this kind of
nanostructured coating, which is nearly VOC-free and based on a water-borne silane system, has good
water repellent functions but, essential for wood, it is open-vapour.
212
Fig. 4 Waterproofing and “lotus effect” on timber surfaces treated with nano-coatings based on silica
nanoparticles. In the central image it is possible to notice the difference between untreated wood (top)
and wood treated with nano-coating (bottom) (from [12]-right, [13]-left and T. Marzi centre])
Fig. 5 Durability test of pine board impregnation after 1 year: board protected with nano-coating
(left) and unprotected board (right) (from [12])
Fig. 6 Durability test of larch wood used for outdoor furniture in the Alpine area impregnated with
different UV absorber coatings (2 traditional and 4 nano-coatings). In the image is visible the
difference after 2 months of outdoor exposure. (from [21])
213
3.2. UV Protection
Ultraviolet rays can change wood structure in the sub-surface layer, causing lignin degradation and
discoloration. A nanoscale protective barrier can increase resistance to ultraviolet radiation, preserving
the appearance and the original colours of wood, which can be very important when dealing with
cultural heritage (Fig. 5).
The small size of the particles makes it possible to offer high protection without affecting the
transparency of the coating. The water based nature of this typology of products, which are volatile
organic compound (VOC) free, is a functional and environmentally friendly solution against rot and
moss build-up [12]. Water absorption is significantly reduced as well. By using this technology, the
product reflects the most damaging waves of the sun’s spectrum, protecting wood for a longer time
from solar radiation.
Furthermore, nanoscale solutions are able to concentrate more active substances in a smaller volume
of liquid. Therefore, it is possible to use less product to achieve better durability [12].
Nano-scale titanium dioxide is a well-known catalyst and photocatalytic material. The associated
literature is extensive [11, 14, 15], analysing its applications as a self-cleaning, UV-absorbing, or
sterilising agent. A key factor is that titania is not degraded by the oxidative reactions and its
photocatalytic action is perpetuated. Furthermore, nano-titania has a very high ratio of surface area to
weight, enhancing its photocatalytic action. Recent research has examined the photochemical stability
of water-based acrylic paints containing anatase and rutile nano-titania [16]. Research carried out at
the Université Laval Québec investigated the properties of nano-coatings using zinc oxide, silica,
alumina and titanium dioxide active nano-particles dispersed in acrylic-based, waterborne, solid-
colour stains for exterior wood [17, 18, 19].
A specific research carried out at the Department of Architecture and Design of Politecnico di Torino
(EU Project Alcotra 2007 “Savoir Bois”) has studied the durability of wood surfaces of different wood
species typical of the Alpine area exposed to outdoor extreme conditions (Fig. 6). Wooden outdoor
furniture realized by local artisans of the Monviso valleys were coated with traditional UV absorbant
coatings and nano-coatings with a view to assessing the best solution that would require the minimal
maintenance. Natural weathering tests were carried out in outdoor exposure in the mountains for a
period of two years [20]. Best results were achieved with a polyurethane resin modified with silica
nanoparticles from sol-gel.
3.3. Biotic decay

Improved protection from biotic decay of wood can be obtained with the use of nano-materials used
as a protective treatment and as a barrier to moisture. In particular, products containing silver
nanoparticles are particularly effective for their antibacterial, anti-microbial and anti-mould properties
and nano-silver is nowadays a constituent of several commercial wood preservatives.
Wood can be attacked by a variety of insects, fungi and other organisms. These decay factors can
successfully be tackled by using preparations based on nano-silver. Specimens coated with nano-silver
214
leave active nanoparticles on the surface. When bacteria come into contact with the protected surface,
the silver binds within the bacteria and causes their enzymes to breakdown, stopping the bacteria and
causing them to die (Fig. 7).
Silver is a well-known biocidal agent and it has been available for over 100 years in colloidal nano-
form [22, 23]. Silver-containing compounds which release silver ions into a moist environment differ
from nano-silver particles, which may act as a catalyst and are often held in a polymeric matrix as a
coating. The release of nano-silver by leaching from paints [24] is often associated with conversion
into less toxic compounds such as silver sulphide, so health risks are reduced.
Nano-silver has been widely utilised to reduce the biodegradation of wood and also to improve
resistance to termite attack [25, 26].
Fig. 7 Scheme of silver nanoparticles action. Fig. 8 Durability test carried out on wood
composite panels protected with nano-silver
coating and unprotected (from [12])
Fig. 9 Test carried out on wood treated with nanostructured coating to increase fire resistance (top
wood) and untreated wood element (bottom) before and after fire action (from [29])
3.4. Fire resistance

Fire safety is an important concern in all types of construction and nanostructured coatings can be
effective also as fire retardants. When heated, wood undergoes thermal degradation and combustion to
produce gases, vapours, tars and char. In order to improve the resistance to fire, timber products are
commonly treated with fire retardants that are typically coated onto the surface of wood (painted,
spayed or dipped), or impregnated into the wood structure using vacuum-pressure techniques or other
techniques as plasma treatments. Fire retardants can usually provide thermal insulation, absorb the
surrounding heat by endothermal reactions, or increase the thermal conductivity of wood in order to
dissipate the heat from the wood surface [27].
215
Traditional fire-retardant treatments, such as inorganic salts, may affect wood in various ways by:
increasing hygroscopicity, reducing strength, and because they lead to dimensional stability changes
depending on treatment, wood degradation, corrosion of metal fasteners, adhesion problems,
increased abrasiveness, leaching of treatments. They also contribute to the production of smoke and
gases toxic and dangerous for people’s health [28]. The growing awareness of environmental issues
and consumer safety of fire retardant products means that some traditional products (i.e. boron and
formaldehyde) are likely to decline. The toxicity of fire retardants plays an important role in health
and safety legislation.
A wide range of fire retardant treatment systems for wood, including nano-composites and layer by
layer applications, have been studied throughout recent years and many others are under development
(Fig. 9).
Fire resistance can be increased through the use of high performance thin film fire retardant coatings
based on nanoparticles of titanium dioxide and silicon dioxide. These products, that may be sprayed
using standard airless equipment, have been formulated to retard the flame spread across a wide
variety of materials and also to suppress the generation of smoke. The two-part epoxy coating creates
a highly adherent fire retardant barrier on wood. In the event of fire, it produces water and gases,
which snuff out oxygen and provides a cooling effect at the flame front. A dense char is formed,
which further protects the surface from combustion [26].
The high thermal conductivity of nano-silver coatings has also been tested to improve heat transfer in
wood and enhance fire resistance [28]. Nano-silver treatment clearly showed potential in improving
some of the fire-retarding properties in solid wood products. Such coating may delay thermal
degradation and carbonization by reducing the accumulation of heat by transferring it. Other
important aspects of fire retardant nano-coatings should still be studied in detail, such as toxic gases
and smoke as well as hygroscopicity.
3.5. Deacidification
Lime has been used traditionally for the treatment of wood in order to provide anti-microbial
properties. It acts as a reflective coating and enhances aesthetic qualities. Nano-limes have been
developed and patented at the University of Florence for the consolidation and conservation of
limestones and lime-based wall paintings [30, 31]. Other researchers [32] have examined the role of
nanoparticles of calcium and magnesium hydroxide and carbonate for neutralising acidification
processes in paper and wood.
Specific research concerning the treatment of the wood was dedicated to the conservation of Vasa
warship (Fig. 10) with nano-limes.
The almost complete hull of the Vasa was recovered from Stockholm harbour in 1961 after 333 years
of immersion. The wood was treated initially with polyethylene glycol to prevent shrinkage, but
subsequent acidification resulted from penetration of hydrogen sulfide in the atmosphere catalysed by
iron elements present in the timber.
A dispersion of nano-lime particles in 2-propanol was impregnated into the surface of oak and pine
samples from the ship and controlled deacidification took place with the formation of calcium sulfate.
Excess nanolime was converted to nano-calcium carbonate, which also acted as alkaline reservoir for
deacidification [33].
216
Recent work at the University of Bath has

