Finite Element II, Solid Mechanics - Victor E.saouma - 2001

Draft
DRAFT
Lecture Notes in:
FINITE ELEMENT II
Solid Mechanics
CVEN 6525
c VICTOR
E. SAOUMA,
SPRING 2001
Dept. of Civil Environmental and Architectural Engineering

University of Colorado, Boulder, CO 80309-0428
Draft
02
Victor Saouma
Finite Elements II; Solid Mechanics
Draft
Contents
1 PREREQUISITE
1.1 Variational Formulations . . . . . . . . . . . . . . . . . . . . .
1.2 Finite Element Formulation . . . . . . . . . . . . . . . . . . .
1.2.1 Strain Displacement Relations . . . . . . . . . . . . .
1.2.1.1 Axial Members . . . . . . . . . . . . . . . . .
1.2.1.2 Flexural Members . . . . . . . . . . . . . . .
1.2.2 Virtual Displacement and Strains . . . . . . . . . . . .
1.2.3 Element Stiness Matrix Formulation . . . . . . . . .
1.2.3.1 Stress Recovery . . . . . . . . . . . . . . . .
1.3 Direct Stiness Method . . . . . . . . . . . . . . . . . . . . .
1.3.1 Global Stiness Matrix . . . . . . . . . . . . . . . . .
1.3.1.1 Structural Stiness Matrix . . . . . . . . . .
1.3.1.2 Augmented Stiness Matrix . . . . . . . . .
1.3.2 Logistics . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.2.1 Boundary Conditions, [ID] Matrix . . . . . .
1.3.2.2
LM Vector . . . . . . . . . . . . . . . . . . .
1.3.2.3 Assembly of Global Stiness Matrix . . . . .
E 1-1 Assembly of the Global Stiness Matrix . . . . . . . .
1.3.2.4 Algorithm . . . . . . . . . . . . . . . . . . .
E 1-2 Direct Stiness Analysis of a Truss . . . . . . . . . . .
E 1-3 Analysis of a Frame with MATLAB . . . . . . . . . .
E 1-4 Analysis of a simple Beam with Initial Displacements
2 INTRODUCTION
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . .
2.2 Elliptic, Parabolic and Hyperbolic Equations . . .
E 2-1 Seepage Problem;(Bathe 1996) . . . . . . .
E 2-2 Diusion Problem; (Bathe 1996) . . . . . .
E 2-3 Wave Equation, (Bathe 1996) . . . . . . . .
2.3 Solution of Discrete-System Mathematical models .
2.3.1 Steady State Problems . . . . . . . . . . . .
2.3.1.1 Elastic Spring . . . . . . . . . . .
2.3.1.2 Heat Transfer . . . . . . . . . . .
2.3.1.3 Hydraulic Network . . . . . . . . .
2.3.1.4 DC Network . . . . . . . . . . . .
2.3.2 Equivalent Truss/Direct Stiness Models
2.3.2.1 Nonlinear Elastic Spring . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
11
. 11
. 11
. 11
. 11
. 14
. 15
. 15
. 16
. 16
. 16
. 17
. 17
. 18
. 18
. 110
. 110
. 110
. 112
. 113
. 118
. 120
.
.
.
.
.
.
.
.
.
.
.
.
.
21
. 21
. 21
. 23
. 25
. 26
. 28
. 28
. 29
. 210
. 211
. 212
. 213
. 215
Draft
02
2.3.3
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 215
. 216
. 216
. 217
. 217
. 218
. 220
. 221
. 221
. 223
. 223
3 FUNDAMENTAL RELATIONS
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.1.1 Indicial Notation . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.1.2 Tensor Notation . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.1.3 Voigt Notation . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.2 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2 Vector Fields; Solid Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1 Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1.1 Force, Traction and Stress Vectors . . . . . . . . . . . . . . .
3.2.1.2 Traction on an Arbitrary Plane; Cauchys Stress Tensor . . .
E 3-1 Stress Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.2 Kinematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3 Fundamental Laws of Continuum Mechanics . . . . . . . . . . . . . .
3.2.3.1 Conservation of Mass; Continuity Equation . . . . . . . . . .
3.2.3.2 Linear Momentum Principle; Equation of Motion . . . . . .
3.2.3.3 Conservation of Energy; First Principle of Thermodynamics
3.2.4 Constitutive Equations . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.4.1 General 3D . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.4.2 Transversly Isotropic Case . . . . . . . . . . . . . . . . . . .
3.2.4.3 Special 2D Cases . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.4.3.1 Plane Strain . . . . . . . . . . . . . . . . . . . . . .
3.2.4.3.2 Axisymmetry . . . . . . . . . . . . . . . . . . . . . .
3.2.4.3.3 Plane Stress . . . . . . . . . . . . . . . . . . . . . .
3.2.4.4 Pore Pressures . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.5 Field Equations for Thermo- and Poro Elasticity . . . . . . . . . . .
3.3 Scalar Field: Diusion Equation . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.1 Heat Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.2 Derivation of the Diusion Problem . . . . . . . . . . . . . . . . . . .
3.3.2.1 Simple 2D Derivation . . . . . . . . . . . . . . . . . . . . . .
3.3.2.2 Generalized Derivation . . . . . . . . . . . . . . . . . . . . .
3.3.2.3 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . .
3.4 Summary and Tonti Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
31
. 31
. 31
. 31
. 33
. 33
. 33
. 35
. 35
. 35
. 36
. 37
. 37
. 310
. 312
. 312
. 313
. 314
. 314
. 315
. 316
. 316
. 316
. 316
. 317
. 317
. 319
. 320
. 321
. 321
. 322
. 324
. 324
2.4
2.5
2.6
Propagation Problems . . . . . . .
2.3.3.1 Dynamic Elastic System
2.3.3.2 Transient Heat Flow . . .
2.3.4 Eigenvalue Problems . . . . . . . .
2.3.4.1 Free Vibration . . . . . .
2.3.4.2 Column Buckling . . . .
Solution Strategies . . . . . . . . . . . . .
2.4.1 Euler Equation . . . . . . . . . . .
E 2-4 Flexure of a Beam . . . . . . . . .
Computer Programs . . . . . . . . . . . .
Examples of applications . . . . . . . . . .
CONTENTS
Victor Saouma
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Draft
CONTENTS
4 MESH GENERATION
4.1 Introduction . . . . . . . . . . . .
4.2 Triangulation . . . . . . . . . . .
4.2.1 Voronoi Polygon . . . . .
4.2.2 Delaunay Triangulation .
4.2.3 MATLAB Code . . . . . .
4.3 Finite Element Mesh Generation
4.3.1 Boundary Denition . . .
4.3.2 Interior Node Generation
4.3.3 Final Triangularization .
03
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5 VARIATIONAL and RAYLEIGH-RITZ METHODS

5.1 Multield Variational Principles . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2 Total Potential Energy Principle . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.1 Static; Euler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.2 Dynamic; Euler/Lagrange . . . . . . . . . . . . . . . . . . . . . . . . . .
E 5-1 Hamiltons Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3 General Hu-Washizu Variational Principle . . . . . . . . . . . . . . . . . . . . .
5.4 Rayleigh Ritz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.1 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
E 5-2 Uniformly Loaded Simply Supported Beam; Polynomial Approximation
E 5-3 Heat Conduction; (Bathe 1996) . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
41
41
42
42
43
43
43
43
44
46
51
. 51
. 52
. 52
. 54
. 55
. 59
. 513
. 515
. 515
. 516
6 INTERPOLATION FUNCTIONS; NATURAL COORDINATE SYSTEMS 61

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.2 Cartesian Coordinate System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.2.1 C 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
6.2.1.1 Truss element . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
6.2.1.2 Generalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
6.2.1.3 Constant Strain Triangle Element . . . . . . . . . . . . . . . . . 63
6.2.1.4 Further Generalization: Lagrangian Interpolation Functions . . . 65
6.2.1.5 Rectangular Bilinear Element . . . . . . . . . . . . . . . . . . . . 66
6.2.1.6 Solid Rectangular Trilinear Element . . . . . . . . . . . . . . . . 67
6.2.2 C 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
6.2.2.1 Flexural . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
6.2.2.2 C 1 : Hermitian Interpolation Functions . . . . . . . . . . . . . . 69
6.2.3 Characteristics of Shape Functions . . . . . . . . . . . . . . . . . . . . . . 610
6.3 Natural Coordinate System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 610
6.3.1 Straight Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 611
6.3.2 Triangular Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 612
6.3.3 Volume Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614
6.3.4 Interpolation Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615
6.4 Pascals Triangle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 616
7 FINITE ELEMENT DISCRETIZATION and REQUIREMENTS
71
7.1 Discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
7.1.1 Discretization of the Variational Statement for the General TPE Variational Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Victor Saouma
Draft
04
7.2
7.3
7.4
7.5
CONTENTS
7.1.2 Discretization of the Variational Statement for the HW Variational Principle73

General Element Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Discretization Error and Convergence Rate . . . . . . . . . . . . . . . . . . . . . 78
Lower Bound Character of Minimum Potential Energy Based Solutions . . . . . 710
Equilibrium and Compatibiliy in the Solution . . . . . . . . . . . . . . . . . . . . 710
8 STRAIGHT SIDED ELEMENTS; 1st GENERATION

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2 Rod Elements . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.1 Truss Element . . . . . . . . . . . . . . . . . . . .
8.2.2 Beam Element . . . . . . . . . . . . . . . . . . . .
8.3 Triangular Elements . . . . . . . . . . . . . . . . . . . . .
8.3.1 Cartesian Coordinate System (CST) . . . . . . . .
8.3.2 Natural Coordinate System . . . . . . . . . . . . .
8.3.2.1 Linear, T3 . . . . . . . . . . . . . . . . .
8.3.2.2 Quadratic Element (T6) . . . . . . . . . .
8.4 Bilinear Rectangular Element . . . . . . . . . . . . . . . .
8.5 Element Assessment . . . . . . . . . . . . . . . . . . . . .
8.5.1 CST . . . . . . . . . . . . . . . . . . . . . . . . . .
8.5.2 BiLinear Rectangular . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
9 ISOPARAMETRIC ELEMENTS; 2nd GENERATION

9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.2 Element Formulation . . . . . . . . . . . . . . . . . . . . . . . .
9.2.1 Bar Element . . . . . . . . . . . . . . . . . . . . . . . .
9.2.2 Quadrilaterals . . . . . . . . . . . . . . . . . . . . . . .
9.2.2.1 Linear Element (Q4) . . . . . . . . . . . . . . .
9.2.2.1.1 Example: Jacobian Operators, (Bathe
9.2.2.2 Quadratic Element . . . . . . . . . . . . . . . .
9.2.2.2.1 Serendipity Element (Q8) . . . . . . .
9.2.2.2.2 Lagrangian element (Q9) . . . . . . .
9.2.2.2.3 Variable (Hierarchical) Element . . .
9.2.3 Triangular Elements . . . . . . . . . . . . . . . . . . . .
9.3 Numerical Integration . . . . . . . . . . . . . . . . . . . . . . .
9.3.1 Newton-Cotes . . . . . . . . . . . . . . . . . . . . . . . .
9.3.2 Gauss-Legendre Quadrature . . . . . . . . . . . . . . . .
9.3.2.1 Legendre Polynomial . . . . . . . . . . . . .
9.3.2.2 Gauss-Legendre Quadrature for n = 2 . . . . .
9.3.3 Rectangular and Prism Regions . . . . . . . . . . . . . .
9.3.4 Triangular Regions . . . . . . . . . . . . . . . . . . . . .
9.4 Stress Recovery; Nodal Stresses . . . . . . . . . . . . . . . . . .
9.5 Nodal Equivalent Loads . . . . . . . . . . . . . . . . . . . . . .
9.5.1 Gravity Load . . . . . . . . . . . . . . . . . . . . . . . .
9.5.2 Traction Load . . . . . . . . . . . . . . . . . . . . . . . .
9.5.3 Initial Strains/Stresses; Thermal Load . . . . . . . . . .
9.6 Computer Implementation . . . . . . . . . . . . . . . . . . . . .
9.6.1 Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . .
9.6.2 MATLAB Code . . . . . . . . . . . . . . . . . . . . . . .
Victor Saouma
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
81
. 81
. 81
. 81
. 82
. 83
. 83
. 85
. 85
. 86
. 87
. 89
. 89
. 810
. . . .
. . . .
. . . .
. . . .
. . . .
1996)
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
91
. 91
. 93
. 93
. 95
. 95
. 910
. 911
. 911
. 913
. 915
. 916
. 917
. 918
. 920
. 921
. 922
. 922
. 922
. 924
. 925
. 925
. 926
. 928
. 929
. 929
. 930
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Draft
CONTENTS
9.6.3
05
9.6.2.1
9.6.2.2
9.6.2.3
9.6.2.4
9.6.2.5
Plott of
stiff.m . . . . .
dmat.m . . . . .
sfr.m . . . . . .
jacob.m . . . . .
bmatps.m . . . .
Shape Functions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 930
. 932
. 933
. 935
. 935
. 936
10 ELEMENT FORMULATION and STRAIN RECOVERY in HW FORMULATION

101
10.1 Element Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
10.2 Strain Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
10.2.1 C-lumping. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
10.2.2 Strain smoothing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
10.2.3 C-splitting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
10.3 Uniqueness and Existence of a Solution . . . . . . . . . . . . . . . . . . . . . . . 105
11 WEIGHTED RESIDUAL METHODS
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.2 General Formulation . . . . . . . . . . . . . . . . . . . . . . . .
11.2.1 Dierential Operators . . . . . . . . . . . . . . . . . . .
11.2.1.1 Application to 1D Axial Member . . . . . . . .
11.2.2 Residual Formulation . . . . . . . . . . . . . . . . . . .
11.3 Weighted Residual Methods . . . . . . . . . . . . . . . . . . . .
11.3.1 Point Collocation Method . . . . . . . . . . . . . . . .
11.3.2 Subdomain Collocation Method . . . . . . . . . . . . .
11.3.3 Least-Squares Method . . . . . . . . . . . . . . . . . .
11.3.4 Galerkin Method . . . . . . . . . . . . . . . . . . . . . .
E 11-1 String Vibration . . . . . . . . . . . . . . . . . . . . . .
11.4 Applications of the Galerkin Method to 3D Elasticity Problems
11.4.1 Derivation of the Weak Form . . . . . . . . . . . . . . .
11.4.2 FE Discretization . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
111
. 111
. 111
. 111
. 112
. 112
. 114
. 114
. 114
. 115
. 115
. 115
. 117
. 117
. 119
12 FINITE ELEMENT DISCRETIZATION OF THE FIELD

12.1 Derivation of the Weak Form . . . . . . . . . . . . . . . . . .
12.2 FE Discretization . . . . . . . . . . . . . . . . . . . . . . . . .
12.2.1 No Convection . . . . . . . . . . . . . . . . . . . . . .
12.2.2 Convection . . . . . . . . . . . . . . . . . . . . . . . .
12.3 Illustrative Examples . . . . . . . . . . . . . . . . . . . . . . .
E 12-1 Composite Wall . . . . . . . . . . . . . . . . . . . . . .
E 12-2 Heat Transfer across a Fin . . . . . . . . . . . . . . . .
12.4 Comparison Between Vector and Scalar Formulations . . . . .
EQUATION
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
.
.
.
.
.
.
.
.
121
. 121
. 122
. 123
. 124
. 124
. 124
. 127
. 1212
13 TOPICS in STRUCTURAL MECHANICS

13.1 Condensation/Substructuring . . . . . . . . .
13.2 Element Evaluation . . . . . . . . . . . . . . .
13.2.1 Patch Test . . . . . . . . . . . . . . .
13.2.2 Eigenvalue Test . . . . . . . . . . . . .
13.2.3 Order of Integration . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
131
. 131
. 132
. 132
. 132
. 134
Victor Saouma
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Draft
06
CONTENTS
13.2.3.1 Full Integration . . . . . . . .

13.2.3.2 Reduced Integration . . . . . .
13.2.3.3 Selective Reduced Integration
13.3 Parasitic Shear/Incompatible Elements . . . . .
13.3.1 Q4, The Problem . . . . . . . . . . . . .
13.3.2 Q6, The Solution . . . . . . . . . . . . .
13.3.3 QM6, Further Enhancements . . . . . .
13.4 Rotational D.O.F. . . . . . . . . . . . . . . . .
13.5 Constraints . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 134
. 135
. 136
. 137
. 137
. 138
. 138
. 139
. 1310
14 GEOMETRIC NONLINEARITY
14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . .
14.1.1 Strong Form . . . . . . . . . . . . . . . . . . . .
14.1.1.1 Lower Order Dierential Equation . . .
14.1.1.2 Higher Order Dierential Equation . .
14.1.2 Weak Form . . . . . . . . . . . . . . . . . . . . .
14.1.2.1 Strain Energy . . . . . . . . . . . . . .
14.1.2.2 Euler Equation . . . . . . . . . . . . . .
14.2 Finite Element Discretization . . . . . . . . . . . . . . .
14.3 Elastic Instability; Bifurcation Analysis . . . . . . . . .
E 14-1 Column Stability . . . . . . . . . . . . . . . . . .
E 14-2 Frame Stability . . . . . . . . . . . . . . . . . . .
14.4 Second-Order Elastic Analysis; Geometric Non-Linearity
E 14-3 Eect of Axial Load on Flexural Deformation . .
E 14-4 Bifurcation . . . . . . . . . . . . . . . . . . . . .
14.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
141
. 141
. 142
. 142
. 143
. 146
. 146
. 147
. 148
. 149
. 1410
. 1413
. 1415
. 1416
. 1419
. 1422
15 PLATES
15.1 Fundamental Relations . . . . . . . . . . . . . . . . .
15.1.1 Equilibrium . . . . . . . . . . . . . . . . . . .
15.1.2 Kinematic Relations . . . . . . . . . . . . . .
15.1.3 Constitutive Relations . . . . . . . . . . . . .
15.2 Plate Theories . . . . . . . . . . . . . . . . . . . . .
15.2.1 Reissner-Mindlin . . . . . . . . . . . . . . . .
15.2.1.1 Fundamental Relations . . . . . . .
15.2.1.2 Dierential Equation . . . . . . . .
15.2.1.3 Variational Formulation . . . . . .
15.2.2 Kirchho . . . . . . . . . . . . . . . . . . . .
15.2.2.1 Fundamental Relations . . . . . . .
15.2.2.2 Dierential Equation . . . . . . . .
15.2.2.3 Stresses . . . . . . . . . . . . . . . .
15.2.2.4 Variational Formulation . . . . . . .
15.2.3 Summary . . . . . . . . . . . . . . . . . . . .
15.3 Finite Element Formulations . . . . . . . . . . . . .
15.3.1 Rectangular Element . . . . . . . . . . . . . .
15.3.1.1 Formulation . . . . . . . . . . . . .
15.3.1.2 Shear Locking . . . . . . . . . . . .
15.3.2 Nonconforming Kirchho Triangular Element
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
151
. 151
. 151
. 154
. 155
. 156
. 156
. 156
. 158
. 158
. 159
. 159
. 1510
. 1510
. 1511
. 1511
. 1513
. 1513
. 1513
. 1516
. 1517
Victor Saouma
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Draft
CONTENTS
15.3.2.1 Formulation . . .
15.3.2.2 Nonconformity . .
15.3.3 Discrete Kirchho Triangle
15.4 Summary . . . . . . . . . . . . . .
07
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 1517
. 1519
. 1520
. 1524
16 MATERIAL NONLINEARITIES
16.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16.1.1 Linearization . . . . . . . . . . . . . . . . . . . . . . .
16.1.2 Solution Strategies . . . . . . . . . . . . . . . . . . . .
16.2 Load Control . . . . . . . . . . . . . . . . . . . . . . . . . . .
16.2.1 Newton-Raphson . . . . . . . . . . . . . . . . . . . . .
16.2.1.1 Newton-Raphson/Tangent Stiness Method .
16.2.1.2 Modied Newton-Raphson . . . . . . . . . .
16.2.1.3 Secant Newton . . . . . . . . . . . . . . . . .
16.2.2 Acceleration of Convergence, Line Search Method . .
16.2.3 Convergence Criteria . . . . . . . . . . . . . . . . . . .
16.3 Direct Displacement Control . . . . . . . . . . . . . . . . . .
16.4 Indirect Displacement Control . . . . . . . . . . . . . . . . . .
16.4.1 Partitioning of the Displacement Corrections . . . . .
16.4.2 Arc-Length . . . . . . . . . . . . . . . . . . . . . . . .
16.4.3 Relative Displacement Criterion . . . . . . . . . . . .
16.4.4 IDC Methods with Approximate Line Searches . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
161
. 161
. 161
. 162
. 164
. 164
. 164
. 166
. 166
. 167
. 1610
. 1610
. 1613
. 1613
. 1615
. 1617
. 1618
A VECTOR OPERATIONS
A.1 Vector Dierentiation . . . . . . . . . . . . . . . . . . . . .
A.1.1 Derivative WRT to a Scalar . . . . . . . . . . . . . .
E A-1 Tangent to a Curve . . . . . . . . . . . . . . . . . .
A.1.2 Divergence . . . . . . . . . . . . . . . . . . . . . . .
E A-2 Divergence . . . . . . . . . . . . . . . . . . . . . . .
A.1.3 Gradient . . . . . . . . . . . . . . . . . . . . . . . . .
A.1.4 Scalar . . . . . . . . . . . . . . . . . . . . . . . . . .
E A-3 Gradient of a Scalar . . . . . . . . . . . . . . . . . .
E A-4 Stress Vector normal to the Tangent of a Cylinder .
A.2 Vector Integrals . . . . . . . . . . . . . . . . . . . . . . . . .
A.2.1 Integral of a Vector . . . . . . . . . . . . . . . . . . .
A.2.2 Line Integral . . . . . . . . . . . . . . . . . . . . . .
A.2.3 Integration by Parts . . . . . . . . . . . . . . . . . .
A.2.4 Gauss; Divergence Theorem . . . . . . . . . . . . . .
A.2.5 Stokes Theorem . . . . . . . . . . . . . . . . . . . .
A.2.6 Green; Gradient Theorem . . . . . . . . . . . . . . .
E A-5 Physical Interpretation of the Divergence Theorem
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
A1
. A1
. A1
. A2
. A3
. A5
. A5
. A5
. A6
. A6
. A9
. A9
. A9
. A9
. A10
. A10
. A10
. A10
B CASE-STUDY: FRACTURING of A DAM DUE TO

B.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . .
B.1.1 Elastic and Thermal Properties . . . . . . . . . .
B.1.2 Loads . . . . . . . . . . . . . . . . . . . . . . . .
B.2 ANALYSIS II; Thermal Shock . . . . . . . . . . . . .
B.2.1 Thermal Analysis . . . . . . . . . . . . . . . . . .
Victor Saouma
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
THERMAL
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
LOAD
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
B1
. B1
. B1
. B2
. B2
. B3
Draft
08
CONTENTS
B.2.2 Stress Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B7

B.2.3 Data Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B9
B.3 CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B9
C MISC.
C1
C.1 Units & Conversion Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C1
C.2 Metric Prexes and Multipliers . . . . . . . . . . . . . . . . . . . . . . . . . . . . C2
Victor Saouma
Draft
List of Figures
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
Summary of Variational Methods . . . . . . . . .

Duality of Variational Principles . . . . . . . . .
Frame Example . . . . . . . . . . . . . . . . . . .
Example for [ID] Matrix Determination . . . . .
Simple Frame Analyzed with the MATLAB Code
. . . . . . . . . . . . . . . . . . . . . . . . . . .
Simple Frame Analyzed with the MATLAB Code
Stiness Analysis of one Element Structure . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
2.11
2.12
Finite Element Process, (Bathe 1996) . . . . . . . . . . . . . . . . . . . . . . . . 22

Seepage Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
One Dimensional Heat Conduction . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Rod subjected to Step Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
System of Rigid Carts Interconnected by Linear Springs, (Bathe 1996) . . . . . . 29
Slab Subjected to Temperature Boundary Conditions, (Bathe 1996) . . . . . . . 210
Pipe Network, (Bathe 1996) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
DC Network, (Bathe 1996) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
Equivalent Trusses/Direct Stiness . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Heat Transfer Idealization in an Electron Tube, (Bathe 1996) . . . . . . . . . . . 217
Stability of a Two Rigid Bars System . . . . . . . . . . . . . . . . . . . . . . . . 218
Idealization, Discretization and Solution of a Numerical Simulation, (Felippa 2000)220
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
Stresses as Tensor Components . . . . . . . . . . . .

