“ho NYY! Finite element analysis in geotechnical engineering Theory [ David M.'Potts and Lidija Zdravkovié Le Imperial College of Science, Technology and Medicine au *“L! ThomasTelfordPublished by Thomas Telford Publishing, Thomas Telford Ltd, 1 Heron Quay, London E14 43D. URL: http://www.t-telford.co.uk Distributors for Thomas Telford books are USA: ASCE Press, 1801 Alexander Bell Drive, Reston, VA 20191-4400, USA Japan: Matuzen Co. Ltd, Book Department, 3-10 Nihonbashi 2-chome, Chuo-ku, Tokyo 103 Australia: DA Books and Journals, 648 Whitehorse Road, Mitcham 3132, Victoria First published 1999 Also available from Thomas Telford Books Finite element analysis in geotechnical engineering: application. ISBN 07277 2783 4 ‘A catalogue record for this book is available from the British Library ISBN: 0 7277 2753 2 © David M. Potts and Lidija Zdravkovié, and Thomas Telford Limited, 1999 All rights, including translation, reserved. Except for fair copying, 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 or otherwise, without the prior written permission of the Books Publisher, Thomas Telford Publishing, Thomas Telford Ltd, 1 Heron Quay, London E14 4D. ‘This book is published on the understanding that the author is/authors are solely responsible for the statements made and opinions expressed in it and that its publication does not necessarily imply that such statements and/or opinions are or reflect the views or opinions of the publishers. Printed and bound in Great Britain by Bookcraft (Bath) LimitedContents Preface Geotechnical analysis 11 1.2 13 L4 1.5 1.10 L.1 Finite 2.1 2.2 2.3 2.4 2.5 Synopsis Introduction Design objectives Design requirements Theoretical considerations 1.5.1 Requirements for a general solution 1.5.2 Equilibrium 1.5.3 Compatibility 1.5.4 Equilibrium and compatibility equations 1.5.5 Constitutive behaviour Geometric idealisation 1.6.1 Plane strain 1.6.2 Axi-symmetry Methods of analysis Closed form solutions Simple methods 1.9.1 Limit equilibrium 1.9.2 Stress field solution 1.9.3 Limit analysis 1.9.4 Comments Numerical analysis 1.10.1 Beam-spring approach 1.10.2 Full numerical analysis Summary element theory for linear materials Synopsis. Introduction Overview Element discretisation Displacement approximationii / Finite element analysis in geotechnical engineering: Theory 2.5.1 Isoparametric finite elements 29 2.6 Element equations 31 2.6.1 Numerical integration 34 2.7 Global equations 36 27.1 The direct stiffness assembly method 36 2.8 Boundary conditions 39 2.9 Solution of global equations 39 2.9.1 Storage of global stiffness matrix 40 2.9.2 Triangular decomposition of the global stiffness matrix 41 2.9.3 Solution of the finite element equations 43 2.9.4 Modification due to displacement boundary conditions 45 2.10 Calculation of stresses and strains 47 2.11 Example 47 2.12 Axi-symmetric finite element analysis 49 2.13. Summary 50 Appendix {1.1 Triangular finite elements 51 Ill Derivation of area coordinates 51 IL1.2_ Isoparametric formulation 53 3. Geotechnical considerations 55 3.1 Synopsis 55 3.2 Introduction 55 3.3 Total stress analysis 56 3.4 Pore pressure calculation 58 3.5 Finite elements to model structural components 61 3.5.1 Introduction 61 Strain definitions 62 Constitutive equation 63 Finite element formulation 64 Membrane elements 67 3.6 Finite elements to model interfaces 68 3.6.1 Introduction 68 3.6.2 Basic theory 69 3.6.3 Finite element formulation 70 3.6.4 Comments 72 3.7 Boundary conditions 2 Introduction 72 Local axes 73 Prescribed displacements 4 Tied degrees of freedom 16 Springs 78 Boundary stresses 80 Point loads 82 Body forces 833.8 Contents / iii 3.7.9 Construction 3.7.10 Excavation 3.7.11 Pore pressures Summary Real soil behaviour 4.1 42 4.3 44 45 4.6 AT Synopsis Introduction Behaviour of clay soils 43.1 Behaviour under one dimensional compression 4.3.2 Behaviour when sheared 4.3.3 Effect of stress path direction 43.4 Effect of the magnitude of the intermediate principal stress 43.5 Anisotropy 43.6 Behaviour at large strains Behaviour of sands 4.4.1 Behaviour under one dimensional compression 4.4.2 Behaviour when sheared 4.4.3 Effect of the magnitude of the intermediate principal stress 444 Anisotropy 4.4.5 Behaviour at large strains Behaviour of soils containing both clay and sand 4.5.1 Comparison of sedimentary soils 4.5.2 Residual soils 4.5.3 Residual strength Concluding remarks Summary Elastic constitutive models 5.1 5.2 53 54 5.5 5.6 5.7 Synopsis Introduction Invariants Elastic behaviour Linear isotropic elasticity Linear anisotropic elasticity Nonlinear elasticity 5.7.1 Introduction 5.7.2 Bi-linear model 5.7.3 K-Gmodel 5.7.4 Hyperbolic model 5.7.5 ‘Small strain stiffness model 5.7.6 Puzrin and Burland model 95 97 97 99 99 100 103 104 10S 105 105 110 11 2 112 114 14 4iv / Finite element analysis in geotechnical engineering: Theory 5.8 Summary 6. Elasto-plastic behaviour 6.1 6.2 63 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12 6.13 6.14 Synopsis Introduction Uniaxial behaviour of a linear elastic perfectly plastic material Uniaxial behaviour of a linear elastic strain hardening plastic material Uniaxial behaviour of a linear elastic strain softening plastic material Relevance to geotechnical engineering Extension to general stress and strain space Basic concepts 6.8.1 Coincidence of axes 6.8.2 A yield function 6.8.3 Aplastic potential function 6.8.4 The hardening/softening rules Two dimensional behaviour of a linear elastic perfectly plastic material Two dimensional behaviour of a linear elastic hardening plastic material Two dimensional behaviour of a linear elastic softening plastic material Comparison with real soil behaviour Formulation of the elasto-plastic constitutive matrix Summary 7. Simple elasto-plastic constitutive models 710 Synopsis Introduction Tresca model Von Mises model Mohr-Coulomb model Drucker-Prager model Comments on simple elastic perfectly plastic models An elastic strain hardening/softening Mohr-Coulomb model Development of the critical state models 791 Basic formulation in triaxial stress space 7.9.2 Extension to general stress space 7.9.3 Undrained strength Modifications to the basic formulation of critical state models 7101 ‘Yield surface on the supercritical side 7.10.2 Yield surface for K,, consolidated soils 133 134 134 135 135 136 136 136 137 138 139 140 141 142 143 146 147 147 147 148 150 151 155 157 158 160 161 166 168 169 169 171TAL Contents / v 7.10.3 Elastic component of the model 7.10.4 Plastic behaviour inside the main yield surface Alternative shapes for the yield and plastic potential surfaces for critical state models TALL Introduction 7.11.2 Development of a new expression in triaxial stress space 7.11.3 Generalisation of the expression 7.12. The effect