examined the effect of application of nano-lime
to the end-grain of Welsh woods, which are used
in flooring applications to improve indentation
hardness [9]. The anti-microbial and
antireflective qualities of nano-lime together with
its deacidifying and breathable qualities could
potentially be the basis for multi-functional
coatings for wood.
Fig.10 Vasa warship in Stockholm

(photo P. Isotalo)
3.6. Environment and Nanotechnology

Nanosized particles could replace current chemical treatments. It was shown that preserving wood
with direct impregnation of titanium dioxide, zinc oxide and other particles improves wood durability.
This will become an important issue especially considering that many countries are already banning
current preservative-treatments of wood that are considered as toxic [5]. Surface treatment nano-
coatings can also significantly reduce the use of chemical cleaners and maintenance costs. As many of
the coatings are water based, they are solvent and VOC free. Furthermore, in contrast to water soluble
biocides, nanoparticles have the advantage that they can be fixed and bounded more effectively in the
coating matrix, and the threat of leaching into the environment is lower, with supposed lower
ecological risks. Therefore, nanostructured coatings have the ability to reduce energy costs and
improve indoor air quality.
The use of engineered nano-materials offers advantages as well as disadvantages from a sustainability
perspective. Life Cycle Assessment (LCA) is a suitable method to assess the environmental
performance of a product or process. A comprehensive material LCA is necessary in order to
understand potential risks [34]. But so far studies applying LCA to the area of nanotechnology have
been scarce, probably because LCA frameworks have a list of issues that need further precision and
investigation in order to be applicable to nanotechnologies [35]. A critical aspect is that at present
there are no long term studies available, even though many research projects (e.g. EU 7FP
“NanoHouse” Project) have been specifically addressed to the assessment of potential environmental
impact of nanotechnologies and nanoparticles’ release from nano-based coatings [36]. Other studies
have been focused on techniques for durability assessment, such as the use of the nanoindentation
technique in conjunction with atomic force microscopy for durability investigation and service life
prediction of nanocomposite coatings for wood [19]. Further research to assess and model
environmental impacts on wood surfaces and timber structures is still needed.
217
4. Next-generation fields of application

Concerning future sectors of application, there is ongoing international research in the following
areas:
4.1. New wood-based composite materials
Nanostructured materials should be applied to the preservation and durability not only of wood but
also of its derivates that nowadays are considered as the sustainable future of this material. New
nanostructured adhesives and resins would allow several improvements. Some examples include an
enhancement of thermal insulation properties, photovoltaic properties or the integration of monitoring
devices directly deposited within the composite. With the incorporation of nanoscience into wood-
based building materials, a new generation of multi-functional, high-performance, low maintenance,
durable building materials and components could be achieved [6].
Several research projects have been applied to the study of wood/polymer/nano-clay composites
obtaining improvements in wood properties, such as surface hardness, modulus of elasticity and
water-repellency (with the addition of hydrophobic nanofillers into the wood structure) [37].
4.2. Monitoring / Maintenance
Monitoring of building heritage is a sector of great interest which foresees the use of devices related
to nanoscience and nanotechnology (e.g. Nanoelectromechanical systems or NEMS). These devices,
already used in other sectors (e.g. medical sector) have the advantage of collecting information and
storing it in-situ for a long period, resulting in improvements in the efficiency of the management,
maintenance and rehabilitation process. Such approaches can be adapted to assess the structure
condition and to produce a "good" or "bad" condition certificate of the construction. Maintenance
actions (if applicable) and/or rehabilitation actions will be based on this information. These shall also
provide the basis for monitoring after the rehabilitation interventions. The increasing requirements
towards safety and security, demand for the application of advanced sensors with greater sensitivity,
better specificity, of smaller sizes, lower power dissipation and advanced remote monitoring, without
expensive and time consuming supervision can be satisfied (Fig. 11).
Nanotechnology is a driving force in the development of advanced sensors, concerning not only the
preparation and investigation of smart nano-sized materials for sensor applications (i.e. nanotubes,
nanowires, …), but also the combination of their performance with integrated circuit (ICs), micro and
nano-optics, Microelectromechanical systems (MEMS) and NEMS, leading to higher levels of
integration, and more effective processing and transmitting of the sensor signals. Recently, nano-
sensors have been tested for relative humidity and temperature monitoring under controlled
environmental conditions. Nanocomposites have been prepared and used to realize prototype devices
proposed for use in art conservation and museum applications [38]. With specific respect to timber,
there is great potential to adopt and integrate existing early-stage nanotechnologies and monitoring
devices with timber structures in mind (e.g. for early detection of decay, monitor moisture content,
sense UV degradation).
218
There is ongoing research focusing on the use of