Cauchys Tetrahedron . . . . . . . . . . . . . . . . .
Flux Through Area dS . . . . . . . . . . . . . . . . .
Equilibrium of Stresses, Cartesian Coordinates . . .
Flux vector . . . . . . . . . . . . . . . . . . . . . . .
Flux Through Sides of Dierential Element . . . . .
*Flow through a surface . . . . . . . . . . . . . . .
Components of Tontis Diagram, (Felippa 2000) . . .
Fundamental Equations of Solid Mechanics and Heat
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 36
. 36
. 311
. 313
. 320
. 322
. 322
. 324
. 325
4.1
4.2
4.3
4.4
Voronoi and Delaunay Tessellation . . . . . . . . . . . . . . . . . . . . . . . .

Control Point for a 2D Mesh . . . . . . . . . . . . . . . . . . . . . . . . . . .
Control Point for a 3D Mesh . . . . . . . . . . . . . . . . . . . . . . . . . . .
A Two Dimensional Triangularization AlgorithmControl Point for a 3D Mesh
.
.
.
.
.
.
.
.
5.1
5.2
Tonti Diagram for the Total Potential Energy, (Cervenka,

J. 1994) . . . . . . . . 54
Tonti Diagram for Hu-Washizu, (Cervenka,

J. 1994) . . . . . . . . . . . . . . . . 512
. . .
. . .
. . .
. . .
. . .
. . .
. . .
. . .
Flow
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 12
. 13
. 17
. 19
. 111
. 113
. 118
. 121
42
44
45
45
Draft
02
LIST OF FIGURES
5.3
Uniformly Loaded Simply Supported Beam Analysed by the Rayleigh-Ritz Method515
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
Axial Finite Element . . . . . . . . . . . . . . . . . . . .

Constant Strain Triangle Element . . . . . . . . . . . . .
Rectangular Bilinear Element . . . . . . . . . . . . . . .
Solid Trilinear Rectangular Element . . . . . . . . . . .
Flexural Finite Element . . . . . . . . . . . . . . . . . .
Shape Functions for Flexure of Uniform Beam Element.
Natural Coordinate System Along a Straight Line . . .
Natural Coordinate System for a Triangle . . . . . . . .
Integration over a Triangle . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 62
. 64
. 66
. 67
. 68
. 610
. 611
. 612
. 613
7.1
7.2
7.3
7.4
Completness and Compatibility .

h, p and r Convergence . . . . .
Interelement Continuity of Stress
Interelement Continuity of Strain
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 78
. 79
. 711
. 711
8.1
8.2
Rectangular Bilinear Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Linear Triangular Element Subjected to Pure Bending . . . . . . . . . . . . . . . 810
9.1
9.2
9.3
9.4
9.5
9.6
9.7
9.8
9.9
9.10
9.11
9.12
9.13
9.14
9.15
9.16
9.17
9.18
9.19
9.20
9.21
Two-Dimensional Mapping of Some Elements . . . . . . . . . . . . . . . . . . . . 91

Actual and Parent Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Iso, Super, and Sub Parametric Elements . . . . . . . . . . . . . . . . . . . . . . 93
Three-Noded Quadratic Bar Element . . . . . . . . . . . . . . . . . . . . . . . . . 93
Four Noded Isoparametric Element . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Dierential Element in Curvilinear Coordinate System . . . . . . . . . . . . . . . 98
Cross Product of Two Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
*Elements with Possible Singular Jacobians . . . . . . . . . . . . . . . . . . . . . 99
Serendipity and Lagrangian Quadratic Quadrilaterals . . . . . . . . . . . . . . . 911
Pascal Triangles for Quadrilateral and Triangle Elements . . . . . . . . . . . . . . 912
Serendipity Isoparametric Quadratic Finite Element: Global and Parent Element 913
Shape Functions for 8 Noded Quadrilateral Element . . . . . . . . . . . . . . . . 914
Shape Functions for 9 Noded Quadrilateral Element . . . . . . . . . . . . . . . . 915
Nodal Numbering for Isoparameteric Elements . . . . . . . . . . . . . . . . . . . 916
Newton-Cotes Numerical integration . . . . . . . . . . . . . . . . . . . . . . . . . 918
Gauss-Legendre Integration Over a Surface . . . . . . . . . . . . . . . . . . . . . 923
Numerical Integration Over a Triangle . . . . . . . . . . . . . . . . . . . . . . . . 923
Extrapolation from 4-Node Quad Gauss Points (Felippa 1999) . . . . . . . . . . . 924
Gravity Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 926
Traction Load in Isoparametric Elements . . . . . . . . . . . . . . . . . . . . . . 927
Traction Load in Contiguous Isoparametric Elements . . . . . . . . . . . . . . . . 928
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
10.1 Patch test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

12.1 Heat Flow in a Thin Rectangular Fin
. . . . . . . . . . . . . . . . . . . . . . . . 129
13.1 Patch Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

13.2 Eigenvectors Corresponding to a) Non-Zero and b) Zero Eigenvalues for a Square
Bilinear element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Victor Saouma
Draft
LIST OF FIGURES
03
13.3 Independent Displacement Modes for a Bilinear Element . . . . . . . . . .

13.4 Hourglas Modes in Under-Integrated Quadratic Element . . . . . . . . . .
13.5 Rectangular Bilinear Element Subjected to Bending; Bilinear Element and
rect Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.6 Displacements Associated with Incompatible Modes for the Q6 Element .
13.7 Side Displacements Induced by Drilling d.o.f. . . . . . . . . . . . . . . . .
. . .
. . .
Cor. . .
. . .
. . .
14.1
14.2
14.3
14.4
. 136
. 136
. 137
. 138
. 139
Level of Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Euler Column . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Simply Supported Beam Column; Dierential Segment; Eect of Axial Force P
Solution of the Tanscendental Equation for the Buckling Load of a Fixed-Hinged
Column . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14.5 Summary of Stability Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . .
. 145
. 1423
15.1
15.2
15.3
15.4
15.5
15.6
15.7
15.8
15.9
.
.
.
.
.
.
.
.
.
. 151
. 152
. 153
. 154
. 156
. 1513
. 1517
. 1520
. 1521
16.1 Test Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

16.2 Newton-Raphson Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16.3 Modied Newton-Raphson Method, Initial Tangent in Increment . . . . . . . .
16.4 Modied Newton-Raphson Method, Initial Problem Tangent . . . . . . . . . .
16.5 Incremental Secant, Quasi-Newton Method . . . . . . . . . . . . . . . . . . . .
16.6 Schematic of Line Search, (Reich 1993) . . . . . . . . . . . . . . . . . . . . . . .
16.7 Flowchart for Line Search Algorithm, (Reich 1993) . . . . . . . . . . . . . . . .
16.8 Divergence of Load-Controled Algorithms . . . . . . . . . . . . . . . . . . . . .
16.9 Hydrostatically Loaded Gravity Dam . . . . . . . . . . . . . . . . . . . . . . . .
16.10Load-Displacement Diagrams with Snapback . . . . . . . . . . . . . . . . . . .
16.11Flowchart for an incremental nonlinear nite element program with indirect displacement control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16.12Two points on the load-displacement curve satisfying the arc-length constraint
16.13Flow chart for line search with IDC methods . . . . . . . . . . . . . . . . . . .
. 163
. 165
. 166
. 167
. 168
. 169
. 169
. 1611
. 1613
. 1613
. 1616
. 1617
. 1620
A.1
A.2
A.3
A.4
A.5
A.6
A.7
A.8
. A2
. A2
. A3
. A4
. A4
. A7
. A8
. A11
Finite Element Formulation . . . . . . . . . . . . . . . .

Stresses in a Plate . . . . . . . . . . . . . . . . . . . . .
Free Body Diagram of an Innitesimal Plate Element .
Displacements in a Plate . . . . . . . . . . . . . . . . . .
Positive Moments and Rotations . . . . . . . . . . . . .
Rectangular Plate Element . . . . . . . . . . . . . . . .
Triangular Plate Element in Natural Coordinate System
Edges of Adjacent Triangular Elements . . . . . . . . .
Discrete Kirchho Triangular Element . . . . . . . . . .
.
.
.
.
.
.
.
.
.
Dierentiation of position vector p . . . . . . . . . . . . .

Curvature of a Curve . . . . . . . . . . . . . . . . . . . . .
Vector Field Crossing a Solid Region . . . . . . . . . . . .
Flux Through Area dA . . . . . . . . . . . . . . . . . . . .
Innitesimal Element for the Evaluation of the Divergence
Radial Stress vector in a Cylinder . . . . . . . . . . . . . .
Gradient of a Vector . . . . . . . . . . . . . . . . . . . . .
Physical Interpretation of the Divergence Theorem . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 141
. 142
. 144
B.1 Boundary Description of Dam for Transient Thermal Analysis . . . . . . . . . . . B4

Victor Saouma
Draft
04
LIST OF FIGURES
B.2 Heat of Hydration Interpolations . . . . . . . . . . . . . . . . . . . . . . . . . . . B6

B.3 Temperature Distribution in the Transient Thermal Analysis at Day 8 . . . . . . B7
B.4 Maximum Principal Stresses and Deformed Mesh at Day 8 . . . . . . . . . . . . B8
Victor Saouma
Draft
List of Tables
1.1
Summary of Variational Terms Associated with One Dimensional Elements . . . 14
3.1
3.2
3.3
Selected Examples of Diusion Problems . . . . . . . . . . . . . . . . . . . . . . . 319

Comparison of Scalar and Vector Field Problems . . . . . . . . . . . . . . . . . . 326
Classication of various Physical Problems, (Kardestuncer 1987) . . . . . . . . . 327
5.1
5.2
Functionals in Linear Elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Comparison Between Total Potential Energy and Hu-Washizu Formulations . . . 512
6.1
6.2
6.3
Characteristics of Beam Element Shape Functions . . . . . . . . . . . . . . . . . 69

Interpretation of Shape Functions in Terms of Polynomial Series (1D & 2D) . . . 616
Polynomial Terms in Various Element Formulations (1D & 2D) . . . . . . . . . . 616
8.1
Shape Functions and Derivatives for T6 Element . . . . . . . . . . . . . . . . . . 86
9.1
9.2
9.3
9.4
Shape Functions, and Natural Derivatives for Q8 Element . . . . . . . . . . . .

Shape Functions for Variable Node Elements . . . . . . . . . . . . . . . . . . .
Weights for Newton-Cotes Quadrature Formulas . . . . . . . . . . . . . . . . .
Integration Points and Weights for Gauss-Quadrature Formulaes Over the Interval [1, 1] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Coordinates and Weights for Numerical Integration over a Triangle . . . . . . .
Natural Coordinates of Bilinear Quadrilateral Nodes . . . . . . . . . . . . . . .
9.5
9.6
. 913
. 916
. 919
. 921
. 923
. 925
10.1 Polynomial orders of the shape functions. . . . . . . . . . . . . . . . . . . . . . . 101

10.2 Table of coecients and spectral radii for CS technique. . . . . . . . . . . . . . 105
12.1 Comparison of Scalar and Vector Field Problems, Revisited . . . . . . . . . . . . 1212
13.1 Full and Reduced Numerical Integrations for Quadrilateral Elements . . . . . . . 135
13.2 Bilinsear and Exact Displacements/Strains . . . . . . . . . . . . . . . . . . . . . 137
15.1 Comparison of Governing Equations in Elasticity and Plate Bending . . . . . . . 1512
15.2 Integration Rules for Mindlin Plate Elements . . . . . . . . . . . . . . . . . . . . 1516
A.1 Similarities Between Multiplication and Dierentiation Operators . . . . . . . . . A1
B.1
B.2
B.3
B.4
B.5
Concrete Material Properties . . . . .

Thermal Properties of the concrete . .
Interface Element Material Properties
Loads applied on the Dam . . . . . . .
Heat of Hydration From the Literature
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. B2
. B2
. B2
. B3
. B5
Draft
02
LIST OF TABLES
B.6 Heat of Hydration Adopted in the Simulation; Days and J/Kg/Day . . . . . . . . B5

B.7 Sresses Along the Interface Element; m] and [Pa] . . . . . . . . . . . . . . . . . . B8
B.8 Crack Opening and Sliding Displacements; [m] . . . . . . . . . . . . . . . . . . . B9
Victor Saouma
Draft
LIST OF TABLES
03
NOTATION
A
c
E
h
I
J
L
Q
t
T
u, v, w
U0
U
U0
U
W
M
P
U
W
SCALARS
Area
Specic heat
Elastic Modulus
Film coecient for convection heat transfer
Moment of inertia
St Venants torsional constant
Length
Rate of internal heat generation per unit volume
Time
Temperature
Translational displacements along the x, y, and z directions
Strain energy density
Strain energy
Complementary strain energy density
Complementary strain energy
Work
Potential energy
Coecient of thermal expansion
Shear modulus
Poissons ratio
mass density
Rotational displacement
Virtual moment
Virtual force
Virtual rotation
Virtual displacement
Virtual curvature
Virtual internal strain energy
Virtual external work
TENSORS order 1
a
b
c
F
p
N
N
p
P
q
R
R
t
Victor Saouma
Vector of coecients in assumed displacement eld

Body force
Nodal coordinates
Unknown element forces and unknown support reactions
Matrix of coecients of a polynomial series
Displacement shape functions
Coordinate shape functions
Element nodal forces = F
Structure nodal forces
Flux per unit area
Structure reactions
Residuals
Traction vector
Draft
04

t
u
(x)
u
u
ue
u
u
V
LIST OF TABLES
Specied tractions along t
Displacement vector
Specied displacements along u
Displacement vector
Nodal element displacements
Nodal displacements in a continuous system
Structure nodal displacements
Shear forces in a plate Vx , Vy
Element nodal displacements
Virtual strain vector
Virtual stress vector
Stress vector
Initial stress vector
TENSORS order 2
d
I
k
kg
K
Kg
lij
M
N
0
k
Element exibility matrix

Idendity matrix
Element stiness matrix
Geometric element stiness matrix
Structure stiness matrix
Structures geometric stiness matrix
Direction cosine of rotated axis i with respect to original axis j
Moments in a plate Mxx , Mxy , Myx , Myy
Membrane forces Nxx , Nxy , Nyx , Nyy
Shear deformations
Transformation matrix
Strain vector
Initial strain vector
Conductivity
Curvature
TENSORS order 4
Constitutive matrix
CONTOURS, SURFACES, VOLUMES
t
u
T
c
q
Surface
Boundary along
Boundary along
Boundary along
Boundary along
Boundary along
Volume of body
which
which
which
which
which
surface tractions, t are specied

displacements, u are specied
temperatures, T are specied
convection ux, qc are specied
ux, qn are specied
FUNCTIONS, OPERATORS
Victor Saouma
Draft
LIST OF TABLES
B
L
.u = div .u
v 2
. . .
05
Neighbour function to u(x)

Variational operator
Discrete strain-displacement operator
Linear dierential operator relating displacement to strains
T
Divergence, (gradient operator) on scalar
x
y
z
u
y
ux
z
Divergence, (gradient operator) on vector x + y + u
z
Euclidian norm.
Innity norm.
PROGRAM ARRAYS
ID
LM
Victor Saouma
Matrix relating nodal dof to structure dof

structure dof of nodes connected to a given element
Draft
06
Victor Saouma
LIST OF TABLES
Draft
LIST OF TABLES
1
2
3
4
5
6
7
8
9
10
12
13
14
15
16
Jan. 16
18
23
25
30
Feb. 1
6
8
13
15
20
22
27
29
Mar. 5
7
12
14
19
21
Apr. 2
4
9
11
16
18
23
25
30
May 2
Victor Saouma
07
Introduction; Course objective; Overview; Notation.

Mathematical Formulations.
Elasticity
Direct Method; Field Eq.
Variational Methods
Mesh Generators; Laboratory
Variational Methods, Mechanics; Laboratory
FE Discretization and Requirements; C0 Elements.
Isoparametric Elements, Bar Element
Isoparamteric Element, Bilinear Element
Isoparameteric Element; Quadratic, Hierarchical Elements;
Numerical integration
Isoparameteric Element; Numerical integration
Laboratory
Weighted Residuals
Galerkin; 3D Elasticity; Field Equation
Field Equation, Theory, application
Field Equation
Exam I
Lab (Field Equations)
Error Analysis
SPRING BREAK
Topics (Condensation, Transformation, Integration, Test)
Order of Integration, Eignevalu tests
Plate Bending
Plate Bending
Plate Bending
Geometric Non Linearity
Geometric Nonlinearity
Material Nonlinearity
Dynamics
Review
6.1-6.14
Ch. 16
9.1-9.6
11.1-11.5
Ch. 14.
Ch. 17
Ch. 13
Draft
08
Victor Saouma
LIST OF TABLES
Draft
Chapter 1
PREREQUISITE
In the rst course (CVEN4525/5525, Finite Element I; Framed Structures), the direct stiness
method was rst introduced (element stiness matrix, transformation matrix, global stiness
matrix assembly, internal force recovery). As an interlude we then covered the exibility method
and stiness-exibility relationship. The second part of the course began with a thorough coverage of variational method (duality between extremization of a functional and a corresponding
euler dierential equation) followed by a rigorous introduction/derivation of the various energy
methods.
1
1.1
Variational Formulations
2 A summary of the various methods introduced in Finite Element I; Framed Structures is

shown in Fig. 1.1, Fig. 1.2, and Table 1.1.
1.2
1.2.1
Finite Element Formulation

Strain Displacement Relations
The displacement at any point inside an element can be written in terms of the shape
functions N and the nodal displacements {}
3
= N{}
(1.1)
= [B]{}
(1.2)
The strain is then dened as:

where [B] is the matrix which relates joint displacements to strain eld.
1.2.1.1
Axial Members

u = (1

x
L)
N
x
L

u1

u2

{}
(1.3-a)
Draft
12
PREREQUISITE
div + b = 0
t
t = 0 t
def
U0 =

0
Natural B.C.
Essential B.C.
Du = 0
u = 0 u
def
Gauss
T d

uT bd t uT
td = 0
Wi We = 0
Principle of Stationary
Potential Energy
= 0
def
= U We

= U0 d ( ui bi d + t ui ti d)
Principle of Complementary
Stationary Potential Energy
= 0
def
= Wi +
We
= U0 d + u ui ti d
Castiglianos
First Theorem
Castiglianos
Second Theorem
Wi
k
=P
k
Wi
k
=
Pk
Rayleigh-Ritz
n
cji ji + j0
uj
cji
Virtual Work

i ti d = 0
ij
ij d u u
Wi We = 0
ij,j = 0
ti = 0 t
Gauss
ij (ui,j + uj,i ) = 0
= 0 u
ui u
U0 =
Principle of Virtual Work

1
2
i=1
=0
i = 1, 2, , n;
j = 1, 2, 3
Figure 1.1: Summary of Variational Methods
Victor Saouma
Draft
1.2 Finite Element Formulation
13
Kinematically Admissible Displacements

Displacements satisfy the kinematic equations
and the the kinematic boundary conditions
Complementary Energy
Principle of Virtual Work
Virtual Work
Potential Energy
Statically Admissible Stresses

Stresses satisfy the equilibrium conditions
and the static boundary conditions
Figure 1.2: Duality of Variational Principles
Victor Saouma
Draft
14
PREREQUISITE
U

P2
dx
0 AE
1
2
Axial

Shear
Virtual Displacement U
General
Linear
L
L
du d(u)
dx
E
Adx

dx dx
0
0
d
...

M2
dx
0 EIz

1
2
Torsion
L
0
M dx
0
T2
dx
GJ
L
0
w(x)v(x)dx
0L
dx d(x )
GJ
dx

dx dx
L
0
T dx
0
M
M
dx

EIz

T
T
dx

GJ

Virtual Force W
i Pi i
M
L i i i
w(x)v(x)dx
Virtual Displacement W
i Pi i
M
L i i i
w(x)v(x)dx
...
M dx

P
M
V xy dx
d2 v d2 (v)
EIz 2
dx
2
dx dx

T dx
0
W
1
i 2 Pi i
i 12 Mi i
...
0L
1
2
Flexure
V xy dx
Virtual Force U
General
Linear
L
L
P
dx
P
dx

AE
0
0

Table 1.1: Summary of Variational Terms Associated with One Dimensional Elements
1
1
du
L
L
=
= x =
dx
N1
N2
x
[B]
1.2.1.2

u1
u2

{}
(1.3-b)
Flexural Members
Using the shape functions for exural elements previously derived in Eq. 6.41 we have:
d2 v
y
=y 2
dx
M
=
EI
d2 v
= y 2
dx
6
2
2
6
L2 (2 1) L (3 2) L2 (2 + 1) L (3 1)
= y

(1.4-a)
(1.4-b)
2 N1
x2
2 N2
x2

[B]
Victor Saouma
2 N3
x2
2 N4
x2

(1.4-c)
v1
1
(1.4-d)
v2
2

{}
Draft
1.2 Finite Element Formulation
1.2.2
15
Virtual Displacement and Strains

= [N]{}
(1.5-a)
= [B]{}
(1.5-b)
(1.5-c)
1.2.3
Element Stiness Matrix Formulation
Let us consider the most general case, or element with:

Initial strain: (temperature eect, support settlement, or other) such that:
x =
x
E

due to load
ix

initial strain
(1.6)
thus:
x = Ex Eix
(1.7)
{} = [D]{} [D]{i }
(1.8)
or in matrix form:
where [D] is the constitutive matrix which relates stress and strain.
Load: q(x) along it.
Let us apply the principle of virtual work.
= W

{}dvol
U =
vol
{} = [D]{} [D]{i }
(1.9-b)
{} = [B]{}
(1.9-d)
{} = [B]{}
(1.9-e)
= [B]
(1.9-a)
(1.9-c)
(1.9-f)
(1.9-g)
Combining Eqns. 1.9-a, 1.9-b, 1.9-c, 1.9-f, and 1.9-e, the internal virtual strain energy is
given by:

T
[B]T [D]{i }dvol
[B] [D][B]{} dvol
U =
vol
vol
[B]T [D][B] dvol{}
[B]T [D]{i }dvol
(1.10-a)
=
vol
vol
the virtual external work in turn is given by:
W =
Victor Saouma

{F}
+ q(x)dx

(1.11)

l
Virt. Nodal Displ. Nodal Force
Draft
16
PREREQUISITE
combining this equation with:

yields:
{} = [N]{}
(1.12)
l
W = {F} + [N]T q(x) dx
(1.13)
Equating the internal strain energy Eqn. 1.10-a with the external work Eqn. 1.13, we obtain:

T
[B] [D][B] dvol{}
[B]T [D]{i }dvol =

vol
vol

[K]
init
{F
}
l
[N]T q(x) dx
(1.14-a)
{F} +
0

{Fe }
where:
The element stiness matrix:

[K] =
[B]T [D][B]dvol
(1.15)
[B]T [D]{i }dvol
(1.16)
vol
Element initial force vector:

{Fi }
=
vol
Element equivalent load vector:

{Fe }
[N] q(x) dx
(1.17)
{} = [D] [N]{}
(1.18)
1.2.3.1
Stress Recovery
Recall that we have:
1.3
1.3.1
{} = [D]{}
{} = [B]{}
Direct Stiness Method

Global Stiness Matrix
4 The physical interpretation of the global stiness matrix K is analogous to the one of the
element, i.e. If all degrees of freedom are restrained, then Kij corresponds to the force along
global degree of freedom i due to a unit positive displacement (or rotation) along global degree
of freedom j.
5
For instance, with reference to Fig. 1.3, we have three global degrees of freedom, 1 , 2 , and
Victor Saouma
Draft
1.3 Direct Stiness Method
17
1A
0
0
1
11
00
00
11
P/2
EI
w
2 1
0
A
B
w
C
111
000
000
111
1
K211
0
K22
C
00
L/2 0000
L/211
0000
1111
1111
00
11
A
1
0
0
1
1
0
11A
00
0
1
00
111 11
000
0
1
00
11
K31
1
0
0
1
1111
0000
0 K11
1
A
11
00
00
11
C
11
00
00
11
11
00
K23
0
1
K32
0
1
0
111K121
000
0
1
01
1
B
0
1
0
1
0
1
A
1
0
0
1
111
000
C
000
111
K 33
1
0
0
1
1111
0000
0 K13
1
1
C
111
000
000
111
Figure 1.3: Frame Example

3 . and the global (restrained or structures) stiness matrix is
K11 K12 K13

K = K21 K22 K23
K31 K32 K33
(1.19)
and the rst column corresponds to all the internal forces in the unrestrained d.o.f. when a
unit displacement along global d.o.f. 1 is applied.
1.3.1.1
Structural Stiness Matrix
6 The structural stiness matrix is assembled only for those active degrees of freedom which
are active (i.e unrestrained). It is the one which will be inverted (or rather decomposed) to
determine the nodal displacements.
1.3.1.2
Augmented Stiness Matrix
7 The augmented stiness matrix is expressed in terms of all the dof. However, it is partitioned into two groups with respective subscript u where the displacements are known (zero
otherwise), and t where the loads are known.

