Introduction To The Variational Formulation in Mechanics Edgardo O Taroco Full Chapter PDF

Introduction to the Variational
Formulation in Mechanics Edgardo O.

Taroco
Visit to download the full and correct content document:
https://ebookmass.com/product/introduction-to-the-variational-formulation-in-mechani
cs-edgardo-o-taroco/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Introduction to Fracture Mechanics Robert O. Ritchie
https://ebookmass.com/product/introduction-to-fracture-mechanics-
robert-o-ritchie/
eTextbook 978-1118763780 Introduction to Cosmetic

Formulation and Technology
https://ebookmass.com/product/etextbook-978-1118763780-
introduction-to-cosmetic-formulation-and-technology/
Introduction to Quantum Mechanics John Dirk Walecka
https://ebookmass.com/product/introduction-to-quantum-mechanics-
john-dirk-walecka/
Introduction to Engineering Fluid Mechanics 1st Edition

Marcel Escudier
https://ebookmass.com/product/introduction-to-engineering-fluid-
mechanics-1st-edition-marcel-escudier/
Work Organization in Practice: A Critical Introduction
to Work Organisation Models P-O Börnfelt
https://ebookmass.com/product/work-organization-in-practice-a-
critical-introduction-to-work-organisation-models-p-o-bornfelt/
PSYCHODYNAMIC FORMULATION : an expanded approach. The

psychodynamic formulation collective John Wiley
https://ebookmass.com/product/psychodynamic-formulation-an-
expanded-approach-the-psychodynamic-formulation-collective-john-
wiley/
Solutions Manual to Introduction to Robotics Mechanics

and Control Third Edition John J. Craig
https://ebookmass.com/product/solutions-manual-to-introduction-
to-robotics-mechanics-and-control-third-edition-john-j-craig/
An Introduction to Statistical Mechanics and

Thermodynamics 2nd Edition Robert H. Swendsen
https://ebookmass.com/product/an-introduction-to-statistical-
mechanics-and-thermodynamics-2nd-edition-robert-h-swendsen/
Variational Calculus with Engineering Applications

Constantin Udriste
https://ebookmass.com/product/variational-calculus-with-
engineering-applications-constantin-udriste/
Introduction to the Variational Formulation in Mechanics
Introduction to the Variational Formulation
in Mechanics
Fundamentals and Applications
Edgardo O. Taroco, Pablo J. Blanco and Raúl A. Feijóo

HeMoLab - Hemodynamics Modeling
Laboratory
LNCC/MCTIC - National Laboratory for
Scientific Computing, Brazil
INCT-MACC - National Institute of Science
and Technology in Medicine Assisted by
Scientific Computing, Brazil
This edition first published 2020
© 2020 John Wiley & Sons Ltd
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,
except as permitted by law. Advice on how to obtain permission to reuse material from this title is available
at http://www.wiley.com/go/permissions.
The right of Edgardo O. Taroco, Pablo J. Blanco and Raúl A. Feijóo to be identified as the authors of this work
has been asserted in accordance with law.
Registered Offices
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Office
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
For details of our global editorial offices, customer services, and more information about Wiley products
visit us at www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that
appears in standard print versions of this book may not be available in other formats.
Limit of Liability/Disclaimer of Warranty

While the publisher and authors have used their best efforts in preparing this work, they make no
representations or warranties with respect to the accuracy or completeness of the contents of this work and
specifically disclaim all warranties, including without limitation any implied warranties of merchantability or
fitness for a particular purpose. No warranty may be created or extended by sales representatives, written
sales materials or promotional statements for this work. The fact that an organization, website, or product is
referred to in this work as a citation and/or potential source of further information does not mean that the
publisher and authors endorse the information or services the organization, website, or product may provide
or recommendations it may make. This work is sold with the understanding that the publisher is not engaged
in rendering professional services. The advice and strategies contained herein may not be suitable for your
situation. You should consult with a specialist where appropriate. Further, readers should be aware that
websites listed in this work may have changed or disappeared between when this work was written and when
it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial
damages, including but not limited to special, incidental, consequential, or other damages.
Library of Congress Cataloging-in-Publication data applied for

HB ISBN: 9781119600909
Cover Design by Wiley

Cover Images: © Sandipkumar Patel/Getty Images, © Pablo Javier Blanco
Set in 10/12pt Warnock by SPi Global, Pondicherry, India
Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY
10 9 8 7 6 5 4 3 2 1
To our families
vii
Contents
Preface xv
Part I Vector and Tensor Algebra and Analysis 1
1 Vector and Tensor Algebra 3

1.1 Points and Vectors 3
1.2 Second-Order Tensors 6
1.3 Third-Order Tensors 17
1.4 Complementary Reading 22
2 Vector and Tensor Analysis 23

2.1 Differentiation 23
2.2 Gradient 28
2.3 Divergence 30
2.4 Curl 32
2.5 Laplacian 34
2.6 Integration 35
2.7 Coordinates 38
Part II Variational Formulations in Mechanics 47
3 Method of Virtual Power 49

3.1 Introduction 49
3.2 Kinematics 50
3.2.1 Body and Deformations 50
3.2.2 Motion: Deformation Rate 55
3.2.3 Motion Actions: Kinematical Constraints 61
3.3 Duality and Virtual Power 66
3.3.1 Motion Actions and Forces 67
3.3.2 Deformation Actions and Internal Stresses 69
3.3.3 Mechanical Models and the Equilibrium Operator 71
viii Contents
3.4 Bodies without Constraints 74

3.4.1 Principle of Virtual Power 75
3.4.2 Principle of Complementary Virtual Power 80
3.5 Bodies with Bilateral Constraints 81
3.6 Bodies with Unilateral Constraints 87
3.7 Lagrangian Description of the Principle of Virtual Power 94
3.8 Configurations with Preload and Residual Stresses 97
3.9 Linearization of the Principle of Virtual Power 100
3.9.1 Preliminary Results 101
3.9.2 Known Spatial Configuration 102
3.9.3 Known Material Configuration 102
3.10 Infinitesimal Deformations and Small Displacements 103
3.10.1 Bilateral Constraints 104
3.10.2 Unilateral Constraints 105
3.11 Final Remarks 106
4 Hyperelastic Materials at Infinitesimal Strains 109

4.1 Introduction 109
4.2 Uniaxial Hyperelastic Behavior 109
4.3 Three-Dimensional Hyperelastic Constitutive Laws 113
4.4 Equilibrium in Bodies without Constraints 116
4.4.1 Principle of Virtual Work 117
4.4.2 Principle of Minimum Total Potential Energy 117
4.4.3 Local Equations and Boundary Conditions 118
4.4.4 Principle of Complementary Virtual Work 120
4.4.5 Principle of Minimum Complementary Energy 121
4.4.6 Additional Remarks 122
4.5 Equilibrium in Bodies with Bilateral Constraints 123
4.6 Equilibrium in Bodies with Unilateral Constraints 128
4.7 Min–Max Principle 131
4.7.1 Hellinger–Reissner Functional 131
4.7.2 Hellinger–Reissner Principle 133
4.8 Three-Field Functional 134
Contents ix
4.9 Castigliano Theorems 136

4.9.1 First and Second Theorems 136
4.9.2 Bounds for Displacements and Generalized Loads 139
4.10 Elastodynamics Problem 144
4.11 Approximate Solution to Variational Problems 148
4.11.1 Elastostatics Problem 148
4.11.2 Hellinger–Reissner Principle 154
4.11.3 Generalized Variational Principle 156
4.11.4 Contact Problems in Elastostatics 158
5 Materials Exhibiting Creep 165

5.2 Phenomenological Aspects of Creep in Metals 165
5.3 Influence of Temperature 168
5.4 Recovery, Relaxation, Cyclic Loading, and Fatigue 170
5.5 Uniaxial Constitutive Equations 173
5.6 Three-Dimensional Constitutive Equations 182
5.7 Generalization of the Constitutive Law 188
5.8 Constitutive Equations for Structural Components 191
5.8.1 Bending of Beams 192
5.8.2 Bending, Extension, and Compression of Beams 195
5.9 Equilibrium Problem for Steady-State Creep 199
5.9.1 Mechanical Equilibrium 199
5.9.2 Variational Formulation 201
5.9.3 Variational Principles of Minimum 205
5.10 Castigliano Theorems 209
5.10.1 First and Second Theorems 209
5.10.2 Bounds for Velocities and Generalized Loads 211
5.11 Examples of Application 214
5.11.1 Disk Rotating with Constant Angular Velocity 214
5.11.2 Cantilevered Beam with Uniform Load 217
5.12 Approximate Solution to Steady-State Creep Problems 219
5.13 Unsteady Creep Problem 225
5.14 Approximate Solutions to Unsteady Creep Formulations 227
6 Materials Exhibiting Plasticity 229

6.2 Elasto-Plastic Materials 229
6.3 Uniaxial Elasto-Plastic Model 235
6.3.1 Elastic Relation 235
6.3.2 Yield Criterion 236
6.3.3 Hardening Law 238
6.3.4 Plastic Flow Rule 240
6.4 Three-Dimensional Elasto-Plastic Model 243
6.4.1 Elastic Relation 244
x Contents
6.4.2 Yield Criterion and Hardening Law 246

6.4.3 Potential Plastic Flow 249
6.5 Drucker and Hill Postulates 253
6.6 Convexity, Normality, and Plastic Potential 255
6.6.1 Normality Law and a Rationale for the Potential Law 255
6.6.2 Convexity of the Admissible Region 257
6.7 Plastic Flow Rule 258
6.8 Internal Dissipation 260
6.9 Common Yield Functions 262
6.9.1 The von Mises Criterion 263
6.9.2 The Tresca Criterion 264
6.10 Common Hardening Laws 266
6.11 Incremental Variational Principles 267
6.11.1 Principle of Minimum for the Velocity 268
6.11.2 Principle of Minimum for the Stress Rate 269
6.11.3 Uniqueness of the Stress Field 270
6.11.4 Variational Inequality for the Stress 270
6.11.5 Principle of Minimum with Two Fields 271
6.12 Incremental Constitutive Equations 272
6.12.1 Constitutive Equations for Rates 273
6.12.2 Constitutive Equations for Increments 275
6.12.3 Variational Principle in Finite Increments 278
Part III Modeling of Structural Components 281
7 Bending of Beams 285

7.2 Kinematics 285
7.3 Generalized Forces 289
7.4 Mechanical Equilibrium 290
7.5 Timoshenko Beam Model 294
8 Torsion of Bars 301

8.2 Kinematics 301
8.3 Generalized Forces 304
8.4 Mechanical Equilibrium 305
8.5 Dual Formulation 309
9 Plates and Shells 315

9.2 Geometric Description 316
9.3 Differentiation and Integration 320
Contents xi
9.4 Principle of Virtual Power 323

9.5 Unified Framework for Shell Models 326
9.6 Classical Shell Models 332
9.6.1 Naghdi Model 332
9.6.2 Kirchhoff–Love Model 335
9.6.3 Love Model 340
9.6.4 Koiter Model 342
9.6.5 Sanders Model 344
9.6.6 Donnell–Mushtari–Vlasov Model 346
9.7 Constitutive Equations and Internal Constraints 347
9.7.1 Preliminary Concepts 348
9.7.2 Model with Naghdi Hypothesis 350
9.7.3 Model with Kirchhoff–Love Hypothesis 357
9.8 Characteristics of Shell Models 360
9.8.1 Relation Between Generalized Stresses 360
9.8.2 Equilibrium Around the Normal 361
9.8.2.1 Kirchhoff–Love Model 361
9.8.2.2 Love Model 362
9.8.2.3 Koiter Model 363
9.8.2.4 Sanders Model 363
9.8.3 Reactive Generalized Stresses 364
9.8.3.1 Reactions in the Naghdi Model 364
9.8.3.2 Reactions in the Kirchhoff–Love Model 366
9.9 Basics Notions of Surfaces 369
9.9.1 Preliminaries 369
9.9.2 First Fundamental Form 370
9.9.3 Second Fundamental Form 372
9.9.4 Third Fundamental Form 375
9.9.5 Complementary Properties 375
Part IV Other Problems in Physics 377
10 Heat Transfer 379

10.2 Kinematics 379
10.3 Principle of Thermal Virtual Power 381
10.4 Principle of Complementary Thermal Virtual Power 386
10.5 Constitutive Equations 388
10.6 Principle of Minimum Total Thermal Energy 390
10.7 Poisson and Laplace Equations 390
11 Incompressible Fluid Flow 393

11.2 Kinematics 394
xii Contents
11.4 Navier–Stokes Equations 403

11.5 Stokes Flow 405
11.6 Irrotational Flow 407
12 High-Order Continua 411

12.2 Kinematics 412
12.4 Dynamics 425
12.5 Micropolar Media 427
12.6 Second Gradient Theory 429
Part V Multiscale Modeling 435
13 Method of Multiscale Virtual Power 439

13.2 Method of Virtual Power 439
13.2.1 Kinematics 439
13.2.2 Duality 442
13.2.4 Equilibrium Problem 446
13.3 Fundamentals of the Multiscale Theory 447
13.4 Kinematical Admissibility between Scales 449
13.4.1 Macroscale Kinematics 449
13.4.2 Microscale Kinematics 451
13.4.3 Insertion Operators 453
13.4.4 Homogenization Operators 456
13.4.5 Kinematical Admissibility 458
13.5 Duality in Multiscale Modeling 462
13.5.1 Macroscale Virtual Power 462
13.5.2 Microscale Virtual Power 464
13.6 Principle of Multiscale Virtual Power 467
13.7 Dual Operators 468
13.7.1 Microscale Equilibrium 468
13.7.2 Homogenization of Generalized Stresses 470
13.7.3 Homogenization of Generalized Forces 472
14 Applications of Multiscale Modeling 475