multifunctional wood-adhesives for structural
health monitoring. Tests have been carried out
with functionalized carbon nanotubes used as
fillers in epoxy resins that can increase the shear
strength and fracture toughness of the composite
together with interesting conductive properties.
Since most of structural timber is glued together
from lamella, the electrical modified wood-
adhesive could be integrated in the production
process of composite timber [39, 40, 41].
Recently, there has been a focus on the
development of civil structures that have
embedded sensors and on-board data processing
capabilities. The fusion of these smart
technologies into infrastructure (e.g. bridges) is
intended to provide better and more timely
information on structural behaviour and
maintenance [42].
Fig.11 Example of scheme of smart-monitoring for
timber structures
4.3. Reinforcement
The application of composite nano-materials for reinforcement of timber structures is a technique not
yet widespread. Some research has been conducted in recent years on the wooden elements of new
constructions [43]. Other research has been focusing on the assessment of epoxy adhesive modified
with different nano- and micro-particles for in-situ bonding of FRP timber connections [44]. The
application of composite nano-materials to reinforce wooden structures is a technique that is not yet
available on the market. It requires therefore to be fully tested before being applied on a large scale. A
research project conducted at Politecnico di Torino and specifically focused on the reinforcement of
timber structures in reference to the use of these methods in the consolidation and reinforcement of
existing timber structures is described in Section 4.4.
4.4. Reinforcement of historic timber structures with carbon nano-composites
4.4.1. Introduction to the research
This part of the chapter describes the experimental investigation carried on the application of a
polymeric resin reinforced with carbon nanotubes (CNTs) for localised interventions on historic
timber structures [10, 45, 46].
The study regards the conservation of a specific wooden joint belonging to traditional constructions
and defines and assesses a methodology of preparation and application of the nano-composite. The
aim was to verify whether the mechanical resistance increased in comparison with traditional
reinforcement methods. Different wood species were considered and the experimentation was carried
out with a view to possible in-situ applications of the technique. Laboratory tests were carried out first
on small wooden samples, and afterwards on full-scale wooden elements [47].
219
The use of nano-materials and composite materials to strengthen wood has been suggested recently by
several researchers who have studied the effects of carbon nanotubes-based composites on the
mechanical characteristics of reinforced wooden elements [43]. In Italy, given the greater relevance of
rehabilitation and conservation, research has focused on the methodologies for consolidating existing
historical structures.
The development of polymers reinforced with nanoparticles is one of the most promising approaches
in the field of future engineering applications. The unique properties of some nanoparticles (carbon
black [43] and CNTs) and the possibility of combining them with traditional reinforcing elements
(glass fibre, carbon fibre or Kevlar) have generated an intense research program in the nano-
composites sector [43, 48, 49]. CNTs have a diameter of several nanometres (1-50 nm) and their
length measures several microns (up to 10 μm); they have very good potential for improving the
electrical and mechanical properties of polymers, even with addition of 0.1% weight content
compared to epoxy resin [43- 50]. The difficulties lie in transferring these remarkable mechanical,
thermal and electrical properties to the polymer matrix. Consequently, the correct dispersion of CNTs
in the polymer and timber interface is a crucial factor [43, 51]. The difficulty in dispersing CNTs
arises from their inherently hydrophobic nature and tendency to agglomerate due to their size and
shape [51]. Different techniques for dispersing them in solvents (acetone, ethanol, …) in addition to
ultrasound, mechanical shaking or a combination of the two techniques, have already been tried
[43,52]. The interface adhesion can be improved by chemically functionalizing the CNTs surfaces;
this generates strong covalent-type of bonding [48, 49, 53, 54, 55].
Referring to the experimentation on wood, CNTs potentially provide a number of advantages:
- they are morphologically and chemically compatible with polymer resins used as bonding materials,
and also with wood. This is due to the fact that CNTs are anatomically similar to strong pipes bonded
with a thermoplastic matrix and equipped with strong dissipative capacity with regards to fracturing
energy;
- they allow the polymer bonding matrix to improve its own inbuilt deformation capacities;
- resin-fibre compounds have excellent mechanical characteristics due to the high specific resistance
of the fibres, which ensure great cohesive strength combined with high ductility. This produces
significant creep resistance within the composite;
- the tubular structure of nanofibre has great permeability to vapour: this is a crucial characteristic
when dealing with large surfaces treated with glue, especially in the case of wood, since any
accumulations of humidity must be easy to disperse in order to avoid biotic degradation;
- CNTs have excellent conductive properties.

The objective of the study is to set up a nano-material, based on polymeric resin reinforced with
CNTs, for the reinforcement of timber structures. The new material could be used as adhesive, surface
consolidant (like in the present case) or as a deep impregnation consolidant in combination with
vacuum techniques [47].
4.4.2. Materials and methods
The first step of the experimentation assessed the efficiency of different impregnation techniques.
CNTs were dispersed by means of an ultrasonic probe for different lengths of time in ethanol and
220
acetone. These solvents are supposed to act as a transport medium to bring CNTs directly into the
wood microstructure. Wood specimens with section size of 2.5 cm x 2.5 cm were sunk into the
suspensions for different times (Fig. 12). Only natural capillarity action was exploited rather than
vacuum or other pressure assisted impregnation techniques. This choice was made in view of future
on-site application.
Fig. 12 Impregnation test at different times: progression of the solution absorption (from [45])
The following wood species were selected for the analysis: Fir (Abies alba Mill.), Douglas-Fir
(Pseudotsuga menziesii), Oak (Quercus robur) and Larch (Larix decidua Mill.). Some samples were
obtained from new timber, others from 18th century structures (oak). Multiwall carbon nanotubes
(MWCNTs) were used for the process. For economical reasons, the Nanocyl® series 7000 [56], not
functionalized (Fig 13), was used to set up the process, while series 3101 functionalized with carboxyl
groups were used for mechanical tests (Nanocyl®-3101 series are purified to greater than 95% carbon
and then functionalized with COOH groups).
a) b) c)
Fig. 13 TEM image (a) and CNTs characteristics of Nanocyl® series 7000 (b) (from [56].
In the subsequent phase a two component solvent-free epoxy resin (Mapei Epojet®) [57] was used for
the dispersion in order to obtain a product for mechanical improvement that could be applied on the
wood surface by painting, or act as a reinforced glue to connect different timber parts. The pre-
measured portions (Part A = resin and Part B =hardener) must be mixed together before using them.
The mix ratio between part A and B was 4 to 1. This resin was selected because of its low Brookfield
viscosity (respectively 500 and 320 mPa.s for part A and B). The efficiency of the impregnation
procedure was assessed using the scanning electron microscope (SEM). The overall mechanical
improvement of the timber specimen was evaluated through a comparison of mechanical tests
performed on untreated samples and on samples impregnated with resin and with CNTs resin coating.
The bending strength was evaluated by standardized three point bending tests on small wooden
samples, and afterwards on full-scale timber elements [47].
221
4.4.3. Carbon nanotubes dispersion