Ktt Ktu
Pt
t ?
=
Ru ?
Kut Kuu
u
(1.20)
We note that Ktt corresponds to the structural stiness matrix.

Victor Saouma
Draft
18
8
PREREQUISITE
The rst equation enables the calculation of the unknown displacements.

t = K1
tt (Pt Ktu u )
(1.21)
The second equation enables the calculation of the reactions

Ru = Kut t + Kuu u
(1.22)
10 For internal book-keeping purpose, since we are assembling the augmented stiness matrix,
we proceed in two stages:
1. First number all the global unrestrained degrees of freedom

2. Then number separately all the global restrained degrees of freedom (i.e. those with
known displacements, zero or otherwise) starting with -1 this will enable us later on to
distinguish the restrained from unrestrained dof.
The element internal forces (axial and shear forces, and moment at each end of the member)
are determined from
11
(e)
pint = k(e) (e)
(1.23)
(e)
at the element level where pint is the six by six array of internal forces, k(e) the element
stiness matrix in local coordinate systems, and (e) is the vector of nodal displacements in local
coordinate system. Note that this last array is obtained by rst identifying the displacements
in global coordinate system, and then premultiplying it by the transformation matrix to obtain
the displacements in local coordinate system.
1.3.2
1.3.2.1
Logistics
Boundary Conditions, [ID] Matrix
12 Because of the boundary condition restraints, the total structure number of active degrees of
freedom (i.e unconstrained) will be less than the number of nodes times the number of degrees
of freedom per node.
13 To obtain the global degree of freedom for a given node, we need to dene an [ID] matrix
such that:
ID has dimensions l k where l is the number of degree of freedom per node, and k is the
number of nodes).
ID matrix is initialized to zero.
1. At input stage read ID(idof,inod) of each degree of freedom for every node such that:

0
if unrestrained d.o.f.
ID(idof, inod) =
(1.24)
1
if restrained d.o.f.
Victor Saouma
Draft
19
2. After all the node boundary conditions have been read, assign incrementally equation
numbers
(a) First to all the active dof
(b) Then to the other (restrained) dof, starting with -1.
Note that the total number of dof will be equal to the number of nodes times the number
of dof/node NEQA.
3. The largest positive global degree of freedom number will be equal to NEQ (Number Of
Equations), which is the size of the square matrix which will have to be decomposed.
14
For example, for the frame shown in Fig. 1.4:
Figure 1.4: Example for [ID] Matrix Determination

1. The input data le may contain:
Node No.
1
2
3
4
[ID]T
000
110
000
100
2. At this stage, the [ID] matrix is equal to:
0 1 0 1
ID = 0 1 0 0
0 0 0 0
3. After we determined the equation numbers, we
1 1
ID = 2 2
3 4
Victor Saouma
would have:
5 3
6 8
7 9
(1.25)
(1.26)
Draft
110
1.3.2.2
PREREQUISITE
LM Vector
15 The LM vector of a given element gives the global degree of freedom of each one of the element
degree of freedoms. For the structure shown in Fig. 1.4, we would have:
LM = 1 2 4 5 6 7 element 1 (2 3)
LM = 5 6 7 1 2 3
element 2 (3 1)
LM = 1 2 3 3 8 9
element 3 (1 4)
1.3.2.3
Assembly of Global Stiness Matrix
16 As for the element stiness matrix, the global stiness matrix [K] is such that Kij is the
force in degree of freedom i caused by a unit displacement at degree of freedom j.
Whereas this relationship was derived from basic analysis at the element level, at the structure level, this term can be obtained from the contribution of the element stiness matrices
[K(e) ] (written in global coordinate system).
17
For each Kij term, we shall add the contribution of all the elements which can connect degree
of freedom i to degree of freedom j, assuming that those forces are readily available from the
individual element stiness matrices written in global coordinate system.
18
19
Kij is non-zero if degree of freedom i and degree of freedom j

1. Are connected by an element.
2. Share a node.
3. Are connected by an element and the corresponding value in the element stiness matrix
in the global coordinate system is zero.
There are usually more than one element connected to a dof. Hence, individual element
stiness matrices terms must be added up.
20
21 Because each term of all the element stiness matrices must nd its position inside the global
stiness matrix [K], it is found computationally most eective to initialize the global stiness
matrix [KS ](N EQAN EQA ) to zero, and then loop through all the elements, and then through
(e)
each entry of the respective element stiness matrix Kij .

(e)
The assignment of the element stiness matrix term Kij (note that e, i, and j are all known
since we are looping on e from 1 to the number of elements, and then looping on the rows and
S is made through
columns of the element stiness matrix i, j) into the global stiness matrix Kkl
the LM vector (note that it is k and l which must be determined).
22
Since the global stiness matrix is also symmetric, we would need to only assemble one side
of it, usually the upper one.
23
24
Contrarily to the previous method, we will assemble the full augmented stiness matrix.
Example 1-1: Assembly of the Global Stiness Matrix
Victor Saouma
Draft
111
50kN
4 kN/m
2
8m
00
11
11
00
00
11
00
11
3m
1
000
111
111
000
000
111
000
111
0001
111
7.416 m
8m
Figure 1.5: Simple Frame Analyzed with the MATLAB Code

As an example, let us consider the frame shown in Fig. 1.5.
The ID matrix is initially set to:
1 0 1
[ID] = 1 0 1
1 0 1
We then modify it to generate the global degrees
4
[ID] = 5
6
(1.27)
of freedom of each node:
1 7
2 8
3 9
(1.28)
Finally the LM vectors for the two elements (assuming that Element 1 is dened from node 1
to node 2, and element 2 from node 2 to node 3):

2
3
4 5 6 1
(1.29)
[LM ] =
1
2
3 7 8 9
Let us simplify the operation by designating the element stiness matrices in global coordinates
as follows:
K (1) =
4 A11
5
A21
6
A31
1
A41
2 A51
3 A61
5
A12
A22
A32
A42
A52
A62
6
A13
A23
A33
A43
A53
A63
1
A14
A24
A34
A44
A54
A64
2
A15
A25
A35
A45
A55
A65
A16
A26
A36
A46
A56
A66
1
1 B11
2
B21
3
B31
7
B41
8 B51
9 B61
2
B12
B22
B32
B42
B52
B62
3
B13
B23
B33
B43
B53
B63
7
B14
B24
B34
B44
B54
B64
8
B15
B25
B35
B45
B55
B65
B16
B26
B36
B46
B56
K (2) =
(1.30-a)
(1.30-b)
B66
We note that for each element we have shown the corresponding LM vector.
Victor Saouma
Draft
112
PREREQUISITE
Now, we assemble the global stiness matrix
K=
A44 + B11 A45 + B12 A46 + B13

A54 + B21 A55 + B22 A56 + B23
A64 + B31 A65 + B32 A66 + B33
A14
A15
A16
A25
A26
A24
A35
A36
A34
B42
B43
B41
B51
B52
B53
B61
B62
B63
A41
A51
A61
A11
A21
A31
0
0
0
A42
A52
A62
A12
A22
A32
0
0
0
A43
A53
A63
A13
A23
A33
0
0
0
B14
B24
B34
0
0
0
B44
B54
B64
B15
B25
B35
0
0
0
B45
B55
B65
B16
B26
B36
0
0
0
B46
B56
B66
(1.31)
We note that some terms are equal to zero because we do not have a connection between
the corresponding degrees of freedom (i.e. node 1 is not connected to node 3).
1.3.2.4
25
Algorithm
The direct stiness method can be summarized as follows:
Preliminaries: First we shall

1. Identify type of structure (beam, truss, grid or frame) and determine the
(a) Number of spatial coordinates (1D, 2D, or 3D)
(b) Number of degree of freedom per node (local and global)
(c) Number of cross-sectional and material properties
2. Determine the global unrestrained and restrained degree of freedom equation numbers for each node, Update the [ID] matrix (which included only 0s and 1s in the
input data le).
Analysis :
1. For each element, determine
(a)
(b)
(c)
(d)
(e)
Vector LM relating local to global degree of freedoms.

Element stiness matrix [k(e) ]
Angle between the local and global x axes.
Rotation matrix [(e) ]
Element stiness matrix in global coordinates [K(e) ] = [(e) ]T [k(e) ][(e) ]
2. Assemble the augmented stiness matrix [K(S) ] of unconstrained and constrained

degree of freedoms.
3. Extract [Ktt ] from [K(S) ] and invert (or decompose into into [Ktt ] = [L][L]T where
[L] is a lower triangle matrix.
4. Assemble load vector {P} in terms of nodal load and xed end actions.
5. Backsubstitute and obtain nodal displacements in global coordinate system.
6. Solve for the reactions.
Victor Saouma
Draft
113
7. For each element, transform its nodal displacement from global to local coordinates
{} = [(e) ]{}, and determine the internal forces [p] = [k]{}.
26
Some of the prescribed steps are further discussed in the next sections.
Example 1-2: Direct Stiness Analysis of a Truss
Using the direct stiness method, analyze the truss shown in Fig. 1.6.
4
5
1
6
8
50k
12
100k
16
16
Figure 1.6:
Solution:
1. Determine the structure ID matrix
Node #
1
2
3
4
5

ID =
Bound.
X
0
0
1
0
0
Cond.
Y
1
0
1
0
0
0 0 1 0 0
1 0 1 0 0
N ode 1 2 3 4

1 2 2 4
1 3 3 5
(1.32-a)
5

6
7
(1.32-b)
2. The LM vector of each element is evaluated next
Victor Saouma
LM 1 = 1 1 4 5
(1.33-a)
LM 2 = 1 1 2 3
(1.33-b)
LM 3 = 2 3 4 5
(1.33-c)
Draft
114
PREREQUISITE
LM 4 = 4 5 6 7
(1.33-d)
LM 5 = 2 3 4 5
(1.33-e)
LM 6 = 2 3 6 7
(1.33-f)
LM 7 = 2 3 2 3
(1.33-g)
LM 8 = 2 3 6 7
(1.33-h)
3. Determine the element stiness matrix of each element in the global coordinate system noting
that for a 2D truss element we have
[K (e) ] = [(e) ]T [k(e) ][(e) ]
2
c
cs c2 cs
EA
s2 cs s2
cs
=
2
c cs c2
cs
L
s2
cs s2 cs
x2 x1
L ;
where c = cos =
Element 1 L = 20 , c =
120
20
= 0.8, s =
Element 2 L = 16 , c = 1 , s = 0 ,
EA
L
1
1
18, 750
1
0
2 18, 750
3
0
Element 3 L = 12 , c = 0 , s = 1 ,
[K3 ] =
2 0
3
0
4 0
5 0
Element 4 L = 16 , c = 1 , s = 0 ,
[K4 ] =
Victor Saouma
(30,000
4
9, 600
7, 200
9, 600
7, 200
ksi)(10 in2 )
20
7, 200
5, 400
7, 200
5, 400
= 15, 000 k/ft.
(1.35)
= 18, 750 k/ft.
[K2 ] =
EA
L
= 0.6,
1
7200
5, 400
7, 200
5, 400
1
1
9, 600
1
7,
200
4 9, 600
5 7, 200
[K1 ] =
(1.34-b)
Y2 Y1
L
s = sin =
160
20
(1.34-a)
EA
L
1
2
0 18, 750
0
0
0
18, 750
0
0
(1.36)
= 25, 000 k/ft.
3
0
25, 000
0
25, 000
EA
L
0
0
0
0
4
5
0
0
0 25, 000
0
0
0 25, 000
(1.37)
= 18, 750 k/ft.
4 18, 750
5
0
6 18, 750
7
0
5
6
0 18, 750
0
0
0 18, 750
0
0
0
0
0
0
(1.38)
Draft

Element 5 L = 20 , c =
115
160
20
[K5 ] =
= 0.8 , s = 0.6 ,
2 9, 600
3
7, 200
4 9, 600
5
7, 200
Element 6 L = 20 , c = 0.8 , s = 0.6 ,
[K6 ] =
2 9, 600
3
7, 200
6 9, 600
7 7, 200
Element 7 L = 16 , c = 1 , s = 0 ,
[K7 ] =
EA
L
2
3
6
7
EA
L
= 15, 000 k/ft.
4
9, 600
7, 200
9, 600
7, 200
7, 200
5, 400
7, 200
5, 400
(1.39)
= 15, 000 k/ft.
3
7, 200
5, 400
7, 200
5, 400
6
9, 600
7, 200
9, 600
7, 200
7, 200
5, 400
7, 200
5, 400
(1.40)
= 18, 750 k/ft.
2
18, 750
3
0
2 18, 750
3
0
Element 8 L = 12 , c = 0 , s = 1 ,
[K8 ] =
EA
L
3
7, 200
5, 400
7, 200
5, 400
EA
L
3
2
0 18, 750
0
0
0 18, 750
0
0
0
0
0
0
(1.41)
= 25, 000 k/ft.
2
3
0
0
0
25, 000
0
0
0 25, 000
6
7
0
0
0 25, 000
0
0
0 25, 000
(1.42)
4. Assemble the global stiness matrix in k/ft Note that we are not assembling the augmented
stiness matrix, but rather its submatrix [Ktt ].
8
>
>
>
>
>
>
>
<
>
>
>
>
>
>
>
:
0
0
100k
0
0
50k
0
9 2
>
>
>
6
>
>
>
>
6
= 6
6
=6
6
>
>
6
>
>
>
4
>
>
;
9, 600 + 18, 750
18, 750
9, 600 + (2) 18, 750
0
7, 200
5, 400 + 25, 000
9, 600
0
0
18, 750 + (2)9, 600
SYMMETRIC
7, 200
0
25, 000
7, 200 7, 200
25, 000 + 5, 400(2)
0
9, 600
7, 200
18, 750
0
18, 750 + 9, 600
0
7, 200
5, 400
0
0
7, 200
25, 000 + 5, 400
(1.43)
5. Convert to k/in and simplify

2, 362.5
1, 562.5
0
800
600
0
0
0
3, 925.0
600
0
0
800
600
0
2,
533.33
0
2,
083.33
600
450
100

=
3, 162.5
0
1, 562.5
0
0
SYMMETRIC
2,
983.33
0
0
2, 362.5
600
50
2, 533.33
0

Pt
Ktt
(1.44)
Victor Saouma

38
>
>
>
7
>
>
7
>
>
7
<
7
7
7
>
>
7
>
>
5>
>
>
:
U1
U2
V3
U4
V5
U6
V7

ut
u1
u2
v3
u4
v5
u6
v7
9
>
>
>
>
>
>
>
=
>
>
>
>
>
>
>
;
Draft
116
PREREQUISITE
6. Invert stiness matrix and solve for displacements

0.0223 in
U1
U
0.00433 in
V3 0.116 in
=
U4
0.0102 in
V
0.0856 in
0.00919 in

U6
V7
0.0174 in
(1.45)
7. Solve for member internal forces (in this case axial forces) in local coordinate systems

V1
c
s c s
u1
=
(1.46)
u2
U
c s
c
s
V2
Element 1
p1
p2
1
0.0223

1 ft
0.8
0.6 0.8 0.6
0
= (15, 000 k/ft)(
)
(1.47-a)
0.8 0.6
0.8
0.6
0.0102
12 in
0.0856

52.1 k
Compression
(1.47-b)
=
52.1 k
Element 2
p1
p2
2
= 18, 750 k/ft(

=
43.2 k
43.2 k
1 ft
)
12 in
0.0233
1 0 1 0
0
1 0
1 0
0.00433
0.116
Tension
(1.48-a)
(1.48-b)
Element 3
p1
p2
3
= 25, 000 k/ft(

=
Victor Saouma
63.3 k
63.3 k
1 ft
)
12 in
0.00433
0
1 0 1
0.116
0 1 0
1
0.0102
0.0856
Tension
(1.49-a)
(1.49-b)
Draft
117
Element 4

0.0102

1 ft
1 0 1 0
0.0856
= 18, 750 k/ft(
)
(1.50-a)
1 0
1 0
0.00919
12 in
0.0174

1.58 k
Tension
(1.50-b)
=
1.58 k
4
p1
p2
Element 5

p1
p2
0.0102

1 ft
0.8
0.6
0.8 0.6
0.0856
= 15, 000 k/ft(
)
(1.51-a)
0.8 0.6 0.8
0.6
0
12 in
0

54.0 k
Compression
(1.51-b)
=
54.0 k
5
Element 6

p1
p2
6
= 15, 000 k/ft(

=
60.43 k
60.43 k
1 ft
)
12 in
0.00433
0.8
0.6 0.8 0.6
0.116
(1.52-a)
0.8 0.6
0.8
0.6
0.00919
0.0174
Tension
(1.52-b)
Element 7

p1
p2
7
= 18, 750 k/ft(

=
6.72 k
6.72 k
1 ft
)
12 in
0.00433
1 0 1 0
0.116
1 0
1 0
0
Compression
(1.53-a)
(1.53-b)
Element 8

p1
p2
8
= 25, 000 k/ft(

=
Victor Saouma
36.3 k
36.3 k
1 ft
)
12 in
0
1 0 1
0 1 0
1

Compression
0
0
0.00919
0.0174
(1.54-a)
(1.54-b)
Draft
118
PREREQUISITE
8. Determine the structures MAXA vector
1 3
9 14
2
5
8
13
19
25
4
7
12
18
24
MAXA =
[K] =
6
11
17
23
10 16 22
15 21
20
1
2
4
6
10
15
20
(1.55)
Thus, 25 terms would have to be stored.
Example 1-3: Analysis of a Frame with MATLAB

The simple frame shown in Fig. 1.7 is to be analyzed by the direct stiness method. Assume:
E = 200, 000 MPa, A = 6, 000 mm2 , and I = 200 106 mm4 . The complete MATLAB solution
is shown below along with the results.
50kN
4 kN/m
2
8m
0
1
1
0
0
1
0
1
0
1
3m
1
111
000
000
111
000
111
0001
111
000
111
7.416 m
8m
Figure 1.7: Simple Frame Analyzed with the MATLAB Code
% zero the matrices

k=zeros(6,6,2); K=zeros(6,6,2); Gamma=zeros(6,6,2);
% Structural properties units: mm^2, mm^4, and MPa(10^6 N/m)
A=6000;II=200*10^6;EE=200000;
% Convert units to meter and kN
A=A/10^6;II=II/10^12;EE=EE*1000;
% Element 1
i=[0,0];j=[7.416,3]; [k(:,:,1),K(:,:,1),Gamma(:,:,1)]=stiff(EE,II,A,i,j);
% Element 2
i=j;j=[15.416,3]; [k(:,:,2),K(:,:,2),Gamma(:,:,2)]=stiff(EE,II,A,i,j);
% Define ID matrix
ID=[
-4 1 -7;
-5 2 -8;
-6 3 -9];
% Determine the LM matrix
LM=[
Victor Saouma
Draft
119
-4 -5 -6 1 2 3;
1 2 3 -7 -8 -9];
% Assemble augmented stiffness matrix
Kaug=zeros(9); for elem=1:2
for r=1:6
lr=abs(LM(elem,r));
for c=1:6
lc=abs(LM(elem,c));
Kaug(lr,lc)=Kaug(lr,lc)+K(r,c,elem);
end
end
end
% Extract the structures Stiffness Matrix
Ktt=Kaug(1:3,1:3);
% Determine the fixed end actions in local coordinate system
fea(1:6,1)=0; fea(1:6,2)=[0,8*4/2,4*8^2/12,0,8*4/2,-4*8^2/12];
% Determine the fixed end actions in global coordinate system
FEA(1:6,1)=Gamma(:,:,1)*fea(1:6,1); FEA(1:6,2)=Gamma(:,:,2)*fea(1:6,2);
% FEA_Rest for all the restrained nodes
FEA_Rest=[0,0,0,FEA(4:6,2)];
% Assemble the load vector for the unrestrained node
P(1)=50*3/8;P(2)=-50*7.416/8-FEA(2,2);P(3)=-FEA(3,2);
% Solve for the Displacements in meters and radians
Displacements=inv(Ktt)*P
% Extract Kut
Kut=Kaug(4:9,1:3);
% Compute the Reactions and do not forget to add fixed end actions
Reactions=Kut*Displacements+FEA_Rest
% Solve for the internal forces and do not forget to include the fixed end actions
dis_global(:,:,1)=[0,0,0,Displacements(1:3)];
dis_global(:,:,2)=[Displacements(1:3),0,0,0]; for elem=1:2
dis_local=Gamma(:,:,elem)*dis_global(:,:,elem);
int_forces=k(:,:,elem)*dis_local+fea(1:6,elem)
end
function [k,K,Gamma]=stiff(EE,II,A,i,j)
% Determine the length
L=sqrt((j(2)-i(2))^2+(j(1)-i(1))^2);
% Compute the angle theta (careful with vertical members!)
if(j(1)-i(1))~=0
alpha=atan((j(2)-i(2))/(j(1)-i(1)));
else
alpha=-pi/2;
end
% form rotation matrix Gamma
Gamma=[ cos(alpha) sin(alpha) 0 0
0
0; -sin(alpha)
cos(alpha) 0 0
0
0; 0
0
1 0
0; 0
0
0 cos(alpha) sin(alpha) 0; 0
0
Victor Saouma
Draft
120
PREREQUISITE
0 -sin(alpha) cos(alpha) 0; 0
0
0 0
1];
% form element stiffness matrix in local coordinate system
EI=EE*II; EA=EE*A; k=[EA/L,
0,
0, -EA/L,
0,
12*EI/L^3, 6*EI/L^2,
0, -12*EI/L^3, 6*EI/L^2;
0,
6*EI/L^2,
4*EI/L,
0, -6*EI/L^2,
2*EI/L;
-EA/L,
0,
0, EA/L,
0,
0;
0, -12*EI/L^3, -6*EI/L^2,
0, 12*EI/L^3, -6*EI/L^2;
0,
6*EI/L^2,
2*EI/L,
0, -6*EI/L^2,
4*EI/L];
% Element stiffness matrix in global coordinate system
K=Gamma*k*Gamma;
0,
0;
This simple proigram will produce the following results:

Displacements =
0.0010
-0.0050
-0.0005
Reactions =
130.4973
55.6766
13.3742
-149.2473
22.6734
-45.3557
int_forces =
int_forces =
141.8530
2.6758
13.3742
-141.8530
-2.6758
8.0315
149.2473
9.3266
-8.0315
-149.2473
22.6734
-45.3557
We note that the internal forces are consistent with the reactions (specially for the second node
of element 2), and amongst themselves, i.e. the moment at node 2 is the same for both elements
(8.0315).
Example 1-4: Analysis of a simple Beam with Initial Displacements
Victor Saouma
Draft
121
The full stiness matrix of a beam element is given by

v1
V1
12EI/L3
M1 6EI/L2
V2 12EI/L3
M2 6EI/L2
1
6EI/L2
4EI/L
6EI/L2
2EI/L
[ke ] =
v2
12EI/L3
6EI/L2
12EI/L3
6EI/L2
6EI/L2
2EI/L
6EI/L2
4EI/L
(1.56)
This matrix is singular, it has a rank 2 and order 4 (as it embodies also 2 rigid body motions).
27
We shall consider 3 dierent cases, Fig. 1.8

-3
1
0
0
1
0
1
0
1
-2
-4
1
1
0
0
1
0
1
-3
1
0
0
1
0
1
1
0
0
1
0
1
-4
-3
1
0
0
1
0
1
0
1
-4
1
0
0
1
0
1
Figure 1.8: Stiness Analysis of one Element Structure

Cantilivered Beam/Point Load
1. The element stiness matrix is
4
6EI/L2
4EI/L
6EI/L2
2EI/L
1
12EI/L3
6EI/L2
12EI/L3
6EI/L2
2
3
6EI/L2
2EI/L 7
7
6EI/L2 5
4EI/L
2. The structure stiness matrix is assembled

1
2
2
1
12EI/L2 6EI/L2
2
2 6
4EI/L
6 6EI/L
K=
34 12EI/L3 6EI/L2
4 6EI/L2
2EI/L
3
12EI/L3
6EI/L2
12EI/L3
6EI/L2
4
3
6EI/L2
2EI/L 7
7
6EI/L2 5
4EI/L
k=
3
2
3 12EI/L3
2
46
6 6EI/L
1 4 12EI/L3
2
6EI/L2
3. The global matrix can be rewritten as