14.2 Solid Mechanics with External Forces 475
14.2.1 Multiscale Kinematics 476
14.2.2 Characterization of Virtual Power 479
14.2.3 Principle of Multiscale Virtual Power 480
14.2.4 Equilibrium Problem and Homogenization 482
Contents xiii
14.2.5 Tangent Operators 487

14.3 Mechanics of Incompressible Solid Media 490
14.3.2 Multiscale Kinematics 493
14.3.3 Principle of Multiscale Virtual Power 495
14.3.4 Incompressibility and Material Configuration 497
Part VI Appendices 501
A Definitions and Notations 503

A.1 Introduction 503
A.2 Sets 503
A.3 Functions and Transformations 504
A.4 Groups 507
A.5 Morphisms 509
A.6 Vector Spaces 509
A.7 Sets and Dependence in Vector Spaces 512
A.8 Bases and Dimension 513
A.9 Components 514
A.10 Sum of Sets and Subspaces 516
A.11 Linear Manifolds 516
A.12 Convex Sets and Cones 516
A.13 Direct Sum of Subspaces 517
A.14 Linear Transformations 517
A.15 Canonical Isomorphism 522
A.16 Algebraic Dual Space 523
A.16.1 Orthogonal Complement 524
A.16.2 Positive and Negative Conjugate Cones 525
A.17 Algebra in V 526
A.18 Adjoint Operators 528
A.19 Transposition and Bilinear Functions 529
A.20 Inner Product Spaces 532
B Elements of Real and Functional Analysis 539

B.1 Introduction 539
B.2 Sequences 541
B.3 Limit and Continuity of Functions 542
B.4 Metric Spaces 544
B.5 Normed Spaces 546
B.6 Quotient Space 549
B.7 Linear Transformations in Normed Spaces 550
B.8 Topological Dual Space 552
B.9 Weak and Strong Convergence 553
xiv Contents
C Functionals and the Gâteaux Derivative 555

C.1 Introduction 555
C.2 Properties of Operator 𝒦 555
C.3 Convexity and Semi-Continuity 556
C.4 Gâteaux Differential 557
C.5 Minimization of Convex Functionals 557
References 559
Index 575
xv
Preface
This book was written intermittently over the period between 1980 and 2016 with an
aim to provide students attending the courses organized by the authors, particularly
for the graduate students at the National Laboratory for Scientific Computing (LNCC),
with the foundational material of Mechanics using a variational tapestry. It is the result
of the knowledge acquired and divulged by E.O.T and R.A.F. since the LNCC was estab-
lished, which was initiated with the creation of the Laboratory of Computing (LAC) of
the Brazilian Center for Research in Physics (CBPF) in 1977, through the foundation of
the Laboratory for Scientific Computing (LCC) in 1980, its conversion into the category
of a national laboratory (LNCC) in 1982 and the definitive move to the city of Petrópolis
in 1998.
Part of the material presented here was used in various courses of theoretical and
applied mechanics organized by E.O.T. and R.A.F. These course were
• 1st Course on Theoretical and Applied Mechanics: Theory of Shells and their Appli-
cations in Engineering (Module I – Basic Principles, July 5 to 30, 1982; Module II –
Mechanical Models, January 3 to February 11, 1983; and Module III – Instability of
Shells, July 4 to 30, 1983).
• 2nd Course on Theoretical and Applied Mechanics: Fundamentals of the Finite Ele-
ment Method and its Applications in Engineering (Module I – Fundamentals of the
Finite Element Method, July 2 to 27, 1984; Module II – Applications of the Finite
Element Method in Solid Mechanics, January 7 to February 1, 1985, Module III –
Modern Aspects of the Finite Element Method, July 1 to 26, 1985).
• 3rd Course on Theoretical and Applied Mechanics: Optimization: Fundamentals and
Applications in Engineering (Module I – Optimization in the Modeling and Analy-
sis of Engineering Problems, July 7 to August 1, 1986; Module II – Optimal Design:
Foundations and Applications, July 6 to 31, 1987).
Unfortunately, and largely due to the long period of time that this document took to
be finished, one of the authors, Prof. Edgardo O. Taroco, left us (passed away in January
2010). Nevertheless, we decided to keep his name among the authors as an acknowledge-
ment to his deep contributions and in honor to his memory, as well as to the friendship
and generosity that he always offered us throughout all these years. Therefore, all errors
(of any kind) are the sole responsibility of P.J.B. and R.A.F.
Part of this material was used by R.A.F. in the course dictated during the Post-Doctoral
Latin American Seminar on Continuum Mechanics and Microstructure, organized by
the National Atomic Energy Commission in Argentina, sponsored by the Organization
xvi Preface
of American States (OAS), and held in Buenos Aires, July-August 1984. It was only after
1998 that the LNCC was moved to Petrópolis, and the LNCC Graduate Program was
initiated. Then, we decided to consolidate the aforementioned texts into a monograph
that would give emphasis to the formulation of the mechanics within a purely variational
structure. More recently, since P.J.B. joined the LNCC in 2009, the idea of finishing this
document, also including topics in other areas of physics as well as the extension of the
variational framework to a multi-scale paradigm, resurfaced. Currently, this material is
being used in several courses in the graduate program in Computational Modeling at
LNCC, Brazil, and also in the graduate program in Mechanical Engineering at National
University of Mar del Plata, Argentina.
Such variational framework used to present the mechanics was chosen not only
for being one of our main areas of research, but also, and fundamentally, because we
strongly believe that this way of looking into the roots of mechanics is the most suitable
and convenient perspective to approach the mathematical modeling. In fact, and as it
will become increasingly clear as we proceed, the foundational pillars upon which the
whole modeling journey rests using this variational tapestry are the following
• The first pillar is related to the description of the kinematics, that is the formalization
of the kinematical hypotheses which provides the definition of the generalized motion
actions and admissible generalized strain rate actions for the model under study.
• The second pillar consists of the mathematical duality postulated between quanti-
ties related to such motion actions and generalized strain rates with, respectively,
the external generalized forces and internal generalized stresses. In this way, forces
and stresses are constructs fully shaped by linear continuous functionals whose argu-
ments are kinematical entities. This aspect establishes a clear difference between the
approach developed in this book and the procedure followed by most of the literature
in the field of continuum mechanics, where forces and stresses are malleable entities
introduced a priori, regardless of the kinematics defined for the physical system.
• The third pillar is the Principle of Virtual Power (or its generalization, the Principle
of Multi-scale Virtual Power in the context of modeling problems with more than
one scale). This principle allows to establish, for the physical system under study, the
concept of mechanical equilibrium between external forces and internal stresses and,
when proper constitutive relations are given, this principle characterizes the general-
ized displacement field for which the associated generalized stress state equilibrates
the external forces.
These three pillars support the realm of so-called Method of Virtual Power (MVP),
which establishes well-defined basic steps targeting a fully consistent modelling tech-
nique. Such variational structure was proposed, although with subtle modifications,
by Prof. Paul Germain [114–118]. Particularly, Prof. Germain (lifetime member of the
French Academy of Science) was invited by us in 1982 to teach the course Four Lectures
on the Foundation of Shell Theory within Module I of the 1st Course on Theoretical and
Applied Mechanics that we organized, precisely, to disseminate these concepts among
students and professors from different Latin American countries. Another professor
who greatly contributed to consolidate this approach in these courses was Prof. Giovanni
Romano (Facoltà di Ingegneria dell’Università degli Studi di Napoli, Italy).
The organization of this book closely follows this spirit. In Part I we present the basic
concepts of vector and tensor algebra and analysis, where the reader is introduced to
Preface xvii
the ubiquitous use of compact vector and tensor notation. This compact notation allows
to go through the basic principles and concepts of the mechanics in a clear and concise
way, without being obscured by the presence of indices, components and metric-related
entities, which should be relegated to their specific role at the time of the calculation.
Part II is dedicated to presenting the Method of Virtual Power (MVP), from the kine-
matics, through the duality and to the Principle of Virtual Power. This Part presents the
variational groundwork which is used as a guiding theme in all that follows. The MVP is
then applied to the most general case in the mechanics of deformable bodies, while its
application to the case of hyperelastic materials, and materials which may experience
creep and plasticity phenomena is also discussed. In Part III we present the applica-
tion of the MVP to the modelling of structural components such as beams, plates and
shells. For these components, it will be clear how forces and stresses emerge as natu-
ral outgrowth of the kinematical hypotheses. In Part IV, the application of the MVP to
other problems from physics is addressed, including heat conduction, incompressible
fluid flow, and high order continua. This will help the reader to illustrate the use of this
unified theoretical framework in problems in which a purely variational approach is sel-
dom encountered in the textbooks. Finally, in Part V, we expand this variational realm
to embrace problems which require a multi-scale paradigm. This extended variational
structure has been called the Method of Multi-scale Virtual Power (MMVP). Pursuing
the same standard, the MMVP allows to provide a convenient and safe tool to substan-
tiate multi-scale models of complex physical systems, allowing to consistently provide
the groundwork on top of which multi-scale homogenization should take place. Finally,
in the Appendices, we present various mathematical concepts and results to make the
document self-contained.
Last but not least, we would like to thank all those who somehow contributed to our
excursion towards this book. In particular we thank Gonzalo R. Feijóo for his contribu-
tion in the first drafting of some chapters of this book, and to Professors Enzo A. Dari
(Bariloche Atomic Centre, Argentina), Sebastián Giusti (National Technological Univer-
sity, Faculty of Córdoba, Argentina), Santiago A. Urquiza (National University of Mar del
Plata, Argentina), Pablo J. Sánchez (National Technological University, Regional Faculty
of Santa Fe, Argentina), Alejandro Clausse (National University of Central Buenos Aires,
Argentina) and Eduardo A. de Souza Neto (Zienkiewicz Center for Computational Engi-
neering, Swansea University, United Kingdom) for their comments and discussions that
definitely enriched us and, therefore, our work. We would also like to thank our Ph.D.
students, particularly to Gonzalo D. Ares, Gonzalo D. Maso Talou, Carlos A. Bulant,
Alonso M. Alvarez and Felipe Figueredo Rocha who, with their criticisms and observa-
tions, have also helped to improve this text.
Petrópolis, Brazil Pablo J. Blanco

March 2018 Raúl A. Feijóo
1
Part I
Vector and Tensor Algebra and Analysis
The goal of the following two chapters, which comprise Part I of this book, is to provide
the reader with the basic concepts and training in the area of tensor algebra (Chapter 1)
and tensor analysis (Chapter 2) which will be omnipresent throughout this book. It is
important to highlight here the use of compact notation (also called intrinsic notation)
when invoking vector and tensors as well as algebraic and differential operations among
them. That is, vectors and tensors are written independently from the adopted coor-
dinate system. In contrast to the use of indicial notation, which puts in evidence the
components of vectors and tensors in the given coordinate system, the compact nota-
tion appears as a clean and elegant form that allows the concepts to be presented without
being obnubilated by the sometimes overwhelming presence of indexes.
At the end of these two chapters, some further reading material is suggested to com-
plement and deepen the concepts presented here. However, we highlight that for the
study of the rest of the book this reading is not mandatory. Moreover, the reader will
also find in Appendices A and B a more detailed exposition of the topics addressed in
what follows, in an attempt to make the book self-contained.
3
Vector and Tensor Algebra
1.1 Points and Vectors

Consider the three-dimensional Euclidean space, denoted by ℰ, whose geometry is built
upon a set of primitive elements called points. Note that ℰ is not a vector space in the
sense of the algebra because the addition of points is a concept without meaning.
The difference between two points X and Y of space ℰ is defined by
v = X − Y, (1.1)
where v is the vector whose origin is in Y and ends at X. All the vectors which can
be determined through the differences between points belonging to ℰ form the set V
associated with ℰ. The set V is a (real) vector space, where the two basic operations
inherent to the notion of vector space are defined, which are (i) the addition of vectors
and (ii) the product of a vector by a real number.
The addition operation between a point Y ∈ ℰ and the vector v ∈ V defines the point
X ∈ ℰ such that (1.1) is verified. This operation allows us to establish a biunivocal cor-
respondence between points of ℰ and vectors of V . In fact, we can arbitrarily pick a
point O from ℰ, and then for each point X ∈ ℰ there exists a unique vector v ∈ V such
that v = X − O.
Space V is called three-dimensional provided there exist in V sets of three vectors
{ei } = {e1 , e2 , e3 } which are linearly independent1 and can span the entire vector space
V , that is, they can generate any vector v ∈ V through the linear combination v = 𝛼i ei .2
Any of these linearly independent sets of vectors {ei } is called a basis for the (real) vector
space V .
Beyond basic operations, multiplication by a real number, and addition of vectors,
the vector space V associated with ℰ is endowed with an inner product, also called the
scalar product, operation between vectors of V . For u, v ∈ V , this operation is denoted
by u ⋅ v, which is geometrically defined by the product of the lengths of the vectors mul-
tiplied by the cosine of the angle between them. This operation satisfies the properties
of the inner product in the sense of the algebra.
1 Vectors {ei } are called linearly independent if 𝛼i ei = 𝟎 implies 𝛼i = 0 for i = 1, 2, 3.