In the first stage, the possibility of deep impregnation exploiting natural capillary action in wood has
been investigated. Four different suspensions have been prepared with different CNT (Nanocyl®
series 7000) concentrations and different solvents (ethanol and acetone). Each wood sample was sunk
into these suspensions for twenty-four hours. The specimen was cut and observed with a SEM in
order to assess the efficiency of impregnation. The image was compared with untreated wood (Fig.
14a).
These tests proved that although the dispersion in the solvent was efficient, and the dimensions of
CNTs were compatible with wood porosity, the overall impregnation technique was not working
satisfactorily [47], since in some of the dispersions the CNTs remained agglomerated (Fig. 14b).
a) b)
Fig. 14 SEM image of larch sample (a); Clustering of CNTs in first dispersion (b) [47]
Therefore, in a second stage, the experimentation focused on the dispersion of CNTs in polymeric
resins: a two-component, low viscosity resin was selected for use for CNT dispersion, with the aim of
having a product that could be applied on wooden surfaces by painting. The main difference between
epoxy resins and the solvents described above pertains to its higher viscosity. Therefore, it was
necessary to find the optimal CNTs content assuming a two hour sonication duration. A CNT weight
content equal to 0.3% with respect to the resin (A+B fraction) was adopted after some trials. It was
decided to disperse first the CNTs in the resin component B, which has the lower viscosity. Then the
two components were mixed together with mechanical stirring at room temperature. The mixture was
allowed to cure for six hours at 60°C. Unfortunately, the modified procedure was not sufficient to
avoid CNTs clustering [58].
4.4.4. Mechanical characterisation of small scale specimen
The overall mechanical improvement of the timber specimen has been evaluated by comparison of
mechanical tests performed on untreated samples, on samples impregnated with resin and on samples
with CNT resin coating. The bending strength was evaluated with standardized three-point bending
tests carried out with displacement control (Fig. 15).
222
a) b)
Fig. 15 Detail of the three-point bending test (a); Load displacement diagrams for four different un-
reinforced timber species (b) (from [47])
All the coated samples were cured at room temperature for seven days, according to the
manufacturer’s recommendations, and to simulate an in-situ intervention. The preliminary results
obtained from the samples tested are in some cases very promising, but in other cases more
ambiguous. Future investigations must be carried out on a larger number of specimens to assure
statistical significance of results. In the case of the 18th century oak samples (Fig. 16a), an increase of
about 19% in the peak load was observed when only resin is used for the coating. If CNTs reinforced
resin is used for the coating, the gain in the peak load increases to 42%. On the other hand, as far as
the Hemlock-Fir samples is concerned (Fig. 16b), an increase of about 45% is obtained in case of
resin coating, regardless the presence of a CNTs reinforcing in the resin.
a) b)
Fig. 16 Load displacement curves for the 18th cent. Oak samples (a) and Hemlock fir samples (b).
Comparison between un-reinforced, resin coated, and CNTs reinforced resin coated(from [47])
The results concerning the assessment of the timber retrofitting with CNTs resin coating are positive,
since these provide an increase in the flexural strength [47].
4.4.5. Tests on structural elements
The methodology was extended to timber structural elements coming from historical structures. Old
decayed timber components of Poplar wood (Populus sp.), belonging to a dismantled timber floor of
the Venaria Reale castle (Italy) have been selected for the experimental campaign. In particular, those
timber elements were strongly affected by brown rot decay and xylophagous insects attack (Anobium,
etc.). Both significantly decreased the mechanical resistance of the wood.
223
Reinforcement and conservation interventions often require the improvement of the load-bearing
capacity of a highly decayed timber component, and often partial substitutions are necessary (e.g.
beam heads). A common intervention is to put in place a new wood prosthesis, which must be
connected with the old, possibly reinforced, remaining part of the structure. Therefore, the effect of
strengthening with CNTs resin impregnation has been studied in combination with the placing of new
wooden prostheses, and tests were carried out on nine composite samples [10].
The numerical analysis was also able to identify the weakest point of the joints, where most often a
crack initiates and propagates [47].
The main connection typologies of timber joints belonging to the historical constructive technique
suitable for rehabilitation of timber elements were considered. After a comparison among the
suggested solutions of historical treatises; the so-called “Jupiter” joint was selected. This joint,
traditionally used to connect two elements subjected to bending, is based on a particular geometric
construction that permits the connection of two elements with a profile very similar to a thunderbolt
(Fig. 17). A protocol for collecting data has been defined, which is particularly important when the
experimental phase is applied to historic timber structures, as e.g. temperature, moisture content,
drilling resistance, and tensile strength. For this last parameter, a preliminary finite element analysis
was performed, accounting for the presence of interfaces, and assessing the magnitude of the initial
stresses that take place when the joint is connected.
The same impregnation method used in the first testing phase on small scale specimens was adopted.
Functionalized MWCNTs (Nanocyl® 3101) were adopted for the impregnation of the decayed timber
elements, dispersed in the same bi-component polymeric resin with low viscosity (used in the measure
of 0.75 wt % in respect of the resin).
Fig. 17 Summary of the combinations adopted Fig. 18 Comparison between the resistance drilling
for three point bending tests on timber profiles of the same specimen (specimen n. T1/C3)
“Jupiter” joints (from [10]) before and after the impregnation with CNTs reinforced
resin (from [10])
224
Special efforts were dedicated to obtain a product that could be applied by brush in case of in-situ
application.
All the “Jupiter” joints were first characterized by visual grading and non-destructive testing
(measurement of the drilling resistance). The characterization was also conducted with destructive
mechanical tests. Wood moisture content was 13±2%. Non-destructive characterization was repeated
after the impregnation procedure by painting. The painting was applied assuming a constant amount
of material applied on each sample. Fig. 18 shows a comparison between the resistance drilling profiles
prior and after the impregnation.
It was possible to detect penetration levels of the CNTs reinforced resin in the wood of ~5 mm. It is
reasonable to assume that, based on the state of degradation of the wood and on the amount of
painting, much higher penetration depths can be achieved.
Destructive four-point bending tests were carried out to assess the mechanical resistance of the
samples. The test geometry was selected to have uniform bending in the joint region. Different
combinations (Fig. 17), with or without the impregnation of polymeric resins and of polymeric resins
reinforced with CNTs, were considered, in order to verify whether there was an improvement of the
structural efficiency. As the tests represent a rehabilitation intervention, realized through an old
wood-new wood prosthesis, part of the joints were made with new wooden elements of the same
wood species (Poplar) (Fig. 19).
4.4.6. Discussion of results
Despite the complexity of variations that influence the results of the experimentation (wood species,
typology of joint, degree of decay), the tests carried out on structural elements have confirmed the
hypothesis of an increase of mechanical performance.
The load displacement curves obtained from the beams connected with “Jupiter” joints have
demonstrated an improvement of the mechanical resistance of the samples impregnated with resin and
CNTs, both compared to decayed wood (from 25% to 35%) and wood impregnated with resin only
(Fig. 20).
The applied methodology can constitute a basis for the definition of protocols for use in practice and
recommendations for standard testing and measurement procedures for polymeric resins reinforced
with CNTs for the rehabilitation of timber structures. The proposed intervention technique refers to
localised structural interventions on some decayed timber parts or connections and not to an overall
intervention. This allows a possible re-tractability of the interventions. The results concerning the
assessment of the timber retrofitting with CNTs resin coating are positive, since they provide an
increase in the flexural strength which is greater or at least equal to the one obtained with the resin
alone.
Further investigations are necessary to prove the statistical significance of the results. In addition,
permeability of the CNTs resin coating must be studied more in detail, in order to obtain CNTs
content that maximizes the coating permeability to vapour.
The results obtained appear to be promising, and the research on the procedure is still in progress to
improve the innovative technique and considering a larger number of samples [47].
225
Fig. 19 Decayed wooden samples; bending tests and scheme for the mechanical resistance on timber
“Jupiter” joints, detail of the brown rot decay (L=1000mm h=100mm B=35mm) (from [10])
Fig. 20 Load displacement diagram: positive effect of