8
9
P
>
>
>
< 0 >
=
12EI/L2
6 6EI/L2
=6
4 12EI/L3
R3 ? >
>
>
>
:
;
6EI/L2
R4 ?
6EI/L2
4EI/L
6EI/L2
2EI/L
12EI/L3
6EI/L2
12EI/L3
6EI/L2
38
6EI/L2 >
1 ?
>
< ?
2EI/L 7
2
7
6EI/L2 5 >
>
: 3
4EI/L
4
9
>
>
=
>
>
;
4. Ktt is inverted (or actually decomposed) and stored in the same global matrix
2
L3 /3EI
6
6
6 L2 /2EI
6
4 12EI/L3
6EI/L2
Victor Saouma
L2 /2EI
12EI/L3
6EI/L2
L/EI
6EI/L2
2EI/L
6EI/L2
2EI/L
12EI/L3
6EI/L2
6EI/L2
4EI/L
3
7
7
7
7
5
Draft
122
PREREQUISITE
5. Next we compute the equivalent load, P t = Pt Ktu u , and overwrite Pt by P t

Pt Ktu u
8
9
>
P >
>
>
>
>
<
=
L3 /3EI
6
6
0
L2 /2EI
6
6
>
>
3
0 >
>
>
>
:
; 4 12EI/L2
0
6EI/L
8
9
>
>
P
>
>
>
>
<
=
0
0
0
>
>
>
:
L2 /2EI
12EI/L3
6EI/L2
L/EI
6EI/L2
2EI/L
6EI/L
2EI/L
6EI/L2
4EI/L
12EI/L
6EI/L2
9
8
P >
>
7>
>
<
7
0 =
7
7> 0 >
>
5>
;
:
0
>
>
>
;

6. Now we solve for the displacement t = K1
tt Pt , and overwrite Pt by t
8
>
1
>
>
<
2
0
0
>
>
>
:
9
>
>
>
=
6
6
6 L2 /2EI
6
4 12EI/L3
>
>
>
;
L3 /3EI
2
8 6EI/L
>
P L3 /3EI
>
>
>
<
2
>
>
>
>
:
P L /2EI
0
0
L2 /2EI
12EI/L3
L/EI
6EI/L2
2EI/L
12EI/L3
6EI/L2
6EI/L2
4EI/L
6EI/L
2EI/L
9
6EI/L2
38
9
>
> P >
>
7>
>
7< 0 =
7
7> 0 >
>
5>
>
: 0 >
;
>
>
>
>
=
>
>
>
>
;
7. Finally, we solve for the reactions, Ru = Kut tt + Kuu u , and overwrite u by Ru

8
P L3 /3EI
>
>
>
< P L2 /2EI
R3
>
>
>
:
R
4
9
>
>
>
=
>
>
>
;
L3 /3EI
L2 /2EI
6
6
6
6 12EI/L3
4
6EI/L
8
P L3 /3EI
>
>
>
2
<
L2 /2EI
L/EI
12EI/L3
6EI/L2
6EI/L2
12EI/L3
2
2EI/L
6EI/L
1
6EI/L2
4EI/L
6EI/L2
2EI/L
4
12EI/L3
6EI/L2
12EI/L3
6EI/L2
9
>
>
>
=
38
6EI/L2 >
>
P L3 /3EI
>
>
>
2EI/L 7
<
7
P L2 /2EI
7
6EI/L2 7 >
5>
>
0
>
>
:
4EI/L
0
P L /2EI
P
>
>
>
>
>
>
:
;
PL
Simply Supported Beam/End Moment

3
2
3 12EI/L3
2
1 6
6 6EI/L
44 12EI/L3
2
6EI/L2
k=
2
3
6EI/L2
2EI/L 7
7
6EI/L2 5
4EI/L

1
2
1
4EI/L
2 6
6 2EI/L
34 6EI/L2
4 6EI/L2
K=
2
2EI/L
4EI/L
6EI/L2
6EI/L2
3
6EI/L2
6EI/L2
12EI/L3
12EI/L3
4
3
6EI/L2
2 7
6EI/L 7
12EI/L3 5
12EI/L3
3. The global stiness matrix can be rewritten as

8 9
0
>
>
>
>
>
>
>
>
< M =
>
>
>
>
:
Victor Saouma
4EI/L
6 2EI/L
6
=
> 4 6EI/L2
R3 ? >
>
>
;
6EI/L2
R4 ?
2EI/L
4EI/L
6EI/L2
6EI/L2
6EI/L2
6EI/L2
12EI/L3
12EI/L3
38
6EI/L2
1 ?
>
>
< ?
6EI/L2 7
2
7
12EI/L3 5 >
3
>
:
3
12EI/L
4
9
>
>
=
>
>
;
9
>
>
>
>
>
=
>
>
>
>
>
;
Draft

4. Ktt is inverted
123
6
6 L/6EI
6
6
4 6EI/L2
6EI/L
L/6EI
6EI/L2
6EI/L2
L/3EI
6EI/L2
6EI/L2
6EI/L2
12EI/L3
12EI/L3
6EI/L2 7
7
12EI/L3 5
12EI/L3
L3 /3EI
7
7
5. We compute the equivalent load, P t = Pt Ktu u , and overwrite Pt by P t

Pt Ktu u
8
>
>
<
0
M
>
>
: 0
0
8
>
>
<
0
M
>
>
: 0
0
9
>
>
=
6
6
L/6EI
6
6
>
>
4
;
6EI/L2
6EI/L2
9
>
>
=
L/6EI
6EI/L2
6EI/L2
L/3EI
6EI/L2
6EI/L2
6EI/L2
6EI/L2
12EI/L3
12EI/L3
12EI/L3
12EI/L3
L3 /3EI
9
8
0 >
>
7>
>
<
7
M =
7
7> 0 >
>
5>
;
:
0
>
>
;

6. Solve for the displacements, t = K1
8
>
1
>
>
<
>
>
>
:
9
>
>
>
=
2
>
0 >
>
;
0
L3 /3EI
6
6 L/6EI
6
6
4 6EI/L2
8 6EI/L
>
M L/6EI
>
>
>
<
2
>
>
>
>
:
L/6EI
6EI/L2
L/3EI
6EI/L2
6EI/L
2
6EI/L
9
M L/3EI
0
0
>
>
>
>
=
12EI/L3
12EI/L3
8
9
0 >
7>
>
>
<
=
7
M
6EI/L2 7
7
> 0 >
>
12EI/L3 5 >
:
;
6EI/L2
12EI/L3
>
>
>
>
;
7. Solve for the reactions, Rt = Kut tt + Kuu u , and overwrite u by Ru

8
M L/6EI
>
>
>
< M L/3EI
>
>
>
:
R1
R2
9
>
>
>
=
>
>
>
;
6
6
6
6
4
L3 /3EI
L/6EI
L/6EI
L/3EI
6EI/L2
6EI/L2
6EI/L2
6EI/L2
6EI/L2
6EI/L2
12EI/L3
12EI/L3
6EI/L2
6EI/L2
12EI/L3
12EI/L3
8
M L/6EI
>
>
>
>
< M L/3EI
9
>
>
>
>
=
>
>
>
>
:
>
>
>
>
;
M/L
M/L
38
>
> M L/6EI
7>
>
7 < M L/3EI
7
7>
5>
0
>
>
:
0
Cantilivered Beam/Initial Displacement and Concentrated Moment

k=
2
2
2 12EI/L3
2
36
6 6EI/L 3
4
4 12EI/L
1
6EI/L2
3
6EI/L2
4EI/L
6EI/L2
2EI/L
4
12EI/L3
6EI/L2
12EI/L3
6EI/L2
1
3
6EI/L2
2EI/L 7
7
6EI/L2 5
4EI/L
3
2EI/L
6EI/L2
4EI/L
6EI/L2
4
3
6EI/L2
12EI/L3 7
7
6EI/L2 5
3
12EI/L
K=
Victor Saouma
1
2
1
4EI/L
2
26
6 6EI/L
4
3 2EI/L
4 6EI/L2
2
6EI/L2
12EI/L3
6EI/L2
12EI/L3
9
>
>
>
>
=
>
>
>
>
;
Draft
124
PREREQUISITE
3. The global matrix can be rewritten as

8
9
M
>
>
>
>
<
=
6EI/L2
12EI/L3
6EI/L2
12EI/L3
4EI/L
6 6EI/L2
R2 ?
=6
R ? > 4 2EI/L
>
>
: 3 >
;
6EI/L2
R4 ?
38
6EI/L2
1 ?
>
>
<
12EI/L3 7
2
7
2 5
6EI/L
>
>
: 3
4
12EI/L3
2EI/L
6EI/L2
4EI/L
6EI/L2
9
>
>
=
>
>
;
4. Ktt is inverted (or actually decomposed) and stored in the same global matrix
2
6EI/L2
L/4EI
6EI/L2
2EI/L
6EI/L2
12EI/L3
6EI/L2
12EI/L3
6EI/L2
4EI/L
6EI/L2
6
6 6EI/L2
6
4 2EI/L
3
7
12EI/L3 7
7
6EI/L2 5
3
12EI/L
5. Next we compute the equivalent load, P t = Pt Ktu u , and overwrite Pt by P t

2
8
9
M >
>
>
>
<
=
Pt Ktu u
2
6
6 6EI/L
>
4 2EI/L
>
;
2
6EI/L
9
0
0
0
>
>
:
8
>
M + 6EI0 /L2 >
>
>
>
>
<
=
0
0
0
>
>
>
:
6EI/L2
2EI/L
6EI/L2
12EI/L3
6EI/L2
12EI/L3
6EI/L2
4EI/L
6EI/L2
12EI/L3
6EI/L2
12EI/L3
L/4EI
38
>
>
7>
<
7
7
5>
>
>
:
0
>
>
>
: 00 >
;
L/4EI
6EI/L2
2EI/L
6EI/L2
12EI/L3
6EI/L2
12EI/L3
6EI/L2
4EI/L
6EI/L2
6
6 6EI/L2
6
4 2EI/L
9
>
>
>
=
>
>
>
;
>
>
>
;

6. Now we solve for the displacements, t = K1
8
9
1 >
>
>
>
<
=
1
0
0
8
>
M L/4EI + 30 /2L
>
>
<
9
>
>
>
=
>
>
>
:
>
>
>
;
0
0
0
38
>
> M + 6EI0 /L2
7>
<
3 7
12EI/L 7
0
6EI/L2 5 >
>
0
>
:
3
0
6EI/L2
12EI/L
9
>
>
>
=
>
>
>
;
7. Finally, we solve for the reactions, Rt = Kut tt + Kuu u , and overwrite u by Ru

8
M L/4EI + 30 /2L
>
>
>
<
9
>
>
>
=
>
>
>
:
>
>
>
;
R2
R3
R4
Victor Saouma
6
6
6
6
6
4
L/4EI
6EI/L2
2EI/L
6EI/L2
6EI/L2
12EI/L3
6EI/L2
12EI/L3 7
2EI/L
6EI/L2
4EI/L
6EI/L2
6EI/L2
12EI/L3
6EI/L2
12EI/L3
8
>
M L/4EI + 30 /2L
>
>
>
<
9
>
>
>
>
=
>
>
>
>
:
8
>
>
>
>
>
<
>
>
>
>
;
0
0
M L/4EI + 30 /2L
3M/2L 3EI0 /L3
>
M/2 3EI0 /L2
>
>
>
>
: 3M/2L + 3EI0 /L3
7
7
7
7
5
9
>
>
>
>
>
=
>
>
>
>
>
;
Draft
Chapter 2
INTRODUCTION
2.1
Introduction
Whereas the rst course focused exclusively on one dimensional rod elements, this course
will greatly expand our horizons by considering introducing a methodology to solve partial
dierential equations, with special emphasis on solid mechanics.
1
The eld of mechanics, can itself be subdivided into four major disciplines:
Theoretical which deals with the fundamental laws and principles of mechanics. A Continuum
Mechanics course is a must.
Applied mechanics seeks to apply the theoretical knowledge to engineering applications. Elasticity or Fracture Mechanics solutions are such an example of applied mechanics.
Computational mechanics combines mathematical models with numerical methods to solve
problems on a digital computer.
Experimental mechanics is conducted exclusively in a laboratory through physical measurements.
3 Any problem characterized by a PDE can be analyzed by the nite element method. The
process of nite element analysis is illustrated by Fig. 2.12.
2.2
Elliptic, Parabolic and Hyperbolic Equations
4 Since the nite element method is a numerical scheme to solve (partial) dierential equations,
let us closely examine some of the major PDE which can be solved.
The general form of a partial dierential equation is (note that we adopt the tensor notation
where u,x = du ):
dx
5
F (x, y, z, , u, u,x , u,y , u,z , , u,xx , u,yy , , u,xy , u,xz , ) = 0

and the order of the PDE is dened by the order of the highest partial derivatives appearing
in the equation. For instance
1 u,xx + 2 u,xy + 3 u,yy + 4 = 0
Draft
22
INTRODUCTION
Physical Problem
Change
Physical Problem
Mathematical Model
FEM Solution of Mathematical Model
Governed by Differential Equations

Assumptions on
Geometry
Kinematics
Material
Loading
Boundary Conditions
Etc.
Improve
Mathematical Model
Finite Element Solution

Choice of:
Finite Elements
Mesh Density
Solution Paramters
Representation of:
Loading
Boundary Conditions
Etc.
Refine mesh,
solution parameters
Assessment of accracy of FEM solution

of mathematical model
Interpretation of Results
Refine Analysis
Design Improvements
Structural Optimization
Figure 2.1: Finite Element Process, (Bathe 1996)
Victor Saouma
Draft
2.2 Elliptic, Parabolic and Hyperbolic Equations
23
is quadratic. It would be linear if i is a function of (x, y) only, otherwise it is nonlinear.

Many signicant physical systems can be described by second order linear partial dierential
equations. The most general form is

2u
2u
u u
2u
+ C(x, y) 2 = x, y, u,
,
(2.1)
A(x, y) 2 + 2B(x, y)
x
xy
y
x y
Where u is the unknown state variable.

7
This equation is classied into three types:

G(x, y) Poisson Equation

0
Laplace Equation
2
u,xx = u,t
Heat Equation
B AC = 0 Parabolic
B 2 AC > 0 Hyperbolic u,xx u,tt = 0 Wave Equation
B 2 AC < 0 Elliptic
u,xx + u,yy =
Note:
1. The Laplace equation (2 u = 0) is a special case of Poissons equation, where the right
hand side is zero. Laplace associated with the equilibrium problem.
2. The Heat equation (Hu,t K 2 2 u = 0) corresponds to exponential decay. Also
referred to as Diusion equation (uid ow through porous media, irrotational uid
ow, Saint Venant torsion of elastic bars...).
3. The Wave equation (u,tt K 2 2 u = 0) corresponds to harmonic motion
9 This classication is established when solving Eq. 2.1 using the method of characteristics
because it is then observed that the character of the solutions is distinctly dierent for the
three categories of equations.
Example 2-1: Seepage Problem;(Bathe 1996)

The idealized dam shown in Fig. 2.2 stands on permeable soil. Formulate the dierential
h
L
q(y+dy)
h2
dy
h1
111
000
000
111
000
111
000
111
Impermeable wall
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
q(x)
q(x+dx)
Seepage
q(y)
y Permeable soil
dx
0000000000000000000000000
1111111111111111111111111
1111111111111111111111111
0000000000000000000000000
Impermeable rock
Figure 2.2: Seepage Problem

Victor Saouma
Draft
24
INTRODUCTION
equation governing the steady-state seepage of water through the soil and give the corresponding
boundary conditions.
Solution:
1. For a typical element of widths dx and dy (and unit thickness), the total ow into the
element must be equal to the total ow out of the element. Hence we have

qy
qx
dx + qx qx +
dy = 0
(2.2-a)
qy qy +
y
x
qy
qx
dydx
dxdy = 0
(2.2-b)
y
x
2. Using Darcys law, the ow is given in terms of the total potential (water elevation),
qx = k
qy = k
(2.3)
where we assume a uniform permeability k. Substituting from Eq. 2.3 into Eq. 2.2-b, we
obtain the Laplace equation

k
2 2
+ 2
x2
y

=0
(2.4)
3. It may be noted that this same equation is also obtained in heat transfer analysis and in
the solution of electrostatic potential and other eld problems.
4. The boundary conditions are no-ow boundary conditions in the soil at x = and
x = +,

= 0;
=0
(2.5)
x x=
x x=+
at the rock-soil interface,

=0
y y=0
(2.6)

=0
y h x+ h ,y=L
(2.7)
and at the dam-soil interface,
10
In addition, the total potential is prescribed at the water-soil interface,

(x, L)|x<(h/2) = h1 and (x, L)|x>(h/2) = h2
(2.8)
The dierential equation in Eq. 2.4 and the boundary conditions in Eq. 2.5 to Eq. 2.8
dene the seepage ow steady-state response.
Victor Saouma
Draft
111
000
000
111
000
111
1.0
dz=1.0
z
q (t)
t>0
111
000
000
111
000
111
1
0
0
1
0
1
0
1
0q(x)
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
11111111
00000000
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
q(x+dx)
dy
=1
.0
y
111
000
000
111
000
000111
111
000
111
000
111
25
dx
Figure 2.3: One Dimensional Heat Conduction
Example 2-2: Diusion Problem; (Bathe 1996)

The very long slab shown in Fig. 2.3 is at a constant initial temperature i when the surface
at x = 0 is suddenly subjected to a constant uniform heat ow input. The surface at x = L of
the slab is kept at the temperature i , and the surfaces parallel to the x, z plane are insulated.
Assuming one-dimensional heat ow conditions, show that the problem-governing dierential
equation is the heat conduction equation
k
2
= c
2
x
t
(2.9)
where is the mass density, c is the heat capacity per unit mass (amount of heat needed to raise
a unit mass by one degree), k is the conductivity, and the temperature is the state variable.
State also the boundary and initial conditions.
Solution:
1. We consider a typical dierential element of the slab. The element equilibrium requirement is that the net heat ow input to the element must equal the rate of heat stored in
the element. Thus

q
dx A = Ac
(2.10)
qA q +
x
t
2. The constitutive relation is given by Fouriers law of heat conduction
q = k
(2.11)
3. Substituting from Eq. 2.11 into Eq. 2.10 we obtain

k
Victor Saouma
= c
2
x
t
(2.12)
Draft
26
INTRODUCTION
1
0
0
1
x
u(x,t)
0
1
0 11
1
00 00
11 11
00 00
11 00
11
dx
F(t)
F0
F0
A + d d x x+dx
dx
Figure 2.4: Rod subjected to Step Load

In this case the element interconnectivity requirements are contained in the assumption
that the temperature be a continuous function of x and no additional compatibility
conditions are applicable.
4. The boundary conditions are (for t > 0)

x
(0,t)
q0 (t)
k
(2.13-a)
|(L,t) = i
(2.13-b)
|(x,0) = i
(2.14)
and the initial condition is
Example 2-3: Wave Equation, (Bathe 1996)

The rod shown in Fig. 2.4 is initially at rest when a load F (t) is suddenly applied at its
free end. Show that the problem-governing dierential equation is the wave equation

1 2u
E
2u
(2.15)
= 2 2 ; and c =
2
x
c t
where E is the Youngs modulus, the mass density, and A the cross sectional area, c corresponds to the velocity of sound in the elastic medium, and u is the state variable. Also state
the boundary and initial conditions.
Solution:
1. The element force equilibrium requirements of a typical dierential element are, using
dAlemberts principle 1 which states that with inertia forces included, a system is in
1
Thinking in terms of equilibrium of forces, it is more appealing to invoke DAlemberts principle of dynamic equilibrium rather than Newtons second law of motion. This principle is based on the notion of a
ctitious inertia force, equal to the product of mass times acceleration and acting in a direction opposite to
the acceleration.
Victor Saouma
Draft
27
equilibrium at each time instant.

A + A
2u
dx A = A 2 dx
x
t
2. The constitutive relation is

=E
u
x
(2.16)
(2.17)
3. Combining Eq. 2.16 and Eq. 2.17 we obtain

1 2u
2u
=
x2
c2 t2
(2.18)
4. The element interconnectivity requirements are satised because we assume that the displacement u is continuous, and no additional compatibility conditions are applicable.
5. The boundary conditions are for t > 0
u|(0,t) = 0

u
= F0
EA
x
(2.19-a)
(2.19-b)
(L,t)
(2.19-c)
and the initial conditions are
u(x,0) = 0

u
= 0
t
(2.20-a)
(2.20-b)
(x,0)
With the above two sets of conditions, the formulation of the problem is complete, and
Eq. 2.18 can be solved for the displacement response of the rod.
10
Observations
Example 2-1 From the rst example we observe

1. The unknown state variables (or their normal derivatives) are given on the boundary.
These problems are for this reason also called boundary value problems, where
we should note that the solution at a general interior point depends on the data at
every point of the boundary.
2. A change in only one boundary value aects the complete solution; for instance, the
complete solution for depends on the actual value of h1 .
3. Elliptic dierential equations generally govern the steady-state response of systems.
Example 2-2, 2-3
Victor Saouma
Draft
28
INTRODUCTION
1. Comparing the governing dierential equations given in the three examples we note
that in contrast to the elliptic equation, the parabolic and hyperbolic equations
Eq. 2.12 and Eq. 2.18 include time as an independent variable and thus dene
propagation problems. These problems are also called initial value problems
because the solution depends on the initial conditions.
2. We may note that analogous to the derivation of the dynamic equilibrium equations
of lumped-parameter models, the governing dierential equations of propagation
problems are obtained from the steady-state equations by including the resistance
to change (inertia) of the dierential elements.
3. The parabolic and hyperbolic dierential equations Eq. 2.12 and Eq. 2.18 would
become elliptic equations if the time-dependent terms were neglected. In this way
the initial value problems would be converted to boundary value problems with
steady-state solutions.
4. The solution of a boundary value problem depends on the data at all points of the
boundary. However, in propagation problem, the solution at an interior point may
depend only on the boundary conditions of part of the boundary and the initial
conditions over part of the interior domain.
2.3
Solution of Discrete-System Mathematical models
Section adapted from (Bathe 1996)

11
The essence of the solution of discrete-system is

1. System idealization: the actual system is idealized as an assemblage of elements.
2. Element equilibrium: the equilibrium requirements of each element are established
in terms of state variables (displacement, temperature, pressure, etc...).
3. Element assemblage: element interconnection requirements are invoked to establish a
set of simultaneous equations in terms of the unknown state variables.
4. Calculation of state variables: The set of linear equations is solved to determine the
state variables at each discretized points.
5. Calculation of ux variable: or derived variables.
12 In the following sections, we shall illustrate, through a number of dierent physical problems,
the solution of discrete-systems. This preliminary exposure is a snap-shot of the type of
problems which can be addressed by the nite element method.
2.3.1
Steady State Problems
13 In this rst class of problem, we shall focus on equilibrium problems, that is problems
where the solution does not change with time.
Victor Saouma
Draft
2.3 Solution of Discrete-System Mathematical models

2.3.1.1
29
Elastic Spring
14 We seek to determine the displacements of each of the rigid carts connected by linear elastic
springs, as well as the force in each spring, Fig. 2.5. Conceptually, this can be viewed as a one
dimensional truss, with the spring stiness k corresponding to AE/L
u 1, R
11
00
00
11
11
00
k1
1111111
0000000
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
1
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
k4
u 2, R
u 3, R
0000000
k1111111
3
0000000
1111111
k2
1111111
0000000
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
2
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
k5
1111111
0000000
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
3
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
Figure 2.5: System of Rigid Carts Interconnected by Linear Springs, (Bathe 1996)
1. Considering the rst two springs, we can write the following equations of equilibrium
(1)
k1 u1 = F1
(2.21-a)
(2)
F1
(2)
F2
(2.21-b)
k2 (u1 u2 ) =
k2 (u2 u1 ) =
2. By extension, we can write the equation of equilibrium for each spring in terms of the
displacement state variables

(2)
1 1
u1
F1
(2.22-a)
=
k
2
(2)
u2
1 1
F2

(4)
1 1
u1
F1
(2.22-b)
= k4
(4)
u3
1 1
F3

(3)
1 1
u1
F1
(2.22-c)
= k3
(3)
u2
1 1
F2

(5)
1 1
u2
F2
(2.22-d)
= k5
(5)
u3
1 1
F3
3. The global equations of equilibrium being
(1)
(2)
(3)
(4)
= R1
F1 + F1 + F1 + F1
(2)
(3)
(5)
F2 + F2 + F2
= R2
(4)
(5)
F3 + F3
= R3
(2.23)
4. We next substitute the equilibrium equations of each element into the three global equations of equilibrium