of the plastic potential in plane strain deformation TAZ Summary Appendix VII.1 Derivatives of stress invariants Appendix VII.2 Analytical solutions for triaxial test on modified Cam clay VIL2.1 Drained triaxial test VIL2.2 Undrained triaxial test Appendix VII.3_ Derivatives for modified Cam clay model Appendix VIL4 Undrained strength for critical state models Advanced constitutive models 8.1 8.2 83 84 8.5 8.6 Synopsis Introduction Modelling of soil as a limited tension material 8.3.1 Introduction 8.3.2 Model formulation 8.3.2.1 Yield surface 8.3.2.2 Plastic potential 8.3.2.3 Finite element implementation Formulation of the elasto-plastic constitutive matrix when two yield surfaces are simultaneously active Lade’s double hardening model 8.5.1 Introduction 8.5.2 Overview of model 8.5.3 Elastic behaviour 8.5.4 Failure criterion 8.5.5 Conical yield function 8.5.6 Conical plastic potential function 8.5.7 Conical hardening law 8.5.8 Cap yield function 8.5.9 Cap plastic potential function 8.5.10 Cap hardening law 8.5.11 Comments Bounding surface formulation of soil plasticity 8.6.1 Introduction 8.6.2 Bounding surface plasticity 172 173 175 175 176 181 181 185 186 187 188 192 197vi / Finite element analysis in geotechnical engineering: Theory 8.7 MIT soil models 8.7.1 Introduction 8. Transformed variables 8.7.3 Hysteretic elasticity 8.7.4 Behaviour on the bounding surface 8.7.5 Behaviour within the bounding surface 8.7.6 Comments 8.8 Bubble models 8.8.1 Introduction 8.8.2 Behaviour of a kinematic yield surface 8.9 Al-Tabbaa and Wood model 8.9.1 Bounding surface and bubble 8.9.2 Movement of bubble 8.9.3 Elasto-plastic behaviour 8.9.4 Comments 8.10 Summary Appendix VIII.1 Derivatives for Lade’s double hardening model Finite element theory for nonlinear materials 91 9.2 93 o4 95 9.6 Synopsis Introduction Nonlinear finite element analysis Tangent stiffness method 94d Introduction 942 Finite element implementation 9.4.3 Uniform compression of a Mohr-Coulomb soil 9.4.4 Uniform compression of modified Cam clay soil Visco-plastic method 9.5.1 Introduction. 9.5.2 Finite element application 9.5.3 Choice of time step 9.5.4 Potential errors in the algorithm 9.5.5 Uniform compression of a Mohr-Coulomb soil 9.5.6 Uniform compression of modified Cam clay soil Modified Newton-Raphson method 9.6.1 Introduction 9.6.2 Stress point algorithms 9.6.2.1 Introduction 9.6.2.2 Substepping algorithm 9.6.2.3 Return algorithm 9.6.2.4 Fundamental comparison 9.6.3. Convergence criteria 9.6.4 Uniform compression of Mohr-Coulomb and modified Cam clay soils 215 215 215 250 260 260Contents / vii 9.7 Comparison of the solution strategies 971 9.7.2 9.7.3 9.7.4 9.7.5 9.7.6 9.8 Summary Appendix IX.1 IX1.1 1X12 IX.1.3 IX.1L.4 IX.1.5 IX.1.6 Appendix 1X.2 1X.2.1 IX.2.2 1X.2.3 X24 Appendix IX.3 1X31 1X32 1X.3.3 1X.3.4 1X.3.5 Introduction Idealised triaxial test Footing problem Excavation problem Pile problem Comments Substepping stress point algorithm Introduction Overview Modified Euler integration scheme with error control Runge-Kutta integration scheme Correcting for yield surface drift in elasto-plastic finite element analysis Nonlinear elastic behaviour Return stress point algorithm Introduction Overview Return algorithm proposed by Ortiz & Simo (1986) Return algorithm proposed by Borja & Lee (1990) Comparison of substepping and return algorithms Introduction Fundamental comparison 1X.3.2.1 Undrained triaxial test 1X.3,2.2 Drained triaxial test Pile problem Consistent tangent operators Conclusions 10. Seepage and consolidation 10.1 Synopsis 10.2 ‘Introduction 10.3. Finite element formulation for coupled problems 10.4 Finite element implementation 10.5 Steady state seepage 10.6 Hydraulic boundary conditions 10.6.1 10.6.2 10.6.3 10.6.4 10.6.5 10.6.6 Introduction Prescribed pore fluid pressures Tied degrees of freedom Infiltration Sources and sinks Precipitation 261 261 263 267 270 273 275 276 277 277 278 280 283 283 296 299 301 303 304 316viii / Finite element analysis in geotechnical engineering: Theory 10.7 Permeability models 318 10.7.1 Introduction 318 10.7.2. Linear isotropic permeability 318 10.7.3 Linear anisotropic permeability 319 10.7.4 Nonlinear permeability related to void ratio 319 10.7.5 Nonlinear permeability related to mean effective stress using a logarithmic relationship 320 10.7.6 Nonlinear permeability related to mean effective stress using a power law relationship 320 10.8 Unconfined seepage flow 320 10.9 Validation example 321 10.10 Summary 323 11. 3D finite element analysis 325 11.1 Synopsis 325 11.2 Introduction 325 11.3. Conventional 3D finite element analysis 326 11.4 _ Iterative solutions 332 11.4.1 Introduction 332 11.4.2 General iterative solution 332 11.4.3 The gradient method 334 11.4.4 The conjugate gradient method 336 11.4.5. Comparison of the conjugate gradient and banded solution techniques 337 11.4.6 Normalisation of the stiffness matrix 341 11.4.7 Comments 342 11.5 Summary 342 12. Fourier series aided finite element method (FSAFEM) 344 12.1 Synopsis 344 12.2 Introduction 344 12.3 The continuous Fourier series aided finite element method 345 12.3.1 Formulation for linear behaviour 345 12.3.2 Symmetrical loading conditions 352 12.3.3 Existing formulations for nonlinear behaviour 354 12.3.4 New formulation for nonlinear behaviour 355 12.3.5 Formulation for interface elements 359 12.3.6 Bulk pore fluid compressibility 361 12.3.7 Formulation for coupled consolidation 364 12.4 Implementation of the CRSAFEM 370 12.4.1 Introduction 370 12.4.2 Evaluating Fourier series harmonic coefficients 371 12.4.2.1 The stepwise linear method 372 12.4.2.2 The fitted method 373Contents / ix 12.4.3 The modified Newton-Raphson solution strategy 374 12.4.3.1 Introduction 374 12.4.3.2 Right hand side correction 375 12.4.4 Data storage 376 12.4.5 Boundary conditions 377 12.4.6 Stiffness matrices 377 12.4.7 Simplification due to symmetrical boundary conditions 378 12.4.7.1 Introduction 378 12.4.7.2 Examples of problems associated with parallel and orthogonal analysis 380 12.5 The discrete Fourier series aided finite element method 385 12.5.1 Introduction. 