reinforcement with resin and CNTs [10]
5. Conclusions
Wood, as a natural nano-composite, can potentially offer important applications in the field of
nanotechnologies. Together with the performances commonly required (long term stability,
durability and weather resistance, good adhesion to the substrate, transparency, sustainability
for the production process, etc.), it is possible, through nano-structured coatings to introduce
additional surface functionality such as self-cleaning, photocatalysis, water resistance, fire
226
resistance, scratch resistance, graffiti resistance, and antibacterial coatings.

The main nano-materials already existing in the market, nano-coatings and their wood surface
protection functions, classification and compatibility with wood have been reviewed in this
chapter. On-going experimental research projects for potential new fields of application in the
sector of architecture, civil engineering and cultural heritage have also been analysed with a
special focus on reinforcement of historical timber structures.
The results of experimental research activity on the application of a polymeric resin reinforced
with CNTs on historical timber structures are promising and constitute a base for future
developments of the application of these nano-composite materials.
Research to assess and model environmental impacts and the economic viability of using nano-scale
materials in wood surfaces and timber structures is still needed in most of the applications.
6. Acknowledgements
The research on reinforcement of historic timber structures with carbon nano-composites was carried
out with the support of Italian Ministry of University and Research, through a PRIN Project (Use of
nanotechnologies in Cultural Heritage for the efficiency of maintenance systems in wooden built
heritage: innovative technologies for restoration) and with EU Regional funds through the M.A.N.
Project (Nanostructured Architectural Maintenance) coordinated by Prof. C. Bertolini Cestari who is
gratefully acknowledged.
7. References
[1] Ashby M.F., Ferreira P.J., Schodek D.L., Nanomaterials, nanotechnologies and design. An
introduction for Engineers and Architects, Elsevier, Paris, 2009.
[2] Bartos P.J.M., De Miguel Y., Porro A. (editors), 2005, Proceedings of the 2nd International
Symposium on Nanotechnology in Construction, 13-16 November 2005, Bilbao (Spain),
RILEM, Bagneux.
[3] Elvin G., “The Nano Revolution. A science that works on the molecular scale is set to transform
the way we build”, Architect Magazine, May 2007.
[4] Blee A., Matisons J.G., “Nanoparticles and the conservation of cultural heritage”, Materials
Forum, Vol. 32, 2008, pp. 121-128.
[5] AA.VV., Nanotechnology for the Forest Products Industry. Vision and Technology Roadmap,
Lansdowne, U.S.A., October 17-19, 2004, TAPPI Press, Atlanta.
[6] Cai Z., Rudie A., Stark N., Sabo R., Ralph S., “New Products and Product Categories in the
Global Forest Sector”, In: Hansen E., Panear R., Vlosky R. (Eds.), The Global Forest Sector:
Changes, Practices, and Prospects, CRC Press, Boca Raton, 2013, pp. 129-149.
[7] Bertolini Cestari C., Marzi T., Seip E., Touliatos P. (editors), Interaction between Science,
Technology and Architecture in timber construction, Elsevier, Paris 2004.
[8] Kasal, B., Tannert, T., In Situ Assessment of Structural Timber, Springer, 2010.
[9] Ansell M. P., “Multi-functional nano-materials for timber in construction”, Proceedings of the
ICE, Construction Materials, Vol. 166, Issue 4, 2013, pp. 248 –256.
[10] Marzi T., Impiego di nanotecnologie nei beni culturali per l'efficienza di sistemi manutentivi del
costruito in legno: tecnologie innovative di recupero (Nanotechnologies / nanosciences: from
wood improvement to the reinforcement of timber and monitoring of interventions), Ph.D.
Thesis in Innovation Technology for Built Environment, tutor Prof. Bertolini Cestari C.,
Politecnico di Torino, 2010.
[11] Nowack B. and Bucheli T.D., “Occurrence, behavior and effects of nanoparticles in the
227
environment”, Environmental Pollution, Vol. 150, No. 1, 2007, pp. 5–22.