(k2 + k3 )
k4
(k1 + k2 + k3 + k4 )
u1 R1
(k2 + k3 + k5 )
k5
u
R
(k2 + k3 )
=
(2.24)
K=
2 2
k4
k5
(k4 + k5 )
u3
R3
Victor Saouma
Draft
210
INTRODUCTION
5. The state variables (nodal displacements) ui can be solved through the inversion of K.
6. The forces in each spring can be determined from equilibrium of each individual spring.
2.3.1.2
Heat Transfer
15 A wall is constructed of two homogeneous slabs in contact. In steady-state conditions the

temperatures in the wall is characterized by the external surface temperature 1 and 3 and the
interface temperature 2 . Establish the equilibrium equations of the problem in terms of these
temperatures when the ambient temperatures 0 and 4 are known.
3
4
111
000
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000000000
111111111
0
1
0
1
000000000111111111111111111
111111111
000000000000000000
000000000
111111111
0
1
0
1
000000000
111111111
000000000000000000
111111111111111111
000000000
111111111
0
1
0
1
000000000
111111111
000000000000000000
111111111111111111
000000000
111111111
0
1
0
1
000000000
111111111
000000000000000000
111111111111111111
2k
3k 00
000000000
111111111
00
11
0
1
0
1
000000000
111111111
000000000000000000
111111111111111111
11
000000000
111111111
00
11
0
1
0
1
000000000
111111111
000000000000000000
111111111111111111
000000000
111111111
0
1
0
1
000000000000000000
111111111111111111
000000000
111111111
000000000
111111111
0
1
0
1
000000000
111111111
000000000000000000
111111111111111111
000000000
111111111
0
1
0
1
000000000
111111111
000000000000000000
111111111111111111
000000000
111111111
0
1
0
1
000000000
111111111
000000000000000000
111111111111111111
000000000
111111111
0
1
0
1
000000000
111111111
000000000000000000
111111111111111111
000000000
111111111
A 111111111
B
C
D
0
1
0
1
000000000000000000
111111111111111111
000000000
000000000
111111111
0
1
0
1
000000000000000000
111111111111111111
000000000
111111111
000000000
111111111
0
1
0
1
000000000000000000
111111111111111111
000000000
111111111
3k
000000000
111111111
2k
000000000
111111111
0
1
000000000000000000
111111111111111111
0
1
000000000
111111111
0
1
000000000000000000
111111111111111111
0
1
000000000
111111111
000000000
111111111
0
1
000000000000000000
0
1
111111111111111111
Conductance
Figure 2.6: Slab Subjected to Temperature Boundary Conditions, (Bathe 1996)
1. The governing equation is the heat conduction law

q = Ak
(2.25)
where q is the total heat ow, A the area, the temperature drop in the direction of
heat ow, and k the conductance or surface coecient.
2. The state variables are 1 , 2 and 3 .
3. We then apply the equilibrium equation for each interface
q1 = 3k(0 1 ) Convection A B
q2 = 2k(1 2 ) Conduction B
q = 3k(2 3 ) Conduction C
3
q4 = 2k(3 4 ) Convection C D
(2.26)
4. Heat ow equilibrium must be satised

q1 = q2 = q3 = q4
Victor Saouma
(2.27)
Draft
211
thus
3k(0 1 ) = 2k(1 2 )
(2.28-a)
2k(1 2 ) = 3k(2 3 )
(2.28-b)
3k(2 3 ) = 2k(3 4 )
(2.28-c)
5. These equations can be rewritten as

5k 2k
0
1 3k0
2k 5k 3k
0
=
2
0
3k 5k
3
2k4
2.3.1.3
(2.29)
Hydraulic Network
In this example, we seek to establish the equations that govern the steady-state pressure
and ow distribution in the hydraulic network shown in Fig. 2.7. We assume the uid to be
incompressible and the pressure drop in a branch to be proportional to the ow q through that
branch, (Darcys law)
d4 p
p = Rq
(2.30)
q=
128L
16
where d is the pipe diameter, the uid kinematic viscosity, L the pipe length, R is the branch
eective resistance coecient.
E
q
R=10b
R=
5b
R=2b
5
R=
R=3b q
Figure 2.7: Pipe Network, (Bathe 1996)
1. We consider each branch of the pipe to constitute an element, and we select as state
variables pA , pC , and pD the pressures at A, C, and D respectively.
2. Thus from equilibrium of each element
q1
q3
q2 |AC
q |
2 DB
q4
Victor Saouma
=
=
=
=
=
pA
10b
pC pD
2b
pA pC
5b
pD
5b
pC pD
3b
(2.31)
Draft
212
INTRODUCTION
3. The conservation mass requirement (or continuity of ow) requires that
Q = q1 + q2
q2 |AC = q3 + q4
q2 |DB = q3 + q4
(2.32)
Substituting, we get

3 2
0
pA 10bQ
6 31 25
p
=
0
C
pD
0
1 1
1
(2.33)
9
6
0
pA 30bQ
6 31 25
p
=
0
C
pD
0
0 25 31
(2.34)
or
4. The pressures are determined through the inversion of the matrix, and the ow can in
turn be determined from Eq. 2.31
2.3.1.4
DC Network
Considering the network shown in Fig. 2.8, determine the steady state current distribution
in the network.
17
I3
I3
I3
6R
2R
I - I3
1
I - I3
2
4R
2
1
I
2R
I2
B
2E
Figure 2.8: DC Network, (Bathe 1996)
1. The state variables will be the currents I1 , I2 , and I3 . Ohms law will be applied
E = RI
(2.35)
where E is the voltage drop across the resistor.

2. The equilibrium equation to be satised across each element interconnection will be
Kirchhos law
= 2E
2RI1 + 2R(I1 I3 )
(2.36)
= E
4R(I2 I3 )
6RI3 + 4R(I3 I2 ) + 2R(I3 I1 ) = 0
Victor Saouma
Draft
213
3. Rewriting this equation in matrix form

4R
0
2R I1 2E
0
I
=
E
4R 4R
2
I3
0
2R 4R 12R
2.3.2
(2.37)
Equivalent Truss/Direct Stiness Models
Each of the preceding problems can be discretized by an equivalent truss framework, and the
direct stiness method applied to assemble the global stiness matrix, Fig. 2.9.
u1 , R1
01
1010
1010
u2 , R2
u3 , R3
3K
2K
3K
2K
1
0
0
1
0
1
0
1
0
1
0110
100
11
1010
1010
2
5b
110
20
1010
10 b
3
2b
3b
u2 , R2
u1 , R1
1
0
0
31
0
1
0
1
5
5b
1
0
0
1
0
1
0
1
0
1
0
1
0
1
0
1
u3 , R3
u2 , R2
01
1010
10
1
2R
1
0
0
21
0
1
0
1
0
11
0
1
0
1
0
1
3
4R
2
2R
u1 , R1
1
0
0
1
0
1
0
31
0
1
0
1
0
1
0
1
0
1
4
6R
1
0
0
1
0
1
0
1
u3 , R3
Figure 2.9: Equivalent Trusses/Direct Stiness
Heat Transfer: where the displacement correspond to the temperature .
Victor Saouma
Draft
214
INTRODUCTION
1. Element stiness matrices

(1)
Spring 1:
Spring 2: 2K

Spring 3: 3K
1 1
1 1
1 1
1 1
Spring 4:
3Ku1 = F
1

(1)
u1
F1
=
(2)
u2
F2

(3)
F2
u2
=
(3)
u3
F3
(4)
2Ku3 = F3
(2.38)
2. Assemble active degrees of freedom

(2)
F1
2K
2K
0
u
1
(2)
(3)
2K 2K + 3K 3K
=
u
F +F
2 (3) 2
2
u3
0
3K
3K
F3
(2.39)
3. Equilbrium at each node
(1)
(2)
(2)
(1)
= R1
F1 + F1
= R1 F1
= 3K0 3Ku1
F1
(2)
(3)
R
=
0
F2 + F2
= R2
2
(3)
(4)
(3)
(4)
= R3 F3
= 2K4 2Ku3
F
F3 + F3
= R3
3
(2.40)
4. Rearrange state variables

5K 2K
0
u1 R1 3K0
2K 5K 3K
u
R
0
=
=
=
2 2
u3
R3
2K4
0
3K 5K
(2.41)
Pipe Network: where the displacement correspond to the pressure p

Spring 1:
Spring 2:
1
5b
Spring 3:
1
2b
Spring 4:
1
3b

1 1
1 1
1 1
1 1
1 1
1 1

1
10b
u1
u2
u2
u3
u2
u3
1
5b u3
Spring 5:
(1)
= F
1

(2)
F1
=
(2)
F2

(3)
F3
=
(3)
F4

(4)
F3
=
(4)
F4
(2.42)
(5)
= F4
1
5b
1
5b
Victor Saouma
1
5b
1
1
1
5b + 2b + 3b
1
f rac12b 3b
(2)
F1
0
u
1
(2)
(3)
(4)
1
1
2b
3b
u
=
F3 + F3 + F3
2
1
1
(3)
(4)
u3
F4 + F4
2b + 3b
(2.43)
Draft
215
(1)
(2)
(2)
(1)
1
= R1
F1 + F1
= R1 F1
= Q 10b
u1
F1
(2)
(3)
(4)
R2 = 0
F3 + F3 + F3
= R2
(3)
(4)
(5)
(3)
(4)
(5)
1
= R3 F4
= 0 5b
u3
F4 + F4
F4 + F4 + F4
= R3
(2.44)
4. Rearrange state variables and multiply by 30

9
6
0
u1 30bQ
6 31 25
u
=
0
2
u3
0
0 25 31
(2.45)
DC Network: where the displacement u correspond to the current intensity I

Spring 1:
(1)
Spring 2: 2R

Spring 3: 4R
1 1
1 1
1 1
1 1
Spring 4:
2Ru1 = F
1

(2)
u1
F1
=
(2)
u2
F3

(3)
F2
u2
=
(3)
u3
F3
(4)
6Ru3 = F3

(2.46)

(2)
F1
2R
0
2R
u1
(3)
0
u
=
4R
4R
F2
2
(2)
(3)
u3
2R 4R 2R + 4R
F3 + F3
(2.47)
(2)
(2)
(1)
+ F1
= R1
= R1 F1
= 2E 2Ru1
F1
(3)
(3)
F2
= R2
F2
= R2
= E
(2)
(3)
(4)
(2)
(3)
(4)
F3 + F3 + F3
= R3
= R3 F3
= 6Ru3
F3 + F3
(2.48)
(1)
F1
4. Rearrange state variables

4R
0
2R u1 2E 3K0
0
u2
0
=
E
=
=
4R 4R

u3
2K4
0
2R 4R 12R
2.3.2.1
2.3.3
(2.49)
Nonlinear Elastic Spring
Propagation Problems
18 The main characteristic of a propagation dynamic problem is that the response of the system
changes with time. In principle, we may apply the same analysis procedure as in steady-state
problems, however in this case the state variables and the the equilibrium relations depend on
time.
Victor Saouma
Draft
216
2.3.3.1
INTRODUCTION
Dynamic Elastic System
Let us reconsider the system of rigid carts previously analysed.We assume the springs to be
massless, but the carts have masses mi .
1. Invoking dAlemberts principle, the interelement
(1)
(2)
(3)
(4)
F1 + F1 + F1 + F1
(2)
(3)
(5)
F2 + F2 + F2
(4)
(5)
F3 + F3
where
u
i =
d2 ui
dt2
for
interconnectivity requirement yields

= R1 (t) m1 u
1
= R2 (t) m2 u
2
= R3 (t) m3 u
3
i = 1, 2, 3
(2.50)
(2.51)
2. Thus we obtain the following system of governing equilibrium equations

M
u + Ku = R(t)
(2.52)
0
m1 0
M = 0 m2 0
0
0 m3
(2.53)
where
3. The equilibrium equation represents a system of ordinary dierential equations of the

second order in time. For the solution of these equations it is also necessary to specify
the initial conditions for u and u at time t = 0 (u0 and u 0 respectively).
2.3.3.2
Transient Heat Flow
Fig. 2.10 illustrates an idealization of the heat ow inside an electron tube. A lament is
heated to a temperature f by an electric current; heat is convected from the lament to the
surrounding gas and is radiated to the wall which also receives heat by convection of the gas.
The wall itself convects heat to the surrounding atmosphere, which is at temperature a . It is
required to formulate the system-governing heat ow equilibrium equations.
1. The state variables are the temperature of the gas, 1 , and the temperature of the wall
2 .
2. The governing equations for heat transfer are

Gas
C1 d1 = k1 (f 1 ) k2 (1 2 )
dt
!
"
d
2
4
4
= kr (f ) (2 ) + k2 (1 2 ) k3 (2 a ) Wall
C2
dt
(2.54)
Note that the rst equation is Newtons law of cooling, and the second is the StefanBoltzman law of radiation.
3. The two equations can be written in matrix form as
C + K = Q

C1 0
C =
0 C2
Victor Saouma
(2.55-a)
(2.55-b)
Draft
217
Wall
Co
n
ve
c ti
on
Gas 1
k2
Atmosphere
Gas
1
0
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
n
tio
ec
nv
Co
1
0
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
Filament f
Radiation kr
Wall 2
Convection
k3
Atmosphere a
Filament
Figure 2.10: Heat Transfer Idealization in an Electron Tube, (Bathe 1996)

k2
(k1 + k2 )
k2
(k2 + k3 )

1
=
2

k1 f "
!
Q =
kr (f )4 (2 )4 + k3 a )

K =
(2.55-c)
(2.55-d)
(2.55-e)
4. We note that because of the radiation boundary conditions, the heat ow equilibrium
equations are nonlinear in .
2.3.4
2.3.4.1
19
Eigenvalue Problems
Free Vibration
The governing equation for the free (undamped) vibration of a structure is

M
u + Ku = 0
(2.56)
where the motion is referred to being free, since there are no applied loads.
20
By assuming a harmonic motion

u = sin t
(2.57)
the natural frequencies and the corresponding mode shapes can be determined from
the generalized eigenvalue problem
(2.58)
2 M = K
or
Since is nontrivial
Victor Saouma
(K 2 M) = 0
(2.59)
|K 2 M| = 0
(2.60)
Draft
218
INTRODUCTION
or with 2 =
|K M| = 0
(2.61)
which is the characteristic equation, and is called the eigenvalue of the equation, and the
structure is said to respond in the mode corresponding to a particular frequency.
For computational purposes, if we premultiply each side of the preceding equation by M1 ,
then
21
(M1 K I) = 0
(2.62)
It should be noted that a zero eigenvalue is obtained for each possible rigid body motion of a
structure that is not completely supported.
22 Depending on which mass matrix is adopted, slightly dierent results are obtained.
In
general, lumped mass matrices approach the exact value (consistent mass matrix) from below.
23 The mode shapes are shapes , anfd give a relative magnitude of the DOF, not the
absolute values (since they are the solution to a set of homogeneous equations).
24 The natural frequencies and mode shapes provide a fundamental description of the vibrating
structure.
2.3.4.2
Column Buckling
25 Next we consider the two rigid bar problem illustrated in Fig. 2.11. For this problem, we
must consider equilibrium of the deformed state (rather than undeformed), because of the large
deformation in presence:
P
P
C
L
B
k
Ak
0000000
1111111
0000 C
1111
1111111
0000000
0000
1111
1
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
2
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
0000000
1111111
0000
1111
B
0000000
1111111
0000
1111
L
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
1
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
1
0
0
1
0
1
0
1
0
1
0
1
0
1
0
0000000
1111111
C 1
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
L
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
2
0P
1
0000000
1111111
0
1
0000000
1111111
0
1
0000000
1111111
0
1
0000000
1111111
0
1
0000000
1111111
0
1
0000000
1111111
0
1
0000000
1111111
0
1
0000000
1111111
0
1
0000000
1111111
B
0000B
1111
0000
1111
0000
1111
0000k( 1111
k(2 - )1 1111
0000
2
0000
1111
0000
1111
0000
1111
0000
1111
P
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
A 1111
k1
0000
1111
0
1
0
1
0
1
0
1
0
1
0
1
P
0
1
0
1
0
1
0
P1
0
1
0
1
1
0
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
)1
1
0
0
1
0
1
0
1
0
1
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
1.618
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000000
1111111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
1
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0 P
1
1
0
0
1
0
1
0
1
0
1
0
1
0
1
000
111
000
111
000
111
000
111
000
111
000
111
000
111
.618
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
1
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
Figure 2.11: Stability of a Two Rigid Bars System

MB = P L2 k(2 1 ) = 0
PL
2
1 + 2 =
k
Victor Saouma
(2.63-a)
(2.63-b)
Draft
26
MA = P L1 + k(2 1 ) k1 = 0
PL
1
21 2 =
k
Those two equations can be cast in matrix form

1
2 1
1
=
2
2
1 1
where = P L/k, this is an eigenvalue formulation and can be rewritten as

0
2 1
1
=
2
0
1 1

2 1

1 1 = 0
2 2 + 2 1 = 0
1,2 =
27
(2.63-d)
(2.64)
(2.65-a)
(2.65-b)
3 5
3 94
=
2
2
(2.65-d)
(2.65-e)
Hence we now have two critical loads:
k
3 5k
= 0.382
2 L
L
k
3+ 5k
= 2.618
2
L
L
Pcr1 =
Pcr2 =
28
(2.63-c)
(2.65-c)
3 + 1 = 0
2
219
(2.66)
(2.67)
We now seek to determine the deformed shape for each of the rst critical loads
3 5
1 =
2
$
#

3 5
1
0
1
2 2
=
3 5
0
2
1
1 2
$
#

1+ 5
1
0
1
2
=
1+ 5
0
2
1 1 2

0
1.618 1
1
=
2
0
1 0.618
(2.68-a)
(2.68-b)
(2.68-c)
(2.68-d)
we now arbitrarily set 1 = 1, then 2 = 1/0.618 = 1.618, thus the rst eigenmode is

Victor Saouma
1
2

=
1
1.618

(2.69)
Draft
220
29
INTRODUCTION
Finally, we examine the second mode shape loads
2 =

1
1
2 3+2 5
=
2
1
1 3+2 5
#
$

1 5
1
1
2
=
2
1 1 12 5

0.618
1
1
=
2
1
1.618
$
3+ 5
2

0
0

0
0

0
0
(2.70-a)
(2.70-b)
(2.70-c)
(2.70-d)
we now arbitrarily set 1 = 1, then 2 = 1/1.618 = 0.618, thus the second eigenmode is

2.4
1
2

=
1
0.618

(2.71)
Solution Strategies
Within the context computer based simulation of a partial dierential equation, we have
three alternatives for a numerical solution
30
Strong Form
Weak Form
FE
Variational Form
Discretization
Mathematical
Model
Discrete
Model
Solution
Physical
System
Idealization
FD
Discrete
Solution
Solution error
Discretization + solution error
Modeling + discretization+ solution error
Figure 2.12: Idealization, Discretization and Solution of a Numerical Simulation, (Felippa 2000)
31 Two of the three alternative forms for discretization have already been introduced in Finite
Element I and are:
Strong Form/Dierential Form, SF/DF which are the Euler Equations corresponding to
the variational formulation. The strong form is represented by a system of ordinary or
partial dierential equations in space and/or time complemented by appropriate boundary
conditions. The strong form is numerically solved by the nite dierence FDM method.
Victor Saouma
Draft
2.4 Solution Strategies
221
Variational Form, VF: is a functional which must be stationary (maximum or minimum),

and which can lead to either the strong form (corresponding Euler equations), or to the
weak form. The variational form is solved by the Rayleigh-Ritz method which leads
to a nite element formulation (FEM).
Weak Form, WF: is a weighted integral equation which relaxes the strong form to be
satised on an average sense inside a nite element through the Weighted Residual
method.
32
Note that:
1. The variational method provides a relatively easy way to construct the system of governing equations. This ease stems from the fact that in the variational formulation scalar
quantities (energies, potentials and so on) are considered rather than vector quantities
(forces, displacements, etc.),
2. Not all problems can be solved by the VF, whereas most can be solved by the WF.
3. In complex problems, a combination of techniques is used
Fluid-structure interaction: FEM for the structure, FDM for the uid.
Structural dynamics: FEM in space, and FDM in time.
2.4.1
33
Euler Equation
Given a functional

(u) =
F (x, u, u )dx
(2.72)
it can be shown that its corresponding Euler equation is given by

d F
F
= 0 in a < x < b
u
dx u
(2.73)
This dierential equation is called the Euler equation associated with and is a necessary
condition for u(x) to extremize .
34 Generalizing for a functional which depends on two eld variables, u = u(x, y) and v =
v(x, y)

(2.74)
=
F (x, y, u, v, u,x , u,y , v,x , v,y , , v,yy )dxdy
There would be as many Euler equations as dependent eld variables

F
F
F
F
2 F
2
2 F
u x u,x y u,y + x2 u,xx + xy u,xy + y 2 u,yy
F
v
F
x v,x
F
y v,y
2 F
x2 v,xx
F
2
xy v,xy
2 F
y 2 v,yy
= 0
= 0
(2.75)
Example 2-4: Flexure of a Beam
Victor Saouma
Draft
222
INTRODUCTION
The total potential energy of a beam supporting a uniform load p is given by

L
L
1
1

M pw dx =
(EIw )w pw dx
=
2
2
0
0

(2.76)
Solution:
1. From Eq. 2.75
&
%
= 12 (EIw )w pw
F
w
'
( = p
F
= (EIw )
w,xx
F
2
x2
(EIw ) = p Euler Equation
2. Alternatively, integrating by part twice

L
L
F
F

w dx
F dx =
w +
=
w
w
0
0
L
(EIw w pw)dx
=
0
L

"
!

L
(EIw ) w pw dx
= (EIw w ) 0
0
L

"
!
L

L

(EIw ) + p wdx = 0
= (EIw w ) 0 [(EIw ) w]|0 +

0
Nat. Ess.
Nat. Ess.
Euler Eq.

BC
Or
(EIw ) = p
(2.77)
(2.78-a)
(2.78-b)
(2.78-c)
(2.78-d)
for all x
which is the governing dierential equation of beams and

Essential
w = 0
w = 0
or
or
Natural
EIw = M = 0
(EIw ) = V = 0
at x = 0 and x = L
3. The variational form is thus given by
L
"
!
(EIw ) + p wdx = 0

0
Euler Eq.
4. The weak form will be given by
L
0
Victor Saouma
"
(EIw ) + p r(x)dx = 0

Euler Eq.
!
(2.79)
(2.80)
Draft
2.5 Computer Programs
2.5
35
223
Computer Programs
Modern programs are composed of three modules

1. Mesh generator
2. Analysis
3. Post-Pocessor
36
Major nite element codes
NASTRAN Originally developed by NASA, primarily used by NASA and its contractors.
Original version public domain, later version (McNeal Schwindler) commercial.
SAP Originally developed by Ed. Wilson at Berkeley. First version public domain, later ones
available only for PC (SAP90). NONSAP is the nonlinear version of SAP
ANSYS Commercial program mostly used in the Nuclear industry.
ABAQUS Probably the most modern and widely used commercial nite element. (Available
in Bechtel Lab).
ALGOR Very widely used PC/based code, mechanical/civil applications.
FEAP Public domain code listed in Zienkiewicz et al., often used in academia as a base for
extension.
MERLIN our very own!.
2.6
Examples of applications
1. Aircraft, automobile, submarine

2. Dam, buildings, bridges
3. Mechanical design and optimization
4. Heat transfer
5. Biomechanics (hip joints)
6. Electrical (design of rotors)
7. Coastal engineering
8. Fluid mechanics
9. Coupled problems
Victor Saouma
Draft
224
Victor Saouma
INTRODUCTION
Draft
Chapter 3
FUNDAMENTAL RELATIONS
3.1
Introduction
Whereas, ideally, a course in Continuum Mechanics should be taken prior to a nite element
course, this is seldom the case. Most often, students have had a graduate course in Advanced
Strength of Materials, which can only provide limited background to a solid nite element
course.
1
2 Accordingly, this preliminary chapter (mostly extracted from the authors lecture notes in
Continuum Mechanics) will partially remedy for occasional deciencies and will be often referenced.
3 It should be noted that most, but not all, of the material in this chapter will be subsequently
referenced.
3.1.1
4
Notation
Dierent set of notations are commonly used in Engineering:
Matrix: Finite Element [A], [], {F }

Indicial: Mechanics cartesian, Fx , ij , Cijkl
Tensorial: Mechanics cartesian/curvilinear, F, , C
Engineering: (Timoshenko/Voigt) Elasticity, x , xy
T
Ax = xi Aij xj
= x
x.A.x

tensor matrix indicial
(3.1)
In the following sections, we shall briey explain the last two.