385 12.5.2 Description of the discrete FFAFEM method 386 12.6 Comparison between the discrete and the continuous FSAFEM 391 12.7 Comparison of CFSAFEM and the 3D analysis 396 12.8 Summary 398 Appendix XII.1 Harmonic coefficients of force from harmonic point loads 399 Appendix XIL.2 Obtaining the harmonics of force from harmonic boundary stresses 400 Appendix X1I.3_ Obtaining the harmonics of force from element stresses 401 Appendix XIL4 Resolving harmonic coefficients of nodal force 403 Appendix XII.5_ Fourier series solutions for integrating the product of three Fourier series 404 Appendix XII.6 Obtaining coefficients for a stepwise linear distribution 405 Appendix XI.7 Obtaining harmonic coefficients for the fitted method 407 References All List of symbols 425 Index 435Preface While the finite element method has been used in many fields of engineering practice for over thirty years, it is only relatively recently that it has begun to be widely used for analysing geotechnical problems. This is probably because there are many complex issues which are specific to geotechnical engineering and which have only been resolved relatively recently. Perhaps this explains why there are few books which cover the application of the finite element method to geotechnical engineering. For over twenty years we, at Imperial College, have been working at the leading edge of the application of the finite clement method to the analysis of practical geotechnical problems. Consequently, we have gained enormous experience of this type of work and have shown that, when properly used, this method can produce realistic results which are of value to practical engineering problems. Because we have written all our own computer code, we also have an in-depth understanding of the relevant theory. Based on this experience we believe that, to perform useful geotechnical finite element analysis, an engineer requires specialist knowledge in a range of subjects. Firstly, a sound understanding of soil mechanics and finite element theory is required. Secondly, an in-depth understanding and appreciation of the limitations of the various constitutive models that are currently available is needed. Lastly, users must be fully conversant with the manner in which the software they are using works. Unfortunately, it is not easy for a geotechnical engineer to gain all these skills, as it is vary rare for all of them to be part of a single undergraduate or postgraduate degree course. It is perhaps, therefore, not surprising that many engineers, who carry out such analyses and/or use the results from such analyses, are not aware of the potential restrictions and pitfalls involved. This problem was highlighted when we recently gave a four day course on numerical analysis in geotechnical engineering. Although the course was a great success, attracting many participants from both industry and academia, it did highlight the difficulty that engineers have in obtaining the necessary skills required to perform good numerical analysis. In fact, it was the delegates on this course who urged us, and provided the inspiration, to write this book. The overall objective of the book is to provide the reader with an insight into the use of the finite element method in geotechnical engineering. More specific aims are:xii / Finite element analysis in geotechnical engineering: Theory - To present the theory, assumptions and approximations involved in finite element analysis; - To describe some of the more popular constitutive models currently available and explore their strengths and weaknesses; - To provide sufficient information so that readers can assess and compare the capabilities of available commercial software; - To provide sufficient information so that readers can make judgements as to the credibility of numerical results that they may obtain, or review, in the future; - To show, by means of practical examples, the restrictions, pitfalls, advantages and disadvantages of numerical analysis. The book is primarily aimed at users of commercial finite element software both in industry and in academia. However, it will also be of use to students in their final years of an undergraduate course, or those on a postgraduate course in geotechnical engineering. A prime objective has been to present the material in the simplest possible way and in manner understandable to most engineers. Consequently, we have refrained from using tensor notation and presented all theory in terms of conventional matrix algebra. When we first considered writing this book, it became clear that we could not cover all aspects of numerical analysis relevant to geotechnical engineering. We reached this conclusion for two reasons. Firstly, the subject area is so vast that to adequately cover it would take many volumes and, secondly, we did not have experience with all the different aspects. Consequently, we decided only to include material which we felt we had adequate experience of and that was useful to a practising engineer. As a result we have concentrated on static behaviour and have not considered dynamic effects. Even so, we soon found that the material we wished to include would not sensibly fit into a single volume. The material has therefore been divided into theory and application, each presented in a separate volume. Volume | concentrates on the theory behind the finite element method and on the various constitutive models currently available. This is essential reading for any user of a finite element package as it clearly outlines the assumptions and limitations involved. Volume 2 concentrates on the application of the method to real geotechnical problems, highlighting how the method can be applied, its advantages and disadvantages, and some of the pitfalls. This is also essential reading for a user of a software package and for any engineer who is commissioning and/or reviewing the results of finite element analyses. This volume of the book (i.e. Volume 1) consists of twelve chapters. Chapter | considers the general requirements of any form of geotechnical analysis and provides a framework for assessing the relevant merits of the different methods of analysis currently used in geotechnical design. This enables the reader to gain an insight into the potential advantage of numerical analysis over the morePreface / xiii ‘conventional’ approaches currently in use. The basic finite element theory for linear material behaviour is described in Chapter 2. Emphasis is placed on highlighting the assumptions and limitations. Chapter 3 then presents the modifications and additions that are required to enable geotechnical analysis to be performed. The main limitation of the basic finite element theory is that it is based on the assumption of linear material behaviour. Soils do not behave in such a manner and Chapter 4 highlights the important facets of soil behaviour that ideally should be accounted for by a constitutive model. Unfortunately, a constitutive model which can account for al] these facets of behaviour, and at the same time be defined by a realistic number of input parameters which can readily be determined from simple laboratory tests, does not exist. Nonlinear elastic constitutive models are presented in Chapter 5 and although these are an improvement over the linear elastic models that were used in the early days of finite element analyses, they suffer severe