2 Throughout this book, unless stated otherwise, Einstein notation is adopted to shorten summation
notations, resulting, for example, in the following lumped notation when indexes are repeated:
∑3
v = i=1 𝛼i ei = 𝛼i ei .
Introduction to the Variational Formulation in Mechanics: Fundamentals and Applications, First Edition.
Edgardo O. Taroco, Pablo J. Blanco and Raúl A. Feijóo.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
4 1 Vector and Tensor Algebra
Also, and through the introduction of this operation, the space V has a topological
structure induced by the inner product through the definition of the norm operation
for a vector v ∈ V
√
||v|| = v ⋅ v. (1.2)
Making use of these operations, vectors u, v ∈ V are said to be orthogonal if u ⋅ v = 0.
Similarly, a basis {ei } of V is called orthogonal if ei ⋅ ej = ||ei ||||ej ||𝛿ij is verified, i, j =
1, 2, 3, where 𝛿ij is the Kronecker symbol.3 Finally, a basis of V is called orthonormal
if it is orthogonal and the norm of the vectors in the basis is unitary, that is, ||ei || = 1,
i = 1, 2, 3.
The matrix [gij ], i, j ∈ {1, 2, 3}, defined by
gij = ei ⋅ ej , i, j = 1, 2, 3, (1.3)
is not singular when the set {ei } is a basis for V . In turn, from the definition it follows
that gij = gji , that is, [gij ] is a symmetric matrix. Let us call
[g ij ] = [gij ]−1 , (1.4)
the inverse of matrix [gij ]. From the definition we obtain
g ik gkj = gik g kj = 𝛿ij , i, j = 1, 2, 3. (1.5)
With these results, we can introduce the dual basis {ei } associated with {ei } as the image
of the linear transformation
[g ij ] ∶V → V ,
(1.6)
ej → ei = g ij ej .
It can be proved that this transformation produces a basis for V . Reciprocally, the appli-
cation of [gij ] over {ei } yields the original basis {ei }. In fact
gij ej = gij g jk ek = 𝛿ik ek = ei . (1.7)
Another useful result is the following
ei ⋅ ej = g ik ek ⋅ ej = g ik gkj = 𝛿ij . (1.8)
A coordinate system consists of a basis {ei } for V , not necessarily orthogonal, and
an arbitrary point O of ℰ called the origin of the coordinate system. When the basis is
orthonormal, the coordinate system is called Cartesian.
Observe that the notion of vector was introduced independently from the adopted
basis, or, equivalently, from the coordinate system. When a basis {ei }, and then its dual
basis {ei }, is chosen, then each vector u ∈ V can be associated with a triple of real num-
bers {u1 , u2 , u3 } called components of u with respect to the basis {ei }, which are defined
as follows
ui = u ⋅ ei i = 1, 2, 3. (1.9)
i i
These are the components with respect to {e } because, as {e } is a basis, it results in
u = 𝛼i ei and then
ui = u ⋅ ei = 𝛼k ek ⋅ ei = 𝛼k 𝛿ki = 𝛼i . (1.10)
3 The Kronecker symbol 𝛿ij is such that 𝛿ii = 1 and 𝛿ij = 0 for i ≠ j.
1.1 Points and Vectors 5
In particular, given a basis {ei }, and its dual {ei }, the components of u with respect to
these bases are usually named as follows
• Components of u with respect to {ei } are contravariant components, and are

defined by
ui = u ⋅ ei i = 1, 2, 3. (1.11)
With these components the vector u can be represented through the linear combina-
tion u = ui ei .
• Components of u with respect to {ei } are covariant components, and are defined by
ui = u ⋅ ei i = 1, 2, 3. (1.12)
With these components the vector u can be represented through the linear combina-
tion u = ui ei .
If the basis is orthonormal, it is easy to show that it is identical to its dual

basis, and therefore there is no distinction between covariant and contravariant
components.
Likewise, given a coordinate system in ℰ, characterized by the basis {ei } and the point
O ∈ ℰ, we can define the coordinates of an arbitrary point X ∈ ℰ as the covariant com-
ponents of the vector X − O from V , that is
Xi = (X − O) ⋅ ei . (1.13)
Thus, the same vector u ∈ V , or the same point X ∈ ℰ, can be associated with differ-
ent triples of components and representations depending upon the chosen coordinate
system.4
To underline the difference between intrinsic and indicial notation, note that
the inner product of vectors u, v ∈ V as a function of their different components
results in
u ⋅ v = ui vi = ui vi = ui g ij vj = ui gij vj , (1.14)
that is, several expressions are possible for the same concept, which may delay the under-
standing progress.
Exercise 1.1 Verify the validity of identities given by (1.14).
4 As stated above, we will employ compact notation, which means that we highlight the entity in detriment
of its components, in this case vector u, or point X. Notation emphasizing components is also called indicial
notation, and it should only be employed during the analysis of a given specific problem, specifically when
calculations are required. In this way, we can present the concepts and operations in a clear and concise
manner, without being confined to a certain coordinate system. This also helps to call the attention to the
fact that, during calculations, we should choose that basis which makes the treatment of the problem as
simple as possible. To sum up, indicial notation is not related to concepts, but to calculus.
1.2 Second-Order Tensors

We will employ the term tensor, or second-order tensor, as a synonym for the linear
transformation between V and V .5 Then, we have
T ∶V → V ,
(1.15)
u → v = Tu,
verifying
T(u + v) = Tu + Tv ∀u, v ∈ V , (1.16)
T(𝛼u) = 𝛼Tu ∀𝛼 ∈ ℝ and ∀u ∈ V . (1.17)
Clearly, expressions (1.16) and (1.17) are equivalent to
T(𝛼u + 𝛽v) = 𝛼Tu + 𝛽Tv ∀𝛼, 𝛽 ∈ ℝ and ∀u, v ∈ V . (1.18)
Taking 𝛼 and 𝛽 zero, we obtain an expression that will be frequently used hereafter,
T𝟎 = 𝟎, (1.19)
where 𝟎 represents the null element in V .
The set of all second-order tensors, that is, the set of all linear transformations between
V and V , will be called Lin
Lin = {T; T ∶ V → V , linear}. (1.20)
Introducing the addition and multiplication by real numbers in Lin defined by
(T + S)u = Tu + Su ∀u ∈ V , (1.21)
(𝛼T)u = 𝛼(Tu) ∀u ∈ V , (1.22)
turns Lin into a vector space, where the null tensor O transforms every vector u ∈ V
into the null vector 𝟎 ∈ V , that is,
Ou = 𝟎 ∀u ∈ V . (1.23)
The identity tensor is denoted by I and is defined by
Iu = u ∀u ∈ V . (1.24)
Given T ∈ Lin, the set of vectors v ∈ V satisfying Tv = 𝟎 is denoted by 𝒩 (T), that is,
𝒩 (T) = {v ∈ V ; Tv = 𝟎}. (1.25)
It is easy to prove that 𝒩 (T) is also a vector subspace of V , called the null space of T
(also the kernel of T).
Given A, B ∈ Lin, the composition of these tensors (composed transformation) is
another tensor (linear transformation) T ∈ Lin such that
Tu = (AB)u = ABu ∀u ∈ V. (1.26)
5 In this chapter we limit the presentation to second- and third-order tensors, and some associated
operations because these are the most used elements along the book. However, in Appendices A and B the
reader will find material related to linear transformations between vector spaces of possibly different
dimensions, which naturally embraces the case of third- and also second-order tensors.
Since in general AB ≠ BA, when the identity is verified we say that A and B are com-
mutative.
From previous definitions, and given arbitrary T, S, D ∈ Lin and 𝛼 ∈ ℝ, then we have
T(SD) = (TS)D = TSD, (1.27)
T(S + D) = TS + TD, (1.28)
𝛼(TS) = T(𝛼S), (1.29)
IT = TI = T. (1.30)
Indeed, for arbitrary v ∈ V , it is
T(SD)v = T(S(Dv)) = (TS)(Dv) = (TS)Dv, (1.31)
from which (1.27) is verified. Also, we have
T(S + D)v = T((S + D)v) = T(Sv + Dv) = TSv + TDv = (TS + TD)v, (1.32)
which verifies (1.28). Similarly,
𝛼(TS)v = 𝛼(T(Sv)) = T(𝛼Sv)) = T(𝛼S)v, (1.33)
and (1.29) holds. Note finally that
ITv = I(Tv) = Tv, (1.34)
and then we have proved (1.30).
The following notation will also be used
n
⏞⏞⏞
Tn = TT...T n ∈ N, (1.35)
with T0 = I.
The transpose of a tensor T is the unique tensor TT satisfying
Tu ⋅ v = u ⋅ TT v ∀u, v ∈ V . (1.36)
Uniqueness is proved assuming that there are two tensor transposes for T, denoted by
TT1 and TT2 , and which will be assumed to be different. From definition (1.36), each tensor
satisfies
Tu ⋅ v = u ⋅ TT1 v ∀u, v ∈ V , (1.37)
Tu ⋅ v = u ⋅ TT2 v ∀u, v ∈ V . (1.38)
Subtracting both expressions yields
u ⋅ (TT1 v − TT2 v) = u ⋅ (TT1 − TT2 )v = 0 ∀u, v ∈ V . (1.39)
Recalling that a ⋅ b = 0, ∀a ∈ V implies b = 𝟎, we have
(TT1 − TT2 )v = 𝟎 ∀v ∈ V , (1.40)
and from the definition of the null tensor (1.23), we arrive at
TT1 = TT2 , (1.41)
which contradicts the fact that both tensors were different. Therefore, there exists a
unique transpose of a tensor.
Using the definition of the transpose of a tensor it is straightforward to conclude that,
for arbitrary A, S ∈ Lin and 𝛼 ∈ ℝ, we obtain
(S + A)T = ST + AT , (1.42)
(𝛼S) = 𝛼S .
T T
(1.43)
In fact, taking arbitrary u, v ∈ V , then
(S + A)T u ⋅ v = u ⋅ (S + A)v = u ⋅ Sv + u ⋅ Av = ST u ⋅ v + AT u ⋅ v
= (ST u + AT u) ⋅ v = (ST + AT )u ⋅ v, (1.44)
and so (1.42) is proved. Analogously,
(𝛼S)T u ⋅ v = u ⋅ (𝛼S)v = 𝛼u ⋅ Sv = 𝛼ST u ⋅ v, (1.45)
and so (1.43) is demonstrated.
As it is easy to see, the transpose operation is a linear transformation between Lin and
Lin. In addition, given arbitrary A, S ∈ Lin, it is
(SA)T = AT ST , (1.46)
(ST )T = S. (1.47)
In fact, for arbitrary u, v ∈ V , we have
(SA)T u ⋅ v = u ⋅ (SA)v = u ⋅ S(Av) = ST u ⋅ Av = AT ST u ⋅ v, (1.48)
and we arrive at (1.46). Analogously
(ST )T u ⋅ v = u ⋅ (ST )v = Su ⋅ v, (1.49)
and then (1.197) is verified.
A tensor S is called symmetric if
Su ⋅ v = u ⋅ Sv ∀u, v ∈ V , (1.50)
T
and in such a case we conclude that S = S . A tensor is called skew-symmetric if
Su ⋅ v = −(u ⋅ Sv) ∀u, v ∈ V , (1.51)
T
which implies S = −S .
The set of all symmetric tensors will be denoted by Sym and the set of all
skew-symmetric tensors will be denoted by Skw. In particular, the null tensor O
is symmetric and skew-symmetric.
Any tensor S ∈ Lin can be univocally represented by the addition of a symmetric ten-
sor (called Ss ) and a skew-symmetric tensor (called Sa ), that is,
S = Ss + Sa , (1.52)
where
1
Ss = (S + ST ), (1.53)
2
1
Sa = (S − ST ), (1.54)
2
are, respectively, called the symmetric component and the skew-symmetric compo-
nent of S.
Since the transpose operation yields a unique tensor transpose, it follows that the
linear combination of a symmetric (skew-symmetric) tensor results in a symmetric
(skew-symmetric) tensor. Then, Sym and Skw are vector subspaces of Lin. Moreover,
from the uniqueness of the decomposition into symmetric and skew-symmetric
components, it is concluded that Lin can be written as the direct sum of these two
subspaces
Lin = Sym ⊕ Skw. (1.55)
Consider an arbitrary W ∈ Skw and an arbitrary T ∈ Lin. Then, it is possible to show
that for any u ∈ V the following holds
u ⋅ Wu = 0, (1.56)
u ⋅ Tu = u ⋅ T u.
s
(1.57)
In fact, for arbitrary u ∈ V we have
u ⋅ Wu = WT u ⋅ u = −Wu ⋅ u, (1.58)
from where (1.56) is proved. Similarly, and using (1.56), we have
u ⋅ Tu = u ⋅ (Ts + Ta )u = u ⋅ Ts u + u ⋅ Tau = u ⋅ Tsu, (1.59)
and (1.57) follows.
The tensor product between two vectors a, b ∈ V is the second-order tensor a ⊗ b
that transforms any vector v ∈ V into vector (b ⋅ v)a, that is,
(a ⊗ b)v = (b ⋅ v)a ∀v ∈ V . (1.60)
From the previous definition, and given arbitrary a, b, c, d ∈ V , we obtain the follow-
ing results
(a ⊗ b)T = b ⊗ a, (1.61)
(a ⊗ b)(c ⊗ d) = (b ⋅ c)(a ⊗ d). (1.62)
In addition, given T ∈ Lin, it can be shown that
T(a ⊗ b) = (Ta) ⊗ b, (1.63)
(a ⊗ b)T = a ⊗ (TT b). (1.64)
Indeed, consider arbitrary u, v ∈ V , using definition (1.60) forward and backward,
we have
(a ⊗ b)T u ⋅ v = u ⋅ (a ⊗ b)v = u ⋅ (b ⋅ v)a = (u ⋅ a)(b ⋅ v) = (b ⊗ a)u ⋅ v, (1.65)
and then (1.61) holds. Now, observe that
(a ⊗ b)(c ⊗ d)u = (a ⊗ b)(d ⋅ u)c = (d ⋅ u)(b ⋅ c)a
= (b ⋅ c)(d ⋅ u)a = (b ⋅ c)(a ⊗ d)u, (1.66)
and we arrive at (1.62). Similarly, we have
T(a ⊗ b)u = (b ⋅ u)Ta = ((Ta) ⊗ b)u, (1.67)
(n ⊗ n)v Figure 1.1 Geometric concept of orthogonal

projection over the plane 𝜋 whose normal vector is n.
v
(I − n ⊗ n)v
which yields (1.63). Lastly, note that