[12] www.nanobiz.com (date of access October 2014)
[13] The NanoPhos Casebook: applications from across the world, Jan 2013,
http://www.nanophos.com/en/ (date of access October 2014)
[14] Auclair N., Riedl B., Blanchard V. and Blanchet P., “Improvement of photoprotection of wood
coatings by using inorganic nanoparticles as ultraviolet absorbers”, Forest Products Journal,
Vol. 61, No. 1, 2011, pp. 20–27.
[15] Blanchard V. and Blanchet P., “Color stability for wood products during use: effects of
inorganic nanoparticles”, BioResources, Vol. 6, No. 2, 2011, pp. 1219–1229.
[16] Allen N.S., Edge M., Ortega A. et al., “Behaviour of nanoparticle (ultrafine) titanium dioxide
pigments and stabilisers on the photooxidative stability of water based acrylic and isocyanate
based acrylic coatings”, Polymer Degradation and Stability, Vol. 78, No. 3, 2002, pp. 467-478.
[17] Vlad-Cristea M., Riedl B. and Blanchet P, “Enhancing the performance of exterior waterborne
coatings for wood by inorganic nanosized UV absorbers”, Progress in Organic Coatings, Vol.
69, 2010, pp. 432–441.
[18] Vlad-Cristea M., Riedl B. and Blanchet P, “Effect of addition of nanosized UV absorbers on the
physicomechanical and thermal properties of an exterior waterborne stain for wood”, Progress
in Organic Coatings, Vol. 72, 2011, pp. 755–762.
[19] Vlad-Cristea M., Riedl B., Blanchet P. and Jimenez-Pique E, “Nanocharacterization techniques
for investigating the durability of wood coatings”, European Polymer Journal, Vol. 48, 2012,
pp. 441–453.
[20] Germak C., Bozzola M., “Savoir Bois: culture and furniture for the mountain”, Paesaggio
Urbano, Vol. 4, 2013, pp. 42-47.
[21] Germak C., Bozzola M., Technical Report of the EU Project Alcotra 2007 “Savoir Bois”,
Politecnico di Torino, 2013.
[22] Nowack B., Krug H.F. and Height M.,” 120 years of nanosilver history: implications for policy
makers”, Environmental Science and Technology, Vol. 45(4), 2011, pp. 1177–1183.
[23] Rai M., Yadav A. and Gade A., “Silver nanoparticles as a new generation of antimicrobials”,
Biotechnology Advances, Vol. 27(1), 2009, pp. 76–83.
[24] Kaegi, R., Sinnet B., Zuleeg S. et al., “Release of Silver nanoparticles from outdoor facades”,
Environmental Pollution, Vol. 158(9), 2010, pp. 2900–2905.
[25] Kartal S.N., Green F. and Clausen C.A., “Do the unique properties of nano-metals affect
leachability or efficacy against fungi and termites?”, International Biodeterioration and
Biodegradation, Vol. 63, No. 6, 2009, pp. 490–495.
[26] Bertolini C., Crivellaro A., Marciniak M., Marzi T., Socha M., “Nanostructured materials for
durability and restoration of wooden surfaces in architecture and civil engineering”, In:
Proceedings of 11th World Conference on Timber Engineering, 2010, Vol. III, pp. 705-706.
[27] Lowden L.A., Hull T.R., “Flammability behaviour of wood and a review of the methods for its
reduction”, Fire Science Reviews, Vol. 2, No. 4, 2013.
[28] Taghiyari H.R., “Fire-retarding properties of nano-silver in solid wood”, Wood Science and
Technology, Vol. 46, 2012, pp. 939-952.
[29] www.percenta.com (date of access October 2014)
[30] Dei L. and Salvadori B., “Nanotechnology in cultural heritage conservation: nanometric slaked
lime saves architectonic and artistic surfaces from decay”, Journal of Cultural Heritage, Vol. 7,
No. 2, 2006, pp.110–115.
[31] Daniele V., Taglieri G. and Quaresima R, “The nanolimes in cultural heritage conservation:
characterisation and analysis of the carbonatation process”, Journal of Cultural Heritage, Vol.
9, No. 3, 2008, pp. 294–301.
[32] Baglioni P. and Giorgi R., “Soft and hard nanomaterials for restoration and conservation of
cultural heritage”, Soft Matter, Vol. 2, no. 4, 2006, pp. 293–303.
228
[33] Baglioni P., Chelazzi D., Giorgi R., “Nanoparticles of calcium hydroxide for wood
conservation. The deacidification of the Vasa warship”, Langmuir, Vol. 21, 2005, pp. 10743-
10748.
[34] Kaiser J-P, Zuin S., Wick P., “Is nanotechnology revolutionizing the paint and lacquer industry?
A critical opinion”, Science of the Total Environment, Vol. 442, 2013, pp. 282-289.
[35] Hischier R., Walser T., “Life cycle assessment of engineered nanomaterials: State of the art and
strategies to overcome existing gaps”, Science of the Total Environment, Vol. 425, 2012, pp.
271-282.
[36] “NanoHouse Dissemination Reports”, http://www-nanohouse-project.eu
[37] Hetzer M., De Kee D, “Wood/polymer/nanoclay composites, environmentally friendly
sustainable technology: A review”, Chemical Engineering Research and Design, Vol. 86, 2008,
pp. 1083-1093.
[38] Avella M., Cocca M., Errico M.E., Gentile G., “Nanotechnologies and Nanosensors: Future
Applications for the Conservation of Cultural Heritage”, In: J.P. Reithmaier, P. Paunovic, W.
Kulish, C. Popov, P. Petkov (edited by), Nanotechnological Basis for Advanced Sensors, NATO
Science for Peace and Security Series B: Physics and Biophysics, Springer, 2011.
[39] Winkler C., Schwarz U., “Multifunctional Wood-Adhesives for Structural Health Monitoring
Purposes”, Materials and Joints in Timber Structures, RILEM Bookseries, Volume 9, 2014, pp.
381-394.
[40] Thostenson, E.T., Chou, T. W., “Real-time in situ sensing of damage evolution in advanced
fiber composites using carbon nanotube networks”, Nanotechnology, Vol.19, 215713, 2008.
[41] Wehnert, F., Heinrich, J., Jansen, I., “Multifunctional adhesives by integration of carbon
nanotubes”, In: EURADH 2012 - 9th European Adhesion Conference, 2012.
[42] Wipf T., Phares B. M., Ritter M., Literature Review and Assessment of Nanotechnology for
Sensing of Timber Transportation Structures Final Report, United States Department of
Agricolture, Forest Service, Forest Products Laboratory, General Technical Report FPL–GTR–
210, April 2012.
[43] Breuer, O., Sundararaj, U., “Big return from small fibers: A review of polymer/carbon nanotube
composites”, Polymer Composites, Vol. 25, No. 6, 2004, pp. 630-645.
[44] Ahmad Z., Ansell M.P., Smedley D., “Epoxy adhesives modified with nano- and micro-
particles for in-situ timber bonding: effect of microstructure on bond integrity”, International
Journal of Mechanical and Materials Engineering, Vol. 5, 2010, pp. 59-67.
[45] Bertolini Cestari C., Invernizzi S., Marzi T., Tulliani J.M., “Nanotechnologies applied to the
restoration and maintenance of wooden built heritage”, In: D’Ayala D., Forde E. (editors),
Structural analysis of historical construction. Preserving safety and significance, Proceedings
of SAHC2008. VI International Conference, Bath; July 2008, Taylor & Francis, London, 2008,
pp. 941-947.
[46] Bertolini Cestari C., Invernizzi S., Marzi T., Tulliani J.M., “Use of nanotechnologies and
nanosciences in cultural heritage for the efficiency of maintenance systems in wooden built
heritage: restoration, conservation, maintenance, monitoring of interventions”, In: Structures
en bois dans le patrimoine bâti, Actes des journées techniques intern. Bois, Les cahiers
d’ICOMOS France, Icomos, Paris, 2009, pp. 87-91.
[47] Bertolini C., Invernizzi S., Marzi T., Tulliani J.M., “The reinforcement of ancient timber-joints
with carbon nano-composites”, MECCANICA, Vol. 48, 2013, pp. 1925-1935.
[48] Gojny FH, Wichmann MHG, Köpke U, Fiedler B, Schulte K., “Carbon nanotube-reinforced
epoxy-composites: enhanced stiffness and fracture toughness at low nanotube content”,
Composites Science & Technology , Vol. 64, 2004, pp. 2363–2371.
[49] Lau K.T., Lui D., “Effectiveness of using carbon nanotubes as nano-reinforcements for
advanced composite structures”, in Carbon, Vol. 40, No. 9, 2002, pp. 1597-1617.
[50] Yeh, M.-K., Hsieh, T.-H, Tai, N.-H, “Fabrication and mechanical properties of multi-walled
carbon nanotubes/epoxy nanocomposites”, Materials Science and Engineering, Vol. 483–484,
2008, pp. 289–292.
[51] Chakraborty AK, Plyhm T, Barbezat M, Necola A, Terrasi GP, “Carbon nanotube (CNT)-epoxy
229
nanocomposites: a systematic investigation of CNT dispersion”, J Nanoparticles Research,

