Beruflich Dokumente
Kultur Dokumente
by:
Prof. Zhaohui Yang; (University of Alaska, Anchorage) Dr. Zhao Cheng; (EarthMechanics Inc.) Dr. Guanzhou Jie; (Wells Fargo Securities) (University of Akron) Prof. Kallol Sett; Prof. Mahdi Taiebat; ( University of British Columbia) Dr. Matthias Preisig (Ecole Polytechnique Fdrale de Lausanne) e e (University of California, Davis) Mr. Nima Tafazzoli;
Version: April 29, 2010, 9:23 Copyright and Copyleft, Boris Jeremi and all contributors c
The use of the modeling and simulation system F I (these lecture notes and accompanying modeling, computational and visualE ization tools) for teaching, research and professional practice is strictly encouraged. Copyright and Copyleft are covered by GPL1 and Woodys (as in Guthrie) license (adapted by B.J.): This work is Copylefted and Copyrighted worldwide, by the Authors, for an indenite period of time, and anybody caught using it without our permission, will be mighty good friends of ourn, cause we dont give a darn. Read it. Learn it. Use it. Hack it. Debug it. Yodel it. Enjoy it. We wrote it, thats all we wanted to do.
Comments are much appreciated! Special thanks to (in chronological order): Miroslav ivkovi (Miroslav ZHivkovi), Dmitry J. Nicolsky, Andrzej Niemunis, Robbie Jaeger, c Perikleous)
(Yiorgos
1 http://www.gnu.org/copyleft/gpl.html
Jeremi et al. c
Contents
1 Introduction 1.1 1.2
(19962003)
13 13 14 14 14 14 14 15 16 17 19 20 23 23 26 34 39 39 42 42 44 45 59 61 65 66 67 69 71
Specialization to Computational Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tour of Computational Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1 1.2.2 1.2.3 Special Equilibrium Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Generalized Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sources of Nonlinearities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Role of Verication and Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . On Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
(1994)
1.3 1.4
Verication and Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.1 1.4.1 1.4.2 1.4.3 Verication and Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2 Finite Element Formulation for Single Phase Material (dry) 2.1 2.2 2.3
Formulation of the Continuum Mechanics Incremental Equations of Motion . . . . . . . . . . . . Finite Element Discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Isoparametric 8 20 Node Finite Element Denition . . . . . . . . . . . . . . . . . . . . . . . .
(19911994200220062010)
Elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Elastoplasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 3.2.2 3.2.3 3.2.4 3.2.5 Constitutive Relations for Innitesimal Plasticity . . . . . . . . . . . . . . . . . . . . . . . On Integration Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Midpoint Rule Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Crossing the Yield Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Singularities in the Yield Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3 3.4
A Forward Euler (Explicit) Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Backward Euler (Implicit) Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1 3.4.2 3.4.3 Single Vector Return Algorithm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Backward Euler Algorithms: Starting Points . . . . . . . . . . . . . . . . . . . . . . . . . Numerical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
4 80 85 89 89 90 91 92 92 93 96
3.4.4 3.4.5 3.5 3.5.1 3.5.2 3.5.3 3.5.4 3.5.5 3.5.6 3.5.7 3.5.8 3.5.9
Consistent Tangent Stiness Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gradients to the Potential Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yield Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plastic Flow Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HardeningSoftening Evolution Laws . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tresca Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . von Mises Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Drucker-Prager Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4 Probabilistic ElastoPlasticity
111
(2006)
5 Stochastic ElasticPlastic Finite Element Method 6 Large Deformation ElastoPlasticity 6.1 6.1.1 6.1.2 6.1.3 6.1.4 6.2 6.2.1 6.2.2 6.2.3 6.2.4 6.2.5 6.2.6 6.2.7 6.3 6.3.1 6.3.2 6.3.3 6.3.4 6.3.5 6.4 6.4.1 6.4.2
Jeremi et al. c
(19962004)
113 115
Continuum Mechanics Preliminaries: Kinematics . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Deformation Gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Strain Tensors, Deformation Tensors and Stretch . . . . . . . . . . . . . . . . . . . . . . 118 Rate of Deformation Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Constitutive Relations: Hyperelasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Isotropic Hyperelasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 VolumetricIsochoric Decomposition of Deformation . . . . . . . . . . . . . . . . . . . . 126 SimoSerrins Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Stress Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 Tangent Stiness Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Isotropic Hyperelastic Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Kinematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Constitutive Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 Implicit Integration Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 Algorithmic Tangent Stiness Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Equilibrium Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Version: April 29, 2010, 9:23
6.4.3 6.4.4
Formulation of NonLinear Finite Element Equations . . . . . . . . . . . . . . . . . . . . 150 Computational Domain in Incremental Analysis . . . . . . . . . . . . . . . . . . . . . . . 151
(1994)
157
The Residual Force Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 Constraining the Residual Force Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 Load Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 Displacement Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 Generalized, HyperSpherical Arc-Length Control . . . . . . . . . . . . . . . . . . . . . . . . . . 160 7.5.1 7.5.2 7.5.3 7.5.4 7.5.5 Traversing Equilibrium Path in Positive Sense . . . . . . . . . . . . . . . . . . . . . . . . 163 Predictor step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 Automatic Increments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 Convergence Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 The Algorithm Progress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
(19892006)
169
The Principle of Virtual Displacements in Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . 169 Direct Integration Methods for the Equations of Dynamic Equilibrium . . . . . . . . . . . . . . . 169 8.2.1 8.2.2 Newmark Integrator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 HHT Integrator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
(19992005)
9 Finite Element Formulation for Porous SolidPore Fluid Systems 9.1 General form of upU Governing Equations 9.1.1 9.1.2 9.1.3 9.2 9.2.1 9.2.2 9.2.3 9.2.4 9.2.5 9.3 9.3.1 9.3.2 9.3.3 9.3.4 9.3.5 9.4 9.4.1
Jeremi et al. c
173
. . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Governing Equations of Porous Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Modied Governing Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 Numerical Solution of solid part equilibrium equation . . . . . . . . . . . . . . . . . . . . 179 Numerical Solution of uid part equilibrium equation . . . . . . . . . . . . . . . . . . . . 180 Numerical Solution of ow conservation equation . . . . . . . . . . . . . . . . . . . . . . 181 Matrix form of the governing equations . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 Choice of shape functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 Verication Example: Drilling of a well . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Verication Example: The Case of a Spherical Cavity . . . . . . . . . . . . . . . . . . . . 190 Verication Example: Consolidation of a Soil Layer . . . . . . . . . . . . . . . . . . . . . 193 Verication Example: Line Injection of a uid in a Reservoir . . . . . . . . . . . . . . . . 200 Verication: Shock Wave Propagation in Saturated Porous Medium . . . . . . . . . . . . 204 Governing Equations of Porous Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
Version: April 29, 2010, 9:23
Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
9.4.2
10 EarthquakeSoilStructure Interaction
209
10.1 Dynamic Soil-Foundation-Structure Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 10.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 10.3 The Domain Reduction Method 10.3.1 Method Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
10.3.2 Method Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 10.4 Numerical Accuracy and Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 10.4.1 Grid Spacing h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 10.4.2 Time Step Length t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 10.4.3 Nonlinear Material Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 10.5 Domain Boundaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 10.6 Verication using one-dimensional Wave Propagation . . . . . . . . . . . . . . . . . . . . . . . . 219 10.6.1 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 10.6.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 10.7 Case History: Simple Structure on Nonlinear Soil . . . . . . . . . . . . . . . . . . . . . . . . . . 224 10.7.1 Simplied Models for Verication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 10.7.2 Full nonlinear 3d Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 11 Parallel Computing in Computational Geomechanics 12 Practical Applications
(1994) (19982000-2005)
247 249
12.1 Consolidation of Clays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 12.2 Staged Construction Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 12.3 Seismic Wave Propagation in Soils (Ground Motions) . . . . . . . . . . . . . . . . . . . . . . . . 252 12.4 Static and Dynamic Behavior of Pile Foundations in Dry and Saturated Soils . . . . . . . . . . . 253 12.5 Static and Dynamics Behavior of Shallow Foundations . . . . . . . . . . . . . . . . . . . . . . . 254 A nDarray 265
A.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 A.2 nDarray Programming Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266 A.2.1 Introduction to the nDarray Programming Tool . . . . . . . . . . . . . . . . . . . . . . . 266 A.2.2 Abstraction Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 A.3 Finite Element Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 A.3.1 Stress, Strain and Elastoplastic State Classes . . . . . . . . . . . . . . . . . . . . . . . . 271 A.3.2 Material Model Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 A.3.3 Stiness Matrix Class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272 A.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 A.4.1 Tensor Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 A.4.2 Fourth Order Isotropic Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
Jeremi et al. c Version: April 29, 2010, 9:23
A.4.3 Elastic Isotropic Stiness and Compliance Tensors . . . . . . . . . . . . . . . . . . . . . . 274 A.4.4 Second Derivative of Stress Invariant . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 A.4.5 Application to Computations in Elastoplasticity . . . . . . . . . . . . . . . . . . . . . . . 276 A.4.6 Stiness Matrix Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 A.5 Performance Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 A.6 Summary and Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 B Useful Formulae 279
B.1 Stress and Strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 B.1.1 Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 B.1.2 Strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282 B.2 Derivatives of Stress Invariants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 C Closed Form Gradients to the Potential Function D Hyperelasticity: Detailed Derivations 287 295
D.1 SimoSerrins Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 D.2 Derivation of 2volW /(CIJ CKL ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 D.3 Derivation of 2isoW /(CIJ CKL ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298 D.4 Derivation of wA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 D.5 Derivation of YAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
Jeremi et al. c
Jeremi et al. c
List of Figures
2.1 2.2 2.3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 Motion of body in stationary Cartesian coordinate system . . . . . . . . . . . . . . . . . . . . . . General three dimensional body . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Isoparametric 820 node brick element in global and local coordinate systems . . . . . . . . . . . The pictorial representation of integration algorithms in computational elastoplasticity: Generalized Midpoint schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pictorial representation of the intersection point problem in computational elastoplasticity . . . . Pictorial representation of the corner point problem in computational elastoplasticity . . . . . . . The pictorial representation of the apex point problem in computational elastoplasticity . . . . . Inuence regions in the meridian plane. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Normalized iso-error maps of Von-Mises model with linear isotropic hardening. . . . . . . . . . . 46 60 62 63 64 72 73 73 74 74 75 24 27 35
Relative iso-error maps of Von-Mises model with linear isotropic hardening. . . . . . . . . . . . . Normalized iso-error maps of Von-Mises model with Armstrong-Frederick kinematic hardening. Relative iso-error maps of Von-Mises model with Armstrong-Frederick kinematic hardening. .
. . .
. . . . . . . . . . . . . . .
3.12 Normalized iso-error maps of Drucker-Prager model with Armstrong-Frederick kinematic hardening. 75 3.13 Relative iso-error maps of Drucker-Prager model with Armstrong-Frederick kinematic hardening. 3.14 Normalized iso-error maps of Dafalias-Manzari model with average elastic moduli. . . . . . . . . 76 76 77 77 78
3.15 Relative iso-error maps of Dafalias-Manzari model with average elastic moduli. . . . . . . . . . . 3.16 Normalized iso-error maps of Dafalias-Manzari model with constant elastic moduli. . . . . . . . . 3.17 Relative iso-error maps of Dafalias-Manzari model with constant elastic moduli. . . . . . . . . .
7
3.18 Typical convergence for Von-Mises model with linear isotropic hardening (tolerance value 110
). 78 79 80 83 91 91
3.19 Typical convergence for Drucker-Prager model with Armstrong-Frederick kinematic hardening (tolerance value 1107 ). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.20 Typical convergence for Dafalias-Manzari model (tolerance value 110
7
). . . . . . . . . . . . .
3.21 Residual norms for typical iteration steps (tolerance value 1106 ). . . . . . . . . . . . . . . . . 3.22 Yield surface patterns in the meridian plane for isotropic granular materials (from Lade (1988b)) . 3.23 Deviatoric trace of typical yield surface for pressure sensitive materials. . . . . . . . . . . . . . . . 9
10
3.24 Various types of evolution laws that control hardening and/or softening of elasticplastic material models: (a) Isotropic (scalar) controlling equivalent friction angle and isotropic yield stress. (b) Rotational kinematic hardening (second order tensor) controlling pivoting around xed point (usually stress origin) of the yield surface. (c) Translational kinematic hardening (second order tensor) controlling translation of the yield surface. (d) Distortional (fourth order tensor) controlling the shape of the yield surface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 3.25 Schematic illustration of the yield, critical, dilatancy, and bounding surfaces in the -plane of deviatoric stress ratio space (after Dafalias and Manzari 2004). 6.1 6.2 6.3 6.4 6.5 6.6 7.1 7.2 7.3 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 . . . . . . . . . . . . . . . . . . 106
Displacement, stretch and rotation of material vector dXI to new position dxi . . . . . . . . . . . 116 Illustration of the equation Fij = Rik Ukj = vik Rkj . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Relative velocity dvi of particle Q at point q relative to particle P at point p. . . . . . . . . . . . 121 Volumetric isochoric decomposition of deformation. . . . . . . . . . . . . . . . . . . . . . . . . . 127 Multiplicative decomposition of deformation gradient: schematics. . . . . . . . . . . . . . . . . . 135 Motion of body in stationary Cartesian coordinate system . . . . . . . . . . . . . . . . . . . . . . 150 Spherical arc-length method and notation for one DOF system. . . . . . . . . . . . . . . . . . . . 159 Inuence of u and f on the constraint surface shape. . . . . . . . . . . . . . . . . . . . . . . . 160 Simple illustration of Bifurcation and Turning point. . . . . . . . . . . . . . . . . . . . . . . . . . 164 Fluid mechanics of Darcys ow (wi ) versus real ow (Ui = wi /n). . . . . . . . . . . . . . . . . . 176 Shape functions used for coupled analysis, displacement u and pore pressure p formulation . . . . 184 Boundary Conditions for Drilling of a Borehole . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186 The mesh generation for the study of borehole problem . . . . . . . . . . . . . . . . . . . . . . . 187 The comparison of radial solid displacement between analytical solution and experimental result for drained behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 The comparison of radial solid displacement between analytical solution and experimental result for undrained behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 The comparison of radial solid displacement between two analytical solutions and expanded boundary190 The mesh generation for the study of spherical cavity . . . . . . . . . . . . . . . . . . . . . . . . 192 The comparison of radial solid displacement between analytical solution and experimental result for drained behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
9.10 The comparison of radial solid displacement between analytical solution and experimental result for undrained behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 9.11 The comparison of radial solid displacement between analytical solution for undrained behavior and experimental result for drained behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 9.12 Consolidation of a Soil Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 9.13 The comparison between analytical solution and experimental results for the normalized p during the consolidation process against normalized depth z = z/h for various normalized t = cf t/h2 . . 199 9.14 The comparison between analytical solution and experimental results for the settlement . . . . . . 199
Jeremi et al. c Version: April 29, 2010, 9:23
11
9.15 The mesh generation for the study of line injection problem . . . . . . . . . . . . . . . . . . . . . 201 9.16 The comparison between analytical solution and experimental result for pore pressure . . . . . . . 202 9.17 The comparison between analytical solution and experimental result for radial displacement . . . . 203 9.18 The comparison between analytical solution and experimental result for radial displacement . . . . 203 9.19 Compressional wave in both solid and uid, comparison with closed form solution. . . . . . . . . . 205 10.1 Large physical domain with the source of load Pe (t) and the local feature (in this case a soil foundationbuilding system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 10.2 Simplied large physical domain with the source of load Pe (t) and without the local feature (in this case a soilfoundationbuilding system. Instead of the local feature, the model is simplied so that it is possible to analyze it and simulate the dynamic response as to consistently propagate the dynamic forces Pe (t). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 10.3 DRM: Single layer of elements between and e is used to create P ef f . . . . . . . . . . . . . . 215 10.4 Resulting acceleration using Linear and Newton-Raphson algorithms . . . . . . . . . . . . . . . . 217 10.5 Absorbing Boundary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 10.6 The analyzed models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 10.7 Using total motions to calculate P ef f . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 10.8 ue and ue obtained from free-eld model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 10.9 As in Figure 10.8 but with absorbing boundary at the base . . . . . . . . . . . . . . . . . . . . . 223 10.10Acceleration time histories and Fourier amplitude spectras . . . . . . . . . . . . . . . . . . . . . 225 10.11Transfer Function between Surface and Base of Soil Layer . . . . . . . . . . . . . . . . . . . . . 228 10.12Displacement Time-Histories at surface of 1d Soil Column . . . . . . . . . . . . . . . . . . . . . 229 10.13Fourier Amplitudes at surface of 1d Soil Column . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 10.14Acceleration time history at lowest free node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 10.15Transfer functions between acceleration at the soil surface and the base . . . . . . . . . . . . . . 232 10.16Dierence between results of analysis with dierent time steps . . . . . . . . . . . . . . . . . . . 233 10.17Averaged dierences between results of analysis with dierent time steps . . . . . . . . . . . . . 234 10.18Two-dimensional quasi-plane-strain model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 10.19The boundary conditions of the 2d model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 10.20Elastic homogeneous free-eld model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 10.21The 2d SFSI-model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 10.22Eigenmodes of SFSI-system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 10.23Fourier amplitude spectrum of input motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 10.24Parametric study of 15 linear structures with varying natural frequency. . . . . . . . . . . . . . . 237 10.25Parametric study of 15 nonlinear structures with varying natural frequency . . . . . . . . . . . . . 238 10.26Displacements in x-direction at the top of the nonlinear structures . . . . . . . . . . . . . . . . . 239 10.27Displacements in x-direction at the top of structures 1 and 2 . . . . . . . . . . . . . . . . . . . . 240 10.28Moments at the base of the linear column . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 10.29Moments at the base of the nonlinear column . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 10.30Fourier amplitude spectra of moments at the base of nonlinear column . . . . . . . . . . . . . . . 242
Jeremi et al. c Version: April 29, 2010, 9:23
12
10.31Average equivalent plastic strain at time t = 12 s . . . . . . . . . . . . . . . . . . . . . . . . . . 243 10.32Average equivalent plastic strain at time t = 14 s . . . . . . . . . . . . . . . . . . . . . . . . . . 243 10.33The full 3d-model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 10.34Resulting displacements and moments for 3d SFSI model . . . . . . . . . . . . . . . . . . . . . . 245 12.1 Finite element for 1D consolidation analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 12.2 Finite element for 1D consolidation analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
Jeremi et al. c
Chapter 1
Introduction
(19962003)
These Lecture Notes are in a perpetual state of change. They were inuenced by some of my reading, in particular by a number of books and lecture notes (Bathe and Wilson Bathe and Wilson (1976), Felippa Felippa (1993), Lubliner Lubliner (1990), Criseld Criseld (1991), Chen and Han Chen and Han (1988), Zienkiewicz and z z Taylor Zienkiewicz and Taylor (1991a,b), Malvern Malvern (1969), Dunica and Kolundija Dunica and Kolundija (1986), Koji Koji (1997), Hjelmstad Hjelmstad (1997), Oberkampf et al. Oberkampf et al. (2002)). c c The intent is to collect writings related to research done within my research group and make it available to students in a number of graduate courses I teach, including Computational Geomechanics (ECI285), Nonlinear Finite Elements for ElasticPlastic Problems (ECI280A) and Dynamic Finite Elements (ECI280B). Each course has its own set of lecture notes, however they are all part of a larger writeup, from which parts are used in compiling notes for each of the courses. A number of research students have also contributed material presented here, and their contributions are acknowledged and much appreciated.
1.1
In this section we start from general mechanics and specialize our interest toward the eld of computational mechanics... Mechanics Computational Mechanics Statics and Dynamics Linear and Nonlinear Analysis Elastic and Inelastic Analysis Discretization Methods 13
14
1.2
In this section we describe various examples of equilibrium path and set up basic terminology. Equilibrium Path
1.2.1
Critical Points Limit Points Bifurcation Points Turning Points Failure Points
1.2.2 1.2.3
Tonti Diagrams
1.3
How do we use experimental simulations to develop and improve models How much can (should) we trust model implementations (verication) How much can (should) we trust numerical simulations (validation) Trusting Simulation Tools Verication: The process of determining that a model implementation accurately represents the developers is solved correctly. Validation: The process of determining the degree to which a model is accurate representation of the real world from the perspective of the intended uses of the model. Physics issue. Validation provides evidence that the correct model is solved.
Jeremi et al. c Version: April 29, 2010, 9:23
conceptual description and specication. Mathematics issue. Verication provides evidence that the model
15
Importance of V & V V & V procedures are the primary means of assessing accuracy in modeling and computational simulations V & V procedures are the tools with which we build condence and credibility in modeling and computational simulations
Maturity of Computational Simulations NRC committee (1986) identied stages of maturity in CFD Stage 1: Developing enabling technologies (scientic papers published) Stage 2: Demonstration of and Condence in technologies and tools (capabilities and limitations of technology understood) Stage 3: Compilation of technologies and tools (capabilities and limitations of technology understood) Stage 4: Spreading of the eective use (changes the engineering process, value exceeds expectations) Stage 5: Mature capabilities (fully dependable, cost eective design applications)
1.3.1
Model Validation
Computer Simulation
Mathematical Model
Continuum Mathematics
Programming
Computer Implementation
Discrete Mathematics
Code Verification
Verication: The process of determining whether the products of a given phase of the software development cycle fulll the requirements established during the previous phase Validation: The process of evaluating software at the end of the software development process to ensure compliance with software requirements.
Jeremi et al. c
16
Other organization have similar denitions: Software quality assurance community American Nuclear Society (safety analysis of commercial nuclear reactors) International Organization for Standardization (ISO) Certication and Accreditation Certication: A written guarantee that a system or component complies with its specied requirements and is acceptable for operational use (IEEE (1990)). Written guarantee can be issued by anyone (code developer, code user, independent code evaluator) Code certication is more formal than validation documentation Accreditation: The ocial certication that a model or simulation is acceptable for use for a specic purpose (DOD/DMSO (1994)) Only ocially designated entities can accredit Normally appointed by the customers/users of the code or legal authority Appropriate for major liability or public safety applications Independence of Computational Condence Assessment 1. V&V conducted by the computational tool developer, No Independence 2. V&V conducted by a user from same organization 3. V&V conducted by a computational tool evaluator contracted by developers organization 4. V&V conducted by a computational tool evaluator contracted by the customer 5. V&V conducted by a computational tool evaluator contracted by the a legal authority High Independence
1.4
Conceptual Model Experimental Data Highly accurate solution Analytical solution Benchmark ODE solution Benchmark PDE solution Computational Model Unit Problems Benchmark Cases Subsystem Cases Computational Solution Complete System
Verification
Validation
Jeremi et al. c
17
Verication: The process of determining that a model implementation accurately represents the developers conceptual description and specication. Identify and remove errors in computer coding Numerical algorithm verication Software quality
Conceptual Model
Computational Solution
Verification
Validation: The process of determining the degree to which a model is accurate representation of the real world from the perspective of the intended uses of the model. Tactical goal:
Strategic goal:
Conceptual Model Experimental Data Computational Model Unit Problems Benchmark Cases Subsystem Cases Computational Solution Complete System
Validation
1.4.1
On Validation
Goals of Validation Quantication of uncertainties and errors in the computational model and the experimental measurements Goals on validation Tactical goal: Identication and minimization of uncertainties and errors in the computational model Strategic goal: Increase condence in the quantitative predictive capability of the computational model Strategy is to reduce as much as possible the following: Computational model uncertainties and errors Random (precision) errors and bias (systematic) errors in the experiments Incomplete physical characterization of the experiment
Jeremi et al. c Version: April 29, 2010, 9:23
18
Conceptual Model
Computational Model
Computational Solution
Comparison of agreement
Validation Procedure Uncertainty Aleatory uncertainty inherent variation associated with the physical system of the environment (variation stochastic uncertainty. Epistemic uncertainty potential deciency in any phase of the modeling process that is due to lack of knowledge (poor understanding of mechanics...). Also known as reducible uncertainty, model form uncertainty and subjective uncertainty Types of Physical Experiments Traditional Experiments Improve the fundamental understanding of physics involved Improve the mathematical models for physical phenomena Assess component performance Validation Experiments Model validation experiments Designed and executed to quantitatively estimate mathematical models ability to simulate well dened physical behavior The simulation tool (SimTool) (conceptual model, computational model, computational solution) is the customer Validation Experiments A validation experiment should be jointly designed and executed by experimentalist and computationalist Need for close working relationship from inception to documentation
Jeremi et al. c Version: April 29, 2010, 9:23
in external excitation, material properties...). Also know known as irreducible uncertainty, variability and
19
Elimination of typical competition between each Complete honesty concerning strengths and weaknesses of both experimental and computational simulations A validation Experiment should be designed to capture the relevant physics Measure all important modeling data in the experiment Characteristics and imperfections of the experimental facility should be included in the model A validation experiment should use any possible synergism between experiment and computational approaches Oset strength and weaknesses of computations and experiments Use high condence simulations for simple physics to calibrate of improve the characterization of the experimental facility Conduct experiments with a hierarchy of physics complexity to determine where the computational simulation breaks (remember, SimTool is the customer!) Maintain independence between computational and experimental results Blind comparison, the computational simulations should be predictions Neither side is allowed to use fudge factors, parameters Validate experiments on unit level problems, hierarchy of experimental measurements should be made which present an increasing range of computational diculty Use of qualitative data (e.g. visualization) and quantitative data Computational data should be processed to match the experimental measurement techniques Experimental uncertainty analysis should be developed and employed Distinguish and quantify random and correlated bias errors Use symmetry arguments and statistical methods to identify correlated bias errors Make uncertainty estimates on input quantities needed by the SimTool
1.4.2
Prediction
Prediction: use of computational model to foretell the state of a physical system under consideration under conditions for which the computational model has not been validated Validation does not directly make a claim about the accuracy of a prediction Computational models are easily misused (unintentionally or intentionally) How closely related are the conditions of the prediction and specic cases in validation database How well is physics of the problem understood
Jeremi et al. c Version: April 29, 2010, 9:23
20
Relation Between Validation and Prediction Quantication of condence in a prediction: How do I quantify validation and its inference value in a predictions? How do I quantify verication and its inference value in a prediction? How far are individual experiments in my validation database from my physical system of interest?
1.4.3
Application Domain
Application Domain
Inference
System complexity
System complexity
System complexity
Validation Domain
System Parameter
Rarely applicable to engineering systems (certainly not for infrastructure objects like bridges, buildings, port facilities, dams...) Even if the engineering system is small, environmental inuences (generalized loads, conditions, wear and tare) are hard to predict Human factors (take Mars rover example with a memory overow, operator forgot to ush the memory...) Inference Based on physics or statistics Validation domain is actually an aggregation of tests and thus might not be convex (bifurcation of behavior) Experimental facilities (e.g. NEES sites) provide for validation domain that are exclusively nonoverlapping with the application domain.
Importance of Models and Numerical Simulations Veried and Validated models can be used for assessing behavior of components or complete systems, with the understanding that the environmental inuences cannot all be taken into the account prior to operation but with a good model, their inuence on system behavior can be assessed as need be (before or after the event)
Jeremi et al. c
21
Prediction under Uncertainty Ever present uncertainty needs to be estimated for predictions Identify all relevant sources of uncertainty Create mathematical representation of individual sources Propagate representation of sources through modeling and simulation process (Probabilistic Elastic Plastic Theory)
Short Course on Verication and Validation in Computational Mechanics, by Dr. Willliam Oberkampf, Sandia National Laboratories July 27th, 2003, Albuquerque, New Mexico. Material from Verication and Validation in Computational Mechanics web site http://www.usacm.org/vnvcsm/ at the USACM. William L. Oberkampf, Timothy G. Trucano, and Charles Hirsch. Verication, validation and predictive capability in computational engineering and physics. In Proceedings of the Foundations for 2002. Johns Hopkins University / Applied Physics Laboratory. Patrick J. Roache. Verication and Validation in Computational Science and Engineering. Hermosa publishers, 1998. ISBN 0-913478-08-3. Boris Jeremi, Gerik Scheuermann, Jan Frey, Zhaohui Yang, Bernd Hamman, Kenneth I. Joy and Hans c Haggen. Tensor Visualizations in Computational Geomechanics. International Journal for Numerical and Analytical Methods in Geomechanics incorporating Mechanics of CohesiveFrictional Materials, Vol 26. Issue 10, pp 925-944, August 2002. Verication and Validation on the 21st Century Workshop, pages 174, Laurel, Maryland, October 22-23
Jeremi et al. c
22
Jeremi et al. c
Chapter 2
Consider1 the motion of a general body in a stationary Cartesian coordinate system, as shown in Figure (2.1), and assume that the body can experience large displacements, large strains, and nonlinear constitutive response. The aim is to evaluate the equilibrium positions of the complete body at discrete time points 0, t, 2t, . . . , where t is an increment in time. To develop the solution strategy, assume that the solutions for the static and kinematic variables for all time steps from 0 to time t inclusive, have been obtained. Then the solution process for the next required equilibrium position corresponding to time t + t is typical and would be applied repetitively until a complete solution path has been found. Hence, in the analysis one follows all particles of the body in their motion, from the original to the nal conguration of the body. In so doing, we have adopted a Lagrangian ( or material ) formulation of the problem. In the Lagrangian incremental analysis approach we express the equilibrium of the body at time t + t using the principle of virtual displacements. Using tensorial notation2 this principle requires that:
t+t
t+t V
ij
t+t eij
t+t
dV = t+t R
(2.1)
where the t+t ij are Cartesian components of the Cauchy stress tensor, the t+t eij are the Cartesian components of an innitesimal strain tensor, and the means variation in i.e.: 1 2 ui t+t x j uj t+t x i 1 2 ui t+t x j uj
t+t x i
t+t eij
(2.2)
1 detailed
2 Einsteins
derivations and explanations are given in Bathe (1982) summation rule is implied unless stated dierently, all lower case indices (i, j, p, q, m, n, o, r, s, t, . . . ) can have values
of 1, 2, 3, and values for capital letter indices will be specied where need be.
23
24
x2 , tx2 , x2
t+ t 0 0 0
P( t+tx1t+t x2 , x3 ) ,
t+ t t+t t+t t t
A V
A 0 V
Configuration at time 0
Configuration at time t
n+1
0
t+t t
x3 ,tx3 , x3 t+ t
xi = x i + u i x i = 0xi + t ui t+t xi = t xi + ui
t+t
x1 ,t x1,t+ tx1
i=1,2,3
Figure 2.1: Motion of body in stationary Cartesian coordinate system It should be noted that Cauchy stresses are body forces per unit area in the conguration at time t + t, and the innitesimal strain components are also referred to this as yet unknown conguration. The right hand side of equation (2.1), i.e. time t + t:
t+t
R is the virtual work performed when the body is subjected to a virtual displacement at
t+t
R=
t+t V
t+t B fi
t+t ut+t ui i
t+t
dV +
t+t S
t+t S fi
ut+t i
t+t
dS
(2.3)
where
t+t B fi
and
spectively, and
component of the virtual displacement vector. A fundamental diculty in the general application of equation (2.1) is that the conguration of the body at time t + t is unknown. The continuous change in the conguration of the body entails some important consequences for the development of an incremental analysis procedure. For example, an important consideration must be that the Cauchy stress at time t + t cannot be obtained by simply adding to the Cauchy stresses at time t a stress increment which is due only to the straining of the material. Namely, the calculation of the Cauchy stress at time t + t must also take into account the rigid body rotation of the material, because the components of the Cauchy stress tensor are not invariant with respect to the rigid body rotations4 .
3 This 4 However,
t+t S fi t+t ui is
are the components of the externally applied body and surface force vectors, rethe inertial body force that is present if accelerations are present3 , ui is the ith
is based on DAlemberts principle. that problem will not be addressed in this work since this work deals with MaterialNonlinearOnly analysis of solids,
Jeremi et al. c
25
The fact that the conguration of the body changes continuously in a large deformation analysis is dealt with in an elegant manner by using appropriate stress and strain measures and constitutive relations. When solving the general problem5 one possible approach6 is given in Simo (1988). Previous discussion was oriented toward small deformation, small displacement analysis leading to the use of Cauchy stress tensor ij and small strain tensor eij . In the following, we will briey cover some other stress and strain measures particularly useful in large strain and large displacement analysis. The basic equation that we want to solve is relation (2.1), which expresses the equilibrium and compatibility requirements of the general body considered in the conguration corresponding to time t + t. The constitutive equations also enter (2.1) through the calculation of stresses. Since in general the body can undergo large displacements and large strains, and the constitutive relations are nonlinear, the relation in (2.1) cannot be solved directly. However, an approximate solution can be obtained by referring all variables to a previously calculated known equilibrium conguration, and linearizing the resulting equations. This solution can then be improved by iterations. To develop the governing equations for the approximate solution obtained by linearization we recall that the solutions for time 0, t, 2t, . . . , t have already been calculated and that we can employ the fact that the 2nd PiolaKirchho stress tensor is energy conjugate to the GreenLagrange strain tensor:
0V
t 0Sij
tij 0 dV = 0
0V
0 0 t x mn 0 xj,n t t t i,m
t 0 xk,i
tekl t
t 0
xl,j
dV =
0V
0 t mn t
temn 0 dV (2.4) 0
because:
t 0 0 xk,l t xl,m
= km
and since:
dV = t t dV
we have:
0V
t 0Sij
tij 0 dV = 0
t
0V
mn temn t dV t
(2.5)
t 0Sij
5 That 6 This
0 0 t x mn 0 xj,n t t t i,m
(2.6)
is, large displacements, large deformations and nonlinear constitutive relations. is still a hot research topic!
Jeremi et al. c
26
where 0 xj,n = t
0 xi t xm
and
represents the ratio of the mass density at time 0 and time t, and the GreenLagrange
t 0ij
1 2
t 0 ui,j
(2.7)
Then, by employing (2.5) we refer the stresses and strains to one of these known equilibrium congurations. The choice lies between two formulations which have been termed total Lagrangian and updated Lagrangian formulations. In the total Lagrangian formulations, also termed Lagrangian formulation, all static and kinematic variables are referred to the initial conguration at time 0, while in the updated Lagrangian formulation all static and kinematic variables are referred to the conguration at time t. Both the total Lagrangian and updated Lagrangian formulations include all kinematic nonlinear eects due to large displacement, large rotations and large strains, but whether the large strain behavior is modeled appropriately depends on the constitutive relations specied. The only advantage of using one formulation rather than the other lies in its greater numerical eciency. Using (2.5) in the total Lagrangian formulations we consider this basic equation:
0V
t+t Sij 0
t+tij 0 dV = t+t R 0
(2.8)
tV
t+t Sij t
t+tij t dV = t+t R t
(2.9)
obtained by linearizing these relations. By comparing the total Lagrangian and updated Lagrangian formulations we can observe that they quite analogous and that, in fact, the only theoretical dierence between the two formulations lies in the choice of dierent reference congurations for the kinematic and static variables. If in the numerical solution the appropriate constitutive tensors are employed, identical results are obtained. Coupling of large deformation, large displacement and material nonlinear analysis is still the topic of research in the research community. Possible direction may be the use of both Lagrangian and Eulerian formulation intermixed in one scheme.
in which t+t R is the external virtual work as dened in (2.3). Approximate solution to the (2.8) and (2.9) can be
2.2
Consider the equilibrium of a general threedimensional body such as in Figure (2.2) (Bathe, 1996). The external forces acting on a body are surface tractions fiS and body forces fiB . Displacements are ui and strain tensor7 is eij and the stress tensor corresponding to strain tensor is ij . Assume that the externally applied forces are given and that we want to solve for the resulting displacements, strains and stresses. One possible approach to express the equilibrium of the body is to use the principle of virtual
7 1 2
(ui,j + uj,i ).
Jeremi et al. c
27
x3
f3B r3 f1B f2B f1
f3
S
f2
r1
r2
x2
x1
Figure 2.2: General three dimensional body displacements. This principle states that the equilibrium of the body requires that for any compatible, small virtual displacements8 imposed onto the body, the total internal virtual work is equal to the total external virtual work. This statement can be mathematically expressed using equation (2.10) for the body at time t + t, and since we are using the incremental approach let us drop the time dimension, so that all the equations are imposed for the given increment9 , at time t + t. The equation is now, using tensorial notation10 :
ij eij dV =
V V
fiB i ui dV + u
fiS ui dS
(2.10)
The internal work given on the left side of (2.10) is equal to the actual stresses ij going through the virtual strains eij that corresponds to the imposed virtual displacements. The external work is on the right side of (2.10) ui . and is equal to the actual (surface) forces fiS and (body) forces fiB i going through the virtual displacements u It should be emphasized that the virtual strains used in (2.10) are those corresponding to the imposed body and surface virtual displacements, and that these displacements can be any compatible set of displacements that satisfy the geometric boundary conditions. The equation in (2.10) is an expression of equilibrium, and for dierent virtual displacements, correspondingly dierent equations of equilibrium are obtained. However, equation (2.10) also contains the compatibility and constitutive requirements if the principle is used in the appropriate manner.
satisfy the essential boundary conditions. + t will be dropped from now one in this chapter. 10 Einsteins summation rule is implied unless stated dierently, all lower case indices (i, j, p, q, m, n, o, r, s, t, . . . ) can have values
9t 8 which
of 1, 2, 3, and values for capital letter indices will be specied where need be.
Jeremi et al. c
28
Namely, the displacements considered should be continuous and compatible and should satisfy the displacement boundary conditions, and the stresses should be evaluated from the strains using appropriate constitutive relations. Thus, the principle of virtual displacements embodies all requirements that need be fullled in the analysis of a problem in solid and structural mechanics. The principle of virtual displacements can be directly related to the principle that total potential of the system must be stationary. In the nite element analysis we approximate the body in Figure (2.2) as an assemblage of discrete nite elements with the elements being interconnected at nodal points on the element boundaries. The displacements measured in a local coordinate system r1 , r2 and r3 within each element are assumed to be a function of the displacements at the N nite element nodal points:
u ua = HI uIa
(2.11)
where I = 1, 2, 3, . . . , n and n is number of nodes in a specic element, a = 1, 2, 3 represents a number of dimensions (can be 1 or 2 or 3), HI represents displacement interpolation vector, uIa is the tensor of global generalized displacement components at all element nodes. The use of the term generalized displacements means that both translations, rotations, or any other nodal unknown are modeled independently. Here specically only translational degrees of freedom are considered. The strain tensor is dened as: 1 (ua,b + ub,a ) 2 1 (a,b + ub,a ) = u 2 1 (HI uIa ),b + (HI uIb ),a = 2 1 ((HI,b uIa ) + (HI,a uIb )) 2
eab =
(2.12)
and the by using (2.11) we can dene the approximate strain tensor: eab eab = = =
(2.13)
0 ab = Eabcd ecd e0 + ab cd
(2.14)
where ab is the approximate Cauchy stress tensor, Eabcd is the constitutive tensor12 , ecd is the innitesimal
0 approximate strain tensor, e0 is the innitesimal initial strain tensor and ab is the initial Cauchy stress tensor. cd
Using the assumption of the displacements within each nite element, as expressed in (2.11), we can now derive equilibrium equations that corresponds to the nodal point displacements of the assemblage of nite elements. We can rewrite (2.10) as a sum13 of integrations over the volume and areas of all nite elements:
ab ab dV m = e
m
11 in
B fa a ua dV m + u
Sm
S fa uS dS m a
(2.15)
12 This
terms of exact stress and strain elds but it holds for approximate elds as well. tensor can be elastic or elastoplastic constitutive tensor. S 13 Or, more correctly as a union m since we are integrating over the union of elements.
Jeremi et al. c
29
where m = 1, 2, 3, . . . , k and k is the number of elements. It is important to note that the integrations in (2.15) are performed over the element volumes and surfaces, and that for convenience we may use dierent element coordinate systems in the calculations. If we substitute equations (2.11), (2.12), (2.13) and (2.14) in (2.15) it follows:
0 Eabcd ecd e0 + ab cd
dV m = (2.16)
Vm
B fa (HI uIa ) dV m
S fa (HI uIa ) dS m
or:
Eabcd
m V
m
0 + ab
dV m =
=
m Vm
B fa (HI uIa ) dV m
S fa (HI uIa ) dS m
(2.17)
We can observe that in the previous equations represents a virtual quantity but the rules for are quite similar to regular dierentiation so that can enter the brackets and virtualize the nodal displacement14 . It thus follows:
Eabcd
m V
m
0 + ab
dV m =
=
m Vm
B fa (HI uIa ) dV m
S fa (HI uIa ) dS m
(2.18)
Let us now work out some algebra in the left hand side of equation (2.18): Eabcd
m V
m
0 Eabcd e0 + ab cd
dV m =
=
m Vm
HJ uJa HI uIa dV m +
m Vm m
S fa (HI uIa ) dS m
(2.19)
and further:
14 since
they are driving variables that dene overall displacement eld through interpolation functions
Jeremi et al. c
30
Eabcd
1 (HI,b uIa + HI,a uIb ) 2 1 (HI,b uIa + HI,a uIb ) 2 1 (HI,b uIa + HI,a uIb ) 2
Vm
dV m + dV m + dV m =
Eabcd e0 cd
Vm 0 ab
+
m
B fa (HI uIa ) dV m
HJ uJa HI uIa dV m
m Vm
+
m Sm
S fa (HI uIa ) dS m
(2.20)
Several things should be observed in the equation (2.20). Namely, the rst three lines in the equation can be simplied if one takes into account symmetries of Eijkl and ij . In the case of the elastic stiness tensor Eijkl major and both minor symmetries exist. In the case of the elastoplastic stiness tensor, such symmetries exists if a ow a rule is associated. If ow rule is nonassociated, only minor symmetries exist while major symmetry is destroyed15 . As a matter of fact, both minor symmetries in Eijkl are the only symmetries we need, and the rst line of (2.20) can be rewritten as:
Eabcd
dV m =
=
m Vm
(HJ,d uJc ) Eabcd (HI,b uIa ) dV m = (HI,b uIa ) Eabcd (HJ,d uJc ) dV m
m Vm
(2.21)
Similar simplications are possible in second and third line of equation (2.20). Namely, in the second line we can use both minor symmetries of Eijkl so that:
Eabcd e0 cd
dV m = (2.22)
=
m Vm
and the third line can be simplied due to the symmetry in Cauchy stress tensor ij as:
15 for
more on stiness tensor symmetries see sections (3.5.1, 3.3 and 3.4)
Jeremi et al. c
31
0 ab
dV m = (2.23)
=
m Vm
0 ab (HI,b uIa ) dV m
+
m Vm
Vm
=
m V
m
(HI uIa ) dV
HJ uJa HI uIa dV m +
m V
m
Sm
or if we leave the unknown nodal accelerations16 uJc and displacements uJc on the left hand side and move all the known quantities on to the right hand side:
HJ ac uJc HI uIa dV m +
m V
m
=
m Vm
B fa (HI uIa ) dV m + m
Sm
+
m V
m
Eabcd e0 cd
(HI,b uIa ) dV m
(2.25)
To obtain the equation for the unknown nodal generalized displacements from (2.25), we invoke the virtual displacement theorem which states that virtual displacements are any, non zero, kinematically admissible displacements. In that case we can factor out nodal virtual displacements uIa so that equation (2.25) becomes:
HJ ac uJc HI dV m +
m Vm B fa HI dV m uIa + m m Vm
uIa =
=
m Vm
Sm
+
m V
m
Eabcd e0 cd
HI,b dV
uIa
(2.26)
is noted that uJc = ac uJa relationship was used here, where ac is the Kronecker delta.
Jeremi et al. c
32
HJ ac HI uJc dV m +
m V
m
=
m V
m
Sm
+
m Vm
Eabcd e0 HI,b dV m cd
(2.27)
Vm
One should also observe that in the rst line of equation (2.27) generalized nodal accelerations uJc and generalized nodal displacements uJc are unknowns that are not subjected to integration so they can be factored out of the integral:
HJ ac HI dV m uJc
m Vm
+
m V
m
=
m Vm
Sm
+
m V
m
Eabcd e0 cd
HI,b dV m
(2.28)
(m)
MIacJ =
Vm
HJ ac HI dV m
(2.29)
(m)
KIacJ =
V
m
(2.30)
(m) B FIa
=
Vm
B fa HI dV m
(2.31)
(m) S FIa
=
Sm
S fa HI dS m
(2.32)
Jeremi et al. c
33
=
V
m
Eabcd e0 HI,b dV m cd
(2.33)
0 (m) mn FIa
=
Vm
0 ab HI,b dV m
(2.34)
where
(m)
(m) B FIa
(m) S FIa
is the tensor
FIa
e0 mn
FIa
0 mn
(m) (m)
MIacJ uJc +
(m)
(m)
KIacJ uJc =
m
(m) B FIa
+
m
(m) S FIa
+
m
0 (m) mn FIa
(2.35)
(2.36)
a, c =
where:
MAacB =
m
(m)
MIacJ
KAacB =
m
(m)
KIacJ
(2.37)
are the system mass and stiness tensors, respectively, uBc is the tensor of unknown nodal accelerations, and uBc is the tensor of unknown generalized nodal displacements, while the load tensor is given as:
FAa =
m
(m) B FIa
+
m
(m) S FIa
+
m
0 (m) mn FIa
(2.38)
After assembling the system of equations in (2.37) it is relatively easy to solve for the unknown displacements uLc either for static or fully dynamic case. It is also very important to note that in all previous equations, omissions of inertial force term (all terms with ) will yield static equilibrium equations. Description of solutions procedures
17 Summation
of the element volume integrals expresses the direct addition of the element tensors to obtain global, system tensors.
This method of direct addition is usually referred to as the direct stiness method.
Jeremi et al. c
34
for static linear and nonlinear problems are described in some detail in chapter 7. In addition to that, solution procedures for dynamic, linear and nonlinear problems are described in some detail in chapter 8. A note on the nal form of the tensors used is in order. In order to use readily available system of equation solvers equation (2.37) will be rewritten in the following from:
MP Q u P + KP Q u P = FQ
(2.39)
where MP Q is system mass matrix, KP Q is system stiness matrix and FQ is the loading vector. Matrix form of equation 2.37, presented as equation 2.39 is obtained attening the system mass tensor MAacB , system stiness tensor KAacB , unknown acceleration tensor uBc , unknown displacement tensor uBc and the system loading tensor FAa . Flattening from the fourth order mass/stiness tensors to two dimensional mass/stiness matrix is done by simply performing appropriate (re) numbering of nodal DOFs in each dimension. Similar approach is used for unknown accelerations/displacements and for loadings. Static Analysis: Internal and External Loads. Internal and external loading tensors can be dened as:
(fIa )int =
(m)
(m)
KIacJ uJc =
m V
m
ab HI,b dV m
(2.40)
(fIa )ext =
m
(m) B FIa
+
m
(m) S FIa
+
m
0 (m) mn FIa
(2.41)
where (fIa )int is the internal force tensor and (fIa )ext is the external force tensor. Equilibrium is obtained when residual: rIa (Jc , ) = (fIa (Jc ))int (fIa )ext u u is equal to zero, r(u, ) = 0. The same equation in attened form yields: (2.42)
(2.43)
2.3
The basic procedure in the isoparametric18 nite element formulation is to express the element coordinates and element displacements in the form of interpolations using the local three dimensional19 coordinate system of the
18 name
isoparametric comes from the fact that both displacements and coordinates are dened in terms of nodal values. Super-
parametric and subparametric nite elements exists also. 19 in the case of element presented here, that is isoparametric 8 20 node nite element.
Jeremi et al. c
35
element. Considering the general 3D element, the coordinate interpolations, using indicial notation20 are:
xi = HA (rk ) xAi
(2.44)
where A = 1, 2, . . . , n and n is the total number of nodes associated with that specic element, xAi is the i-th coordinate of node A, i = 1, 2, 3, k = 1, 2, 3 and HA are the interpolation functions dened in local coordinate system of the element, with variables r1 , r2 and r3 varying from 1 to +1. The interpolation functions HA for the isoparametric 820 node are the so called serendipity interpolation functions mainly because they were derived by inspection. For the nite element with nodes numbered as in Figure (2.3) they are given21 in the following set of formulae:
r3
3 4 11 12 15 16 1 17 13 5 19 7 9 14 18 6 10 2
r 2
x2 r1 20
8
x1 x3
Figure 2.3: Isoparametric 820 node brick element in global and local coordinate systems
2 isp (20 ) (1 + r1 ) (1 r2 ) 1 r3 4 2 isp (18 ) (1 r1 ) (1 + r2 ) 1 r3 H18 = 4 2 isp (16 ) (1 + r1 ) 1 r2 (1 r3 ) H16 = 4 2 isp (14 ) (1 r1 ) 1 r2 (1 r3 ) H14 = 4 2 isp (12 ) (1 + r1 ) 1 r2 (1 + r3 ) H12 = 4 2 isp (10 ) (1 r1 ) 1 r2 (1 + r3 ) H10 = 4
H20 =
isp (19 ) (1 r1 ) (1 r2 ) 4 isp (17 ) (1 + r1 ) (1 + r2 ) H17 = 4 2 isp (15 ) 1 r1 (1 r2 ) H15 = 4 2 isp (13 ) 1 r1 (1 + r2 ) H13 = 4 2 isp (11 ) 1 r1 (1 r2 ) H11 = 4 2 isp (9 ) 1 r1 (1 + r2 ) H9 = 4 H19 =
2 1 r3 2 1 r3
(1 r3 ) (1 r3 ) (1 + r3 ) (1 + r3 )
20 Einsteins
summation rule is implied unless stated dierently, all lower case indices (i, j, p, q, m, n, o, r, s, t, . . . ) can have values
of 1, 2, 3, and values for capital letter indices will be specied where need be. 21 for more details see Bathe (1982).
Jeremi et al. c
36
(1 + r1 ) (1 r2 ) (1 r3 ) h15 h16 h20 + 8 2 (1 r1 ) (1 r2 ) (1 r3 ) h14 h15 h19 H7 = + 8 2 (1 r1 ) (1 + r2 ) (1 r3 ) h13 h14 h18 + H6 = 8 2 (1 + r1 ) (1 + r2 ) (1 r3 ) h13 h16 h17 H5 = + 8 2 (1 + r1 ) (1 r2 ) (1 + r3 ) h11 h12 h20 H4 = + 8 2 (1 r1 ) (1 r2 ) (1 + r3 ) h10 h11 h19 H3 = + 8 2 (1 r1 ) (1 + r2 ) (1 + r3 ) h10 h18 h9 + H2 = 8 2 (1 + r1 ) (1 + r2 ) (1 + r3 ) h12 h17 h9 H1 = + 8 2 H8 = where r1 , r2 and r3 are the axes of natural, local, curvilinear coordinate system and isp (nod num) is boolean function that returns +1 if node number (nod num) is present and 0 if node number (nod num) is not present. To be able to evaluate various important element tensors22 , we need to calculate the straindisplacement transformation tensor23 . The element strains are obtained in terms of derivatives of element displacements with respect to the local coordinate system. Because the element displacements are dened in the local coordinate system, we need to relate global x1 , x2 and x3 derivatives to the r1 , r2 and r3 derivatives. In order to obtain derivatives with respect to global coordinate system, i.e. following form:
xa
(2.45)
xa = = Jak rk rk xa xa
(2.46)
where Jak is the Jacobian operator relating local coordinate derivatives to the global coordinate derivatives:
Jak
x 1 r1 xa = = x1 r2 rk x1 r3
x2 r1 x2 r2 x2 r3
x3 r1 x3 r2 x3 r3
are dened in chapter (2.2).