limitations. The majority of constitutive models currently in use are based on the framework of elasto-plasticity and this is described in Chapter 6. Simple elasto-plastic models are then presented in Chapter 7 and more complex models in Chapter 8. To use these nonlinear constitutive models in finite element analysis requires an extension of the theory presented in Chapter 2. This is described in Chapter 9 where some of the most popular nonlinear solution strategies are considered. It is shown that some of these can result in large errors unless extreme care is exercised by the user. The procedures required to obtain accurate solutions are discussed. Chapter 10 presents the finite element theory for analysing coupled problems involving both deformation and pore fluid flow. This enables time dependent consolidation problems to be analysed. Three dimensional problems are considered in Chapter 11. Such problems require large amounts of computer resources and methods for reducing these are discussed. In particular the use of iterative equation solvers is considered. While these have been used successfully in other branches of engineering, it is shown that, with present computer hardware, they are unlikely to be economical for the majority of geotechnical problems. The theory behind Fourier Series Aided Finite Element Analysis is described in Chapter 12. Such analysis can be applied to three dimensional problems which Possess an axi-symmetric geometry but a non axi-symmetric distribution of material properties and/or loading. It is shown that analyses based on this approach can give accurate results with up to an order of magnitude saving in computer resources compared to equivalent analyses performed with a conventional three dimensional finite element formulation. Volume 2 of this book builds on the material given in this volume. However, the emphasis is less on theory and more on the application of the finite element method in engineering practice. Topics such as obtaining geotechnical parametersxiv / Finite element analysis in geotechnical engineering: Theory from standard laboratory and field tests and the analysis of tunnels, earth retaining structures, cut slopes, embankments and foundations are covered. A chapter on benchmarking is also included. Emphasis is placed on explaining how the finite element method should be applied and what are the restrictions and pitfalls. In particular, the choice of suitable constitutive models for the various geotechnical boundary value problems is discussed at some length. To illustrate the material presented, examples from the authors experiences with practical geotechnical problems are used. Although we have edited this volume, and written much of the content, several of the chapters involve contributions from our colleagues at Imperial College. All the numerical examples presented in both this volume and Volume 2 of this book have been obtained using the Authors’ own computer code. This software is not available commercially and therefore the results presented are unbiased. As commercial software has not been used, the reader must consider what implications the results may have on the use of such software. London David M. Potts November 1998 Lidija Zdravkovié1. Geotechnical analysis 1.1. Synopsis This chapter considers the analysis of geotechnical structures. Design requirements are presented, fundamental theoretical considerations are discussed and the various methods of analysis categorised. The main objective of the chapter is to provide a framework in which different methods of analysis may be compared. This will provide an insight into the potential advantages of numerical analysis over the more ‘conventional’ approaches currently in use. _— os Cut slope Embankment Raft foundation Piled foundation Gravity wall Embedded wall Figure 1.1: Examples of geotechnical structures2 | Finite element analysis in geotechnical engineering: Theory 1.2 Introduction Nearly all civil engineering structures involve the ground in some way. Cut slopes, earth and rockfill embankments, see Figure 1.1, are made from geological materials. The soil (or rock) provides both the destabilising and stabilising forces which maintain equilibrium of the structure. Raft and piled foundations transfer loads from buildings, bridges and offshore structures to be resisted by the ground, Retaining walls enable vertical excavations to be made. [n most situations the soil provides both the activating and resisting forces, with the wall and its structural support providing a transfer mechanism. Geotechnical engineering, therefore, plays a major role in the design of civil engineering structures. The design engineer must assess the forces imposed in the soil and structural members, and the potential movements of both the structure and the surrounding soil. Usually these have to be determined under both working and ultimate load conditions. Traditionally geotechnical design has been carried out using simplified analyses or empirical approaches. Most design codes or advice manuals are based on such approaches. The introduction of inexpensive, but sophisticated, computer hardware and software has resulted in considerable advances in the analysis and design of geotechnical structures. Much progress has been made in attempting to model the behaviour of geotechnical structures in service and to investigate the mechanisms of soil-structure interaction. At present, there are many different methods of calculation available for analysing geotechnical structures. This can be very confusing to an inexperienced geotechnical engineer. This chapter introduces geotechnical analysis. The basic theoretical considerations are discussed and the various methods of analysis categorised. The main objectives are to describe the analysis procedures that are in current use and to provide a framework in which the different methods of analysis may be compared. Having established the place of numerical analysis in this overall framework, it is then possible to identify its potential advantages. 