(a ⊗ b)Tu = (b ⋅ Tu)a = (TT b ⋅ u)a = (a ⊗ (TT b))u, (1.68)
and then (1.64) is proved.
Analogously to the definition of the inner product in V , it is also possible to define
the inner product in Lin. Consider two elements T, S ∈ Lin which can be written in the
forms T = t1 ⊗ t2 and S = s1 ⊗ s2 , respectively, with t1 , t2 , s1 , s2 ∈ V . Then, we define
the inner product T ⋅ S in Lin as
T ⋅ S = (t1 ⊗ t2 ) ⋅ (s1 ⊗ s2 ) = (t1 ⋅ s1 )(t2 ⋅ s2 ). (1.69)
With definition (1.69), it is straightforward to prove that, for arbitrary T ∈ Lin and
u, v ∈ V , the following result holds
T ⋅ (u ⊗ v) = u ⋅ Tv. (1.70)
In fact, putting T = t1 ⊗ t2 , and making use of definitions (1.69) and (1.60) we obtain
T ⋅ (u ⊗ v) = (t1 ⊗ t2 ) ⋅ (u ⊗ v) = (t1 ⋅ u)(t2 ⋅ v)
= u ⋅ [(t2 ⋅ v)t1] = u ⋅ [(t1 ⊗ t2 )v] = u ⋅ Tv. (1.71)
Let us denote n the unit normal vector to the plane 𝜋 (see Figure 1.1). The tensor n ⊗ n
applied over any vector v ∈ V gives
(n ⊗ n)v = (n ⋅ v)n, (1.72)
which is the orthogonal projection of v ∈ V over the direction n. In turn, the
second-order tensor P = I − n ⊗ n applied over any vector v ∈ V yields
Pv = (I − n ⊗ n)v = v − (n ⋅ v)n, (1.73)
which is the orthogonal projection of v over the plane 𝜋.
It can be appreciated that tensor P is symmetric and also verifies P2 = P. In fact
PT = (I − n ⊗ n)T = IT − (n ⊗ n)T = I − n ⊗ n = P, (1.74)
and, for arbitrary v ∈ V , it is
P2 v = P(Pv) = P[v − (n ⋅ v)n] = Pv − (n ⋅ v)Pn = Pv, (1.75)
and the previous statements hold.
Tensors satisfying these two properties, that is, P ∈ Sym and P2 = P, are called
orthogonal projection tensors. Examples of this kind of tensor are
I, I − n ⊗ n, n ⊗ n. (1.76)
It is possible to show that
dim(Lin) = dim V × dim V = 9, (1.77)
and if we take a basis {ei } for V , the sets
{(ei ⊗ ej ); i, j = 1, 2, 3}, (1.78)
{(ei ⊗ ej ); i, j = 1, 2, 3}, (1.79)
{(e ⊗ ej ); i, j = 1, 2, 3},
i
(1.80)
{(e ⊗ e ); i, j = 1, 2, 3},
i j
(1.81)
are different possible bases for Lin. In this way, any tensor T ∈ Lin can be expressed by a
unique linear combination of the element of the chosen basis. The components of tensor
T in the chosen basis are defined in an analogous manner to that for vectors. Hence, we
have
Tij = T ⋅ (ei ⊗ ej ) = ei ⋅ Tej , (1.82)
T ij = T ⋅ (ei ⊗ ej ) = ei ⋅ Tej , (1.83)
T.ji = T ⋅ (e ⊗ ej ) = e ⋅ Tej ,
i i
(1.84)
.j
Ti = T ⋅ (ei ⊗ ej ) = ei ⋅ Tej , (1.85)
where the inner product between elements of Lin is defined in (1.69). Also, this inner
product can be introduced in terms of the trace operation, as shown below. This way,
the representation of T ∈ Lin in terms of these components is given by
.j
T = Tij (ei ⊗ ej ) = T ij (ei ⊗ ej ) = T.ji (ei ⊗ ej ) = Ti (ei ⊗ ej ), (1.86)
where the implicit summation of repeated indices is considered. Coefficients Tij , T ij , T.ji
.j
and Ti are, respectively, covariant, contravariant, and mixed components of tensor T.
In particular, the identity tensor I is
I = ei ⊗ ei , (1.87)
again with implicit summation over index i. In fact, given arbitrary u = uj ej = uj ej ∈ V ,
we have
Iu = (ei ⊗ ei )uj ej = uj ei 𝛿ij = ui ei = u, (1.88)
or equivalently,
Iu = (ei ⊗ ei )uj ej = uj g ij ei = uj ej = u. (1.89)
For a Cartesian basis {ei } for V , there is no difference between components of T. In
this case, we simply refer to the Cartesian components of the tensor.
The advantage of employing compact notation is again evident when comparing to
indicial notation. A tensor is a concept (linear transformation in V ) which does not
depend on the basis chosen for V . The same is valid for the trace operation and inner
product between tensors.
To further illustrate the conceptual aspects highlighted by the intrinsic notation, let
us introduce some of the definitions already presented in terms of both notations.
• Tensor. Compact notation: T ∈ Lin. Indicial notation
Tij = T kl gki glj = T.jk gki = Ti.k gkj , (1.90)
.j
T ij = Tkl g ki g lj = T.ki g kj = Tk g ki , (1.91)
T.ji ki
= Tkj g = T gkj = ik
Tk.l g ki glj , (1.92)
.j
Ti = Tik g kj = T kj gki = T.lk gki glj . (1.93)
• Application of a tensor over a vector. Compact notation: u = Tv, u, v ∈ V and T ∈
Lin. Indicial notation (just some of the possible expressions)
.j .j
ui = Tij vj = Tij g jk vk = Ti vj = Ti gjk vk , (1.94)
.j
i ij
u = T vj = T gjk v = ij k
T.ji vj = Ti g jk vk . (1.95)
• Composition of tensors. Compact notation: T = AB, T, A, B ∈ Lin. Indicial notation
Tij = Aik Bk.j = A.ki Bkj , (1.96)
.j
T ij = Aik Bk = Ai.k Bkj , (1.97)
T.ji = Ai.k Bk.j = A Bkj , ik
(1.98)
.j .j
Ti = A.ki Bk = Aik Bkj , (1.99)
and variants including the tensor or gij . g ij
• Tensor product between vectors. Compact notation: u ⊗ v, u, v ∈ V . Indicial nota-
tion
(u ⊗ v)ij = ui vj , (1.100)
(u ⊗ v) = u v ,ij i j
(1.101)
(u ⊗ v)i.j = u vj ,
i
(1.102)
.j
(u ⊗ v)i = ui vj . (1.103)
• Symmetric component of a tensor. Compact notation: S ∈ Sym ⇔ S = ST , and equiv-
alently Sa = O. Indicial notation
Sij = Sji =⇒ [Sij ] = [Sij ]T , (1.104)
S ij = S ji =⇒ [S ij ] = [S ij ]T , (1.105)
S.ji = Sj.i . (1.106)
Note that for a symmetric tensor the matrix of covariant components is also sym-
metric, and the same holds for contravariant components. In contrast, the matrix
representation of a symmetric tensor in mixed components is not symmetric in gen-
eral. Indeed, since S.ji = Sj.i, we conclude that the same components are symmetrically
placed in the two (different) matrix representations of the tensor.
• Skew-symmetric component of a tensor. Compact notation: W ∈ Skw ⇔ W = −WT ,

and then Ws = O. Indicial notation
Wij = −Wji =⇒ [Wij ] = −[Wij ]T =⇒ Wii = 0, (1.107)
W ij = −W ji =⇒ [W ij ] = −[W ij ]T =⇒ W ii = 0, (1.108)
W.ji = −Wj.i . (1.109)
That is, for a skew-symmetric tensor the matrix representations in covariant and
contravariant components are skew-symmetric, while the matrix representations in
mixed components are not skew-symmetric. As before, the expression W.ji = −Wj.i
indicates that these coefficients are symmetrically placed, but in the two different
matrix representations in mixed components.
The trace operation of a second-order tensor T ∈ Lin is a linear functional which asso-
ciates to each tensor T ∈ Lin a real number denoted by trT, that is,
tr ∶ Lin → ℝ,
(1.110)
T → trT,
with the property
tr(a ⊗ b) = a ⋅ b ∀a, b ∈ V . (1.111)
From the linearity of the trace operation it follows that
.j
trT = T ij tr(ei ⊗ ej ) = Tij tr(ei ⊗ ej ) = T.ji tr(ei ⊗ ej ) = Ti tr(ei ⊗ ej )
= T ij gij = Tij g ij = T.ii = Ti.i. (1.112)
These expressions allow us to evaluate trT in terms of the components of the tensor.
Evidently, the result is independent of the adopted basis.
As the transpose operation is a linear operation, we have
trT = trTT , (1.113)
tr(AB) = tr(BA), (1.114)
trI = 3. (1.115)
Indeed, putting T = t1 ⊗ t2 , A = a1 ⊗ a2 , and B = b1 ⊗ b2 , and recalling that I = ei ⊗
ei , it is
trT = tr(t1 ⊗ t2 ) = t1 ⋅ t2 = t2 ⋅ t1 = tr(t2 ⊗ t1 ) = trTT , (1.116)
and then (1.113) follows. Also, using (1.62) and (1.113) yields
tr(AB) = tr[(a1 ⊗ a2 )(b1 ⊗ b2 )] = tr[(a2 ⋅ b1 )(a1 ⊗ b2 )] = (a2 ⋅ b1 )(a1 ⋅ b2 )
= tr[(a1 ⋅ b2 )(a2 ⊗ b1 )] = tr[(a1 ⋅ b2 )(b1 ⊗ a2 )]
= tr[(b1 ⊗ b2 )(a1 ⊗ a2 )] = tr(BA), (1.117)
and (1.114) is proved. Finally, with definition (1.87), we have
trI = tr(ei ⊗ ei ) = ei ⋅ ei = 𝛿ii = 3, (1.118)
thus proving (1.115).
Notice that using the trace operation makes it possible to define an inner product in
Lin
S ⋅ T = tr(ST T) ∀S, T ∈ Lin. (1.119)
With this definition, the following properties are satisfied
S ⋅ T = T ⋅ S, (1.120)
S⋅S≥0 ∀S ∈ Lin, (1.121)
S ⋅ S = 0 ⇔ S = O, (1.122)
and the inner product, in turn, induces a norm in Lin
√ √
‖S‖ = S ⋅ S = tr(ST S). (1.123)
Expressions of the inner product in terms of some of the different tensor components
are
.j .j
A ⋅ T = Aij T ij = Aij Tij = Ai T.ji = Ai.j Ti , (1.124)
and variants also involving g ij or gij .
Now, consider arbitrary a, b, c, d ∈ V , T, A, B ∈ Lin, S ∈ Sym and W ∈ Skw. Then,
some properties of the inner product in Lin are
I ⋅ T = trT, (1.125)
A ⋅ BT = BT A ⋅ T, (1.126)
T ⋅ (u ⊗ v) = u ⋅ Tv, (1.127)
(a ⊗ b) ⋅ (c ⊗ d) = (a ⋅ c)(b ⋅ d), (1.128)
S ⋅ T = S ⋅ Ts , (1.129)
W⋅T=W⋅T , a
(1.130)
S ⋅ W = 0, (1.131)
T ⋅ A = 0 =⇒ T = O, (1.132)
T ⋅ S = 0 =⇒ T ∈ Skw, (1.133)
T ⋅ W = 0 =⇒ T ∈ Sym, (1.134)
A⋅B=A ⋅B +A ⋅B .
s s a a
(1.135)
In fact, from definition (1.119), expression (1.125) follows directly because
I ⋅ T = tr(IT T) = trT. (1.136)
Identity (1.126) is also proved directly, since
A ⋅ BT = tr(AT BT) = tr[(BT A)T T] = BT A ⋅ T. (1.137)
Exercise 1.2 Prove that identities (1.127)–(1.135) hold using the properties of the trace
operation.
The determinant of a tensor A ∈ Lin can be defined as