3.1.1.1
Indicial Notation
5 Whereas the Engineering notation may be the simplest and most intuitive one, it often leads
to long and repetitive equations. Alternatively, the tensor will lead to shorter and more compact
forms.
Draft
32
6 While working on general relativity, Einstein got tired of writing the summation symbol with
)
its range of summation below and above (such as n=3
i=1 aij bi ) and noted that most of the time
the upper range (n) was equal to the dimension of space (3 for us, 4 for him), and that when
the summation involved a product of two terms, the summation was over a repeated index
) (i
in our example). Hence, he decided that there is no need to include the summation sign
if
there was repeated indices (i), and thus any repeated index is a dummy index and is summed
over the range 1 to 3. An index that is not repeated is called free index and assumed to take
a value from 1 to 3.
7
Hence, this so called indicial notation is also referred to Einsteins notation.
The following rules dene indicial notation:

1. If there is one letter index, that index goes from i to n (range of the tensor). For instance:
a1
a
i = 1, 3
(3.2)
ai = ai = a1 a2 a3 =
2
a3
assuming that n = 3.
2. A repeated index will take on all the values of its range, and the resulting tensors summed.
For instance:
a1i xi = a11 x1 + a12 x2 + a13 x3
(3.3)
3. Tensors order:
First order tensor (such as force) has only one free index:
ai = ai = a1 a2 a3
(3.4)
other rst order tensors aij bj , Fikk , ijk uj vk

Second order tensor (such as stress or strain) will have two free indeces.
D11 D22 D13

Dij = D21 D22 D23
D31 D32 D33
(3.5)
other examples Aijip , ij uk vk .

A fourth order tensor (such as Elastic constants) will have four free indeces.
4. Derivatives of tensor with respect to xi is written as , i. For example:
xi
= ,i
vi
xi
= vi,i
vi
xj
= vi,j
Ti,j
xk
= Ti,j,k
(3.6)
Usefulness of the indicial notation is in presenting systems of equations in compact form. For
instance:
x1 = c11 z1 + c12 z2 + c13 z3

x = c21 z1 + c22 z2 + c23 z3
(3.7)
xi = cij zj
2
x3 = c31 z1 + c32 z2 + c33 z3
9
Victor Saouma
Draft
3.1 Introduction
33
Similarly:
Aij = Bip Cjq Dpq
3.1.1.2
A11
A12
21
A22
=
=
=
=
B11 C11 D11 + B11 C12 D12 + B12 C11 D21 + B12 C12 D22
B11 C11 D11 + B11 C12 D12 + B12 C11 D21 + B12 C12 D22
B21 C11 D11 + B21 C12 D12 + B22 C11 D21 + B22 C12 D22
B21 C21 D11 + B21 C22 D12 + B22 C21 D21 + B22 C22 D22
(3.8)
Tensor Notation
10 In tensor notation the indices are not shown. While cartesian indicial equations apply only to
Cartesian coordinates, expressions in tensor notation are independent of the coordinate system
and apply to other coordinate systems such as cylindrical, curvilinear, etc.
11 In tensor notation, we indicate tensors of order one or greater in boldface letters. It is usually
recommended to use lower case letters for tensors of order one, and upper case letters for tensors
of order 2 and above.
12 We distinguish between tensor notation and matrix notation by using dots and colons between
terms, as in a.b ai bi , (the . indicates a contraction of a pair of repeated indices which appear
in the same order), and A : B Aij Bij . or ij = Cijkl kl is equivalent to = C : .
3.1.1.3
Voigt Notation
In nite element, symmetric second-order tensors are often written as column matrices. This
conversion, and the one of other higher-order tensors into column matrices are called Voigt
notation.
13
The Voigt rule depends on whether the

kinematic one (such as strain).

11 12
21 22

11 12
21 22
14
3.1.2
tensor is a kinetic quantity (such as stress) or a

11 1
22
=
= {}
2
12
3

11 1
22
=
= {}
2
212
3
(3.9-a)
(3.9-b)
Tensors
We generalize the concept of a vector by introducing the tensor (T), which essentially exists
to operate on vectors v to produce other vectors (or on tensors to produce other tensors!). We
designate this operation by Tv or simply Tv.
15
16
We hereby adopt the tensor (or dyadic) notation for tensors as linear vector operators
u = Tv or ui = Tij vj
u = vS where S = T
Victor Saouma
(3.10-a)
T
(3.10-b)
Draft
34
17 Whereas a tensor is essentially an operator on vectors (or other tensors), it is also a physical
quantity, independent of any particular coordinate system yet specied most conveniently by
referring to an appropriate system of coordinates.
18 Tensors frequently arise as physical entities whose components are the coecients of a linear
relationship between vectors.
19 A tensor is classied by the rank or order.
A Tensor of order zero is specied in any
coordinate system by one coordinate and is a scalar. A tensor of order one has three coordinate
components in space, hence it is a vector. In general 3-D space the number of components of
a tensor is 3n where n is the order of the tensor.
20
A force and a stress are tensors of order 1 and 2 respectively.
21
The sum of two (second order) tensors is simply dened as:

Sij = Tij + Uij
22
(3.11)
The multiplication of a (second order) tensor by a scalar is dened by:

Sij = Tij
(3.12)
In a contraction, we make two of the indices equal (or in a mixed tensor, we make a subscript
equal to the superscript), thus producing a tensor of order two less than that to which it is
applied. For example:
2 0
Tij Tii ;
2 0
ui vj ui vi ;
mr
r
(3.13)
4 2
Amr
..sn A..sm = B.s ;
3 1
Eij ak Eij ai = cj ;
mp
Ampr
Ampr
qs
qr = Bq ; 5 3
23
The outer product of two tensors (not necessarily of the same type or order) is a set of
tensor components obtained simply by writing the components of the two tensors beside each
other with no repeated indices (that is by multiplying each component of one of the tensors by
every component of the other). For example
24
ai bj
i
Bj.k
= Tij
= C
(3.14-a)
i.k
.j
(3.14-b)
vi Tjk = Sijk
(3.14-c)
The inner product is obtained from an outer product by contraction involving one index
from each tensor. For example
25
ai bi
(3.15-a)
ai Ejk ai Eik = fk
(3.15-b)
ai bj
Eij Fkm Eij Fjm = Gim

i
A
Victor Saouma
Bi.k
Bi.k
=D
(3.15-c)
(3.15-d)
Draft
3.2 Vector Fields; Solid Mechanics
26
35
The scalar product of two tensors is dened as

T : U = Tij Uij
(3.16)
in any rectangular system.
3.2
3.2.1
3.2.1.1
27
Vector Fields; Solid Mechanics

Kinetics
Force, Traction and Stress Vectors
There are two kinds of forces in continuum mechanics
body forces: act on the elements of volume or mass inside the body, e.g. gravity, electromagnetic elds. dF = bdV ol.
surface forces: are contact forces acting on the free body at its bounding surface. Those will
be dened in terms of force per unit area.
28 The surface force per unit area acting on an element dS is called traction or more accurately
stress vector.

tdS = i
tx dS + j
ty dS + k
tz dS
(3.17)
Most authors limit the term traction to an actual bounding surface of a body, and use the term
stress vector for an imaginary interior surface (even though the state of stress is a tensor and
not a vector).
The traction vectors on planes perpendicular to the coordinate axes are particularly useful.
When the vectors acting at a point on three such mutually perpendicular planes is given, the
stress vector at that point on any other arbitrarily inclined plane can be expressed in terms
of the rst set of tractions.
29
30 A stress, Fig 3.1 is a second order cartesian tensor, ij where the 1st subscript (i) refers to
the direction of outward facing normal, and the second one (j) to the direction of component
force.

11 12 13
t1
t
(3.18)
= ij = 21 22 23 =
2
31 32 33
t3
In fact the nine rectangular components ij of turn out to be the three sets of three vector
components (11 , 12 , 13 ), (21 , 22 , 23 ), (31 , 32 , 33 ) which correspond to the three tractions
t1 , t2 and t3 which are acting on the x1 , x2 and x3 faces (It should be noted that those tractions
are not necesarily normal to the faces, and they can be decomposed into a normal and shear
traction if need be). In other words, stresses are nothing else than the components of tractions
(stress vector), Fig. 3.1.
31
32 The state of stress at a point cannot be specied entirely by a single vector with three
components; it requires the second-order tensor with all nine components.
Victor Saouma
Draft
36
X3
X3
V3
33
32
t3
23
31
t2
13
V2
X2
22
V1
X1
21
X2
(Components of a vector are scalars)
12
t1
11
X 1 Stresses as components of a traction vector

(Components of a tensor of order 2 are vectors)
Figure 3.1: Stresses as Tensor Components

3.2.1.2
Traction on an Arbitrary Plane; Cauchys Stress Tensor
33 Let us now consider the problem of determining the traction acting on the surface of an
oblique plane (characterized by its normal n) in terms of the known tractions normal to the
three principal axis, t1 , t2 and t3 . This will be done through the so-called Cauchys tetrahedron
shown in Fig. 3.2, and will be obtained without any assumption of equilibrium and it will apply
X2
-t
1
-t
t*n S
h N
S3
n
A
X1
V
*
-t 2 S2
*
X3
Figure 3.2: Cauchys Tetrahedron

in uid dynamics as well as in solid mechanics.
34
This equation is a vector equation, and the corresponding algebraic equations for the com-
Victor Saouma
Draft
37
ponents of tn are
tn1
tn2
tn3
Indicial notation tni
Tensor notation
tn
=
=
=
=
=
11 n1 + 21 n2 + 31 n3
12 n1 + 22 n2 + 32 n3
13 n1 + 23 n2 + 33 n3
ji nj
n = T n
(3.19)
We have thus established that the nine components ij are components of the second order
tensor, Cauchys stress tensor.
35
Example 3-1: Stress Vectors

if the stress tensor at point P is given by

7 5 0
t1
t
= 5 3 1 =
2
t3
0
1 2
(3.20)
We seek to determine the traction (or stress vector) t passing through P and parallel to the
plane ABC where A(4, 0, 0), B(0, 2, 0) and C(0, 0, 6).
Solution:
The vector normal to the plane can be found by taking the cross products of vectors AB and
AC:

e 1 e2 e3

(3.21-a)
N = ABAC = 4 2 0
4 0 6
(3.21-b)
= 12e1 + 24e2 + 8e3
The unit normal of N is given by
n=
6
2
3
e1 + e2 + e3
7
7
7
Hence the stress vector (traction) will be
7 5 0
37 76 27 5 3 1 = 97
0
1 2
and thus t = 97 e1 + 57 e2 +
3.2.2
36
(3.22)
5
7
10
7
(3.23)
10
7 e3
Kinematic
In continuum mechanics, we dierentiate between
Material coordinates (X1 , X2 , X3 ) dened in the undeformed original coordinate system

which gives rise to the Lagrangian coordinate system.
Victor Saouma
Draft
38
Spatial Coordinates (x1 , x2 , x3 ) dened in the deformed coordinate system. This gigives rise
to the Eulerian coordinate system
1. If both the displacement gradients and the displacements themselves are small, then
ui
ui
Xj xj and thus the Eulerian and the Lagrangian innitesimal strain tensors may be
.
taken as equal Eij = Eij
2. If the displacement gradients are small, but the displacements are large, we should use
the Eulerian innitesimal representation.
3. If the displacements gradients are large, but the displacements are small, use the Lagrangian nite strain representation.
4. If both the displacement gradients and the displacements are large, use the Eulerian nite
strain representation.
37
The Lagrangian nite strain tensor can be written as

1
Eij =
2
or:
38
uj
ui
uk uk
+
+
Xj
Xi Xi Xj

or E =
1
(uX + X u + X uuX )

2
J+Jc
J c J
#

$
u1 2
u1
1
u2 2
u3 2
+
+
+
E11 =
X1 2
X1
X1
X1

1 u1
u2
u2 u2
u3 u3
1 u1 u1
+
+
+
+
E12 =
2 X2 X1
2 X1 X2 X1 X2 X1 X2
=
(3.24)
(3.25-a)
(3.25-b)
(3.25-c)
If large deformation is accounted for (such as in buckling), the Eulerian nite strains are:
xx =
u 1
+
x 2
yy =
v 1
+
y 2
zz =
xy =
xz =
yz =
Victor Saouma
#
#
u
x
2

+
v
x
2

+
w
x
2 $

$
2
u 2
v
w 2
+
+
y
y
y
#

$

u 2
w 1
v 2
w 2
+
+
+
z
2
z
z
z

u u u v v w w
1 v
+
+
+
+
2 x y
x y
x y
x y

1 w u u u v v w w
+
+
+
+
2 x
z
x z
x z
x z

1 w v u u v v w w
+
+
+
+
2 y
z
y z
y z
y z
(3.26)
(3.27)
(3.28)
(3.29)
(3.30)
(3.31)
Draft
39
or
ij =
1
(ui,j + uj,i + uk,i uk,j )
2
(3.32)
From this equation, we note that:

1. We dene the engineering shear strain as
ij = 2ij
(i = j)
(3.33)
2. If the strains are given, then these strain-displacements provide a system of (6) nonlinear
partial dierential equation in terms of the unknown displacements (3).
3. ik is the Green-Lagrange strain tensor.
4. The strains have been expressed in terms of the coordinates x, y, z in the undeformed state,
i.e. in the Lagrangian coordinate which is the preferred one in structural mechanics.
5. Alternatively we could have expressed ds 2 ds2 in terms of coordinates in the deformed
state, i.e. Eulerian coordinates x , y , z , and the resulting strains are referred to as the
Almansi strain which is the preferred one in uid mechanics.
6. In most cases the deformations are small enough for the quadratic term to be dropped,
the resulting equations reduce to
xx =
yy =
zz =
xy =
xz =
yz =
u
x
v
y
w
z
u
v
+
x y
w u
+
x
z
w v
+
y
z
(3.34)
(3.35)
(3.36)
(3.37)
(3.38)
(3.39)
or
ij =
1
(ui,k + uk,i )
2
(3.40)
which is called the Cauchy strain

39
In nite element, the strain is often expressed through the linear operator L
= Lu
Victor Saouma
(3.41)
Draft
310
or
3.2.3
xx
yy
zz
xy
xz
yz

=
y
0

0
0
0

x
ux
uy
uz

(3.42)
Fundamental Laws of Continuum Mechanics
In this section, we will derive dierential equations governing the way stress and deformation
vary at a point and with time. They will apply to any continuous medium, and yet we will
not have enough equations to determine unknown tensor eld. For that we need to wait for
the next section where constitututive laws relating stress and strain will be introduced. Only
with constitutive equations and boundary and initial conditions would we be able to obtain a
well dened mathematical problem to solve for the stress and deformation distribution or the
displacement or velocity elds.
40
41 In this section we shall summarize the dierential equations which express locally the conservation of mass, momentum and energy.
These dierential equations of balance will be derived from integral forms of the equation of
balance expressing the fundamental postulates of continuum mechanics.
42
43 Conservation laws constitute a fundamental component of classical physics. A conservation law establishes a balance of a scalar or tensorial quantity in voulme V bounded by a surface
S. In its most general form, such a law may be expressed as

d
AdV +
dS
=
AdV
(3.43)
dt V

S
V
Rate of variation
Exchange by Diusion
Source
where A is the volumetric density of the quantity of interest (mass, linear momentum, energy,
...) a, A is the rate of volumetric density of what is provided from the outside, and is the
rate of surface density of what is lost through the surface S of V and will be a function of the
normal to the surface n.
44 Hence, we read the previous equation as: The input quantity (provided by the right hand
side) is equal to what is lost across the boundary, and to modify A which is the quantity of
interest. The dimensions of various quantities are given by
dim(a) = dim(AL3 )
(3.44-a)
dim() = dim(AL
2 1
(3.44-b)
dim(A) = dim(AL
3 1
(3.44-c)
Hence this section will apply the previous conservation law to mass, momentum, and energy.
The resulting dierential equations will provide additional interesting relation with regard to
45
Victor Saouma
Draft
311
the imcompressibiltiy of solids (important in classical hydrodynamics and plasticity theories),

equilibrium and symmetry of the stress tensor, and the rst law of thermodynamics.
46 The enunciation of the preceding three conservation laws plus the second law of thermodynamics, constitute what is commonly known as the fundamental laws of continuum
mechanics.
47 Prior to the enunciation of the rst conservation law, we need to dene the concept of ux
across a bounding surface.
The ux across a surface can be graphically dened through the consideration of an imaginary surface xed in space with continuous medium owing through it. If we assign a
positive side to the surface, and take n in the positive sense, then the volume of material
owing through the innitesimal surface area dS in time dt is equal to the volume of the cylinder
with base dS and slant height vdt parallel to the velocity vector v, Fig. 3.3 (If vn is negative,
48
vn dt
vdt
dS
Figure 3.3: Flux Through Area dS
then the ow is in the negative direction). Hence, we dene the volume ux as

Volume Flux =
vndS =
S
vj nj dS
(3.45)
where the last form is for rectangular cartesian components.

49
We can generalize this denition and dene the following uxes per unit area through dS:

Mass Flux =
vndS =
vj nj dS
(3.46)
S
S

v(vn)dS =
vk vj nj dS
(3.47)
Momentum Flux =
S
S

1 2
1
v (vn)dS =
vi vi vj nj dS (3.48)
Kinetic Energy Flux =
2
S2

S
qndS =
qj nj dS
(3.49)
Heat ux =
S
S
JndS =
Jj nj dS
(3.50)
Electric ux =
S
Victor Saouma
Draft
312
3.2.3.1
Conservation of Mass; Continuity Equation

vi
d
d
+
+ v = 0
= 0 or
dt
xi
dt
(3.51)
The vector form is independent of any choice of coordinates. This equation shows that the
divergence of the velocity vector eld equals (1/)(d/dt) and measures the rate of ow of
material away from the particle and is equal to the unit rate of decrease of density in the
neighborhood of the particle.
50 If the material is incompressible, so that the density in the neighborhood of each material
particle remains constant as it moves, then the continuity equation takes the simpler form
vi
= 0 or v = 0
xi
(3.52)
this is the condition of incompressibility

3.2.3.2
Linear Momentum Principle; Equation of Motion
51 The momentum principle states that the time rate of change of the total momentum of a
given set of particles equals the vector sum of all external forces acting on the particles of the
set, provided Newtons Third Law applies. The continuum form of this principle is a basic
postulate of continuum mechanics.

d
tdS +
bdV =
vdV
(3.53)
dt V
S
V
Then we substitute ti = Tij nj and apply the divergence theorm (Appendix A, (Schey 1973)) to
obtain

dvi
Tij
dV
(3.54-a)
+ bi dV =
x
j
V
V dt

dvi
Tij
dV = 0
(3.54-b)
+ bi
xj
dt
V
or for an arbitrary volume
Tij
dv
dvi
or T + b =
+ bi =
xj
dt
dt
(3.55)
which is Cauchys (rst) equation of motion, or the linear momentum principle, or

more simply equilibrium equation.
52
When expanded in 3D, this equation yields:

T11 T12 T13
+
+
+ b1 = 0
x1
x2
x3
T21 T22 T23
+
+
+ b2 = 0
x1
x2
x3
T31 T32 T33
+
+
+ b3 = 0
x1
x2
x3
Victor Saouma
(3.56-a)
Draft
313
We note that these equations could also have been derived from the free body diagram
shown in Fig. 3.4 with the assumption of equilibrium (via Newtons second law) considering
an innitesimal element of dimensions dx1 dx2 dx3 . Writing the summation of forces, will
yield
53
(3.57)
Tij,j + bi = 0
where is the density, bi is the body force (including inertia).
yy+ yy d y
y
dy
yx+
yx y
d
y
xx+
xx
xy
xy+
xx d x
x
xy x
d
x
yx
yy
dx
Figure 3.4: Equilibrium of Stresses, Cartesian Coordinates
3.2.3.3
Conservation of Energy; First Principle of Thermodynamics
The rst principle of thermodynamics relates the work done on a (closed) system and the
heat transfer into the system to the change in energy of the system. We shall assume that
the only energy transfers to the system are by mechanical work done on the system by surface
traction and body forces, by heat transfer through the boundary.
54
If mechanical quantities only are considered, the principle of conservation of energy for
the continuum may be derived directly from the equation of motion given by Eq. 3.55. This is
accomplished by taking the integral over the volume V of the scalar product between Eq. 3.55
and the velocity vi .

dvi
dV
(3.58)
vi Tji,j dV +
bi vi dV =
vi
dt
V
V
V
55
Applying the divergence theorem,

dU
dW
dK
+
=
+Q
dt
dt
dt
Victor Saouma
(3.59)
Draft
314
this equation relates the time rate of change of total mechanical energy of the continuum on
the left side to the rate of work done by the surface and body forces on the right hand side.
56 If both mechanical and non mechanical energies are to be considered, the rst principle states
that the time rate of change of the kinetic plus the internal energy is equal to the sum of the
rate of work plus all other energies supplied to, or removed from the continuum per unit time
(heat, chemical, electromagnetic, etc.).
57 For a thermomechanical continuum, it is customary to express the time rate of change of
internal energy by the integral expression

d
dU
udV
(3.60)
=
dt
dt V
where u is the internal energy per unit mass or specic internal energy. We note that U
appears only as a dierential in the rst principle, hence if we really need to evaluate this
quantity, we need to have a reference value for which U will be null. The dimension of U is one
of energy dim U = M L2 T 2 , and the SI unit is the Joule, similarly dim u = L2 T 2 with the
SI unit of Joule/Kg.
3.2.4
Constitutive Equations
ceiinosssttuu
Hooke, 1676
Ut tensio sic vis
Hooke, 1678
3.2.4.1
58
General 3D
The Generalized Hookes Law can be written as:

ij = Dijkl kl
(3.61)
i, j, k, l = 1, 2, 3
The (fourth order) tensor of elastic constants Dijkl'has 81((34 ) components however, due to
distinct elastic terms.
the symmetry of both and , there are at most 36 9(91)
2
59
For the purpose of writing Hookes Law, the double indexed system is often replaced by a
simple indexed system with a range of six:
60
62 =36

k = Dkm m
61
k, m = 1, 2, 3, 4, 5, 6
(3.62)
In terms of Lames constants, Hookes Law for an isotropic body is written as

Tij = ij Ekk + 2Eij

1
Tij
ij Tkk
Eij =
2
3 + 2
Victor Saouma
or
or
T = IE + 2E
1
IT +
T
E=
2(3 + 2)
2
(3.63)
(3.64)
Draft

62
In terms of engineering constants:

1
E
63
315
; =
(3 + 2)
2( + )
E
E
; = G =
(1 + )(1 2)
2(1 + )
(3.65)
(3.66)
Similarly in the case of pure shear in the x1 x3 and x2 x3 planes, we have

21 = 12 = all other ij = 0
212 =
G
(3.67-a)
(3.67-b)
and the is equal to the shear modulus G.