1.3. Design objectives When designing any geotechnical structure, the engineer must ensure that it is stable. Stability can take several forms. Firstly, the structure and support system must be stable as a whole. There must be no danger of rotational, vertical or translational failure, see Figure 1.2. Secondly, overall stability must be established. For example, if a retaining structure supports sloping ground, the possibility of the construction promoting an overall slope failure should be investigated, see Figure 1.3. Figure 1.2: Local stabilityGeotechnical analysis / 3 The loads on any structural elements involved in the construction must also be calculated, so that these may be designed to carry them safely. Movements must be estimated, both of the structure and of the ground. This is particularly important if there are adjacent buildings and/or sensitive services. For example, if an excavation is to be made in an urban area close to existing services and buildings, see Figure 1.4, one of the key design constraints is the effect that the excavation has on the adjacent structures and services. It may be necessary to predict any structural forces induced in these existing structures and/or services. As part of the design process, it is necessary for an engineer to perform calculations to provide estimates of the above quantities. Analysis provides the mathematical framework for such calculations. A good analysis, which simulates real behaviour, allows the engineer to Figure 1.4: Interaction of structures understand problems better. While an important part of the design process, analysis only provides the engineer with a tool to quantify effects once material properties and loading conditions have been set. The design process involves considerably more than analysis. Slip surface Figure 1.3: Overall stability ° Services Tunnel 1.4 Design requirements Before the design process can begin, a considerable amount of information must be assembled. The basic geometry and loading conditions must be established. These are usually defined by the nature of the engineering project. A geotechnical site investigation is then required to establish the ground conditions. Both the soil stratigraphy and soil properties should be determined. In this respect it will be necessary to determine the strength of the soil and, if ground movements are important, to evaluate its stiffness too. The position of the ground water table and whether or not there is underdrainage or artesian conditions must also be established. The possibility of any changes to these water conditions should be investigated. For example, in many major cities around the world the ground water level is rising.4 / Finite element analysis in geotechnical engineering: Theory The site investigation should also establish the location of any services (gas, water, electricity, telecommunications, sewers and/or tunnels) that are in the vicinity of the proposed construction. The type (strip, raft and/or piled) and depth of the foundations of any adjacent buildings should also be determined. The allowable movements of these services and foundations should then be established. Any restrictions on the performance of the new geotechnical structure must be identified. Such restrictions can take many different forms. For example, due to the close proximity of adjacent services and structures there may be restrictions imposed on ground movements. Once the above information has been collected, the design constraints on the geotechnical structure can be established. These should cover the construction period and the design life of the structure. This process also implicitly identifies which types of structure are and are not appropriate. For example, when designing an excavation, if there is a restriction on the movement of the retained ground, propped or anchored embedded retaining walls are likely to be more appropriate than gravity or reinforced earth walls. The design constraints also determine the type of design analysis that needs to be undertaken. 1.5 Theoretical considerations 1.5.1 Requirements for a general solution In general, a theoretical solution must satisfy Equilibrium, Compatibility, the material Constitutive behaviour and Boundary conditions (both force and displacement). Each of these conditions is considered separately below. 1.5.2 Equilibrium To quantify how forces are transmitted through a continuum engineers use the concept of stress (force/unit area). The magnitude and direction of a stress and the manner in which it varies spatially indicates how the forces are transferred. However, these stresses cannot vary randomly but must obey certain rules. Before considering the concept of stresses, an analogous example of the problem of water flowing through a tank full of sand is presented in Figure 1.5. The tank full of sand has one inlet and two outlets. This figure indicates vectors of water velocity at discrete points within the tank. The size of the arrows represents the magnitude of the flow velocity, while their orientation shows the direction of flow. Due to the closer proximity of the left hand outlet to the inlet, more water flows in this direction than to the right hand outlet. As would be expected, the flows are very small in regions A, B and C. Such a result could be observed by using a transparent tank and injecting dye into the flow. Tank Jj Inet Outlet | |! outtet Figure 1.5: Flow trajectoriesGeotechnical analysis / 5 Similarly, a concrete beam, supported by two reactions on its lower surface and loaded by a load L on its upper surface, is presented in Figure 1.6. Clearly, for overall equilibrium the reactions must be 2L/3 and L/3. What is not so clear, however, is how the load is transferred through the beam. It is not possible to see how the load is transmitted in this case. As noted above, engineers use the concept of stress to investigate the load transfer. Figure 1.6: Stress trajectories Stresses are essentially fictitious quantities. For example, the manner in which the major principal stress varies through the beam is given in Figure 1.6. The length of the trajectories represents the magnitude of the stress and their orientation its direction. Whereas the velocity of flow is a vector with essentially three components, one in each of the Cartesian coordinate directions, stress is a tensor consisting of six components. In the same way as there are rules which govern the flow of water through the tank, there are also rules which control the manner in which the stress components vary throughout the concrete beam. Neglecting inertia effects and all body forces, except self weight, stresses iv a soil mass must satisfy the following three equations (Timoshenko and Goodier (1951)): Load: L Beam 3 Reaction: 20/3 Reaction: L/3 8G, 2% OTe yg We Fe En Oo 9 ay Ot, O% | 80; 4 . The following should be noted: - self weight, y, acts in the x direction; - compressive stresses are assumed positive; - the equilibrium Equations (1.1) are in terms of total stresses; - stresses must satisfy the boundary conditions (ic. at the boundaries the stresses must be in equilibrium with the applied surface traction forces). Figure 1.7: Stresses on a typical element 1.5.3. Compatibility Physical compatibility Compatible deformation involves no overlapping of material and no generation of holes. The physical meaning of compatibility can be explained by considering a6 / Finite element analysis in geotechnical