1[ ]
det A = 2tr(A3 ) + (trA)3 − 3(trA)tr(A2 ) . (1.138)
6
For arbitrary A, T ∈ Lin, and 𝛼 ∈ ℝ, the determinant operation satisfies
det(AT) = det(TA), (1.139)
det(AT) = (det A)(det T), (1.140)
det(𝛼T) = 𝛼 3 det T, (1.141)
det I = 1. (1.142)
The inverse of a tensor T is also a tensor T−1 with the property
T−1 T = I. (1.143)
A tensor T is called invertible if det T ≠ 0, and if T is invertible, its transpose tensor TT
is also invertible, verifying
(TT )−1 = (T−1 )T = T−T . (1.144)
Tensor Q is called an orthogonal tensor if it preserves the inner product between
vectors, that is,
Qu ⋅ Qv = u ⋅ v ∀u, v ∈ V . (1.145)
The necessary and sufficient condition for a tensor Q to be orthogonal is
QQT = QT Q = I, (1.146)
or, equivalently,
QT = Q−1 . (1.147)
The set of all orthogonal tensors is denoted by Orth.
Exercise 1.3 Show that a tensor is orthogonal if and only if QQT = QT Q = I.
Any orthogonal tensor with positive determinant is called a rotation tensor. In partic-
ular, the set of all rotation tensors is denoted by Rot.
A tensor T is called positive definite if
v ⋅ Tv > 0 ∀v ∈ V v ≠ 𝟎, (1.148)
v ⋅ Tv = 0 ⇔ v = 𝟎. (1.149)
Given a basis {ek }, it is said that all the bases {ek } have the same orientation as {ek }
if they can be obtained from a rotation of the latter. That means
ek = Rek k = 1, 2, 3. (1.150)
Since any orthogonal tensor Q is a rotation R, or it is a rotation multiplied by −1, there
exist two classes of bases each associated with an orientation. Hereafter we will assume
that one of these orientations has been chosen.
The cross product u × v between vectors u, v ∈ V , whose angle between them is 𝜃, is

another vector w such that
w ⟂ to the plane defined by u and v, (1.151)
||w|| = ||u||||v|| sin 𝜃, (1.152)
{u, v, w} have the same orientation as the adopted basis. (1.153)
Then, it follows that
u × v = −(v × u), (1.154)
u × u = 𝟎, (1.155)
u ⋅ v × w = w ⋅ u × v = v ⋅ w × u. (1.156)
Given three arbitrary elements u, v, w ∈ V , the cross product satisfies
u × (v × w) = (u ⋅ w)v − (u ⋅ v)w, (1.157)
or equivalently,
u × (v × w) = (v ⊗ w)a u. (1.158)
Exercise 1.4 Prove that (1.157) holds.
When u, v, w are linearly independent, the value of the product u ⋅ (v × w) represents

the volume of the parallelepiped 𝒫 determined by the vectors u, v, w. Then, given a
tensor T, it is not difficult to verify that
Tu ⋅ Tv × Tw
det T = , (1.159)
u⋅v×w
from where it follows that
vol(T(𝒫 ))
det T = , (1.160)
vol(𝒫 )
which provides a geometric interpretation of the determinant of a tensor T, and where
T(𝒫 ) is the image of the parallelepiped 𝒫 under the linear transformation (tensor) T
and vol(⋅) stands for the volume.
Given a second-order tensor field T ∈ Lin, we introduce the principal invariants of T
as follows
I1 (T) = trT, (1.161)
1
I2 (T) = (tr(T2 ) − (trT)2), (1.162)
2
I3 (T) = det T. (1.163)
The invariance property is established by the fact that
Ii (T) = Ii (QTQT ) ∀Q ∈ Orth i = 1, 2, 3. (1.164)
Then, any tensor T admits the representation established by the Cayley–Hamilton
theorem [132]
det(𝛼I − T) = 𝛼 3 − 𝛼 2I1 (T) − 𝛼I2(T) − I3 (T). (1.165)
By putting
p(𝛼) = det(𝛼I − T), (1.166)
it can be shown that all second-order tensors satisfy the following characteristic
equation
p(T) = O. (1.167)
Therefore, we have
T3 − I1 (T)T2 − I2 (T)T − I3 (T)I = O, (1.168)
and taking the trace of this equation, and considering the definition of the invariants
(1.161), (1.162) and (1.163), yields
1
tr(T3 ) − (trT)tr(T2 ) + ((trT)3 − (trT)tr(T2 )) = 3 det T, (1.169)
2
or equivalently,
1 1 1
det T = tr(T3 ) + (trT)3 − (trT)tr(T2), (1.170)
3 6 2
which is exactly (1.138).
Let T = Tij (ei ⊗ ej ) be a second-order tensor given in any orthonormal basis {ei } for
V , then it results in
det T = 𝜀ijk Ti1Tj2 Tk3 = 𝜀ijk T1i T2j T3k , (1.171)
where 𝜀ijk is the permutation symbol6 defined by
⎧1 if (i, j, k) is an even permutation of (1, 2, 3),

⎪
𝜀ijk = ⎨−1 if (i, j, k) is an odd permutation of (1, 2, 3), (1.172)
⎪0 otherwise.
⎩
Finally, given a tensor W ∈ Skw, there exists a unique vector w ∈ V such that
Wv = w × v ∀v ∈ V . (1.173)
Vector w associated with the skew-symmetric tensor W is called the axial vector of W.
From this definition, it follows that the kernel of W ∈ Skw, denoted by 𝒩 (W), is the
one-dimensional subspace of V spanned by the axial vector w of W, that is,
𝒩 (W) = {v ∈ V ; Wv = 𝟎} = {v ∈ V ; v = 𝜆w, ∀𝜆 ∈ ℝ}. (1.174)
1.3 Third-Order Tensors

In this section we will extend the concept of linear transformations to embrace linear
operators between V and Lin, and vice versa. This context gives rise to the notion of
third-order tensors.
6 Actually, this is a very peculiar third-order tensor, also known as the Levi–Civita symbol.
In short, third-order tensors are linear applications of the form

S ∶ Lin → V ,
(1.175)
T → ST,
as well as
S ∶ V → Lin,
(1.176)
a → Sa,
which implies that these entities can be applied to elements (vectors) of V , and also to
elements (second-order tensors) of Lin. The set, actually vector space when endowed
with the standard operations, of all third-order tensors will be called Lin. The null ele-
ment in Lin is denoted by 0, and is such that
0a = O ∀a ∈ V , (1.177)
0T = 𝟎 ∀T ∈ Lin. (1.178)
However, applications (1.175) and (1.176) are not fully characterized yet. Consider
a ∈ V , T = t1 ⊗ t2 ∈ Lin and S = s1 ⊗ s2 ⊗ s3 ∈ Lin, with t1 , t2 , s1 , s2 , s3 ∈ V . Hence, we
define application (1.175) as follows
ST = (s1 ⊗ s2 ⊗ s3 )(t1 ⊗ t2 ) = (s2 ⋅ t1 )(s3 ⋅ t2 )s1 , (1.179)
and, analogously, application (1.176) is defined by
Sa = (s1 ⊗ s2 ⊗ s3 )a = (s3 ⋅ a)(s1 ⊗ s2 ). (1.180)
In the space Lin the transpose operation of a second-order tensor was defined as
a ⋅ Tb = TT a ⋅ b. In the same manner, we can define different transpose operations for
elements in Lin. We then define the transpose operations7
a ⋅ ST = ST a ⋅ T, (1.181)
1
Sa ⋅ T = a ⋅ S T. T (1.182)
1
Let us show that these third-order tensors are unique, and also that (S )T = S holds. ToT
do this, assume that ST1 ≠ ST2 , then

a ⋅ ST = ST1 a ⋅ T ∀a ∈ V , ∀T ∈ Lin,
a ⋅ ST = ST2 a ⋅T ∀a ∈ V , ∀T ∈ Lin.
Subtracting both expressions above, we obtain
(ST1 − ST2 )a ⋅ T = 0 ∀a ∈ V , ∀T ∈ Lin, (1.183)
7 An important point to stress here is that the definition of the transpose operation is not unique in the case
of third-order tensors. As a matter of fact, the transpose operation could equivalently be defined as
Sa ⋅ T = a ⋅ ST T,
1
a ⋅ ST = S T a ⋅ T.
1
Comparing the definitions, we see that the operation (⋅)T defined here is equivalent to the operation (⋅) T
defined in the main text (also equivalent to ((⋅)T )T ). In any case, the subsequent mathematical developments
have to consistently follow the chosen definition of this operation.
which implies
(ST1 − ST2 )a = O ∀a ∈ V , (1.184)
and, from (1.177), we finally conclude that
ST1 = ST2 , (1.185)
leading to a contradiction.
1 1
Analogously, assuming that there exist S1T ≠ S2T , a similar argument yields
1 1
a ⋅ (S1T − S2T )T = 0 ∀a ∈ V , ∀T ∈ Lin, (1.186)

from which we obtain
1 1
(S1T − S2T )T = 𝟎 ∀T ∈ Lin, (1.187)

and, using (1.178), we obtain a contradiction, from which it follows that
1 1
S1T = S2T . (1.188)

Finally, from the transpose definition we have
1 1
Sa ⋅ T = a ⋅ S T T = (S T )T a ⋅ T, (1.189)
and then
1
(S − (S T )T )a ⋅ T = 0 ∀a ∈ V , ∀T ∈ Lin, (1.190)
that is,
1
(S − (S T )T )a = O ∀a ∈ V , (1.191)
and, similarly, from (1.177), we conclude that
1
S = (S T )T . (1.192)
Let us see now that, if a third-order tensor is of the form S = s1 ⊗ s2 ⊗ s3 , the following
is verified
(s1 ⊗ s2 ⊗ s3 )T = s2 ⊗ s3 ⊗ s1 . (1.193)
To obtain this result, consider definitions (1.179) and (1.180). Then, for arbitrary a ∈ V
and T = t1 ⊗ t2 ∈ Lin, we have
a ⋅ ST = a ⋅ (s1 ⊗ s2 ⊗ s3 )(t1 ⊗ t2 ) = (s2 ⋅ t1 )(s3 ⋅ t2 )(a ⋅ s1 )
= (s2 ⊗ s3 ⊗ s1 )a ⋅ (t1 ⊗ t2 ) = (s1 ⊗ s2 ⊗ s3 )T a ⋅ (t1 ⊗ t2 ) = ST a ⋅ T,
(1.194)
and (1.193) follows.
In the same manner, it can be proved that
1
(s1 ⊗ s2 ⊗ s3 ) T = s3 ⊗ s1 ⊗ s2 . (1.195)
In fact, with (1.179) and (1.180) we get
Sa ⋅ T = (s1 ⊗ s2 ⊗ s3 )a ⋅ (t1 ⊗ t2 ) = (s3 ⋅ a)(s1 ⋅ t1 )(s2 ⋅ t2 )
1
= a ⋅ (s3 ⊗ s1 ⊗ s2 )(t1 ⊗ t2 ) = a ⋅ (s1 ⊗ s2 ⊗ s3 ) T (t1 ⊗ t2 ), (1.196)
proving then (1.195).
With the previous results, we can directly obtain the following relations
1
(ST )T = S T , (1.197)
((ST )T )T = S. (1.198)
Exercise 1.5 Prove that (1.197) and (1.198) are verified.
Let a ∈ V and T = t1 ⊗ t2 ∈ Lin be arbitrary, then we can build an element of Lin

using the tensor product as follows
a ⊗ T = a ⊗ t1 ⊗ t 2 , (1.199)
T ⊗ a = t1 ⊗ t2 ⊗ a. (1.200)
In this way, considering u, v ∈ V , T, S ∈ Lin and S ∈ Lin, it is quite straightforward to
show that
(u ⊗ T)T = T ⊗ u, (1.201)
1
(T ⊗ u) = u ⊗ T,
T (1.202)
(u ⊗ T)S = (T ⋅ S)u, (1.203)
(u ⊗ T)v = u ⊗ (Tv), (1.204)
(T ⊗ u)S = TSu, (1.205)
(T ⊗ u) S = TSu,
T T T
(1.206)
(T ⊗ u)v = (u ⋅ v)T, (1.207)
(T ⊗ v) u = (T u) ⊗ v,
T T
(1.208)
(T ⊗ v)T S = TT ST v, (1.209)
T T
(S u)v = (Sv) u, (1.210)
S(v ⊗ u) = (Su)v. (1.211)
Exercise 1.6 Prove that tensor identities (1.201)–(1.211) hold.
The vector space Lin can also be endowed with an inner product structure. Consider
the elements T, S ∈ Lin, with T = t1 ⊗ t2 ⊗ t3 and S = s1 ⊗ s2 ⊗ s3 , t1 , t2 , t3 , s1 , s2 , s3 ∈
V . Then, we define the inner product T ⋅ S as follows
T ⋅ S = (t1 ⊗ t2 ⊗ t3 ) ⋅ (s1 ⊗ s2 ⊗ s3 ) = (t1 ⋅ s1 )(t2 ⋅ s2 )(t3 ⋅ s3 ). (1.212)
So, for S, T ∈ Lin (second-order tensors), it is possible to show that
(u ⊗ S) ⋅ (v ⊗ T) = (u ⋅ v)(S ⋅ T), (1.213)
(u ⊗ S) ⋅ (T ⊗ v) = TSv ⋅ u, (1.214)
(u ⊗ S) ⋅ (T ⊗ v) = STu ⋅ v.T
(1.215)
Exercise 1.7 Prove that tensor identities (1.213)–(1.215) hold.