(2.47)
0 0 22 i.e. (m)K (m)F B , (m)F S , (m)F mn , (m)F mn , that IacJ , Ia Ia Ia Ia `` ` 1 23 from the equation ab = 2 HI,b uIa + HI,a uIb
Jeremi et al. c
37
The existence of equation (2.45) requires that the inverse of Jak exists and that inverse exists provided that there is a onetoone24 correspondence between the local and the global coordinates of element. It should be pointed out that except for the very simple cases, volume and surface element tensor25 integrals are evaluated by means of numerical integration26 Numerical integration rules is quite a broad subject and will not be covered here27 .
24 unique. 25 as 26 GaussLegendre,
dened in chapter (2.2) by equations (2.30), (2.31), (2.32), (2.33) and (2.34). NewtonCoates, Lobatto are among most used integration rules. 27 nice explanation with examples is given in Bathe (1982).
Jeremi et al. c
38
Jeremi et al. c
Chapter 3
(In collaboration with Prof. Zhaohui Yang, Dr. Zhao Cheng, Prof. Mahdi Taiebat and Doctoral Student Nima Tafazzoli)
3.1
Elasticity
In linear elasticity the relationship between the stress tensor ij and the strain tensor kl can be represented in the following form: ij = (ij ) (3.1)
If we assume the existence of a strain energy function1 W (ij ) then the stress strain relation is: ij = W (ij ) ij (3.2)
The introduction of the strain energy density function into elasticity is due to Green, and elastic solids for which such a function is assumed to exist are called Green elastic or hyperelastic solids. Linearization of an elastic continuum is carried out with respect to a reference conguration which is stress free at temperature T0 , so that 0ij = 0. If we denote as Eijkl an isothermal modulus tensor, then under isothermal conditions, we obtain the generalized Hookes law: ij = Eijkl kl (3.3)
39
40
The Eijkl is called the elastic constants tensor of fourth order2 . It has 81 independent components in total. It has symmetries with respect to pairs of indices ij and kl and these symmetries reduce the number of independent components to 36. There is an additional symmetry conditions: Eijkl = 2W ij kl =
=0
2W kl ij
(3.4)
=0
thus we have Eijkl = Eklij and the number of independent components is reduced to 21 (Spencer, 1980). We will restrain our considerations to the isotropic case. The most general form of the isotropic tensor of rank 4 has the following representation: I4 = ij kl + ik jl + il jk (3.5)
If Eijkl has this form then in order to satisfy the symmetry condition3 Eijkl = Ejikl we must have = . The symmetry condition4 Eijkl = Eklji is then automatically satised. The elastic constant tensor has the following form: Eijkl = ij kl + (ik jl + il jk ) (3.6)
and E and are Youngs Modulus and Poissons ratio respectively. The symmetric part of the fourth order unit tensor is :
sym Iijkl =
1 (ik jl + il jl ) 2
(3.8)
and can be found as multiplier of in equation (3.186). Equation (3.186) can be written in terms of E and as: Eijkl = E 2 (1 + ) 2 ij kl + ik jl + il jk 1 2 (3.9)
The same relation in terms of bulk modulus K and shear modulus G is: 2 Eijkl = Kij kl + G ij kl + ik jl + il jk 3 where K and G are given as: 2 K =+ 3
2 also 3 4
(3.10)
G=
(3.11)
Jeremi et al. c
41
The relation between the strain tensor, kl and the stress tensor, ij is: kl = Dklpq pq (3.12)
where Dklpq is the elastic compliance fourth order tensor, dened as: Dklpq = 1 kl pq + (kp lq + kq lp ) 2 (3 + 2) 4 (3.13)
It is worthwhile noting that the part adjacent to the inverse of the bulk modulus K: (kl pq ) controls the volumetric response and that the part adjacent to the inverse of the shear modulus G: 1 1 kl pq + (kp lq + kq lp ) 3 2 controls the shear response! This note will prove useful later on. Linear transformation of the stress tensor pq into itself, i.e. ij is dened as: ij = Eijkl kl = Eijkl Dklpq pq (3.16)
Linear Elastic Model. Linear elastic law is the simplest one and assumes constant Youngs modulus E and constant Poissons Ration . Nonlinear Elastic Model #1. This nonlinear model (Janbu, 1963), (Duncan and Chang, 1970) assumes dependence of the Youngs modulus on the minor principal stress 3 = min in the form E = Kpa 3 pa
n
(3.18)
Here, pa is the atmospheric pressure in the same units as E and stress. The two material constants K and n are constant for a given void ratio.
Jeremi et al. c Version: April 29, 2010, 9:23
42
Nonlinear Elastic Model #2. If Youngs modulus and Poissons ratio are replaced by the shear modulus G and bulk modulus K the nonlinear elastic relationship can be expressed in terms of the normal eective mean stress p as G and/or K = AF (e, OCR)pn (3.19)
where e is the void ratio, OCR is the overconsolidation ratio and p = ii /3 is the mean eective stress (Hardin, 1978). Lades Nonlinear Elastic Model. Lade and Nelson (1987) and Lade (1988a) proposed a nonlinear elastic model based on Hookes law in which Poisson ratio is kept constant. According to this model, Youngs modulus can be expressed in terms of a power law as: E = M pa I1 pa
2
1+ + 6 1 2
J2D p2 a
(3.20)
where I1 = ii is the rst invariant of the stress tensor and J2D = (sij sij )/2 is the second invariant of the deviatoric stress tensor sij = ij kk ij /3. The parameter pa is atmospheric pressure expressed in the same unit as E, I1 and J2D and the modulus number M and the exponent are constant, dimensionless numbers.
3.2
3.2.1
Elastoplasticity
Constitutive Relations for Innitesimal Plasticity
A wide range of elastoplastic materials can be characterized by means of a set of constitutive relations of the general form: ij = e + p ij ij (3.21)
(3.22)
dp = d ij
(3.23)
(3.24) e and p denotes the total, elastic and plastic strain tensor, ij is ij ij
indices in q replaces n indices6 . Equation (3.21) expresses the commonly assumed additive decomposition of
5 In
the Cauchy stress tensor, and q signies some suitable set of internal variables5 . The asterisk in the place of
the innitesimal strain tensor into elastic and plastic parts. Equation (3.22) represents the generalized Hookes
the simplest models of plasticity the internal variables are taken as either plastic strain components p or the hardening ij variables dened, for example as a function of inelastic (plastic) work, i.e. = f (W p ). See Lubliner (1990) page 115. p 6 for example ij if the variable is ij , or nothing if the variable is a scalar value, i.e. .
Jeremi et al. c
43
law7 which linearly relates stresses and elastic strains through a stiness modulus tensor Eijkl . Equation (3.23) expresses a generally associated or non-associated ow rule for the plastic strain and (3.24) describes a suitable set of hardening laws, which govern the evolution of the plastic variables. In these equations, mij is the plastic ow direction, h the plastic moduli and d is a plastic parameter to be determined with the aid of the loadingunloading criterion, which can be expressed in terms of the KarushKuhnTucker condition (Karush, 1939; Kuhn and Tucker, 1951) as: F (ij , q ) 0 d 0 F d = 0 (3.25)
(3.26)
(3.27)
In the previous equations F (ij , q ) denotes the yield function of the material and (3.25) characterizes the corresponding elastic domain, which is presumably convex. Along any process of loading, conditions (3.25), (3.26) and (3.27) must hold simultaneously. For F < 0, equation (3.27) yields d = 0, i.e. elastic behavior, while plastic ow is characterized by d > 0, which with (3.27) is possible only if the yield criterion is satised, i.e. F = 0. From the latter constraint, in the process of plastic loading the plastic consistency conditions8 is obtained in the form: dF = where : nij = F ij (3.29) F F dij + dq = nij dij + dq = 0 ij q (3.28)
Equation (3.28) has the eect of conning the stress trajectory to the yield surface9 . It is worthwhile noting that nij and are normals to the yield surface in stress space and the plastic variable space respectively. An interesting alternative way of representing nonassociated ow rules can be found in Runesson (1987). A ctitious plastic strain derived from associated ow rule, ep is introduced. This ctitious plastic strain is assumed ij to be related to the real plastic strain p , which is derived from a nonassociated ow rule10 through the linear ij transformation: ep = Aijkl p ij kl (3.31)
F q
(3.30)
Linear transformation tensor Aijkl may be state dependent in general case, and it reduces to the symmetric part of the fourth order identity tensor11 for the case of associated plasticity.
Eq. 3.185 order accuracy condition. 9 Since it is only linear expansion stress trajectory is conned to the tangential plane only. 10 as in equation 3.23. ` sym 1 11 A ijkl Iijkl 2 ik jl + il jk .
8 rst 7 also
Jeremi et al. c
44
It is often of interest to model deviatoric strains by an associated ow rule while the volumetric part is nonassociated. For this case, Aijkl can be formulated as: 1 1 (ij kl ) + (ik jl + il jk ) 3 2
Aijkl =
(3.32)
1 1 (ij kl ) + (ik jl + il jk ) (3.33) 1+ 3 2 and it is obvious that the nonassociated ow rule is obtained with = 0 and the associated ow rule with = 0. A1 = ijkl It is useful to choose 0 and retain nice, positive denite properties of adjusted constitutive tensors later. Let the ij norm, be the complementary energy norm12 :
2
= ij Dijkl kl
(3.34)
1 where Dijkl is the elastic compliance tensor ( Dijkl = Eijkl ), and let us introduce the adjusted complementary
ij
= ij (Aijkl Dklmn ) mn = ij
Dijmn mn
(3.35)
where ADijmn is the elastic compliance tensor transformed with respect to the nonassociativity involved. It is
sym clear that when Aijkl Iijkl = A ij 2
ij
3.2.2
On Integration Algorithms
In the section Constitutive Relations for Innitesimal Plasticity we have summarized constitutive equations that are capable of representing a wide variety of elastoplastic materials. The problem in Computational Elasto plasticity is to devise accurate and ecient algorithms for the integration of such constitutive relations. In the context of nite element analysis using isoparametric elements, the integration of constitutive equations is carried out at Gauss points. In each step the deformation increments are given or known, and the unknowns to be found are updated stresses and plastic variables. According to Ortiz and Popov (1985) an acceptable algorithm should satisfy: consistency with the constitutive relations to be integrated or rst order accuracy, Numerical stability, incremental plastic consistency A nonrequired but desirable feature to be added to the above list is: higher13 order accuracy First two conditions are needed for attaining convergence for the numerical solution as the step or increment becomes vanishingly small. The third condition is the algorithmic counterpart of the plastic consistency condition and requires that the state of stress computed from the algorithm be contained within the elastic domain.
12 This 13 at
Jeremi et al. c
45
3.2.3
A class of algorithms for integrating constitutive equations with potential to satisfy the above mentioned conditions are the Generalized Midpoint rule algorithms. They are given in the following form:
n+1
ij = Eijkl
n+1
kl n+1p kl
(3.36)
n+1 p ij
= np + n+mij ij
(3.37)
n+1
q = nq + n+h
(3.38)
Fn+1 = 0 where:
n+
(3.39)
n+1
ij , (1 ) nq +
n+1
(3.40)
n+
h = h (1 ) nij +
n+1
ij , (1 ) nq +
n+1
(3.41)
It is quite clear that the case = 0 corresponds to the Forward Euler approach14 , the case = 1 corresponds to the Backward Euler approach15 , and the case = 1/2 to the Crank Nicholson scheme. Equations (3.36), (3.37), (3.38), (3.39), (3.40) and (3.41) are the nonlinear algebraic equations to be solved for the unknowns
n+1
ij ,
n+1 p n+1 ij , q
and . From the Figure (3.1)16 it can be seen that the Generalized Midpoint rule may be
pred
regarded as a returning mapping algorithm in which the elastic predictor surface along the ow direction evaluated at the midpoint (
n+
ij ,
n+
q ).
Accuracy Analysis Bearing in mind the context of the displacement based nite element analysis the integration of constitutive equations is performed for the given strain increment. The updated strains n+1 ij = ij (tn + t) may be viewed as the known function of the step size t. The remaining updated variables
n+1
ij ,
as well as
the incremental plastic parameter become functions of t implicitly dened through equations (3.36), (3.37), (3.38) and (3.39). It should be clear from (3.36), (3.37), (3.38) and (3.39) that as t 0 than and thus the limiting values of
14 explicit
n+1
ij ,
n+1 p n+1 ij , q
ij n ij ,
scheme. scheme. 16 it should be pointed out that the vectors, as drawn on this gure, are pointing in the right direction only if we assume that
15 implicit
Eijkl Iijkl . For any general elasticity tensor Eijkl all vectors are dened in the Eijkl metric, so the term normal, as we are used to it, does not apply here.
Jeremi et al. c
46
m ij
m ij
predictor ij
n+1
m ij
n+1
Q=0
n+1 n+
cross ij
n+1
ij
ij
ij
ij
ij
2 1 3
F=0
n+
F=0 F=0
n+1
t0 t0 t0
lim
n+1
ij
= = = =
ij
lim
n+1 p ij n+1
n p ij n
lim
q (3.42)
t0
lim
It can also be argued that, by virtue of the implicit function theorem ((Abraham et al., 1988) Chapter 2.5), n+1ij ,
n+1 p n+1 ij , q
and are dierentiable functions of t, if the functions n+mij , n+h and F are suciently smooth.
Sucient smoothness will be assumed as needed. First Order Accuracy. First order accuracy17 of the algorithm, dened by the equations (3.36), (3.37), (3.38)
n+1
and (3.39) with the constitutive equations given by (3.21), (3.22), (3.23) and (3.24) necessitates that the numerically integrated variables
n p ij
ij ,
n+1 p ij
and
n+1
and q (t + t) to within second order terms in the Taylors expansion around the initial state n ij = ij (t), = p (t) and n q = q (t) in t. First order accuracy can be written in the following form: ij d ij d (nij ) = = Eijkl d (t) d (t)
n+1 p d (nij ) d nij d (t) d (t)
t0
lim
(3.43)
17 rst
order consistency.
Jeremi et al. c
47
t0
lim
n+1 p ij
d (t)
(3.44)
t0
lim
n+1
(3.45)
(3.46)
and the plastic parameter d (n) /d (t) is determined with the aid of the plastic consistency condition at t: (nF ) dij (nF ) dq dij dq d (nF ) = + = nnij + n =0 d (t) ij d (t) q d (t) d (t) d (t)
(3.47)
It is now rather straightforward to check whether the Generalized Midpoint rule satises the consistency conditions as given by (3.43), (3.44), (3.45) and (3.46). We can proceed further on by dierentiating (3.36), (3.37), (3.38) and (3.39) with respect to t18 : kl d n+1p kl d (t) d (t)
n+1
ij = Eijkl d (t)
n+1
(3.48)
d d d (t)
n+1 p ij
d (t)
n+
d d (t) d
n+1
n+
mij +
mij +
mij ij
n+1
ij mij + d (t) q
(3.49)
n+1
d d d (t)
d (t)
n+
n+1 p q
= d
d d (t)
n+1
n+
h +
h +
h ij
n+1
ij h + d (t) q
(3.50)
n+1
d
18
(3.51)
bearing in mind that values at t are constants and that only variables at t + t are changing with respect to t.
Jeremi et al. c
48
where
n+
mij and
n+
By taking t to the limit value, t 0, in the (3.48), (3.49), (3.50) and (3.51) and using the relations from
t0
lim
ij = Eijkl d (t)
n+1
d (nkl ) d d (t)
n+1 p kl t=0
d (t)
(3.52)
t0
lim
n+1 p ij
d (t)
t0
lim
n+1
t0
lim
F (nF ) = d (t) ij
n+1
t0
lim
ij d (t)
n+1
(nF ) q
t0
lim
q d (t)
n+1
=0
(3.55)
In the previous equations it is quite clear that since t = 0, then equations (3.42) hold and since the variables
n
ij , n p and n q are constant with respect to the change in t, the result follows readily, i.e. the Midpoint rule ij
To investigate second order accuracy of the algorithm given by (3.36), (3.37), (3.38)
n+1
and (3.39) together with the constitutive equations given by (3.21), (3.22), (3.23) and (3.24) we shall proceed in the following manner. Second order accuracy actually means that the numerically integrated variables
n+1 p ij
ij ,
and
n+1
terms in the Taylors expansion around the initial state ij = ij (t), n p = p (t) and n q = q (t) in t. This ij ij
n
p ij
t0
lim
d2
n+1
ij
2
d (t)
2
= (3.56)
Eijkl
t0
lim
d2
n+1
kl
2
d (t)
lim
d2 (np ) kl d (t)
2
t0
= Eijkl
d2 (nkl ) d (t)
d2 (np ) kl d (t)
2
Jeremi et al. c
49
t0
lim
d2
d (t)
n+1 p ij 2
d2 d (t)
2 t0
lim
n+1
mij + lim d2
2
d (t) d n mij d (nij ) d ( ) mij d (nq ) ( mij ) + + 2 d (t) ij |n d (t) q |n d (t) d (t)
2 n
(nmij ) +
t0
lim
d2
d (t)
n+1 p q 2
d2 d (t)
2 t0
lim
n+1
h + lim d
2 2
t0
n+1
d (t) d2 d (t)
2
(nh ) +
(nh ) +
dt
n+1
t0
lim
d2 () d (t)
2
d2 (n ) d (t)
2
(3.59)
2
and the plastic parameter d2 (n ) /d (t) is determined with the aid of the second order oscillatory satisfaction of the plastic consistency condition:
d2 (nF ) d (t)
2
dnij dt
+
n
d ( ) dt
(3.60)
Now we can proceed by taking the second derivative of the equations (3.36), (3.37), (3.38) and (3.39) or use the already derived rst derivatives from equations (3.48), (3.49), (3.50) and (3.51), and then dierentiate them again so that we get:
d2
n+1
ij
2
d (t)
= Eijkl
d2
n+1
kl
2
d (t)
d2
d (t)
n+1 p kl 2
(3.61)
Jeremi et al. c
50
d (t)
n+
n+1 p ij 2
mij + + (3.62)
ij mij + d (t) q d
n+1
n+1
d
n+1
q d (t) d q d (t)
n+1
n+1
mij ij
n+1
ij mij + d (t) q
n+1
d (t)
n+
n+1 p q 2
h + + (3.63)
ij h + d (t) q d
n+1
n+1
d
n+1
q d (t) d q d (t)
n+1
n+1
h ij
n+1
ij h + d (t) q
n+1
d2 d
n+1
n+1
F
2
nij d
dij + d
n+1
n+1
d (t) 2 n+1 d ij
2
= + =0 (3.64)
dij
d (t) 2 n+1 d q
If we drive t to the limit, namely by taking limt0 and keeping in mind equations (3.42) and the assumed consistency of the algorithm19 as given by the equations (3.43), (3.44), (3.45) and (3.46) one nds: d2
n+1
t0
lim
ij
2
d (t)
= Eijkl
d2 (nkl ) d (t)
2
lim
d2
t0
d (t)
n+1 p kl 2
(3.65)
t0
lim
d2
d (t)
n
n+1 p ij 2
t0
lim
d2
n+1
n+
d (t)
mij + 2
d (n ) d (t)
mij ij
19 actually
Jeremi et al. c
51
t0
lim
d2
d (t)
n
n+1 p q 2
t0
lim
d2
n+1
n+
d (t)
h + 2
d (n ) d (t)
h ij
h d (nij ) + d (t) q
t0
lim
d2
n+1
F
2
d (t) 2 n+1 d ij
2
= + =0 (3.68)
d (t) 2 n+1 d q
By comparing equations (3.65), (3.66), (3.67) and (3.68) with the second order accuracy condition stated in equations (3.56), (3.57), (3.58) and (3.59) it is quite clear that the second order accuracy is obtained i20 = 1/2 ! accurate for = 1/2. However, one should not forget that these results are obtained for the limiting case t 0, The conclusion is that the Midpointrule algorithm is consistent21 for all [0, 1] and it is second order
i.e. the strain increments are small and tend to zero. Numerical Stability Analysis
Numerical stability of an algorithm plays a central role in approximation theory for initial value problems. In fact, it can be stated that consistency and stability are necessary and sucient conditions for convergence of an algorithm as the time step tends to zero. In the approach presented by Ortiz and Popov (1985) a new methodology is proposed by which the stability properties of an integration algorithm for elastoplastic constitutive relations can be established. Our attention is conned to perfect plasticity and a smooth yield surface. The purpose of the following stability analysis is to determine under what conditions a nite perturbation in the initial stresses is diluted by the algorithm. In other words:
n (2) n (1) ij , ij
(3.69)
(2) n+1 (1) ij and n+1ij are two sets of updated n (2) ij , respectively, and all of the previous stress
values are assumed to lie on the yield surface. Stability in the sense of equation (3.69) is referred to as large scale
20 21 rst
Jeremi et al. c
52
stability. It is shown in Helgason (1978)22 that for nonlinear initial value problems dened on Banach manifolds, consistency and large scale stability with respect to a complete metric are sucient for convergence. The task of directly establishing estimates of the type expressed in (3.69) is rather dicult, and so despite the conceptual appeal of large scale stability, simplied solutions are sought. It should be recognized that attention simplication is founded on the fact that the dilution or attenuation, by the algorithm of innitesimal perturbations: can be restricted to innitesimal perturbation in the initial conditions of the type nij nij + d (nij ). This
n+1
ij d nij
(3.70)
with respect to some suitable norm , of small scale stability, implies large scale stability in the sense of equation (3.69). Let the
2
ij
= ij Dijkl kl
(3.71)
dened as
1 where Dijkl is the elastic compliance tensor (Dijkl = Eijkl ), and let the distance d (, ) on the yield surface be
d ij , ij
(1)
(2)
= inf
ij (s) ds
(3.72)
where the innum is taken over all possible stress paths on the yield surface that are joining two stress states, namely ij and ij . It can be found in Helgason (1978) that for a smooth yield surface, equation (3.72) denes the geodesic distance which endows the yield surface with a complete metric structure. Suppose that we have any two initial states of stress nij and nij and let
(1) (2) n+1 (1) ij (1) (2)
and
n+1 (2) ij
be the
(2)
corresponding updated values, respectively, and all the previous stress states are assumed to lie on the yield surface. Then, according to Helgason (1978), there exists a unique geodesic curve that joins nij and nij for which the innum in equation (3.72) is attained. If n is such a curve, then by denition:
(1)
n (1) n (2) ij , ij
=
n
ij (s) ds
(3.73)
Let the new curve n+1 be the transform of curve n by the algorithm. By denition n+1 lies on the yield surface and joins two stress states
n+1 (1) ij
and
n+1 (2) ij .
=
n+1
ij (s) ds
(3.74)
Under the assumption of small scale stability of the algorithm one can write:
(3.75)
Jeremi et al. c
53
for every pair of corresponding points sn and sn+1 on n and n+1 respectively, so it follows:
n+1
ij (sn+1 ) ds
ij (sn ) ds
(3.76)
n (1) n (2) ij , ij
(3.77)
which proves large scale stability. The main conclusion of the above argument may be stated as follows: small scale stability in the energy norm is equivalent to large scale stability in the associated geodesic distance. The previous result is of practical importance, since it shows that the stability analysis for the integration algorithm in elastoplasticity can be carried out by the assessment of small scale stability. The small scale stability analysis of the Generalized Midpoint rule is necessary to determine how the algorithm propagates innitesimal perturbations in the initial conditions. By dierentiating equations (3.36), (3.37), (3.38) and (3.39)and considering that we are dealing with perfectly plastic case here so that
n+1
n+1
ij = Eijkl d
n+1 p kl
(3.78)
d (n ij ) = Eijkl d (np ) kl
(3.79)
n+1 p ij
n p ij
=d
n+
mij + d
n+
mij
(3.80)
n+1
F =
F ij
d
n+1
n+1
ij = n+1 nij d
n+1
ij = 0
(3.81)
Let us now examine the shape of d (n+mij ) having in mind the original denition23 given in equation (3.40):
n+
n+1
ij , (1 ) nq +
n+1
n+
mij = (1 )
d (nkl ) +
n+
d
n+
n+1
kl
23
the remark about restraining analysis to perfectly plastic case still holds, so that
Jeremi et al. c
54
Mijkl =
mij kl
d d
n+
n+1 p ij
n p ij n+1
mij + (1 )
n+
Mijkl d (nkl ) +
n+
Mijkl d
kl (3.82)
1 By using equations (3.78) and (3.79) and knowing that Eijkl = Dijkl one can write:
n+1 p ij
= Dijkl d
n+1
kl
n p ij
= Dijkl d(nkl )
Dijkl d d
n+
n+1
kl + Dijkl d (nkl ) =
n+
mij + (1 )
n+
Mijkl d (nkl ) +
Mijkl d
n+1
kl
n+1
kl :
Dijkl + Dijkl (1 )
n+
n+
Mijkl
n+1
kl = mij
Mijkl
d (nkl ) d
n+
and by denoting :
ijkl = Dijkl (1 )
n+
Mijkl
Jeremi et al. c
55
ijkl = Dijkl +
n+
Mijkl
it follows:
n+1
n+
mij
(3.83)
n+1
n+1
F = n+1 nkl d
n+1
kl = 0
one gets:
d
n+1
n+1
F = =0 (3.84)
n+
mij
(3.85)
or24 :
n+1
d =
(3.86)
kl from (3.83) and the solution for d from (3.86) one can nd:
n+1
n+1
n+m
1 pqrs
d (nrs ) (3.87)
24
Jeremi et al. c
56
n+1
n+1
kl = 1 ijkl ijkl
ks rl
n+m
1 pqrs
d (nrs )
(3.88)
n+1
klrs = ks rl
n+m
1 pqrs
(3.89)
n+1
(3.90)
In order to derive the estimate of the type (3.70) from (3.90) we shall proceed in the following way. The norm of a tensor is dened as:
Aijkl = sup
Aijkl kl kl
(3.91)
Aijkl kl Aijkl
kl
Bijkl
(3.92)
it follows:
n+1
kl
(3.93)
n+1
kl
1 ijkl ijkl
klrs
d (nrs )
(3.94)
klrs
Jeremi et al. c
57
(klrs ) 1 ijkl
n+
mij =
(3.95)
(klrs ) (rs ) = rs
(3.96)
n+1
nrs . From these properties and the denition in equation (3.91) it follows
klrs 1
(3.97)
is symmetric and positive denite everywhere on the yield surface. The assumption is valid, if the ow direction mij is derived from the convex potential function, which is a rather common feature among yield criteria. It is now clear that :
1 ijkl = ijkl
max
ij max ij
(3.98)
where
max
(ijkl ijkl ) kl = 0
(3.99)
max
ij
max
ij = maxij Dijkl
max
kl = 1
(3.100)
If we denote:
Jeremi et al. c Version: April 29, 2010, 9:23
58
n+
= maxij
n+
Mijkl
max
kl
(3.101)
n+
from equations (3.98), (3.100), (3.101) and from the denition26 of ijkl and ijkl , it follows:
1 ijkl = ijkl
1 (1 ) (n+) 1 + (n+)
(3.102)
n+1
kl
1 (1 ) (n+) 1 + (n+)
d (nrs )
(3.103)
n+
1 1 (1 ) (n+) n+) 1+ (
n+
n+
(3.104)
n+1
kl
d (nrs )
(3.105)
1 1
(3.106)
min =
1 2
(3.107)
25 because n+M 26 27
ijkl is derived`from a convex potential function. ` ijkl = Dijkl (1 ) n+Mijkl and ijkl = Dijkl + n+Mijkl that is d n+1ij d nij
Jeremi et al. c
59
< 1/2 the Generalized Midpoint rule is only conditionally stable. To obtain a stability condition for 1/2
The conclusion is that the Generalized Midpoint rule is unconditionally stable for 1/2. In the case when
1 (1 ) (n+) 1 1 + (n+)
max
1 2 for 2 (1 2)
(3.108)
critical
3.2.4
Midpoint rule algorithms in computational elastoplasticity require28 the evaluation of the intersection29 stress. Despite the appeal of the closed form solution, as found in Biani (1989), and numerical iterative procedures c c as found in Marques (1984) and Nayak and Zienkiewicz (1972), for some yield criteria30 the solution is not that simple to nd. Special problems arises, even with the numerical iterative methods in the area of a apex. The apex area problems are connected to the derivatives of yield a function. Having in mind the before mentioned problems, a dierent numerical scheme, that does not need derivatives, was sought for solving this problem. One possible solution was found in Press et al. (1988b) in the form of an excellent algorithm that combines root bracketing, bisection, and inverse quadratic interpolation to converge from a neighborhood of a zero crossing. The algorithm was developed in the 1960s by van Wijngaarden, Dekker and others at the Mathematical Center in Amsterdam. The algorithm was later improved by Brent, and so it is better known as Brents method. The method is guaranteed to converge, so long as the function31 can be evaluated within the initial interval known to contain a root. While the other iterative methods that do not require derivatives32 assume approximately linear behavior between two prior estimates, inverse quadratic interpolation uses three prior points to t an inverse quadratic function33 , whose value at y = 0 is taken as the next estimate of the root x. Lagranges classical formula for interpolating the polynomial of degree N 1 through N points y1 = f (x1 ), y2 = f (x2 ), . . . y3 = f (x3 ) is given by:
P (x) = + +
(3.109)
28
29 contact,
except for the fully implicit Backward Euler algorithm. penetration point, i.e the point along the stress path where F = 0 or the point where stress state crosses from the
elastic to the plastic region. 30 namely for the MRS-Lade elastoplastic model. 31 in our case yield function F ( ). ij 32 false position and secant method. 33 x as a quadratic function of y.
Jeremi et al. c
60
F(
contact
ij)=0
F( ij )
predictor
ij ij F=0
predictor
F( ij)
1 elastic region
Figure 3.2: The pictorial representation of the intersection point problem in computational elastoplasticity: which must be resolved for the Forward and Midpoint schemes If the three point pairs are [a, f (a)], [b, f (b)], [c, f (c)], then the interpolating formula (3.109) yields:
x= + +
(y f (a)) (y f (b)) c+ (f (c) f (a)) (f (c) f (b)) (y f (b)) (y f (c)) a+ (f (a) f (b)) (f (a) f (c)) (y f (c)) (y f (a)) b (f (b) f (a)) (f (b) f (a))
(3.110)
By setting y = 0, we obtain a result for the next root estimate, which can be written as:
x=b+
f (b) f (a)
f (a) f (c)
f (a) f (c)
(c b) 1
f (b) f (c)
f (b) f (c)
(b a)
f (b) f (a)
(3.111)
In practice b is the current best estimate of the root and the term:
Jeremi et al. c Version: April 29, 2010, 9:23
61
f (b) f (a)
f (a) f (c)
f (a) f (c)
(c b) 1
f (b) f (c)
f (b) f (c)
(b a)
f (b) f (a)
is a correction. Quadratic methods34 work well only when the function behaves smoothly. However, they run serious risk of giving bad estimates of the next root or causing oating point overows, if divided by a small number f (a) 1 f (c) f (b) 1 f (c) f (b) 1 f (a)
Brents method prevents against this problem by maintaining brackets on the root and checking where the interpolation would land before carrying out the division. When the correction of type (3.112) would not land within bounds, or when the bounds are not collapsing rapidly enough, the algorithm takes a bisection step. Thus, Brents method combines the sureness of bisection with the speed of a higher order method when appropriate.
3.2.5
Corner Problem Some yield criteria are dened with more that one yield surface35 . We will restrict our attention to a twosurface yield criterion36 . Koiter has shown in Koiter (1960) and Koiter (1953) that in the case when two yield surfaces are active, the plastic strain rate from equation (3.23) can be derived as follows:
dp = dcone ij where
cone
cone
cap
mij (ij , q )
(3.112)
mij (ij , q ) are normals to the potential functions at a corner, which belongs to
the yield functions that are active, i.e. Fcone and Fcap . We now observe that we have two nonnegative plastic multipliers dcone and dcap instead of one. We must require that at the end of the loading step37 , neither of the two yield functions is violated. These multipliers dcone and dcap can be determined from the conditions:
Fcone
n+1
ij , n+1q = 0
(3.113)
Fcap
n+1
ij , n+1q = 0
(3.114)
34
Newtons method for example. example MRS-Lade yield criterion has two surfaces. 36 having in mind MRS-Lade cone-cap yield criterion. 37 after stress correction, i.e. return to the yield surface(s).
35 for
Jeremi et al. c
62
ij
m ij
cap
predictor
cone
nij
nij = m ij
cap
corner
ij
Qcone =0 p =0 Fcone
n+1
ij
start
Figure 3.3: Pictorial representation of the corner point problem in computational elastoplasticity: Yield surfaces with singular points noting that by virtue of equation (3.112) we have at the corner singular point:
n+1
cone
cap
mkl
(3.115)
The full algorithm for the Backward Euler scheme is derived in Section (??). Apex Problem The apex problem, as depicted in Figure (3.4) is solved in an empirical fashion. Rather than facing the complexity of solving a complex dierential geometry problem38 the stress point that is situated in the gray apex region is immediately returned to the apex point. In the case when the hardening rule for the cone portion has developed to the stage that it aects the size of that cone portion of the yield criterion and not the position of intersection with the hydrostatic axis, then all stress returns from any part of apex gray region will be to the apex point itself. This strategy was used by Criseld
P Koiters work described in Koiter (1960) and Koiter (1953) and the fact that the sum dp = k dk (Fk /ij ) can be ij R p transformed into the integral equation dij = d (F /ij )|aroundapex where the integration should be carried out innitesimally
38 using
Jeremi et al. c
63
Fcone=0
ij
predictor n
ij
p
n+1
ij
Figure 3.4: The pictorial representation of the apex point problem in computational elastoplasticity: Yield surfaces with singular points (1987). Nevertheless, the problem of integrating the rate equations in the apex gray region is readily solvable for the piecewise at yield criteria39 by using Koiters conditions as found in Koiter (1960) and Koiter (1953). The apex problem for yield criteria that are smooth and dierentiable everywhere except at the apex point, is solvable by means of dierential geometry. Further work is needed for solving the problem, when the yield surface is not piecewise at in the apex vicinity.
Inuence Regions in Meridian Plane In order to dene which surface is active and which is not for the current state of stress, a simple two dimensional analysis will be conducted. The fortunate fact for the MRS-Lade material model is that such an analysis can be stress invariants p, q and for the current state of stress40 , calculate the position of the apex and corner points in p q space for given the , calculate the two dimensional gradients at these points, perform linear transformation of the current stress state41 to the new coordinate systems, and then check for the values of p i , i = 1, 2, 3, 4, where p i is the transformed pi axis. The angle is dened as the angle between the p axis and the tangent to the potential function. In Appendix
- Coulomb for example. using equations (3.167) and (3.168) as dened in section (3.4.5). 41 now in p, q and space.
40 by 39 Mohr
conducted in the p q meridian plane, only, i.e. the value for can be frozen. The concept is to calculate the
Jeremi et al. c
64
q q q
([p,q] v [p,q])
+90o p
tran
q
tran
p
2
p q
1
p
3
p
1
p0
cap
pcap p4
q4
Figure 3.5: Inuence regions in the meridian plane for the cone/cap surface of the MRS-Lade material model. (??) the gradients to the cone portion of the potential surface are dened as:
Qcone = n cone p
Qcone q = g() 1 + q qa
g() m q 1 + + qa
q qa
1+m
while, in Appendix (??) the gradients of the cap portion of the yield/potential surface is dened as:
2 (p pm ) Qcap = p p2 r
2 g()2 q 1 + Qcap = q fr 2
q qa
2m
2 g()2 m q 2 1 + + fr 2 qa
q qa
1+2 m
Jeremi et al. c
65
(3.116)
Care must be exercised with regard to which potential function is to be used in angle calculations. It should be mentioned that for the cap portion, the angle at the corner is = 0 , while at the tip of the cap, the angle is = 90 . If a new denition, as found in Ferrer (1992), is used for the cone potential function, where n is variable and n 0 as p pcap , then the corner gray region is empty. The linear transformation42 between coordinate systems p q and p q is dened as: p q
= arctan
Q p Q q
90
(3.117)
and by using that linear transformation, one can check the region where our current stress state, in p, q and space, belongs. Figure (3.5) depicts the transformation scheme and the new coordinate systems at three important points43 .
cos
sin
sin cos
tran
p q
tran
(3.118)
3.3
n The explicit algorithm (Forward Euler) is based on using the starting point (the state stress ij and internal n variable space q on the yield surface) for nding all the relevant derivatives and variables.
The Explicit algorithm can be derived by starting from a rst order Taylor expansion about starting point
n n (ij , q ):
F new
= F old + =
n
nmn
F F d (nmn ) + mn n q dmn + h d = 0
dq = (3.119)
mn = Emnpq d (pq ) d p pq
and rotation. the apex point, corner point and the cap tip point.
Jeremi et al. c
66
d =
crosn mn
With this solution for d one can obtain the increments in stress tensor and internal variables as
n
nn ab
(3.120)
dqA =
crosn mn
(3.121)
where n() denotes the starting elasticplastic point for that increment where the combined stressinternal variable state nes the yield surface. It should be noted that the explicit algorithm performs only one step of the computation and does not check on the convergence of the provided solutions. This usually results in the slow drift of the stressinternal variable point from the yield surface for monotonic loading. It also results in spurious plastic deformations during elastic unloading during cycles of loadingunloading.
ep Continuum Tangent Stiness Tensor. The continuum tangent stiness tensor (contEpqmn ) is obtained from
the explicit (forward Euler) integration procedure (Jeremi and Sture, 1997): c
cont ep Epqmn
= Epqmn
(3.122)
ep ep It is important to note that continuum tangent stiness (contEpqmn ) posses minor symmetries (contEpqmn = cont ep Eqpmn ep ep ep = contEpqnm ), while major symmetry (contEpqmn = contEmnpq ), is only retained for associated elastic
3.4
In previous sections, the general theory of elastoplasticity was presented. The accuracy and stability for the general Midpoint rule algorithm has been shown. In this chapter, the focus is on the Backward Euler algorithm, which is derived from the general Midpoint algorithm by setting = 1. The advantage of the Backward Euler scheme over other midpoint schemes is that the solution is sought by using the normal44 at the nal stress state. By implicitly assuming that such a stress state exists, the Backward Euler scheme is guaranteed to provide a solution, despite the size of the strain step45 . However, it was shown in section (3.2.3) that the Backward Euler algorithm is only accurate to the rst order.
44 m = Q ij ij 45 large strain step
increments were tested, the scheme converged to the solution even for deviatoric strain steps of 20% in magnitude.
Jeremi et al. c
67
ij = predij Eijkl
n+1
mkl
(3.123)
Q kl
where predij = Eijkl kl is the elastic trial stress state, Q is the plastic potential function and n+1mkl = is the gradient to the plastic potential function in the stress space at the nal stress position, and
pred
n+1
ij = nij + Eijkl
pred
kl
(3.124)
is the elastic predicted (trial) stress state. An initial estimate for the stress surface.
n+1
does not satisfy the yield condition, so some kind of iterative scheme is necessary to return the stress to the yield
3.4.1
ij is not in a corner or apex gray regions, a single vector return to the yield surface is
possible. In order to derive such a scheme for a single vector return algorithm, a tensor of residuals rij will be :
pred
rij = ij
pred
ij Eijkl mkl
(3.125)
This tensor represents the dierence between the current stress state ij and the Backward Euler stress state ij Eijkl mkl . The trial stress state
pred
ij is kept xed during the iteration process. The rst order Taylor series expansion
new
old
where dij is the change in ij , d() is the change in , and The goal is let
new
mkl qA dqA
= oldrij + dij + d() Eijkl mkl + Eijkl Similarly, qA = nqA + hA rA will be dened as: rA = qA (nqA + hA ) and nqA is kept xed during iteration, that = oldrA + dqA d() hA
46 By
(3.128)
(3.129)
hA hA dij + dqB ij qB
(3.130)
n+1 ()
Jeremi et al. c
68
From equation 3.127 and 3.130, one obtains mkl mkl d s I + Eijkl mn Eijkl qA mn ijmn hA A dqB ij AB hB q E m oldr ijkl kl ij +d() = + hA oldrA Since f (ij , qA ) = 0, one obtains = oldf + nmn dmn + B dqB From equations 3.131 and 3.132, d() = 1 mkl oldr s kl Iijmn + Eijkl mn Eijkl mA ij q B hA hA oldrA AB qB ij 1 mkl mkl E m s I + Eijkl mn Eijkl qA ijkl kl ijmn hA hA hA AB
ij qB
(3.131)
(3.132)
old
f nmn
nmn
(3.133)
k+1 = k + d()k
(3.134)
The iterative procedure is continued until the yield criterion f = 0, rij = , and rA = are satised within
47
plays an important role in the implicit algorithm. It should be mentioned here that the above denition is a simplied expression for very general model with various isotropic and kinematic hardening. Specically, if there is no hardening, C=
mkl s Iijmn + Eijkl mn 1
In Equation 3.133, the generalized matrix C, which is dened by 1 mkl mkl s Iijmn + Eijkl mn Eijkl qA C= hA A ij AB hB q
(3.135)
(3.136)
If there is only one isotropic internal variable q, 1 mkl s Iijmn + Eijkl mn Eijkl mkl q C= h h ij 1 q
(3.137)
For only one kinematic internal variable ij , 1 mkl mkl s Eijkl mn Iijmn + Eijkl mn C= s Iijmn hmn hmn ij ij
47
(3.138)
Jeremi et al. c
69
For one isotropic variable q and one kinematic variable ij , mkl mkl s Eijkl mn Iijmn + Eijkl mn Eijkl mkl q h h C= ij 1 h ij q s Iijmn hmn hmn hmn ij q ij or for two kinematic variables zij and ij , mkl mkl s Eijkl zmn Iijmn + Eijkl mn z h hz s C= ij Iijmn zmn h h mn mn ij zmn If we dene n mn n= B
(3.139)
mkl Eijkl mn h ij
z
(3.140)
s Iijmn
h mn ij
(3.141)
E m ijkl kl m= hA
old
(3.142)
r=
old
ij
old
d() = and
f nT C old r nT C M
rA
(3.143)
(3.144)
3.4.2
d mn = C dqB
old
r + d()m
(3.145)
Some remarks are necessary in order to clarify the Backward Euler Algorithm. It is a well known fact that the rate of convergence of the Newton - Raphson Method , or even obtaining convergence at all, is closely tied to the starting point for the iterative procedure. Bad initial or starting points might lead our algorithm to an oscillating solution, i.e. the algorithm does not converge. In the following, starting points for the Newton - Raphson iterative procedure will be established for one and twovector return algorithms. Single Vector Return Algorithm Starting Point. One of the proposed starting points (Criseld, 1991) uses the normal at the elastic trial point48 order Taylor expansion about point
pred new pred pred
ij . A rst
ij yields: d
pred pred
= =
pred old
+ +
F mn
pred
mn +
F qA
dqA =
pred
pred old
nmn dmn + A hA d = 0
(3.146)
48 I
Jeremi et al. c
70
pred
It is assumed that the total incremental strain kl is applied in order to reach the point condition kl = . From the dierential form of equation (3.36) it follows: d
pred
ij , i.e.
pred
ij =
Eijkl kl so that any further stress relaxation toward the yield surface takes place under zero total strain
mn = Emnpq d
pred p pq
pred
pq d
pred p pq pred
= mpq
= Emnpq d
= Emnpq d
pred
and it follows:
pred old
d =
predn mn
Emnpq
predm
pq
A hA
With this solution for d we can obtain the starting point for the Newton-Raphson iterative procedure
start
mn = Emnpq
pred
pred old
pq Emnpq
pred pq
predn mn
Emnpq
predm
This starting point in six dimensional stress space will in general not satisfy the yield condition F = 0, but it will provide a good initial guess for the upcoming Newton-Raphson iterative procedure. It should be mentioned, however, that this scheme for returning to the yield surface is the well known Radial Return Algorithm , if the yield criterion under consideration is of the von Mises type. In the special case the normal at the elastic trial point
pred
A hA
mpq
(3.147)
n+1
ij , the return is
exact, i.e. the yield condition is satised in one step. Another possible and readily available starting point can be obtained by applying one Forward Euler step49 . To be able to use the Forward Euler integration scheme, an intersection point has to be found. The procedure for calculating intersection points is given in section (3.2.4). A rst order Taylor expansion about intersection point F new = F old + = F F d (crossmn ) + mn cross qA cross nmn dmn + A hA d = 0
cross
ij yields: dqA =
cross
(3.148)
mn = Emnpq d (pq ) d p pq
more steps for really large strain increments, for example over 10% in deviatoric direction. What has actually been done is to
divide the region into several parts and depending on the curvature of the yield surface in deviatoric plane, use dierent schemes and dierent number of subincrements ( the more curved, the more subincrements) to get the rst, good initial guess. In the region around = 0, one step of the semi Backward Euler scheme is appropriate, but close to = /3 the Forward Euler subincrementation works better.
Jeremi et al. c
71
and it follows
cross
d =
crosn mn
With this solution for d we can obtain the starting point for the Newton-Raphson iterative procedure
start cross crossn ab
cross
mpq
(3.149)
This starting point in sixdimensional stress space will again not satisfy the yield condition50 F = 0, but will provide a good initial estimate for the upcoming Newton-Raphson iterative procedure.
3.4.3
Numerical Analysis
In this section, the accuracy analysis of the implicit algorithm is assessed. Examples of simple models (VonMises and Drucker-Prager) for accuracy analysis are demonstrated to verify the validation of the general implicit algorithm. Convergence performance analysis is conducted. More details on accuracy analysis and consistent tangent stiness are explained. Numerical simulation examples are demonstrated using the implemented framework. Special concerns are on the comparison of experimental data and numerical results of Dafalias-Manzari model. Error Assessment There are various error measures for the integration algorithms. Simo and Hughes (1998), Manzari and Prachathananukit (2001) used the relative stress norm by Equation 3.150,
r
(ij ij )(ij ij ) pq pq
(3.150)
c where ij is the exact stress solution, and ij the calculated stress solution. Alternatively, Jeremi and Sture
(3.151)
= ij Dijkl kl , and Dijkl is the elastic compliance fourth-order tensor, punit is the unit energy
The relative stress norm by Equation 3.150 is more reasonable since two points having the same ij ij
for the yield criteria that have at yield surfaces ( in the stress invariant space) so that the rst order Taylor linear
expansion, is exact.
Jeremi et al. c
72
but dierent pq pq .
In this work, we use these two error measure methods, but for simplicity, Equation 3.151 is modied into r =
(ij ij )(ij ij ) 0 0 pq pq
(3.152)
0 0 where pq pq is evaluated at some non-zero initial isotropic stress state. That is, the normalized error is evaluated
in Equation 3.152. The one-step predicted stress state point for the implicit algorithm is within the range of 0.1 p 100 kPa, 0 q 100 kPa, 0 /3. The exact solution is actually unknown for most elastoplastic problems. Here the exact solution is simply replaced by 50 substep solution of the explicit algorithm in the same one-step prediction incremental. All these error evaluations are within the material constitutive level. The rst test examples are Von-Mises models with the uniaxial yield strength k = 50 kPa, with linear elasticity parameters are Youngs modulus E = 1105 kPa, and Poissons ratio = 0.25. Figures 3.6 and 3.7 show the iso-error maps for the Von-Mises model with linear isotropic hardening. The linear hardening modulus H = 2104 kPa. The blue lines represents the yield surface boundary. It can be seen that the error magnitudes are as small as 1010 to 109 , which implies that the solutions by implicit algorithm for this linear isotropic hardening Von-Mises model are numerically accurate if one realized that the machine oating errors cannot be avoided.
100
1e0
1.5
01 0
6e010 5e010
1e009
80
5e
80
09
00
5e
2e010
010
1e010
q (kPa)
40
q (kPa) (a) At = 0
60
60
40
20
20
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
Figure 3.6: Normalized iso-error maps of Von-Mises model with linear isotropic hardening. Figures 3.8 and 3.9 show the iso-error maps for the Von-Mises model with Armstrong-Frederick translational kinematic hardening. The hardening parameters are ha = 5104 kPa and Cr = 2.5103 . It can be seen that errors are very small which proves the good performance of the implicit algorithm. The iso-error map gives a good trend, i.e., the further away from the yield surface, the errors become more pronounced; the normalized errors are pressure-independent, which ts well the feature of Von-Mises model; the iso-error lines in the q gure are parallel to the yield surface and are independent of the Lodes angle , which again ts well with Von-Mises
Version: April 29, 2010, 9:23
Jeremi et al. c
73
100
2e 0 09 1 .5 e 0 09
1e
0 0
100
80
q (kPa)
40
q (kPa)
60
5e
10
0 e
09
80
20
(a) At = 0
010 5e
5e010
1e010
60
40
20
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
Figure 3.7: Relative iso-error maps of Von-Mises model with linear isotropic hardening. model which is only q-related.
100 0.016 0.014 0.012 0.01 0.008 0.006 0.004 0.002 100 0.016 0.014 0.012 0.01 0.008 0.006 0.004 0.002
0.01
80
0.01
80
q (kPa)
40
q (kPa) (a) At = 0
60
60
40
20
20
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
Figure 3.8: Normalized iso-error maps of Von-Mises model with Armstrong-Frederick kinematic hardening. The second test examples are Drucker-Prager model with yield surface constant q/p = 0.8. Linear elasticity parameters are Youngs modulus E = 1105 kPa, and Poissons ratio = 0.25. The iso-error maps for perfectly plastic Drucker-Prager model are shown in Figures 3.10 and 3.11. The blue lines represents the yield surface boundary. It can be seen that the error magnitudes are as small as 1011 to 109 . Again, these errors are so small that we can consider that the implicit algorithm give accurate solutions numerically. Another Drucker-Prager model is with Armstrong-Frederick rotational kinematic hardening, and the parameters are ha = 20, Cr = 2. The iso-error maps are shown in Figures 3.12 and 3.13. Unlike Von-Mises model, the normalized errors are pressure-dependent, which ts well the feature of Drucker-Prager model; the iso-error lines
Jeremi et al. c Version: April 29, 2010, 9:23
74
100
0 0. 0. .03 04 02
100
0.0
80
80
.01
q (kPa)
40
q (kPa) (a) At = 0
60
60
40
20
20
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
Figure 3.9: Relative iso-error maps of Von-Mises model with Armstrong-Frederick kinematic hardening.
8 1e009 e 0 10 1e009 8e010 6e010 4e010 2e010
100
100
80 09 1.4e0
80
1.2e009
q (kPa)
40
q (kPa) (a) At = 0
6e010 4e010
60
60
40
10
2e0
20
20
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
Figure 3.10: Normalized iso-error maps of Drucker-Prager perfectly plastic model. in the q gure are parallel to the yield surface and are independent of the Lodes angle , which still ts well
with Drucker-Prager model which does not consider the third stress invariant, Lodes angle . From the relative iso-error maps in Figure 3.13, very dense iso-error lines are investigated in the region of small pressure, which is
evidently due to the cone apex singularity of Drucker-Prager yield surface. From the error analysis by the above Von-Mises and Drucker-Prager models, One nds that the implemented implicit algorithm can oer accurate solutions for simple models with simple hardening laws, e.g. Von-Mises model with linear hardening and Drucker-Prager model with perfectly plastic hardening (no hardening). Complicated hardening laws increases the error even for simple plastic models, although the errors are still small. These observations match the well known conclusion that the error of the implicit algorithm is pretty dependent on the smoothness of the solution. The implemented implicit algorithm proves very robust for Von-Mises and DruckerPrager model with simple or complicated hardening laws.