engineering: Theory plate composed of smaller plate elements, as shown in Figure 1 .8a. After straining, the plate elements may be so distorted that they form the array shown in Figure 1.8b. This condition might represent failure by rupture. Alternatively, deformation might be such that the various plate elements fit together (i.e. no holes created or overlapping) as shown in Figure 1.8c. This condition represents a compatible deformation. a) Original ) Non-compatible ©) Compatible Figure 1.8: Modes of deformation Mathematical compatibility The above physical interpretation of compatibility can be expressed mathematically, by considering the definition of strains. If deformations are defined by continuous functions u, v and win the x, y and z directions respectively, the strains (assuming small strain theory and a compression positive sign convention) are defined as (Timoshenko and Goodier (1951)): cceue eovas ow ax” ‘i 2 oe oy 6: : (12) =o _ Gu, __ bw _ dv, 2 _ Sw ou Var Be By 1 eB ae Fle Be Be As the six strains are a function of only three displacements, they are not independent. It can be shown mathematically that for a compatible displacement field to exist, all the above components of strain and their derivatives must exist (are bounded) and be continuous to at least the second order. The displacement field must satisfy any specified displacements or restraints imposed on the boundary. 1.5.4 Equilibrium and compatibility conditions Combining the Equilibrium (Equations (1.1)) and Compatibility conditions (Equations (1.2)), gives: Unknowns: 6 stresses + 6 strains + 3 displacements =15 Equations: 3 equilibrium + 6 compatibility =9 To obtain a solution therefore requires 6 more equations. These come from the constitutive relationships.Geotechnical analysis / 7 1.5.5 Constitutive behaviour This is a description of material behaviour. In simple terms it is the stress - strain behaviour of the soil. It usually takes the form of a relationship between stresses and strains and therefore provides a link between equilibrium and compatibility. For calculation purposes the constitutive behaviour has to be expressed mathematically: Ag,| [Du Di Dis Dis Dis Die || AF Ao,| | Dr Dar Das Drs Drs Das |) 46, Ao, |_| Ds Diz Dss Dss Das Drs || Ae, At, {| Ds De Da Du Dis Dis|) Ary (13) At,,| | Ds1 Ds Dss Dss Dss Dss || Ay, Ar,| [De De Des Dr Des Des]| Ay, or Ao = [D] Ae For a linear elastic material the [D] matrix takes the following form: d-“) “ Hu 0 0 0 “ d-4) “u 0 0 0 E Hw H (-) 0 0 0 (4) G+) 0 0 0 (2-0 0 ; 0 0 0 0 @/2-p) 0 0 0 0 0 0 (/2-) where E and yw are the Young’s Modulus and Poisson’s ratio respectively. However, because soil usually behaves in a nonlinear manner, it is more realistic for the constitutive equations to relate increments of stress and strain, as indicated in Equation (1.3), and for the [D] matrix to depend on the current and past stress history. The constitutive behaviour can either be expressed in terms of total or effective stresses. If specified in terms of effective stresses, the principle of effective stress (o =a'+0,) may be invoked to obtain total stresses required for use with the equilibrium equations: Ao’ =[D'] Ae; Ao,=[D/ Ae; therefore Ao=([D'}+[D,]) Ae (1.5) where [Dy] is a constitutive relationship relating the change in pore fluid pressure to the change in strain. For undrained behaviour, the change in pore fluid pressure is related to the volumetric strain (which is small) via the bulk compressibility of the pore fluid (which is large), see Chapter 3.8 / Finite element analysis in geotechnical engineering: Theory 1.6 Geometric idealisation In order to apply the above concepts to a real geotechnical problem, certain assumptions and idealisations must be made. In particular, it is necessary to specify soil behaviour in the form of a mathematical constitutive relationship. It may also be necessary to simplify and/or idealise the geometry and/or boundary conditions of the problem. 1.6.1 Plane strain Due to the special geometric characteristics of many of the physical problems treated in. soil mechanics, additional simplifications of considerable magnitude can be applied. Problems, such as the analysis of retaining walls, continuous footings, and the stability of slopes, generally have one dimension very large in comparison with the other two, see Figure 1.9. Hence, if the force and/or applied displacement boundary conditions are perpendicular to, and independent of, this dimension, all cross sections will be the same. If the z dimension of the problem is large, and it can be assumed that the state existing in the x-y plane holds for all planes parallel to it, the displacement of any x-y cross section, relative to any parallel x-y cross section, is zero. This means that w=0, and the displacements u and v are independent of the z coordinate. The conditions consistent with these approximations are said to define the very important case of plane strain: Figure 1.9: Examples of plane strain ew _ av aw _ du oS oe 0 : ay & a (1.6) The constitutive relationship then reduces to: Ao] [Pn Ds Da Aoy! | Dy Dz Dy Ifa | Ag, |_| Dy, Dy, Ds, Ae. (1.7) Aty{ | Day Dig Dag |"? At,.| | Ds, Dsy Dey |[A%» [Ary] [Dar Diz Dos However, for elastic and the majority of material idealisations currently used to represent soil behaviour Ds,=Ds,=Ds,=D5,=Ds:=D,i=0, and consequently Ar,-At,=0. This results in four non-zero stress changes, Ao,, Ag,, Ag, and Ar, . It is common to consider only the stresses o,, o, and t,, when performing analysis for plane strain problems. This is acceptable if D,,, Dy2, Dy4, Dz), Din, DaasGeotechnical analysis / 9 Dy, Diz and D,, are not dependent on ¢,. This condition is satisfied if the soil is assumed to be elastic. It is also true if the Tresca or Mohr-Coulomb failure condition is adopted (see Chapter 7) and it is assumed that the intermediate stress oy=0,. Such an assumption is usually adopted for the simple analysis of geotechnical problems. It should be noted, however, that these are special cases. 1.6.2 Axi-symmetry Some problems possess rotational symmetry. For example, a uniform or centrally loaded circular footing, acting on a homogeneous or horizontally layered foundation, has rotational symmetry about a vertical axis through the centre of the foundation. Cylindrical triaxial samples, single piles and caissons are other examples where such symmetry may exist, see Figure 1.10. 