64
65
66
Hookes law for isotropic material in terms of engineering constants becomes

E
ij +
ij kk
+
I
ij =
or =
(3.68)
1+
1 2
1+
1 2
1+
1+
ij ij kk or =
I
(3.69)
ij =
E
E
E
E
When the strain equation is expanded in 3D cartesian coordinates it would yield:
0
0
0
xx
xx
0
0
0
yy
yy
1 1
0
0
0
zz
zz
=
0
0 1+
0
0
xy (2xy )
xy
E 0
(2
)
0
0
0
0
1
+
yz
yz
yz
zx
zx (2zx )
0
0
0
0
0
1+
If we invert this equation, we obtain
xx
1
xx
yy
yy
(1+)(12)
zz
zz
1
0
0
xy
xy (2xy )

(2 )
0
G 0 1 0
yz
yz yz
zx
zx (2zx )
0 0 1
(3.70)
(3.71)
3.2.4.2
67
Transversly Isotropic Case
For transversely isotropic, we can express the stress-strain relation in tems of
Victor Saouma
xx
yy
zz
xy
yz
xz
=
=
=
=
=
=
a11 xx + a12 yy + a13 zz

a12 xx + a11 yy + a13 zz
a13 (xx + yy ) + a33 zz
(3.72)
2(a11 a12 )xy
a44 xy
a44 xz
Draft
316
and
a11 =
1
;
E
a12 =
;
E
a13 =

;
E
a33 =
1
;
E
a44 =
(3.73)
where E is the Youngs modulus in the plane of isotropy and E the one in the plane normal
to it. corresponds to the transverse contraction in the plane of isotropy when tension is
applied in the plane; corresponding to the transverse contraction in the plane of isotropy
when tension is applied normal to the plane; corresponding to the shear moduli for the plane
of isotropy and any plane normal to it, and is shear moduli for the plane of isotropy.
3.2.4.3
68
Special 2D Cases
Often times one can make simplifying assumptions to reduce a 3D problem into a 2D one.
3.2.4.3.1 Plane Strain 69 For problems involving a long body in the z direction with no
variation in load or geometry, then zz = yz = xz = xz = yz = 0. Thus, replacing into Eq.
3.71 we obtain
(1
xx
xx
(1 )
0
yy
(3.74)
=
yy
(1
+
)(1
2)
zz
xy
12
xy
0
0
2
3.2.4.3.2
and
Axisymmetry
70
In solids of revolution, we can use a polar coordinate sytem
rr =
=
zz =
rz =
71
u
r
u
r
w
z
u w
+
z
r
The constitutive relation is again analogous to 3D/plane strain
0
rr
E
zz

1
0
=

(1 + )(1 2)
1
0
rz
12
0
0
0
2
(3.75-a)
(3.75-b)
(3.75-c)
(3.75-d)
rr
zz
rz
(3.76)
3.2.4.3.3 Plane Stress 72 If the longitudinal dimension in z direction is much smaller than
in the x and y directions, then yz = xz = zz = xz = yz = 0 throughout the thickness.
Victor Saouma
Draft
317
Again, substituting into Eq. 3.71 we obtain:
1
0
xx
xx
1
=
yy
1
0
yy
1 2
1
xy
0 0
xy
2
1
zz =
(xx + yy )
1
3.2.4.4
(3.77-a)
(3.77-b)
Pore Pressures
73 In porous material, the water pressure is transmitted to the structure as a body force of
magnitude
bx =
p
x
by =
p
y
(3.78)
where p is the pore pressure.

The eective stresses are the forces transmitted between the solid particles and are dened
in terms of the total stresses and pore pressure p
74

ij
= ij + mij p
mT = [ 1, 1, 0 ]
(3.79)
i.e simply removing the hydrostatic pressure component from the total stress.
3.2.5
Field Equations for Thermo- and Poro Elasticity
Adapted from (Reich 1993)

75
In the absence of thermal/initial stresses and pore pressures, the eld equations are written
as

+ bi = 0 Equilibrium
ij,j

ij nj ti = 0 Natural B.C.
(3.80)
where bi and ti are the body forces and surface tractions respectively.
These equations will form the basis of the variational formulation of the nite element
method.
76
77 To account for the eect of thermal/initial stresses and pore pressures we seek to modify Eq.
3.80, in such a way that bi and ti are replaced by b i and t i .
Thermal strains are caused by a change in temperature with respect to the stress-free condition.
78
79 In Thermo- or poro-elasticity problems, the thermal strains and pore pressures are treated
as initial strains and stresses, respectively.
In porous media, the total stress is equal to the sum of and p. This last term is the pore
pressures and acts only in the voids of the material, and the eective stresses act only on
the skeleton of the material, (Terzaghi and Peck 1967).
80
Victor Saouma
Draft
318
81 It must be noted that the pore pressures p being considered in this discussion and throughout
the remainder of this course are the steady state pore pressures; excess pore pressures resulting
from dilatant behavior in the skeleton of the material are not considered.
82
Step by step generalization of the eld equations:
Initial thermal strain is caused by a change in temperature

0ij = T ij
(3.81)
where T is the temperature (change), the coecient of thermal expansion, ij is the

Kronecker delta,
= D
0
Thermal stress is simply given by ij
ijkl kl or
Net strain ij = ij 0ij or

Net eective stress

= T Dijkl kl
ij
(3.82)
ij = kl T kl
(3.83)

= Dijkl (kl 0kl )
ij
(3.84)
bi = p,i
(3.85)
0
= p ij
ij
(3.86)
Pore pressure/body force relation
Pore pressure
p the pore pressure dened using the compression positive sign convention (the minus
sign corrects the discrepancy in the sign conventions)1

Eective body force b i = bi ij,j
b i = bi p,i Dijkl T,i kl
(3.87)
t i = ti + p ni + T Dijkl ni kl
(3.88)
Eective surface traction:
83
The eective stress principle for the case of combined thermo- and poro-elasticity is

0
ij = ij
ij
+ ij

(3.89)
ij
and in the absence of both initial strains and

ij
Clearly, in the absence of initial stresses ij =

stresses ij = ij .
1
Pore pressures are typically dened using the sign convention for soil mechanics in which compression is
positive, but in the sign convention for standard solid mechanics tension is considered to be positive.
Victor Saouma
Draft
3.3 Scalar Field: Diusion Equation
319
84 The general stress-strain relationship obtained by substituting the constutive law into the
eective stress principle is
0
ij = Dijkl (kl 0kl ) + ij
(3.90)
85 When thermal strains and pore pressures are considered in combination the constitutive law
is dened as a simple combination of Equations ?? and ??
ij = Dijkl (kl T kl ) p ij
(3.91)
86 The equilibrium equation and natural boundary conditions, respectively, can be rewritten in

terms of the eective stresses ij
where
3.3

:
ij,j
+ b i = 0 Equilibrium

t : ij nj t i = 0 Natural B.C.
(3.92)
b i = bi p,i Dijkl T,i kl

t i = ti + p ni + T Dijkl ni kl
(3.93)
Scalar Field: Diusion Equation
87 Scalar eld problems are encountered in almost all branches of engineering and physics. Most
of them can be viewed as special forms of the general Helmholtz equation given by

kx
+
ky
+
kz
+ Q = c
(3.94)
x
x
y
y
z
z
t
where (x, y, z) is the eld variable to be solved.

88
Table 3.1 illustrates selected examples of the diusion equation.

Equation
Heat Flow
Fluid ow through
porous media
Diusion
Saint-Venant
Torsion
Temperature
T
Piezometric
head h
ion
concentration
Prandtls
stress function
div(Dr) + Q = 0
D
Q
Thermal
Heat
conductivity supply/sink
Permeability Fluid
Coecients
supply
Permeability Ion
coecients
supply
1
Rate
of
twist 2
q
Heat
ux
Volume
ux
Ion
ux
Constitutive law
Fourrier
q = DrT
Darcy
q = Dr
Fick
q = Dr
Hooke
Table 3.1: Selected Examples of Diusion Problems
Victor Saouma
Draft
320
3.3.1
89
Heat Transfer
There are three fundamental modes of heat transfer:
Conduction: takes place when a temperature gradient exists within a material and is governed
by Fouriers Law, Fig. 3.5 on q :
Figure 3.5: Flux vector

T
x
T
= ky
y
qx = kx
(3.95)
qy
(3.96)
where T = T (x, y) is the temperature eld in the medium, qx and qy are the componenets
of the heat ux (W/m2 or Btu/h-ft2 ), k is the thermal conductivity (W/m.o C or Btu/hT
ft-o F) and T
x , y are the temperature gradients along the x and y respectively. Note
that heat ows from hot to cool zones, hence the negative sign.
Convection: heat transfer takes place when a material is exposed to a moving uid which is
at dierent temperature. It is governed by the Newtons Law of Cooling
q = h(T T ) on c
(3.97)
where q is the convective heat ux, h is the convection heat transfer coecient or lm
coecient (W/m2 .o C or Btu/h-ft2 .o F). It depends on various factors, such as whether
convection is natural or forced, laminar or turbulent ow, type of uid, and geometry of
the body; T and T are the surface and uid temperature, respectively. This mode is
considered as part of the boundary condition.
Radiation: is the energy transferred between two separated bodies at dierent temperatures
by means of electromagnetic waves. The fundamental law is the Stefan-Boltmans Law of
Thermal Radiation for black bodies in which the ux is proportional to the fourth power
of the absolute temperature, which causes the problem to be nonlinear. This mode will
not be covered.
Victor Saouma
Draft
3.3.2
321
Derivation of the Diusion Problem
Consider a solid through which there is a ow q of some quantity, mass, heat, chemical, etc...
The rate of transfer per unit area is q which is nothing else than the ux.
90
The direction of ow (i.e. ux) is in the direction of maximum potential (such as temperature, piezometric head, or ion concentration) decrease (Fourrier, Darcy, Fick...). Hence the
ux is equal to the gradient of the scalar quantity.
q
x
x
qy
= D
= D
(3.98)
q=
y
qz
91
D is a three by three (symmetric) constitutive/conductivity matrix

The conductivity can be either
Isotropic
Anisotropic
Orthotropic
1 0 0
D = k 0 1 0
0 0 1
(3.99)
kxx kxy kxz

D = kyx kyy kyz
kzx kzy kzz
(3.100)
0
kxx 0
D = 0 kyy 0
0
0 kzz
(3.101)
Note that for ow through porous media, Darcys equation is only valid for laminar ow.
3.3.2.1
92
Simple 2D Derivation
If we consider a unit thickness, 2D dierential body of dimensions dx by dy, Fig. 3.6 then
1. Rate of heat generation/sink is
I2 = Qdxdy
(3.102)
2. Heat ux across the boundary of the element is shown in Fig. ?? (note similarity with
equilibrium equation)

qy
qx
qy
qx
dx qx dx dy + qy +
dy qy dy dx =
dxdy +
dydx
qx +
I1 =
x
y
x
y
(3.103)
3. Change in stored energy is
d
.dxdy
(3.104)
dt
where we dene the specic heat c as the amount of heat required to raise a unit mass by
one degree.
I3 = c
Victor Saouma
Draft
322
q + qy dy
y
y
qx
qx +
qx
x dx
dy
qy
dx
Figure 3.6: Flux Through Sides of Dierential Element

Figure 3.7: *Flow through a surface
93 From the rst law of thermodynamics, energy produced I2 plus the net energy across the
boundary I1 must be equal to the energy absorbed I3 , thus
I1 + I2 I3 = 0
qy
d
qx
dxdy +
dydx + Qdxdy c dxdy = 0

x
y
dt

I2
94
(3.105-b)
I3
I1
3.3.2.2
(3.105-a)
Generalized Derivation
The amount of ow per unit time into an element of volume and surface is

q(n)d =
D.nd
I1 =
(3.106)
where n is the unit exterior normal to , Fig. 3.7

95
Using the divergence theorem

vnd =
Eq. 3.106 transforms into
div vd
(3.107)
div (D)d
(3.108)

I1 =
Furthermore, if the instantaneous volumetric rate of heat generation or removal at a point

x, y, z inside is Q(x, y, z, t), then the total amount of heat/ow produced per unit time is

Q(x, y, z, t)d
(3.109)
I2 =
96
Victor Saouma
Draft
323
Finally, we dene the specic heat of a solid c as the amount of heat required to raise a unit
mass by one degree. Thus if is a temperature change which occurs in a mass m over a time
t, then the corresponding amount of heat that was added must have been cm, or

cd
(3.110)
I3 =
97
where is the density, Note that another expression of I3 is t(I1 + I2 ).

98 The balance equation, or conservation law states that the energy produced I2 plus the net
energy across the boundary I1 must be equal to the energy absorbed I3 , thus
I1 + I2 I3 = 0

d = 0
div (D) + Q c
t
(3.111-a)
(3.111-b)
but since t and are both arbitrary, then
=0
t
(3.112)
(3.113)
qx qy
+
+ Q = c
x
y
t
(3.114)
div (D) + Q c
or
div (D) + Q = c
This equation can be rewritten as
1. Note the similarity between this last equation, and the equation of equilibrium
2 ux
xx xy
+
+ bx = m 2
x
y
t
xy
2 uy
yy
+
+ by = m 2
y
x
t
(3.115-a)
(3.115-b)
2. For steady state problems, the previous equation does not depend on t, and for 2D
problems, it reduces to

kx
+
ky
+Q=0
(3.116)
x
x
y
y
3. For steady state isotropic problems,
Q
2 2 2
+ 2 + 2 =
2
x
y
z
k
(3.117)
which is Poissons equation in 3D.
Victor Saouma
Draft
324
4. If the heat input Q = 0, then the previous equation reduces to

2 2 2
+ 2 + 2 =0
x2
y
z
(3.118)
which is an Elliptic (or Laplace) equation. Solutions of Laplace equations are termed
harmonic functions (right hand side is zero) which is why Eq. 3.116 is refered to as the
quasi-harmonic equation.
5. If the function depends only on x and t, then we obtain

=
kx
+Q
c
t
x
x
(3.119)
which is a parabolic (or Heat) equation.

3.3.2.3
Boundary Conditions
The boundary conditions, for thermal problems, are mainly of three kinds:
Essential: Temperature prespecied on T
Natural: Flux
Specied Flux prescribed on q , qn = cst
Convection Flux prescribed on c , qc = h(T T ). Note this type of boundary condition is analogous to the one in structural mechanics where we have an inclined
support on rollers.
similar boundary conditions can be written for uid ow.
3.4
99
Summary and Tonti Diagrams
The analogy between scalar and vector problems is shown in Table 3.2.
To graphically illustrate the inter-relationship between eld equations and variables, Tonti
diagrams, Fig. 3.8 have been used.
100
Specified
Primary Variable
Essential
B.C.
(EBC)
Primary Variable
Kinematic
Equation
(KE)
Intermediate
Variable
Field
Equations
Constitutive
Equations
(CE)
Source Function
Balance
Equations
(BE)
Flux
Variable
Natural
B.C.
(NBC)
Specified
Flux Variable
Figure 3.8: Components of Tontis Diagram, (Felippa 2000)
Victor Saouma
Draft
3.4 Summary and Tonti Diagrams
325
Essential B.C.
ui : u
Body Forces
State variable
bi
ui
Balance
Kinematics
LT + b = 0
" = Lu
Flux variable
ij
Constitutive Rel.
ij = Dijkl kl
Interm. variable
ij
Natural B.C.
t i : t
Heat Supply
Essential B.C.
T : T
Balance
r.q
State variable
= Q
Flux variable
q
Constitutive Rel.
q = Dg
Interm. variable
g=rT
Natural B.C.
qn : q c
Figure 3.9: Fundamental Equations of Solid Mechanics and Heat Flow
Victor Saouma
Draft
326
Scalar
Conduction
T
g
q
div q + Q = 0
g = T
q = Dg

Fourrier
Vector
Solid Mechanics
Variables
u
Field Equations
LT + b = 0
= Lu
=D
State
Intermediate
Flux
Balance
kinematic
Constitutive
Hooke
Boundary Conditions
T on T
u on u
Essential BC
qn = qT n on q c
t = n on t
Natural BC (Flux)
Table 3.2: Comparison of Scalar and Vector Field Problems
101 However, we here expand on the standard Tonti diagram to graphically show the fundamental equations of solid mechanics and heat ow in Fig. 3.9.
Similarly, many other physical problems can be solved by the nite element method, Table
3.3.
102
Victor Saouma
Victor Saouma
Magnetostatic
Flow through
porous media
Electrostatics
Fluid ow
Heat transfer
Electric potential
Magnetic
potential
Conservation of electric ux
Conservation of magnetic potential
Magnetic ux
Electric ux
Flow rate
Shear stress
Magnetic permeability
Permitivity
Permeability
Thermal conductivity
Viscosity
Material Constants
Youngs
Modulus,
Poissons
ratio
Electric conductivity
Shear
Current
Charge
Fluid sources
External electric charge

-2 angle of
twist
Internal or external heat
Body forces
Body force or
traction
Source
Table 3.3: Classication of various Physical Problems, (Kardestuncer 1987)
Hydraulic head
Velocity
Heat ux
Rate of twist
Electric ux
Stress or strain
Displacement
or forces
Voltage or intensity
Stress or warping function
Temperature
Flux
State Vriable
Conservation of momentum
Conservation of mass
Equilibrium of currents
Conservation of potential energy
Conservation of energy
Electric
work
Torsion
net-
Conservation Principle
Equilibrium of forces
Physical
Problem
Deformation of
an elastic body
Maxwells law
Coulombs law
Darcys law
Stokes law
Fourriers law
Hookes law
Kirchhos law
Constitutive
Equation
Hookes law
Draft
3.4 Summary and Tonti Diagrams
327
Draft
328
Victor Saouma
Draft
Chapter 4
MESH GENERATION
Requires further editing
4.1
Introduction
1 Finite element mesh generation is now an integral part of a nite element analysis. With the
increased computational capabilities, increasingly more complex structures are being analysed.
Those structures must be discretized.
2 The task is one of developing a mathematical model (discretization or tessalation) of a continuum model. This is not only necessary in nite elment analysis, but in computer graphics/rendering also.
3 In computer graphics, we focus on the boundary representation, and assign colors and shades
on the basis of light source and outward normal direction of the polygon.
Hence, in the most general case, meshing can be dened as the process of breaking up a physical domain into smaller sub-domains (elements) in order to facilitate the numerical solution of a
partial dierential equation. Surface domains may be subdivided into triangle or quadrilateral
shapes, while volumes may be subdivided primarily into tetrahedra or hexahedra shapes. The
shape and distribution of the elements is ideally dened by automatic meshing algorithms.
4
1. Point placement, followed by triangularization (discussed below).

2. Sub-domain removal. Elements are gradually removed from the domain, one ata time,
until the whole domain is decomposed int nite elements.
3. Recursive subdivision. The domain is broken into simpler parts until the individual parts
are single elements or simple regions, that can be meshed directly, for instance by the
conformal mapping algorithm.
4. Hierarchical decomposition. The basic principle of a quadtree (or hierarchical decomposition) is to cover a planar region of interest by a square, then recursively partition squares
into smaller squares until each square contains a suitably uniform subset of the input.
Draft
42
4.2
MESH GENERATION
Triangulation
5 The concept of Voronoi diagrams rst appeared in works of Descartes as early as 1644.
Descartes used Voronoi-like diagrams to show the disposition of matter in the solar system and
its environs.
6 The rst man who studied the Voronoi diagram as a concept was a German mathematician
G. L. Dirichlet. He studied the two- and three dimensional case and that is why this concept
is also known as Dirichlet tessellation. However it is much better known as a Voronoi diagram
because another German mathematician M. G. Voronoi in 1908 studied the concept and dened
it for a more general n-dimensional case.
7 Very soon after it was dened by Voronoi it was developed independently in other areas like
meteorology and crystalography. Thiessen developed it in meteorology in 1911 as an aid to
computing more accurate estimates of regional rainfall averages. In the eld of crystalography
German researchers dominated and Niggli in 1927 introduced the term Wirkungsbereich (area
of inuence) as a reference to a Voronoi diagram.
During the years this concept kept being rediscovered over and over again in dierent elds
of science and today it is extensively used in about 15 dierent elds of sciences. Some of them
being mathematics, computer science, biology, cartography, physiology and many others.
8
Vb
Va
Vc
1
4f
5
6
6f
Voronoi
Delaunay
Figure 4.1: Voronoi and Delaunay Tessellation
4.2.1
Voronoi Polygon
9 Given a nite set of poits in the plane, the idea is to assign to each point a region of inuence
in such a way that the regions decompose the plane, (Voronoi 1907).
Victor Saouma
Draft
43
To describe a specic way to do that, let S be a subset of R 2 (S R 2 ). We dene the

Voronoi region of p S as the set of points x R that are at least as close to p as to any
other points in S:
*
+
Vp = x R2 | ||x p|| x 1 , q S
(4.1)
10
Each point x R 2 has at least one nearest point in S, so it lies in at least one Voronoi region.
Two Voronoi regions lie on opposite sides of the perpendicular bisector separating the two
generating points.
4.2.2
Delaunay Triangulation
The dual of the Voronoi diagram is obtained by drawing straight Delaunay edges connecting
points p, q S if and only if their Voronoi regions intersect along a common line segment. Thus
in general, the Delaunay edges decompose the convex hull of S into triangular regions which
are referred to as Delaunay triangles, (Delaunay 1934).
11
12 Using Eulers relation, it can be shown that a planar graph with n 3 vertices has at most
3n 6 edges and at most 2n 4 faces. THe same bounds hold for the number of Delaunay
edges and triangles.
Each Voronoi vertex u = Va Vb Vc is the center of a circle with radius = u a =

u b = u c . The circle is empty because it encloses no point of S.
13
14 Additional detailed information on Voronoi tesselation, (Okabe, Boots, Sugihara and Chiu
2000) is an excellent reference.
4.2.3
MATLAB Code
rand(state,4);
x = rand(1,3);
y = rand(1,3);
TRI = delaunay(x,y);
subplot(1,2,1),...
trimesh(TRI,x,y,zeros(size(x))); view(2),...
axis([0 1 0 1]); hold on;
plot(x,y,o);
set(gca,box,on);
[vx, vy] = voronoi(x,y,TRI);
subplot(1,2,2),...
plot(x,y,r+,vx,vy,b-),...
axis([0 1 0 1])
4.3
4.3.1
Finite Element Mesh Generation

Boundary Denition
15 In order to discretize the continuum into a nite element mesh, rst key geometrical information of the must be specied hierarchically:
Victor Saouma
Draft
44
MESH GENERATION
Vertices: with nodal coordinates, and approximate desired element size in the immediate
vicinity (thus describing the mesh density).
Edges: which connect vertices. Those can be either linear segments, polylines, or curves.
Surfaces: composed of edges, dened counterclockwise.
Volumes: (3D only) composed of surfaces.
16
Associated with surfaces (2D0, or volumes (3D) are dierent material properties.
17
Examples of the hierarchical boundary denition is shown in Fig. 4.2 and 4.3.
Figure 4.2: Control Point for a 2D Mesh
4.3.2
Interior Node Generation
18 Once the boundary has been dened, we need to insert internal nodes at a spacing which
respect the required mesh density. There are numerous techniques to insert those internal
nodes. We present one approach, FIg. 4.4
1. Decompose the region into a disjoint ensemble of subregions with equal mesh density.
2. Shrink the mesh to avoid elements near the boundary with very acute interior angles.
3. Starting with the rst zone, circumscribe it by the smallest possible rectangle.
4. Superimpose a square rectangular grid over the circumscribing rectangle.
Victor Saouma
Draft
45
Figure 4.3: Control Point for a 3D Mesh
Boundary discretization
Zone II, r2
Zone III, r3
Zone I, r1
Region to be meshed
Zonal decomposition with nodal density
Boundary shrinking by
r
Generation of internal nodes
in zone I within shrunk boundary
Disk
r
Figure 4.4: A Two Dimensional Triangularization AlgorithmControl Point for a 3D Mesh
Victor Saouma
Draft
46
MESH GENERATION
5. Use a random number generator to randomly generate one interior node in each square.
A disk of radius r centered at each node is used to test that no other surrounding nodes
are enclosed in the disk. If so, the node in question is regenerated.
19
An alternative approach consists in, (Cervenka,

J. 1994)
1. Generating a triangularization compatible with the initial nodes.
2. Check lengths of the edges. If an edge does not satisfy the prescribed size r, a new node
is inserted in the center of the edge. The prescribed size is interpolated between those of
the vertices at each end of the edge.
3. Repeat this operation until convergence.
4. Smoothen the elemnts to assure appropriate aspect ratios.
4.3.3
20
Final Triangularization
With boundary and interior nodes generated:

1. Determine the Voronoi polygons
2. Perform a Delaunay triangularization
3. Smoothen the mesh to ensure that all generated elements have a satisfactory aspect ratio.
21 It should be noted that recent algorithms, which can generate quadrilateral elements out of
the Delaunay triangularization have recently emerged.
Victor Saouma
Draft
Chapter 5
VARIATIONAL and
RAYLEIGH-RITZ METHODS
5.1
Multield Variational Principles
A Multield variational principle is one that has more than one master eld (or state
variable), that is more than one unknown eld is subject to independent variations.
1
2 In linear elastostatics, we can have displacement, u, strains , or stress as potential candidates for master elds. Hence seven combinations are possible, (Felippa 2000), Table 5.1.
3 In this course, we shall focus on only the Total Potential Energy, and the Hu-Washizu variational principles.
Y
Y
Y
Y
Y
Y
Y
Y
Name
Single Field
Total Potential Energy
Total Complementary Potential Energy
No name
Two Fields
Hellinger-Reissner
de Veubeke
No name
Three Fields
Hu-Washizu
Table 5.1: Functionals in Linear Elasticity
Draft
52
5.2
Total Potential Energy Principle
5.2.1
4
VARIATIONAL and RAYLEIGH-RITZ METHODS
Static; Euler
The expression for the total potential energy (TPE) is given by
We
U
def
def
def
U We

uT bd +
uT td + uP
t

1
T
T
D d
D 0 d +
T 0 d
2
(5.1-a)
(5.1-b)
(5.1-c)
the functional for the general form of the potential energy variational principle is obtained
1
=
2

D d
T
D 0 d +

T

0 d

T
u bd
uT td uP
t

T
(5.2)
We
5 A variational statement is obtained by taking the rst variation of and setting this scalar
quantity equal to zero.
The variational statement for the general form of the potential energy functional (i.e. Equation 5.2) is

T
T
T
T
D d
D 0 d +
0 d
u bd
uT td = 0
(5.3)
=
6
which is the Principle of Virtual Work.