Vol. 13, 2011, pp. 6493-6506.
[52] Chen H., Jacobs O., Wu W., Rüdiger G., Schädel B., “Effect of dispersion method on
tribological properties of carbon nanotube reinforced epoxy resin composites”, Polymer
Testing, Vol. 26, 2007, pp. 351–360.
[53] Miyagawa H., Drzal L.T., “Thermo-physical impact properties of epoxy nanocomposites
reinforced by singlewall carbon nanotubes”, Polymer, Vol. 45, 2004, pp. 5163–5170.
[54] Lau K., Lu M., Lam C., Cheung H., Sheng F., Li H. (2005), “Thermal and mechanical
properties of single-walled carbon nanotube bundle-reinforced epoxy composites: the role of
solvent for nanotube dispersion”, Composites Science & Technology, Vol. 65, 2005, pp. 719–
725.
[55] Miyagawa H., Rich M.J., Drzal L.T., “Thermophysical properties of epoxy nanocomposites
reinforced by carbon nanotubes and vapour grown carbon fibers”, Thermochim Acta, Vol. 442,
2006, pp. 67–73.
[56] http://www.nanocyl.com (date of access October 2014)
[57] http://www.mapei.it/Referenze/Multimedia/367_Epojet_GB.pdf (date of access October 2014)
[58] Bertolini Cestari C, Invernizzi S., Marzi T., Tulliani J.M., “Nano-technologies/Smart-materials
in timber constructions belonging to cultural heritage”, In: Proceedings of 11th World
Conference on Timber Engineering, 2010, Vol. IV, pp. 761–762.
230
Outlook
Dietsch, P., Harte, A.M.
Wood features a multitude of positive characteristics in view of its application as building

material and it is undisputed that good design of timber structures should still aim at
minimizing stresses for which timber only features small capacities and brittle failure
mechanisms (e.g. tensile stresses perpendicular to the grain and shear), thereby avoiding or at
least minimizing the necessity for reinforcement. Despite this, there is still a multitude of
reasons necessitating the structural reinforcement of timber buildings, not only in existing
buildings but also for new structures.
In this report, current and emerging methods that are available to repair or enhance the
structural performance of timber structures are summarised. Criteria for selection and
examples are given for the implementation of the various reinforcement methods. For all
construction works it has to be verified that essential requirements like mechanical resistance,
stability and safety in use are met. The required performance is commonly verified by
complying with corresponding harmonized technical rules for the structural design as well as
products used in construction works. In cases where harmonized technical rules or technical
approvals are not available, an approval in the individual case or comparable have to be
sought.
Various reinforcement methods presented in this report still lack harmonized technical rules.
According to the European position on future standardization, harmonized technical rules
shall be prepared for “common design cases” and shall contain “only commonly accepted
results of research and validated through sufficient practical experience”. The target audience
for such rules is “competent civil, structural and geotechnical engineers, typically qualified
professionals able to work independently in relevant fields”. When preparing items for
standardization it is important to take into account the approach within the internationally
accepted system for standardization. Comparable to the approach taken by scientists when
tackling a problem - (1) methods, (2) materials (3) results - the system for standardization is
based on a 3-step-pyramid. This pyramid is based on (1) test standards (containing rules on
how to test products). Relating to these, product standards (2) are developed (giving strength
and stiffness parameters, boundary conditions and rules for production and quality control).
The design standards (3) represent the tip of this pyramid (providing design equations and
formulating specific requirements in e.g. spacing, edge distance, minimum anchorage length,
etc.). When developing design rules it a precondition to also develop (1) test procedures as
well as (2) a product standard on the product or system used. Without the latter, rules in a
design standard cannot be used since the basic parameters are missing, in other words, the
pyramid will not be complete if one element is missing.
231
List of contributors
Laurent Bléron University of Lorraine,
ENSTIB-LERMAB, Epinal, France
Jean-François Bocquet University of Lorraine,

Jorge M. Branco ISISE, Dept. Civil Eng., University of Minho

Guimarães, Portugal
Reinhard Brandner Institute of Timber Engineering and Wood Technology

Graz University of Technology, Austria
Wen-Shao Chang Department of Architecture and Civil Engineering,

University of Bath, UK
Jian-Fei Chen Queens University Belfast,

Belfast, Northern Ireland, UK
Thierry Descamps URBAINE, Dept. of Structural Mech. and Civil Eng., University of
Mons, Mons, Belgium
Philipp Dietsch Chair of Timber Structures and Building Construction,

Technische Universität München, Germany
Bettina Franke Bern University of Applied Sciences

Biel/Bienne, Switzerland
Steffen Franke Bern University of Applied Sciences

Biel/Bienne, Switzerland
Alessandra Gubana University of Udine

Udine, Italy
Annette Harte College of Engineering and Informatics

National University of Ireland, Galway, Ireland
Robert Jockwer ETH Swiss Federal Institute of Technology,

Zurich, Switzerland
Robert Kliger Chalmers University of Technology,

Gothenburg, Sweden
232
Damien Lathuillière University of Lorraine,

Tanja Marzi Politecnico di Torino, Department of Architecture and Design, Torino,

Italy
Daniel McPolin Queen’s University Belfast,

Belfast, UK
Caoimhe O’Neill Queen’s University Belfast,

Belfast, Northern Ireland, UK
Maria Adelaide Parisi Politecnico di Milano,

Milano, Italy
Maurizio Piazza Università degli Studi di Trento

Trento, Italy
Benedetto Pizzo CNR-IVALSA, Trees and Timber Institute, Sesto Fiorentino (FI), Italy
Rajčić Vlatka University of Zagreb,

Zagreb, Croatia
Kay-Uwe Schober Mainz University of Applied Sciences

Mainz, Germany
Erik Serrano Structural Mechanics

Lund University, Sweden
Dave Smedley Rotafix Ltd.