Another transpose operation that can be introduced in Lin is the following. Consider
an arbitrary S = s1 ⊗ s2 ⊗ s3 , then we define
St = (s1 ⊗ s2 ⊗ s3 )t = s1 ⊗ s3 ⊗ s2 . (1.216)
With this operation, for an arbitrary a ∈ V , the following holds
St a = (s1 ⊗ s2 ⊗ s3 )t a = (a ⋅ s2 )(s1 ⊗ s3 ), (1.217)
and also
1
(St )T = (S T )t . (1.218)
Also, consider two arbitrary vectors u, v ∈ V , then the next result holds
(St u)v = (Sv)u. (1.219)
Moreover, and similarly to the composition of second-order tensors (composition of

linear transformations), it is possible to define the composition of third-order tensors.
Consider arbitrary S = s1 ⊗ s2 ⊗ s3 ∈ Lin and T = t1 ⊗ t2 ∈ Lin. Then we can define the
composition of S with T yielding an element in Lin. This can be fully characterized as
follows
S ○ T = (s1 ⊗ s2 ⊗ s3 ) ○ (t1 ⊗ t2 ) = (s3 ⋅ t1 )(s1 ⊗ s2 ⊗ t2 ). (1.220)
Clearly, this operation gives S ○ T ∈ Lin. Differently, for arbitrary S, R ∈ Lin, we can pro-
ceed as
S ○ R = (s1 ⊗ s2 ⊗ s3 ) ○ (r1 ⊗ r2 ⊗ r3 ) = (s2 ⋅ r1 )(s3 ⋅ r2 )(s1 ⊗ r3 ), (1.221)
noting that this case results in S ○ R ∈ Lin.
With these definitions we obtain, for example,
(u ⊗ T) ○ (S ⊗ v) = (T ⋅ S)(u ⊗ v). (1.222)
This introduction should help the reader to become familiar with the mathemati-
cal manipulation of vector and tensor algebra, as well as illustrating the introduction
of definitions and their consequences, always working with intrinsic notations, that is,
independently from the coordinate system.
In effect, it is possible to extend all the operations with vectors and second-order ten-
sors to second- and third-order tensors. An example is the definition of the cross product
between a second-order tensor S = s1 ⊗ s2 and a vector a, or between two second-order
tensors S = s1 ⊗ s2 and T = t1 ⊗ t2 , for instance
S × a = (s1 ⊗ s2 ) × a = s1 ⊗ (s2 × a), (1.223)
a × S = a × (s1 ⊗ s2 ) = (a × s1 ) ⊗ s2 , (1.224)
S × T = (s1 ⊗ s2 ) × (t1 ⊗ t2 ) = (s1 × t1 ) × (s2 × t2 ). (1.225)
With these definitions we can obtain, for example, the following result
S × a = −[a × ST ]T . (1.226)
Furthermore, we can define operations for a second-order tensor with the structure
S = s1 ⊗ s2 such as the following
S× = (s1 ⊗ s2 )× = s1 × s2 , (1.227)
which is called the Gibbs product, and this verifies
(ST )× = −S× . (1.228)
In addition, given the axial vector a of an arbitrary A ∈ Skw, and letting B be arbitrary,
the following result holds
A ⋅ B = −a ⋅ B× . (1.229)
The treatment of higher-order tensors, fourth-order tensors for example, is entirely

analogous. Nevertheless, as we increase the order of the tensor it is evident that the
number of possible operations and number of alternative equivalent definitions grows.
The simplest examples are the different transpose operations that can be introduced.
For instance, consider the fourth-order tensor 𝔻
𝔻 = d1 ⊗ d 2 ⊗ d 3 ⊗ d 4 . (1.230)
The following transpose operations can be defined
𝔻T = (d1 ⊗ d2 ⊗ d3 ⊗ d4 )T = d3 ⊗ d4 ⊗ d1 ⊗ d2 , (1.231)
𝒯 𝒯
𝔻 = (d1 ⊗ d2 ⊗ d3 ⊗ d4 ) = d2 ⊗ d1 ⊗ d3 ⊗ d4 , (1.232)
𝔻𝔗 = (d1 ⊗ d2 ⊗ d3 ⊗ d4 )𝔗 = d1 ⊗ d2 ⊗ d4 ⊗ d3 , (1.233)
𝔻 = (d1 ⊗ d2 ⊗ d3 ⊗ d4 ) = d4 ⊗ d2 ⊗ d3 ⊗ d1 ,
T T
(1.234)
among others.
1.4 Complementary Reading

For the reader interested in deepening their knowledge of the topics briefly addressed
in this chapter, we recommend reading the appendices, particularly Appendix A, where
the reader will find concepts and additional information that will ease the reading of
the following works: [23], [47], [84], [93–95], [106], [132], [135], [171], [265], [299]
and [300].
23
Vector and Tensor Analysis
2.1 Differentiation
In this chapter we will introduce a notation for the concept of differentiation abstract
enough to include in the definition functions f of the kind
D(f ) ⊂ ℝ, ℰ, V , Lin, Lin, (2.1)
R(f ) ⊂ ℝ, ℰ, V , Lin, Lin, (2.2)
where D(f ) and R(f ) are, respectively, the domain of f and the image under the transfor-
mation f , also called the codomain of f .
Let 𝒳 and 𝒴 be two normed vector spaces. For the purpose of the present chapter, we
will consider that these spaces are finite-dimensional. Suppose f ∶ 𝒳 → 𝒴 is defined
in a neighborhood of the origin (null element) of 𝒳 . Then, f (u) approaches the origin
faster than u if
‖f (u)‖𝒴
lim = 0. (2.3)
u→0
u≠0
‖u‖𝒳
This will be represented by
f (u) = o(u), u → 0, (2.4)
or, equivalently, by
f (u) = o(u). (2.5)
Similarly, f (u) = g(u) + o(u) stands for
f (u) − g(u) = o(u). (2.6)
Note that this last definition has a clear meaning even when R(f ) and R(g) are subsets of
ℰ because in such cases f (u) − g(u) ∈ V .
Example 2.1 Let 𝜙 ∶ ℝ → ℝ be such that for each t ∈ ℝ we have 𝜙(t) = t 𝛼 . Then,
𝜙(t) = o(t) if 𝛼 > 1.
Consider a function g defined in an open set ℐ of ℝ, whose image is in the set ℝ

(g is a real-valued function), or in V (g is a vector-valued function), or in Lin (g is
a second-order tensor-valued function), or in Lin (g is a third-order tensor-valued
Introduction to the Variational Formulation in Mechanics: Fundamentals and Applications, First Edition.
Edgardo O. Taroco, Pablo J. Blanco and Raúl A. Feijóo.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
24 2 Vector and Tensor Analysis
̇
function), or in ℰ (g is said to be point-valued function). The derivative of g at t, g(t), if
exists, it is defined by
d 1
̇ =
g(t) g(t) = lim [g(t + 𝛼) − g(t)]. (2.7)
dt 𝛼→0 𝛼
Observe that if R(g) ⊂ ℰ then (g(t + 𝛼) − g(t)) ∈ V , and then the derivative is a vector
field. Likewise, the derivative of a real-valued function is a real number, the derivative of
a vector-valued function will be a vector field, and that of a tensor-valued function will
be a tensor field.
Function g is said to be differenciable at t ∈ ℐ if there exists the derivative at t. Thus,
from the definition of the derivative we have that
1
̇
lim [g(t + 𝛼) − g(t) − 𝛼 g(t)] = 0, (2.8)
𝛼→0 𝛼
or in an equivalent manner
̇ = o(𝛼),
g(t + 𝛼) − g(t) − 𝛼 g(t) (2.9)
from where
̇ + o(𝛼).
g(t + 𝛼) − g(t) = 𝛼 g(t) (2.10)
̇ is linear with respect to 𝛼, then, g(t + 𝛼) − g(t) can be written as the sum
Clearly, 𝛼 g(t)
of a linear term in 𝛼 and a term that approaches zero faster than 𝛼. This is a rather
useful definition for a derivative. Therefore, we define the derivative of g at t as the linear
transformation in ℝ that approaches g(t + 𝛼) − g(t) for small values of 𝛼.
Hence, consider two finite-dimensional normed vector spaces 𝒳 and 𝒴 , 𝒜 is an open
set of 𝒳 and let g ∶ 𝒜 → 𝒴 be a function. Function g is said to be differentiable at x ∈ 𝒜
if there exists the linear transformation
𝒟 g(x) ∶ 𝒳 → 𝒴 ,
(2.11)
u → 𝒟 g(x)[u],
such that
g(x + u) = g(x) + 𝒟 g(x)[u] + o(u), u → 0. (2.12)
The linear transformation 𝒟 g(x) is then called the derivative of g at x. Given that in a
finite-dimensional normed space all norms are equivalent, 𝒟 g(x) is independent from
the chosen norms for 𝒳 and 𝒴 . In turn, if there exists 𝒟 g(x), it is unique. In fact, for
each u ∈ 𝒳 it is
1 d |
𝒟 g(x)[u] = lim [g(x + 𝛼u) − g(x)] = g(x + 𝛼u)|| . (2.13)
𝛼→0 𝛼 d𝛼 |𝛼=0
If g is differentiable for all x ∈ 𝒜 , 𝒟 g represents the transformation which, for each
x ∈ 𝒜 , gives 𝒟 g(x) ∈ Lin(𝒳 , 𝒴 ), where Lin(𝒳 , 𝒴 ) is the space whose elements are
the linear transformations from 𝒳 into 𝒴 . This space is also finite-dimensional, in fact,
dim(Lin(𝒳 , 𝒴 )) = (dim 𝒳 )(dim 𝒴 ), and then we can define a norm. This fact allows us
to have notions of continuity and differentiability for 𝒟 g. In particular, g is said to be reg-
ular, or of class C 1 in 𝒜 if g is differentiable for all points x ∈ 𝒜 and if 𝒟 g is continuous.
Also, g is said to be of class C 2 in 𝒜 if g is of class C 1 and 𝒟 g is also regular, and so on.
Let us consider some examples which will be useful for the forthcoming develop-
ments.
• Consider
g ∶ V → ℝ,
(2.14)
v → g(v) = v ⋅ v.
Then
g(v + u) = (v + u) ⋅ (v + u) = v ⋅ v + 2v ⋅ u + u ⋅ u = g(v) + 2v ⋅ u + o(u).

(2.15)
From where
𝒟 g(v)[u] = 2v ⋅ u. (2.16)
• Consider
g ∶ Lin → Lin,
(2.17)
A → g(A) = A2 = AA.
Then
g(A + U) = (A + U)2 = A2 + (AU + UA) + o(U). (2.18)
From where
𝒟 g(A)[U] = AU + UA. (2.19)
• Consider
g ∶ 𝒳 → 𝒴 linear,
(2.20)
x → g(x).
Then
g(x + u) = g(x) + g(u). (2.21)
From where
𝒟 g(x)[u] = g(u). (2.22)
Frequently, it will be necessary to calculate the derivative of a product 𝜋(f , g) between
two functions f and g. In particular, we have already presented some product operations,
for example:
• Product of a real number 𝛼 by a vector v ∈ V . In this case we have 𝜋(𝛼, v) = 𝛼v.
• Inner product between two vectors in V . In this case 𝜋(u, v) = u ⋅ v.
• Inner product between two tensors, then 𝜋(A, T) = A ⋅ T.
• Tensor product between vectors, then 𝜋(u, v) = u ⊗ v.
• Application product of a second-order tensor T over a vector v. In this case it is
𝜋(T, v) = Tv.
All these operations have a common feature: they are all linear with respect to each
argument. In this case the operation is said to be bilinear. Hence, and seeking a general
rule, let us consider the following abstract product operation
𝜋 ∶ ℱ ×𝒢 → 𝒴 bilinear,
(2.23)
(f , g) → 𝜋(f , g),
where ℱ , 𝒢 and 𝒴 are finite-dimensional normed vector spaces. Consider now the
functions
f ∶ 𝒜 ⊂ 𝒳 → ℱ, g ∶ 𝒜 ⊂ 𝒳 → 𝒢, (2.24)
where 𝒳 is a finite-dimensional vector space and 𝒜 is an open subset of 𝒳 . With these
elements, the product h = 𝜋(f , g) is the function
h ∶ 𝒜 → 𝒴,
(2.25)
x → h(x) = 𝜋(f (x), g(x)).
Therefore, we have the following differentiation rule.
Product Rule. Let f and g be differentiable functions at x ∈ 𝒜 . Then, its product h =
𝜋(f , g) is also differentiable at x, and it is such that
𝒟 h(x)[u] = 𝜋(𝒟 f (x)[u], g(x)) + 𝜋(f (x), 𝒟 g(x)[u]). (2.26)
In particular, when 𝒜 is an open subset of ℝ, the product rule and the fact that 𝜋 is
bilinear lead to
̇
h(t) = 𝜋(ḟ (t), g(t)) + 𝜋(f (t), g(t)).
̇ (2.27)
Making use of this result, we obtain the following proposition. Let 𝜙, v, w, S and T
smooth functions in an open subset of ℝ such that the image of 𝜙 is ℝ, that of v, w is V ,
and for S, T it is Lin. Then, it is straightforward to prove that
̇ ̇ + 𝜙v,
(𝜙w) = 𝜙v ̇ (2.28)
̇
(v ⋅ w) = v̇ ⋅ w + v ⋅ w,
̇ (2.29)
̇ ̇ + TS, ̇
(TS) = TS (2.30)
̇
(T ⋅ S) = Ṫ ⋅ S + T ⋅ S,
̇ (2.31)
̇ ̇ + Tv.
(Tv) = Tv ̇ (2.32)
Exercise 2.1 Prove identities (2.28)–(2.32).
Another result frequently used in practice is the chain rule. Let 𝒳 , 𝒴 and ℱ be
finite-dimensional vector spaces (here we include the Euclidean point space ℰ). 𝒜 and
𝒞 are open subsets of 𝒳 and 𝒴 , respectively. Let f and g be functions such that
g ∶ 𝒜 → 𝒴, f ∶ 𝒞 → ℱ, (2.33)
and where R(g) ⊂ 𝒞 . Then, we have the following differentiation rule for the composi-
tion of functions.
Chain Rule. Let g and f be differentiable functions at x ∈ 𝒜 and y = g(x), correspond-
ingly. Then, the composition
h=f ○ g ∶𝒜 → ℱ, (2.34)
defined by
h(x) = f (g(x)), (2.35)
is differentiable at x, and the derivative is given by
𝒟 h(x) = 𝒟 f (y) ○ 𝒟 g(x), (2.36)
and its differential at x in the direction of u ∈ 𝒳 is
𝒟 h(x)[u] = 𝒟 f (g(x))[𝒟 g(x)[u]]. (2.37)
In particular, when 𝒳 ≡ ℝ, g and therefore h are real-valued functions. Putting t instead

of x in the previous expression we have
𝒟 h(t)[𝛼] = 𝒟 f (g(t))[𝒟 g(t)[𝛼]], (2.38)
which leads us to
[ ]
d d d
h(t) = f (g(t)) = 𝒟 f (g(t)) g(t) . (2.39)
dt dt dt
With this result we can prove that, for instance, the following holds
̇ ̇ T.
(ST ) = (S) (2.40)
Exercise 2.2 Prove identity (2.40).
A particularly useful result is to calculate the derivative of the determinant of a

second-order tensor. In order to do this we recall (1.165) for a second-order tensor T,
and with 𝛼 = −1, which leads to
1
det(I + T) = 1 + trT + ((trT)2 + tr(T2 )) + det T, (2.41)
2
and hence
det(I + T) = 1 + trT + o(T). (2.42)
Consider now the following
det(A + B) = det((I + BA−1 )A) = (det A) det(I + BA−1 ). (2.43)
Using now (2.42) into (2.43) we obtain
det(A + B) = det A(1 + tr(BA−1 ) + o(BA−1 ))