Jeremi et al. c Version: April 29, 2010, 9:23
75
100
2e009 1e009
100
009 2e
1e0094e009 3e009
1.4e010
8e011 6e011
80
80
q (kPa)
40
q (kPa) (a) At = 0
60
60
20
100
80
q (kPa)
40
q (kPa)
60
20
Figure 3.12: Normalized iso-error maps of Drucker-Prager model with Armstrong-Frederick kinematic hardening. Figures 3.14 and 3.15 present the iso-error maps of Dafalias-Manzari model. The initial void ratio is 0.8, and the other parameters are from Dafalias and Manzari (2004). The blue lines represents the yield surface boundary Dafalias-Manzari model are not parallel to the yield surface and are dependent of the Lodes angle , which was when the predicted stress q close to 100 kPa, or or about 100 times the yield strain increment, the relative errors can reach up to 100%, which implies that even for implicit algorithm, Dafalias-Manzari model still requires small strain increments. However, when q < 30 kPa, or about 30 times of the yield strain increment, the relative errors are less than 5%, excepts at the region close to the yield surface apex. It should be pointed out that errors for the complex Dafalias-Manzari model are much bigger than those of simple models (e.g. Von-Mises and Drucker-Prager), due to its high non-linearity. However, if the predicted stress
Jeremi et al. c Version: April 29, 2010, 9:23
(slope ratio m = 0.01). Unlike Von-Mises and Drucker-Prager models, the iso-error lines in the q gure of
one of the highlighting improvements upon the previous version (Manzari and Dafalias, 1997). From Figure 3.15,
5e009 e009 6
0 0
40
20
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
100
80
60
40
(a) At = 0
20
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
76
100 0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001
80
100
80
q (kPa)
40
q (kPa) (a) At = 0
60
60
40
20
20
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
Figure 3.13: Relative iso-error maps of Drucker-Prager model with Armstrong-Frederick kinematic hardening. (or in other words, the strain increment) is small enough, the algorithm errors are within a small tolerant range.
100
0.25
0.25
100
0. 2
0.25
0.2
80 0.2
80
0.2
0.15
0.15
0.1
0.15
q (kPa)
q (kPa)
60
60
40
0.2 0.15 0. 1
0.1
40
0.1
0.05
0.05
20
0.05
20
(a) At = 0
0.2 0.15 0.1 0.05
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
Figure 3.14: Normalized iso-error maps of Dafalias-Manzari model with average elastic moduli. Figures 3.14 and 3.15 are based on an approach of averaged elastic moduli. Instead, Figures 3.16 and 3.17 present iso-error maps based on constant elastic moduli approach. The averaged elastic moduli approach improves the accuracy against the averaged elastic moduli approach. More explanation on these two approaches and other performance analysis of Dafalias-Manzari model will be in section ??. Constitutive Level Convergence In the implemented implicit algorithm, the iteration continues until the absolute value of yield function and the residue norm of considering variables are less than some small tolerances, or if by equations, |f | T ol1; rnorm = r T ol2
Jeremi et al. c
(3.153)
Version: April 29, 2010, 9:23
77
100
100
0.2
0.25
0.25
0.2
1 0.8 0.6
80
0.4
80
0.2
0.15
0.15
q (kPa)
q (kPa)
0.2
60
60
40
0.1
40
0.1
0.05
0.05
0.2
20
(a) At = 0
20
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
Figure 3.15: Relative iso-error maps of Dafalias-Manzari model with average elastic moduli.
3
0.35 0.4
0.2
100
100
0.
0.25
0.2
0.1
80
80
0.2
0.2
0.15
0.2
q (kPa)
q (kPa)
0.15
60
0.3
0.2
0.15
60
0.1
0.1
0.25
40
0.
1
0.05
0.2
0.1
40
0.1
0.05
0.1 0.15
20
0.05
0.2
0.15 0.1
0.05
20
(a) At = 0
0.2 0.25
50
100
150
200
0.2
0.4
p (kPa)
(rad)
0.6
0.8
Figure 3.16: Normalized iso-error maps of Dafalias-Manzari model with constant elastic moduli. Three examples including simple Von-Mises model with linear isotropic hardening, relative complicated DruckerPrager model with Armstrong-Frederick kinematic hardening, and even more complicated Dafalias-Manzari model considering fabric dilation eect are presented here to show the constitutive level convergence performances for Iteration number 0 represents the virtual iteration number before return mapping implicit iteration cycle. |f | at iteration number 0 thus means |f | at the rst predicted stress for each load increment; there is no value of rnorm the implemented implicit algorithm. In all these examples, both |f | and rnorm v.s. iteration numbers are plotted.
at iteration number 0. A tolerance of T ol1 = T ol2 = 1107 is for both |f | and rnorm . The iteration stops when |f | T ol1 and rnorm = T ol2 are satised, even if there is only one iteration number. The initial stress
is an isotropic stress state of p0 = 100 kPa. The undrained-like load increment is adopted by strain control as ij are strain components.
Jeremi et al. c
11 = 222 = 233 = n, where n is the load increment number and is the strain increment interval,
78
100
0.25
100
1.6 1.4
0.25
0 .4
0.2
80
80
0.2
0.2
0.15
0.15
0.1
q (kPa)
q (kPa)
60
60
40
0.2
0.1
40
0.05
20
0.05
0.8
(a) At = 0
0.2 0.4 0.6 50 100 150 200
20
0.2
0.4
p (kPa)
(rad)
0.6
0.8
Figure 3.17: Relative iso-error maps of Dafalias-Manzari model with constant elastic moduli. Figure 3.18 shows the typical constitutive level convergence performance for Von-Mises model with linear isotropic hardening. The input parameters are Youngs Modulus E = 1105 kPa, Poissons ratio = 0.25, the material strength k = 50 kPa, and the linear isotropic hardening modulus H = 2104 kPa. The strain increment interval is set 2104 . It can be seen that for this simple example, only two iteration steps are needed and words, the stresses are exactly at the yield surface and the residue norm is zero.
10 0 10
1
|f | and rnorm are far smaller than the tolerances and in fact close to the machine oating error value, or in other
|f| rnorm
10
Tolerance Line
10
10
10
15
Iteration Number
Figure 3.18: Typical convergence for Von-Mises model with linear isotropic hardening (tolerance value 1107 ). Figure 3.19 shows the typical constitutive level convergence performance for Drucker-Prager model with
Jeremi et al. c Version: April 29, 2010, 9:23
79
Armstrong-Frederick kinematic hardening. The input parameters are Youngs Modulus E = 1104 kPa, Poissons ratio = 0.25, the material q/p ratio is 0.8, and the Armstrong-Frederick parameters are ha = 20, Cr = 2. The the increasing iteration number; However, |f | and rnorm show dierent rates; |f | needs 5 iteration steps while strain increment interval is set 2104 . For this example, both |f | and rnorm are stably decreasing with
rnorm needs 7 iteration steps; The convergence rate of rnorm lags behind that of |f |.
10 0 10
|f| rnorm
10
Tolerance Line
10
10
10
15
Iteration Number
Figure 3.19: Typical convergence for Drucker-Prager model with Armstrong-Frederick kinematic hardening (tolerance value 1107 ).
Figure 3.20 shows the typical constitutive level convergence performance for the complicated Dafalias-Manzari model considering fabric dilation eect. The input parameters are as in Table ??, and the initial void ration is set as 0.8. Dierent from the above examples, The strain increment interval is set a much smaller value of However, |f | and rnorm show dierent rates; |f | needs less iteration steps than rnorm ; The convergence rate of 1105 . In this example, again, both |f | and rnorm are stably decreasing with the increasing iteration number;
fabric dilation eect, the typical constitutive level convergence performance is similar to that of Drucker-Prager model with Armstrong-Frederick kinematic hardening, but with much smaller strain increment interval. From the above examples, it is clear that the simpler the model is , the better constitutive level convergence
rnorm lags behind that of |f |. It should be mentioned here for this complicated Dafalias-Manzari model considering
performances are observed. This is consistent to the error assessment in section 3.4.3. Generally, the implemented implicit algorithm shows stable constitutive level convergence performances provided an appropriate small strain increment interval for the material model.
Jeremi et al. c Version: April 29, 2010, 9:23
80
10 0 10
|f| rnorm
10
Tolerance Line
10
10
10
15
Iteration Number
Figure 3.20: Typical convergence for Dafalias-Manzari model (tolerance value 1107 ).
3.4.4
The nal goal in deriving the Backward Euler scheme for integration of elastoplastic constitutive equations is to use that scheme in nite element computations. If the Newton Raphson iterative scheme is used at the global
ep equilibrium level then the use of the so called traditional tangent stiness tensor51 Eijkl destroys the quadratic
rate of asymptotic convergence of the iterative scheme. In order to preserve such a quadratic rate, a consistent, also called algorithmic, tangent stiness tensor is derived. The consistent tangent stiness tensor make use of derivatives of direction52 normal to the potential function, and they are derived at the nal, nal at each iteration, that converges to the nal stress point on the yield surface, stress point. The traditional forward scheme has a constant derivative, mij that is evaluated at the intersection point. It appears that Simo and Taylor (1985) and Runesson and Samuelsson (1985) have rst derived the consistent tangent stiness tensor. Other interesting articles on the subject can be found in Simo and Taylor (1986), Simo and Govindjee (1988), Jetteur (1986), Braudel et al. (1986), Criseld (1987), Ramm and Matzenmiller (1988) and Mitchell and Owen (1988). As a consequence of consistency, the use of the consistent tangent stiness tensor signicantly improves the convergence characteristics of the overall equilibrium iterations, if a Newton - Raphson scheme is used for the latter. Use of the consistent tangent stiness tensor yields a quadratic convergence rate of Newton - Raphson equilibrium iterations. In what follows, two derivations are given, namely the consistent tangent stiness tensor for single and twovector return algorithms. The concept of consistent linearization was introduced by Hughes and Pister (1978), while detailed explanation is given by Simo and Hughes (1998). The consistent tangent stiness leads to quadratic convergence rates at
51 the 52 m ij
one obtained with the Forward Euler method, i.e. where parameter = 0. = Q/ij , i.e. mij /kl = 2 Q/ij kl .
Jeremi et al. c
81
global level. It should be mentioned that there are various equivalent forms of consistent tangent stiness depending on the specic implicit algorithm equations. For instance, Simo and Hughes (1998), and Belytschko et al. (2001) derived the consistent tangent stiness by taking current plastic strain as unknown and seeking its derivatives in the stress space; Prez-Foguet and Huerta and Prez-Foguet et al. (2000) used the numerical dierentiation e e to calculate the consistent tangent stiness in a compact matrix-vector form; Choi (2004) adopted the compact matrix-vector form by Prez-Foguet and Huerta and Prez-Foguet et al. (2000) but taking current plastic strain e e as unknown and seeking its derivatives in the elastic strain space. Slightly dierent from the above strategies, in this work the implicit algorithm is adopting the traditional form but taking current stress as unknown and seeking its derivatives in the stress space. Provided these dierences, the consistent tangent stiness in this work is slightly dierent from those in the above work.
Single Vector Return Algorithm. In implicit algorithm, a very important advantage is that the implicit algorithm may lead to consistent (algorithmic) tangent stiness (Equation 3.161), a concept against continuum tangent stiness (Equation ??). The concept of consistent linearization was introduced in Hughes and Pister (1978), more details on consistent tangent stiness were explained in Simo and Hughes (1998). The consistent tangent stiness leads to quadratic convergence rates at global level. It should be mentioned that there are various equivalent forms of consistent tangent stiness depending on the specic implicit algorithm equations. For instance, Simo and Hughes (1998), and Belytschko et al. (2001) derived the consistent tangent stiness by taking current plastic strain as unknown and seeking its derivatives in the stress space; Prez-Foguet and Huerta and Prez-Foguet et al. (2000) used the numerical dierentiation e e to calculate the consistent tangent stiness in a compact matrix-vector form; Choi (2004) adopted the compact matrix-vector form by Prez-Foguet and Huerta and Prez-Foguet et al. (2000) but taking current plastic strain e e as unknown and seeking its derivatives in the elastic strain space. Slightly dierent from the above strategies, in this work (section ??) the implicit algorithm is adopting the traditional form but taking current stress as unknown and seeking its derivatives in the stress space. Provided these dierences, the consistent tangent stiness in this work is slightly dierent from those in the above work. The detail derivation will be followed. When seeking the algorithmic tangent stiness, we look into the explicit expression of dij /dpred . At the mn same time, the internal variables are initialized the values at the previous time step, in other words, they are xed within the time step when seeking the algorithmic tangent stiness. Linearize Equation 3.123, one obtains dij = Eijkl dpred d() Eijkl mkl Eijkl kl Similarly, linearize Equation 3.128, one obtains dqA = d() hA +
Jeremi et al. c
(3.154)
hA hA dij + dqB ij qB
(3.155)
Version: April 29, 2010, 9:23
82
From equation 3.154 and 3.155, one obtains mkl d s kl Iijmn + Eijkl mn Eijkl mA mn q hA hA dqB AB qB ij pred E m E ijkl kl ijkl dkl +d() = hA 0
(3.156)
Linearize the yield function f (ij , qA ) = 0, one obtains nmn dmn + B dqB = 0 or in a simplied form d mn nT =0 dqB
If one use the denitions of 3.135, 3.142 and 3.141, Equation 3.156 can be simplied to d E dpred mn ijkl kl C1 + d()m = dqB 0
(3.157)
(3.158)
(3.159)
Substitute expression 3.160 into 3.157, one obtains d E pred T Cmn C ij ijmn dmn = C T dqA n Cm 0
From Equations 3.157 and 3.159, one obtain E pred T n C ijmn dmn d() = T n Cm 0
(3.160)
(3.161)
This equation gives the explicit expression of the consistent tangent stiness dij /dpred for the implicit algorithm. mn From section ??, if there are interactions between internal variables, the implicit algorithm will become very complicated. Simple models (e.g. Von-Mises model, or sometimes termed as J2 model) have been proved ecient and good performance by the implicit algorithm (Simo and Hughes, 1998). Evidently, the implicit algorithm
is mathematically based on the Newton-Raphson nonlinear equation solving method as well as the Eulerian backward integration method. Theoretically, the Newton-Raphson method may have quadratic convergence rate. However, Newton-Raphson method is not unconditional stable, and sometimes the iteration will diverge (Press et al., 1988a). Any bad starting point, non-continuous derivatives around solution, high nonlinearity, and interactions between internal variables, will deteriorate the implicit algorithm performance. A complicated model cannot guarantee good performance or quadratic convergence by the implicit algorithm Criseld (1997). The task to obtain the analytical expressions (Equations 3.136 to 3.140) may prove exceeding laborious for complicated plasticity models Simo and Hughes (1998). The explicit and implicit algorithm performances for the simple model such as Von-Mises and Drucker-Prager models should not be considered novel, however, these examples can verify the validation of the implemented algorithms in the general framework of NewTemplate3Dep. To demonstrate the implicit algorithm performance,
Jeremi et al. c Version: April 29, 2010, 9:23
83
an one-element (111) triaxial compression example of Drucker-Prager model with linear kinematic hardening, the elastic Youngs modulus is 20000 kPa, Poissons ratio is 0.25, yield surface constant q/p = 0.8, linear kinematic hardening modulus is 2000 kPa. the number of the total incremental steps is 100 with 0.0005 axial displacement increment in each step. Table 3.1 and Figure 3.21 shows the number of iteration per step and the typical current residual norms by output interface of OpenSees are shown in Table 3.2. It can be seen that for this problem, only 1-3 global Newton iteration steps are needed and most iteration steps are 2. The iteration convergence rates are quadratic and almost quadratic. Table 3.1: Number of Newton iteration/step. Step Iterations 1-4 3 5-71 2 72-100 1
Table 3.2: Residual norms for typical iteration steps (tolerance value 1106 ). Step Residual Norm 4 6.80E-05 1.32E-06 3.47E-07 20 4.50E-05 4.73E-09 40 1.26E-06 2.36E-10 80 9.48E-09
10
10
Residual Norm
10
10
10
10
10
10
Iteration Number
Figure 3.21: Residual norms for typical iteration steps (tolerance value 1106 ). The performances analysis of the implemented algorithms for the much more complicated Dafalias-Manzari model need additional eorts. The reasons are (1) the elasticity of Dafalias-Manzari model is nonlinear; (2) the yield surface of Dafalias-Manzari model has very small slope; (3) the third stress invariant, or Lodes angle is considered; (4) the void ratio (volumetric strain) is involved in the constitutive relations; (5) complicated formulations.
Jeremi et al. c Version: April 29, 2010, 9:23
84
The elastic nonlinearity creates an elastic stiness that is not a constant during prediction and return mapping iterations. Borja and Lee (1990); Borja (1991) used a secant elastic stiness for the fully implicit algorithm of Cam-Clay elastic-plastic model. Similar strategy was presented by Manzari and Prachathananukit (2001) in the fully implicit algorithm of Manzari-Dafalias model (1997), which was the simplied version of Dafalias-Manzari model (2004). However, this method requires a direct integration for predicted elastic bulk or shear modulus before calculating the secant stiness. Dierent models have dierent integration equations, which makes this method lose generality. The simplest way is to assume the elastic stiness is constant. However, constant elastic stiness assumption may lead to too much errors. In our implementation, we use average elastic stiness instead of secant elastic stiness for general non-linear elasticity. The elastic stiness can be expressed by:
1 Eijkl
= E 1 (ij )
1 = ij + Eijkl kl (k) (k+1)
(k)
ij
(k+1)
2 Eijkl
= E 2 (ij =
Eijkl
1 2 0.5(Eijkl + Eijkl )
From Figures 3.14 and 3.15 to Figures 3.16 and 3.17, The averaged elastic moduli approach proves improved accuracy against the averaged elastic moduli approach. The very small slope of yield surface of Dafalias-Manzari model can introduce another problem. Suppose the initial stress point is at p = p0 , q = 0, = 0, the volumetric strain increment is zero, and the deviatoric strain increment is q . The small slope of yield surface of Dafalias-Manzari model means a small value of q yield (Manzari and Prachathananukit, 2001) named it yield strain increment) will make the material yield. For Dafalias-Manzari model, q yield mp0 G
(3.166)
where G is dened by Equation 3.297. For Toyoura sand, q yield can be easily calculated by the material
0.5 6 0.5 parameters given by Dafalias and Manzari (2004), e.g. q p0 for e = 0.8 if p is in yield = kyield p0 = 310
terms of kPa. Although other soils may have dierent kyield from Toyoura sand, it should be no far away from that of Toyoura sand. The small value of kyield can hint us that Dafalias-Manzari model requires very small strain increment for good performance of elastoplastic calculation, which should be considered as one featured shortcoming of Dafalias-Manzari model. Although the simple models, e.g. Von-Mises and Drucker-Prager model, the explicit formulation of the algorithmic (consistent) tangent stiness can be easily obtained by Equation 3.161, the explicit expression of the algorithmic tangent stiness for Dafalias-Manzari model is very laborious to be obtained, for the reason that the
in void ratio is involved in various constitutive relations, the elasticity is not linear, and the ij heavily depends on
the reverse loading paths. Its algorithmic tangent stiness is only approximate by using the general calculation framework for all material models. The global performance will be eected although the local performance proves well, which again limit the permissible strain increments.
Jeremi et al. c Version: April 29, 2010, 9:23
85
3.4.5
In the derivation of the Backward Euler algorithm and the Consistent Tangent Matrix it is necessary to derive the rst and the second derivatives of the potential function. The function Q is the function of the stress tensor ij and the plastic variable tensor qA . Derivatives with respect to the stress tensor ij and plastic variable tensor qA are given here. It is assumed that any stress state can be represented with the three stress invariants p, q and given in the following form: 3 3 cos 3 = 2
1 p = I1 3
q=
3J2D
J3D (J2D )3
(3.167)
I1 = kk
J2D =
1 sij sij 2
J3D =
1 sij = ij kk ij 3
(3.168)
Stresses are here chosen as positive in tension. The denition of Lodes angle in equation (3.167) implies that = 0 denes the meridian of conventional triaxial extension (CTE), while = /3 denotes the meridian of conventional triaxial compression (CTC). The Potential Function is given in the following form:
Q = Q(p, q, ) The complete derivation of the closed form gradients is given in Appendix C. Analytical Gradients The rst derivative of the function Q in stress space is:
(3.169)
Q Q p Q q Q = + + ij p ij q ij ij
(3.170)
and subsequently the rst derivatives of the chosen stress invariants are
p 1 = ij ij 3
(3.171)
3 1 q = sij ij 2 q
(3.172)
Jeremi et al. c
86
(3.173)
where:
tij =
J3D ij
+ +
2 Q q 2 Q 2 Q p + + p2 mn pq mn p mn 2 2 Q q 2 Q Q p + + 2 qp mn q q mn mn 2 Q p 2 Q q 2 Q + + p mn q mn 2 mn
2Q pq mn 2p p Q + pq p pq mn 2q Q q + pq q pq mn 2 Q + pq pq mn
= + + (3.174)
2p = pq mn
(3.175)
2q 31 = pq mn 2q
1 pm nq pq nm 3
9 1 smn spq 4 q3
(3.176)
2 = pq mn + 9 cos 3 1 27 cos 3 81 + spq tmn + spq smn + 2 q 4 sin (3) 4 q 4 sin3 3 4 q 5 sin3 3 243 cos 3 1 81 cos2 3 81 tpq tmn + tpq smn + 4 q 5 sin 3 4 q 5 sin3 3 4 q 6 sin3 3 3 cos (3) 1 9 + 2 ppqmn wpqmn 2 q sin (3) 2 q 3 sin (3)
(3.177)
where:
Jeremi et al. c Version: April 29, 2010, 9:23
87
wpqmn =
and:
ppqmn =
spq = mn
1 mp nq pq mn 3
(3.178)
Finite Dierence Gradients After having developed the closed form, analytical derivatives53 the author of this thesis asked himself: is there a simpler way of nding these derivatives? One of the proposed ways to check the analytical solution is found in Dennis and Schnabel (1983). Dennis and Schnabel proposes the nite dierence method for approximating derivatives if these derivatives are not analytically available and as a tool to check your analytical derivatives if they are derived. Another good reason for developing alternative gradients is that for = 0, /3 gradients are not dened, i.e. indenite terms as 0/0 are appearing. One possible solution is the use of lHospitals rule. This has been done in Peri (1991). The solution to the problem in this work went in a dierent direction, i.e. instead of aiming for the c analytical form, numerical derivatives are derived. We should recall that for a function f of a single variable, the nite dierence approximation to f (x), by using forward nite dierence approach, is given by:
a=
f (x + h) f (x) h
(3.179)
where h is a vanishingly small quantity. The same denition was used in deriving the nite dierence approximation for the rst derivative of the yield function F and potential function Q. The rst derivative of F ( or Q ) with respect to the stress tensor ij for diagonal elements is54 :
53 see 54 no
Appendix (C). sum convention implied, just the position of the element.
Jeremi et al. c
88
approx.
F,ii =
(3.180)
approx.
F,ij =
(3.181)
where hij is the step size which, because of nite precision arithmetic, is a variable56 . The accuracy of the nite dierence approximation to the analytical derivatives is closely bound to the step size hij . It is suggested in Dennis and Schnabel (1983)[section 5.4.] that for functions given by the simple formula, the number h should be h = macheps, while for more complicated functions that number should be larger. Here macheps is the so called machine epsilon. It is dened as the smallest distinguishable positive number57 , such that 1.0 + macheps > 1.0 on the given platform. For example58 , on the Intel x86 platform59 that in the case of yield or potential functions the best approximation of analytical gradients is obtained by using h = macheps 103 . The three order of magnitude increase in the nite dierence step is due to a rather complicated60 formula for yield and potential functions. The error in the approximation, be after the N
th approx.
macheps = 1.08E 19 while on the SUNSparc and DEC platforms macheps = 2.22E 16. It has been found
F,ij is found to
Second derivative approximations for one variable function are given in the form:
a=
(f (x + hi ei + hj ej ) f (x + hi ei )) (f (x + hj ej ) f (x)) hi hj
(3.182)
If the rst derivatives are available in closed form, one could use equations (3.180) and (3.181) just by replacing the function values with tensor values for analytical derivatives61 . However, if the analytic derivatives are not available, one has to devise a formula that will create a fourth order tensor from the changes in two dimensional stress tensors, ij and kl . Using the scheme employed in equation (3.182) the following scheme has been devised:
Q,ijkl = (Q(mn + hij + hkl ) Q(mn + hij )) (Q(mn + hkl ) Q(mn )) hij hkl (3.183)
55 since 56 it
approx.
the stress tensor ij is symmetric, change in one non-diagonal element triggers the other to be changed as well. is actually one small number, h, that is multiplied with the current stress value so that the relative order of magnitude is
retained. 57 in a given precision, i.e. oat ( real*4 ), double ( real*8 ) or long double ( real*10 ). 58 the precision sought was double ( real*8 ). 59 PC computers. 60 One should not forget that we work with six dimensional tensor formulae directly. 61 see Dennis and Schnabel (1983), section 5.6.
Jeremi et al. c
89
Special considerations are necessary in order to retain symmetry of the fourth order tensor. At the moment it has not been possible to gure out how to build the nite dierence approximation to the second derivatives of yield/potential functions for a general stress state. The only nite dierence approximation of the second derivatives that appears to have worked was the one devised in principal stress space. Namely, diagonal elements of the analytical and the approximate gradients matched exactly, but development of non-diagonal elements, and the whole scheme of symmetrizing the fourth order approximation, still remain a mystery. However, some pattern was observed in nondiagonal elements, and the work on symmetrizing it is in progress. For many dierent potential functions (or yield functions) the only task left would be the derivation of the rst derivatives of F and Q and the second derivatives of Q with respect to p, q and , namely the rst derivatives
Q Q q and 2Q 2 . If the Q p ,
and
Q Q p , q
and
2Q 2Q 2Q 2Q 2Q 2Q 2Q 2Q p2 , pq , p , qp , q 2 , q , p , q
and
potential function is twice dierentiable with respect to the stress tensor ij , and if it is continuous
3.5
In this section we present elements of general elasticplastic material models for geomaterials. We describe various forms of the yield functions, plastic ow directions and hardening and softening laws.
3.5.1
Elasticity
In elasticity the relationship between the stress tensor ij and the strain tensor kl can be represented in the following form: ij = (ij ) In its simplest (incremental) form it reads ij = Eijkl kl (3.185) (3.184)
where Eijkl is the fourth order elastic stiness tensor with 81 independent components in total. The elastic stiness tensor features both minor symmetry Eijkl = Ejikl = Eijlk and major symmetry Eijkl = Eklij (Jeremi and Sture, 1997). The number of independent components for such elastic stiness tensor is 21 (Spencer, c 1980). Most of the models used to describe elastic behavior of soils assume isotropic behavior. The most general form of the isotropic elastic stiness tensor of rank 4 has the following representation: Eijkl = ij kl + (ik jl + il jk ) where and are the Lam coecients: e = E (1 + ) (1 2) ; = E 2 (1 + ) (3.187) (3.186)
90
The elastic isotropic behavior of soils obeys Hookes law with a constant Poissons ratio. The variation of the Youngs modulus E is usually assumed to be a function of the stress state. To this end we use four dierent elastic laws. Linear Elastic Model. Linear elastic law is the simplest one and assumes constant Youngs modulus E and constant Poissons Ration . Nonlinear Elastic Model #1. This nonlinear model (Janbu, 1963), (Duncan and Chang, 1970) assumes dependence of the Youngs modulus on the minor principal stress 3 = min in the form E = Kpa 3 pa
n
(3.188)
Here, pa is the atmospheric pressure in the same units as E and stress. The two material constants K and n are constant for a given void ratio. It should be noted that this material model was developed in order to capture the reduction of soil stiness with increasing strain without any reference to plasticity. However, the model can be used (with some care) to represent nonlinear elastic behavior of soils. Nonlinear Elastic Model #2. If Youngs modulus and Poissons ratio are replaced by the shear modulus G and bulk modulus K the nonlinear elastic relationship can be expressed in terms of the normal eective mean stress p as G and/or K = AF (e, OCR)pn (3.189)
where e is the void ratio, OCR is the overconsolidation ratio (related to the overburden stress) and p = ii /3 is the mean eective stress (Hardin, 1978). Lades Nonlinear Elastic Model. Lade and Nelson (1987) and Lade (1988a) proposed a nonlinear elastic model based on Hookes law in which Poisson ratio is kept constant. According to this model, Youngs modulus can be expressed in terms of a power law as: E = M pa I1 pa
2
1+ + 6 1 2
J2D p2 a
(3.190)
where I1 = ii is the rst invariant of the stress tensor and J2D = (sij sij )/2 is the second invariant of the deviatoric stress tensor sij = ij kk ij /3. The parameter pa is atmospheric pressure expressed in the same unit as E, I1 and J2D and the modulus number M and the exponent are constant, dimensionless numbers.
3.5.2
Yield Functions
The typical plastic behavior of frictional materials is inuenced by both normal and shear stresses. It is usually assumed that there exists a yield surface F in the stress space that encompasses the elastic region. States of stress inside the yield surface are assumed to be elastic (linear or nonlinear). Stress states on the surface are assumed to produce plastic deformations. Yield surfaces for geomaterials are usually shaped as asymmetric tar drops with smoothly rounded triangular cross sections. In addition to that, simpler yield surfaces, based on
Jeremi et al. c Version: April 29, 2010, 9:23
91
the DruckerPrager cone or MohrCoulomb hexagon can also be successfully used if matched with appropriate hardening laws. Yield surface shown in Figure 3.22 Lade (1988b) represent typical meridian plane trace for an isotropic granular material. Line BC represents stress path for conventional triaxial compression test. Figure 3.23
Figure 3.22: Yield surface patterns in the meridian plane for isotropic granular materials (from Lade (1988b)) represents the view of the yield surface traces in the deviatoric plane.
2
low confinment trace high confinment trace
Figure 3.23: Deviatoric trace of typical yield surface for pressure sensitive materials.
3.5.3
Plastic ow directions are traditionally derived from a potential surface which to some extent reassembles the yield surface. Potential surfaces for metals are the same as their yield surfaces but experimental evidence suggests that it is not the case for geomaterials. The nonassociated ow rules, used in geomechanics, rely on the potential surface, which is dierent from the yield surface, to provide the plastic ow directions. It should be noted that
Jeremi et al. c Version: April 29, 2010, 9:23
92
the potential surface is used for convenience and there is no physical reason to assume that the plastic strain rates are related to a potential surface Q (Vardoulakis and Sulem, 1995). Instead of dening a plastic potential, one may assume that the plastic ow direction is derived from an tensor function which does not have to possess a potential function.
3.5.4
The change in size and/or shape of the yield and potential surfaces is controlled by the hardeningsoftening evolution laws. Physically, these laws control the hardening and/or softening process during loading. Depending on the evolution type they control, these laws can be in general separated into isotropic and kinematic (also called anisotropic). The isotropic evolution laws control the size of the yield surface through a single scalar variable. This is usually related to the Coulomb friction or to the mean stress values at isotropic yielding. The nonisotropic evolution laws can be further specialized to rotational, translational kinematic and distortional. It should be noted that all of the kinematic evolution laws can be treated as special case of the general, distortional laws (Baltov and Sawczuk, 1965). Figure 3.24 depicts various types of evolution laws (for the control of hardening softening) in the meridian plane62 .
q q
a)
q
p c
b)
q
c)
d)
Figure 3.24: Various types of evolution laws that control hardening and/or softening of elasticplastic material models: (a) Isotropic (scalar) controlling equivalent friction angle and isotropic yield stress. (b) Rotational kinematic hardening (second order tensor) controlling pivoting around xed point (usually stress origin) of the yield surface. (c) Translational kinematic hardening (second order tensor) controlling translation of the yield surface. (d) Distortional (fourth order tensor) controlling the shape of the yield surface.
3.5.5
Tresca Model
The rst yield criteria in the metal plasticity is Tresca yield criteria. Tresca yield criteria states that when the maximum shear stress or, the half dierence of the maximum and minimum principal stresses, reaches the shear
62 The
meridian plane is chosen just for illustration purposes, similar sketch can be produced in deviatoric plane as well.
Jeremi et al. c
93
strength, s , the material will begin yielding. It is can be expressed by the yield function f = |max | s = 1 |1 3 | s = 0 2 (3.191)
Tresca yield surface in the principal stress space is a regular hexagonal cylinder. It is implied that the intermediate principal stress plays no role in the yielding for Tresca yield criteria.
3.5.6
Experimental data showed that for most metals, von Mises yield criteria is more accurate than Tresca criteria. von Mises yield function can be expressed by f = 3J2 k 2 = 0 or if extended to include the kinematic hardening, f= 3 (sij ij )(sij ij ) k 2 = 0 2 (3.193) (3.192)
where k is the scalar internal variable; its initial value is the uniaxial tension strength. ij is the tensor internal variable termed back stress. Similar to sij , ij is also a deviatoric symmetric tensor. Although von Mises model is mainly for the metal plasticity analysis, for undrained analysis in geomechanics, c von Mises model can be approximately used to simulate the undrained behaviors, (Yang and Jeremi, 2002), (Yang and Jeremi, 2003). c The stress derivative of the yield function is f = 3(sij ij ) ij From Equation 3.194, it is easily to derive that f = 3(sij ij ) ij and f = 2k k They will be used in Equation ?? as specic forms of f /qA . If the associated plastic ow rule g = f is assumed, then mij = g = 3(sij ij ) ij (3.197) (3.196) (3.195) (3.194)
(3.198)
(3.199)
Jeremi et al. c
94
It is interesting that from the Equation 3.197, von Mises model gives p = p = mii = 3(sii ii ) = 0 v ii (3.200)
which accords with the phenomena that no plastic volumetric strain occurs for metals. It is implied that the isotropic stress (hydrostatic pressure) can never make the metal yield for this yield criteria. von Mises model is therefore pressure-independent. If k is assumed a linear relation to the equivalent plastic strain p , or by the equation q k = Hs p = Hs eq 2 dev dev m mij 3 ij
0.5
(3.201)
where Hs is the linear hardening/softening modulus to the equivalent plastic strain, the corresponding hA in Equation ?? for the evolution law is then h = Hs 2 dev dev m mij 3 ij
0.5
(3.202)
where mdev is the deviatoric plastic ow, and if it is associated plasticity, ij h = 2Hs k If ij is assumed a linear relation to the plastic strain tensor p , or by the equation ij ij = Ht p = Ht mij ij if it is associated plasticity, ij = 3Ht (sij ij ) for the evolution law is then hij = Ht mij if it is associated plasticity, hij = 3Ht (sij ij ) 1966), ij = 2 ha p cr p ij ij eq 3 (3.208) (3.207) (3.206) (3.205) (3.204) (3.203)
where Ht is the linear hardening/softening modulus to plastic strain tensor, the corresponding hA in Equation ??
A saturation-type kinematic hardening rule is the Armstrong-Frederick hardening (Armstrong and Frederick,
if it is associated plasticity, ij = [ha (sij ij ) 2cr kij ] evolution law is then hij = 2 ha mij cr meq ij 3 (3.210) (3.209)
where ha and cr are material constants. The corresponding hA in Equation ?? for the Armstrong-Frederick
where meq is the equivalent plastic ow, and if it is associated plasticity, hij = 2ha sij 2(ha + cr k)ij
Jeremi et al. c
(3.211)
Version: April 29, 2010, 9:23
95
Yield and Plastic Potential Functions: von Mises Model (form I) Yield function and related derivatives f= 3 [(sij ij ) (sij ij )] k 2 = 0 2 (3.212)
(3.215)
(3.216)
mij = 3im jn ij mn mn
(3.217)
mij =0 k
(3.218)
mij = 3im jn mn Yield and Plastic Potential Functions: von Mises Model (form II) Yield function and related derivatives f = [(sij ij ) (sij ij )]
Jeremi et al. c
0.5
(3.219)
3 k=0 2
(3.220)
Version: April 29, 2010, 9:23
96
f skl 0.5 = (skl kl ) [(smn mn ) (smn mn )] ij ij 1 0.5 = ki lj kl ij (skl kl ) [(smn mn ) (smn mn )] 3 = (sij ij ) [(smn mn ) (smn mn )]
0.5
(3.221)
(3.222)
f = k
3 2
(3.223)
Plastic ow (associated plasticity) and related derivatives mij = f 0.5 = (sij ij ) [(smn mn ) (smn mn )] ij (3.224)
mij = mn
mij =0 k
(3.226)
mij 0.5 1.5 = im jn [(srs rs ) (srs rs )] +(sij ij ) (smn mn ) [(srs rs ) (srs rs )] (3.227) mn
3.5.7
Drucker-Prager Model
Drucker and Prager (1950) proposed a right circle cone to match with the Mohr-Coulomb irregular hexagonal pyramid, which can be expressed by f = I1 + J2 = 0 (3.228)
(3.229)
where and are material constants. By coinciding Drucker-Prager cone with the outer apexes of the Mohr-Coulomb hexagon locus, we get the compressive cone of Drucker-Prager model, with the constants as =
Jeremi et al. c
(3.230)
Version: April 29, 2010, 9:23
97
By coinciding Drucker-Prager cone with the inner apexes of the Mohr-Coulomb hexagon locus, we get the tensile cone of Drucker-Prager model, with the constants as 6 cos 2 sin , = c = 3(3 + sin ) 3(3 + sin ) We can also get the mean cone of the compressive and tensile cone, with the constants as 3 sin 2 3 cos = , = c 9 sin2 9 sin2 Another inner-tangent cone to the the Mohr-Coulomb pyramid, with the constants as = tan 9 + 12 tan
2
(3.231)
(3.232)
, =
3c 9 + 12 tan2
(3.233)
Obviously, in practice and are not directly obtained from experiments. They are functions of MohrCoulomb parameters, the cohesion c and the friction angle , which can be determined by experiments. The shape of Drucker-Prager yield surface has dierent types. They only partially satisfy the above requirements for locus in the plane: they do not coincide with both compressive and tensile experimental points. A useful formulation on Equation 3.228 is f sij = ij + ij 2 J2 For cohesionless sands, k = 0, Drucker-Prager yield function can thus be simplied as f = I1 + J2 = 0 (3.235) (3.234)
or in terms of p and q, f = q Mp = 0 If Equation 3.230 is adopted, then M can be easily derived as M= 6 sin 3 sin (3.237) (3.236)
If the kinematic hardening is taken account, Equation 3.236 can be extended into f= 3 [(sij pij )(sij pij )] M 2 p2 = 0 2 (3.238)
Useful formulations for this yield function are 2 f = 3ij + smn mn + M 2 p ij s ij 3 f = 3pij s ij where sij = sij pij . mij = (3.239)
(3.240)
(3.242)
Version: April 29, 2010, 9:23
98
Yield and Plastic Potential Functions: Drucker-Prager Model (form I) Yield function and related derivatives f= 3 [(sij pij ) (sij pij )] k 2 p2 = 0 2 (3.243)
p 3 smn 3 p f 2 2mn = (smn pmn ) + (smn pmn ) 2k 2 p ij 2 ij 2 ij ij 1 1 2 = 3 mi nj mn ij (smn pmn ) + 3 mn ij (smn pmn ) + k 2 pij 3 3 3 2 = 3 (sij pij ) + mn (smn pmn ) ij + k 2 pij 3
(3.244)
f = 3p (sij pij ) ij
(3.245)
f = 2kp2 k Plastic ow (associated plasticity) and related derivatives mij = f 2 = 3 (sij pij ) + rs (srs prs ) ij + k 2 pij ij 3
(3.246)
(3.247)
mij =3 mn
4 mij = kpij k 3
(3.249)
mij = 3pim jn + rm sn (srs prs ) ij rs prm sn ij mn = 3pim jn + (smn pmn ) ij mn pij = 3pim jn + smn ij 2pmn ij
Jeremi et al. c
(3.250)
Version: April 29, 2010, 9:23
99
Yield and Plastic Potential Functions: Drucker-Prager Model (form II) Yield function and related derivatives f = [(sij pij ) (sij pij )]
0.5
2 kp = 0 3
(3.251)
f = ij =
p smn mn ij ij
0.5
2 p k 3 ij 2 kij 27 (3.252)
f 0.5 = pmi nj (smn pmn ) [(srs prs ) (srs prs )] ij = p (sij pij ) [(srs prs ) (srs prs )]
0.5
(3.253)
f = k
2 p 3
(3.254)
Plastic ow (associated plasticity) and related derivatives mij = 1 f 0.5 = (sij pij ) + pq ij (spq ppq ) [(srs prs ) (srs prs )] + ij 3 2 kij (3.255) 27
mij = mn
1 1 mi nj mn ij + mn ij 3 3
0.5
1 1 1 + pq ij mp nq mn pq + mn pq 3 3 3 1 1 mr ns mn rs + mn rs 3 3
1.5
[(srs prs ) (srs prs )] 1 (sij pij ) + pq ij (spq ppq ) 3 (srs prs ) [(stu ptu ) (stu ptu )]
(3.256)
mij = k
2 ij 27
(3.257)
mij 1 1 0.5 = pmi nj + mp nq ij (spq ppq ) ppq ij mp nq [(srs prs ) (srs prs )] mn 3 3 1 1.5 (sij pij ) + pq ij (spq ppq ) [prm sn (srs prs )] [(stu ptu ) (stu ptu )] 3 (3.258)
Jeremi et al. c Version: April 29, 2010, 9:23
100
Hardening and Softening Functions Isotropic Hardening and related derivatives Linear Isotropic Hardening (Linear Eeq) k = Hmequivalent = H 2 mij mij 3
0.5
(3.259)
2 mmn mmn 3
0.5
(3.260)
2 mmn mmn 3
0.5
(3.261)
2 mmn mmn 3
0.5
(3.262)
Kinematic Hardening and related derivatives Linear Kinematic Hardening (Linear Eij) 1 ij = Hmdev = H mij mkl kl ij ij 3 (3.263)
ij =H mn
1 mkl mij kl ij mn 3 mn
(3.264)
ij =H k
1 mkl mij kl ij k 3 k
(3.265)
ij =H mn
1 mkl mij kl ij mn 3 mn
(3.266)
ij
(3.267)
2 mkl mkl 3
0.5
ij
(3.268)
mij ij = Ht mn mn
Jeremi et al. c
(3.269)
Version: April 29, 2010, 9:23
101
3.5.8
The pioneering research work on the critical state soil mechanics by the researchers in Cambridge University (Roscoe et al., 1958), (Roscoe and Burland, 1968), (Wood, 1990)) has made great contribution on the modern soil elastoplastic models. The original Cam Clay model (Roscoe et al., 1963), and later the modied Cam Clay model (Schoeld and Wroth, 1968) were within the critical state soil mechanics framework. We focus on only the modied Cam Clay model and herein the word modied is omitted to shorten writing.
Critical State
ec = ec,r c ln pc where ec is the critical void ratio at the critical mean eective stress pc c is the normal consolidation slope.
The critical state soil mechanics assumes that the normal consolidation line (NRL) is parallel to the CSL, which is expressed by e = e ln p (3.271)
where e is the intercept on the NRL at p = 1. is the normal consolidation slope or the elastoplastic slope of e ln p relation, and c = . The unloading-reloading line (URL) take the similar form but with dierent slope by e = e ln p (3.272)
where e is the intercept on the URL at p = 1. c is the normal consolidation slope or the elastoplastic slope of e ln p relation. Elasticity The elastic bulk modulus K can be directly derived from the Equation 3.272 and takes the form K= (1 + e)p (3.273)
If a constant Poissons ratio is assumed, since the isotropic elasticity needs only two material constants, the shear elastic modulus can be obtained in terms of K and by G= 3(1 2)(1 + e) 3(1 2) K= p 2(1 + ) 2(1 + ) (3.274)
Alternatively, a constant shear elastic modulus G can be assumed and then the Poissons ratio is expressed in terms of K and G as =
63 In
3K 2G 2(G + 3K)
(3.275)
this chapter, only single-phase (dry phase) is studied, the total and eective stresses are thus identical, e.g. p = pc . c
Jeremi et al. c
102
Yield Function
2 f = q 2 Mc [p(p0 p)] = 0
where Mc is the critical state stress ration in the q p plane, and the p0 is the initial internal scalar variable, The gradient of the yield surface to the stress can be obtained as f q p 1 2 2 = 2q Mc (2p p0 ) = 3sij + Mc (p0 2p)ij ij ij ij 3 where q/ij and p/ij are independent of the yield function. The gradient of the yield surface to p0 will be used in the integration algorithm, and can be expressed by f 2 = Mc p p0 (3.278) (3.277)
Plastic Flow The plastic ow of the Cam Clay model is associated with its yield function, in other words, the plastic ow is dened by the potential function, g, which is assumed the same as the yield function, f .
2 g = f = q 2 Mc [p(p0 p)] = 0
(3.279)
The stress gradient to the yield surface can be obtained as mij = g q p 1 2 2 = 2q + Mc (2p p0 ) = 3sij + Mc (p0 2p)ij ij ij ij 3 (3.280)
It can dene the plastic dilation angle , which is related to the ratio of plastic volumetric and deviatoric strain (Wood, 1990), by tan =
2 p Mc (p0 2p) v p = q 2q
(3.281)
It is interesting to nd that from Equation 3.281, when p < p0 /2, the plastic dilation angle is positive; when p > p0 /2, the plastic dilation angle is negative. If p = p0 /2, the plastic dilation angle is zero, which is corresponding to the critical state. This is evidently more realistic than Drucker-Prager model, whose associated plastic ow always gives positive plastic dilation angle. Evolution Law The evolution law of the Cam Clay model is a scalar one, which can be expressed by p0 = (1 + e)p0 p v (3.282)
With this scalar evolution law, the change of p0 is decided by the change of plastic volumetric strain. When it reaches the critical state, or when there is no plastic volumetric strain, the evolution of p0 will cease. From Equation 3.282, one gets (1 + e)p0 mii p0 = h= (3.283)
or by dilation angle, h=
(3.285)
Version: April 29, 2010, 9:23
Jeremi et al. c
103
Yield and Plastic Potential Functions: Cam-Clay Model Yield function and related derivatives
2 f = q 2 Mc [p(p p0 )] = 0
(3.286)
(3.287)
f 2 = Mc p p0 Plastic ow (associated plasticity) and related derivatives mij = 1 2 f = 3sij + Mc (p0 2p) ij ij 3
(3.288)
(3.289)
(3.290)
mij 1 2 = Mc ij p0 3
(3.291)
Isotropic Hardening and related derivatives (CC Ev) Note the due to the current denition of p (i.e. p = 1 ii ), a minus sign appears in from of the evolution of p0 as follows: 3 p0 = (1 + e) p0 mii (1 + e) p0 2 Mc (2p p0 ) = (3.292) (3.293)
p0 (1 + e) p0 2 = Mc ij
2 ij 3
(3.294)
2 (1 + e) 2 p0 = Mc (p p0 ) p0
Jeremi et al. c
(3.295)
104
3.5.9
Dafalias-Manzari Model
Within the critical state soil mechanics framework, Manzari and Dafalias (1997) proposed a two-surface sand model. This model considered the eects of the state parameter on the behaviors of the dense or loose sands. The features of this model include successfully predicting the softening at the dense state in drained loading, and also softening at the loose state but in the undrained loading. Dafalias and Manzari (2004) later presented an improved version. This version introduced the fabric dilatancy tensor which has a signicant eect on the contraction unloading response. It is also considered the Lodes angle eect on the bounding surface, which produces more realistic responses in non-triaxial conditions. Here only the new version is summarized. The compression stress is assumed negative here, which is dierent from the original reference by Dafalias and Manzari (2004). Critical State Instead of using the most common linear line of critical void ration vs. logarithmic critical mean
eective stress, the power relation recently suggested by Li and Wang (1998) was used: ec = ec,r c pc Pat
(3.296)
where ec is the critical void ratio at the critical man eective stress p , ec,r is the reference critical void ratio, c c and (for most sands, = 0.7) are material constants, and Pat is the atmospheric pressure for normalization. Elasticity The elastic incremental moduli of shear and bulk, are following Richart et al. (1970): G = G0 (2.97 e)2 (1 + e) p Pat
0.5
Pat , K =
where G0 is a material constant, e is the void ratio, and is the Poissons ratio. The isotropic hypoelasticity is then dened by ee = ij p sij , e = v 2G K The yield function is dened by 2 mp = 0 3 (3.299) (3.298)
2(1 + ) G 3(1 2)
(3.297)
where sij is the deviatoric stress tensor, ij is the deviatoric back stress-ratio tensor, m is a material constant,
|| = sij pij = [(sij pij )(sij pij )]0.5 The gradient of the yield surface to the stress can be obtained as f 1 = nij + (pq npq + ij 3 2 m)ij 3
(3.300)
(3.301)
where rij = sij /p is the normalized deviatoric stress tensor, and nij is the unit gradient tensor to the yield surface dened by nij =
Jeremi et al. c
sij pij ||
(3.302)
Version: April 29, 2010, 9:23
105
The tensor of nij is to dened n , the Lodes angle of the yield gradient, by the equation cos 3n = 6nij njk nki where 0 n /6 and n = 0 at triaxial compression and n = /6 at triaxial extension. stress ratio, and Me , the triaxial extension critical stress ratio. M = Mc g(n , c), g(n , c) = Me 2c , c= (1 + c) (1 c) cos 3n Mc (3.305) (3.304)
The critical stress ratio M at any stress state can be interpolated between Mc , the triaxial compression critical
The line from the origin of the plane parallel to nij will intersect the bounding, critical and dilation surfaces
b c d at three image back-stress ratio tensor ij , ij , and ij respectively (Figure 3.25), which are expressed as b ij =
2 b 3
nij
(3.306)
c ij =
nij 2 d 3
(3.307)
d ij =
nij
(3.308)
where = e ec is the state parameter; nb and nd are material constants. Plastic Flow The plastic strain is given by 1 p = Rij = (Rij + Dij ) ij 3 The deviatoric plastic ow tensor is 1 Rij = Bnij + C(nik nkj ij ) 3 where B =1+ 31c g cos 3n , C = 3 2 c 31c g 2 c (3.311) (3.310) (3.309)
2 d ij nij 3
(3.312)
where Ad = A0 (1 + zij nij ) A0 is a material constant, and zij is the fabric dilation tensor. The Macauley brackets if x > 0 and x = 0, if x 0.
Jeremi et al. c Version: April 29, 2010, 9:23
106
Figure 3.25:
Schematic illustration of the yield, critical, dilatancy, and bounding surfaces in the -plane of
deviatoric stress ratio space (after Dafalias and Manzari 2004). Evolution Laws This model has two tensorial evolution internal variable, namely, the back stress-ratio tensor ij and the fabric dilation tensor zij . The evolution law for the back stress-ratio tensor ij is
b 2 ij = [ h(ij ij )] 3
(3.314)
where ij is the initial value of ij at initiation of a new loading process and is updated to the new value when the denominator of Equation 3.315 becomes negative. b0 is expressed by b0 = G0 h0 (1 ch e) p Pat
0.5
(3.316)
where h0 and ch are material constants. The evolution law for the fabric dilation tensor zij is zij = cz D (zmax nij + zij ) where cz and zmax are material constants. Analytical Derivatives for the Implicit Algorithm When implemented into an implicit algorithm for the Dafalias-Manzari model, some complicated additional analytical derivatives are needed. This section gives the analytical derivatives expressions based on the tensor calculus.