9, t Circular footing Pile Triaxial sample Figure 1.10: Examples of axi-symmetry In this type of problem it is usual to carry out analyses using cylindrical coordinates r (radial direction), z (vertical direction) and @ (circumferential direction). Due to the symmetry, there is no displacement in the @ direction and the displacements in the r and z directions are independent of @ and therefore the strains reduce to (Timoshenko and Goodier (1951)): ou = (1.8) where w and v are the displacements in the r and z directions respectively. This is similar to the plane strain situation discussed above and, consequently, the same arguments concerning the [D] matrix apply here too. As for plane strain, there are four non-zero stress changes, Ag,, Ao,, Ao, and Ar, .10 / Finite element analysis in geotechnical engineering: Theory 1.7. Methods of analysis As noted above, fundamental considerations assert that for an exact theoretical solution the requirements of equilibrium, compatibility, material behaviour and boundary conditions, both force and displacement, must all be satisfied. It is therefore useful to review the broad categories of analysis currently in use against these theoretical requirements. Current methods of analysis can be conveniently grouped into the following categories: closed form, simple and numerical analysis. Each of these categories is considered separately. The ability of each method to satisfy the fundamental theoretical requirements and provide design information are summarised in Tables 1.1 and 1.2 respectively. Table 1.1: Basic solution requirements satisfied by the various methods of analysis METHOD OF SOLUTION REQUIREMENTS. ANALYSIS Boundary E conditions 2 = : 3 Constitutive Force | Disp gs behaviour Closed form s S| Linear elastic s | Closed fo Limit Rigid with a failure equilibrium s NS _| criterion S__|_NS | Rigid with a failure Stress field s criterion s NS g | Lower 2 | bound} _S Ideal plasticity with |_s_| ns | as Upper associated flow rule bound || _NS NS s Soil modelled by Beam-Spring springs or elastic approaches s interaction factors s s Full Numerical analysis s Any s S - Satisfied; NS - Not SatisfiedGeotechnical analysis / 11 Table 1.2: Design requirements satisfied by the various methods of analysis DESIGN REQUIREMENTS = Wall & Adjacent Stability supports structures METHOD OF ANALYSIS z : yl i BIL 3 ~| é oy = 5 @ 5 es 2 a 3 & 2 & =a| 9 BE lee] 2 [Be] 2 ee) 2 | 2 )22) 2 | FE) & Bal] a ae a ae a Closed form (Linear elastic) [|No [No __|No Yes __|Yes 2 2 Limit 52 |E ; equilibrium ys [#3 [2 No_|No g 2 ge (8 a2 a Stress field ys [23 (13 No__|No r 2 2 Lower ge |e & [bound_ves [23 13 No _|No £3 g 2 2 2 #2 |25 * {Upper 22 |22 |S |82 bound_flyes [23 [28 [5% [5% [No [No Beam-Spring | approaches yes _|No__[No___|Yes__}¥es__[No__|No Full Numerical analysis Yes Yes Yes Yes [ves Yes Yes 1.8 Closed form solutions For a particular geotechnical structure, if it is possible to establish a realistic constitutive model for material behaviour, identify the boundary conditions, and combine these with the equations of equilibrium and compatibility, an exact theoretical solution can be obtained. The solution is exact in the theoretical sense, but is still approximate for the real problem, as assumptions about geometry, the applied boundary conditions and the constitutive behaviour have been made in idealising the real physical problem into an equivalent mathematical form. In principle, it is possible to obtain a solution that predicts the behaviour of a problem12 / Finite element analysis in geotechnical engineering: Theory from first loading (construction/excavation) through to the long term and to provide information on movements and stability from a single analysis. A closed form solution is, therefore, the ultimate method of analysis. In this approach all solution requirements are satisfied and the theories of mathematics are used to obtain complete analytical expressions defining the full behaviour of the problem. However, as soil is a highly complex multi-phase material which behaves nonlinearly when loaded, complete analytical solutions to realistic geotechnical problems are not usually possible. Solutions can only be obtained for two very simple classes of problem. Firstly, there are solutions in which the soil is assumed to behave in an isotropic linear elastic manner. While these can be useful for providing a first estimate of movements and structural forces, they are of little use for investigating stability. Comparison with observed behaviour indicates that such solutions do not provide realistic predictions. Secondly, there are some solutions for problems which contain sufficient geometric symmetries that the problem reduces to being essentially one dimensional. Expansion of spherical and infinitely long cylindrical cavities in an infinite elasto-plastic continuum are examples. 1.9 Simple methods To enable more realistic solutions to be obtained, approximations must be introduced. This can be done in one of two ways. Firstly, the constraints on satisfying the basic solution requirements may be relaxed, but mathematics is still used to obtain an approximate analytical solution. This is the approach used by the pioneers of geotechnical engincering, Such approaches are considered as ‘simple methods’ in what follows. The second way, by which more realistic solutions can be obtained, is to introduce numerical approximations. All requirements of a theoretical solution are considered, but may only be satisfied in an approximate manner. This latter approach is considered in more detail in the next section. Limit equilibrium, Stress field and Limit analysis fall into the category of ‘simple methods’. All methods essentially assume the soil is at failure, but differ in the manner in which they arrive at a solution. 1.9.1 Limit equilibrium In this method of analysis an ‘arbitrary’ failure surface is adopted (assumed) and equilibrium conditions are considered for the failing soil mass, assuming that the failure criterion holds everywhere along the failure surface, The failure surface may be planar, curved or some combination of these. Only the global equilibrium of the ‘blocks’ of soil between the failure surfaces and the boundaries of the problem are considered. The internal stress distribution within the blocks of soil is not considered. Coulomb’s wedge analysis and the methed of slices are examples of limit equilibrium calculations.Geotechnical analysis / 13 Example: Critical height of a vertical cut Failure criterion: t= c'+o' tang! Figure 1.11: Failure mechanism for limit equilibrium solution The actual distributions of o and t along the failure surface ‘ac’, presented in Figure 1.11, are unknown. However, if / is the length of the failure surface ‘ac’, then: i ; ; i [rdf = fed! + fo'tang' dl = cil + tang! lo'dl (1.9) ° ° ° 0 where c’ and g’ are the soil’s cohesion and angle of shearing resistance respectively. Applying equilibrium to the wedge ‘abc’, i.e. resolving forces normal and tangential to failure surface ‘ac’, gives: ' Jo'dl = Wsing 0 1 Jrdl = WoosB ° Noting that W=Y4yFPtanf and [=Hicosf, Equations (1.9) and (1.10) can be combined to give: (1.10) a 2c' cosy’ 1° feos F= 9) snp au The value of the angle # which produces the most conservative (lowest) value of His obtained from dH/ap=0: OH _ -2 c' cosg’ cos(2h + op y(sinB cos( f+ o" (1.12) Equation (1.12) equals zero if cos(2/+9")=0. Therefore f = 1/4-9'/2 Substituting this angle into Equation (1.11) yields the Limit equilibrium value of Ay: 2c’ cose’ dct = —tan(r/4+e'/2) (1.13) Y Gppee etter eer eosantn titi Geos(a /4+ p'/2) sin(n/4—g'/2)14./ Finite element analysis in geotechnical engineering: Theory In terms of /otal stress, the equation reduces to: 45S, y Aip = (14) where S, is the undrained strength. Note: This solution is identical to the upper bound solution obtained assuming a planar sliding surface (see Section 1.9.3). The lower bound solution gives half the above value. 