= det A + (det A)tr(BA−1 ) + (det A)o(BA−1 ). (2.44)
Clearly, from definition (2.12), we obtain
𝒟 (det A)[B] = (det A)tr(BA−1 ). (2.45)
From this result, it is straightforward to see that

̇ ̇ −1 ).
(det A) = (det A)tr(AA (2.46)
2.2 Gradient
In this section the definition given in (2.11) will be made specific for functions defined
in the open set 𝛺 of the Euclidean space ℰ where an origin O will be adopted, allow-
ing the biunivocal correspondence between points X ∈ ℰ and vectors x = X − O ∈ V .
Function 𝜙 in 𝛺 is said to be real-valued, vector-valued, tensor-valued (of any order),
or point-valued if the elements in the image are, for each x ∈ 𝛺, real numbers, vectors,
tensors (of any order), or points from ℰ (equivalently vectors from V ). For such func-
tions 𝜙 (assumed to be differentiable) we have seen that 𝒟 𝜙(x) ∈ Lin(𝒱 , 𝒴 ), where 𝒱
is the normed space associated with V , and where the normed space 𝒴 will denote the
normed spaces associated with ℝ, V , Lin, Lin, accordingly.
This linear application 𝒟 𝜙(x) will be called the gradient (and will be denoted by ∇𝜙(x)
or simply by ∇𝜙) of the field 𝜙 which, as seen in (2.12), is such that it approaches the
value of 𝜙 at x + u ∈ 𝛺 with an error of order o(u), in the following sense
𝜙(x + u) = 𝜙(x) + ∇𝜙(x)[u] + o(u). (2.47)
If we adopt an orthonormal basis for V , say {ei }, the previous expression is equivalent
to
1
∇𝜙(x)[ek ] = lim [𝜙(x + 𝛼ek ) − 𝜙(x)]. (2.48)
𝛼→0 𝛼
As can be easily seen, the limit is simply the partial derivative of 𝜙 with respect to xk , at
point x, that is
𝜕𝜙(x)
∇𝜙(x)[ek ] = . (2.49)
𝜕xk
Observe that, for a field in a finite-dimensional space of dimension dim 𝒴 , the
gradient is a transformation which belongs to a space of dimension (dim 𝒴 )(dim V ).
Result (2.49) allows us to easily define the representation of real-valued functions,
vector-valued functions, and tensor-valued functions, both in compact (intrinsic)
notation as well as in indicial notation, as shown in the following.
• Let 𝜙 be a real-valued function (also called scalar field) 𝜑, then
𝜕𝜑
∇𝜑 = e. (2.50)
𝜕xk k
• Let 𝜙 be a vector-valued function (also called a vector field) u, then
𝜕u 𝜕uj
∇u = ⊗ ek = (e ⊗ ek ). (2.51)
𝜕xk 𝜕xk j
• Let 𝜙 be a second-order tensor-valued function (also called a second-order tensor
field) S, then
𝜕S 𝜕Sij
∇S = ⊗ ek = (e ⊗ ej ⊗ ek ). (2.52)
𝜕xk 𝜕xk i
• Let 𝜙 be a third-order tensor-valued function (also called a third-order tensor field)
S, then
𝜕S 𝜕Sijr
∇S = ⊗ ek = (e ⊗ ej ⊗ er ⊗ ek ). (2.53)
𝜕xk 𝜕xk i
2.2 Gradient 29
• Let 𝜙 be a fourth-order tensor-valued function (also called a fourth-order tensor field)

𝕊, then
𝜕𝕊 𝜕Sijmn
∇𝕊 = ⊗ ek = (e ⊗ ej ⊗ em ⊗ en ⊗ ek ). (2.54)
𝜕xk 𝜕xk i
In intrinsic notation, the gradient of a vector field u can be defined as the second-order
tensor ∇u that satisfies
(∇u)T a = ∇(u ⋅ a), (2.55)
for any arbitrary constant vector a ∈ V .
Analogously, the gradient of a second-order tensor field S can be defined as the
third-order tensor ∇S that satisfies
(∇S)T a = ∇(ST a), (2.56)
for any arbitrary constant vector a ∈ V .
Alternatively, we can define the gradient of the tensor field S as the third-order tensor
∇S such that, for any arbitrary constant a ∈ V verifies
(∇S)t a = ∇(Sa). (2.57)
Now, consider a vector field of the form 𝜙u, with constant and arbitrary u ∈ V . Then,
the following holds
∇(𝜙u) = u ⊗ ∇𝜙. (2.58)
In fact, considering an arbitrary constant a ∈ V , from (2.55) note that
[u ⊗ (∇𝜙)]T a = [(∇𝜙) ⊗ u]a = (∇𝜙)u ⋅ a = ∇(𝜙u ⋅ a) = [∇(𝜙u)]T a. (2.59)
Then, expression (2.58) holds.
Consider now that we have a second-order tensor with the structure S = u ⊗ v. Then,
the following result can be proved
∇(u ⊗ v) = (u ⊗ ∇v) + ((∇u)T ⊗ v)T . (2.60)
Indeed, from definition (2.56), notice that
[∇(u ⊗ v)]T a = ∇((v ⊗ u)a) = ∇((a ⋅ u)v) = (a ⋅ u)∇v + v ⊗ ∇(a ⋅ u)
= (∇v ⊗ u)a + v ⊗ ((∇u)T a) = [(∇v ⊗ u) + (v ⊗ (∇u)T )]a. (2.61)
Then, using (1.195), this implies that
1 1 1
∇(u ⊗ v) = ((∇(u ⊗ v))T ) T = (∇v ⊗ u) T + (v ⊗ (∇u)T ) T
= (u ⊗ ∇v) + ((∇u)T ⊗ v)T , (2.62)
then identity (2.60) is verified.
As a consequence, we can define the second gradient of a vector field u as the
third-order tensor ∇∇u which satisfies the following
(∇∇u)T a = ∇((∇u)T a) = ∇(∇(u ⋅ a)), (2.63)
for any arbitrary constant a ∈ V . In a completely equivalent manner, ∇∇u can be defined
as
(∇∇u)a = ∇((∇u)a). (2.64)
Several results can be proved involving different types of functions. In effect, for 𝜑,
u (v), S (T) and S scalar, vector, tensor (second and third order), respectively, the follow-
ing results hold
∇(ST u) = (∇S)T u + ST ∇u, (2.65)
t
∇(Su) = (∇S) u + S∇u, (2.66)
∇(𝜑u) = 𝜑∇u + u ⊗ ∇𝜑, (2.67)
∇(v ⋅ u) = (∇v) u + (∇u) v,
T T
(2.68)
∇(𝜑S) = 𝜑(∇S) + S ⊗ ∇𝜑, (2.69)
1 1
∇(S ⋅ T) = (∇S) T + (∇T) S,
T T (2.70)
1
(∇(∇u)) = ((∇∇u)T )t ,
T (2.71)
1
T
(∇(∇u)) = ∇((∇u) ).
T (2.72)
Exercise 2.3 Prove the validity of expressions (2.65)–(2.72).
2.3 Divergence
With the definition of the gradient of a vector field, as well as the gradient of a tensor
field (of any order), we can introduce a new linear operator called divergence of 𝜙 and
denoted by div 𝜙. This operator can be represented by a (linear) operation involving the
gradient ∇𝜙 and the second-order identity tensor I, which takes the gradient ∇𝜙, which
belongs to a space of dimension (dim 𝒴 )(dim V ), into a space of dimension dim 𝒴 . We
can reinterpret the divergence operation as opposed to the gradient in this specific sense.
Then, we have
div 𝜙 = c(∇𝜙, I), (2.73)
which corresponds to the following specific definitions.
• Let 𝜙 be a vector field u, then
div u = ∇u ⋅ I, (2.74)
which in Cartesian coordinates is
𝜕u 𝜕uk
div u = ⋅e = . (2.75)
𝜕xk k 𝜕xk
• Let 𝜙 be a second-order tensor field S, then
div S = (∇S)I, (2.76)
𝜕S 𝜕Sjk
div S = ek = e. (2.77)
𝜕xk 𝜕xk j
2.3 Divergence 31
• Let 𝜙 be a third-order tensor field S, then

div S = (∇S)I, (2.78)
𝜕S 𝜕Sijk
div S = ek = (e ⊗ ej ). (2.79)
𝜕xk 𝜕xk i
• Let 𝜙 be a fourth-order tensor field 𝕊, then
div 𝕊 = (∇𝕊)I, (2.80)
𝜕𝕊 𝜕Sijrk
div 𝕊 = ek = (e ⊗ ej ⊗ er ). (2.81)
𝜕xk 𝜕xk i
Alternatively, but in a completely analogous way, we can define the divergence of a

second-order tensor field S as the unique vector field div S with the property
(div S) ⋅ a = div (ST a), (2.82)
for any arbitrary constant a ∈ V .
Similarly, for a third-order tensor S we can define
1
(div S) ⋅ A = div (S T A), (2.83)
for any arbitrary constant A ∈ Lin.
Now, it is possible to show that
div (ST) = (∇S)T + Sdiv T. (2.84)
In fact, using definitions (2.74) and (2.82) we have
div (ST) ⋅ a = div ((ST)T a) = div (TT ST a) = I ⋅ [TT ∇(ST a)] + ST a ⋅ div T
= T ⋅ ∇(ST a) + a ⋅ S div T = (∇S)T ⋅ a + a ⋅ S div T
= [(∇S)T + S div T] ⋅ a, (2.85)
and identity (2.84) follows.
Consider now the identity tensor T = I, then it follows directly that
div S = (∇S)I, (2.86)
which establishes the connection with definition (2.76).
Consider now a second-order tensor with the structure S = u ⊗ v, then from the pre-
vious results it follows that
div (u ⊗ v) = u div v + (∇u)v. (2.87)
In fact, from definition (2.76) and using expression (2.60) yields
div (u ⊗ v) = (∇(u ⊗ v))I = (u ⊗ ∇v)I + ((∇u)T ⊗ v)T I
= u(∇v ⋅ I) + (∇u I)v = u div v + (∇u)v, (2.88)
and (2.87) follows.
Another random document with
no related content on Scribd:
PLEASE READ THIS BEFORE YOU DISTRIBUTE OR USE THIS WORK
To protect the Project Gutenberg™ mission of promoting the free

distribution of electronic works, by using or distributing this work (or
any other work associated in any way with the phrase “Project
Gutenberg”), you agree to comply with all the terms of the Full
Project Gutenberg™ License available with this file or online at
www.gutenberg.org/license.
Section 1. General Terms of Use and

Redistributing Project Gutenberg™
electronic works
1.A. By reading or using any part of this Project Gutenberg™
electronic work, you indicate that you have read, understand, agree
to and accept all the terms of this license and intellectual property
(trademark/copyright) agreement. If you do not agree to abide by all
the terms of this agreement, you must cease using and return or
destroy all copies of Project Gutenberg™ electronic works in your
possession. If you paid a fee for obtaining a copy of or access to a
Project Gutenberg™ electronic work and you do not agree to be
bound by the terms of this agreement, you may obtain a refund from
the person or entity to whom you paid the fee as set forth in
paragraph 1.E.8.
1.B. “Project Gutenberg” is a registered trademark. It may only be

used on or associated in any way with an electronic work by people
who agree to be bound by the terms of this agreement. There are a
few things that you can do with most Project Gutenberg™ electronic
works even without complying with the full terms of this agreement.
See paragraph 1.C below. There are a lot of things you can do with
Project Gutenberg™ electronic works if you follow the terms of this
agreement and help preserve free future access to Project
Gutenberg™ electronic works. See paragraph 1.E below.
1.C. The Project Gutenberg Literary Archive Foundation (“the
Foundation” or PGLAF), owns a compilation copyright in the
collection of Project Gutenberg™ electronic works. Nearly all the
individual works in the collection are in the public domain in the
United States. If an individual work is unprotected by copyright law in
the United States and you are located in the United States, we do
not claim a right to prevent you from copying, distributing,
performing, displaying or creating derivative works based on the
work as long as all references to Project Gutenberg are removed. Of
course, we hope that you will support the Project Gutenberg™
mission of promoting free access to electronic works by freely
sharing Project Gutenberg™ works in compliance with the terms of
this agreement for keeping the Project Gutenberg™ name
associated with the work. You can easily comply with the terms of
this agreement by keeping this work in the same format with its
attached full Project Gutenberg™ License when you share it without
charge with others.
1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside the
United States, check the laws of your country in addition to the terms
of this agreement before downloading, copying, displaying,
performing, distributing or creating derivative works based on this
work or any other Project Gutenberg™ work. The Foundation makes
no representations concerning the copyright status of any work in
any country other than the United States.
1.E. Unless you have removed all references to Project Gutenberg:
1.E.1. The following sentence, with active links to, or other

immediate access to, the full Project Gutenberg™ License must
appear prominently whenever any copy of a Project Gutenberg™
work (any work on which the phrase “Project Gutenberg” appears, or
with which the phrase “Project Gutenberg” is associated) is
accessed, displayed, performed, viewed, copied or distributed:
This eBook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this eBook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.
1.E.2. If an individual Project Gutenberg™ electronic work is derived

from texts not protected by U.S. copyright law (does not contain a
notice indicating that it is posted with permission of the copyright
holder), the work can be copied and distributed to anyone in the
United States without paying any fees or charges. If you are
redistributing or providing access to a work with the phrase “Project
Gutenberg” associated with or appearing on the work, you must
comply either with the requirements of paragraphs 1.E.1 through
1.E.7 or obtain permission for the use of the work and the Project
Gutenberg™ trademark as set forth in paragraphs 1.E.8 or 1.E.9.
1.E.3. If an individual Project Gutenberg™ electronic work is posted

with the permission of the copyright holder, your use and distribution
must comply with both paragraphs 1.E.1 through 1.E.7 and any
additional terms imposed by the copyright holder. Additional terms
will be linked to the Project Gutenberg™ License for all works posted
with the permission of the copyright holder found at the beginning of
this work.
1.E.4. Do not unlink or detach or remove the full Project