Jeremi et al. c Version: April 29, 2010, 9:23
(3.317)
107
Analytical expression of
mij : kl
mij mn
= B
(3.318)
where 1 1 1 1 nij Is ij mn + ij mn nij nmn (ab nab )nij mn = mn || ijmn 3 3 3 D mn and B mn C mn = 3 2 1c c 3 2 1c c g cos 3 cos 3 + g mn mn
d mn
(3.319)
Ad mn
2 d ab nab 3
Ad
d nab 2 ab 3 mn mn
(3.320)
(3.321)
(3.322)
(3.323)
mn g mn cos 3 mn Ad mn
c 3Pat
p Pat
mn
(3.324)
= g2
1c 2c
cos 3 mn
(3.325)
(3.326)
(3.327)
and dene X = 1 if X > 0, and X = 0 if X 0. mij : Analytical expression of kl mij mn nij B nik 1 C + nij +C nkj + (nik nkj ij ) mn mn mn 3 mn 1 D + ij 3 mn
= B
(3.328)
where nij mn
Jeremi et al. c
(3.329)
Version: April 29, 2010, 9:23
108
D mn
Ad mn
2 d ab nab 3
Ad
(3.330) and B mn C mn
d mn
3 2
1c c 3 2 1c c
g cos 3 cos 3 + g mn mn
d mn
(3.331)
(3.332)
= Mc exp (nd ) 1c 2c
g mn
(3.333)
g mn cos 3 mn Ad mn
= g2
cos 3 mn
(3.334)
(3.335)
(3.336)
Analytical expression of
1 D ij 3 zmn
(3.337)
Ad zmn
2 d ab nab 3
(3.338)
(3.339)
Analytical expression of
Aij mn
h 2 3 mn
(3.340) where
b = Mc exp (nb ) mn
g nb g mn mn
(3.341)
(3.342)
Version: April 29, 2010, 9:23
109
Aij mn
2 3 +
where
b g = Mc exp (nb ) mn mn
(3.345)
(3.346)
Aij zmn
(3.347)
Analytical expression of
Zij mn
(3.348)
Analytical expression of
Zij mn
D nij + zmax D mn mn
(3.349)
Analytical expression of
Zij zmn
D zmn
(3.350)
Jeremi et al. c
110
Jeremi et al. c
Chapter 4
Probabilistic ElastoPlasticity
(In collaboration with Dr. Kallol Sett)
(2004)
111
112
Jeremi et al. c
Chapter 5
113
114
Jeremi et al. c
Chapter 6
6.1
6.1.1
In modeling the material nonlinear behavior of solids, plasticity theory is applicable primarily to those bodies that can experience inelastic deformations considerably greater than the elastic deformation. If the resulting total deformation, including both translations and rotations, are small enough, we can apply small deformation theory in solving these problems. If, however strains and rotations are nite, one must resort to the theory of large deformations. In doing so, we will be using two sets of representations1 , namely: Material coordinates in the undeformed conguration, also called Lagrangian coordinates, Spatial coordinates in the deformed conguration, also called Eulerian coordinates. Figure 6.1 shows the displacement of a particle from its initial position XI to the current position xi , dened by the deformation equation: xi = xi (X1 , X2 , X3 , t) The initial position XI of the particle now occupying the position xi is given by the Eulerian equation: (6.1)
(6.2)
Malvern (1969).
115
116
d xi d XI
XI
ui
xi
Figure 6.1: Displacement, stretch and rotation of material vector dXI to new position dxi .
xi = XI + ui
XI = xi ui
(6.3)
6.1.2
Deformation Gradient
The deformation gradients are the gradients of the functions on the righthand side of equations (6.1) and (6.2). To emphasize the dierence between the material, Lagrangian setting and the spatial, Eulerian setting, we will use capital letters for the material coordinate indices and lower case letters for the spatial coordinate indices. We limit our work to the rectangular Cartesian coordinates, thus simplifying the tensor notation to the covariant indices only. The deformation gradient is dened as the twopoint tensor whose rectangular Cartesian components are the partial derivatives:
FkK =
xk = xk,K XK
(6.4)
The deformation gradient FkK transforms (convects) on an arbitrary innitesimal material vector dXI at XI to associate it with a vector dxi at xi : xk dXK = xk,K dXK XK
(6.5)
The the spatial deformation gradients are tensors referred to the deformed, Eulerian conguration: (FKk )
1
XK = XK,k xk
1
Similarly to the deformation gradient FkK , spatial deformation gradient (FKk ) nitesimal material vector dxi at xi to associate it with a vector dXI at XI : dXK = (FKk )
Jeremi et al. c
1
dxk =
(6.7)
Version: April 29, 2010, 9:23
= ik
and
(FIj )
FjK = IK
The Jacobian of the mapping (6.4) can be represented as: J = det (FkK ) = 1 eijk eP QR FiP FjQ FkR 6 (6.9)
The relative deformation gradient fkm is the gradient for the relative motion function:
(6.10)
(6.11)
If the xed reference position XI , the current position xi and the variable position i are all referred to the rectangular Cartesian coordinate system, the chain rule of dierentiation yields:
k xm k = XI xm XI
or
(6.12)
The polar decomposition theorem permits the unique representation2 : Fij = Rik Ukj = vik Rkj (6.13)
where Ukj , vik are positive denite symmetric tensors, called right stretch tensors and left stretch tensors, respectively, and Rkj is an orthogonal tensor such that:
Rik Rjk = ij
and also
Rki Rkj = ij
(6.14)
Equation (6.13), as well as Figure 6.1.2 demonstrate that the motion and deformation of an innitesimal volume element at Xi consist of consecutive applications of: a stretch by Ukj , a rigid body rotation by Rik , a rigid body translation to xi or alternatively: a rigid body translation to xi a rigid body rotation by Rkj , a stretch by vik ,
2 referring
xi and Xi to the same reference axes and using lower case indices for both. This reference to the same coordinate
system will be applied only for the polar decomposition example presented here.
Jeremi et al. c
118
d xi d XI
XI
ui
xi
Figure 6.2: Illustration of the equation Fij = Rik Ukj = vik Rkj .
6.1.3
The strain tensors EIJ and eij are dened so that they give the change in the square length of the material vector dXI . For the Lagrangian formulation we write:
(ds)2 (dS)2 = 2dXI EIJ dXJ and for the Eulerian formulation:
(6.15)
(6.16)
The deformation tensors CIJ and cij are connecting the squared lengths in Lagrangian and Eulerian congurations. The Green deformation tensor3 CIJ , referred to the undeformed conguration, gives the new squared length (ds)2 of the element into which the given element dXI is deformed: (ds)2 = dXI CIJ dXJ The Cauchy deformation tensor cij , sometimes also denoted as4 (bij ) an element dxi identied in the deformed conguration: (dS)2 = dxi cij dxj Substituting equation (6.17) into (6.15) yield: 2EIJ = CIJ IJ
3 Also 4 Another
(6.17)
1
(6.18)
(6.19)
called right CauchyGreen tensor. name for bij is Finger deformation tensor or left CauchyGreen tensor.
Jeremi et al. c
119
and similarly, substituting equation (6.18) into (6.16) we obtain: 2eij = ij cij By using equation (6.5) we can express (ds)2 as: (ds)2 = dxk dxk = (FkI dXI )(FkJ dXJ ) = (xk,I dXI )(xk,J dXJ ) = dXI (FkI FkJ )dXK = dXI CIJ dXK (6.21) (6.20)
so we have obtained the connection between the deformation tensor CIJ and the deformation gradient FkI in the form: CIJ = (FkI FkJ ) = xk,I dXI xk,J dXJ (6.22)
Similarly, by using equation (6.7) and the expression for (dS)2 we can establish the connection between the deformation tensor cij and the deformation gradient FKi as: (dS)2 = dSK dXK = (FKi dxi )(FKj dxj ) = (XK,i dxi )(XK,j dxJ ) = dxi (FKi FKj )dxk = dxi cij dxk cij = (FKi )
1 1 FKj
(6.23)
The expressions for the strain tensors in Lagrangian and Eulerian description5 is obtained from equations (6.19) and (6.20): 1 1 1 1 ij (FKi ) (FKj ) ((FkI FkJ ) IJ ) ; E: eij = 2 2 If one starts from the displacement equation (6.3), referenced to the same axes for both XI and xi L: EIJ = xI = XI + uI ; XI = xI uI 1 ((FKI FKJ ) IJ ) = 2 (6.24)
the general expression for the Lagrangian strain tensor EIJ in terms of displacements is: EIJ =
1 ((KI + uK,I ) (KJ + uK,J ) IJ ) = 2 1 (KI KJ + KI uK,J + uK,I KJ + uK,I uK,J IJ ) = 2 1 (IJ + uI,J + uJ,I + uK,I uK,J IJ ) = 2 1 (uI,J + uJ,I + uK,I uK,J ) 2 Similarly, the general expression for the Eulerian strain tensor eij in terms of displacements is: eij = 1 1 1 ij (Fki ) (Fkj ) = 2
(6.25)
1 (ij (ki uk,i ) (kj uk,j )) = 2 1 (ij ki kj + ki uk,j + uk,i kj uk,i uk,j ) = 2 1 (ij ij + ui,j + uj,i uk,i uk,j ) = 2 1 (ui,j + uj,i uk,i uk,j ) 2
5 Lagrangian
(6.26)
Jeremi et al. c
120
It is worthwhile noting that equations (6.25) and (6.26) represent the complete nite strain tensor. They involve only linear and quadratic terms in the components of displacement gradients. The stretch is a measure of extension of an innitesimal element and is a function of direction of an element, in either deformed or undeformed conguration. By denoting NI a unit vector in the undeformed conguration and ni a unit vector in the deformed conguration, we denote material stretch as (N ) of those elements whit initial direction NI and spatial stretch (n) of those elements with initial direction ni . By dividing equations (6.15) and (6.16) by (ds)2 and (dS)2 respectively and by using: NI = dXI dS and ni = dxi ds (6.27)
we obtain the Cartesian form of stretch in the Lagrangian and Eulerian descriptions: L: 2 ) = (N dXJ dXI CIJ dS dS and E: 2 = (n) dxi dxj cij ds ds (6.28)
General strain tensors can be dened by considering a scale function (Hill, 1978) for the stretch. Scale function is any smooth, monotonic function of stretch f () such that: f () ; [0, ) subject to f (1) = 0, f (1) = 1 (6.29)
Scale function is often taken in the form (2m 1)/2m, where m may have any value. If we choose m to be an integer, the corresponding strain tensor is: EIJ =
2m UIJ IJ 2m
where
(6.30)
Table 6.1 shows dierent Lagrangian strain measures obtained for a particular choice of parameter m. Table 6.1: Dierent Lagrangian strain measures. Strain measure name GreenLagrange Almansi Biot Hencky parameter m 1 -1 1/2 0
m expression for EIJ
(6.31)
Version: April 29, 2010, 9:23
121
6.1.4
The rate of deformation tensor6 describes the tangent motion in terms of velocity components vi = dxi /dt. The spatial coordinates are: vi = vi (x1 , x2 , x3 , t) (6.32)
q Q d xi d Xi P ui
vi+ d vi
p vi d vi
Figure 6.3: Relative velocity dvi of particle Q at point q relative to particle P at point p. In Figure 6.3 the dashed lines represents the trajectories of particles P and Q. The velocity vectors vi at p and vi + dvi at q are tangent to the two trajectories. The relative velocity components dvi of particle at q relative to the particle at p are given by: dvk = vk dxm = vk,m dxm = Lkm dxm xm (6.33)
The spatial gradient of the velocity Lkm can be decomposed as the sum of the symmetric, rate of deformation tensor Dkm , and a skew symmetric spin tensor Wkm as follows: Lkm = where: Dkm = 1 (Lkm + Lmk ) = Dmk 2 and Wkm = 1 (Lkm Lmk ) = Wmk 2 (6.35) 1 1 (Lkm + Lmk ) + (Lkm Lmk ) = Dkm + Wkm 2 2 (6.34)
An alternate way of deriving the rate of deformation tensor goes as follows. The rate of change of squared length (ds) is given as: d (ds) d (ds) =2 ds dt dt since (ds) = dxk dxk it follows: d (dxk ) d (ds) =2 dxk dt dt
6 Also
(6.36)
(6.37)
Jeremi et al. c
122
d =
xk Xm
dt
dt
dXm +
xk d Xm d (dXm ) xk = dXm dt Xm dt
(6.38)
since d (dXm ) /dt 0, because the initial relative position vector dXm does not change with time. By intervk dXm Xm xk dt
dt
dXm =
where
vk =
(6.39)
From equation (6.33) dvk = Lkm dxm and equation (6.39) it follows that: vk dXm = Lkm dxm Xm
2
(6.40)
and then the equation (6.37) becomes: d (dxk ) d (ds) =2 dxk = 2dxk vk,m dxm dxk = 2dxk Lkm dxm dxk = dt dt = 2dxk Dkm dxm dxk + 2dxk Wkm dxm dxk = 2dxk Dkm dxm dxk since dxk dxm dxm dxk and Wkm is skew symmetric such that Wkm = Wmk . Finally we obtain: d (ds) = 2dxk Dkm dxm dt
2 2
(6.41)
(6.42)
and thus it follows that the rate of change of the squared length (ds) of the material instantaneously occupying any innitesimal relative position dxk at point p is determined by the tensor Dkm at point p. In order to compare the strain rate to the rate of deformation, we dierentiate equation (6.15) with respect to time: d (ds)2 (dS)2 d (dXI EIJ dXJ ) =2 = dt dt d (ds)2 d (EIJ ) = 2dXI dXJ = dt dt since (dS)2 and dXI are constant through time. From the equations (6.42) and (6.43) it follows that: d (ds) = 2dxk Dkm dxm d = 2 (dXI FIk ) Dkm (FmJ dXJ ) = 2dXI (FIk Dkm FmJ ) dXJ dt (6.44) and from equations (6.43) and (6.44) it follows that: dEIJ = FIk Dkm FmJ dt or inversely: Dkm = (FIk )
1 2
(6.43)
(6.45)
dEIJ 1 (FmJ ) dt
(6.46)
To obtain the rate of change of the deformation gradient we start from equations (6.4) and dierentiate it with respect to time:
xk d XK dFkK = dt dt
dxk dt XK
Jeremi et al. c
123
(6.48)
6.2
6.2.1
A material is called hyperelastic or Green elastic, if there exists an elastic potential function W , also called the strain energy function per unit volume of the undeformed conguration, which represents a scalar function of strain of deformation tensors, whose derivatives with respect to a strain component determines the corresponding stress component. The most general form of the elastic potential function, is described in equation 6.49, with restriction to pure mechanical theory, by using the axiom of locality and the axiom of entropy production7 : W = W (XK , FkK ) (6.49)
By using the axiom of material frame indierence8 , we conclude that W depends only on XK and CIJ , that is: W = W (XK , CIJ ) or: W = W (XK , cij ) (6.50)
By assuming hyperelastic response, the following are the constitutive equations for the material stress tensors: 2. PiolaKirchho stress tensor: SIJ = 2 W CIJ (6.51)
Marsden and Hughes (1983) pp. 190. Marsden and Hughes (1983) pp. 194.
Jeremi et al. c
124
Kirchho stress tensor ij = 2 W W = 2 FiA (FjB )t = FiA (FjB )t SAB bij CAB (6.54)
Material tangent stiness relation is dened from: dSIJ = 2 where LIJKL = 4 2W CIJ CKL (6.56) 2W 1 dCKL = LIJKL dCKL CIJ CKL 2 (6.55)
mation gradient:
The spatial tangent stiness tensor Eijkl is obtained by the following pushforward operation with the deforEijkl = FiI FjJ (FkK )t (FlL )t LIJKL (6.57)
6.2.2
Isotropic Hyperelasticity
In the case of material isotropy, the strain energy function W (XK , CIJ ) belongs to the class of isotropic, invariant scalar functions. It satises the relation: W (XK , CKL ) = W XK , QKI CIJ (QJL )
t
(6.58)
where QKI is the proper orthogonal transformation. If we choose QKI = RKI , where RKI is the orthogonal rotation transformation, dened by the polar decomposition theorem in equation (6.13), then: W (XK , CKL ) = W (XK , UKL ) = W (XK , vkl ) (6.59)
Right and left stretch tensors, UKL , vkl have the same principal values9 i ; i = 1, 3 so the strain energy function W can be represented in terms of principal stretches, or similarly in terms of principal invariants of deformation tensor: W = W (XK , 1 , 2 , 3 , ) = W (XK , I1 , I2 , I3 ) where: I1 I2 I3
def
(6.60)
def
def
(6.61)
9 Principal
stretches.
Jeremi et al. c
125
Left and right CauchyGreen tensors were dened by equations (6.22) and (6.23), respectively as: CIJ = (FkI )t FkJ ; (c1 )ij = bij = FiK (FjK )t (6.62)
The spectral10 decomposition theorem for symmetric positive denite tensors11 states that: CIJ = 2 NI A
(A)
NJ
(A) A
where
A = 1, 3
(6.63)
and NI are the eigenvectors12 of CIJ . Values 2 are the roots of the characteristic polynomial A P (2 ) = 6 + I1 6 I2 4 + I3 = 0 A A A A It should be noted that no summation is implied over indices in parenthesis13 . The mapping of the eigenvectors can be deduced from equation (6.5) and is given by (A) ni where ni
(A) (A) def
(6.64)
= FiJ NJ
(A)
(6.65)
FiJ = A ni NJ
3 (A)
(6.66)
RiJ =
A=1
ni NJ
(A)
(6.67)
bij = 2 ni nj A
(A) (A) A
(6.68)
Spectral decomposition from equation (6.63) is valid for the case of nonequal principal stretches, i.e. 1 = 2 = 3 . If two or all three principal stretches are equal, we shall introduce a small perturbation to the numerical values for principal stretches in order to make them distinct. The case of two or all three values of principal stretches being equal is theoretically possible and results for example from standard triaxial tests or isotropic compression tests. However, we are never certain about equivalence of two numerical numbers, because of the nite precision arithmetics involved in calculation of these numbers. From the numerical point of view, two number are equal if the dierence between them is smaller than the machine precision (macheps) specic to the computer platform on which computations are performed. Our perturbation will be a function of the macheps. The characteristic polynomial P (2 ) from equation (6.64) can be solved14 for A : A 1 A = 3 where = arccos
10 See
2 I1 + 2 I1 3I2 cos
+ 2A 3
(6.69)
(6.70)
CIJ
in this work, we shall make an eort to represent all the tensorial equations in indicial form. 14 See also Morman (1986) and Schellekens and Schellekens and Parisch (1994).
example, in the present case NI is the Ath eigenvector with members N1 , N2 and N3 , so that the actual equation P (A) (A) 2 N (A) N (A) can also be written as CIJ = A=3 2 NI NJ . In order to follow the consistency of indicial notation = A A=1 (A) I J
A
(A)
(A)
(A)
(A)
Jeremi et al. c
126
Recently, Ting (1985) and Morman (1986) have used Serrins representation theorem in order to devise a
m useful representation for generalized strain tensors15 EIJ and eij through CIJ and bm . Morman (1986) has ij
shown that bm can be stated as ij 2 (b2 )ij (I1 2 (A) bij + I3 (A) ij bm = 2m ij A 24 I1 2 + I3 2 (A) (A) (A)
2 (b2 )ij I1 2 (A) bij + I3 (A) ij
(6.71)
(A) (A)
By comparing equations (6.71) and (6.68) it follows that the Eulerian eigendiad ni nj
(A) (A) ni nj
2 (A)
(6.73)
where it was used that: CIJ = (FiI )1 (b2 )ij (FjJ )t IJ = (FiI )1 bij (FjJ )t (C 1 )IJ = (FiI )1 ij (FjJ )t It should be noted that the denominator in equations (6.72) and (6.73) can be written as: 24 I1 2 + I3 2 = 2 2 (A) (A) (A) (B) (A) 2 2 (A) (C)
def
= D(A)
(6.77)
where indices A, B, C are cyclic permutations of 1, 2, 3. It follows directly from the denition of D(A) in equation and (6.68) we can obtain: (6.77) that 1 = 2 = 3 D(A) = 0 for equations (6.72) and (6.73) to be valid. Similarly to equations (6.63) (C 1 )IJ = 2 NI A
(A)
NJ
(A) A
(6.78) (6.79)
(b1 )ij = 2 ni nj A
(A) (A) A
6.2.3
It proves useful to separate deformation in volumetric and isochoric parts by a multiplicative split of a deformation gradient as FiI = Fi
vol
FI
where
1 Fi = FiI J 3
vol
FI = J 3 I
(6.80)
where x represents an intermediate conguration such that deformation XI x is purely volumetric and x xi is purely isochoric. It also follows from equation (6.80) that FI and FiI have the same eigenvectors.
15 Dened
Jeremi et al. c
127
x
vol
FI FI i
iso
F i xi
XI
Figure 6.4: Volumetric isochoric decomposition of deformation. The isochoric part of the Green deformation tensor CIJ , dened in equation (6.63) can be dened as
2 (A) (A) CIJ = J 3 CIJ = 2 NI NJ A
(6.81)
while the isochoric part of the Finger deformation tensor bij can be dened similarly as
2 ij = J 3 bij = 2 n(A) n(A) b A i j
(6.82)
(6.83)
The free energy W is then decomposed additively as: W XK , (A) = isoW XK , (A) + volW (XK , J) (6.84)
6.2.4
SimoSerrins Formulation
In Section (6.2.2) we have presented the most general form of the isotropic strain energy function W in terms of of principal stretches: W = W (XK , 1 , 2 , 3 , ) (6.85)
It was also shown in Section (6.2.1) that it is necessary to calculate the gradient W /CIJ in order to obtain 2. PiolaKirchho stress tensor SIJ and accordingly other stress measures. Likewise, it was shown that the material tangent stiness tensor LIJKL (as well as the spatial tangent stiness tensor Eijkl ) requires second order second order tensor MIJ
16 See
derivatives of strain energy function 2 W /(CIJ CKL ). In order to obtain these quantities we introduce16 a
(A)
Runesson (1996).
Jeremi et al. c
128
MIJ
(A)
def
(A)
NJ
(A)
(6.86) (FjJ )t
from (6.73)
= =
ni nj
(A) (A)
where D(A) was dened by equation (6.77). With MIJ dened by equation (6.86), we get from equation (6.63) that: CIJ = 4 MIJ A
(A) A
(6.87)
and also from equation (6.78) it follows that: (C 1 )IJ = MIJ + MIJ + MIJ It can also be concluded that: IJ = 2 MIJ + 2 MIJ + 2 MIJ = 2 MIJ A (1) (2) (3) since, from the orthogonal properties of eigenvectors
3 (1) (2) (3) (A) A (1) (2) (3)
(6.88)
(6.89)
IJ =
A=1
NI
(A)
NJ
(A)
= NI
(A) A
NJ
(A) A
(6.90)
We are now in a position to dene the SimoSerrin fourth order tensor MIJKL as: MIJKL
(A) def
MIJ = CKL 1 (A) (A) IIJKL KL IJ + 2 (A) IJ MKL + MIJ KL + D(A) I3 2 (C 1 )IJ (C 1 )KL + (A) 2 I3 (A)
(A)
(A)
(6.91)
6.2.5
Stress Measures
In Section (6.2.1) we have dened various stress measures in terms of derivatives of the free energy function W . With the free energy function decomposition, as dened in equation (6.84) we can appropriately decompose all the previously dened stress measures: 2. PiolaKirchho stress tensor: SIJ = W isoW volW =2 +2 CIJ CIJ CIJ iso vol = SIJ + SIJ 2
(6.92)
Jeremi et al. c
129
TIJ + volTIJ
PiJ + volPiJ
Kirchho stress tensor ab = 2 isoW volW W = FaI (FbJ )t SIJ = 2FaI (FbJ )t + 2FaI (FbJ )t eij CIJ CIJ ab + volab (6.95)
The derivative of the volumetric part of the free energy function is volW (J) volW (J) J 1 volW (J) = = J (C 1 )IJ CIJ J CIJ 2 J (6.96)
where equation (D.9) was used, while the derivative of the isochoric part of the free energy function yields isoW ((A) ) isoW ((A) ) (A) 1 isoW ((A) ) 1 (A) (A) = = (A) (MIJ )A = wA (MIJ )A CIJ (A) CIJ 2 (A) 2 (6.97) where equation (D.7) was used and wA is derived in Appendix D.4 as: wA = isoW ((A) ) B isoW ((A) ) 1 isoW ((A) ) (A) = (A) B + (A) 3 B B (A) (6.98)
(6.99)
The derivative of the free energy is then: W ((A) ) CIJ = = volW ((A) ) isoW ((A) ) + CIJ CIJ 1 volW (J) 1 (A) J (C 1 )IJ + wA (MIJ )A 2 J 2
(6.100)
It is obvious that the only material dependent parts are derivatives in the form volW /J and wA , while the rest is independent of which hyperelastic material model we choose.
Jeremi et al. c Version: April 29, 2010, 9:23
130
6.2.6
The free energy function decomposition (6.84) is used together with the appropriate denitions made in section (6.2.1) toward the tangent stiness operator decomposition LIJKL = vol LIJKL + iso LIJKL = 4 The volumetric part 2
vol
(6.101)
1 4
J2
2 volW volW +J JJ J
and the complete derivation is again given in appendix D.2. The isochoric part 2
iso
2isoW ((A) ) 1 1 (B) (A) (A) = YAB (MKL )B (MIJ )A + wA (MIJKL )A CIJ CKL 4 2 and the complete derivation is given in the appendix (D.3). Finally, one can write the volumetric and isochoric parts of the tangent stiness tensors as:
vol
LIJKL =
J2
volW (J) 1 volW (J) (C 1 ) 2volW (J) 1 (C )KL (C 1 )IJ + J (C )KL (C 1 )IJ + 2J IIJKL JJ J J (6.104)
(B)
(A)
(A)
(6.105)
In a similar manner to the stress denitions it is clear that the only material model dependent parts are YAB material model. This observation has a practical consequence in that it is possible to create a template derivations and wA . The remaining second and fourth order tensors MIJ and MIJKL are independent of the choice of the
(A) (A)
for various hyperelastic isotropic material models. Only rst and second derivatives of strain energy function with respect to isochoric principal stretches (A ) and Jacobian (J) are needed in addition to the independent tensors, for the determination of various stress and tangent stiness tensors.
6.2.7
The strain energy function for isotropic solid in terms of principal stretches is represented as: W = W (1 , 2 , 3 )
Jeremi et al. c
(6.106)
Version: April 29, 2010, 9:23
131
The only restriction is that W is a symmetric function of 1 , 2 , 3 , although an appropriate natural conguration condition requires that: W (1, 1, 1) = 0 and W (1, 1, 1) =0 i (6.107)
The strain energy function W can either be regarded as a function of principal stretches or the principal invariants of stretches17 : I1 I2 I3 = 2 + 2 + 2 1 2 3 = 2 2 + 2 2 + 2 2 1 2 3 1 2 3 = 2 2 2 1 2 3 (6.108)
A slightly more general formulation is obtained by using principal stretches in the strain energy function denition. A widely exploited family of compressible hyperelastic models18 are dened (Ogden, 1984) as an innite series in powers of (I1 3), (I2 3) and (I3 1) as: W =
N p,q,r=0
(6.109)
The regularity condition that W is continuously dierentiable an innitely number of times is satised. The requirement that energy vanishes in the reference conguration is met provided c000 = 0. Reference conguration is stress free i c100 + 2c010 + c001 = 0. Isochoric deviatoric decoupling is possible by setting cpqr = 0 (r = 1, 2, 3, ...) and cpqr = 0 (p, q = 1, 2, 3, ..) to obtain: W = isoW + volW where:
iso
(6.110)
W =
N p,q=0
vol
W =
N r=0
(6.111)
In what follows, we will present a number of widely used strain energy functions for isotropic elastic solids. Ogden Model A very general set of hyperelastic models was dened by Ogden (1984). The strain energy is expressed as a function of principal stretches as: W =
N r=1
cr (r + r + r 3) 3 2 1 r
(6.112)
W =
N r=1
cr r
r + r + r 1 3 2 1
(6.113)
17 See
18 Used
Jeremi et al. c
132
Derivatives needed for building tensors wA and YAB are given by the following formulae: isoW = A 2
iso N r=1 N r=1 r 2
cr
r 1
(6.114)
2 A
cr (r 1)
(6.115)
(6.116)
The general isotropic hyperelastic model dened in terms of invariants of principal stretches contains the Neo Hookean model as special cases. The isochoric part of NeoHookean isotropic elastic model can be obtained by selecting N = 1, q = 0, cp00 = G/2, to get:
iso
W =
G 2 2 2 1 + 2 + 3 3 2 Kb 2 2 2 1 2 3 1 2
2
(6.117)
while the volumetric part can be dened by choosing N = 2, c001 = 0, c002 = Kb /2, as:
vol
W =
Kb 2 J 1 2
(6.118)
where G and Kb are the shear and bulk moduli respectively. Derivatives needed for building tensors wA and YAB are given by the following formulae: isoW = G A A 2 W =G 2 A
iso
(6.119)
(6.120)
2 isoW =0 A B MooneyRivlin Model Mooney proposed a strain energy function for isochoric behavior of the form:
iso
(6.121)
= =
N n=0 N n=0
2n 2n 2n an 2n + 2n + 2n 3 + an 1 + 2 + 3 3 1 2 3 an 2n + 2n + 2n 3 1 2 3 2 3
2n
+ an
+ 3 1
2n
+ 1 2
2n
(6.122)
with an and bn being the material parameters and a volume preserving constrain a = 1/(b c , and a, b, c are cyclic permutations of (1, 2, 3). A more general form was proposed by Rivlin:
iso N p,q=0
W =
(6.123)
Jeremi et al. c
133
which is actually quite similar to the isochoric part of the general isotropic representation from equation (6.111). Both Mooney and Rivlin strain energy functions become similar, if one chooses to set N = 1 and c10 = C1 and c01 = C2 to obtain:
iso
= =
C1 2 + 2 + 2 3 + C2 2 + 2 + 2 3 1 2 3 1 2 3 C1 I1 3 + C2 I2 3 (6.124)
Derivatives needed for building tensors wA and YAB are given by the following formulae: isoW = 2 C1 A 2 C2 3 A A 2 W = 2 C1 + 6 C2 4 A 2 A
iso
(6.125)
(6.126)
2 isoW =0 A B Logarithmic Model By choosing an alternative set of isochoric principal stretch invariants in the form: ln I1 = = 2 ln 1 ln 1
2 2
(6.127)
+ 2 ln 2
2
+ 2 ln 3
2
+ ln 2
2
+ ln 3
2
ln I2
= =
4 ln 2 ln 2
2
ln 3
2
+ 4 ln 3
2
2 2
ln 1
2 2
+ 4 ln 1 ln 2
2
ln 2
ln 3
+ ln 3
ln 1
+ ln 1
(6.128)
(6.129)
ln ln The general representation of the isochoric part of the strain energy function in terms of I1 and I2 was proposed by Simo and Miehe (1992). A somewhat simpler isochoric strain energy function can be presented in the form:
iso 2 2 2
W =G
ln 1
+ ln 2
+ ln 3
(6.130)
W =
Kb 2 (ln J) 2
(6.131)
Derivatives needed for building tensors wA and YAB are given by the following formulae: isoW =2G A 2
iso
(6.132)
2
2 A
Jeremi et al. c
= 2 G
(6.133)
Version: April 29, 2010, 9:23
134
2 isoW =0 A B d
vol
(6.134)
W = Kb J 1 ln J dJ
vol
(6.135)
d2
dJ 2
= Kb J 2 Kb J 2 ln J
(6.136)
SimoPister Model Another form or a volumetric part of strain energy function was proposed by Simo and Pister (1984) in the form: Wvol (J) = 1 Kb J 2 1 2 ln J 4 (6.137)
The rst and second derivatives with respect to J are then given as:
2 + 2 J Kb dvolW (J) = J dJ 4 2vol 2 + J22 Kb d W (J) = dJ 2 4
(6.138) (6.139)
6.3
6.3.1
The mathematical structure and numerical analysis of classical small deformation elastoplasticity is generally well established. However, development of large deformation elasticplastic algorithms for isotropic and anisotropic material models is still a research area. Here, we present a new integration algorithm, based on the multiplicative decomposition of the deformation gradient into elastic and plastic parts. The algorithm is novel in that it is designed to be used with isotropic as well as anisotropic material models. Consistent derivation is based on the c idea from the innitesimal strain algorithm developed earlier by Jeremi and Sture (1997). The algorithm is not an extension of earlier developments, but rather a novel development which consistently utilizes Newtons method for numerical solution scheme for integrating pertinent constitutive equations. It is also shown that in the limit, the proposed algorithm reduces to the small strain counterpart. In what follows, we briey introduce the multiplicative decomposition of the deformation gradient and pertinent constitutive relations. We then proceed to present the numerical algorithm and the algorithmic tangent stiness tensor consistent with the presented algorithm.
6.3.2
Kinematics
An appropriate generalization of the additive strain decomposition is the multiplicative decomposition of displacement gradient. The motivation for the multiplicative decomposition can be traced back to the early works of Bilby et al. (1957), and Krner (1960) on micromechanics of crystal dislocations and application to continuum o modeling. In the context of large deformation elastoplastic computations, the work by Lee and Liu (1967), Fox (1968) and Lee (1969) stirred an early interest in multiplicative decomposition.
Jeremi et al. c Version: April 29, 2010, 9:23
135
The appropriateness of multiplicative decomposition technique for soils may be justied from the particulate nature of the material. From the micromechanical point of view, plastic deformation in soils arises from slipping, crushing, yielding and plastic bending19 of granules comprising the assembly20 . It can certainly be argued that deformations in soils are predominantly plastic, however, reversible deformations could develop from the elasticity of soil grains, and could be relatively large when particles are locked in high density specimens.
dx
Current Configuration
Fe
F e
-1
F p
X x
0
dX
X
dx
Figure 6.5: Multiplicative decomposition of deformation gradient: schematics. The reasoning behind multiplicative decomposition is a rather simple one. If an innitesimal neighborhood of a body xi , xi + dxi in an inelastically deformed body is cutout and unloaded to an unstressed conguration, it would be mapped into xi , xi + di . The transformation would be comprised of a rigid body displacement21 x and purely elastic unloading. The elastic unloading is a ctitious one, since materials with a strong Baushingers eect, unloading will lead to loading in an other stress direction, and, if there are residual stresses, the body
plate like clay particles. also Lambe and Whitman (1979) and Sture (1993). 21 Translation and rotation.
20 See 19 For
Jeremi et al. c
136
must be cutout in small pieces and then every piece relieved of stresses. The unstressed conguration is thus incompatible and discontinuous. The position xi is arbitrary, and we may assume a linear relationship between dxi and di , in the form22 : x
dxi
(6.140)
is not to be understood as a deformation gradient, since it may represent the incompatible, dis-
continuous deformation of a body. By considering the reference conguration of a body dXi , then the connection to the current conguration is23 :
Fij dXj
(6.141)
(6.142)
p The plastic part of the deformation gradient, Fkj represents micromechanically, the irreversible process of
slipping, crushing dislocation and macroscopically the irreversible plastic deformation of a body. The elastic part,
e Fki represents micromechanically a pure elastic reversal of deformation for the particulate assembly, macroscop-
ically a linear elastic unloading toward a stress free state of the body, not necessarily a compatible, continuous deformation but rather a ctitious elastic unloading of small cut outs of a deformed particulate assembly or continuum body.
6.3.3
Constitutive Relations
We propose the free energy density W , which is dened in , as follows e e 0 W (Cij , ) = 0 W e (Cij ) + 0 W p ( ) (6.143)
e e where W e (Cij ) represents a suitable hyperelastic model in terms of the elastic right deformation tensor Cij , whereas W p ( ) represents the hardening. It has been shown elsewhere (Runesson, 1996), that the pertinent dissipation inequality becomes D = Tij Lp + ij
K 0
(6.144)
where Tij is the Mandel stress24 and Lp is the plastic velocity gradient dened on . ij We now dene elastic domain B as B = {Tij , K | (Tij , K ) 0} (6.145)
When is isotropic in Tij (which is the case here) in conjunction with elastic isotropy, we can conclude that Tij is symmetrical and we may replace Tij by ij in .
22 referred 23 See
Jeremi et al. c
137
As to the choice of elastic law, it is emphasized that this is largely a matter of convenience since we shall be dealing with small elastic deformations. Here, the NeoHookean elastic law is adopted. The generic situation e is Tij = Tij (Ukl , J e ), where we have used the isochoric/volumetric split of the elastic right stretch tensor as U e = U e (J e )1/3 .
kl kl
The constitutive relations can now be written as e Tij = Tij (Ukl , J e ) p Lp := Fik ij
p Fjk 1
K = K ( ) = K , (0) = 0
(6.149)
p e where Fik = (Fli )1 Flk is the plastic part of the deformation gradient.
6.3.4
(6.149): p Fik
The incremental deformation and plastic ow are governed by the system of evolution equations (6.147) and
1
p Fjk
= Mij
(6.150)
(0) = 0
(6.151)
(6.152)
By using the multiplicative decomposition e p e Fij = Fik Fkj Fik = Fij and equation (6.152) we obtain
n+1 e Fij p Fkj 1
(6.153)
= =
n+1
p Fim (nFmk )
exp n+1Mkj
(6.155)
= =
Jeremi et al. c
exp n+1Mlj
138
By recognizing that the exponent of a tensor can be expanded in Taylor series25 exp n+1Mlj = lj n+1Mlj + 1 n+1Mls 2 n+1Msj + (6.157)
and by using the rst order expansion in the equation (6.156), we obtain
n+1 e Cij
= = =
ir n+1Mir
n+1 e,tr Cil n+1 e,tr Cij
n+1Mir Mir
n+1Mir
Mlj (6.158)
+2
n+1
Remark 6.3.1 The Taylors series expansion from equation (6.157) is a proper approximation for the general nonsymmetric tensor Mlj . That is, the approximate solution given by equation (6.158) is valid for a general anisotropic solid. This contrasts with the spectral decomposition family of solutions26 which are restricted to isotropic solids. Remark 6.3.2 Taylors series expansion27 is proper for small values of plastic ow tensor n+1Mlj . This is indeed the case for small increments, when 0 which are required for following the equilibrium path for pathdependent solids. Remark 6.3.3 In the limit, when the displacements are suciently small, the solution (6.158) collapses to
Fij ij
lim ij + 2n+1ij =
Mlj
n+1
Mir rl + 2n+1e,tr rl
Mlj
Mlj Mir
n+1 tr n+1 rl Mlj
Mil
n+1
Mlj + 22
n+1
n+1Mij
(6.159)
which is a small deformation elastic predictorplastic corrector equation in strain space. In working out the small deformation counterpart (6.159) it was used that
Fij ij
lim
n+1 e Cij
= ij + 2n+1ij
n+1
2 n+1tr il
Mlj
n+1
Mij (6.160)
25 See 26 See
for example Pearson (1974). Simo (1992). 27 It should be called MacLaurins series expansion, since expansion is about zero plastic ow state (no incremental plastic deformation).
Jeremi et al. c
139
By neglecting the higher order term with 2 in equation (6.158), the solution for the right elastic deformation tensor n+1C e can be written as
ij n+1 e Cij
e,tr = n+1Cij
n+1
Mir
e,tr + n+1Cil
n+1
Mlj
(6.161)
= n +
(6.162)
n+1
Remark 6.3.4 It is interesting to note that equation (6.161) resembles the elastic predictorplastic corrector equation for small deformation elasticplastic incremental analysis. That resemblance will be used to build an iterative solution algorithm in the next section. The incremental problem is dened by equations (6.161), (6.162), and the constitutive relations
n+1 SIJ
=2
W CIJ W
(6.163)
n+1
n+1
K =
(6.164)
n+1
n+1 = 0
(6.165)
Remark 6.3.5 The Mandel stress tensor Tij can be obtained from the second PiolaKirchho stress tensor Skj and the right elastic deformation tensor C e as
ik
(6.167)
This set of nonlinear equations will be solved with a Newton type procedure, described in the next section. e,tr For a given n+1Fij , or n+1Cij , the upgraded quantities n+1SIJ and n+1K can be found, then the appropriate pullback to B0 or pushforward28 to B will give
n+1 n+1
SIJ and
n+1
ij (6.168)
SIJ =
n+1 p FjJ
n+1
e ij = n+1FiI
n+1 SIJ
n+1 e 1 FjJ
(6.169)
= =
n+1
Mir
e,tr + n+1Cil
n+1
Mlj (6.170)
n+1Zij
28 See
Appendix ??.
Jeremi et al. c
140
is used as a starting point for a Newton iterative algorithm. In previous equation, we have introduced tensor Zij to shorten writing. The trial right elastic deformation tensor is dened as
p FrM (nFiM )
FrS
n p FjS
(6.171)
e Rij = Cij
current
n+1Zij
(6.172)
BackwardEuler
Tensor Rij represents the dierence between the current right elastic deformation tensor and the Backward Euler right elastic deformation tensor. The trial right elastic deformation tensor n+1C e,tr is maintained xed during the
ij
iteration process. The rst order Taylor series expansion can be applied to the equation (6.172) in order to obtain
new old the iterative change, the new residual Rij from the old Rij
n+1Zij n+1Zij new old e dK dTmn + Rij = Rij + dCij + d() n+1Zij + K Tmn
(6.173)
By using that
e Csk
Tsn = Skn
(6.174)
we can write
dTmn
e e = dCmk Skn + Cmk dSkn 1 e e e from (6.55) = dCmk Skn + Cmk Le knpq dCpq 2 e knpq dCpq e e 1 Tsn + 1 Cmk Le e from (6.174) = dCmk Csk 2
(6.175)
Version: April 29, 2010, 9:23
Jeremi et al. c
141
+ +
= + + +
n+1Zij e e 1 Tsn + dCmk Csk Tmn n+1Zij e e 1 e Cmk Lknpq dCpq + 2 Tmn n+1Zij dK K
= + + +
n+1Zmn Tik
e Csj
e Tsk dCij +
(6.176)
n+1Zmn e 1 e Tsk dCij + Csj Tik n+1Zpq e e 1 e Cmk Lknij dCij + + 2 Tmn n+1Zij dK + K + =
old Rij + d() n+1Zij
e Tsk dCij +
e knij dCij e Cmk Le n+1Zij dK + K n+1 e 1 Tsk + 1 Zmn C e Le ) dC e Csj pk kqij ij 2 Tpq (6.177)
+ (im nj +
n+1Zmn Tik
Jeremi et al. c
142
(6.178)
(6.179)
or, by rearranging indices e dCpq = (Tmnpq ) By using that dK = K K Q Q d = d() = d() H K K Q n+1Zmn d() H K K (6.181)
1 old Rmn d() n+1Zmn
(6.180)
(6.182)
(Tij , K ) = + + = + + = + +
(Tij , K ) + (Tij , K ) dTmn Tmn (Tij , K ) dK K old (Tij , K ) + ij , K ) (T e dCmk Tmn (Tij , K ) dK K old (Tij , K ) + ij , K ) (T e 1 Csq Tpn
old
e Csk
e Tsn dCpq
1 (Tij , K ) e e e Cmk Lknpq dCpq 2 Tmn (Tij , K ) + dK K old = (Tij , K ) + (Tij , K ) e 1 1 (Tij , K ) e e + Tsn + Cmk Lknpq Csq 2 Tpn Tmn (Tij , K ) + dK K By using (6.181), equation (6.183) becomes
new
e dCpq (6.183)
(Tij , K ) = +
old
e dCpq (6.184)
d()
Jeremi et al. c
(Tij , K ) H K K
143
(6.185)
e and with the solution for dCpq from (6.182), (6.184) becomes
new
n+1Zmn H K K (6.186)
(Tij , K ) H K K
d() =
1 old Fpq (Tmnpq ) Rmn n+1 Zmn 1 n+1Z H + H mn Fpq (Tmnpq ) K K K K old 1
Fpq (Tmnpq )
(6.187)
Remark 6.3.6 In the perfectly plastic case, the increment inconsistency parameter d() is
old
d() =
Fpq (Tmnpq )
Fpq (Tmnpq )
old Rmn
1 n+1 Zmn
(6.188)
Remark 6.3.7 In the limit, for small deformations, isotropic response, the increment inconsistency parameter d() becomes
old
d() =
(nmn Emnpq )
pm nq +
mmn Eijpq ij
n+1
old Rmn
nmn Emnpq
mpq mp qn + Eijmn ij
mmn + H K K
(6.189)
since in the limit, as deformations are getting small Tmnpq Fpq Zpq Rpq pm nq + mmn Eijpq ij
Upon noting that residual Rpq is dened in strain space, the increment inconsistency parameter d() compares c exactly with its small strain counterpart (Jeremi and Sture, 1997). The procedure described below summarizes the implementation of the return algorithm.
Jeremi et al. c Version: April 29, 2010, 9:23
144
e Given the right elastic deformations tensor nCpq and a set of hardening variables nK at a specic
n+1
fij = ij + ui,j =
n+1 e fir nFrk
fir
n+1
e fkl nFlj
(6.192)
Compute the trial elastic second PiolaKirchho stress and the trial elastic Mandel stress tensor
n+1 e,tr Sij n+1 e,tr Tij
(6.193) (6.194)
n+1 tr
e,tr (Tij , K ). If
= = =
K Tij
n+1
n e,tr Tij
and exit constitutive integration procedure. Return Algorithm If yield criteria has been violated (n+1tr > 0) proceed to step 1. step 1. k th iteration. Known variables
n+1 e(k) Cij
n+1 (k)
n+1
(k) K
n+1
(k)
= =
(n+1Tij
e(k) n+1
(k) K )
n+1(k)n+1Zij
(k)
(k)
= = = =
K Tij
n+1 n+1
(k)
Exit constitutive integration procedure. step 3.29 If convergence is not achieved, i.e. (k) > N T OL or Rij stiness tensor Lijkl (k) Lijkl = 4
29 From
(k)
2W Ckl
e(k)
e(k) Cij
(6.195)
step 3. to step 9. all of the variables are in intermediate n + 1 conguration. For the sake of brevity we are omitting
superscript n + 1.
Jeremi et al. c
145
step 4. Compute the incremental inconsistency parameter d((k+1) ) d((k+1) ) = (k) Fmn where ,(k) H (k) = H (k) K (Tij , K ) = Tpn
(k) (k)
(k) Fmn Rmn (k) Zmn (k) (k) (k) (k) H + H Zmn (k) (k) Fmn (k) K K
(k)
(6.196)
Tmnpq (k)
Fpq
e,(k) Csq
(k)
(k) (k) (Tij , K ) e,(k) e,(k) (k) + 1 Tsn Cmk Lknpq 2 Tmn
1
Tmnij = im nj + (k)
Zmn (k) T
ik
e,(k) Csj
(k)
step 5. Updated the inconsistency parameter (k+1) (k+1) = (k) + d((k+1) ) step 6. Updated the right deformation tensor, the hardening variable and the Mandel stress (6.197)
e,(k+1) = dCpq
(k) Tmnpq 1 (k) (k) Rmn d((k+1) ) n+1Zmn + (k)
(k)
(k+1) d = d((k+1) )
,(k) K
(k)
(6.199)
(k+1) dK = d((k+1) ) H
,(k) K
1
(6.200)
e,(k) Csk
(6.201)
(k+1) e,(k+1) , hardening variable K (k+1) step 7. Update right deformation tensor Cpq and Mandel stress Tmn
e,(k+1) Cpq
(k+1) (k+1) K
(k+1) Tmn
(6.202)
step 8. evaluate the yield function and the residual (k+1) = (Tij
Jeremi et al. c
e(k+1) (k+1) , K ) ; Rij (k+1)
= Cij
e,(k+1)
(6.203)
Version: April 29, 2010, 9:23
146
(k+1)
e,(k+1) = Cpq
(k+1) = (k+1) = K
(k) Tmn
(k+1) = Tmn
(6.204)
6.3.5
(6.205)
and taking the rst order Taylor series expansion we obtain e dCij e,tr = dCij d() Zij e,tr = dCij d() Zij Zij e e dCmk Csk Tmn Zij dK K Zij Zij dK dTmn K Tmn
1
from (6.175)
(6.206)
Zij 1 e e e e C L dCpq Tsn dCmk + d() Tmn 2 mk knpq Zij e,tr = dCij d() Zij + d() H K K
(6.207)
or as e e,tr dCij (Tmnij ) = dCij d() Zij + d() where Tmnij = im nj + (k) Zmn (k) T
ik (k)
Zij H K K
(6.208)
e,(k) Csj
(k)
The solution for the increment in right elastic deformation tensor is then e dCij = (Tmnij )
Jeremi et al. c
1
Zij H K K
(6.209)
Version: April 29, 2010, 9:23
147
We next use the rst order Taylor series expansion of yield function d(Tij , K ) = 0 mn dTmn + K dK T 1 e e 1 e e e Tsn + Cmk Lknpq dCpq + dCmk Csk dK 2 K Tmn e e 1 Tsn dC e + 1 C e Le dK dCpq + Csq pq 2 Tmn mk knpq K Tpn e e 1 Tsn + 1 C e Le d() H dCpq Csq mk knpq 2 Tmn K K e d() H Fpq dCpq K K = = = = = 0 (6.210) where Fpq = Tpn e Csq
1
Tpn
(6.211)
e By using solution for dCij from 6.209 we can write Fpq (Tmnpq )
1
Zmn H K K d() H K K
0 (6.212)
We are now in the position to solve for the incremental inconsistency parameter d() d() = Fpq (Tmnpq )
1
e,tr dCmn
(6.213)
n+1Zmn H + H K K K K (6.214)
and by using 6.209 we can write e dCpq = (Tmnpq ) Then e e,tr dCpq = Ppqvt dCvt
Jeremi et al. c
1
mv nt
1
Fop (Trsop )
rv st
Fop (Trsop )
rv st Zij H K K
(6.217)
Version: April 29, 2010, 9:23
148
where Ppqvt = Fop (Trsop ) rv st Zmn 1 Fop (Trsop ) rv st Zij + H K K n+1Zmn Fab (Tvtab )1 Zmn H = (Tmnpq )1 mv nt K K (Tmnpq )
1
mv nt
(6.218) Algorithmic tangent stiness tensor Lijkl (in intermediate conguration ) is then dened as LAT S = Le knvt knpq Ppqvt conguration 0
n+1 AT S Lijkl
(6.219)
Pullback to the reference conguration 0 yields the algorithmic tangent stiness tensor Lijkl in reference
p = n+1Fim n+1 p n+1 p n+1 p n+1 AT S Fjn Fkr Fls Lmnrs
(6.220)
Remark 6.3.8 In the limit, for small deformations, isotropic response, the Algorithmic Tangent Stiness tensor LAT S becomes ijkl
AT S lim LAT S = Evtpq = vtpq
Eknpq 1 ncd Ecdab 1 Hmr mrpq vtab ncd Rcdvt Rknmr Hmr Rknvt Eknpq 1 vtpq Zmn e Csq Tpk mmn e E rs kspq
1
(6.221)
pm nq + pm nq +
1
e Csb Tad
n+1Zmn H + H K K K K mmn H + H nab Eabpq 1 mmn nab Eabpq 1 mnpq mnpq K K K K mmn + H H nab Eabpq 1 mnpq mmn K K K K nab Rabmn Hmn + H (6.225) K K lim Fpq (Tmnpq ) Zmn Fpq (Tmnpq )
Version: April 29, 2010, 9:23
149
It is noted that the Algorithmic Tangent Stiness tensor given by 6.221 compares exactly with its small strain c counterpart (Jeremi and Sture, 1997).
6.4
6.4.1
We here present a detailed formulation of a material and geometric nonlinear static nite element system of equations. The conguration of choice is material or Lagrangian. Eulerian and mixed EulerianLagrangian conguration will be mentioned as need be.