1.9.2 Stress field solution In this approach the soil is assumed to be at the point of failure everywhere and a solution is obtained by combining the failure criterion with the equilibrium equations. For plane strain conditions and the Mohr-Coulomb failure criterion this gives the following: i 7 Equilibrium equations: ao, er, ee Care) a Oy ~< (1.15) eats a | ay Mohr-Coulomb failure criterion [eS eeeeee re (from Figure 1.12): i of 0} = 2ceosg! H(o} +.) sing’ NY Le (1.16) ho fr Noting that: a Ss s=eclcotgy + (0) +03) Figure 1.12: Mohr’s circle of =clcotg’ + ¥(o,'+o,') stress t= Kolo )=[K(o, oy) +P" and substituting in Equation (1.16), gives the following alternative equations for the Mohr-Coulomb criterion: t= ssing (1.17) Lye, 0, +23,2% = [c’cotg' + 4(o,'+0,)] sing’ (1.18) The equilibrium Equations (1.15) and the failure criterion (1.18) provide three equations in terms of three unknowns. It is therefore theoretically possible to obtain a solution. Combining the above equations gives:Geotechnical analysis / 15 (1 +sing’cos26) 2 + sing’ sin2o 2 +. 25 sing'(cos20 22 — sin29 22, = 0 ar ay ay ax ; i 7 (1.19) sing’ sin20-25 + (1— sing’ cos20) 25 + 28 sing'(sin2022 + cos29 28) = y ax a ay ox ‘These two partial differential equations can be shown to be of the hyperbolic type. A solution is obtained by considering the characteristic directions and obtaining equations for the stress variation along these characteristics (Atkinson and Potts (1975)). The differential equations of the stress characteristics are: 2 = tan[O~(x/4—9!/2)] : (1.20) - = tan[O+(n/4~9'/2)] Along these characteristics the following equations hold: ds —2s tang’ d0 = (dy tang’ dv) a ds + 2stang'd@ = y(dy + tang’ dx) Equations (1.20) and (1.21) provide four differential equations with four unknowns x, y, s, and @ which, in principle, can be solved mathematically. However, to date, it has only been possible to obtain analytical solutions for very simple problems and/or if the soil is assumed to be weightless, y=0. Generally, they are solved numerically by adopting a finite difference approximation. Solutions based on the above equations usually only provide a partial stress field which does not cover the whole soil mass, but is restricted to the zone of interest. In general, they are therefore not Lower bound solutions (see Section 1.9.3). The above equations provide what appears to be, and some times is, static determinacy, in the sense that there are the same number of equations as unknown stress components. In most practical problems, however, the boundary conditions involve both forces and displacements and the static determinacy is misleading. Compatibility is not considered in this approach. Rankine active and passive stress fields and the earth pressure tables obtained by Sokolovski (1960, 1965) and used in some codes of practice are examples of stress field solutions. Stress fields also form the basis of analytical solutions to the bearing capacity problem. 1.9.3. Limit analysis The theorems of limit analysis (Chen (1975)) are based on the following assumptions: - Soil behaviour exhibits perfect or ideal plasticity, work hardening/softening does not occur. This implies that there is a single yield surface separating elastic and elasto-plastic behaviour.16 / Finite element analysis in geotechnical engineering: Theory - The yield surface is convex in shape and the plastic strains can be derived from the yield surface through the normality condition. - Changes in geometry of the soil mass that occur at failure are insignificant. This allows the equations of virtual work to be applied. With these assumptions it can be shown that a unique failure condition will exist. The bound theorems enable estimates of the collapse loads, which occur at failure, to be obtained. Solutions based on the ‘safe’ theorem are safe estimates of these loads, while those obtained using the ‘unsafe’ theorem are unsafe estimates. Use of both theorems enable bounds to the true collapse loads to be obtained. Unsafe theorem An unsafe solution to the true collapse loads (for the ideal plastic material) can be found by selecting any kinematically possible failure mechanism and performing an appropriate work rate calculation. The loads so determined are either on the unsafe side or equal to the true collapse loads. This theorem is often referred to as the ‘Upper bound” theorem. As equilibrium is not considered, there is an infinite number of solutions which can be found. The accuracy of the solution depends on how close the assumed failure mechanism is to the real one. Safe theorem Ifa statically admissible stress field covering the whole soil mass can be found, which nowhere violates the yield condition, then the loads in equilibrium with the stress field are on the safe side or equal to the true collapse loads. ‘This theorem is often referred to as the ‘Lower bound’ theorem. A statically admissible stress field consists of an equilibrium distribution of stress which balances the applied loads and body forces. As compatibility is not considered, there is an infinite number of solutions. The accuracy of the solution depends on how close the assumed stress field is to the real one. If safe and unsafe solutions can be found which give the same estimates of collapse loads, then this is the correct solution for the ideal plastic material. It should be noted that in such a case all the fundamental solution requirements are satisfied. This can rarely be achieved in practice. However, two such cases in which it has been achieved are (i) the solution of the undrained bearing capacity of a strip footing, on a soil with a constant undrained shear strength, S, (Chen (1975)), and (ii) the solution for the undrained lateral load capacity of an infinitely long rigid pile embedded in an infinite continuum of soil, with a constant undrained shear strength (Randolph and Houlsby (1984).