Abercrave, Swansea, UK
René Steiger Swiss Federal Laboratories for Materials Testing and Research
(EMPA), Dübendorf, Switzerland
Mislav Stepinac University of Zagreb,

Zagreb, Croatia
Robert Widmann Swiss Federal Laboratories for Materials Testing and Research
Qingfeng Xu Shanghai Research Institute of Building Sciences,

Shanghai, China
233
List of reviewers
Martin Ansell Department of Mechanical Engineering, University of Bath, UK
Francisco Arriaga Martitegui Departamento de Ingeniería y Gestión Forestal y Ambiental,

Universidad Politécnica de Madrid, Spain
Jorge Branco ISISE, Deptartment of Civil Engineering, University of Minho,

Guimarães, Portugal
Clara Certari Bertolini Department of Architecture and Design,

Politecnico di Torino, Italy
Helena Cruz LNEC, Timber Division, Lisbon, Portugal
Dina D’Ayala Civil, Environmental and Geomatic Engineering, University College

London, UK
Philipp Dietsch Chair of Timber Structures and Building Construction
Technische Universität München, Germany
Steffen Franke Bern University of Applied Sciences, Biel/Bienne, Switzerland
Annette Harte College of Engineering and Informatics,

National University of Ireland, Galway, Ireland
André Jorissen Technical University of Eindhoven, Eindhoven, Netherlands
Bo Kasal Department of Organic Materials and Faculty of Wood Materials, TU

Braunschweig, Germany
Robert Kliger Chalmers University of Technology, Gothenburg, Sweden
Daniel McPolin Queen’s University Belfast, Belfast, UK
Maurizio Piazza Università degli Studi di Trento, Trento, Italy
Vlatka Rajčić University of Zagreb, Zagreb, Croatia
Jakub Sandak National Research Council, Institute of Tree and Timber IVALSA,
Italy.
Erik Serrano Structural Mechanics, Lund University, Sweden
René Steiger Swiss Federal Laboratories for Materials Testing and Research,
Jean Marc Tulliani Department of Applied Science and Technology,
Politecnico di Torino, Italy
Robert Widmann Swiss Federal Laboratories for Materials Testing and Research
David Yeomans Consulting Engineer, Banbury, UK
234
235

HarteA - Dietschp.edsreinforcementoftimberstructures STAR

Uploaded by

Document Information

Original Title

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

HarteA - Dietschp.edsreinforcementoftimberstructures STAR

Uploaded by

Copyright:

Available Formats

Reinforcement of Timber Structures

Copyright Shaker Verlag 2015

Part I: Reinforcement Applications 3

2 Reversible timber-to-timber strengthening interventions on wooden floors 25

4 Analysis and strengthening of carpentry joints 55

5 Reinforcement of dowel type connections 77

6 Seismic Strengthening of Timber Structures 89

Part II: Reinforcement Methods 111

8 Glued-in rods 133

9 Reinforcement with self-tapping screws 161

11 Nanotechnologies for reinforcement and protection of timber structures:

List of contributors 232

Annette M. Harte, Philipp Dietsch, Editors

The maintenance and rehabilitation of existing buildings is an area of increasing importance

In conclusion, the reinforcement of existing timber structures is important from economic,

Steffen Franke1, Bettina Franke2, Annette M. Harte3

1 Typology of timber beams

Tab. 1 Classification of timber beams

Tab. 2 Overview of timber beam forms

Fig. 4 Cracks in grain direction at a glulam Fig. 5 Delamination at a glulam member

Tab. 4 Characteristics of cracks and their distributions [1], [2]

2.3 Bending failure

2.4 Compression failure

2.5 Tension failure

2.6 Shear failure

2.7 Insects and fungi

3.2 Repair of shrinkage cracks or delaminations

3.3 Replacement of damaged or decayed parts

Fig. 19 Beam end repair using bonded in rods or plates

Fig. 20 Beam end repair using bonded in rods or plates

Fig. 21 GFRP rods inserted in ends of propped beams [13]

Fig. 22 External reinforcement arrangements

Fig. 23 Internal reinforcement arrangements

3.5 Reinforcement in tension perpendicular to the grain

Ft ,90,d  1,3 Vd 31     2 1    

For the case of beams with a round or rectangular hole,

Fig. 26 Reinforcement of openings

Fig. 27 Reinforcement of pitched camber beams

The design tensile stress, σt,90,d, may be calculated as

3.6 Shear reinforcement

3.7 Reinforcement in compression perpendicular to the grain

nRd  kc,90  lef ,1  b  f c,90,d 

Fig. 28 Reinforcement for compression perpendicular to the grain

3.8 General remarks

Fig. 31 Long-term monitoring of Sins Bridge, [52]

4.2 Sins Bridge, Switzerland – Flexural reinforcement [51], [52]

[38] Blumer, H., Spannungsberechnungen an anisotropen Kreisbogenscheiben und Sattelträgern

2.1 Timber flanges connected to main beams by hardwood dowels

2.2 Timber flanges connected to main beams by self-tapping screws

Fig.4 Load configuration and monitoring device set up (mm)(from [27,28])

Fig. 5 Load-Deflection diagrams of XLam-to-Timber composite section beams

2.4 General Remarks

3. In-plane stiffness upgrade by using timber or timber based elements.

New layer of floorboards

Existing layer of floorboards

New layer of floorboards

Masonry wall to stiffening layer connection

Fig. 10 Specimens used in the full-scale tests (from[33])

Fig. 11 Setup of the floor test (from [33])

Fig. 13 Flooring board layout (from [35])

1. The need to reinforce/repair timber columns and walls

1.2 Mechanical failure

1.4 Need for higher strength