6.4.2
Equilibrium Equations
The local form of equilibrium equations in material format (Lagrangian) for static case can be written as: PiJ,J 0 bi = 0 (6.226)
where PiJ = SIJ (FiI )t and SIJ are rst and second PiolaKirchho stress tensors, respectively and bI are body forces. Weak form of equilibrium equations is obtained by premultiplying 6.226 with virtual displacements ui and integrating by parts on the initial conguration B0 (initial volume V0 ): ui,j Pij dV =
V0 V0
0 ui bi dV
ui ti dS
S0
(6.227)
It proves benecial to rewrite Lagrangian format of weak form of equilibrium equilibrium by using symmetric second PiolaKirchho stress tensor Sij : ui,j Fjl Sil dV
V0
= = = = = (6.228)
V0
V0
1 (ui,l + ui, juj,l ) + (ul,i + ul,j uj,i ) Sil dV V0 2 1 ((ui,l + ul,i ) + (ui,j uj,l + ul,j uj,i )) Sil dV V0 2
where we have used the symmetry of Sil , denition for deformation gradient Fki = ki + uk,i . We have also conveniently dened dierential operator Eil (ui , ui )
(6.229)
Version: April 29, 2010, 9:23
150
6.4.3
Consider the motion of a general body in a stationary Cartesian coordinate system, as shown in Figure (6.6), and assume that the body can experience large displacements, large strains, and nonlinear constitutive response. The aim is to evaluate the equilibrium positions of the complete body at discrete time points 0, t, 2t, . . . , where t is an increment in time. To develop the solution strategy, assume that the solutions for the static and kinematic variables for all time steps from 0 to time t inclusive, have been obtained. Then the solution process for the next required equilibrium position corresponding to time t + t is typical and would be applied repetitively until a complete solution path has been found. Hence, in the analysis one follows all particles of the body in their motion, from the original to the nal conguration of the body. In so doing, we have adopted a Lagrangian ( or material ) formulation of the problem.
x2 , tx2 , x2
t+ t 0 0 0
P( t+tx1t+t x2 , x3 ) ,
t+ t t+t t+t t t
A V
A 0 V
Configuration at time 0
Configuration at time t
n+1
0
t+t t
x3 ,tx3 , x3 t+ t
xi = x i + u i x i = 0xi + t ui t+t xi = t xi + ui
t+t
x1 ,t x1,t+ tx1
i=1,2,3
Figure 6.6: Motion of body in stationary Cartesian coordinate system Weak format of the equilibrium equations can be obtained by premultiplying 6.226 with virtual displacements ui and integrating by parts. We obtain the virtual work equations in the Lagrangian format: ui,j Pij dV =
V0 V0
0 ui bi dV
ui ti dV
S0
(6.230)
Virtual work equations can also be written in terms of second PiolaKirchho stress tensor SIJ as: ui,j Fjl Sil dV =
V0 V0
0 ui bi dV +
S0
ui ti dV
(6.231)
which after some algebraic manipulations, and after observing that SIJ = SJI yields (SEE ABOVE!) By intro ducing a dierential operator E(u1 , u2 ) as: 1 Eil (1ui , 2ui ) = 2
Jeremi et al. c
1
ui,l + 1ul,i +
1 2
(6.232)
Version: April 29, 2010, 9:23
151
virtual work equation 6.229 can be written as: Eil (ui , ui )Sil dV
V0
=
V0
0 ui bi dV +
S0
ui ti dV
(6.233)
(6.234)
(6.235)
(k)
W ext (ui )
0 ui
c
n+1 0 bi
dV
ui
c
n+1 0 ti
dS
(6.236)
6.4.4
In this Chapter we elaborate on the choice of Total Lagrangian (TL) formulations as computational domain. We also choose Newton type procedure for satisfying equilibrium, i.e. virtual work for a given computational domain. Given the displacement eld ui (Xj ), in iteration k, the iterative change ui ui
(k+1) (k)
= ui
(k)
+ ui
(6.237)
) W (ui , ui ) + W (ui , ui ; ui )
(k)
(k)
(6.238)
Here, W (ui , ui ) is the virtual work expression W (ui , ui ) = W (ui , ui )int + W ext (ui ) with W int (ui , n+1ui ) = 0
(k) (k) (k) Eij (ui , n+1ui ) n+1Sij dV 0 0 (k) (k)
(k)
(6.239)
(6.240)
W ext (ui )
0 ui
c
n+1 0 bi
dV
ui
c
n+1 0 ti
dS
(6.241)
(k)
=
c
(6.242) Here we have used dSij = 1/2 Lijkl dCkl = Lijkl Ekl (ui , ui ).
Jeremi et al. c Version: April 29, 2010, 9:23
152
In order to obtain expressions for stiness matrix we shall work on 6.242 in some more details. To this end, (6.242) can be rewritten by expanding denitions for E as W (ui , ui ; ui ) = 1 4
c (k)
((uj,i + ui,j ) + (uj,r ur,i + ui,r ur,j )) Lijkl ((uk,l + ul,k ) + (uk,s us,l + ul,s us,k )) dV + +
c
((uj,i + ui,j ) + (uj,r ur,i + ui,r ur,j )) Lijkl ((uk,l + ul,k ) + (uk,s us,l + ul,s us,k )) dV (6.244)
2W (ui , ui ; ui ) =
(k)
(6.245)
+
c
+
c
1 (uk,l + ul,k ) dV 2 1 Lijkl (uk,s us,l + ul,s us,k ) dV 2 1 1 (uj,r ur,i + ui,r ur,j ) Lijkl (uk,s us,l + ul,s us,k ) dV 2 2 1 1 (uj,r ur,i + ui,r ur,j ) Lijkl (uk,l + ul,k ) dV 2 2 Lijkl
(6.246)
plasticity lead to the loss of major symmetry (Lijkl =Lklij ). Moreover, it can be shown (i.e. Jeremi and Sture c (1997)) that there is algorithmic induced symmetry loss even for associated ow rules. With the minor symmetry of Lijkl one can write (6.246) as: 1W (ui , ui ; ui ) = +
c (k) c
(Lijkl = Ljikl = Lijlk ). However, Major symmetry cannot be guaranteed. Nonassociated ow rules in elasto-
It should be noted that the Algorithmic Tangent Stiness (ATS) tensor Lijkl poses both minor symmetries
ui,j Lijkl ul,k dV ui,j Lijkl uk,s ul,s dV ui,r ur,j Lijkl uk,s ul,s dV ui,r ur,j Lijkl ul,k dV (6.247)
+
c
+
c
Similarly, by observing symmetry of second PiolaKirchho stress tensor Sij we can write 2W (ui , ui ; ui ) =
Jeremi et al. c
(k)
(6.248)
Version: April 29, 2010, 9:23
153
Weak form of equilibrium expressions (i.e. (6.236) and (6.236) ) for internal (W int ) and external (W ext ) virtual work, with the above mentioned symmetry of Sij can be written as W int (ui , n+1ui ) = 0
(k)
ui,j Sij dV +
c c
(6.249)
W ext (ui )
0 ui bi dV
ui ti dS
c
(6.250)
Standard nite element discretization of displacement eld yields: ui ui = HI uIi functions and uIi are nodal displacements. With this approximation, we have: 1W (ui , ui ; ui ) = +
c (k) c
(6.251)
where ui is the approximation to exact, analytic (if it exists) displacement eld ui , HI are standard FEM shape
(HI,j uIi ) Lijkl (HQ,k Ql ) dV u (HI,j uIi ) Lijkl (HJ,k uJs ) (HQ,s Ql ) dV u (HI,r uIi ) (HJ,j uJr ) Lijkl (HJ,k uJs ) (HQ,s Ql ) dV u (HI,r uIi ) (HJ,j uJr ) Lijkl (HQ,k Ql ) dV u (6.252)
+
c
+
c (k)
2W (ui , ui ; ui ) =
(k)
(6.253)
(6.254)
W ext (ui )
0 (HI uIi ) bi dV
(HI uIi ) ti dS
c
(6.255)
Upon noting that virtual nodal displacements uIi are any nonzero, continuous displacements, and since they occur in all expressions for linearized virtual work (from Equations (6.238), (6.239), (6.240), (6.241) and (6.242)) they can be factored out so that we can write (while remembering that W 1 + W 2 + W ext + W int = 0: (HI,j ) Lijkl (HQ,k Ql ) dV u (HI,j ) Lijkl (HJ,k uJs ) (HQ,s Ql ) dV u (HI,r ) (HJ,j uJr ) Lijkl (HJ,k uJs ) (HQ,s Ql ) dV u (HI,r ) (HJ,j uJr ) Lijkl (HQ,k Ql ) dV u (HI,l ) (HQ,j Ql ) Sij dV u
c
+
c
+
c
+
c
+ +
c
(HI,j ) Sij dV +
c
=
c
0 (HI ) bi dV +
(6.256)
Version: April 29, 2010, 9:23
Jeremi et al. c
154
By rearranging previous equations one can write: HI,j Lijkl HQ,k dV + HI,j Lijkl HJ,k uJs HQ,s dV + HI,l HQ,j Sij dV
c
+
c
Ql u
+
c
=
c
0 (HI ) bi dV +
(6.257)
The vectors of external and internal forces are fint fext = = (W int (ui , n+1ui )) 0 (ui ) (W ext (ui )) (ui )
(k)
(6.258) (6.259)
stress tensor Sij with respect to the right deformation tensor Ckl dSij = 1 Lijkl dCkl with dCkl = 2 Ekl (dui , ui ) 2
The Algorithmic Tangent Stiness (ATS) tensor LAT S is dened as a linearization of second PiolaKirchho ijkl
(6.260)
Then, the global algorithmic tangent stiness matrix (tensor) is given as Kt = (W (ui , ui ; ui )) (ui )
(k)
(6.261)
The iterative change in displacement vector ui is obtained by setting a linearized virtual work to zero W (ui , ui
(k+1)
) = 0 W (ui , ui ) = W (ui , ui ; ui )
(k)
(k)
(6.262)
Total Lagrangian Format The undeformed conguration 0 is chosen as the computational domain (c = 0 ). The iterative displacement ui is obtained from the equation W (ui , n+1ui ) = W (ui , ui ; n+1ui )
Jeremi et al. c
(k) (k)
(6.263)
Version: April 29, 2010, 9:23
155
=
c
and
0 ui
c
n+1
bi dV
ui
c
n+1
ti dS
(6.264)
W (ui , ui ; n+1ui ) =
(k)
(k) (k) (k) Eij (ui , n+1ui ) n+1Lijkl Ekl (ui , n+1ui ) dV
+
c
(k)
(6.265)
In the case of hyperelasticplastic response, second PiolaKirchho stress n+1Sij is obtained by integrating the constitutive law, described in Chapter 6.3. It should be noted that by integrating in the intermediate conguration, we obtain Mandel stress n+1Tij and subsequently30 the second PiolaKirchho stress Skj . The ATS tensor Lijkl conguration we need to perform a pull-back from the intermediate conguration to the initial one
n+1 p Sij = n+1Fip n+1 p n+1 Fjq Spq
is then obtained based on Skj . In order to obtain second PiolaKirchho stress Skj and ATS tensor in initial
(6.266) (6.267)
n+1
p Lijkl = n+1Fim
30 S kj
` 1 Tij = Cik
Jeremi et al. c
156
Jeremi et al. c
Chapter 7
7.1
In previous Chapters we have derived the basic equations of materially nonlinear analysis of solids. Discretization of such problems by nite element methods results in a set of nonlinear algebraic equations called residual force equations: r(u, ) = fint (u) fext = 0 (7.1)
where fint (u) are the internal forces which are functions of the displacements, u, the vector fext is a xed external loading vector and the scalar is a loadlevel parameter that multiplies fext . Equation (7.1) describes the case of proportional loading in which the loading pattern is kept xed. All solution procedures of practical importance are strongly rooted in the idea of advancing the solution by continuation. Except in very simple problems, the continuation process is multilevel and involves hierarchical breakdown into stages, incremental steps and iterative steps. Processing a complex nonlinear problem generally involves performing a series of analysis stages. Multiple control parameters are not varied independently in each stage and may therefore be characterized by a single stage control parameter . Stages are only weakly coupled in the sense that end solution of one may provide the starting point for another.
7.2
Various forms of path following methods1 have stemmed from the original work of Riks (1972), Riks (1979) and Wempner (1971). They aimed at nding the intersection of equation (7.1) with s = constant where s is the
1 also
called arc-length methods with various methods of approximating the exact length of an arc.
157
158
s= where:
ds
(7.2)
ds =
2 u T 2 2 2 du Sdu + d f uref
(7.3)
a = (s)2 (l)2 =
2 u T 2 2 2 u Su + f uref
(l)
(7.4)
where l is the radius of the desired intersection3 and represents an approximation to the incremental arc length. Scaling matrix S is usually diagonal non-negative matrix that scales the state vector u and uref is a reference value with the dimension of uT Su. It is important to note that the vector u and scalar are incremental and not iterative values, and are starting from the last converged equilibrium state. The main essence of the arc-length methods is that the load parameter becomes a variable. With load parameter variable we are dealing with n + 1 unknowns4 . In order to solve this problem we have n equilibrium equations (7.1) and the one constraint equation (7.4). We can solve the augmented system of n + 1 equations by applying the Newton-Raphson5 method to equations (7.1) and (7.4)
rnew (u, )
2 u 2 uT Su + 2 f = 0 u2 ref
(7.6)
is the tangent stiness matrix. The aim is to have rnew (u, ) = 0 and anew = 0 so the
bit dierent form in that it is scaled with scaling matrix S, introduced by Felippa (1984). Figure (7.1). 4 n unknown displacement variables and on extra unknown in the form of load parameter. 5 By using a truncated Taylor series expansion.
3 See
Kt
2
ref
2 2 u2u uT S 2 f
fext
rold aold
(7.7)
Jeremi et al. c
159
Load f
1 f ext 1 f ext 2 fext ( u2 , 2 fext) ( u3 , 3 fext )
2 f ext 3 fext
0 f ext
Equilibrium Path
l
Constraint Hypersurface ( u0 , 0 fext) 0 fext u0 u
1
u2
Displacement
u u 1
0
u 2
u 3
Figure 7.1: Spherical arc-length method and notation for one DOF system. One can solve previous system of two equations for u and :
Kt
2
ref
2 2 u2u uT S 2 f
fext
rold aold
(7.8)
K=
2 2 u2u uT S 2 f
fext
(7.9)
= K1
rold aold
(7.10)
It should be mentioned that the augmented stiness matrix remains non-singular even if Kt is singular.
augmented Jacobian.
Jeremi et al. c
160
In section (7.2) we have introduced an constraining equation that is intended to reduce the so called drift error in the incremental nonlinear nite element procedure. The constraining equation was given in a rather general form. Some further comments and observations are in order. By assigning various numbers to parameters u , f , S and uref one can obtain dierent constraining schemes from (7.4).
Load f f = u f < u f << u
Load f f = u
Equilibrium Path
f >> u f > u
Equilibrium Path Constraint Hypersurfaces
l
Constraint Hypersurfaces
Displacement u
Displacement u
Figure 7.2: Inuence of u and f on the constraint surface shape. Coecients u and f may not be simultaneously zero. Useful choices for S are I, Kt and diag (Kt ). If S = I and uref = 1 the method is called the arclength method7 . If we choose S = diag (Kt ) nice scaling is obtained8 but otherwise no physical meaning can be attributed to this scaling type. With S = Kt and f 0 general equation (7.4) can be further specialized to load () control with u 0; f 1 and state control9 with one ends up with something very similar to the external work constraint of Bathe and Dvorkin (1983). A rather
associated with the case in which only one of the components of the displacement vector u10 is specied. This
u 1; f 0 and S = I. In the nite element literature, the term displacement control has been traditionally may be regarded either as a variant of state control, in which a norm that singles out the ith component is used, or as a variant of the control if the control parameter is taken as ui . It is, of course, possible to make the previous parameters variable, functions of dierent unknowns. For example if one denes uref = uT Su then close to the limit point u 0
7 It 8 For
2 u u2 ref
2 f that makes our constraint from equation (7.4) behave like state
actually reduces to the original work of fRiks (1972), Riks (1979) and Wempner (1971). example if FEM model includes both translational and rotational DOFs. 9 That is the cylindrical constraint, or general displacement control. 10 Say u . i
Jeremi et al. c
161
control. One important aspect of scaling constraint equations by using S = diag (Kt ) or S = Kt is the possibility of nonpositive denite stiness matrix Kt . It usually happens that after the limit point is passed, at least one of the eigenvalues of Kt is nonpositive, thus rendering the constraint hypersurface nonconvex. In order to get better control of the solution to the system of equations (7.10) one may directly introduce the constraint from equation (7.6) by following the approach proposed by Batoz and Dhatt (1979), as described by Criseld (1991) and Felippa (1993). The iterative displacement u is split into two parts, and with the Newton change at the new unknown load level:
(7.11)
u = K1 r uold , = K1 fint (uold ) new fext t t = K1 fint (uold ) old + fext = K1 t t fint (uold ) old fext fext (7.12) = K1 r uold , old fext = K1 rold + K1 fext = u + ut t t t
where ut = K1 fext is the displacement vector corresponding to the xed load vector fext , and u is an iterative t change that would stem from the standard load-controlled Newton-Raphson, at a xed load level old . With the solution11 for the u from (7.12), the new incremental displacements are:
(7.13)
where is the only unknown. The constraint from equation (7.4) can be used here, and by rewriting it as:
= (l)
(7.14)
then by substituting unew from equation (7.13) into equation (7.14) and by recalling that new = old + one ends up with the following quadratic scalar equation:
2 u uold + u + ut 2 uref
S uold + u + ut + old +
2 f
= (l)
Jeremi et al. c
162
2 u T 2 2 ut Sut + f uref
+ + =0 (7.16)
+2 + or:
2 u uold + u 2 uref T
S uold + u l2 + old
2 f
a1 2 + 2a2 + a3 = 0 where:
2 u 2 uT Sut + f t u2 ref
(7.17)
a1 2a2 a3
= = =
2 u uold + u u2 ref
S uold + u l2 + old
2 f
a2 +
= 2 =
a2
(7.18)
(7.19)
unew = uold + u + ut
(7.20)
An ambiguity is introduced in the solution for from (7.18). The tangent at the regular point on the equilibrium path has two possible directions, which generally intersect the constraint hypersurface at two points. However, some exceptions from that rule are possible, so the solutions from (7.18) can be categorized as: Real roots of opposite signs. This occurs when the iteration process converges normally and there is no schemes proposed below. Real roots of equal sign opposite to that of old . This usually happens when going over a at limit point.
limit or turning point enclosed by the constraint hypersurface. The root is chosen by applying one of the
Jeremi et al. c
163
Real roots of equal sign same as that of old . This is an unusual case. It may signal a turning point or be triggered by erratic iteration behavior. Complex roots. This is an unusual case too. It may signal a sharp turning point, a bifurcation point, erratic or divergent iterates.
For the rst two cases, the correct can be chosen by applying one of the following schemes.
7.5.1
Positive External Work The simplest rule requires that the external work expenditure over the predictor step be positive:
T T W = fext u = fext K1 fext > 0 t
(7.21)
T The simple conclusion is that should have the sign of fext K1 fext . This condition is particularly eective at t
limit points. However, it fails when fext and K1 fext are orthogonal: t
T fext K1 fext = 0 t
(7.22)
This can happen at: Bifurcation points, Turning points, The treatment of bifurcation points is of a rather special nature and is left for the near future. Turning points12 can be traversed by a modication of a previous rule, as described in the next section. Angle Criterion Near a turning point application of the positive work rule (7.21) causes the path to double back upon itself. When it crosses the turning point it reverses so the turning point becomes impassable. Physically, a positive work rule is incorrect because in passing a turning point the structure releases external work until another turning point is encountered. To pass a turning point imposing a condition on the angle of the prediction vector proves more eective. The idea is to compute both solutions 1 and 2 and then both pnew and unew : 1 1
unew = uold + u + 1 ut 1
(7.23)
unew = uold + u + 2 ut 2
12 One
(7.24)
might ask why treating turning points in a material nonlinear analysis?. The answer is rather simple: try to prevent all
unnecessary surprises. For a good account of some of surprises in material nonlinear analysis one might take a look at some examples in Criseld (1991) pp. 270.
Jeremi et al. c
164
Load f
Limit Point
Primary Equilibrium Path Turning Point Bifurcation Point Secondary Equilibrium Path
Displacement u
Figure 7.3: Simple illustration of Bifurcation and Turning point. The one that lies closest to the old incremental step direction uold is the one sought. This should prevent the solution from double backing. The procedure can be implemented by nding the solution with the minimum angle between uold and pnew , and hence the maximum cosine of the angle: uold uold
T
cos =
uold + u + ut old + u + u u t
(7.25)
where u = K1 rold and ut = K1 fext . Once the turning point has been crossed, the work criterion should t t be reversed so the external work is negative. By directly introducing the constraint from equation (7.6) and following the method through equations (7.12) to (7.25) a limitation is introduced. Precisely at the limit point, Kt is singular and the equations cannot be solved. However, Batoz and Dhatt (1979) and Criseld (1991) report that no such problem has occurred, because one appears never to arrive precisely at limit point.
7.5.2
point:
Predictor step
The predictor solution is achieved by applying one forward Euler, explicit step from the last obtained equilibrium
up = K1 qe = p K1 fext = p ut t t
Jeremi et al. c
(7.26)
Version: April 29, 2010, 9:23
165
where Kt is the tangent stiness matrix at the beginning of increment. Substituting equation (7.26) into the constraint equation (7.14) one obtains:
= = = (l)
2
2 p
2 u T 2 2 ut Sut + f uref
(7.27)
p =
l
2 u u2 ref
(7.28)
2 |uT Sut | + f t
where l > 0 is the given increment length. The absolute value of |uT Sut | is needed if the stiness matrix t is chosen as a scaling matrix, i.e. S = Kt , since, after passing limit point, the stiness matrix is nonpositive denite so uT Sut 0. The question of choosing the right sign + or in (7.28) is still a vigorous research t pivot during factorization of the tangent stiness matrix Kt . If more than one pivot happens to be negative, one is advised14 to stop the analysis and try to restart from previously converged solution with smaller step size.
topic. In the simplied procedure13 the negative sign is chosen with respect to the occurrence of one negative
7.5.3
Automatic Increments
A number of workers have advocated dierent strategies for controlling the step length size. In this work we will follow the strategy advocated by Criseld (1991). The idea is to nd the new incremental length by applying:
lnew = lold
Idesired Iold
(7.29)
where lold is the old incremental factor for which Iold iterations were required, Idesired is the input, desired number of iterations15 and the parameter n is set to
1 2
7.5.4
Convergence Criteria
Introduction of an iterative scheme calls for the introduction of an iteration termination test. There are several convergence criteria that can be applied.
is not guaranteed to work if one takes into account bifurcation phenomena. more details see Criseld (1991). 15 Say I desired 3.
14 For 13 Which
Jeremi et al. c
166
Displacement Convergence Criteria. The change in the last correction u of the state vector u, measured in an appropriate norm, should not exceed a given tolerance u . For example, if we use Euclidean norm16 the termination criteria can be written as: u
scaled
(u) S (u) u
(7.30)
Scaling matrix S is used in order to ensure that for a problem involving mixed variables17 , all parameters have the same dimension. Here, an obvious choice for the scaling matrix is S = diag(K1 ). If, on the t other hand we dont have mixed variables in state vector u then the simplest choice for scaling matrix is S = I. Residual Convergence Criteria. Since the residual r(u, ) measures the departure from the equilibrium path, an appropriate convergence test would be to compare Euclidean norm of residual with some predened tolerance: r(u, )
scaled
(r) S (r) r
(7.31)
Here, an obvious choice for scaling matrix is S = diag(Kt ) Energy Based Convergence Criteria. The two previous convergence criteria can be combined in a single work change criterion: (u) (r) =
T
(u) (r) u r
(7.32)
A word of caution is appropriate at this point. As pointed out by Criseld (1991), it follows that: (u) (r) = (u)
T T
(7.33)
the stiness of Kt . This merely implies that a stationary energy position has been reached in the current iterative direction, u. This can occur when the solution is still far away from equilibrium. Since u and r usually have physical dimensions, so do necessarily u and r . For a general purpose implementation of NewtonRaphson iteration this dependency on physical units is undesirable and it is more convenient
where the iterative change was (u) = K1 r. It should be noted that equations (7.33) give a measure of t
to work with ratios that render the u and r dimensionless. Displacement Convergence Criteria can be rendered dimensionless by using ratio of scaled Euclidean norm of iterative displacement u norm of total displacement u u scaled u u scaled The similar approach can be used for Residual Convergence Criteria: r scaled predictor r scaled
16 The 17 For
scaled
scaled :
(7.34)
(7.35)
Jeremi et al. c
167
Another important thing to be considered is Divergence. The NewtonRaphson scheme is not guaranteed to converge. Some sort of divergence detection scheme is therefor necessary in order to interrupt an erroneous iteration cycle. Divergence can be diagnosed if either of following inequalities occur: u scaled gu u scaled r scaled gr rpredictor scaled where gu and gr are dangerous growth factors that can be set to, for example gu = gr = 100. In some cases, the NewtonRaphson iteration scheme will neither diverge nor converge, but rather exhibit oscillatory behavior. To avoid excessive bouncing around, a good practice is to put upper limit to the number of iterations performed in one iteration cycle. Typical limits to the iteration number are 20 to 50. (7.36)
(7.37)
7.5.5
The progress of the scheme will be briey described, in relation with the Figure (7.1). The procedure starts from a previously converged solution (u0 , 0 fext ). An incremental, tangential predictor step u1 , 1 is obtained18 and the next point obtained is (u1 , 1 fext ). The rst iteration would then use quadratic equation 7.17 where constants a1 , a2 and a3 should be computed with uold = u1 and old = 1 , to calculate 1 and19 u1 = K1 r (u1 , 1 ) + 1 K1 fext . After these values are calculated, the updating procedure20 would lead to: t t
2 = 1 + 1
and
u2 = u1 + u1
(7.38)
When added to the displacements u0 and load level 0 , at the end of the previous increment this process would lead to the next point (u2 , 2 fext ). The next iteration would then again use quadratic equation 7.17 where constants a1 , a2 and a3 should be computed with uold = u2 and old = 2 , to calculate 3 and u3 = u + 2 ut . After these values are calculated, the updating procedure would lead to:
3 = 2 + 2
and
u3 = u2 + u2
(7.39)
When added to the displacements u0 and load level 0 , at the end of the previous increment this process would lead to the next point (u3 , 3 fext ).
18 As
19 From
Jeremi et al. c
168
Jeremi et al. c
Chapter 8
8.2
8.2.1
Newmark Integrator
The Newmark time integration method (Newmark, 1959) uses two parameters, and , and is dened by the following two equations:
n+1
n+1
x]
(8.1) (8.2)
These equatons give relationship between knowns variables at time step n to the unknown variables at next time step n + 1. Method is in general an implicit one, except when both and are zero. There are several possible implementation methods for Newmark Integrator. One of them, the predictorcorrector method is dened throuh a predictors: 1 n+1 x = n x + t n x + t2 ( ) n x 2 n+1 x = n x + t (1 ) n x and then a correctors:
n+1 n+1 n+1
(8.3) (8.4)
x = x =
x + t2 x + t
n+1
n+1
n+1
170
n+1
x,
n+1
R=M
n+1
x+C
n+1
x + K
x + F (n+1 x)
n+1
(8.7)
It can be solved using Newton integration method M + tC + t2 K = n+1 R x Equation 8.3 to 8.8 constitute an iterative solving procedure. An alternative Newmark integration approach is to use displacement as the basic unknowns, and the following dierence relations are used to relate
n+1 n+1
(8.8)
x and
n+1
x to
n+1
x = x =
n+1
t 1 t2
n+1
x nx + 1 x nx
x+ 1
(8.9) (8.10)
n+1
1 n 1 x+ 1 t 2
n+1
n n x+ 1 x+ 1 t 2 1 n 1 n 1 n = x x x+ 1 t 2 t2
(8.11) (8.12)
x +
(8.13) (8.14)
The Newton integration method becomes M C + + K x = n+1 R t2 t Equation 8.11 to 8.15 constitute an iterative solving procedure Argyris and Mlejnek (1991b). If the parameters and satisfy 1 1 1 , = ( + )2 2 4 2 (8.16) (8.15)
it is unconditionally stable and second-order accurate. Any value greater than 0.5 will introduce numerical damping. Well-known members of the Newmark time integration method family include: trapezoidal rule or average acceleration method for = 1/4 and = 1/2, linear acceleration method for = 1/6 and = 1/2, and (explicit) central dierence method for = 0 and = 1/2. If and only if = 1/2, the accuracy is second-order Hughes (1987).
8.2.2
HHT Integrator
Numerical damping introduced in the Newmark time integration method will degrade the order of accuracy. The Hilber-Hughes-Tailor (HHT) time integration -method (Hilber et al., 1977), (Hughes and Liu, 1978a) and
Jeremi et al. c Version: April 29, 2010, 9:23
171
(Hughes and Liu, 1978b) using an alternative residual form by introducing an addition parameter to improve the performance:
n+1
R=M
n+1
x + (1 + )F (n+1 x, n+1 x) F (n x, n x)
n+1
(8.17)
but retaining the Newmark nite-dierence formulas 8.1 and 8.2 or 8.9 and 8.10. If = 0, Equation 8.17 is reduced to Equation 8.7, and this special case of the HHT time integration method is exactly the Newmark time integration method. Decreasing value increase numerical dissipation (Hughes, 1987). The iteration method for HHT time integration is similar to that of Newmark time integration. Due to the change of Equation 8.17, Equations 8.8 and 8.15 respectively become M + (1 + )tC + (1 + )t2 K = n+1 R x for acceleration iteration and (1 + )C M + + (1 + ) x = n+1 R t2 t for displacement iteration. If the parameters , and satisfy 1/3 0, = 1 (1 2), 2 = 1 (1 )2 4 (8.20) (8.19) (8.18)
it is unconditionally stable and second-order accurate (Hughes, 1987). There are dierent denotation meaning of the parameter , e.g. the parameter in the HHT integrator codes of OpenSees equals to the conventional plus one.
Jeremi et al. c
172
Jeremi et al. c
Chapter 9
9.1
9.1.1
For single-phase material encountered in structural mechanics, the response under ultimate load (failure) can be predicted using very simple calculations, at least for static problems. But for soil mechanics, simple, limit-load calculations can not be fully justied under static situation. However, for problems of soil dynamics, the use of simplied methods is almost never admissible. As the strength of the soil can be determined once the pore water pressures are known, it is possible to reduce the soil mechanics problem to that of the behavior a single phase. Then we can use again the simple, single-phase analysis approaches. Now we will introduce the concept of eective stress. The relationship between eective stress, total stress and pore pressure is (assume tensile components of stress as positive and compressive pressure, p is positive) (Zienkiewicz et al., 1999a) ij = ij + ij p
(9.1)
where ij is eective stress tensor, ij is total stress tensor, ij is Kronecker delta. ij = 1, when i=j, and KS is the bulk modulus of the solid particle. For most of the soil mechanics problems, as the bulk modulus KS of the solid particles is much larger that that of the whole material, 1 can be assumed. Equation (9.1) becomes ij = ij + ij p
ij = 0, when i = j. For isotropic materials, = 1 KT /KS . KT is the total bulk modulus of the solid matrix,
(9.2)
9.1.2
Before proceeding with the general equations, we introduce the associated notation. In this we have 173
174
ij ui wi p
the total Cauchy stress in the mixture at any instant the displacement of the solid skeleton the pseudo-displacement of the uid phase relative to the skeleton of solid the pore water pressure the strain increment of the solid phase the rotation increment of the solid phase
1 ij = 2 (dui,j + duj,i )
ij = 1 (dui,j duj,i ) 2 , s , f n
the densities of the mixture, and the solid phase and water respectively
the porosity the rate at which volume of water changes per unit total volume of mixture
= wi,i
With these denitions, and others introduced as necessary, we can write the governing equation of the coupled system. The equilibrium equation of the mixture The overall equilibrium or momentum balance equation for the soil-uid mixture can be written as (Zienkiewicz et al., 1999a) ij,j i f [wi + wj wi,j ] + bi = 0 u (9.3)
Where ui is the acceleration of the solid part, bi is the body force per unit mass, wi + wj wi,j is the uid acceleration relative to the solid part, wi is local acceleration, wj wi,j is convective acceleration. The underlined terms in the above equation represent the uid acceleration relative to the solid and convective terms of this acceleration. Generally this acceleration is so small that we shall frequently omit it. And for static problems, equation(9.3) only consists of the rst and last terms. For fully saturated porous media (no air inside), from denition = = = = = Mt Vt Ms + Mf Vt Vs s + Vf f Vt Vf Vt Vf f + s Vt Vt nf + (1 n)s (9.4)
= nf + (1 n)s
where Mt , Ms and Mf are the mass of total, solid part and uid part respectively. Vt , Vs and Vf are the volume of total, solid part and uid part respectively.
Jeremi et al. c Version: April 29, 2010, 9:23
175
The equilibrium equation of the uid For the pore uid, the equation of momentum balance can be written as (Zienkiewicz et al., 1999a) p,i Ri f ui f [wi + wj wi,j ]/n + f bi = 0 (9.5)
where R is the viscous drag forces. It should be noted that the underlined terms in (9.5) represent again the convective uid acceleration and are generally small. Also note that, throughout report, the permeability k is used with dimensions of [length]3 [time]/[mass], which is dierent from the usual soil mechanics convention, the permeability has the same dimension of velocity, i.e. [length]/[time]. Their values are related by k = K/f g, where g is the gravitational acceleration at which the is measured. Assuming the Darcy seepage law: nw = Ki, here i is the head gradient. Seepage force is then R = f gi. R can be written as
1 Ri = kij wj
or
Ri = k 1 wi
(9.6)
kij or k are Darcy permeability coecients for anisotropic and isotropic conditions respectively.
Flow conservation equation The nal equation is supplied by the mass conservation of the uid ow (Zienkiewicz et al., 1999a) wi,i + ii + p f + n + s0 = 0 Q f (9.7)
The rst term of equation(9.7) is the ow divergence of a unit volume of mixture. The second term is the volume change of the mixture. In the third term, Q is relative to the compressibility of the solid and uid. The underlined terms represent change of density and rate of volume expansion of the solid in case of thermal changes. They are generally negligible. 1 n n n 1n + + = Q Kf Ks Kf Ks where Ks and Kf are the bulk moduli of the solid and uid phases respectively. Thus, we got the total mixture equilibrium equation (9.3), uid equilibrium equation(9.5) and the ow conservation equation (9.7) for saturated soil. By omitting the convective acceleration (the underline terms in (9.3) and (9.5)), density variation and the volume expansion due to the thermal change (the underline terms in (9.7)), the equations of the total coupled system can be further simplied, they are summarized as below ij,j i f wi + bi = 0 u p,i Ri f ui p =0 Q f wi + f bi = 0 n (9.9) (9.8)
(9.10)
wi,i + ii +
Jeremi et al. c
(9.11)
176
Figure 9.1: Fluid mechanics of Darcys ow (wi ) versus real ow (Ui = wi /n).
9.1.3
Solid part equilibrium equation First of all, a new variable Ui is introduced in place of the relative pseudo-displacement wi , it is dened as Ui = ui + UiR = ui + wi n (9.12)
Insertion of denition (9.12) into equation (9.9)(9.10) and subtraction of [n (9.10)] from equation (9.9)
ij,j i + bi + np,i + nRi + nf ui nf bi = 0 u By substituting = (1 n)s + nf ij,j (1 n)s ui nf ui + (1 n)s bi + nf bi + np,i + nRi + nf ui nf bi = 0 ij,j + np,i + (1 n)s bi (1 n)s ui + nRi = 0 By using the denition of eective stress (9.1),(9.13) becomes ij,j ( n)p,i + (1 n)s bi (1 n)s ui + nRi = 0 Fluid part equilibrium equation The uid part equilibrium equation can be obtained simply by [n (9.10],i.e. np,i nRi nf ui f wi + nf bi = 0 wi np,i nRi nf (i + u ) + nf bi = 0 n From equation (9.12), we have wi Ui = ui + n so that equation (9.16) becomes: np,i + nf bi nf Ui nRi = 0
Jeremi et al. c
(9.13)
(9.14)
(9.15)
(9.16)
(9.17)
(9.18)
Version: April 29, 2010, 9:23
177
Mixture balance of mass By dierentiating equation (9.12) in time space, we have wi,i = nUi,i nui,i Notice that ii = ui,i , so that equation (9.19) becomes wi,i = nUi,i nii By substituting (9.20) to (9.11), we obtain nUi,i nii + ii + or: 1 nUi,i = ( n)ii + p Q below ij,j ( n)p,i + (1 n)s bi (1 n)s ui + nRi = 0 np,i + nf bi nf Ui nRi = 0 nUi,i = ( n)ii + 1 p Q
(9.19)
(9.20)
p =0 Q
(9.21)
(9.22)
Thus we got the whole set of modied governing equations (9.15),(9.18) and (9.22). They are summarized as
From the modied equation set (9.23),(9.24) and (9.25), we can see that only ui occurs in the rst equation,and i in the second, thus leading to a convenient diagonal form in discretization. only U Now we have a complete equation system given by (9.23),(9.24) and (9.25). With the basic denitions introduced earlier, there are three essential 1. solid displacement u 2. pore pressure p 3. absolute uid displacement U The boundary conditions imposed on these variables will complete the problem. These boundary conditions are: For the momentum balance part, on boundary t , traction ti (t)(or ij nj ), where ni is the i-th component of the normal to the boundary is specied. On boundary u , the displacement ui is given. For the uid part, again the boundary is divided into two parts. On p , the pressure p is specied, on w , the normal outow wn is specied. For impermeable boundary a zero value for the outow should be specied. The boundary conditions can be summarized below ti ui = t u on = u = w (9.26)
Version: April 29, 2010, 9:23
= ij nj = ti = ui on
Jeremi et al. c
178
and p nT w = p w on on = p = w (9.27)
= p = wn
9.2
The solutions to the problems governed by the modied governing equation set (9.23),(9.24) and (9.25) can be found by solving partial dierential equations, which can be written as A + B + L() = 0 (9.28)
where A, B are matrices of constant, and L is an operator involving spatial dierentials. The dot notation represents the time dierentiation. is a vector of dependent variables, in this problem, we say it represents the displacement u or the pressure p. The nite element solution of a problem will always proceed as the following procedures: 1. Discretize or approximate the unknown functions by a nite set of parameters k and shape function Nk . They are specied in space dimensions. Thus, we can write
n
h = =
Nk k
k=1
(9.29)
2. Insert the value of the approximating function into the dierential equations to obtain a residual, then we can write a set of weighted residual equations in the form
(9.30)
In nite element method, the weighting functions Wj are usually identical to the shape functions. The solid displacement ui , the pore pressure p, and the absolute uid displacement Ui can be approximated using shape functions and nodal values. ui
u = NK uKi
p p = NK p K
Ui
U = NK U Ki
(9.31)
p u U where NK , NK and NK are shape functions for solid displacement, pore pressure and uid displacement respec-
tively, uKi ,pK ,U Ki are nodal values of solid displacement, pore pressure and uid displacement respectively.
Jeremi et al. c Version: April 29, 2010, 9:23
179
9.2.1
u To obtain the numerical solution of the rst equation, we premultiply (9.23) by NK and integrate over the domain.
=
t
u NK nj ij d
u NK,j ij d u NK,j ij d
=
t
u NK (ti + ni p)d
= = = =
u = (f1 ) K EP u
(9.32)
where K EP is the stiness matrix of the solid part,ni is the direction of the normal on the boundary. Second term of (9.23) becomes
u NK ( n)p,i d
u NK ( n)ni pd + u NK ( n)ni pd + [
(9.33)
(9.34)
u u NK (1 n)s ij NL d]uLj
where Ms is the mass matrix of solid part. By substituting equations (9.6) and (9.12),last term of (9.23) (Damping Matrix) becomes
u NK Ri d = 1 u NK nkij wj d 1 u NK n2 kij Uj d 1 u NK n2 kij uj d 1 U u NK n2 kij NL d]uL j
= =
1 U u NK n2 kij NL d]U L j [
= C2 U C1 u
Jeremi et al. c
180
(9.37)
(9.38)
(9.39)
in index form EP KKijL (G1 )KiL pL + (C2 )KijL U Lj (C1 )KijL uLj + (Ms )KijL uKi = (f s )Ki Where K EP = (K EP )KimP G1 = (G1 )KiM C2 = (C2 )KijL C1 = (C1 )KijL Ms = (Ms )KijL f = (f s )Ki =
u u NK,j Dijml NP,l d p u NK,i ( n)NM d 1 U u NK n2 kij NL d 1 u u NK n2 kij NL d
(9.40)
= =
(9.41)
9.2.2
(9.42)
(9.43)
U nNK ni pd +
U nNK,i pd
= (f1 )Ki + [
p U nNK,i NM d]pM
= (f1 )Ki + (G2 )KiM pM = (f1 )Ki + (G2 )p Second term of (9.43) is then
U NK f bi d = (f2 )Ki
(9.44)
(9.45)
Version: April 29, 2010, 9:23
Jeremi et al. c
181
U U NK nf ij NL d]U Lj
1 U U NK n2 kij NL d]U Lj
(9.47)
+[
1 u U NK n2 kij NL d]uLj
(9.49)
(9.50)
(9.51)
(G2 )KiM pM (C2 )T LjiK uLj + (C3 )KijL U Lj + (Mf )KijL ULj = (ff )Ki where (ff )Ki G2
T C2
(9.52)
= = = = =
C3 Mf
(C3 )KijL =
1 U U NK n2 kij NL d U U NK nf ij NL d
(Mf )KijL =
(9.53)
9.2.3
By integrating (9.25) in time and noticing ii = ui,i , we can obtain nUi,i = ( n)ii + (9.54)
p By multiplying (9.54) by NM and integrating over the domain, rst term of (9.54) becomes
(9.55)
Version: April 29, 2010, 9:23
Jeremi et al. c
182
= =
p u NM ( n)NL,j d]uLj
1 p N d]pN Q M
= P N M pM = P p
(9.57)
The equation (9.54) becomes GT U + G T u + P p = 0 1 2 in index form (G2 )LiK U Li + (G1 )LiK uLi + PKL pL = 0 (9.59) (9.58)
9.2.4
The numerical forms of governing equations (9.38),(9.50) and (9.58) can be written together in the matrix form as Ms 0 0 0 C1 u p + 0 0 T Mf C2 U 0 0 C2 0 0 C3 K EP u p + GT 1 0 U G1 G2 P 0 u
0 0
GT p 2 0 U fs = 0 ff
(9.60)
or in index form (Ms )KijL 0 0 0 0 0 0 0 (C1 )KijL uLj p N + 0 (C2 )LjiK U Lj (G1 )KiM (G2 )KiL PM N 0 0 (C2 )KijL 0 uLj p N U
(9.61)
Jeremi et al. c
183
where Ms Mf C1 C2 C3 K
EP
= = = = = = = =
(Ms )KijL =
u u NK (1 n)s ij NL d U U NK nf ij NL d
(Mf )KijL =
(C1 )KijL =
(C2 )KijL =
(C3 )KijL =
(K
EP
)KijL =
G1 G2 P
(G1 )KiM =
(G2 )KiM =
p NN
= PN M =
1 p N d Q M (9.62)
f Ki f Ki
solid
f luid
u (f1 )Ki u (f4 )Ki u (f5 )Ki U (f1 )Ki U (f2 )Ki
=
t
u NK nj ij d u NK ( n)ni pd
=
p
u NK (1 n)s bi d U nNK ni pd
=
p
U nNK f bi d
(9.63)
Here we have N u ,N p ,N U as the shape functions of skeleton, pore pressure and uid, ,s ,f are the density of the symbol ni is the direction of the normal on the boundary, u is the displacement of the solid part, p is pore pressure coupled system which can be written in a familiar form as M x + C x + Kx = f total, the solid and the uid phases, respectively, n is the porosity, and by its denition = (1 n)s + nf , the and U is the absolute displacement of the uid part. Equation (9.60) represents the general form (u p U ) for
(9.64)
where x represents the generalized unknown variable. Solution of this equation for each time step will render unknown eld for given initial and boundary conditions.
Jeremi et al. c Version: April 29, 2010, 9:23
184
9.2.5
Isoparametric elements are used in previous sections, where the coordinates are interpolated using the same shape functions as for the unknown. This mapping allows the use of elements of more arbitrary shape than simple forms such as rectangles and triangles. But in static or dynamic undrained analysis the permeability (and compressibility) matrices are zero, i.e.(Q ,and P 0),resulting in a zero-matrix diagonal term in the equation(9.61). The matrix to be solved is the same as that in the solutions of problems of incompressible elasticity or uid
mechanics. Actually a wide choice of shape functions is available if the limiting(undrained) condition is never imposed. Due to the presence of rst derivatives in space in all the equations it is necessary to use Co -continuous interpolation functions and the suitable element forms are shown in Fig.9.2.
Figure 9.2: Shape functions used for coupled analysis, displacement u and pore pressure p formulation
9.3
Examples
In the derivation of the governing equations of the governing equations of the coupled system, there are no limitations on the compressibility of the uid involved, so a general complete u-p-U form is presented. While in the general soil mechanics, the notion of uid incompressibility are often assumed. As a consequence of the incompressible uid, the compressive waves in the uid can be neglected and there is only one dilational wave existing in the mixture. And the pore pressure generation in the compressible uid usually is slightly higher than that of the incompressible uid because of the existence of the oscillatory waves (Zienkiewicz and Shiomi, 1984). Special attention should also be paid to the concept of incompressible uid. The compressibility of the porous media is governed by the compressibility of the solid skeleton, permeability and volume factions, instead of the uid compressibility. Before proceeding to the analysis, we have to make the following assumptions: For high-frequency components, the permeability remains constant, the dependency of the permeability on the frequency should be neglected. Unless specied, all the models in this report are elastic isotropic.
Jeremi et al. c Version: April 29, 2010, 9:23
185
In order to illustrate the performance of the formulation and versatility of the numerical implementation, several elastic one-dimensional problems are presented in the following sections, including: 1. Drilling of a borehole 2. The case of a spherical cavity 3. Consolidation of a soil layer 4. Line injection of a uid in a reservoir 5. Shock wave propagation
9.3.1
The Problem Let us consider an innite half space domain composed of an isotropic, homogeneous and saturated thermoporoelastic material. At its reference state, it is assumed that the temperature, uid pressure and stress elds are uniform and equal respectively, to T0 , p0 and 0 = 0 1(with 0 < 0). At time = 0, an innite cylinder of radius r0 is instantaneously drilled parallel to the vertical axis Oz. It is lled with a uid of the same nature as that saturating the porous medium but at a dierent pressure and temperature at the values of p1 and T1 respectively. The interface r = r0 between the well and the porous medium is assumed to be in thermodynamic equilibrium. In cylindrical coordinates (r, , z), the boundary conditions can be summarized as follows (see Fig.9.3): t0 t>0 0 = 0 1 rr (r0 ) = p1 p(r) = p0 T (r) = T0 (9.65)
r (r0 ) = rz (r0 ) = 0
rr (r ) 0
Analytical Solution Since the well is assumed to be innite long in its vertical axis Oz, the analysis is performed under plane strain hypothesis(zz = 0). Therefore, = r (r)er p = p(r) T = T (r) (9.67)
in which r is the radial displacement. In cylindrical coordinates, Eqn. 9.103 yields rr = r r = r r other ij = 0 (9.68)
Based on the constitutive equations from Coussy (1995), it follows that rr = 0 + 0 ( r r r + ) + 2 b(p p0 ) 3K0 (T T0 ) r r r r r r + ) + 2 b(p p0 ) 3K0 (T T0 ) r r r (9.69)
= 0 + 0 (
Jeremi et al. c
(9.70)
Version: April 29, 2010, 9:23
186
(9.71)
other
ij = 0
(9.72)
Finally combined with the Eqns. 9.69-9.72, it yields the near eld or long-term solution (Coussy, 1995) r = 0 + p1 r0 2 r0 [b(p1 p0 ) + 30 K0 (T1 T0 )] r r0 + ( ) 2 r 2(0 + 2) r0 r r0 2 [b(p1 p0 ) + 3K0 (T1 T0 )] r0 2 {0 }(1 2 ) 2 r 0 + 2 r r0 2 [b(p1 p0 ) + 3K0 (T1 T0 )] r0 2 (1 + 2 ) 2 r 0 + 2 r (9.73)
rr = p1
(9.74)
= (20 + p1 )
(9.75)
zz = 0
(9.76)
And the diusion process can be achieved if that the time are large enough with respect to the characteristics diusion time relative to point r. When the boundary conditions for r = r0 in uid pressure and temperature which are p = p1 and T = T1 apply for the whole model, the following equations correspond to the undrained solution of the instantaneous drilling of a borehole in an innite elastic medium. r = 0 + p1 r0 2 2 r rr = 0 (0 + p1 ) r0 2 r2 r0 2 r2 (9.77)
Version: April 29, 2010, 9:23
= 0 + ( 0 + p1 )
Jeremi et al. c
zz = 0
187 Symbol E Ks Kf s f n Value 0.2 1.2E+6 3.6E+7 1.0E+17 2.7 1.0 0.4 Units kN/m2 kN/m2 kN/m2 ton/m3 ton/m3 -
Parameter Poisson Ratio Youngs Modulus Solid Bulk Modulus Fluid Bulk Modulus Solid Density Fluid Density Porosity
Discussion of the Results As the problem is Axisymmetric, we construct the model as a quarter of a donut. The inside diameter of the donut is 10 cm and the outside diameter is 1 m. To accommodate both the plain strain hypothesis and the geometry of the element for nite element, the thickness of the model is chosen to be 5 cm. The nal mesh is generated as Fig.9.4. And the boundary conditions is as follows: As a consequence of plain strain problem, all the movements for solid and uid in vertical direction Oz are suppressed; the solid and uid displacement for the nodes along the X axis and Y axis are xed in Y and X direction respectively for the reason of axisymmetry; the nodes along the outside perimeter are xed in the solid and uid displacement with the assumption of innite medium. the pressure is translated into nodal forces and applied on the nodes along the inside perimeter. For simplicity, the hydrostatic stress 0 is equal to zero and with the assumption of thermodynamic equilibrium through the process, the temperature factor can be neglected. Also the initial uid pressure p0 is set to be 0 kPa. The analytical solution is studied below using the following set of parameters shown in Table 9.1.