Gutenberg™ License terms from this work, or any files containing a
part of this work or any other work associated with Project
Gutenberg™.
1.E.5. Do not copy, display, perform, distribute or redistribute this

electronic work, or any part of this electronic work, without
prominently displaying the sentence set forth in paragraph 1.E.1 with
active links or immediate access to the full terms of the Project
Gutenberg™ License.
1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if you
provide access to or distribute copies of a Project Gutenberg™ work
in a format other than “Plain Vanilla ASCII” or other format used in
the official version posted on the official Project Gutenberg™ website
(www.gutenberg.org), you must, at no additional cost, fee or expense
to the user, provide a copy, a means of exporting a copy, or a means
of obtaining a copy upon request, of the work in its original “Plain
Vanilla ASCII” or other form. Any alternate format must include the
full Project Gutenberg™ License as specified in paragraph 1.E.1.
1.E.7. Do not charge a fee for access to, viewing, displaying,

performing, copying or distributing any Project Gutenberg™ works
unless you comply with paragraph 1.E.8 or 1.E.9.
1.E.8. You may charge a reasonable fee for copies of or providing

access to or distributing Project Gutenberg™ electronic works
provided that:
• You pay a royalty fee of 20% of the gross profits you derive from
the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information
about donations to the Project Gutenberg Literary Archive
Foundation.”
• You provide a full refund of any money paid by a user who

notifies you in writing (or by e-mail) within 30 days of receipt that
s/he does not agree to the terms of the full Project Gutenberg™
License. You must require such a user to return or destroy all
copies of the works possessed in a physical medium and
discontinue all use of and all access to other copies of Project
Gutenberg™ works.
• You provide, in accordance with paragraph 1.F.3, a full refund of

any money paid for a work or a replacement copy, if a defect in
the electronic work is discovered and reported to you within 90
days of receipt of the work.
• You comply with all other terms of this agreement for free
distribution of Project Gutenberg™ works.
1.E.9. If you wish to charge a fee or distribute a Project Gutenberg™

electronic work or group of works on different terms than are set
forth in this agreement, you must obtain permission in writing from
the Project Gutenberg Literary Archive Foundation, the manager of
the Project Gutenberg™ trademark. Contact the Foundation as set
forth in Section 3 below.
1.F.
1.F.1. Project Gutenberg volunteers and employees expend

considerable effort to identify, do copyright research on, transcribe
and proofread works not protected by U.S. copyright law in creating
the Project Gutenberg™ collection. Despite these efforts, Project
Gutenberg™ electronic works, and the medium on which they may
be stored, may contain “Defects,” such as, but not limited to,
incomplete, inaccurate or corrupt data, transcription errors, a
copyright or other intellectual property infringement, a defective or
damaged disk or other medium, a computer virus, or computer
codes that damage or cannot be read by your equipment.
1.F.2. LIMITED WARRANTY, DISCLAIMER OF DAMAGES - Except

for the “Right of Replacement or Refund” described in paragraph
1.F.3, the Project Gutenberg Literary Archive Foundation, the owner
of the Project Gutenberg™ trademark, and any other party
distributing a Project Gutenberg™ electronic work under this
agreement, disclaim all liability to you for damages, costs and
expenses, including legal fees. YOU AGREE THAT YOU HAVE NO
REMEDIES FOR NEGLIGENCE, STRICT LIABILITY, BREACH OF
WARRANTY OR BREACH OF CONTRACT EXCEPT THOSE
PROVIDED IN PARAGRAPH 1.F.3. YOU AGREE THAT THE
FOUNDATION, THE TRADEMARK OWNER, AND ANY
DISTRIBUTOR UNDER THIS AGREEMENT WILL NOT BE LIABLE
TO YOU FOR ACTUAL, DIRECT, INDIRECT, CONSEQUENTIAL,
PUNITIVE OR INCIDENTAL DAMAGES EVEN IF YOU GIVE
NOTICE OF THE POSSIBILITY OF SUCH DAMAGE.
1.F.3. LIMITED RIGHT OF REPLACEMENT OR REFUND - If you

discover a defect in this electronic work within 90 days of receiving it,
you can receive a refund of the money (if any) you paid for it by
sending a written explanation to the person you received the work
from. If you received the work on a physical medium, you must
return the medium with your written explanation. The person or entity
that provided you with the defective work may elect to provide a
replacement copy in lieu of a refund. If you received the work
electronically, the person or entity providing it to you may choose to
give you a second opportunity to receive the work electronically in
lieu of a refund. If the second copy is also defective, you may
demand a refund in writing without further opportunities to fix the
problem.
1.F.4. Except for the limited right of replacement or refund set forth in
paragraph 1.F.3, this work is provided to you ‘AS-IS’, WITH NO
OTHER WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR ANY PURPOSE.
1.F.5. Some states do not allow disclaimers of certain implied

warranties or the exclusion or limitation of certain types of damages.
If any disclaimer or limitation set forth in this agreement violates the
law of the state applicable to this agreement, the agreement shall be
interpreted to make the maximum disclaimer or limitation permitted
by the applicable state law. The invalidity or unenforceability of any
provision of this agreement shall not void the remaining provisions.
1.F.6. INDEMNITY - You agree to indemnify and hold the

Foundation, the trademark owner, any agent or employee of the
Foundation, anyone providing copies of Project Gutenberg™
electronic works in accordance with this agreement, and any
volunteers associated with the production, promotion and distribution
of Project Gutenberg™ electronic works, harmless from all liability,
costs and expenses, including legal fees, that arise directly or
indirectly from any of the following which you do or cause to occur:
(a) distribution of this or any Project Gutenberg™ work, (b)
alteration, modification, or additions or deletions to any Project
Gutenberg™ work, and (c) any Defect you cause.
Section 2. Information about the Mission of

Project Gutenberg™
Project Gutenberg™ is synonymous with the free distribution of
electronic works in formats readable by the widest variety of
computers including obsolete, old, middle-aged and new computers.
It exists because of the efforts of hundreds of volunteers and
donations from people in all walks of life.
Volunteers and financial support to provide volunteers with the

assistance they need are critical to reaching Project Gutenberg™’s
goals and ensuring that the Project Gutenberg™ collection will
remain freely available for generations to come. In 2001, the Project
Gutenberg Literary Archive Foundation was created to provide a
secure and permanent future for Project Gutenberg™ and future
generations. To learn more about the Project Gutenberg Literary
Archive Foundation and how your efforts and donations can help,
see Sections 3 and 4 and the Foundation information page at
www.gutenberg.org.
Section 3. Information about the Project
Gutenberg Literary Archive Foundation
The Project Gutenberg Literary Archive Foundation is a non-profit
501(c)(3) educational corporation organized under the laws of the
state of Mississippi and granted tax exempt status by the Internal
Revenue Service. The Foundation’s EIN or federal tax identification
number is 64-6221541. Contributions to the Project Gutenberg
Literary Archive Foundation are tax deductible to the full extent
permitted by U.S. federal laws and your state’s laws.
The Foundation’s business office is located at 809 North 1500 West,

Salt Lake City, UT 84116, (801) 596-1887. Email contact links and up
to date contact information can be found at the Foundation’s website
and official page at www.gutenberg.org/contact
Section 4. Information about Donations to

the Project Gutenberg Literary Archive
Foundation
Project Gutenberg™ depends upon and cannot survive without
widespread public support and donations to carry out its mission of
increasing the number of public domain and licensed works that can
be freely distributed in machine-readable form accessible by the
widest array of equipment including outdated equipment. Many small
donations ($1 to $5,000) are particularly important to maintaining tax
exempt status with the IRS.
The Foundation is committed to complying with the laws regulating

charities and charitable donations in all 50 states of the United
States. Compliance requirements are not uniform and it takes a
considerable effort, much paperwork and many fees to meet and
keep up with these requirements. We do not solicit donations in
locations where we have not received written confirmation of
compliance. To SEND DONATIONS or determine the status of
compliance for any particular state visit www.gutenberg.org/donate.
While we cannot and do not solicit contributions from states where

we have not met the solicitation requirements, we know of no
prohibition against accepting unsolicited donations from donors in
such states who approach us with offers to donate.
International donations are gratefully accepted, but we cannot make

any statements concerning tax treatment of donations received from
outside the United States. U.S. laws alone swamp our small staff.
Please check the Project Gutenberg web pages for current donation
methods and addresses. Donations are accepted in a number of
other ways including checks, online payments and credit card
donations. To donate, please visit: www.gutenberg.org/donate.
Section 5. General Information About Project

Gutenberg™ electronic works
Professor Michael S. Hart was the originator of the Project
Gutenberg™ concept of a library of electronic works that could be
freely shared with anyone. For forty years, he produced and
distributed Project Gutenberg™ eBooks with only a loose network of
volunteer support.
Project Gutenberg™ eBooks are often created from several printed

editions, all of which are confirmed as not protected by copyright in
the U.S. unless a copyright notice is included. Thus, we do not
necessarily keep eBooks in compliance with any particular paper
edition.
Most people start at our website which has the main PG search
facility: www.gutenberg.org.
This website includes information about Project Gutenberg™,

including how to make donations to the Project Gutenberg Literary
Archive Foundation, how to help produce our new eBooks, and how
to subscribe to our email newsletter to hear about new eBooks.

Introduction To The Variational Formulation in Mechanics Edgardo O Taroco Full Chapter PDF

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

Introduction To The Variational Formulation in Mechanics Edgardo O Taroco Full Chapter PDF

Uploaded by

Copyright:

Available Formats

Introduction to the Variational

Formulation in Mechanics Edgardo O.

Introduction to Fracture Mechanics Robert O. Ritchie

eTextbook 978-1118763780 Introduction to Cosmetic

Introduction to Quantum Mechanics John Dirk Walecka

Introduction to Engineering Fluid Mechanics 1st Edition

PSYCHODYNAMIC FORMULATION : an expanded approach. The

Solutions Manual to Introduction to Robotics Mechanics

An Introduction to Statistical Mechanics and

Variational Calculus with Engineering Applications

Fundamentals and Applications

Edgardo O. Taroco, Pablo J. Blanco and Raúl A. Feijóo

Limit of Liability/Disclaimer of Warranty

Library of Congress Cataloging-in-Publication data applied for

Cover Design by Wiley

Set in 10/12pt Warnock by SPi Global, Pondicherry, India

Part I Vector and Tensor Algebra and Analysis 1

1 Vector and Tensor Algebra 3

2 Vector and Tensor Analysis 23

Part II Variational Formulations in Mechanics 47

3 Method of Virtual Power 49

3.4 Bodies without Constraints 74

4 Hyperelastic Materials at Inﬁnitesimal Strains 109

4.9 Castigliano Theorems 136

5 Materials Exhibiting Creep 165

6 Materials Exhibiting Plasticity 229

6.4.2 Yield Criterion and Hardening Law 246

Part III Modeling of Structural Components 281

7 Bending of Beams 285

8 Torsion of Bars 301

9 Plates and Shells 315

9.4 Principle of Virtual Power 323

Part IV Other Problems in Physics 377

10 Heat Transfer 379

11 Incompressible Fluid Flow 393

11.4 Navier–Stokes Equations 403

12 High-Order Continua 411

Part V Multiscale Modeling 435

13 Method of Multiscale Virtual Power 439

14 Applications of Multiscale Modeling 475

14.2.5 Tangent Operators 487

Part VI Appendices 501

A Deﬁnitions and Notations 503

B Elements of Real and Functional Analysis 539

C Functionals and the Gâteaux Derivative 555

Petrópolis, Brazil Pablo J. Blanco

Vector and Tensor Algebra and Analysis

Vector and Tensor Algebra

1.1 Points and Vectors

1 Vectors {ei } are called linearly independent if 𝛼i ei = 𝟎 implies 𝛼i = 0 for i = 1, 2, 3.

• Components of u with respect to {ei } are contravariant components, and are

If the basis is orthonormal, it is easy to show that it is identical to its dual

Exercise 1.1 Verify the validity of identities given by (1.14).

1.2 Second-Order Tensors

(n ⊗ n)v Figure 1.1 Geometric concept of orthogonal

which yields (1.63). Lastly, note that

• Skew-symmetric component of a tensor. Compact notation: W ∈ Skw ⇔ W = −WT ,

The determinant of a tensor A ∈ Lin can be deﬁned as

Exercise 1.3 Show that a tensor is orthogonal if and only if QQT = QT Q = I.

The cross product u × v between vectors u, v ∈ V , whose angle between them is 𝜃, is

Exercise 1.4 Prove that (1.157) holds.

When u, v, w are linearly independent, the value of the product u ⋅ (v × w) represents

⎧1 if (i, j, k) is an even permutation of (1, 2, 3),