Figure 9.4: The mesh generation for the study of borehole problem In the analysis, ten loading cases for nal uid pressure from 10 kPa to 100 kPa are studied. And by
Jeremi et al. c Version: April 29, 2010, 9:23
188
manipulating the permeability, it is possible to investigate both the drained behavior and undrained behavior. For comparison between the close solution and experimental result is shown in Fig.9.5. From the results, we can see the drained behavior, we choose the permeability as k = 3.6 104 m/s, which is a typical value for sand, the that along the inside perimeter, the close solution and experimental result provide very good agreement to each other. But as the increase of the radius, we can see the analytical solution is getting more and more distant from the experimental results. In another word, the analytical solution can be interpreted as that with the increase from the loading surface, the radial displacement is larger. This is unreasonable in the point of view in soil mechanics. While the experimental result show the eect that with the increase of the radial distance, the radial displacement is decreasing. (The input les for the above examples are: ValidationExamples/WellDrilling/Drained
x 10
3
3.5
r=0.1m, close solution r=0.1m, experimental r=0.5m, close solution r=0.5m, experimental r=0.9m, close solution r=0.9m, experimental
1.5
0.5
10
20
30
40
50 60 Pressure (kPa)
70
80
90
100
Figure 9.5: The comparison of radial solid displacement between analytical solution and experimental result for drained behavior For the undrained behavior, the permeability of k = 3.6 108 m/s is selected as a representative value for
typical clayey soil. The comparison between the close solution and experimental result is provided as well. From the Fig.9.6 we can see that, the analytical solution is linearly away from the experimental result by a ratio of approximately 1.6. It should also be noticed that the close solution of the drained and undrained behavior for the nodes along the inside perimeter are exactly the same, which is contradictory to the denition of drained and undrained behavior. For the drained behavior, as the water easily dissipate from the soil body, the problem can be treated with the knowledge of continuum mechanics using the parameters of the solid skeleton. While for
Jeremi et al. c Version: April 29, 2010, 9:23
189
the undrained behavior, with the involvement of the pore water, the elastic parameters for the mixture should be dierent, so the response will not be the same as well. As a result of this, the experimental results give a more reasonable conclusion. (The input les for the above examples are: ValidationExamples/WellDrilling/Undrained
1 x 10
3
0.8
r=0.1m, close solution r=0.1m, experimental r=0.5m, close solution r=0.5m, experimental r=0.9m, close solution r=0.9m, experimental
0.6
0.4
0.2
10
20
30
40
50 60 Pressure (kPa)
70
80
90
100
Figure 9.6: The comparison of radial solid displacement between analytical solution and experimental result for undrained behavior As for the drained behavior, the uid totally ows out of the soil body and all excessive pore pressure dissipates, there is small coupling between the solid and uid phase. We can use the continuum mechanics to treat this problem. Here introduces a problem of an innite cylindrical tube, with the inner radius R1 and outer radius R0 , subjected to an internal pressure P1 and an external pressure P2 . The displacement eld as follows (S.Timoshenko and D.H.Young, 1940): r = r R 1 2 P1 R0 2 ( + ) r 2(R0 2 R1 2 ) + P1 R 1 2 2 r (9.78)
With P0 = 0 and take the limit of R0 , we can obtain the following equation: r = (9.79)
which is identical to Eq.9.74. Also to minimize the eect of innite boundary, we introduce the result from another model which is exactly the same as the previous one besides the expansion of the outer radius to 30m. At the nal uid pressure of 50 kPa, the results are shown in Fig.9.7. From the plot we can make a conclusion that
Jeremi et al. c Version: April 29, 2010, 9:23
190
the undrained analytical solution from Coussy (1995) is actually the drained solution and the undrained solution still needs to be investigated. (The input les for the above examples are:
ValidationExamples/WellDrilling/Comparison
x 10
0 0.1
0.2
0.3
0.4
0.7
0.8
0.9
Figure 9.7: The comparison of radial solid displacement between two analytical solutions and expanded boundary
9.3.2
The Problem Considering a medium composed of an isotropic, homogeneous, saturated thermoporoelastic material. In its initial state, it is assumed that the temperature, uid pressure and stress elds are uniform and equal respectively, to T0 , p0 and 0 = 0 1(with 0 < 0). At time t= 0, a spherical cavity of radius r0 is immediately drilled and lled with the same saturating uid in the medium. For t > 0, the temperature and the pressure of the uid are kept constant with the value of T1 and p1 respectively. The interface r = r0 between the well and the porous medium is assumed to be in the thermodynamic equilibrium. In spherical coordinates (r, , ), the boundary conditions can be summarized as follows:
Jeremi et al. c Version: April 29, 2010, 9:23
191
t0 t>0
0 = 0 1 rr (r0 ) = p1
p(r) = p0
T (r) = T0
(9.80)
r (r0 ) = r (r0 ) = 0
rr (r ) 0
to be satised by the solution. It is used to model that at the point far from the disturbed area, the state of the medium are held as its initial state. Analytical Solution This is a problem of spherical symmetry. The radial displacement is the only non-zero displacement and all the elds are r and t dependent. Therefore, = r (r)er p = p(r) T = T (r) (9.82)
Strictly speaking, the expressions for r are not boundary conditions. They are complementary conditions
in which r is the radial displacement. In spherical coordinates, Eqn.9.82 yields rr = r r = r r other ij = 0 (9.83)
Based on the constitutive equations from Coussy (1995), it follows that rr = 0 + 0 ( r r r + ) + 2 b(p p0 ) 3K0 (T T0 ) r r r r r r + ) + 2 b(p p0 ) 3K0 (T T0 ) r r r (9.84) (9.85) (9.86)
== = 0 + 0 ( other ij = 0
Finally combined with the Eqns. 9.83-9.86, it yields the near eld or long-term solution (Coussy, 1995) r = r0 [b(p1 p0 ) + 3K0 (T1 T0 )] r0 2 0 + p1 r0 3 + (1 2 ) 4 r2 2(0 + 2) r r0 3 r0 3 2[b(p1 p0 ) + 3K0 (T1 T0 )] r0 r0 3 + 0 (1 3 ) ( 3) r3 r 0 + 2 r r r0 3 [b(p1 p0 ) + 3K0 (T1 T0 )] r0 r0 2 r0 3 0 (1 + 3 ) [ (1 + 2 )] 3 2r 2r 0 + 2 r r (9.89) And the diusion process can be achieved if that the time are large enough with respect to the characteristics diusion time relative to point r. When the boundary conditions for r = r0 in uid pressure and temperature which are p = p1 and T = T1 apply for the whole model, the following equations correspond to the undrained solution of the instantaneous drilling of a borehole in an innite elastic medium. r = 0 + p1 r0 3 4 r2 rr = p1 r0 3 r0 3 + 0 (1 + 3 ) 3 r r (9.90)
Version: April 29, 2010, 9:23
(9.87)
rr = p1
(9.88)
= = p1
= = p1
Jeremi et al. c
r0 3 r0 3 + 0 (1 + 3 ) 2r3 2r
192
Discussion of the Results The model is constructed as a quarter of a half ball. The cavity radius is 10cm. As the outside boundary is xed, to minimize the possibility of the sudden increase of the uid bulk modulus, the outside radius of the sphere is set to be 2 m. The nal mesh is generated as Fig.9.8. And the following boundary conditions apply: The nodes on XZ and Y Z plane are xed for solid and uid displacement in Y and X direction respectively; the vertical solid and uid displacement for the nodes on the XY plane are suppressed; for the nodes along the outside surface, to satisfy the complementary conditions, all the solid and uid displacements are set to be zero as well. The pressure is translated in to nodal forces and applied in the radial direction. For simplicity, the hydrostatic stress 0 is equal to zero and with the assumption of thermodynamic equilibrium through the process, the temperature factor can be neglected. Also the initial uid pressure p0 is set to be 0 kPa. The analytical solution is studied below using the following set of parameters shown in Table 9.2. (The input les for the above examples are: ValidationExamples/SphericalCavity/Drained ValidationExamples/SphericalCavity/Undrained
Figure 9.8: The mesh generation for the study of spherical cavity As the same procedure in the previous drilling of borehole problem, we compared both the drained and and k = 3.6 108 m/s respectively. In drained behavior, we can see along the cavity surface, the experimental the decrease of the radial displacement for close solution is much smaller that of the experimental results. For the undrained behavior, we can see the radial displacement of the experimental results are always smaller than the close solution. Again it should be noted that the close solutions for the drained and undrained behavior along the cavity surface are exactly the same. This can be explained in the same way as the previous drilling of the borehole problem. When the experimental results from drained behavior are compared with the analytical undrained solution, it is observed they provide good agreement to each other as well.
Jeremi et al. c Version: April 29, 2010, 9:23
undrained behavior. The drained and undrained behavior are tested by the permeability of k = 3.6 104 m/s
result of the radial displacement match the analytical solution very well. While with the increase of the radius,
193 Symbol E Ks Kf s f n Value 0.2 1.2E+6 3.6E+7 1.0E+17 2.7 1.0 0.4 Units kN/m2 kN/m2 kN/m2 ton/m3 ton/m3 -
Parameter Poisson Ratio Youngs Modulus Solid Bulk Modulus Fluid Bulk Modulus Solid Density Fluid Density Porosity
x 10
r=0.1m, close solution r=0.1m, experimental r=0.15m, close solution r=0.15m, experimental r=0.25m, close solution r=0.25m, experimental
10
20
30
40
50 60 Pressure (kPa)
70
80
90
100
Figure 9.9: The comparison of radial solid displacement between analytical solution and experimental result for drained behavior
9.3.3
The Problem The consolidate process can be dened as follows: When a soil layer is subjected to an external loading, immediately the water will alone sustain this load and cause the build-up the excessive pore water pressure. In the progress of the ow of the water to the surface, the load is gradually transferred to the soil skeleton and the excessive pore water pressure will dissipate. At the same time, the settlement of the soil layer occurs. As settlement is usually a major concern in geotechnical engineering, this is a key problem in soil mechanics. Consider a soil layer composed of an isotropic, homogeneous and saturated thermoporoelastic material. The layer has a thickness of h in the Oy direction and of innite extent in the two other directions Ox and Oy. The
Jeremi et al. c Version: April 29, 2010, 9:23
194
x 10
r=0.1m, close solution r=0.1m, experimental r=0.15m, close solution r=0.15m, experimental r=0.25m, close solution r=0.25m, experimental
5 6 Pressure (kPa)
10 x 10
4
Figure 9.10: The comparison of radial solid displacement between analytical solution and experimental result for undrained behavior layer is underlain by a rigid and impervious base at y = 0. And the top surface at y = h is so perfectly drained that the pore pressure is held constant as zero. At the initial state of the soil layer, the thermal eects are neglected so that the boundary conditions follow that:
t0
y=h y=0
p=0 p =0 z
(9.91)
At time t = 0, tan instantaneous vertical load ey is suddenly applied on the top surface y = h, the induced
t>0
y=h
ey = ey
(9.92)
y=0
Jeremi et al. c
=0
(9.93)
Version: April 29, 2010, 9:23
195
x 10
r=0.1m, undrained(c) r=0.1m, drained(e) r=0.15m, undrained(c) r=0.15m, drained(e) r=0.25m, undrained(c) r=0.25m, drained(e)
5 6 Pressure (kPa)
10 x 10
4
Figure 9.11: The comparison of radial solid displacement between analytical solution for undrained behavior and experimental result for drained behavior
Jeremi et al. c
196
y=0
w ey = wy = 0
(9.94)
The problem is then to determine the new elds of uid pressure, stress and displacement induced by the external loading. Analytical Solution Since this is a one-dimensional problem, the only non-zero displacement is the vertical displacement y . But in particular the uid pressure depends only on y and t.
= y (y, t)ey
p = p(y, t)
(9.95)
yy = (0 + 2)
y + bp y (9.96)
xx = zz =
2b 0 yy p ( + 2) 0 + 2
And because the uid pressure p must be an ordinary function of time t, although the derivative of p is innite at time t = 0 according to the consolidation equation (Coussy, 1995), the discontinuity of the uid pressure p at time t = 0 must satisfy
p(y, t = 0+ ) =
0 (1 )(1 20 )b
(9.97)
where and 0 are the drained and undrained Poisson ratio, respectively. For time t > 0, the vertical stress yy = is constant in time and space, therefore the diusion equation reads
t>0
cm
2 p= p 2 y t
(9.98)
p(y, t) =
Each term of the series decreases exponentially with respect to the ratio consolidation time
Jeremi et al. c
197
h2 cm
cm = kM
0 + 2 + 2
(9.100)
where and 0 are the drained and undrained Lame coecient, respectively. Given by the Eqn.9.96, the only non-zero displacement y satises
y 1 = (yy + bp) y 0 + 2
(9.101)
By substituting the value of of the vertical stress and expression of (9.101), the series converges and it can integrated term by term yielding
y (y, t) =
s(t) = s + (s0+ s ) h + 2
n=0
2 (2n
s0+ =
s =
h 0 + 2
(9.103)
Discussion of the Results Twenty 1D nite elements are used to model the horizontal layer. The height of the soil column is 20 m and the height of each element is 1 m. The material properties, shown in Table 9.3, are chosen as representative values for natural soil deposit. A uniform vertical pressure of 200 kPa is applied on the top surface of the soil column. The following boundary conditions apply: As the bottom of the soil column is modelled as an undeformable and impermeable layer, both the solid and uid displacements are xed. The pore pressure is kept constant as zero at the top surface of the soil column because of the perfectly drained condition. For the reason of 1D consolidation problem, all the lateral movement of the solid and uid phase are suppressed so that the vertical displacement is the only non-zero displacement for the intermediate nodes. To capture both the long term(t >0.1 sec) and short term (t <0.1 sec) response of the soil column, two dierent time steps are adopted: 0.1 sec and 0.005 sec. In order to observe the dissipation of the excessive pore water pressure in a reasonable and convenient period, we select k = 3.6 104 m/s as the value for the permeability. To cure the articial oscillation, some numerical damping is introduced into the analysis by using = 0.6 and = 0.3025 in Based on the above parameters, the other relative parameters can be calculated as follows: The bulk modulus of the mixture: K= E = 6.67 105 kP a 3(1 2)
Version: April 29, 2010, 9:23
Jeremi et al. c
198 Symbol E Ks Kf s f n Value 0.2 1.2E+6 3.6E+7 1.0E+17 2.7 1.0 0.4 Units kN/m2 kN/m2 kN/m2 ton/m3 ton/m3 -
Parameter Poisson Ratio Youngs Modulus Solid Bulk Modulus Fluid Bulk Modulus Solid Density Fluid Density Porosity
The undrained bulk modulus of the mixture: N= Kf N Ks = 6.19 107 kP a M = = 6.19 107 kP a bn Kf + N n
h2 = 8.04s cf
In Fig.9.13, the normalized uid pressure p is plotted against the location z for various normalized times t. For normalized time t = 0.001 (natural time t=0.008 sec), only the nodes close to the top free ow surface display the dissipation of the pore pressure, for normalized depth z < 0.15, the pore pressure equals to the applied external pressure. The experimental result provide good agreement with the analytical solution except at the end of the curve, the reason could be the small time step we adopt in this analysis to show the undrained behavior. With the increase of the normalized time, we can clearly see the tendency of the dissipation of the water. At normalized time t = 1.0 (natural time t= 8 sec), the maximum normalized pore pressure is only about 0.11. We can say that our model can eectively show the process of the dissipation of the pore pressure. Now let us turn to the settlement(see Fig.9.14), the experimental results match the analytical solution very well, and at natural time t= 20 sec, the soil layer almost nish the primary consolidation and reach the primary settlement. (The input les for the above examples are: ValidationExamples/Consolidation
Jeremi et al. c Version: April 29, 2010, 9:23
199
Normalized Pressure 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.1
0.2
Normalized Time = 0.001(c) Normalized Time = 0.001(e) Normalized Time = 0.1(c) Normalized Time = 0.1(e) Normalized Time = 0.5(c) Normalized Time = 0.5(e) Normalized Time = 1.0(c) Normalized Time = 1.0(e)
0.4
0.5
0.6
0.7
0.8
0.9
Figure 9.13: The comparison between analytical solution and experimental results for the normalized p during the = cf t/h2 consolidation process against normalized depth z = z/h for various normalized t
3
x 10
0.5
1 Settlement (m)
1.5
2.5
10 Time (sec)
12
14
16
18
20
Figure 9.14: The comparison between analytical solution and experimental results for the settlement
Jeremi et al. c Version: April 29, 2010, 9:23
200
9.3.4
The Problem Liquid water is usually injected into a reservoir from a primary well in order to recover the oil from a secondary well in petroleum engineering. This induces a problem of injecting a uid into a cylindrical well of negligible dimensions. Consider a reservoir of innite extent composed of an isotropic, homogeneous and saturated poroelastic material. Through a cylindrical well of negligible dimensions, the injection of the same uid is performed in all directions orthogonal to the well axis forming the Oz axis of coordinates. As a result of the axisymmetry and cylindrically innite, all quantities spatially depends on r only. The injection starts at time t = and stops at time t = . The ow rate of uid mass injection is constant and equal to q.As a nite amount of of uid mass is injected instantaneously(i.e. q as 0). Analytical Solution This is a problem of cylindrically symmetry. Consequently the cylindrical coordinates(r,,z)is adopted. The vector of relative ow of uid mass w reads
w = w(r, t)er where er is the unit vector along the radius. Using the uid mass balance relationship, it yields
(9.104)
r 0
r, t
(9.105)
In addition, we require the uid ow to reduce to zero innitely far from the well
(9.106)
p=
4f l kt 0
exp(
r =
bM 2f l ( + 2)r 0
Using the constitutive equation, the stress eld can be derived as follows:
rr = 2 zz =
Jeremi et al. c
r r
= 2
r 2b p r 0 + 2 (9.108)
Version: April 29, 2010, 9:23
2b p 0 + 2mu
201
Discussion of the Results As a result of axisymmetry, the model can be constructed as a quarter of a pie. The radius of the pie is 1 m and the thickness of the pie is 5 cm. A cylindrical well is drilled at the center of the pie, and its radius is 1 cm, which can be neglected in dimension when compared with the whole pie. The nal mesh is shown as Fig.9.15. And the boundary conditions is as follows: As a consequence of plain strain problem, all the movements for solid and uid in vertical direction Oz are suppressed; the solid and uid displacement for the nodes along the X axis and Y axis are xed in Y and X direction respectively for the reason of axisymmetry; the nodes along the outside perimeter are xed in the solid and uid displacement with the assumption of innite medium. To the dierence with the previous problems, the traction boundary conditions are applied on the uid displacement. It should be noted that the mention in the above equations is the volume of the uid injected per unit of vertical well length and has a unit of m3 /m. In order to generate the volume of 1 cm3 /m, the corresponding uid displacement of the nodes along the well has been calculated and applied as a step function at the time of 0 sec. For simplicity, the initial uid pressure p0 is set to be 0 kPa. The analytical solution is studied below using the following set of parameters shown in Table 9.4. (The input les for the above examples are: ValidationExamples/LineInjection
Figure 9.15: The mesh generation for the study of line injection problem In the analysis, the pore pressure and the radial displacement are studied. The results are recorded from three points at the radius of 10 cm, 50 cm and 85 cm. The close solution and experimental results are shown in Fig.9.22 and Fig.9.23. As the time step is set to be 1 sec, the rst data point starts at the time of 1 sec. From the pore pressure plot we can see that the build-up of the pore pressure reach the peak value of 34 kPa at the radius of 85 cm. With the decrease of the radius, the pore pressure decreases as well. This can be explained by the fact that the closer the point to the injection location, the earlier and the larger load is applied, so the pore pressure dissipates faster. And as time passes by, we can see the pore pressure progressively dissipates and nally almost reaches the same value within the model. The same phenomena can be been observed from the radial solid displacement. The maximum solid displacement occurs at the radius of 85 cm, which means more coupling between the solid and uid phase, as consequence, the pore pressure should have the largest value. This
Jeremi et al. c Version: April 29, 2010, 9:23
202 Symbol E Ks Kf Ku K s f cf n k Value 0.2 1.2E+6 3.6E+7 1.0E+17 6.0E+7 6.7E+5 2.7 1.0 0.4973 0.4 3.6E-6 Units kN/m2 kN/m2 kN/m2 kN/m2 kN/m2 ton/m3 ton/m3 m2 /s m/s
Parameter Poisson Ratio Youngs Modulus Solid Bulk Modulus Fluid Bulk Modulus Undrained Bulk Modulus Bulk Modulus Solid Density Fluid Density Fluid Diusivity coecient Porosity Permeability
Table 9.4: Material Properties used to study the line injection problem
corresponds to the previous result. With the increase of the time, the radial solid displacement get closer to zero, which means the uid moves out the solid skeleton.
35
30
Radius = 10cm(c) Radius = 10cm(e) Radius = 50cm(c) Radius = 50cm(e) Radius = 85cm(c) Radius = 85cm(e)
20
15
10
10
15 Time (sec)
20
25
30
Figure 9.16: The comparison between analytical solution and experimental result for pore pressure
Jeremi et al. c
203
x 10
Radius = 10cm(c) Radius = 10cm(e) Radius = 50cm(c) Radius = 50cm(e) Radius = 85cm(c) Radius = 85cm(e)
10
15 Time (sec)
20
25
30
Figure 9.17: The comparison between analytical solution and experimental result for radial displacement
x 10
Radius = 10cm(c) Radius = 10cm(e) Radius = 50cm(c) Radius = 50cm(e) Radius = 85cm(c) Radius = 85cm(e)
10
15 Time (sec)
20
25
30
Figure 9.18: The comparison between analytical solution and experimental result for radial displacement
Jeremi et al. c
204
Table 9.5: Simulation parameters used for the shock wave propagation verication problem. Parameter Poisson ratio Youngs modulus Solid particle bulk modulus Fluid bulk modulus Solid density Fluid density Porosity Newmark parameter Symbol E Ks Kf s f n Value 0.3 1.2 106 kN/m2 3.6 107 kN/m2 2700 kg/m3 1000 kg/m3 0.4 0.6
9.3.5
In order to verify the dynamic behavior of the system, an analytic solution developed by Gajo (1995) and Gajo and Mongiovi (1995) for 1D shock wave propagation in elastic porous medium was used. A model was developed consisting of 1000 eight node brick elements, with boundary conditions that mimic 1D behavior. In particular, no displacement of solid (ux = 0, uy = 0) and uid (Ux = 0, Uy = 0) in x and y directions is allowed along the height of the model. Bottom nodes have full xity for solid (ui = 0) and uid (Ui = 0) displacements while all the nodes above base are free to move in z direction for both solid and uid. Pore uid pressures are free to develop along the model. Loads to the model consist of a unit step function (Heaviside) applied as (compressive) displacements to both solid and uid phases of the model, with an amplitude of 0.001 cm. The upU model dynamic system of equations was integrated using Newmark algorithm (see section ??). Table 9.5 gives relevant parameters for this verication. Two set of permeability of material were used in our verication. The rst model had permeability set k = 106 cm/s which creates very high coupling between porous solid and pore uid. The second model had permeability set to k = 102 cm/s which, on the other hand creates a low coupling between porous solid and pore uid. Comparison of simulations and the analytical solution are presented in Figure 9.19.
9.4
9.4.1
u-p Formulation
Governing Equations of Porous Media
The formulation given here is based on Zienkiewicz et al. (1999b). The rst governing equation of porous media is total momentum balance equation: ij,j i + bi = 0 u
where ij = ij pij and = (1 n)s + nf .
(9.109)
The second governing equation is the uid mass balance equation: (kij (p,j + f bj )),i + ui,i + p =0 Qsf (9.110)
Version: April 29, 2010, 9:23
Jeremi et al. c
205
1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 K=10 cm/s, FEM 6 K=10 cm/s, Closed Form 2 K=10 cm/s, FEM 2 K=10 cm/s, Closed Form 4 6 8 10 time (sec) 12 14
6
1.6 1.4 Solid Displ. (x10 cm) 1.2 1 0.8 0.6 0.4 0.2 0 4 6 8 10 time (sec) 12 14 Fluid Displ. (x103cm)
kij kij = gf f
Figure 9.19: Compressional wave in both solid and uid, comparison with closed form solution. where kij = (9.111)
and kij is the permeability in Darcys law with the same unit as velocity.
Qsf =
Ks Kf Ks + K f
(9.112)
is the total compression modulus, Ks and Kf are the solid and uid compression modulus respectively. The boundary conditions are ij nj = ti ui = ui ni wi = ni kij (p,j + f bj ) = w = q p=p where w is the outow and q is the inux. on on on on = t = u = w = p (9.113) (9.114) (9.115) (9.116)
9.4.2
The solid displacement ui and the pore pressure p can be approximated using shape functions and nodal values: ui p
u = NK uKi
(9.117) (9.118)
p NL pL
206
Numerical solution of the total momentum balance The numerical solution of the total momentum balance is
u NK (ij,j i + bi ) d = 0 u
(9.119)
=
t
= = = =
u NK,j ij d +
u NK,j Dijml ml d
+[
p u NK,i NN d]N p
+[
p u NK,i NN d]N p
u = f1 (K ep ) + Qp u
(9.120)
= = [
(9.122)
(9.123)
or
u u M u Qp + (K ep ) = f1 + f2 = f u u
(9.124)
Numerical solution of the uid mass balance The numerical solution of the uid mass balance is
p NM
p Qsf
d = 0
(9.125)
Version: April 29, 2010, 9:23
Jeremi et al. c
207
(9.126)
=
w
=
w
= =
p (f1 )M p (f1 )M
+[
(9.127)
= =
p u NM NL,j d]uLj
p d Qsf
= =
(9.129)
(9.130)
or
p p H p + QT u + S p = f1 + f2 = f p
(9.131)
Matrix form of the governing equations Combine equation (9.123) and (9.130), we obtain uLj 0 0 MKiLj 0 + QLjM SM N pN 0 0 (K ep )KiLj QKiN uLj = + 0 HM N pN
uLj pN
u fKi p fM
Q H u p fu fp (9.132)
(9.133)
208
where fu fp and fu 1 fu 2 fp 1 fp 2 M Q S H
u (f1 )Ki = u (f2 )Ki = p (f1 )M = p (f2 )M = u NK ti d u NK bi d p NM wd
(9.134) (9.135)
MKiLj = ij
QKiN =
p u NK,i NN d
SM N =
p NM
1 N p d Qsf N
HM N =
Jeremi et al. c
Chapter 10
EarthquakeSoilStructure Interaction
(2002)
(In collaboration with Dr. Matthias Preisig and Dr. Guanzhou Jie)
10.1
Current design practice for structures subject to earthquake loading regards dynamic SFSI to be mainly benecial to the behavior of structures (Jeremi and Preisig, 2005). Including the exibility of the foundation reduces the c overall stiness of a system and therefore reduces peak loads caused by a given ground motion. Even if this is true in most cases there is the possibility of resonance occurring as a result of a shift of the natural frequencies of the SFS-system. This can lead to large inertial forces acting on a structure. As a result of these large inertial forces caused by the structure oscillating in its natural frequency the structure as well as the soil surrounding the foundation can undergo plastic deformations. This in turn further modies the overall stiness of the SFS-system and makes any prediction on the behavior very dicult. Dynamic SFSI also becomes important in the design of large infrastructure projects. As authorities and insurance companies try to introduce the concept of performance based design to the engineering community more sophisticated models are needed in order to obtain the engineering demand parameters (EDPs). A good numerical model of a soil-foundation-structure system can therefore not only prevent the collapse or damage of a structure but also help to save money by optimizing the design to withstand an earthquake with a certain return period. A variety of methods of dierent levels of complexity are currently being used by engineers. In the following an overview over the most important ones is presented. A more thorough discussion on methods and specic aspects of dynamic SFSI is available in Wolf (1985) and more recently in Wolf and Song (2002). No SFSI The ground motion is applied directly to the base of the building. Alternatively, instead of applying the ground motion directly to the base of the structure, eective earthquake forces proportional to the base 209
210
acceleration can be applied to the nodes. This procedure is reasonable only for exible structures on very sti soil or rock. In this case the displacement of the ground doesnt get modied by the presence of the structure. For stier structures on soil the ground motion has to be applied to the soil. The model has to incorporate propagation of the motion through the soil, its interaction with the structure and the radiation away from the structure. Direct methods Direct methods treat the SFS-system as a whole. The numerical model incorporates the spatial discretization of the structure and the soil. The analysis of the entire system is carried out in one step. This method provides most generality as it is capable of incorporating all nonlinear behavior of the structure, the soil and also the interface between those two (sliding, uplift). Substructure methods Substructure methods refer to the principle of superposition. The SFS-system is generally subdivided into a structure part and a soil part. Both substructures can be analyzed separately and the total displacement can be obtained by adding the contributions at the nodes on the interface. This method reduces the size of the problem considerably. As the time needed for an analysis doesnt increase linearly with an increasing number of equations the substructure method is much faster than the direct method. The biggest drawback of the method however is the fact that linearity is assumed. For nonlinear systems the substructure method cannot be used. For the direct method dierent levels of sophistication are possible: Foundation stiness approach The behavior of the soil is accounted for by simple mechanical elements such as springs, masses and dash pots. Dierent congurations of the subsoil can be taken into account by connecting several springs, masses and dash pots whose parameters have been determined by a curve tting procedure Wolf (1994). This approach is very popular among structural engineers as it is relatively easy to be integrated in a commonly used nite element code. Other methods use frequency dependent springs and dash pots and therefore require an analysis in frequency domain. Relatively complex congurations of layered subsoil and embedded foundations can be modeled with good accuracy by replacing the (elastic) soil with a sequence of conical rods Wolf and Song (2002) and Wolf and Preisig (2003). p-y methods Attempts have been made to apply the static p-y approach for evaluating lateral loading on pile foundations to dynamic problems. E. and H. (2002) lists several references and provides a parametric study of single piles and pile groups in dierent soil types under simplied loading cases. Even if p-y curves are widely used for estimating lateral loading on piles they are rarely used in full dynamic soil-structure interaction analysis. Current work trying to implement these methods into nite element codes is likely to make them more popular with the engineering community.
Jeremi et al. c Version: April 29, 2010, 9:23
211
Full 3d Full nonlinear three-dimensional modeling of dynamic soil-foundation-structure interaction can be regarded as the brute force approach. Displacements and forces can be obtained not only for the structure as in the above mentioned methods but also for the soil. In spite of the computational resources and modeling eort required for an analysis it is the only method that remains valid for all kinds of problems involving material nonlinearities, contact problems, dierent loading cases and complex geometries.
10.2
Introduction
The Domain Reduction Method (DRM) was developed recently by Bielak et al. (2003); Yoshimura et al. (2003)). It is a modular, two-step dynamic procedure aimed at reducing the large computational domain to a more manageable size. The method was developed with earthquake ground motions in mind, with the main idea to replace the force couples at the fault with their counterpart acting on a continuous surface surrounding local feature of interest. The local feature can be any geologic or man made object that constitutes a dierence from the simplied large domain for which displacements and accelerations are easier to obtain. The DRM is applicable to a much wider range of problems. It is essentially a variant of globallocal set of methods and as formulated can be used for any problems where the local feature can be bounded by a continuous surface (that can be closed or not). The local feature in general can represent a soilfoundationstructure system (bridge, building, dam, tunnel...), or it can be a crack in large domain, or some other type of inhomogeneity that is fairly small compared to the size of domain where it is found. In what follow, the DRM is developed in a somewhat dierent way than it was done in original papers by Bielak et al. (2003); Yoshimura et al. (2003)). The main features of the DRM are then analyzed and appropriate practical modeling issues addressed.
10.3
A large physical domain is to be analyzed for dynamic behavior. The source of disturbance is a known time history of a force eld Pe (t). That source of loading is far away from a local feature which is dynamically excited by Pe (t) (see Figure 10.1). The system can be quite large, for example earthquake hypocenter can be many kilometers away from the local feature of interest. Similarly, the small local feature in a machine part can be many centimeters away from the source of dynamic loading which inuences this local feature. In this sense the term large domain is relative to the size of the local feature and the distance to the dynamic forcing source. It would be benecial not to analyze the complete system, as we are only interested in the behavior of the local feature and its immediate surrounding, and can almost neglect the domain outside of some relatively close boundaries. In order to do this, we need to somehow transfer the loading from the source to the immediate vicinity of the local feature. For example we can try to reduce the size of the domain to a much smaller model bounded by surface as shown in Figure 10.1. In doing so we must ensure that the dynamic forces Pe (t) are appropriately propagated to the much smaller model boundaries .
Jeremi et al. c Version: April 29, 2010, 9:23
212
ui
ub
P (t) e
ue
Figure 10.1: Large physical domain with the source of load Pe (t) and the local feature (in this case a soil foundationbuilding system.
10.3.1
Method Formulation
In order to appropriately propagate dynamic forces Pe (t) one actually has to solve the large scale problem which will include the eects of the local feature. Most of the time this is impossible as it involves all the complexities of large scale computations and relatively small local feature. Besides, the main goal of presented developments is to somehow reduce the large scale domain as to be able to analyze in details behavior of the local feature. In order to propagate consistently the dynamic forces Pe (t) we will make a simplication in that we will replace a local feature with a simpler domain that is much easer to be analyzed. That is, we replace the local feature (bridge, building. tunnel, crack) with a much simpler geometry and material. For example, Figure 10.2 shows a simplied model, without a foundationbuilding system. The idea is to simplify the model so that it is much easier
u0 i
u0 b u0 e
P (t) e
Figure 10.2: Simplied large physical domain with the source of load Pe (t) and without the local feature (in this case a soilfoundationbuilding system. Instead of the local feature, the model is simplied so that it is possible to analyze it and simulate the dynamic response as to consistently propagate the dynamic forces Pe (t).
to consistently propagate the dynamic forces to the boundary . The notion that it is much easier to propagate those dynamic forces is of course relative. This is still a very complex problem, but at least the inuence of local feature is temporarily taken out. It is convenient to name dierent parts of domain. For example, the domain inside the boundary is named 0 . The rest of the large scale domain, outside boundary , is then named + . The outside domain + is still
Jeremi et al. c Version: April 29, 2010, 9:23
213
the same as in the original model, while the change, simplication, is done on the domain inside boundary . The displacement elds for exterior, boundary and interior of the boundary are ue , ub and ui , on the original domain. The equations of motions for the complete system can be written as M u + K u = Pe (10.1) or if written for each domain (interior, boundary and exterior of ) separately, the equations obtain the following form: M bi 0
Mii Mib + Mbb + Mbb + Meb
0
+ Mbe + Mee
In these equations, the matrices M and K denote mass and stiness matrices respectively; the subscripts i, e, and b refer to nodes in either the interior or exterior domain or on their common boundary; and the superscripts and + refer to the domains over which the various matrices are dened. The previous equation can be separated provided that we maintain the compatibility of displacements and equilibrium. The resulting two equations of motion are Kii Kib ui 0 Mii Mib ui , = + Kbi Kbb ub Pb Mbi Mbb ub
+ Mbb + Meb + Mbe + Mee
ui Kii ub + Kbi ue 0
0
+ Kbe + Kee
ui 0 = (10.2) ub 0 ue Pe
in
(10.3)
and
Compatibility of displacements is maintained automatically since both equations contain boundary displacements ub (on boundary ), while the equilibrium is maintained through actionreaction forces Pb . In order to simplify the problem, the local feature is removed from the interior domain. Thus, the interior domain is signicantly simplied. In other words, the exterior region and the material therein are identical to those of the original problem as the dynamic force source. On the other hand, the interior domain (denoted as 0 ), is simplied, the localized features is removed (as seen in gure 10.2). For this simplied model, the displacement eld (interior, boundary and exterior, respectively) and action
0 reaction forces are denoted by u0 , u0 , u0 and Pb . The entire simplied domain 0 and + is now easier to e i b
u K + b + bb + ue Keb
+ Kbe + Kee
u P b b , = u e Pe
in+
(10.4)
analyze. The equations of motion in + for the auxiliary problem can now be written as: + + + + u0 u0 P 0 Mbb Mbe b Kbb Kbe b b + + = M + M + u0 K + u0 Pe K
eb ee e eb ee e
(10.5)
Since there was no change to the exterior domain + (material, geometry and the dynamic source are still the same) the mass and stiness matrices and the nodal force Pe are the same as in Equations (10.3) and (10.4).
Jeremi et al. c Version: April 29, 2010, 9:23
214
Second part of previous equation (10.5) can be used to obtain the dynamic force Pe as
+ + + + Pe = Meb u0 + Mee u0 + Keb u0 + Kee u0 b e b e
(10.6)
The total displacement, ue , can be expressed as the sum of the free eld u0 (from the background, simplied e model) and the residual eld we (comming from the local feature) as following: ue = u0 + we e (10.7)
It is important to note that this is just a change of variables and not an application of the principle of superposition. The residual displacement eld, we is measured relative to the reference free eld ue . 0 By substituting Equation (10.7) in Equation (10.2) one ui Mii Mib 0 M M + M + M + + ub bb bb be bi 0 + + ue + we Mee 0 Meb obtains:
Kii Kbi Kib + Kbb + Kbb + Keb
0
+ Kbe + Kee
ui ub u0 e
which, after moving the free eld motions Mii Mib 0 M M + M + M + bb bb be bi + + Mee 0 Meb
u0 to the right hand side, becomes e 0 Kib ui Kii + + + Kbi Kbb + Kbb Kbe ub + + Kee 0 Keb we 0
+ we 0 (10.8) 0 Pe
By substituting Equation (10.6) in previous Equation (10.9), the right hand side can now be written as ui Kii ub + Kbi 0 we ui + + = ub Kbe Kbb + Kbb + + we Keb Kee 0 + 0 + 0 Mbe ue Kbe ue + + Meb u0 + Keb u0 b b
Kib
(10.9)
M bi 0
Mii
0
+ Mbe + Mee
(10.10)
The right hand side of equation (10.10) is the dynamically consistent replacement force (so called eective force, P ef f for the dynamic source forces Pe . In other words, the dynamic force Pe was consistently replaced by the eective force P ef f : Pief f 0 ef f ef f + 0 + = P = Mbe ue Kbe u0 P e b ef f + + Meb u0 + Keb u0 b Pe b
(10.11)
Jeremi et al. c
215
10.3.2
Method Discussion
The Equation (10.11) shows that the eective nodal forces P ef f
involve only the sub-matrices Mbe , Kbe , Meb , Keb . These matrices vanish everywhere except the single layer of nite elements in domain + adjacent to . The signicance of this is that the only waveeld (displacements and accelerations) needed to determine eective forces P ef f is that obtained from the simplied (auxiliary) problem at the nodes that lie on and between boundaries and e , as shown in Figure 10.3.
Local feature
u0 i u0 e
e
u0 b
P (t) e
+
u0 e
+
Figure 10.3: DRM: Single layer of elements between and e is used to create P ef f .
Only residual waves outgoing. Another interesting observation is that the solution to problem described in Equation (10.10) comprises full unknowns (displacements and accelerations) inside and on the boundary (ui and ub respectively). On the other hand, the solution for the domain outside single layer of nite elements (outside e ) is obtained for the residual unknown (displacement and accelerations) eld, we only. This residual unknown eld is measured relative to the reference free eld of unknowns (see comments on page 214). That eectively means that the solution to the equation Equation (10.10) outside the boundary e will only contain additional waves eld resulting from the presence of a local feature. This in turn means that if the interest is in behavior of local feature and the surrounding media (all within boundary ) one can neglect the behavior of the full model (outside e in + ) and provide appropriate supports (including xity and damping) at some distance from the boundary e into region + . This is signicant for a number of reasons: large models can be reduced in size to encompass just a few layers of elements outside boundary e smaller then the distance to the dynamic source force Pe (earthquake hypocenter). the residual unknown eld can be monitored and analyzed for information about the dynamic characteristics the solution for we has all the characteristics of the additional wave eld stemming from the local feature. Inside domain can be inelastic. In all the derivations in section 10.3 no restriction was made on the type of material inside the plastic bowl (inside e ). That is, the assumption that the material inside is linear elastic
Jeremi et al. c Version: April 29, 2010, 9:23
(signicant reduction for, say earthquake problems where the size of a local feature is orders of magnitudes
of the local feature. Since the residual wave eld is we is measured relative to the reference free eld ue , 0
216
is not necessary as the DRM is not relying on principle of superposition. The Equation 10.7 was only describing the change of variables, and clearly there was no use of the principle of superposition, which is only valid for linear elastic solids and structures. It is therefor possible to assume that the derivations will still be valid with any type of material (linear or nonlinear, elastic or inelastic) inside e . With this in mind, the DRM becomes a very powerful method for analysis of soilfoundationstructure systems.
10.4
The accuracy of a numerical simulation of dynamic SFSI is controlled by two main parameters: a) the spacing of the nodes of the nite element model (h) and b) the length of the time step t. Assuming that the numerical method converges toward the exact solution as t and h go toward zero the desired accuracy of the solution can be obtained as long as sucient computational resources are available.
10.4.1
Grid Spacing h
In order to represent a traveling wave of a given frequency accurately about 10 nodes per wavelength are required (Bathe and Wilson, 1976; Hughes, 1987; Argyris and Mlejnek, 1991b). Fewer than 10 nodes can lead to numerical damping as the discretization misses certain peaks of the wave. In order to determine the appropriate maximum grid spacing the highest relevant frequency fmax that is present in the model needs to be found by performing a Fourier analysis of the input motion. Typically, for seismic analysis fmax is about 10 Hz. By choosing the wavelength min = v/fmax , where v is the wave velocity, to be represented by 10 nodes the smallest wavelength that can still be captured partially is = 2h, corresponding to a frequency of 5 fmax . The maximum grid spacing should not exceed h v = 10 10 fmax (10.12)
where v is the lowest wave velocity that is of interest in the simulation. Generally this is the shear wave velocity.
10.4.2
The time step t used for numerically solving nonlinear vibration or wave propagation problems has to be limited for two reasons. The stability requirement depends on the numerical procedure in use and is usually formulated in the form t/Tn < value. Tn denotes the smallest fundamental period of the system. Similar to the spatial discretization Tn needs to be represented by about 10 time steps. While the accuracy requirement provides a measure on which higher modes of vibration are represented with sucient accuracy, the stability criterion needs to be satised for all modes. If the stability criterion is not satised for all modes of vibration, then the solution may diverge. In many cases it is necessary to provide an upper bound to the frequencies that are present in a system by including frequency dependent damping to the model. The second stability criterion results from the nature of the nite element method. As a wave front progresses in space it reaches one point after the other. If the time step in the nite element analysis is too large the wave front can reach two consecutive elements at the same moment. This would violate a fundamental property of
Jeremi et al. c Version: April 29, 2010, 9:23
217
wave propagation and can lead to instability. The time step therefore needs to be limited to t < h v (10.13)
10.4.3
If nonlinear material models are used the considerations for stability and accuracy as stated above dont necessarily remain valid. Especially modal considerations need to be examined further for these cases. It is however save to assume that the natural frequencies decrease as plastic deformations occur. The minimum time step required to represent the natural frequencies of the dynamic system can therefore taken to be the same as in an elastic analysis.
10 8 6 Acceleration [m/s2] 4 2 0 2 4 6 8 Linear algorithm NewtonRaphson algorithm
A high frequency component is introduced due to plastic slip and counter balancing of the resulting displacement. This is especially true if a linear algorithm with no iterations within one time step is used. Figure 10.4 shows a part of an acceleration time history from an analysis involving elastic-plastic material. It can be seen that the out-of-balance forces at the end of a time step can be quite large if a linear algorithm is used. While the Newton-Raphson algorithm minimizes out-of-balance forces within one time step the linear algorithm requires several time steps to return to a stable equilibrium path. The frequencies corresponding to these peaks are typically of the order of 1 /(a few t). Normally the time step is small enough so that these frequencies dont interfere with the input motion. They can be prevented from propagating through the model by an appropriate choice of algorithmic or material damping. For stability the time step used in a nonlinear analysis needs to be smaller than in a linear elastic analysis. By how much it has to be reduced is dicult to predict as this depends on many factors such as the material model, the applied loading or the numerical method itself. Argyris and Mlejnek (1991a) suggest the time step to be reduced by 60% or more compared to the time step used in an elastic analysis. The best way to determine whether the time step is appropriate for a given analysis consists in running a second analysis with a reduced time step.
Jeremi et al. c Version: April 29, 2010, 9:23
218
10.5
Domain Boundaries
One of the biggest problems in dynamic SFSI in innite media is related to the modeling of domain boundaries. Because of limited computational resources the computational domain needs to be kept small enough so that it can be analyzed in a reasonable amount of time. By limiting the domain however an articial boundary is introduced. As an accurate representation of the soil-structure system this boundary has to absorb all outgoing waves and reect no waves back into the computational domain. The most commonly used types of domain boundaries are presented in the following: Fixed or free By xing all degrees of freedom on the domain boundaries any radiation of energy away from the structure is made impossible. Waves are fully reected and resonance frequencies can appear that dont exist in reality. The same happens if the degrees of freedom on a boundary are left free, as at the surface of the soil. A combination of free and fully xed boundaries should be chosen only if the entire model is large enough and if material damping of the soil prevents reected waves to propagate back to the structure. Absorbing Lysmer Boundaries A way to eliminate waves propagating outward from the structure is to use Lysmer boundaries. This method is relatively easy to implement in a nite element code as it consists of simply connecting dash pots to all degrees of freedom of the boundary nodes and xing them on the other end (Figure 10.5).
1 0 1 0 1 0 1 0
Cs Cp
11 00 11 00 11 00 11 00 11 00 11 00
Cs Cs
Cs
1 0 1 0 1 0 1 0
Cp Cs
Cs
11 00 11 00 11 00 11 00
1 0 1 0 1 0
Cp
11 00 11 00 11 00 11 00 11 00 11 00
111 000 1 0 1 0 1 0
Cp Cs
Cs
11 00 11 00 11 00 11 00
Figure 10.5: Absorbing boundary consisting of dash pots connected to each degree of freedom of a boundary node
Lysmer boundaries are derived for an elastic wave propagation problem in a one-dimensional semi-innite bar. It can be shown that in this case a dash pot specied appropriately has the same dynamic properties as the bar extending to innity (Wolf, 1988). The damping coecient C of the dash pot equals C = Ac
Jeremi et al. c
(10.14)
Version: April 29, 2010, 9:23
219
where A is the section of the bar, is the mass density and c the wave velocity that has to be selected according to the type of wave that has to be absorbed (shear wave velocity cs or compressional wave velocity cp ). In a 3d or 2d model the angle of incidence of a wave reaching a boundary can vary from almost 0 up to nearly 180 . The Lysmer boundary is able to absorb completely only those under an angle of incidence of 90 . Even with this type of absorbing boundary a large number of reected waves are still present in the domain. By increasing the size of the computational domain the angles of incidence on the boundary can be brought closer to 90 and the amount of energy reected can be reduced. In Section 10.7.1 the Lysmer boundary is tested on a two-dimensional model. More sophisticated boundaries modeling wave propagation toward innity For a spherical cavity involving only waves propagating in radial direction a closed form solution for radiation toward innity, analogous to the Lysmer boundary for wave propagation in a prismatic rod, exists (Sections 3.1.2 and 3.1.3 in Wolf (1988)). Since this solution, in contrast to the Lysmer boundary, includes radiation damping it can be thought of as an ecient way of eliminating reections on a semi-spherical boundary surrounding the computational domain. More generality in terms of absorption properties and geometry of the boundary are provided by the various boundary element methods (BEM) available in the literature.
10.6
10.6.1
The model used for verication of the DRM consists of 13 8-node brick elements (Figure 10.6). Each element
Model 1 (FreeField)
Interior Domain
Boundary Element
Boundary Element
Exterior Domain
Jeremi et al. c
220
is a cube of 1 meter side length. The four nodes in a horizontal plane are constrained to move together in all directions. Model 1 represents the free eld whereas Model 2 has a column with a lumped mass attached to its top. The translational degrees of freedom of the base of the beam-column are connected to the nodes at the top of the soil column, the rotational degrees of freedom are xed. The natural period of the pile, assuming it is xed at the soil surface, is 0.5 seconds. As input motion, a sinusoidal horizontal ground motion with a period of 1 second has been applied to the base of the soil column, resulting in an upward propagating shear wave. The following material properties have been used in the analysis: Youngs modulus E = 1944 kP a, Poissons ratio = 0.35 and specic density = 1.8 kg/m3 . The resulting shear wave velocity is 20m/s. The lowest natural frequencies are 3 Hz for model 1 and 0.31 Hz for model 2.
10.6.2
Results
As a rst test Model 2 is analyzed by prescribing the sinusoidal motion to the base of the model. The displacement and acceleration time histories of the nodes of the boundary layer are recorded. In a second analysis these time histories are used to calculate P ef f which then is applied to the nodes of the boundary layer. As can be seen in Figure 10.7 the results are identical in the interior domain. In the exterior domain the result from the DRManalysis shows no motion at all. This is the expected result since the large domain used to obtain ue and ue and the local domain are identical. The residual eld in the exterior domain we therefore vanishes. As a second test the same analysis as above is performed, but this time ue and ue from a free eld analysis have been used to calculate P ef f (Figure 10.8). In the interior domain the results from the DRM- and from the base-shaking analysis are identical at the beginning and start to diverge after about 2 seconds. Since the free-eld motions arent compatible with the soil-structure system, the motions in the exterior domain dont vanish. Those downward propagating waves, which result from interaction with the structure, will not be canceled out by P ef f in the boundary layer and will therefore leave the interior domain. At the base they will eventually be reected and sent upward into the interior domain, where undesired interference occurs. In order to reduce this undesired interference absorbing boundaries of the Lysmer type can be used (Section 10.5). The result of the previous analysis, this time with absorbing boundaries, is given in Figure 10.9. The Lysmer boundary reduces the waves in the exterior domain considerably. The displacement in the exterior domain is now entirely due to the outgoing wave caused by interaction of the free-eld motion with the beam-column at the surface. It can be seen that the residual displacement eld contains the frequency of the sinusoidal input motion and the natural frequency of the SFSI-model (0.31 Hz).
Jeremi et al. c
221
Interior node of boundary layer 1.5 1 xDisp. [m] xDisp. [m] 0.5 0 0.5 1 1.5 0 2 Time [s] 4 6 2 1 0 1 2 3
2 Time [s]
Base, z = 13 m
0.5
2 Time [s]
Jeremi et al. c
222
Interior node of boundary layer 1.5 1 xDisp. [m] xDisp. [m] 0.5 0 0.5 1 1.5 0 2 Time [s] 4 6 3 2 1 0 1 2 3 0
2 Time [s]
Base, z = 13 m
0.5
2 Time [s]
Jeremi et al. c
223
Interior node of boundary layer 1.5 1 xDisp. [m] xDisp. [m] 0.5 0 0.5 1 1.5 0 2 Time [s] 4 6 2 1 0 1 2 3
2 Time [s]
Base, z = 13 m
0.5
2 Time [s]
Figure 10.9: As in Figure 10.8 but with absorbing boundary at the base
Jeremi et al. c
224
10.7
10.7.1
Due to the complexity of full scale nite element models it is helpful to perform preliminary tests on simplied models in order to verify the adequacy of the time and mesh discretization with respect to the input motion. It also provides good insight in the performance of the nonlinear material model. To achieve this a series of tests on a one-dimensional soil column have been proposed: Static pushover test on nonlinear soil column Through the static pushover test the behavior of the nonlinear material model can be veried. Dynamic test of elastic soil column By applying an earthquake motion to the elastic soil column it can be tested whether the selected grid spacing is capable of representing the motion correctly without ltering out any relevant frequencies. This test also allows to choose appropriate damping parameters. It should be noted that this is additional (small) damping that is used for stability of the numerical scheme and should not be relied upon to provide major energy dissipation. Major energy dissipation should be coming from inelastic deformations of the SFS system. Dynamic test of nonlinear soil column Finally the stability and the accuracy of the numerical method can be examined by applying the earthquake motion to the nonlinear column of soil. A second analysis with a time step reduced by 50% should not give a signicantly dierent result. Furthermore it will be examined how propagation through an elastic-plastic material will change the frequency content of the motion.
Model Description The one-dimensional soil column used for verication has the same depth and element sizes as the 2d and 3d models that will be addressed later. Its total depth is 10.5 meters and it consists of a single stack of 8-node brick elements of 1.5 meters side length. In order to achieve one-dimensional wave propagation in vertical direction the movement of four nodes at each level of depth is constrained to be equal. The input motion is applied to the four nodes at the base of the model. As input motions four time histories from the Northridge Earthquake are selected (Figure 10.10). The material properties of the soil are given in Table 10.7.1. Friction angle Undrained shear strength cu Mass density Shear wave velocity vs
Jeremi et al. c
225
LA University Hospital 1 0 1 0 10 20 30 Lake Hughes 40 Fourier Amplitude [g] Fourier Amplitude [g] 0.03 0.02 0.01 0 0.03 0.02 0.01 0 0.04 0.03 0.02 0.01 0 0.06 0.04 0.02 0 0 5 10 15 20 25 Acceleration [g]
10
15
20
25
Acceleration [g]
10
15
20
25
1 0 1 0 10 20 Time [s] 30 40
Acceleration [g]
Acceleration [g]
10 15 Frequency [Hz]
20
25
Figure 10.10: Acceleration time histories and Fourier amplitude spectras of the selected ground motions
The discretization parameters, the time step t and the maximum grid spacing h, are determined following the guidelines outlined in Section 10.4. This yields a maximum grid spacing of h vs 200 = = 2m 10 fmax 10 10 (10.15)
For the following analysis h = 1.5 m is selected. The maximum time step is t 1.5 h = = 0.0075 s vs 200 (10.16)
Taking into account a further reduction of the time step by about 60% due to the use of nonlinear material models t = 0.002 s is chosen. Static Pushover Test on Elastic-Plastic Soil Column For the static pushover test, an elastic perfectly plastic Drucker-Prager material model as specied in Table 10.7.1 is used. After applying self weight a horizontal load of 100 kN is applied to a surface node in increments of 0.1 kN. The system of equations is solved using a full Newton-Raphson algorithm. The predicted shear strength of the
Jeremi et al. c Version: April 29, 2010, 9:23
226
rst element that is expected to fail, the one at the surface, is: f = = = cu + z g tan 10 + 0.75 x 1.8 x 9.81 tan 37 19.98 kPa (10.17)
where z is the depth of the center of the rst element. Self weight produces the following stresses in the element at the surface: x = y z = = 8.83 kPa 13.24 kPa
(10.18)
z x 2
A (10.19)
44.7 kN
The static failure load is underestimated by about 6%. This accuracy is acceptable for the given model because the boundary conditions cannot assure constant stresses at a given depth (no shear stress is applied to the lateral surfaces). Dynamic Test on Elastic Soil Column In order to test the spatial discretization of the model an earthquake motion is propagated through an elastic soil column. The grid spacing of the nite element mesh can be considered suciently ne if frequencies up to fmax = 10 Hz are represented accurately in the numerical analysis. A good way to verify this is to calculate transfer functions between the base and the surface of the soil column. Because transfer functions dont depend on the input motion they can easily be compared with closed form solutions. The transfer function of a soil deposit describes the amplication between the frequencies of the motion at the base and at the soil surface: T F () = u(z = 0, ) u(z = H, ) (10.20)
where z is the depth measured from the surface and H is the thickness of the soil deposit above the bedrock. = 2 f is the circular frequency. For elastic soil with viscous damping the wave equation can be written as (Kramer, 1996) 2u 3u 2u = G 2 + 2 2 t z z t (10.21)
Version: April 29, 2010, 9:23
Jeremi et al. c
227
where is the frequency independent hysteretic material damping. After solving the wave equation the transfer function can be written as T F () = 1 cos H/vs (10.23)
vs =
G =
G(1 + i 2 )
(10.24)
In a nite element model with mass- and stiness proportional Rayleigh damping the damping coecient is constant. Therefore the hysteretic material damping ratio needs to be frequency dependent in order to satisfy Equation 10.22. Solving Equation 10.22 for and substituting it into Equation 10.24 and then into Equation 10.23 yields a new transfer function: T F () = cos H 1 G + i (10.25)
Figure 10.11 shows a comparison between the closed form solution and the numerical transfer functions obtained from the nite element analysis. Rayleigh damping is used to obtain the damping matrix C: C = M + K (10.26)
The analysis are performed using stiness proportional Rayleigh damping of = 0.001 and = 0.01. No mass proportional damping is applied ( = 0). The damping coecients of the closed form solution are chosen to be = G . It can be seen that the numerical transfer functions are very close to the closed form solutions for = G. The peak corresponding to the second natural frequency of the soil layer is slightly shifted to the right in the result of the FE analysis. For the FE analysis the Rayleigh damping cannot be reduced any further as the solution would become unstable. This result proves that a FE analysis involving Rayleigh damping with = 0 and = / G is equivalent to the closed form solution of the wave equation with frequency-dependent hysteretic material damping. Based on the above observations a stiness proportional Rayleigh damping of = 0.01 is selected for the nite element analysis. This choice damps frequencies above 10 Hz appropriately. Dynamic Test on Elastic-Plastic Soil Column As the next step an elastic-plastic material model of Drucker-Prager type with kinematic strain hardening has been selected. Previous analysis involving material with isotropic hardening have proved to be unsuitable because energy can only be dissipated as the yield surface expands. For dynamic problems this can lead to an unreasonably large extension of the yield surface, especially if resonance frequencies are present. Therefore only kinematic hardening has been selected in this analysis.
Jeremi et al. c Version: April 29, 2010, 9:23
228
Solution obtained from FE analysis
15 Stiffness proportional Rayleigh Damping: = 0.001 Stiffness proportional Rayleigh Damping: = 0.01 10
10
15
15 Frequency dependent Hysteretic Material Damping = 0.001*G Frequency dependent Hysteretic Material Damping = 0.01*G No Damping 10
5 Frequency [Hz]
10
15
Figure 10.11: Transfer Function between Surface and Base of Soil Layer
The analysis were performed with four dierent ground motions using time steps of t = 0.002s and t = 0.001s. A linear integrator without iterations within a time step was used. All ground motions were scaled to a maximum acceleration of 1g. For comparison the analysis were also performed on elastic material. Figure 10.12 shows the displacement time histories at the surface for all four ground motions. While the overall shapes of the displacements are the same as for the elastic case there is some residual plastic displacement resulting in the time histories of the Century City motions. The Fourier amplitude spectras of the acceleration recorded at the surface (Figure 10.13) have the same general shape for the case of elastic and elastic-plastic material. The amplication at the rst resonance frequency (f = 4.75Hz) is bigger in the elastic analysis. Higher frequencies resulting from plastic slip are damped out eectively in the nonlinear analysis. Figure 10.14 shows the acceleration time history at the upper node of the lowest element, that is the rst free node above the base. The record shows large peaks of the order of about 6 g. These peaks are caused by plastic slip and counter balancing of the resulting plastic deformation. The periods of the peaks are of the order of a few time steps, they add a very high frequency component to the acceleration. Because these frequencies are due to a purely numerical phenomenon, they should not be allowed to propagate through the model. This can be achieved easily by specifying an appropriate numerical procedure (Newmark with appropriate combination of and ) or with Rayleigh Damping. As for the elastic model transfer functions were also computed for the nonlinear model. In Figure 10.15 the transfer functions between the acceleration at the soil surface and the base are compared. The functions for the
Jeremi et al. c Version: April 29, 2010, 9:23
229
Displacement [m]
0.1 0
Elastic ElasticPlastic
15 Time [s]
20
25
30
Displacement [m]
Displacement [m]
Displacement [m]
10 Time [s]
15
20
25
Figure 10.12: Displacement Time-Histories at surface of 1d Soil Column, elastic and elastic-plastic material
nonlinear model are not smooth anymore but the general shape is the same as for the linear elastic model, i.e. the rst natural frequency of the layer is clearly visible. The peaks that are present in the range of 25 Hz are purely numerical as they appear due to the division by a very small value. A second set of analysis performed with half the time step of the previous analysis gives an idea of the accuracy of the numerical method. In Figure 10.16 the dierence between the displacement (or acceleration) of the analysis with t = 0.002s and t = 0.001s, divided by the corresponding maximum value is given for the entire time history: d = d0.002 (t) d0.001 (t) a0.002 (t) a0.001 (t) and a = max d0.001 max a0.001 (10.27)
In Figure 10.17 an integral measure for the dierence in displacements and accelerations between the two analysis is given for all depths. The integral measures are dened as dif fd = dif fa = 1 1 max |d| T 1 1 max |a| T
T 0 T 0
The integral dierences in accelerations are quite large in the elements that are close to the base, that is where the motion is applied. Toward the surface the dierence becomes smaller than 1%. This is a result of the fact
Jeremi et al. c Version: April 29, 2010, 9:23
230
0.6 0.4 0.2 LA University Hospital 0 0.6 0.4 0.2 Lake Hughes 0 1 0 5 Frequency [Hz] 10 15 0 5 Frequency [Hz] 10 15
0.5
Century City 090 0 0.6 0.4 0.2 Century City 360 0 0 5 Frequency [Hz] 10 15 Elastic ElasticPlastic 0 5 Frequency [Hz] 10 15
that most of the plastic deformation occurs near the base which represents an undesired boundary eect. Again this result underlines the importance of an appropriate choice of the size of the computational domain. With a point wise dierence not exceeding 5% for accelerations and 2% for displacements the time step t = 0.002s is suciently small to ensure stable and accurate results. 2d Model A 2d-model is proposed as a simplication of the full 3d-model. Representing a cross section of the full model it is expected to provide insight into its dynamic behavior while requiring considerably less computational resources. The 2d-model consists of one slice of eight-node brick elements as shown in Figure 10.18. The nodes of the two lateral faces are constrained to move together in x- and z-direction, the out-of-plane displacement in y-direction is xed. The model approximates a plane strain situation. The earthquake motion is applied to the model by the DRM method. Input Motions As input for the 2d model the motion from the Northridge earthquake recorded at LA University Hotel (Figure 10.10) is used. The acceleration time history is scaled to a peak ground acceleration of 1 g. Motion is applied in x-direction only, that is, this is a 1D wave propagation.
Jeremi et al. c Version: April 29, 2010, 9:23
231
60 40 Acceleration [m/s ]
2
Acceleration time histories at all the nodes of the boundary layer are obtained by vertically propagating a plane wave using the program SHAKE91 (Idriss and Sun, 1992). Because the free-eld model has to match the properties of the free eld as represented by the nite element model for the reduced domain, only linear elastic material without strain dependent reduction of shear modulus and a constant amount of hysteretic material damping is used in the SHAKE91-analysis. The earthquake motion obtained in this way corresponds to a shear wave propagating upward through a homogeneous linearly elastic half space. The acceleration time histories from the SHAKE91-analysis are then integrated twice to obtain displacements. Before integration the acceleration and velocity time histories are transformed into Fourier space, multiplied with a high pass lter and transformed back into time domain. Then a simple parabolic baseline correction is performed in order to obtain zero initial, nal and mean values. Boundary Conditions Dierent boundary conditions are tested on the free-eld model. First all outside boundary nodes are fully xed as shown in Figure 10.19 a). Then they are released and attached to dash pots that are both perpendicular and tangential to the boundary (schematically shown in Figure 10.19 b)). The dash pots perpendicular to the boundary are specied to absorb p-waves, those tangential to the boundary to absorb s-waves. Because in this conguration no displacement constraint is imposed to the model on the faces at x = 10.5 meter the horizontal forces in the model with xed boundaries and applying them with opposite sign to the model with absorbing boundaries. The horizontal displacements after applying self weight should be very small. This conguration of boundary conditions has no xed point in x-direction. Because the dash pots only provide
Jeremi et al. c Version: April 29, 2010, 9:23
at-rest soil pressure has to be applied to the corresponding nodes manually. This is done by recording the reaction
232
5 4 TF [] 3 2 1 0 5 4 TF [] 3 2 1 0 5 4 TF [] 3 2 1 0 5 4 TF [] 3 2 1 0 0 5 0 5 0 5 0 5
LA University Hospital
Elastic ElasticPlastic
20
25
20
25
20
25
10 15 Frequency [Hz]
20
25
Figure 10.15: Transfer functions between acceleration at the soil surface and the base
resistance to high velocity motions the model is very sensitive to low frequency components of the motion. The slightest imbalance in acceleration causes the entire model to move as a rigid body in x-direction. To avoid this to happen the node at the center of the base (x = 0.0 m, z = -10.5 m) is fully xed in the following analysis. Figure 10.20 shows results from a free-eld analysis on a homogeneous elastic model. Displacements on an exterior boundary node as well as transfer functions between a point at the surface and a point on the exterior boundary of the plastic bowl are presented for the two congurations of boundary conditions shown in Figure 10.19. It can be seen that the displacements outside the plastic bowl in the model using absorbing boundary conditions are much larger compared to the model with xed boundaries. This result is as expected considering the immediate proximity of the boundary. It also gives an idea about the constraints the xed boundary imposes on the motions. The transfer function in Figure 10.20 b) is dened as the ratio between the Fourier amplitude spectra of a point at the surface and a point on the exterior boundary: T F () = A1 () A2 () (10.31)
where A1 () is the Fourier amplitude spectrum of the acceleration time history at the point (x,z) = (0,0)m and A2 () the corresponding spectrum at the point (x,z) = (9.0,-7.5)m. The gure shows that the large peak representing the rst natural frequency of the system, corresponding to a standing shear wave in a soil layer of 10.5 meter depth, gets reduced considerably by the absorbing boundary. An energy build-up in the model due
Jeremi et al. c Version: April 29, 2010, 9:23
233
5 5 Difference [%]
5 5 Difference [%]
5 5 Difference [%]
Figure 10.16: Dierence between results of analysis with dierent time steps, in percent of the maximum value
to reection of waves on the model boundaries can be reduced eectively with the conguration of boundary conditions shown in Figure 10.19 b). By releasing the xed node at (x,z) = (0,-10.5)m the resonance peak could be reduced by another 10% approximately, however at the cost of remaining permanent displacements at the end of the analysis. Alternatively to imposing a rigid constraint to a single node at the base the model can be prevented to move horizontally as a rigid body through uniaxial springs. This gives the possibility to adjust the frequency of the eigenmode that corresponds to a vertically propagating plane shear wave. By appropriately choosing the spring constants the model can therefore be adjusted in such a way that it represents the natural frequency of a soil deposit on bedrock. Structure Four very simple structures are chosen to illustrate the eects of dynamic SFSI. A beam-column element of length L and moment of inertia Iy is xed to a footing. A lumped mass of M = 100, 000 kg is added to the translational
Jeremi et al. c Version: April 29, 2010, 9:23
234
Displacements
1 Difference [%]
0.5
0 9 7.5
6 4.5
3 1.5
Depth [m]
Century City 360 Lake Hughes Century City 090 LA University Hospital
Accelerations
6 4.5
3 1.5
Depth [m]
Century City 090 Century City 360 Lake Hughes LA University Hospital
Figure 10.17: Averaged dierences between results of analysis with dierent time steps
1.5 m
10.5 m
7 elements
21 m 14 elements
degrees of freedom of the top of the structure. The footing is 0.5 m deep, spans over four soil elements and is rigidly connected to the adjacent soil nodes. Its Youngs modulus is chosen large enough so that the footing
Jeremi et al. c Version: April 29, 2010, 9:23
235
a)
b)
10 Displacement [m]
x 10
b)
10
15 Time [s]
20
25
30
maximum value = 5 5
a)
10 12 Frequency [Hz]
14
16
18
20
Figure 10.20: Elastic homogeneous free-eld model: a) Displacements of an exterior boundary node (x,z) = (9.0,0.0) m, b) Acceleration transfer function between surface and depth
A()1 A()2
can be considered rigid. The mass density of the footing is = 2400 kg/m , the column is considered massless. The moment connection between the nodes of the footing, having 3 (translational) degrees of freedom, and the 6 degrees of freedom of the nodes of the column is assured by a very sti beam element that is connected to a node at the bottom and a node at the top of the footing. The column is then simply connected to the upper node of this auxiliary beam element. The parameters of the four columns are chosen such that the second natural frequency, that is the natural frequency attributed to bending of the column (Figure 10.22 b)), is evenly distributed over the frequency range of the input motion (Figure 10.23). Structure 4 is designed such that its second natural frequency matches the largest spike in the input motion. Table 10.1 lists the properties of the structures used in the analysis. For the nonlinear columns a strain hardening material is chosen that consists of an initial elastic branch with tangent modulus E and a post-yield branch with tangent modulus 0.2 E. The Youngs modulus for all four structures is E = 210 GPa. The yield stress fy for structures 1, 2 and 4 is 20 MPa and for structure 3 it is 2 MPa.
Jeremi et al. c Version: April 29, 2010, 9:23
236
mass
a)
b)
0.04 0.035 Fourier Amplitude [g] 0.03 0.025 0.02 0.015 0.01 0.005 0 0 1 2 3 Frequency [Hz] 4 5 6
f1
n
f4 f2
n n
f3
n
Figure 10.23: Fourier amplitude spectrum of input motion with second natural frequencies of the 4 SFSI-systems
Jeremi et al. c
237 Stiness E Iy [M N m2 ] 1680 5670 13440 5670 Mass [kg] 100,000 100,000 100,000 100,000 Yield Moment [kNm] 800 1,800 320 1,800
Structure 1 2 3 4
Linear structure
0 0
Figure 10.24: Parametric study of 15 linear structures with varying natural frequency.
Structure with Fixed Base To begin with a parametric study of a series of structures with varying stinesses is analyzed. The stiness is varied by changing the width of the column section. The dierent structures are expected to respond specically to the frequency range of the input motion that is in the neighborhood of the natural frequency of the column. The input motion that is applied at the base of the structure has been recorded in a previous free-eld analysis of the 2d-model. The results of this parametric study are shown in Figures 10.24 and 10.25 for linear and nonlinear structures, respectively. The Fourier amplitudes spectras of the acceleration at the top of the structure are plotted for 15 structures with variable natural frequency fn . A line of equal frequency is also provided. The input motion is plotted in the background of the gure. It can be seen that the maxima of the frequency spectras are almost perfectly aligned along the line of equal frequency. This is even more obvious in the case of a linear structure. In
Jeremi et al. c Version: April 29, 2010, 9:23
238
0 0 3.5 f [Hz]
Figure 10.25: Parametric study of 15 nonlinear structures with varying natural frequency
that case the responses of the structures are very narrow banded. As the structure remains elastic the top of the structure oscillates mainly in its initial natural frequency. Lower and higher frequencies are eliminated to a great extent. In the case of the inelastic structure there is clearly more damping and reduction of the responses at some (most) frequencies. The nonlinearity in structure is producing a longer eective period for the structure, and that eective period changes during shaking. This in turn widens the frequency range of structural response. That is, the response is lower, but the frequency characteristic is (much) wider. A series of xed-base analysis is also performed on the four structures mentioned in Section 10.7.1. The rst natural frequencies of the four structures with its base xed, corresponding to the second mode of vibration of the SFSI-model, are given in Table 10.2. It can be seen that the inuence of the soil on the natural frequency of the SFSI-system increases as the overall stiness of the structure increases. Structure 1 2 3 4 1st natural frequency of xed-base system [Hz] 2.71 9.82 25.1 5.75 2nd natural frequency of SFSI-system [Hz] 2.07 3.89 5.53 3.52
Jeremi et al. c
239
Results The results of the SFSI- as well as the xed-base analysis are presented in the following. The displacements at the top of the nonlinear structures are recorded and plotted in Figure 10.26. It can be seen that the results from
Displacement [m] 0.1 0 0.1 0 Displacement [m] Structure 2 5 10 15 Time [s] 20 25 30 Fixed Base SFSI Structure 1 Fixed Base SFSI
0.1 0 0.1 0
5 Structure 3
10
15 Time [s]
20
25
Displacement [m]
0.1 0 0.1 0
5 Structure 4
10
15 Time [s]
20
25
Displacement [m]
0.1 0 0.1 0
10
15 Time [s]
20
25
30
the SFSI- and the xed-base-model dier considerably in terms of maximum as well as permanent displacement. In contrast to this the displacements at the base of the column are almost identical for the two models (results not plotted). Figure 10.27 displays the displacements at the top of structures 1 and 2 for all the combinations of linear and nonlinear soil and structures that have been analyzed. Due to the low yield moment the permanent displacement for structure 1 is relatively large in the analysis involving nonlinear columns. The results involving nonlinear columns on linear and on nonlinear soil are very similar in their overall shape, however permanent deformations are very dierent. It seems that the forces that trigger plastic deformations in the column strongly depend on the behavior of the soil beneath the foundation. In order to investigate the forces causing plastic deformations in the structures we look at the base moments between foundation and column. In Figure 10.28 the moments at the base of the linear structures are plotted. For structures 1 and 4 the moments for the xed-base model are higher than for the SFSI-model. This means that in this case neglecting the eects of SFSI leads to a conservative design. Structures 2 and 3 however have to resist higher moments when SFSI is taken into account. Because the SFSI-system is more exible than the xed-base structure its modes of vibration are excited by a dierent range of frequencies contained in the input
Jeremi et al. c Version: April 29, 2010, 9:23
240
30
Structure 1
25
Structure 2
20
15 Time [s]
Fixed Base nonlinear column SFSI nonlinear soil, linear column SFSI nonlinear soil, nonlinear column SFSI linear soil, nonlinear column
Fixed Base SFSI nonlinear soil, linear column SFSI nonlinear soil, nonlinear column SFSI linear soil, nonlinear column
10
0.1
0.05
0.15
0.05
0.1
0.1
0.1
motion. For a particular motion this can lead to resonance of the SFSI-system. This result is in contradiction with current engineering practice suggesting that neglecting SFSI in general leads to a more conservative design. Figure 10.29 shows the moments at the base of structures 1 to 4, this time for the analysis involving nonlinear column elements. The evolution of the second natural frequency of the SFSI-system is also provided as a qualitative indication for when plastic deformations occur. The base moments for structures 1 and 3 in the xed-base- and the SFSI-analysis are very similar. Due to the low yield moment of the structure no resonance with the input motion occurs as a lot of energy is dissipated through plastic deformation. Figure 10.30 shows an interesting aspect of nonlinear SFSI. In the analysis involving elastic-plastic soil the Fourier amplitudes of the moment at the base of the structure are reduced in the neighborhood of the natural frequency of the system. This is most likely due to dissipation of energy caused by elastic-plastic deformations in the soil that, in their turn, are a result of large loads provoked by resonance between the SFSI-system and the input motion. As a measure of the plastic strain occurring beneath the footing the equivalent plastic strains averaged over
Jeremi et al. c Version: April 29, 2010, 9:23
0.15
0.15
0.05
0.15
Displacement [m]
0.05
10
15 Time [s]
20
25
30
Displacement [m]
241
10 Time [s]
15
20
10 Time [s]
15
20
3 Structure 3 2 Moment [MNm] 1 0 1 2 3 5 10 Time [s] 10 Structure 4 Moment [MNm] 5 0 5 10 Fixed Base SFSI max/min(Fixed Base) max/min(SFSI) 15 20 Fixed Base SFSI max/min(Fixed Base) max/min(SFSI)
10 Time [s]
15
20
all the Gauss points are calculated. The results are given at t = 12 s and at t = 14 s, that is shortly before and after the largest plastic deformation occurs (Figures 10.31 and 10.32). Plastic strains are larger in the analysis involving an elastic structure. This reects the fact that elastic structures dont dissipate any energy by themselves. For structure 2 no signicant dierence can be observed because of its high yield moment. Structure 4 is characterized by the same yield moment, its slightly smaller natural frequency however causes resonance with the input motion which leads to larger plastic strains beneath the footing. The largest plastic strains develop in the layer of elements adjacent to the boundary layer. This can be due to an input motion that isnt fully compatible with the elastic properties of the DRM-model. It should be possible to reduce these undesired plastic strains by either increasing the size of the soil model or by selecting a method to obtain the free-eld motions that represents the soil properties of the DRM-model more closely.
Jeremi et al. c Version: April 29, 2010, 9:23
242
8 6 4 2 0 20 16 12 8 4 0 20 15 12.5 10 7.5 5 2.5 5 10 Time [s] 15 0 20 16 Structure 4 12 8 4 0 20 Natural Frequency [Hz] Fixed Base SFSI max/min(Fixed Base) max/min(SFSI) Natural Frequency [Hz] Structure 3 Fixed Base SFSI max/min(Fixed Base) max/min(SFSI) Natural Frequency [Hz] Structure 2 Fixed Base SFSI max/min(Fixed Base) max/min(SFSI) Natural Frequency [Hz]
10 Time [s]
15
4 Moment [MNm] 2 0 2 4
10 Time [s]
15
10 5 0 5 10
10 Time [s]
15
x 10
0.5
1.5
3.5
4.5
Figure 10.30: Fourier amplitude spectra of moments at the base of nonlinear column, SFSI-analysis
Jeremi et al. c
243
Structure 2 Structure 3 Structure 4
Inelastic Column
Elastic Column
Structure 1
Inelastic Column
Elastic Column
Structure 1
Structure 2
Structure 3
Structure 4
10.7.2
The 2d SFSI-model presented in the previous section is extended to a 3d model in the following. The goal is to show that the considerations for accuracy and stability of the numerical method obtained from the 1d-model remain valid for the 3d-model. Even if the simplicity of the analyzed problem doesnt necessarily justify the additional computational eort it is important to show that it is possible to obtain reliable results for a problem that involves the following elements: 3d model with about 700 elements, 960 nodes and 2700 equations Elastic-plastic soil (Drucker-Prager with kinematic hardening) Nonlinear structure (bilinear material model) Ground motion applied through the Domain Reduction Method (DRM) Absorbing boundary of Lysmer type Description of Model The 3d model is based on the 2d model shown in Figure 10.21. In y-direction 6 more slices of 7 x 14 elements are added (Figure 10.33). The x-z plane at y = 0 represents a plane of symmetry. Lysmer boundaries are attached
Jeremi et al. c Version: April 29, 2010, 9:23
244
10.5 m mass
10.5 m
7 elements
21 m 14 elements y 7 elements
z x
to all outside boundaries with the exception of the plane of symmetry and the soil surface. The main dierence to the 2d model is that 3d wave propagation is possible which leads to higher radiation damping. The structure was chosen to have the same geometric and material properties as Structure 4 in the previous section. Results Some results of the 3d-analysis together with the corresponding data of the 2d analysis are presented in Figure 10.34. Due to limited memory only the rst 20 seconds of the time history were processed. A more ecient implementation of the application of eective forces for the DRM-method inside the nite element code should solve this problem. The analysis took 66 hours to nish. The displacements obtained at the top of the structure as well as the moments at its base are very close to the results of the 2d-analysis. This shows that the analysis provides reliable results for a full 3d nonlinear SFSI problem. The amplitude of the base moment is at several instances larger for the 3d-model than for the 2d-model. This can be explained with the fact that more energy is present in the 3d-model whereas the energy the structure can absorb is the same as in the 2d-model. Also it is obvious that the natural frequencies of the 3d-model are not exactly the same as for the 2d-model and therefore changes the dynamic behavior in a way that is almost impossible to predict beforehand. Because of the simple geometry of the problem the 2d-model is absolutely sucient for analyzing the forces acting on the structure. If one is interested in the stress history in the soil surrounding the footing then the 3d-model can provide valuable additional information.
Jeremi et al. c
245
0.1
10 Time [s]
15
20
25
Figure 10.34: Top: Displacements at the top of Structure 4, Bottom: Moments at the base of Structure 4
Jeremi et al. c
246
Jeremi et al. c
Chapter 11
247
248
Jeremi et al. c
P0 K
Chapter 12
J I
(1994)
Practical Applications
12.1 Consolidation of Clays
H G F E D C
the 1 dimensional model used for consolidation verication is shown in Figure 12.2 below.
The model was veried against closed form solutions for elastic material available in
249
P0
Computational Geomechanics: Lecture Notes
250
K J I H G F E
Figure 12.2: Finite element for 1D consolidation analysis.
D C B
Jeremi et al. c
251
12.2
Jeremi et al. c
252
12.3
(In collaboration with Mr. Matthias Preisig, Mr. Guanzhou Jie and Mr. Kallol Sett)
Jeremi et al. c
253
12.4
Static and Dynamic Behavior of Pile Foundations in Dry and Saturated Soils
(In collaboration with Mr. Guanzhou Jie and Mr. Zhao Cheng)
Jeremi et al. c
254
12.5
(In collaboration with Mr. George Hue and Mr. Guanzhou Jie)
Jeremi et al. c
Bibliography
R. Abraham, J. Marsden, and T. Ratiu. Manifolds, Tensor Analysis, and Applications, volume 75 of Applied Mathematical Sciences. Springer Verlag, second edition, 1988. Working Paper for Draft Proposed International Standard for Information SystemsProgramming Language C++. ANSI/ISO, Washington DC, April 1995. Doc. No. ANSI X3J16/95-0087 ISO WG21/N0687. J. Argyris and H.-P. Mlejnek. Dynamics of Structures. North Holland in USA Elsevier, 1991a. J. Argyris and H.-P. Mlejnek. Dynamics of Structures. North Holland in USA Elsevier, 1991b. P. Armstrong and C. Frederick. A mathematical representation of the multiaxial bauschinger eect. Technical Report RD/B/N/ 731,, C.E.G.B., 1966. A. Baltov and A. Sawczuk. A rule of anisotropic hardening. Acta Mechanica, I(2):8192, 1965. K.-J. Bathe. Finite Element Procedures in Engineering Analysis. Prentice Hall Inc., 1996. ISBN 0-13-301458-4. K.-J. Bathe. Finite Element Procedures in Engineering Analysis. Prentice Hall Inc., 1982. K.-J. Bathe and E. Dvorkin. On the automatic solution of nonlinear nite element equations. Computers & Structures, 17(5-6):871879, 1983. K.-J. Bathe and L. Wilson, Edward. Numerical Methods in Finite Element Analysis. Prentice Hall Inc., 1976. J. Batoz and G. Dhatt. Incremental displacement algorithms for nonlinear problems. International Journal for Numerical Methods in Engineering, 14:12621267, 1979. Short Communications. T. Belytschko, W. K. Liu, and B. Moran, editors. Nonlinear Finite Element for Continua and Structures. John Wiley & Sons Ltd., New York, 2001. P. N. Biani. Exact evaluation of contact stress state in computational elasto plasticity. Engineering Computac c tions, 6:6773, 1989. J. Bielak, K. Loukakis, Y. Hisada, and C. Yoshimura. Domain reduction method for threedimensional earthquake modeling in localized regions. part I: Theory. Bulletin of the Seismological Society of America, 93(2):817824, 2003. 255
256
B. A. Bilby, L. R. T. Gardner, and A. N. Stroh. Continuous distributions of dislocations and the theory of plasticity. In IX e Congr`s International de M`chanique Appliqu`e, volume VIII, pages 3544, Universit` de Bruxelles, 50. e e e e Av`nue Franklin Roosevelt, 1957. e G. Booch. Object Oriented Analysis and Design with Applications. Series in ObjectOriented Software Engineering. Benjamin Cummings, second edition, 1994. R. I. Borja. Cam clay plasticity, part II: Implicit integration of constitutive equation based on a nonlinear elastic stress predictor. Computer Methods in Applied Mechanics and Engineering, 88:225240, 1991. R. I. Borja and S. R. Lee. Camclay plasticity, part I: Implicit integration of elastoplastic constitutive relations. Computer Methods In Applied Mechanics and Engineering, 78:4972, 1990. H. J. Braudel, M. Abouaf, and J. L. Chenot. An implicit and incremental formulation for the solution of elastoplastic problems , by the nite element method. Computers Structures, 22(5):801814, 1986. W. F. Chen and D. J. Han. Plasticity for Structural Engineers. Springer Verlag, 1988. C. H. Choi. Physical and Mathematical Modeling of Coarse-Grained Soils. PhD thesis, Department of Civil and Environmental Engineering, University of Washington, Seattle, Washington, 2004. J. O. Coplien. Advanced C++, Programming Styles and Idioms. Addison Wesley Publishing Company, 1992. O. Coussy. Mechanics of Porous Continua. John Wiley and Sons, 1995. ISBN 471 95267 2. M. A. Criseld. NonLinear Finite Element Analysis of Solids and Structures Volume 1: Advanced Topics. John Wiley and Sons, Inc. New York, 605 Third Avenue, New York, NY 101580012, USA, 1997. M. A. Criseld. Consistent schemes for plasticity computation with the Newton Raphson method. Computational Plasticity Models, Software, and Applications, 1:136160, 1987. M. A. Criseld. NonLinear Finite Element Analysis of Solids and Structures Volume 1: Essentials. John Wiley and Sons, Inc. New York, 605 Third Avenue, New York, NY 101580012, USA, 1991. Y. F. Dafalias and M. T. Manzari. Simple plasticity sand model accounting for fabric change eects. Journal of Engineering Mechanics, 130(6):622634, 2004. J. E. Dennis, Jr. and R. B. Schnabel. Numerical Methods for Unconstrained Optimization and Nonlinear Equations. Prentice Hall , Engelwood Clis, New Jersey 07632., 1983. P. Donescu and T. A. Laursen. A generalized objectoriented approach to solving ordinary and partial dierential equations using nite elements. Finite Elements in Analysis and Design, 22:93107, 1996. J. M. Duncan and C.-Y. Chang. Nonlinear analysis of stress and strain in soils. Journal of Soil Mechanics and Foundations Division, 96:16291653, 1970. S. Dunica and B. Kolundija. Nelinarna Analiza Konstrukcija. Gradjevinski fakultet Univerziteta u Beogradu, z Bulevar revolucije 73 i IRO Nauna knjiga, Beograd, UzunMirkova 5, 1986. Nonlinear Analysis of Structures, c In Serbian.
Jeremi et al. c Version: April 29, 2010, 9:23
257
M. Y. E. and E. N. M. H. Dynamic analysis of laterally loaded pile groups in sand and clay. Canadian Geotechnical Journal, 39(6):13581383, 2002. B. Eckel. Using C++. Osborne McGraw Hill, 1989. D. Eyheramendy and T. Zimmermann. Objectoriented nite elements II. a symbolic environment for automatic programming. Computer Methods in Applied Mechanics and Engineering, 132:277304, 1996. C. A. Felippa. Dynamic relaxation under general incremental control. In W. K. Liu, T. Belytschko, and K. C. Park, editors, Innovative Methods for Nonlinear Problems, pages 103133. Pineridge Press, Swansea U.K., 1984. C. A. Felippa. Nonlinear nite element methods. Lecture Notes at CU Boulder, January-May 1993. P. A. Ferrer. Elastoplatic characterization of granular materials. Masters thesis, University of Puerto Rico, 1992. B. W. R. Forde, R. O. Foschi, and S. F. Steimer. Object oriented nite element analysis. Computers and Structures, 34(3):355374, 1990. N. Fox. On the continuum theories of dislocations and plasticity. Quarterly Journal of Mechanics and Applied Mathematics, XXI(1):6775, 1968. A. Gajo. Inuence of viscous coupling in propagation of elastic waves in saturated soil. ASCE Journal of Geotechnical Engineering, 121(9):636644, September 1995. A. Gajo and L. Mongiovi. An analytical solution for the transient response of saturated linear elastic porous media. International Journal for Numerical and Analytical Methods in Geomechanics, 19(6):399413, 1995. B. O. Hardin. The nature of stressstrain behavior of soils. In Proceedings of the Specialty Conference on Earthquake Engineering and Soil Dynamics, volume 1, pages 390, Pasadena, 1978. S. Helgason. Dierential Geometry, Lie Groups, and Symmetric Spaces. Pure and Applied Mathematics. Academic Press, 1978. H. M. Hilber, T. J. R. Hughes, and R. L. Taylor. Improved numerical dissipation for time integration algorithms in structural dynamics. Earthquake Engineering and Structure Dynamics, pages 283292, 1977. R. Hill. Aspects of invariance in solid mechanics. In C.-S. Yih, editor, Advances in Applied Mechanics, volume 18, pages 172. Academic Press, 1978. K. Hjelmstad. Fundamentals of Structural Mechanics. PrenticeHall, 1997. ISBN 0-13-485236-2. T. Hughes. The Finite Element Method ; Linear Static and Dynamic Finite Element Analysis. Prentice Hall Inc., 1987. T. Hughes and K. Pister. Consistent linearization in mechanics of solids and structures. Computers and Structures, 8:391397, 1978. T. J. R. Hughes and W. K. Liu. Implicit-explicit nite elements in transient analaysis: implementation and numerical examples. Journal of Applied Mechanics, pages 375378, 1978a.
Jeremi et al. c Version: April 29, 2010, 9:23
258
T. J. R. Hughes and W. K. Liu. Implicit-explicit nite elements in transient analaysis: stability theory. Journal of Applied Mechanics, pages 371374, 1978b. I. M. Idriss and J. I. Sun. Excerpts from USERS Manual for SHAKE91: A Computer Program for Conducting Equivalent Linear Seismic Response Analyses of Horizontally Layered Soil Deposits. Center for Geotechnical Modeling Department of Civil & Environmental Engineering University of California Davis, California, 1992. N. Janbu. Soil compressibility as determined by odometer and triaxial tests. In Proceedings of European Conference on Soil Mechanics and Foundation Engineering, pages 1925, 1963. B. Jeremi. nDarray programming tool. Object Oriented Approach to Numerical Computations in Elastoplactity, c Reference Manual University of Colorado at Boulder, December 1993. B. Jeremi and M. Preisig. Seismic soil foundation structure interaction: Numerical modeling issues. In SEI c Structures Congress, 2005. B. Jeremi and S. Sture. Implicit integrations in elastoplastic geotechnics. International Journal of Mechanics c of CohesiveFrictional Materials, 2:165183, 1997. B. Jeremi and S. Sture. Tensor data objects in nite element programming. International Journal for Numerical c Methods in Engineering, 41:113126, 1998. P. Jetteur. Implicit integration algorithm for elastoplasticity in plane stress analysis. Engineering Computations, 3:251253, 1986. W. Karush. Minima of functions of several variables with inequalities as side constraints. Masters thesis, University of Chicago, Chicago, IL., 1939. A. Koenig. C++ columns. Journal of Object Oriented Programming, 1989 - 1993. W. T. Koiter. General theorems for elastic-plastic solids. In I. N. Sneddon and R. Hill, editors, Progress in Solid Mechanics, pages 165221. North Holland, 1960. W. T. Koiter. Stress - strain relations, uniqueness and variational theorems for elastic - plastic materials with singular yield surface. Quarterly of Applied Mathematics, 2:350354, 1953. M. Koji. Computational Procedures in Inelastic Analysis of Solids and Structures. Center for Scientic Research c of Srbian Academy of Sciences and Arts and University of Kragujevac and Faculty of Mechanical Engineering University of Kragujevac, 1997. ISBN 86-82607-02-6. S. L. Kramer. Geotechnical Earthquake Engineering. Prentice Hall, Inc, Upper Saddle River, New Jersey, 1996. E. Krner. Algemeine kontinuumstheorie der versetzungen und eigenspanungen. Archive for Rational Mechanics o and Analysis, 4(4):273334, 1960. H. W. Kuhn and A. W. Tucker. Nonlinear programming. In J. Neyman, editor, Proceedings of the Second Berkeley Symposium on Mathematical Statistics and Probability, pages 481 492. University of California Press, July 31 August 12 1950 1951.
Jeremi et al. c Version: April 29, 2010, 9:23
259
P. V. Lade. Model and parameters for the elastic behavior of soils. In Swoboda, editor, Numerical Methods in Geomechanics, pages 359364, Innsbruck, 1988a. Balkema, Rotterdam. P. V. Lade. Eects of voids and volume changes on the behavior of frictional materials. International Journal for Numerical and Analytical Methods in Geomechanics, 12:351370, 1988b. P. V. Lade and R. B. Nelson. Modeling the elastic behavior of granular materials. International Journal for Numerical and Analytical Methods in Geomechanics, 4, 1987. W. T. Lambe and R. V. Whitman. Soil Mechanics, SI Version. John Wiley and Sons, 1979. E. H. Lee. Elasticplastic deformation at nite strains. Journal of Applied Mechanics, 36(1):16, 1969. E. H. Lee and D. T. Liu. Finitestrain elasticplastic theory with application to planewave analysis. Journal of Applied Physics, 38(1):1927, January 1967. X. S. Li and Y. Wang. Linear representation of steady-state line for sand. Journal of Geotechnical and Geoenvironmental Engineering, 124(12):12151217, 1998. J. Lubliner. Plasticity Theory. Macmillan Publishing Company, New York., 1990. L. E. Malvern. Introduction to the Mechanics of a Continuous Medium. In Engineering of the Physical Sciences. Prentice Hall Inc., 1969. M. T. Manzari and Y. F. Dafalias. A critical state twosurface plasticity model for sands. Gotechnique, 47(2): e 255272, 1997. M. T. Manzari and R. Prachathananukit. On integartion of a cyclic soil plasticity model. International Journal for Numerical and Analytical Methods in Geomechanics, 25:525549, 2001. J. M. M. C. Marques. Stress computations in elastoplasticity. Engineering Computations, 1:4251, 1984. J. E. Marsden and T. J. R. Hughes. Mathematical Foundations of Elasticity. Prentice Hall Inc,, 1983. local CM65; 4QA 931.M42 ; ISBN 0-13-561076-1. G. R. Miller. An object oriented approach to structural analysis and design. Computers and Structures, 40(1): 7582, 1991. G. P. Mitchell and D. R. J. Owen. Numerical solution for elasto - plastic problems. Engineering Computations, 5:274284, 1988. K. N. Morman. The generalized strain meassures with application to nonhomogeneous deformation in rubberlike solids. Journal of Applied Mechanics, 53:726728, 1986. G. C. Nayak and O. C. Zienkiewicz. Elasto - plastic stress analysis a generalization for various constitutive relations including strain softening. International Journal for Numerical Methods in Engineering, 5:113135, 1972. N. M. Newmark. A method of computation for structual dynamics. Journal of the Engineering Mechanics Division, ASCE, pages 6794, 1959.
Jeremi et al. c Version: April 29, 2010, 9:23
260
W. L. Oberkampf, T. G. Trucano, and C. Hirsch. Verication, validation and predictive capability in computational engineering and physics. In Proceedings of the Foundations for Verication and Validation on the 21st Century Workshop, pages 174, Laurel, Maryland, October 22-23 2002. Johns Hopkins University / Applied Physics Laboratory. R. W. Ogden. NonLinear Elastic Deformations. Series in mathematics and its applications. Ellis Horwood Limited, Market Cross House, Cooper Street, Chichester, West Sussex, PO19 1EB, England, 1984. M. Ortiz and E. P. Popov. Accuracy and stability of integration algorithms for elastoplastic constitutive relations. International Journal for Numerical Methods in Engineering, 21:15611576, 1985. C. E. Pearson, editor. Handbook of Applied Mathematics. Van Nostrand Reinhold Company, 1974. A. Prez-Foguet and A. Huerta. Plastic ow potetnial for the cone region of the MRSLade model. e A. Prez-Foguet, A. Rodr e guez-Ferran, and A. Huerta. Numerical dierentiation for non-trivial consistent tangent matrices: an application to the mrs-lade model. International Journal for Numerical Methods in Engineering, 48:159184, 2000. D. Peri. Localized Deformation and Failure Analysis of Pressure Sensitive Granular Material. PhD thesis, c University of Colorado at Boulder, 1991. R. M. V. Pidaparti and A. V. Hudli. Dynamic analysis of structures using objectoriented techniques. Computers and Structures, 49(1):149156, 1993. W. H. Press, B. P. Flannery, S. A. teukolsky, and W. T. Vetterling. Numerical Recipes in C. Cambridge University Press, Cambridge, 1988a. W. H. Press, B. P. Flannery, S. A. Teukolvsky, and W. T. Vetterling. Numerical Recipes in C, The Art of Scientic Computing. Cambridge University Press, 1988b. E. Ramm. Strategies for tracing the nonlinear response near limit points. In W. Wunderlich, E. Stein, and K.-J. Bathe, editors, Nonlinear Finite Element Analysis in Structural Mechanics, pages 6389. SpringerWerlag Berlini Heidelberg New York, 1981. E. Ramm. The Riks/Wempner approach an extension of the displacement control method in nonlinear analysis. In E. Hinton, D. Owen, and C. Taylor, editors, Recent Advances in NonLinear Computational Mechanics, chapter 3, pages 6386. Pineridge Press, Swansea U.K., 1982. E. Ramm and A. Matzenmiller. Consistent linearization in elasto - plastic shell analysis. Engineering Computations, 5:289299, 1988. F. E. Richart, J. J. R. Hall, and R. D. Woods. Vibration of Soils and Foundations, International Series in Theoretical and Applied Mechanics. Prentice-Hall, Englewood Clis, N.J., 1970. E. Riks. The application of Newtons method to the problems of elastic stability. Journal of Applied Mechanics, 39:10601066, December 1972.
Jeremi et al. c Version: April 29, 2010, 9:23
261
E. Riks. An incremental approach to the solution of snapping and buckling problems. International Journal for Solids and Structures, 15:529551, 1979. A. D. Robison. C++ gets faster for scientic computing. Computers in Physics, 10(5):458462, Sept/Oct 1996. K. H. Roscoe and J. B. Burland. On the generalized stressstrain behavior of wet clays. In J. Heyman and F. A. Leckie, editors, Engineering Plasticity, pages 535609, 1968. K. H. Roscoe, A. N. Schoeld, and C. P. Wroth. On the yielding of soils. Geotechnique, 8(1):2253, 1958. K. H. Roscoe, A. N. Schoeld, and A. Thurairajah. Yielding of clays in states wetter than critical. Gotechnique, e 13(3):211240, 1963. K. Runesson. Constitutive theory and computational technique for dissipative materials with emphasis on plasticity, viscoplasticity and damage: Part III. Lecture Notes, Chalmers Technical University, Gteborg, Sweden, o September 1996. K. Runesson. Implicit integration of elastoplastic relations with reference to soils. International Journal for Numerical and Analytical Methods in Geomechanics, 11:315321, 1987. K. Runesson and A. Samuelsson. Aspects on numerical techniques in small deformation plasticity. In G. N. P. J. Middleton, editor, NUMETA 85 Numerical Methods in Engineering, Theory and Applications, pages 337347. AA.Balkema., 1985. J. C. J. Schellekens and H. Parisch. On nite deformation elasticity. Technical Report ISD Nr. 94/3, Institute for Statics and Dynamics of Aerospace Structures, Pfaenwaldring 27, 70550 Stutgart, 1994. A. N. Schoeld and C. P. Wroth. Critical State Soil Mechanics. McGraww Hill, 1968. S.-P. Scholz. Elements of an object oriented FEM++ program in C++. Computers and Structures, 43(3): 517529, 1992. J. C. Simo. A framework for nite strain elastoplasticity based on maximum plastic dissipation and the multiplicative decomposition: Part i. continuum formulation. Computer Methods in Applied Mechanics and Engineering, 66:199219, 1988. TA345. C6425. J. C. Simo. Algorithms for static and dynamic multiplicative plasticity that preserve the classical return mapping schemes of the innitesimal theory. Computer Methods in Applied Mechanics and Engineering, 99:61112, 1992. J. C. Simo and S. Govindjee. Exact closed - form solution of the return mapping algorith in plane stress elasto viscoplasticity. Engineering Computations, 5:254258, 1988. J. C. Simo and T. J. R. Hughes, editors. Computational Inelasticity. Springer-verlag, New York, 1998. J. C. Simo and C. Miehe. Associative coupled thermoplasticity at nite strains: Formulation, numerical analysis and implementation. Computer Methods in Applied Mechanics and Engineering, 98:41104, 1992.
Jeremi et al. c Version: April 29, 2010, 9:23
262
J. C. Simo and K. S. Pister. Remarks on rate constitutive equations for nite deformations problems: Computational implications. Computer Methods in Appliied Mechanics and Engineering, 46:201215, 1984. J. C. Simo and R. L. Taylor. Quasiincompressible nite elasticity in principal stretches. continuum basis and numerical algorithms. Computer Methods in Applied Mechanics and Engineering, 85:273310, 1991. J. C. Simo and R. L. Taylor. Consistent tangent operators for rate-independent elastoplasticity. Computer Methods in Applied Mechanics and Engineering, 48:101118, 1985. J. C. Simo and R. L. Taylor. A returning mapping algorithm for plane stress elastoplasticity. International Journal for Numerical Methods in Engineering, 22:649670, 1986. A. Spencer. Continuum Mechanics. Longman Mathematical Texts. Longman Group Limited, 1980. S.Timoshenko and D.H.Young. Engineering Mechanics. McGraw-Hill Book Company, London, 1940. B. Stroustrup. The Design and Evolution of C++. AddisonWesley Publishing Company, 1994. S. Sture. Engineering properties of soils. Lecture Notes at CU Boulder, January - May 1993. T. C. T. Ting. Determination of C 1/2 , C 1/2 and more general isotropic tensor functions of C. Journal of Elasticity, 15:319323, 1985. I. Vardoulakis and J. Sulem. Bifurcation Analysis in Geomechanics. Blackie Academic & Professional, 1995. ISBN 0-7514-0214-1. Various Authors. The C++ report: Columns on C++, 1991-. T. Veldhuizen. Expression templates. C++ Report, 7(5):2631, June 1995. T. Veldhuizen. Rapid linear algebra in C++. Dr. Dobbs Journal, August 1996. G. A. Wempner. Discrete approximations related to nonlinear theories of solids. International Journal for Solids and Structures, 7:15811599, 1971. J. P. Wolf. Dynamic Soil-Structure-Interaction. Prentice-Hall, Englewood Clis (NJ), 1985. J. P. Wolf. Soil-Structure-Interaction Analysis in Time Domain. Prentice-Hall, Englewood Clis (NJ), 1988. J. P. Wolf. Foundation Vibration Analysis Using Simple Physical Models. Prentice-Hall, Englewood Clis (NJ), 1994. J. P. Wolf and M. Preisig. Dynamic stiness of foundation embedded in layered halfspace based on wave propagation in cones. Earthquake Engineering and Structural Dynamics, 32(7):10751098, JUN 2003. J. P. Wolf and C. M. Song. Some cornerstones of dynamic soil-structure interaction. Engineering Structures, 24 (1):1328, 2002. D. M. Wood. Soil Behaviour nad Critical State Soil Mechanics. Cambridge University Press, 1990.
Jeremi et al. c Version: April 29, 2010, 9:23
263
Z. Yang and B. Jeremi. Numerical analysis of pile behaviour under lateral loads in layered elastic-plastic soils. c International Journal for Numerical and Analytical Methods in Geomechanics, 26(14):13851406, 2002. Z. Yang and B. Jeremi. Numerical study of the eective stiness for pile groups. International Journal for c Numerical and Analytical Methods in Geomechanics, 27(15):12551276, Dec 2003. C. Yoshimura, J. Bielak, and Y. Hisada. Domain reduction method for threedimensional earthquake modeling in localized regions. part II: Verification and examples. Bulletin of the Seismological Society of America, 93(2): 825840, 2003. G. W. Zeglinski, R. S. Han, and P. Aitchison. Object oriented matrix classes for use in a nite element code using C++. International Journal for Numerical Methods in Engineering, 37:39213937, 1994. O. Zienkiewicz, A. Chan, M. Pastor, B. Schreer, and T. Shiomi. Computational Geomechanicswith Special Reference to Earthquake Engineering. John Wiley and Sons Ltd., Bans Lane, Chichester, West Sussex PO19 IUD, England, 1999a. O. C. Zienkiewicz and T. Shiomi. Dynamic behaviour of saturated porous media; the generalized Biot formulation and its numerical solution. International Journal for Numerical Methods in Engineering, 8:7196, 1984. O. C. Zienkiewicz and R. L. Taylor. The Finite Element Method, volume 1. McGraw - Hill Book Company, fourth edition, 1991a. O. C. Zienkiewicz and R. L. Taylor. The Finite Element Method, volume 2. McGraw - Hill Book Company, Fourth edition, 1991b. O. C. Zienkiewicz, A. H. C. Chan, M. Pastor, B. A. Schreer, and T. Shiomi. Computational Geomechanics with Special Reference to Earthquake Engineering. John Wiley and Sons., 1999b. ISBN 0-471-98285-7.
Jeremi et al. c
264
Jeremi et al. c
Appendix A
nDarray
Material in this chapter is based on the following publications Jeremi (1993); Jeremi and Sture (1998). c c This section describes a programming tool, nDarray, which is designed using an Object Oriented Paradigm (OOP) and implemented in the C++ programming language. Finite element equations, represented in terms of multidimensional tensors are easily manipulated and programmed. The usual matrix form of the nite element equations are traditionally coded in FORTRAN, which makes it dicult to build and maintain complex program systems. Multidimensional data systems and their implementation details are seldom transparent and thus not easily dealt with and usually avoided. On the other hand, OOP together with ecient programming in C++ allows building new concrete data types, namely tensors of any order, thus hiding the lower level implementation details. These concrete data types prove to be quite useful in implementing complicated tensorial formulae associated with the numerical solution of various elastic and elastoplastic problems in solid mechanics. They permit implementing complex nonlinear continuum mechanics theories in an orderly manner. Ease of use and the immediacy of the nDarray programming tool in constitutive driver programming and in building nite element classes will be shown.
A.1
Introduction
In implementing complex programming systems for nite element computations, the analyst is usually faced with the challenge of transforming complicated tensorial formulae to a matrix form. Considerable amount of time in solving problems by the nite element method is often devoted to the actual implementation process. If one decides to use FORTRAN, a number of nite element and numerical libraries are readily available. Although quick results can be produced in solving simpler problems, when implementing complex small deformation elastoplastic or large deformation elastic and elastoplastic algorithms, C++ provides clear benets. Some of the improvements C++ provides over C and FORTRAN are classes for encapsulating abstractions, the possibility of building userdened concrete data types and operator overloading for expressing complex formulae in a natural way. In the following we shall show that the nDarray tool will allow analysts to be a step closer to the problem space and a step further away from the underlying machine. As most analysts know, the intention (Stroustrup, 1994) behind C++ was not to replace C. Instead, C was 265
266
extended with far more freedom given to the program designer and implementor. In C and FORTRAN, large applications become collections of programs and functions, order and the structure are left to the programmer. The C++ programming language embodies the OOP, which can be used to simplify and organize complex programs. One can build a hierarchy of derived classes and nest classes inside other classes. A concern in C and FORTRAN programming languages is handling data type conicts and data which are being operated on or passed. The C++ programming language extends the denition of type to include abstract data types. With abstract data types, data can be encapsulated with the methods that operate on it. The C++ programming language oers structure and mechanisms to handle larger, more complex programming systems. Object Oriented technology, with function and operator overloading, inheritance and other features, provides means of attacking a problem in a natural way. Once basic classes are implemented, one can concentrate on the physics of a problem. By building further abstract data types one can describe the physics of a problem rather that spend time on the lower level programming issues. One should keep in mind the adage, credited to the original designer and implementor of C++ programming language, Bjarne Stroustrup: C makes it easy to shoot yourself in the foot, C++ makes it harder, but when you do, it blows away your whole leg. Rather than attempting here to give a summary of Object Oriented technology we will suggest useful references for readers who wish to explore the subject in greater depth (Booch, 1994). The current language denition is given in the Working Paper for Draft Proposed International Standard for Information SystemsProgramming Language C++ (ANS, 1995). Detailed description of language evolution and main design decisions are given by Stroustrup (1994). Useful sets of techniques, explanations and directions for designing and implementing robust C++ code are given in books (Coplien, 1992) (Eckel, 1989) and journal articles (Koenig, 1989 - 1993) (Various Authors, 1991-). Increased interest in using Object Oriented techniques for nite element programming has resulted in a number (Donescu and Laursen, 1996) (Eyheramendy and Zimmermann, 1996) (Forde et al., 1990) (Miller, 1991) (Pidaparti and Hudli, 1993) (Scholz, 1992) (Zeglinski et al., 1994) of experimental developments and implementations. Programming techniques used in some of the papers are inuenced by the FORTRAN programming style. Examples provided in some of the above mentioned papers are readable by C++ experts only. It appears that none of the authors have used Object Oriented techniques for complex elastoplasticity computations.
A.2
A.2.1
The nDarray programming tool is a set of classes written in the C++ programming language. The main purpose of the package is to facilitate algebraic manipulations with matrices, vectors and tensors that are often found in computer codes for solving engineering problems. The package is designed and implemented using the Object Oriented philosophy. Great care has been given to the problem of crossplatform and crosscompiler portability. Currently, the nDarray set of classes has been tested and running under the following C++ compilers: Sun CC on SunOS and Solaris platforms, IBM xlC on AIX RISC/6000 platforms,
Jeremi et al. c Version: April 29, 2010, 9:23
267
Borland C++ and Microsoft C++ on DOS/Windows platforms, CodeWarrior C++ on Power Macintosh platform, GNU g++ on SunOS, SOLARIS, LINUX, AIX, HPUX and AMIGA platforms.
A.2.2
Abstraction Levels
nDarray tool has the following simple class hierarchy: nDarray_rep, nDarray matrix vector tensor Indentation of class names implies the inheritance level. For example, class vector is derived from class matrix, which, in turn is derived from classes nDarray and nDarray rep. The idea is to subdivide classes into levels of abstraction, and hide the implementation from end users. This means that the end user can use the nDarray tool on various levels. At the highest level of abstraction, one can use tensor, matrix and vector objects without knowing anything about the implementation and the inner workings. They are all designed and implemented as concrete data types. In spite of the very powerful code that can be built using Object Oriented technology, it would be unwise to expect prociency in Object Oriented techniques and the C++ programming language from end users. It was our aim to provide power programming with multidimensional data types to users with basic knowledge of C. At a lower abstraction level, users can address the task of the actual implementation of operators and functions for vector, matrix and tensor classes. A number of improvements can be made, especially in optimizing some of the operators. The lowest level of abstraction is associated with nDarray and nDarray rep classes. Arithmetic operators1 are implemented at this level. Next, classes are described from the base and down the inheritance tree. Later we focus our attention on nDarray usage examples. Our goal is to provide a useful programming tool, rather than to teach OOP or to show C++ implementation. For readers interested in actual implementation details, source code, examples and makeles are available at http://sokocalo.engr.ucdavis.edu/~jeremic nDarray rep class The nDarray rep class is a data holder and represents an n dimensional array object. A simple memory manager, implemented with the reference counting idiom (Coplien, 1992) is used. The memory manager uses rather inecient builtin C memory allocation functions. Performance can be improved if one designs and implements
1 Like
Jeremi et al. c
268
specially tailored allocation functions for fast heap manipulations. Another possible improvement is in using memory resources other than heap memory. Sophisticated memory management introduced by the reference counting is best explained by Coplien (1992). The nDarray rep class is not intended for standalone use. It is closely associated with the nDarray class. The data structure of nDarray rep introduces a minimal amount of information about a multidimensional array object. The actual data are stored as a onedimensional array of double numbers. Rank, total number of elements, and array of dimensions are all that is needed to represent an multidimensional object. The data structure is allocated dynamically from the heap, and memory is reclaimed by the system after the object has gone out of scope.
nDarray class The nDarray class together with the nDarray rep class represents the abstract base for derived multidimensional data types: matrices, vectors and tensors. Objects derived from the nDarray class are generated dynamically by constructor functions at the rst appearance of an object and are destroyed at the end of the block in which the object is referenced. The reference counting idiom provides for the objects life continuation after the end of the block where it was dened. To extend an objects life, a standard C++ compiler would by default call constructor functions, thus making the entire process of returning large objects from functions quite inecient. By using reference counting idiom, destructor and constructor functions manipulate reference counter which results in a simple copying of a pointer to nDarray rep object. By using this technique, copying of large objects is made very ecient.
constructor function nDarray(int rank_of_nD=1, double initval=0.0) nDarray(int rank_of_nD, const int *pdim, double *val) nDarray(int rank_of_nD, const int *pdim, double initval) nDarray(const char *flag, int rank_of_nD, const int *pdim) nDarray(const nDarray & x) nDarray(int rank_of_nD, int rows, int cols, double *val) nDarray(int rank_of_nD, int rows, int cols, double initval) description default from array from scalar value unit nDarrays copy-initializer special for matrix special for matrix
Table A.1: nDarray constructor functions. Objects can be created from an array of values, or from a single scalar value, as shown in Table A.1. Some of the frequently used multidimensional arrays are predened and can be constructed by sending the proper ag to the constructor function. For example by sending the I ag one creates Kronecker delta ij and by sending e ag, one creates a rank 3 Levi-Civita permutation tensor eijk . Functions and operators common to multidimensional data types are dened in the nDarray class, as described in Table A.2. These common operators and functions are inherited by derived classes. Occasionally, some of the functions will be redened, overloaded in derived classes. In tensor multiplications we need additional information about indices. For example Cil = (Aijk + Bijk ) Djkl
Jeremi et al. c
coded
269 right value nDarray nDarray nDarray nDarray nDarray nDarray nDarray double nDarray description nDarray assignment nDarray addition nDarray addition unary minus nDarray subtraction nDarray subtraction scalar multiplication (from left) scalar multiplication (from right) nDarray comparison reference to members of nDarray members of nDarray trace of square nDarray eigenvalues of rank 2 square nDarray eigenvectors of rank 2 square nDarray general p-th norm of nDarray square root of nDarray generic print function
operator or function = + += -= * * == val(...) cval(...) trace() eigenvalues() eigenvectors() General_norm() nDsqrt() print(...)
left value nDarray nDarray nDarray nDarray nDarray double nDarray nDarray nDarray nDarray nDarray nDarray nDarray nDarray nDarray nDarray
Table A.2: Public functions and operators for nDarray class. ijk indices, to be used for multiplication with Djkl . It is interesting to note (Koenig, 1989 - 1993) that operator += is dened as a member and + is dened as an inline function in terms of += operator. Matrix and Vector Classes The matrix class is derived from the nDarray class through the public construct. It inherits common operators and functions from the base nDarray class, but it also adds its own set of functions and operators. Table (A.3) summarizes some of the more important additional functions and operators for the matrix class. The vector class operator or function = * transpose() determinant() inverse() left value matrix matrix matrix matrix matrix right value matrix matrix description matrix assignment matrix multiplication matrix transposition determinant of a matrix matrix inversion
Table A.3: Matrix class functions and operators (added on nDarray class denitions). denes vector objects and is derived and inherits most operators and data members from the matrix class. Some
Jeremi et al. c Version: April 29, 2010, 9:23
270
functions, like copy constructor, are overloaded in order to handle specics of a vector object.
Tensor Class The main goal of the tensor class development was to provide the implementing analyst with the ability to write the following equation directly into a computer program:
1 dmn = oldrij Tijmn d Eijkl n+1 1 mkl Tijmn
as:
Instead of developing theory in terms of indicial notation, then converting everything to matrix notation and then implementing it, we were able to copy formulae directly from their indicial form to the C++ source code. In addition to the denitions in the base nDarray class, the tensor class adds some specic functions and operators. Table A.4 summarizes some of the main new functions and operators. The most signicant addition is operator or function + * transpose0110() transpose0101() transpose0111() transpose1100() transpose0011() transpose1001() transpose11() symmetrize11() determinant() inverse() left value tensor tensor tensor tensor tensor tensor tensor tensor tensor tensor tensor tensor tensor right value tensor tensor tensor description tensor addition tensor subtraction tensor multiplication Aijkl Aikjl Aijkl Ailkj Aijkl Ailjk
Aijkl Ajikl Aijkl Aijlk aij aji Aijkl Aljki symmetrize second order tensor determinant of 2nd order tensor tensor inversion (2nd, 4th order)
Table A.4: Additional and overloaded functions and operators for tensor class.
the tensor multiplication operator. With the help of a simple indicial parser, the multiplication operator contracts or expands indices and yields a resulting tensor of the correct rank. The resulting tensor receives proper indices, and can be used in further calculations on the same code statement.
Jeremi et al. c Version: April 29, 2010, 9:23
271
A.3
A.3.1
The next step in our development was to use the nDarray tool classes for constitutive level computations. The simple extension was design and implementation of innitesimal stress and strain tensor classes, namely stresstensor and straintensor. Both classes are quite similar, they inherit all the functions from the tensor class and we add some tools that are specic to them. Both stress and strain tensors are implemented as full second order 3 3 tensors. Symmetry of stress and strain tensor was not used to save storage space. Table A.5 summarizes some of the main functions added on for the stresstensor class. operator or function Iinvariant1() Iinvariant2() Iinvariant3() Jinvariant2() Jinvariant3() deviator() principal() sigma_octahedral() tau_octahedral() xi() rho() p_hydrostatic() q_deviatoric() theta() description rst stress invariant I1 second stress invariant I2 third stress invariant I3 second deviatoric stress invariant J2 third deviatoric stress invariant J3 stress deviator principal stresses on diagonal octahedral mean stress octahedral shear stress HaighWestergard coordinate HaighWestergard coordinate hydrostatic stress invariant deviatoric stress invariant stress invariant (Lodes angle)
Table A.5: Additional methods for stress tensor class. Further on, we dened an elastoplastic state, which according to incremental theory of elastoplasticity with internal variables, is completely dened with the stress tensor and a set of internal variables. This denition led us to dene an elastoplastic state termed class ep state. Objects of type ep state contain a stress tensor and a set of scalar or tensorial internal variables2 .
A.3.2
With all the previous developments, the design and implementation of various elastoplastic material models was not a dicult task. A generic class Material Model denes techniques that form a framework for small deformation elastoplastic computations. Table A.6 summarizes some of the main methods dened for the Material Model class
2 Internal
variables can be characterized as tensors of even order, where, for example, zero tensor is a scalar internal variable
associated with isotropic hardening and second order tensors can be associated with kinematic hardening.
Jeremi et al. c
272
in terms of yield (F ) and potential (Q) functions. operator or function F dFods dQods d2Qods2 dpoverds dqoverds dthetaoverds d2poverds2 d2qoverds2 d2thetaoverds2 ForwardPredictorEPState BackwardEulerEPState ForwardEulerEPState BackwardEulerCTensor ForwardEulerCTensor description F Yield function value F /ij Q/ij 2 Q/ij kl p/ij q/ij /ij 2 p/ij kl 2 q/ij kl 2 /ij kl Explicit predictor elastoplastic state Implicit return elastoplastic state Explicit return elastoplastic state Algorithmic tangent stiness tensor Continuum tangent stiness tensor
Table A.6: Some of the methods in material model class. It is important to note that all the material model dependent functions are dened as virtual functions. Integration algorithms are designed and implemented using template algorithms, and each implemented material model appends its own yield and potential functions and appropriate derivatives. Implementation of additional material models requires coding of yield and potential functions and respective derivative functions.
A.3.3
Starting from the incremental equilibrium of the stationary body, the principle of virtual displacements and with expressed as (Zienkiewicz and Taylor, 1991a) HI,b Eabcd HJ,d dV m uJc =
m V
m
the nite element approximation of the displacement eld u ua = HI uIa , the weak form of equilibrium can be fa HI dV m or (fIa (Jc ))int = (fIa )ext u
where Eabcd is the constitutive tangent stiness tensor3 . The element stiness tensor is recognized as
e kaIcJ =
HI,b
Vm
tan
Eabcd HJ,d dV m
This generic form for the nite element stiness tensor is easily programmed with the help of the nDarray tool. A simple implementation example is provided later. It should be noted that the element stiness tensor in this case is a fourdimensional tensor. It is the task of the assembly function to collect proper terms for addition in a global stiness matrix.
3 Which
may be continuum or algorithmic (Jeremi and Sture, 1997) tangent stiness tensor c
Jeremi et al. c
273
A.4
A.4.1
Examples
Tensor Examples
Some of the basic tensorial calculations with tensors are presented. Tensors have a default constructor that creates a rst order tensor with one element initialized to 0.0:
tensor t1;
Here, DefDim2, DefDim3 and DefDim4 are arrays of dimensions for the second, third and fourth order tensor4 . A fourth order tensor with 0.0 value assignment and dimension 3 in each order (3 3 3 3) is constructed in the following way:
tensor ZERO(4,DefDim4,0.0);
Tensors can be multiplied using indicial notation. The following example will do a tensorial multiplication of previously dened tensors t2 and t4 so that tst1 = t2ij t4ijkl t4klpq t2pq . Note that the memory is dynamically allocated to accept the proper tensor dimensions that will result from the multiplication5
tensor tst1 = t2("ij")*t4("ijkl")*t4("klpq")*t2("pq");
Inversion of tensors is possible. It is dened for 2 and 4 order tensors only. The fourth order tensor inversion is done by converting it to matrix, inverting that matrix and nally converting matrix back to tensor.
tensor t4inv_2 = t4.inverse();
There are two builtin tensor types, Levi-Civita permutation tensor eijk and Kronecker delta tensor ij
tensor e("e",3,DefDim3); tensor I2("I", 2, DefDim2); // Levi-Civita permutation tensor // Kronecker delta tensor
this case dimensions are 3 in every order. this case it will be zero dimensional tensor with one element. 6 Machine epsilon (macheps) is dened as the smallest distinguishable positive number (in a given precision, i.e. oat (32 bits),
double (64 bits) or long double (80 bits), such that 1.0 + macheps > 1.0 yields true on the given computer platform. For example, double precision arithmetics (64 bits), on the Intel 80x86 platform yields macheps=1.08E-19 while on the SUN SPARCstation and DEC platforms macheps=2.22E-16.
Jeremi et al. c
274
A.4.2
Some of the fourth order tensors used in continuum mechanics are built quite readily. The most general representation of the fourth order isotropic tensor includes the following fourth order unit isotropic tensors7
tensor I_ijkl = I2("ij")*I2("kl");
The resulting tensor I_ijkl will have the correct indices, I ijklijkl = I2ij I2kl . Note that I_ijkl is just a name for the tensor, and the _ijkl part reminds the implementor what that tensor is representing. The real indices, ijkl in this case, are stored in the tensor object, and can be used further or changed appropriately. The next tensor that is needed is a fourth order unit tensor obtained by transposing the previous one in the minor indices,
tensor I_ikjl = I_ijkl.transpose0110();
while the third tensor needed for representation of general isotropic tensor is constructed by using similar transpose function
tensor I_iljk = I_ijkl.transpose0111();
Creating a symmetric and skew symmetric unit fourth order tensors gets to be quite simple by using tensor addition and scalar multiplication
tensor I4s = (1./2.)*(I_ikjl+I_iljk);
Another interesting example is a numerical check of the e identity (Lubliner, 1990) (eijm eklm = ik jl il jk )
tensor id = e("ijm")*e("klm") - (I_ikjl - I_iljk); if ( id == ZERO ) printf(" e-delta identity HOLDS !! ");
A.4.3
The linear isotropic elasticity tensor Eijkl can be built from Youngs modulus E and Poissons ratio
double Ey = 20000; // Youngs modulus of elasticity double nu = 0.2; // Poissons Ratio tensor E = ((2.*Ey*nu)/(2.*(1.+nu)*(1-2.*nu)))*I_ijkl + (Ey/(1.+nu))*I4s;
One can multiply the two and check if the result is equal to the symmetric fourth order unit tensor
7 Remember
Jeremi et al. c
275
tensor test = E("ijkl")*D("klpq"); if ( test == I4s ) printf(" test == I4s else printf(" test == I4s NOTTRUE "); TRUE (OK up to sqrt(macheps)) ");
The linear isotropic elasticity and compliance tensors can be obtained in a dierent way, by using Lam constants e and
double lambda = nu * Ey / (1. + nu) / (1. - 2. * nu); double mu = Ey / (2. * (1. + nu)); tensor E = lambda*I_ijkl + (2.*mu)*I4s; tensor D = (-nu/Ey)*I_ijkl // stiffness tensor + (1./(2.*mu))*I4s; // compliance tensor
A.4.4
As an extended example of nDarray tool usage, the implementation for the second derivative of the stress invariant (Lode angle) is presented. The derivative is used for implicit constitutive integration schemes applied to three invariant material models. The original equation reads:
2 pq mn
= + + 1 27 cos 3 9 cos 3 81 + spq tmn + spq smn + 2 q 4 sin (3) 4 q 4 sin3 3 4 q 5 sin3 3 1 81 cos2 3 243 cos 3 81 + tpq tmn + tpq smn 5 sin 3 4 q 4 q 5 sin3 3 4 q 6 sin3 3 1 9 3 cos (3) ppqmn wpqmn 2 q 2 sin (3) 2 q 3 sin (3)
3 3 cos 3 = 2
1 sij = ij kk ij 3
1 ppqmn = mp nq pq mn 3
tensor Yield_Criteria::d2thetaoverds2(stresstensor & stress) { tensor ret( 4, DefDim4, 0.0); tensor I2("I", 2, DefDim2); tensor I_pqmn = I2("pq")*I2("mn"); tensor I_pmqn = I_pqmn.transpose0110(); double J2D = stress.Jinvariant2(); tensor s = stress.deviator(); tensor t = s("qk")*s("kp") - I2*(J2D*(2.0/3.0));
Jeremi et al. c
276
double theta = stress.theta(); double q_dev = stress.q_deviatoric(); //setting up some constants double c3t double s3t double s3t3 double q3 double q4 double q5 double q6 = cos(3*theta); = sin(3*theta); = s3t*s3t*s3t; = q_dev * q_dev * q_dev; = q3 * q_dev; = q4 * q_dev; = q5 * q_dev;
double tempss = -(9.0/2.0)*(c3t)/(q4*s3t)-(27.0/4.0)*(c3t/(s3t3*q4)); double tempst = +(81.0/4.0)*(1.0)/(s3t3*q5); double tempts = +(81.0/4.0)*(1.0/(s3t*q5))+(81.0/4.0)*(c3t*c3t)/(s3t3*q5); double temptt = -(243.0/4.0)*(c3t/(s3t3*q6)); double tempp double tempw = +(3.0/2.0)*(c3t/(s3t*q_dev*q_dev)); = -(9.0/2.0)*(1.0/(s3t*q3));
tensor s_pq_d_mn = s("pq")*I2("mn"); tensor s_pn_d_mq = s_pq_d_mn.transpose0101(); tensor d_pq_s_mn = I2("pq")*s("mn"); tensor d_pn_s_mq = d_pq_s_mn.transpose0101(); tensor p = I_pmqn // finally ret = (s("pq")*s("mn")*tempss + s("pq")*t("mn")*tempst + t("pq")*s("mn")*tempts + t("pq")*t("mn")*temptt + p*tempp + w*tempw ); return ret; } - I_pqmn*(1.0/3.0); tensor w = s_pn_d_mq+d_pn_s_mq - s_pq_d_mn*(2.0/3.0)-d_pq_s_mn*(2.0/3.0);
A.4.5
A useful application of the previously described classes is for elastoplastic computations. If the Newton iterative scheme is used at the global equilibrium level, then in order to preserve a quadratic rate, a consistent, algorithmic tangent stiness (ATS) tensor should be used. For a general class of threeinvariant, nonassociated, hardening or softening material models, ATS is dened (Jeremi and Sture, 1997) as: c
cons ep Epqmn
= Rpqmn
Hkl = n+1mkl +
mkl h q
Rmnkl =
n+1
Tijmn
Eijkl
Jeremi et al. c
277
tensor Eel = StiffnessTensorE(Ey,nu); tensor I2("I", 2, DefDim2); tensor I_ijkl = I2("ij")*I2("kl"); tensor I_ikjl = I_ijkl.transpose0110(); tensor m = Criterion.dQods(final_stress); tensor n = Criterion.dFods(final_stress); double lambda = current_lambda_get(); tensor d2Qoverds2 = Criterion.d2Qods2(final_stress); tensor T = I_ikjl + Eel("ijkl")*d2Qoverds2("klmn")*lambda; tensor Tinv = T.inverse(); tensor R = Tinv("ijmn")*Eel("ijkl"); double h_ = h(final_stress);
double xi_ = xi(final_stress); double hardMod_ = h_ * xi_; tensor d2Qodqast2 = d2Qoverdqast2(final_stress); tensor H = m + d2Qodqast2 * lambda * h_; // tensor upper = R("pqkl")*H("kl")*n("ij")*R("ijmn"); double lower = (n("ot")*R("otpq"))*H("pq")).trace(); lower = lower + hardMod_; tensor Ep = upper*(1./lower); tensor Eep = R - Ep; // elastoplastic ATS constitutive tensor
This ATS tensor can be used further in building nite element stiness tensors, as will be shown in our next example.
A.4.6
individual contributions are summed into the element stiness tensor. This process can be implemented on a integration point level by using the nDarray tool as
K = K + H("Ib") * E("abcd") * H("Jd") * weight ;
It is interesting to note the lack of loops at this level of implementation. However, there exists a loop over integration points which contributes stiness to the element tensor.
A.5
Performance Issues
In the course of developing the nDarray tool, execution speed was not a priority or issue that we tried to perfect. The benet of being able to implement and test various numerical algorithms in a straightforward manner was the main concern. The eciency of the nDarray tool when compared with FORTRAN or C was never assessed. In all honesty, some of the formulae implemented in C++ with the help of the nDarray tool would be dicult
Jeremi et al. c Version: April 29, 2010, 9:23
278
to implement in FORTRAN or C. The entire question of eciency of the nDarray as compared to FORTRAN or C codes might thus remain unanswered for the time being. The eciency of C++ for numerical computations has been under consideration (Robison, 1996) for some time now. Poor eciency and possible remedies for improving eciency of C++ computations has been reported in literature (Robison, 1996) (Veldhuizen, 1995) (Veldhuizen, 1996). Novel techniques, such as Template Expressions (Veldhuizen, 1995) can be used to achieve and sometimes surpass the performance of handtuned FORTRAN or C codes.
A.6
A novel programming tool, named nDarray, has been presented which facilitates implementation of tensorial formulae. It was shown how OOP and ecient programming in C++ allows building of new concrete data types, in this case tensors of any order. In a number of examples these new data types were shown to be useful in implementing tensorial formulae associated with the numerical solution of various elastic and elastoplastic problems with the nite element method. The nDarray tool is been used in developing of the F Mtools tools library. The E F Mtools tools library includes a set of nite elements, various solvers, solution procedures for nonlinear nite E element system of equations and other useful functions.
Jeremi et al. c
Appendix B
Useful Formulae
B.1 Stress and Strain
This section reviews small deformation stress and strain measures used in this report.
B.1.1
Stress
In this work, the tensile stress is assumed positive, and in general we follow classical strength of materials (mechanics of materials) conventions for stress and strain. The stress tensor ij is given as xx xy yy zy xz x xy y yz zx
= yx zx
yz = xy zx zz
yz z
(B.1)
In small deformation theory, this stress is symmetric, that is, xy = yx , yz = zy , and zx = xz . There are only six independent components and sometimes the stress can be expressed in the vector form = {xx , yy , zz , xy , yz , zx } (B.2)
The principle stresses 1 , 2 , and 3 (1 2 3 ) are the eigenvalues of the symmetric tensor ij in Equation B.1 and can be obtained by solving the equation xx yx zx xy yy zy xz yz zz =0 (B.3)
280
The three rst-type stress invariants are then I1 = ii = xx + yy + zz = 1 + 2 + 3 1 ij ji 2 2 2 2 = (xx yy + yy zz + zz xx ) + (xy + yz + zx ) (B.6) (B.5)
I2
(B.7)
The stress ij can be decomposed into the hydrostatic stress m ij and deviatoric stress sij as ij = m ij + sij , with the denitions m = 1 1 I1 , sij = ij kk ij 3 3 (B.8)
where ij is the Kronecker operator such that ij = 1 for i = j and ij = 0 for i = j. Since both hydrostatic and deviatoric stresses are stress tensors, they have their own coordinate-independent stress invariants respectively. The three invariants of the hydrostatic stress are I1 = 3m = I1 , 1 2 1 3 2 3 I2 = 3m = I1 , I3 = m = I 3 27 1 (B.9)
Since I1 , I2 and I3 are all simple functions of I1 , the hydrostatic stress state can therefore be represented by only one variable I1 . The three eigenvalues of the deviatoric stresses sij are called principal deviatoric stresses, with the order s1 s2 s3 . The three invariants of the deviatoric stress are J1 = sii = 0 1 sij sji 2 1 2 I + I2 3 1 1 [(1 2 )2 + (2 3 )2 + (3 1 )2 ] 6 (sxx syy + syy szz + szz sxx ) + (s2 + s2 + s2 ) xy yz zx 1 2 2 2 (s + s2 + s3 ) = (s1 s2 + s2 s3 + s3 s1 ) 2 1 (B.10)
J2
= = = = =
(B.11)
Version: April 29, 2010, 9:23
Jeremi et al. c
281
J3
1 sij sjk ski = det(sij ) 3 1 2 3 1 1 3 = I3 + I1 I2 + I1 = I3 I1 J2 I1 3 27 3 27 1 (21 2 3 )(22 3 1 )(23 1 2 ) = 27 = sxx syy szz + 2sxy syz szx (sxx s2 + syy s2 + szz s2 ) yz zx xy = s1 s2 s3 (B.12)
The deviatoric stress state can therefore be represented by only two variables J2 and J3 . Combining hydrostatic and deviatoric stress, we can conclude that the stress state can be be represented by three variables I1 , J2 and J3 . Using the three invariants (I1 , J2 , J3 ) or its equivalents instead of the nine components of ij is widely used in geomechanics. The stress state may also be described in three dimensional space (p, q, ), dened as p q 1 = I1 3 3J2 = = 1 arccos 3 (B.13) 3 3 J3 3 2 J2 (B.14) (B.15)
where ij is the stress Lodes angle (0 ij /3). A stress state with = 0 corresponds to the meridian of conventional triaxial compression (CTC), while = /3 to the meridian of conventional triaxial extension (CTE). The relationship between (1 , 2 , 3 ) cos 1 2 2 = p + q cos ( 2 ) 3 3 2 3 cos ( + ) 3 and (p, q, ) is
(B.16)
The line of the principal stress space diagonal is called hydrostatic axis. Any plane perpendicular to the
hydrostatic axis is an deviatoric plane, or plane. The Haigh-Westergaard three dimensional stress coordinate system (, , ) Chen and Han (1988), is dened as I1 = = 3p 3 = 2J2 = 2 q 3
(B.17) (B.18)
The Haigh-Westergaard invariants have physical meanings. is the distance of the deviatoric plane to the origin of the Haigh-Westergaard coordinates, and is the distance of a stress point to the hydrostatic line and represents the magnitude of the deviatoric stress. The projections of the axes 1 , 2 and 3 on the deviatoric plane are
assumed 1 , 2 and 3 respectively. (, ) is the polar coordinate system in the deviatoric plane with the 1 the
polar axis and the polar angle. The relationship between (1 , 2 , 3 ) and (, , ) is cos 1 1 2 = + 2 cos ( 2 ) 3 3 3 2 3 cos ( + ) 3
Jeremi et al. c
(B.19)
282
B.1.2
Strain
Point P (xi ) and nearby point Q(xi + dxi ) displace due to applied loading to new positions P (xi + Ui ) and Q(ui + (ui /xj )dxj ). We can dene a displacement gradient tensor ui,j as ui,j = ui xj the displacement gradient can decomposed into the symmetric and antisymmetric parts 1 1 u1,1 u1,2 u1,3 2 (u1,2 + u2,1 ) 2 (u1,3 + u3,1 ) 1 1 u2,2 (u2,3 + u3,2 ) u2,2 u2,3 = 2 (u2,1 + u1,2 ) 2 1 1 u3,3 u3,2 u3,3 2 (u3,1 + u1,3 ) 2 (u3,2 + u2,3 ) 1 1 0 2 (u1,2 u2,1 ) 2 (u1,3 u3,1 ) 1 + 1 (u2,1 u1,2 ) 0 2 2 (u2,3 u3,2 ) 1 1 0 2 (u3,1 u1,3 ) 2 (u3,2 u2,3 ) (B.20)
(B.21)
or
(B.22)
(B.23)
wij =
(B.24)
The symmetric part of the deformation gradient tensor, ij , is the small deformation strain tensor 1 , while the antisymmetric part of the deformation gradient tensor, wij , is the rotation motion tensor. The matrix form of the strain ij is xx xy = zx
xy yy yz
zx
yz = zz
x
1 2 xy 1 2 zx
1 2 xy
y
1 2 yz
1 2 zx 1 2 yz
(B.25)
(B.26)
Note that the engineering shear strain ij is the double of the corresponding strain component ij .
invariants I1 , I2 , and I3 , dened as I1
Similar to the stress tensor, the strain tensor also has three principle strains i (1 2 3 ), and three strain = ii = xx + yy + zz = 1 + 2 + 3 (B.27)
1 Here
the second and higher order derivative terms are neglected due to the small deformation assumption.
Jeremi et al. c
283
I2
1 ij ji 2 = (xx yy + yy zz + zz xx ) + (2 + 2 + 2 ) xy yz zx (B.28)
= (1 2 + 2 3 + 3 1 )
I3
The rst strain invariant is also called the volumetric strain v . The strain ij can be decomposed into the hydrostatic strain m ij and deviatoric strain eij through ij = m ij + eij where: m = 1 1 I1 , eij = ij kk ij 3 3 (B.30)
Since both hydrostatic and deviatoric strains are strain tensors, they have their own strain invariants respectively. The three invariants of the hydrostatic strain are
I1 = 3m = I1 ,
1 3 1 (I ) I2 = 32 = (I1 )2 , I3 = 3 = m m 3 27 1
(B.31)
The hydrostatic strain state can therefore be represented by only one variable I1 .
The three eigenvalues of the deviatoric strains eij are called principal deviatoric strains, with the order e1 e2 e3 . The three invariants of the deviatoric strain are
J1
= eii = 0 = = = = = 1 eij eji 2 1 2 (I ) + I2 3 1 1 [(1 2 )2 + (2 3 )2 + (3 1 )2 ] 6 (exx eyy + eyy ezz + ezz exx ) + (e2 + e2 + e2 ) xy yz zx 1 2 2 2 (e + e2 + e3 ) = (e1 e2 + e2 e3 + e3 e1 ) 2 1
(B.32)
J2
(B.33)
J3
1 eij ejk eki = det(eij ) 3 2 1 1 1 = I3 + I1 I2 + (I1 )3 = I3 I1 J2 (I1 )3 3 27 3 27 1 = (21 2 3 )(22 3 1 )(23 1 2 ) 27 = exx eyy ezz + 2exy eyz ezx (exx e2 + eyy s2 + ezz e2 ) yz zx xy = = e1 e2 e3 (B.34)
The deviatoric strain state can therefore be represented by only two variables J2 and J3 .
Jeremi et al. c
284
Combining the hydrostatic and deviatoric strain, we can conclude that the strain state can be be represented
by three variables I1 , J2 and J3 .
Strain state may also be represented with another three invariant (p , q , ), dened as p q
= I1 = v
(B.35) (B.36) 3 3 2
J3 (J2 )3
= =
J2
1 arccos 3
(B.37)
where is the strain Lodes angle and 0 /3. The relationship between (1 , 2 , 3 ) and (p , q , ) is 1 3 q cos ( 2 cos ( + cos 2 ) 3 2 ) 3
2 = 1 p + 3 3
(B.38)
B.2
In this part of the Appendix, we shall derive some useful formulae, that are rarely found2 in texts treating elasto plastic problems in mechanics of solid continua. First derivative of I1 with respect to stress tensor ij :
kk I1 = = ij ij ij
( 1 smn snm ) 1 smn 1 snm J2D = 2 = snm + smn = ij ij 2 ij 2 ij = (nm 1 kk nm ) snm 1 3 smn = smn = (ni jm nm ki jk )smn = ij ij 3 1 1 = (ni jm nm ij )smn = ni jm snm nm ij smn = sij 3 3
because
Jeremi et al. c
285
( 1 sij sjk ski ) 1 sij 1 sjk 1 ski J3D = 3 = sjk ski + sij ski + sij sjk = pq pq 3 pq 3 pq 3 pq =
1 (ij 3 kk ij ) sij 1 sjk ski = sjk ski = (ip qj ij kp qk )sjk ski = pq pq 3 2 1 = ip qj sjk ski ij kp qk sjk ski = sqk skp pq J2D = tpq 3 3
First derivative of spq with respect to stress tensor mn , or second derivative of J2D with respect to stress tensors pq and mn :
pq 1 pq kk spq 3 = = mn mn =
1 mp nq 3 pq mn mn = mn 1 mp nq pq mn = ppqmn 3
2 sqk skp 3 pq J2D tpq (sqk skp ) 2 pq J2D 3 = = = mn mn mn mn (sqk skp ) 2 J2D sqk skp 2 = pq = skp + sqk pq smn = mn 3 mn mn mn 3 1 2 1 qm nk qk nm skp + sqk km np kp nm pq smn = 3 3 3 1 2 1 = qm snp sqp nm + sqm np sqp nm pq smn = 3 3 3 2 2 = snp qm + sqm np sqp nm pq smn = wpqmn 3 3
since 1 s s 3 ij kp qk jk ki
1 s s 3 kp qk ik ki
1 s s 3 qp ik ki
2 J 3 pq 2D
Jeremi et al. c
286
Eijkl Dklpq 2 2 ij kl + ik jl + il jk kl pq + kp lq + kq lp 1 2 1+ 1 (ik jl kp lq + il jk kp lq + ik jl kq lp + il jk kq lp ) 4 (ij kl kp lq + ij kl kq lp ) (ik jl kl pq + il jk kl pq ) + 2 (1 2) 2 (1 + ) 2 ij kl kl pq (1 2) (1 + ) 1 (ip jq + iq jp ) 2 + (ij kq kp + ij kp kq ) (il jl pq + il jl pq ) 2 (1 2) 2 (1 + ) 3 2 ij pq (1 2) (1 + ) 1 (ip jq + iq jp ) 2 (ij pq + ij pq ) (ij pq + ij pq ) + 2 (1 2) 2 (1 + ) 3 2 ij pq (1 2) (1 + ) 1 (ip jq + iq jp ) 2 + ij pq ij pq (1 2) 2 (1 + ) 3 2 ij pq (1 2) (1 + ) 1 (ip jq + iq jp ) 2 (1 + ) (1 2) 3 2 + ij pq (1 2) (1 + ) 1 (ip jq + iq jp ) 2 E 1+ 2 (1 + ) 2E
= = + = + = + = + = + =
sym = Iijpq
Jeremi et al. c
Appendix C
1 p = I1 3
q=
3J2D
J3D (J2D )3
(C.1)
I1 = kk
J2D =
1 sij sij 2
J3D =
1 sij = ij kk ij 3
(C.2)
and stresses are chosen as positive in tension. One can write the Potential Function in the following form:
Q = Q(p, q, )
(C.3)
and the derivation follows. Hopefully the pace of derivation is rather slow, thus little explanation will be given until the end of the derivation. Chain rule of dierentiation yields:
Q p Q q Q Q = + + ij p ij q ij ij
(C.4)
288
1 ( 3 kk ) 1 p = = ij ij ij 3
(C.5)
(C.6)
(C.7) 3 3 J3D 2 2 3/2 ) 3 3 2 2 J3D ij 1 (J2D 1 (J2D )3 )3 J2D 3 J3D 2 ij 3 + J3D sij 2 1 (J2D )5 1 (J2D )5 =
J2D
3 3 J3D 2 J 3/2 2D
tij
3 J3D 2
1 (J2D )5
sij
1 (J2D )3
tij
(C.8)
Second derivatives of the potential function Q using again the chain rule of dierentiation are as follows:
Jeremi et al. c Version: April 29, 2010, 9:23
289
Q pq 2Q = pq mn mn
= +
Q pq
Q p p pq
Q q q pq
mn
Q p
mn +
Q q
2p Q p + + pq p pq mn 2q Q q + + pq q pq mn 2 Q + = pq pq mn 2p Q p + + pq p pq mn 2q Q q + + pq q pq mn 2 Q + pq pq mn
mn
Q
mn
2 Q p 2 Q q 2 Q + + p2 mn pq mn p mn + + 2 Q q 2 Q 2 Q p + + 2 qp mn q q mn mn 2 2 Q q 2 Q Q p + + p mn q mn 2 mn
1 2 1 kk 3 kp qk 2p 3 = = = pq mn pq mn mn
3 1 2 J2D
pq 2q = = pq mn mn 1 1 23 J spq 3 1 spq 3 J2D 2D = + spq = mn 2 J2D mn 2 mn 3 1 1 1 pm nq pq km nk + smn spq = 3 3 2 2 J2D 3 1 1 3 1 3 pm nq pq nm smn spq = 2 J2D 3 4 J2D 9 1 1 31 pm nq pq nm smn spq 2q 3 4 q3
2 pq mn
290
pq
= AS spq + AT tpq
where:
AS =
AT =
9 1 2 q 3 sin (3)
pq 2 = = pq mn mn
(AS spq ) mn .
Since:
Version: April 29, 2010, 9:23
291
spq +
3 cos (3) ppqmn = 2 q 2 sin (3) 1 27 cos 3 9 cos 3 81 spq smn spq smn + spq tmn + 2 q 4 sin (3) 4 q 4 sin3 3 4 q 5 sin3 3 3 cos (3) ppqmn = 2 q 2 sin (3) 9 cos 3 27 cos 3 + 4 sin (3) 2q 4 q 4 sin3 3 spq smn + 81 1 s t + 5 sin3 3 pq mn 4 q 3 cos (3) ppqmn 2 q 2 sin (3)
where:
ppqmn =
spq = mn
1 mp nq pq mn 3
3. cot(3 ) AS = q q3
AS 4.5 csc(3 ) = q2
(AT tpq ) mn :
292
AT q q mn
13.5 csc(3 ) 3 1 smn + q4 2q 1 3 cos (3) 9 tpq + s t 2 sin (3) mn 3 sin (3) mn 2q 2q 1 9 wpqmn = + 2 q 3 sin (3)
81 243 cos 3 1 81 cos2 3 t s tpq tmn tpq smn + 5 sin 3 5 sin3 3 pq mn 4 q 4 q 4 q 6 sin3 3 9 1 3 wpqmn = 2 q sin (3) 81 1 81 cos2 3 + 4 q 5 sin 3 4 q 5 sin3 3 tpq smn 243 cos 3 tpq tmn 4 q 6 sin3 3 9 1 3 wpqmn 2 q sin (3)
where:
wpqmn =
13.5 csc(3 ) AT = q q4
Jeremi et al. c
293
2 = pq mn (AS spq ) (AT tpq ) + = mn mn + 1 27 cos 3 9 cos 3 81 + s t + spq smn + 4 sin (3) 4 sin3 3 5 sin3 3 pq mn 2q 4 q 4 q 1 81 cos2 3 243 cos 3 81 + tpq tmn + tpq smn 5 sin 3 5 sin3 3 4 q 4 q 4 q 6 sin3 3 1 9 3 cos (3) ppqmn wpqmn + 2 3 sin (3) 2 q sin (3) 2q
Jeremi et al. c
294
Jeremi et al. c
Appendix D
D.1
SimoSerrins Formula
MIJ CKL
In order to derive the analytical gradient of the fourth order tensor MIJKL = (D.1)
MIJ CKL
1 D(A) +
(A) (C 1 )IJ 2 I3 2 (C 1 )IJ 23 I3 (C 1 )IJ + (A) I3 (A) (A) CKL CKL CKL ` 1 D(A) 2 1 2 CIJ I1 2 )IJ (A) IJ + I3 (A) (C D(A) CKL
IIKJL
(D.3)
Derivatives (A) /CKL can be found by starting from equation for CIJ (6.63) and dierentiating it (A) (A) (A) (A) (A) (A) + 2 dNI NJ + 2 NI dNJ dCIJ = 2A d(A) NI NJ A A
A A A (A)
(D.4)
(A)
dNI
(A)
NI
(A)
(D.5)
dCIJ NI
(A)
= 2A d(A)
(D.6)
NJ
(A)
(A)
(D.7)
295
Computational Geomechanics: Lecture Notes It can be proved1 that I2 I3 I1 = IJ ; = I1 KL CKL ; = I3 (C 1 )KL CKL CKL CKL and since I3 = J 2 J 1 = J (C 1 )KL CKL 2
296
(D.8)
(D.9)
With this in mind, equation (D.2) can be rewritten as: MIJ CKL 1 D(A) 1 (A) IIKJL KL IJ + 22 M IJ + (A) 2 KL
+ I3 2 (C 1 )IJ (C 1 )KL 2 I3 (C 1 )IJ MKL + (A) (A) 1 ` 1 (C )IK (C 1 )JL + (C 1 )IL (C 1 )JK 2 I3 + (A) 2 1 D(A) MIJ D(A) CKL where the denition of MIJ from equation (6.86) was used and also: (C 1 )IJ 1` (C 1 ) = (C 1 )IK (C 1 )JL + (C 1 )IL (C 1 )JK = IIJKL CKL 2 Relation (D.11) can be obtained if one starts from the identity:
(A)
(D.10)
(D.11)
CIJ (C 1 )JK = IK
(D.12)
dCIJ (C 1 )JK + CIJ d(C 1 )JK = 0 d(C 1 )JK = (C 1 )JM dCM N (C 1 )N K = 1` = (C 1 )JM (C 1 )KN + (C 1 )JN (C 1 )KM dCM N 2 (C 1 )JK 1` = (C 1 )JM (C 1 )KN + (C 1 )JN (C 1 )KM CM N 2 The derivative of D(A) , that was dened in equation (6.77)as
(D.13)
(D.14)
1 See
Jeremi et al. c
297
is given by:
D(A) CKL
= = = =
83 (A)
(A) (A) (A) I1 I3 2 2(A) I1 + 2 23 I3 (A) CKL CKL (A) CKL CKL (A) CKL
(A) (A) (A)
44 MKL KL 2 2 I1 MKL + I3 (C 1 )KL 2 2 I3 MKL (A) (A) (A) (A) (A) (A) 44 2 I1 2 I3 MKL KL 2 + I3 (C 1 )KL 2 (A) (A) (A) (A) (A)
D(A) MKL KL 2 + I3 (C 1 )KL 2 (A) (A) (A)
(D.15)
where D(A) = 44 2 I1 2 I3 . With the previous derivations, equation (D.10) can be written in expanded form (A) (A) (A)
as:
MIJ CKL
1 D(A)
+ I3 2 (C 1 )IJ (C 1 )KL 2 I3 (C 1 )IJ MKL + (A) (A) 1 ` 1 1 1 + (C )IK (C )JL + (C )IL (C 1 )JK 2 I3 (A) 2
(A)
(A)
(D.16)
If one collects similar terms, equation (D.16), also known as SimoSerrins formula can be written in the nal form as:
MIJ CKL
MIJKL = 1 (A) (A) IIKJL KL IJ + 2 + (A) IJ MKL + MIJ KL D(A) 1 ` 1 (C )IK (C 1 )JL + (C 1 )IL (C 1 )JK + I3 2 (C 1 )IJ (C 1 )KL + (A) 2 (A) (A) (A) (A) 2 I3 (C 1 )IJ MKL + MIJ (C 1 )KL D(A) MIJ MKL (A)
(D.17)
D.2
The volumetric part 2volW /(CIJ CKL ) can be derived by starting from the equation (6.96): Jeremi et al. c Version: April 29, 2010, 9:23
298
J (C 1 )IJ
CKL
vol ` W (C 1 )IJ J 1 1 volW (J) 1 volW J (C 1 )IJ + (C 1 )IJ + J 2 CKL 2 J CKL 2 J CKL = ` 2 vol W J 1 volW 1 1 volW 1 (C 1 ) J (C 1 )IJ + J (C 1 )KL (C 1 )IJ + J IIJKL 2 J J CKL 2 J 2 2 J = 1 volW 1 volW (C 1 ) 1 2 2 volW J (C 1 )KL (C 1 )IJ + J (C 1 )KL (C 1 )IJ + J IIJKL 4 J J 4 J 2 J = 2 volW 1 volW (C 1 ) volW J2 (C 1 )KL (C 1 )IJ + J +J IIJKL J J J 2 J
1 4
(D.18)
D.3
The isochoric part 2isoW /(CIJ CKL ) can be derived by starting from equation (6.97) 2isoW ((A) ) CIJ CKL (A) wA (MIJ )A 1 2 CKL (MIJ )A 1 1 wA (A) (MIJ )A + wA 2 CKL 2 CKL 1 wA B 1 (A) (A) (MIJ )A + wA (MIJKL )A 2 B CKL 2 1 1 wA 1 (B) (A) (A) (B) (MKL )B (MIJ )A + wA (MIJKL )A 2 B 2 2 1 1 (B) (A) (A) YAB (MKL )B (MIJ )A + wA (MIJKL )A 4 2 (D.19) Jeremi et al. c Version: April 29, 2010, 9:23
(A)
299
where equation (D.7) was used and tensor YAB is dened as: YAB = wA (B) B (D.20)
D.4
Derivation of wA
wA = isoW isoW B A = B (A) A (A) (D.21)
(D.22)
From the denition of B in equation (D.22) it follows 1 1 1 B J 3 1 B = B + J 3 = J 3 1 B + J 3 B(A) (A) (A) (A) (A) 3 since
4 4 1 1 1 1 2 3 1 J 3 = J 3 = J 3 J 1 = J 3 1 (A) (A) (A) 3 (A) 3 3 1 1
(D.23)
(D.24)
then wA = =
iso 1 1 1 3 isoW 1 W J + J 3 B (A) B (A) B B(A) (A) 3 isoW 1 isoW (A) B + 3 B (A)
(D.25)
D.5
Derivation of YAB
wA (B) B
and by using equation D.25 wA = we can write: YAB = wA D (B) D B (D.28) 1 isoW isoW (A) C + 3 C (A) (D.27)
Jeremi et al. c
Computational Geomechanics: Lecture Notes By rst considering wA / D we get: iso iso 1 W C + W (A) C 3 (A) wA = D D 1 isoW C isoW (A) 1 2isoW 2isoW C (A) + = + (A) D (A) D 3 C D 3 C D = = 1 2isoW 1 isoW isoW 2isoW C (A) + (A)D CD + 3 C D 3 C (A) D (A) 1 isoW isoW 2isoW 1 2isoW C (A) + (A)D + 3 C D 3 D (A) D (A)
300
(D.29)
Next, from equation D.23, we have that 1 1 1 D = J 3 1 D + J 3 D(B) (B) (B) 3 (D.30)
and by multiplying the result for wA / D from equation D.29 and the result for D /(B) from equation D.30 we obtain: wA D = D (B) + +
1 1 2isoW C J 3 1 D (B) 9 C D 1 isoW 1 1 J 3 (B) D 9 D 1 1 2isoW (A) J 3 1 D (B) 3 (A) D 1 1 isoW (A)D J 3 1 D (B) 3 (A)
+ +
1 1 2isoW C J 3 3 C (B)
1 isoW 1 J 3 3 (B) +
1 2isoW (A) J 3 (A) (B)
(D.31)
where equation D.22 was used. The nal form for YAB is obtained by multiplying equation D.31 with (B) to obtain: YAB = + + wA D = D B (B) 1 2isoW C 1 D (B) (B) 9 C D 1 isoW 1 D (B) 9 D (B)
1 2isoW (A) 1 D (B) (B) 3 (A) D 1 isoW 1 (A) (B) 3 (A) (B) +
By recognizing that 1 (B) 1 and after rearranging elements, we can nally write the equation for YAB as: (B) Jeremi et al. c Version: April 29, 2010, 9:23
301
YAB = +
isoW 2isoW + (A) (A)(B) (B) (A) (B) (A) (B) 2isoW isoW 2isoW isoW C (B) + (B) + (A) (A) D + C (B) (B) (A) D (A) 2iso isoW 1 W C D + D 9 C D D 1 3 ! (D.33)
Jeremi et al. c