7 We rewrite the strain-displacement relations in terms of a linear dierential operator L
(Eq. 3.41)
= Lu
(5.4)
where L is a linear dierential operator and u is the

0
x
y
0
0
L=
y x

z 0
0 z
displacement vector dened in Eq. 3.42:
0
0

z
(5.5)
8 Since the dierential operator L is linear, the variation of the strains can be expressed in
terms of the variation of the displacements u
= (Lu) = Lu
9
(5.6)
We consider two forms of the variational statement
Victor Saouma
Draft
5.2 Total Potential Energy Principle
53
1. In terms of strains: which is more suitable for the derivation of the corresponding Euler
equations,

(Lu) D d
T
(Lu) D 0 d +

U

uT bd
uT td = 0

t

(Lu)T 0 d

(5.7)
We
2. In terms of displacements: (using Equation 5.7) which is more suitable for the subsequent discretization.

(Lu) D(Lu)d
T
=
uT bd
(Lu) D 0 d +
(Lu)T 0 d

We
uT td = 0

(5.8)
10 To obtain the Euler equations for the general form of the potential energy variational principle
the volume integrals dening the virtual strain energy U in Equation 5.7 must be integrated by
parts in order to convert the variation of the strains (Lu) into a variation of the displacements
u.
11
Integration by parts of these integrals using Greens theorem (Eq. 1.46)

,

S
R
Sd + RSnx d
R d =
x
(5.9)
yields
,
(Lu) D d =

u G(D )d
T
uT LT (D )d
(Lu)T D 0 d =
uT G(D 0 )d uT LT (D 0 )d
,

T
T
(Lu) 0 d =
u G 0 d uT LT 0 d
(5.10-a)

(5.10-b)
(5.10-c)
where G is a transformation matrix containing the direction cosines for a unit normal vector
such that the surface tractions t are dened as t = G and the surface integrals are over the
entire surface of the body .
12
Substituting Equation 5.10-a into Equation 5.7, the variational statement becomes

= uT {LT [D( 0 ) + 0 ] +b}d

Victor Saouma
Draft
54
^
t
û
Figure 5.1: Tonti Diagram for the Total Potential Energy, (Cervenka,
J. 1994)

+
uT {G [D( 0 ) + 0 ] t}d = 0

(5.11)
13 Since u is arbitrary the expressions in the integrands within the braces must both be equal
to zero for to be equal to zero. Recognizing that the stress-strain relationship appears in
both the volume and surface integrals, the Euler equations are
(BE): Equilibrium
(NBC): Natural B.C.
LT + b = 0 on
G t = 0 on t
(5.12)
(5.13)
where the rst Euler equation is the equilibrium equation and the second Euler equation denes
the natural boundary conditions. The natural boundary conditions are dened on t rather
than because both the applied surface tractions t and the matrix-vector product G are
identically zero outside t .
In general, only certain forms of dierential equations are Euler equations of a variational
functional.
14
15 For the elastostatic problem, it is possible to start from the Euler equations, and then derive
the total potential energy functional by performing the operations just presented in reverse
order.
16 The Tonti diagram for the TPE is shown in Fig. 5.2. In this diagram, strong connections are
shown by solid lines, weak connections by spring-like symbols, boxes with solid lines denote the
primary unknown eld, variables inside dashed boxes are internally derived elds, and shaded
boxes indicate prescribed elds.
5.2.2
Dynamic; Euler/Lagrange
In dynamics, we dene the Lagrangian function as the dierence between the kinetic and
the potential energies
LK
(5.14)
17
and Hamiltons principle states that The motion of a particle acted on by conservative forces
between two arbitrary instants of time t1 and t2 is such that the line integral over the Lagrangian
function is an extremum for the path motion.
t2
Ldt
(5.15)
I
t1
Victor Saouma
Draft
55
is an extremum.
If the path can be dened in terms of generalized coordinates qi (i = 1, 2, 3), then it can be
shown that
t2
L(q1 , q2 , q3 , q1 , q2 , q3 )dt = 0
(5.16-a)
I =
18
t1
t2
3
t1 i=1
qi
d
dt
L
qi

qi dt
(5.16-b)
If all qi are linearly independent (i.e. no constraints among them), the variations qi are
independent of t, except qi = 0 at t1 and t2 . Therefore, the coecients of q1 , q2 , and q3
vanish separately yielding the Lagrangian equations of motion, also known as the EulerLagrange equations.

d L
L
(5.17)
= 0,
i = 1, 2, 3
qi dt qi
19
Hamiltons principle can be expressed as:

t2

(K )dt +
H =

H
t1
t2
(K )dt +
t1
t2
Wnc dt
(5.18)
t1
t2
Wnc dt = 0
(5.19)
t1
where K is the kinetic energy, the potential energy, and Wnc the work done by nonconservative
forces acting on the system (including damping).
It should be noted that TPE is a special case of Hamiltons principle in the absence of
kinetic energy.
The Kinetic energy is given by

u u
. d
(5.20)
K=
2 t t
A conservative force is one for which the sum of the potential and kinetic energies is conserved).
Example 5-1: Hamiltons Principle

For a uniform cross section bar of length L, cross sectional area A, Youngs modulus E, and
mass density xed at one end, and connected to a rigid support at the other by a spring with
stiness k
k
A, E
L
Victor Saouma
Draft
56
1. Show that the kinetic energy and the strain energy are given by

u 2
dx
K =
t

L
k
EA u 2
d + [u(L)]2
U =
x
2
0 2

A
0 2
(5.21)
(5.22)
2. For Wnc = 0 derive an expression for the rst variation of the Hamiltons functional H
3. If we are interested in determining the periodic motion, which has the form
u(x, t) = u0 (x)eit
(5.23)
where is the frequency of natural vibration, and u0 (x) is the amplitude. Show that
#

2 $

L
du
1
k
0
A 2 (u0 )2 EA
dx [u0 (L)]2 = 0
(5.24)
dx
2
0 2
4. Show that the Euler equations of the preceding functional are

du0
d
EA
+ A 2 u0 = 0 for 0 < x < L
dx
dx

du0
+ ku0
= 0 for x = L
EA
dx
5. Rewrite Eq. 5.24 in terms of x =
x
L,
u=
u0
L,
kL
EA ,
(5.25)
(5.26)
2 L2
E
6. Show that if this problem was to be solved by the Rayleigh-Ritz method with
u = c1 x + c2 x2
(5.27)
then a nontrivial solution c1 = 0 and c2 = 0 exists if an only if

152 640 + 2400 = 0
(5.28)
7. Solve for 1 and 2 , and compare with the exact solution given by
+ tan = 0
(5.29)
Solution:
1. Kinetic energy

L
u 2
A u 2
d =
dx
K =
t
t
0 2

L

k
E u 2
EA u 2
d =
dx + [u(L)]2
U =
2
x
2
x
2
0
Victor Saouma
(5.30)
(5.31)
Draft
57
2. For Wnc = 0

H
t2
(K )dt
2 $
t2 L # 2
u
1
u
A
dx k[u(L)]2
EA
2
t
x
0
t1
(5.32)
t1
dt
(5.33)
3. If we adopt
u(x, t) = u0 (x)eit
(5.34)
where is the frequency of natural vibration, and u0 (x) is the amplitude. Then substituting, we obtain

#
$

L
k
u0 2
1
2
2
dx [u0 (L)]2 = 0
(5.35)
A [u0 (x)] EA
2
x
2
0

t2
where
exp2it dt being nonzero is factored out.
t1
4. To determine the corresponding Euler Equation

#

$

L
du0 2
1
k
2
2
A (u0 ) EA
dx [u0 (L)]2 =
2
dx
2
0

L
u0
u0
dx ku0 (L)u0 (L) = 0

A 2 u0 u0 EA
x
x
0
(5.36)
(5.37)
but
L

L

u0
u0
u0

EA
dx = EA
u0
x
x
x
0

v

v
Substituting

L
0

u0
EA
u0 dx
x
x

v
(5.38)
u

u0 u0
dx ku0 (L)u0 (L)

(5.39)
A [u0 (x)] EA
x x
0
L

L

u0
du0
d
EA
EA
u0 ku0 (L)u0 (L) = 0(5.40)
A[u0 (x)] +
=
dx
x
dx
0
0

L
or
d
dx
Victor Saouma

du0
EA
+ A 2 u0 = 0 for 0 < x < L
dx

du0
+ ku0
= 0 for x = L
EA
dx
(5.41)
(5.42)
Draft
58

2
kL
dx = Ldx, u = uL0 , and = EA
, = EL we have
2 $

L #
1
du
k
0
A 2 (u0 )2 EA
dx [u0 (L)]2
dx
2
0 2
2 $

1 #
du0
k
1 EA
(Lu)2 EA
Ldx [Lu0 (1)]2
=
2
L
dx
2
02
#

2 $
1
du0
kL2
1
EAL (u)2
[u0 (1)]2
dx
=
2
dx
2
0
#

$
1
2 1 du0 2
KL2
dx [u0 (1)]2
=
(u)
2
dx
0
#
#
$

$
1
KL2
2 1 du0 2
2
(u)

dx [u0 (1)]
=0
2
dx
0
#
$

du0 2
1 1
2
(u)
dx [u0 (1)]2 = 0
2 0
dx
2
5. Substituting x =
x
L,
(5.43)
(5.44)
(5.45)
(5.46)
(5.47)
(5.48)
6. Using the Rayleigh-Ritz method with

u = c1 x + c2 x2
(5.49)
and rewriting the previous equation in terms of u

#
2 $
du
1 1
(u)2
dx [u(1)]2 = 0
2 0
dx
2
1
"
!
(c1 x + c2 x2 )(c1 x + c2 x2 ) (c1 + 2c2 x)(c1 + 2c2 x) dx
(5.50)
(c1 + c2 )(c1 + c 2) = 0
(5.51)
Collecting the coecients of c1 and c2 and setting them to zero separately, we have
1
"
!
(5.52)
(c1 x + c2 x2 )x (c1 + 2c2 x) dx (c1 + c2 ) = 0
c1 :
0
1
"
!
(5.53)
(c1 x + c2 x2 )x2 (c1 + 2c2 x)2x dx (c1 + c2 ) = 0
c2 :
0
Integrating, reducing and compacting

2 2
2 73
1
3
1
4
A nontrivial solution exists if and only if

2
3

2
4
1
4
1
5

c1
c2

2 4
7
=0
3 5

=
0
0

(5.54)
(5.55)
or
152 640 + 2400 = 0
Victor Saouma
(5.56)
Draft
5.3 General Hu-Washizu Variational Principle
59
This quadratic equation has two roots

1 = 4.1545
2 = 38.512
(5.57)
which correspond to

2.038
1 =
L

E
6.206
2 =
L
(5.58)
7. The exact solution is given by given by

+ tan = 0
(5.59)
whose rst two roots are

2.02875
1 =
L
5.3

E
4.91318
2 =
L
(5.60)
General Hu-Washizu Variational Principle
20 The Hu-Washizu (HW) variational principle is a three-eld variational principle in which

the displacements, strains, and stresses are treated as independent elds (as opposed
to only the displacement in the total potential energy principle).
21 Naturally, the two additional eld variables, with respect to the TPE variational principle,
appear not only in the functional, but also in the discretized system of equations. Consequently,
for a domain with a given discretization the discrete system of equations derived from the HW
variational principle will be much larger than the discrete system of equations derived from the
TPE variational principle.
22 With the increased number of equations, signicant improvements in accuracy can be observed for the solution obtained from the discrete form of the HW variational principle compared
to the solution obtained from the discrete form of the TPE variational principle for the same
discretization. This means that coarse discretizations can be used with the discrete form of the
HW variational principle to obtain the same degree of accuracy that would be observed with
much ner discretizations using the TPE variational principle.
23 The functional for the HW variational principle is derived from the functional for the TPE
variational principle by imposing the strain-displacement equation as a nite subsidiary condition using the method of Lagrange multipliers.
24
The nite subsidiary condition or constraint is written in residual form as

Lu = 0
(5.61)
and enforced in an average sense over the entire body .
Victor Saouma
Draft
510
By imposing the strain-displacement equation as a constraint C 0 continuous strains and

stresses are obtained in the discrete form of the varitional statements, as opposed to the discontiuous strains and stresses obtained in the discrete form of the varitional statement for the
TPE variational principle.
25
26
The constrained functional is written as
HW = TPE +
T (L u ) d
(5.62)
Where is the Lagrange multiplier and to be consistent with the integrals in the TPE functional
(i.e. Equation 5.2) the Lagrange multiplier must have the units of stress. Since this is the case,
will be used for the Lagrange multiplier instead of the more typical such that the physical
meaning of the Lagrange multiplier is more apparent.
27
The functional for the HW variational principle thus becomes

1
T
T
D d
D 0 d +
T 0 d
HW =
2

U

uT b d
uT t d + T (L u ) d

t

We
Constraint
(5.63)
28 A variational statement is obtained by taking the rst variation of the functional and setting
this scalar quantity equal to zero.
The rst variation of the HW functional, with terms arranged according to which eld
variable is varied, is

T
T
(L u) d
u b d
uT t d
HW =

t

(5.64)
T D d
T D 0 d +
T 0 d
T d
+

T (L u ) d = 0
+
29
Note that the 4th and 7th term were added and cancell each others, and that we are not using
Eq. 5.6 in this formulation.
Since u, , and are independent eld variables, terms involving u, , and must add up
to zero individually and are, therefore grouped together to form three separate variational statements (analogous to the method of separation of variables in the solution of partial dierential
equations)

(Lu)T d uT b d
uT t d = 0 (5.65)
t

T [D( 0 ) + 0 ] d = 0 (5.66)

T (L u ) d = 0 (5.67)
30
Victor Saouma
Draft
5.3 General Hu-Washizu Variational Principle
511
31 To obtain the corresponding Euler equations for the general form of the HW variational
principle the volume integral in Equation 5.65 containing the variation of the strains (Lu)
dened in terms of the displacements u must be integrated by parts using Greens theorem in
order to obtain a form of the variational statement in terms of the variation of the displacements
u.
32
Integration by parts (Eq. 1.39) of this integral yields

,

T
T
(L u) d =
u G d
uT LT d
(5.68)
where G is a transformation matrix containing the direction cosines for a unit normal vector
such that the surface tractions t are dened as t = G and the surface integral is over the
entire surface of the body .
33
Substituting Equation 5.68 into Equation 5.65, the rst variational statement becomes

T
T
uT (G t) d = 0
(5.69)
u (L + b) d +
Since u is arbitrary the expressions in the integrands within the parentheses must both be
equal to zero for the sum of the integrals to be equal to zero.
34 Likewise, and are also arbitrary and the expressions within the braces in the second
variational statement (i.e. Equation 5.66) and within the parentheses in the third variational
statement (i.e. Equation 5.67) must both be equal to zero for the integral to be equal to zero.
35
The Euler equations for the HW functional are

(BE): Equilibrium
(CE): Stress-Strain
(KE): Strain-Displacement
(NBC): Natural B.C.
LT + b = 0
on
(5.70)
D ( 0 ) + 0 = 0 on
(5.71)
Lu = 0
G t = 0
on
(5.72)
on t
(5.73)
where the rst Euler equation is the equilibrium equation; the second Euler equation is the
stress-strain relationship; the third Euler equation is the strain-displacement equation; and the
fourth Euler equation denes the natural boundary conditions.
The natural boundary conditions are dened on t rather than because both the applied
surface tractions t and the matrix-vector product G are identically zero outside t . Starting
from the Euler equations, it is possible to derive the HW functional by performing the operations
just presented in reverse order.
36
37 This last set of four Euler equations, should be compared with the two (Eq. 5.12 and 5.13)
obtained from the original TPE. The additional two equations bring into play stress-strain and
strain displacement. Also, whereas the original formulation (Eq. 5.12 and 5.13) was in terms
of the displacement only (u), the Hu-Washizu formulation is in terms of three independent
variables (u, and ), Table 5.2.
38
The Tonti diagram for the HW is shown in Fig. ??.
Victor Saouma
Draft
512
Equilibrium
Stress-Strain
Strain-Displacement
Natural B.C.
Euler Equations
LT + b = 0
D ( 0 ) + 0 = 0
Lu = 0
G t = 0
Variables
Displacement
Strain
Stress
TPE
HW
Y
N
N
Y
Y
Y
Y
Y
Y
N
N
Y
Y
Y
Table 5.2: Comparison Between Total Potential Energy and Hu-Washizu Formulations
^
u
^
t
Figure 5.2: Tonti Diagram for Hu-Washizu, (Cervenka,

J. 1994)
Victor Saouma
Draft
5.4 Rayleigh Ritz
5.4
513
Rayleigh Ritz
39 In the principle of virtual displacement (or minimum total potential energy), the Euler equations are the equilibrium equations. The Euler equations are usually in the form of dierential
equations that are not always solvable by exact methods of solutions.
Indeed we do have available a number of approximate methods such as the nite dierence
but an alternative approach is to bypass entirely the Euler equations and start directly from a
variational statement of the problem to the solution of the Euler equation.
40
41 Such a direct method was rst proposed by Lord Rayleigh, and then independently generalized by Ritz.
Hence, this section will briey review the Rayleigh-Ritz method (which solves the variational
problem, but which does not require a topological discretization (i.e. nite element). The
similarity between those two approaches should by now be apparent. In both cases we are
dealing with an approximate method.
42
43 Application of the principle of total potential energy (or virtual displacement) requires an
assumed displacement eld. This displacement eld can be approximated by interpolation
functions written in terms of:
1. Unknown polynomial coecients most appropriate for continuous systems, and the
Rayleigh-Ritz method
(5.74)
y = a1 + a2 x + a3 x2 + a4 x3
A major drawback of this approach, is that the coecients have no physical meaning.
2. Unknown nodal deformations most appropriate for discrete systems and Potential Energy based formulations
y = u = N1 u1 + N2 u2 + . . . + Nn un
(5.75)
44 This chapter will focus on the rst type of approximation, whereas subsequent ones will
adopt an approximation based on the nodal deformations.
45
In general, we may adopt an approximate displacement eld by a function

u1
u2
u3
i=1
n
i=1
n
c1i 1i + 10
(5.76-a)
c2i 2i + 20
(5.76-b)
c3i 3i + 30
(5.76-c)
i=1
where cji denote undetermined parameters, and are appropriate functions of positions.
46
should satisfy three conditions

1. Be continous.
Victor Saouma
Draft
514
2. Must be admissible, i.e. satisfy the essential boundary conditions (the natural boundary
conditions are included already in the variational statement. However, if also satisfy
them, then better results are achieved).
3. Must be independent and complete (which means that the exact displacement and their
derivatives that appear in can be arbitrary matched if enough terms are used. Furthermore, lowest order terms must also be included).
In general is a polynomial or trigonometric function.
We determine the parameters cji by requiring that the principle of virtual work for arbitrary
variations cji . or
47
(u1 , u2 , u3 ) =
n
i=1
c1
c1i i
+ 2 c2i + 3 c3i
ci
ci

=0
(5.77)
for arbitrary and independent variations of c1i , c2i , and c3i , thus it follows that
cji
=0
i = 1, 2, , n; j = 1, 2, 3
(5.78)
Thus we obtain a total of 3n linearly independent simultaneous equations. From these displacements, we can then determine strains and stresses (or internal forces). Hence we have replaced
a problem with an innite number of d.o.f by one with a nite number.
48
Some general observations

1. cji can either be a set of coecients with no physical meanings, or variables associated
with nodal generalized displacements (such as deection or displacement).
2. If the coordinate functions satisfy the above requirements, then the solution converges
to the exact one if n increases.
3. For increasing values of n, the previously computed coecients remain unchanged.
4. Since the strains are computed from the approximate displacements, strains and stresses
are generally less accurate than the displacements.
5. The equilibrium equations of the problem are satised only in the energy sense = 0
and not in the dierential equation sense (i.e. in the weak form but not in the strong one).
Therefore the displacements obtained from the approximation generaly do not satisfy the
equations of equilibrium.
6. Since the continuous system is approximated by a nite number of coordinates (or d.o.f.),
then the approximate system is stier than the actual one, and the displacements obtained
from the Ritz method converge to the exact ones from below.
7. If the functional is quadratic, then
Kc + f = 0
c
(5.79)
and K will always be symmetric

Victor Saouma
Draft
5.4 Rayleigh Ritz
515
y
w
x
111
000
111
000
L
Figure 5.3: Uniformly Loaded Simply Supported Beam Analysed by the Rayleigh-Ritz Method
5.4.1
Examples
Example 5-2: Uniformly Loaded Simply Supported Beam; Polynomial Approximation

For the uniformly loaded beam shown in Fig. 5.3
let us assume a solution given by the following innite series:
v = a1 x(L x) + a2 x2 (L x)2 + . . .
(5.80)
for this particular solution, let us retain only the rst term:
v = a1 x(L x)
(5.81)
We observe that:
1. The essential B.C. are immediately satised at both x = 0 and x = L.
= 0 (If we had left v in terms of a1 and a2 we

2. We can keep v in terms of a1 and take a
1
should then take both a1 = 0, and a2 = 0 ).
3. Or we can solve for a1 in terms of vmax (@x =

= U W =
L
o
Recalling that:
M
EIz
Victor Saouma
L
2)
M2
dx
2EIz
and take

vmax
= 0.
wv(x)dx
(5.82)
2
= d v2 , the above simplies to:
dx
$

2
L#
EIz d2 v
wv(x) dx
=
2
dx2
0

L
EIz
2
(2a1 ) a1 wx(L x) dx
=
2
0
L3
L3
EIz 2
4a1 L a1 w
+ a1 w
=
2
2
3
3
a
wL
1
= 2a21 EIz L
6
(5.83)
(5.84)
Draft
516
If we now take
a1
= 0, we would obtain:
4a1 EIz l
wL3
6
= 0
wL2
24EIz
a1 =
(5.85)
Having solved the displacement eld in terms of a1 , we now determine vmax at

wL4
v =
24EI
z
x
x2
2
L L
L
2:
a1
wL4
96EIz
(5.86)
4
wL
exact = 5 wL =
This is to be compared with the exact value of vmax
384 EIz
76.8EIz which constitutes
17% error.
wL2
w
& a2 = 24EI
Note: If two terms were retained, then we would have obtained: a1 = 24EI
z
z
exact . (Why?)
and vmax would be equal to vmax
Example 5-3: Heat Conduction; (Bathe 1996)

Considering the slab in Example 2-2, and assuming
(t) = 1 (t) + 2 (t)x + 3 (t)x2
(5.87)
where 1 (t), 2 (t), and 3 (t) are the undetermined parameters. Use the Rayleigh-Ritz method
procedure to generate the governing heat transfer equilibrium equations, using the following
functional
2
L
L
1
k
dx
q B dx |x=0 q0
(5.88)
=
2
x
0
0
with the essential boundary condition |x=L = i
Solution:
1. Substituting
L
L
"
!
"
1 !
k (2 )2 + 42 3 x + 4(3 )2 x2 dx
1 + 2 x + 3 x2 q B dx 1 q0 (5.89)
=
0 2
0
2. Invoking
= 0;
1
we obtain
0 0
k 0 L
0 L2
Victor Saouma
= 0;
2
=0
3
L B
0
1
0 q dx + q0
L
B
2
=
L2
0 xq dx

4 3
L 2 B
3
3L
0 x q dx
(5.90)
(5.91)
Draft
5.4 Rayleigh Ritz
517
3. In this problem q0 varies with time, hence heat capacity must be accounted for
q B = c
4. Substituting
0 0
k 0 L
0 L2
0
L
1
1 2
2
+ c 2 L
2
4 3
1 3
L
3
3
3L
1 2
2L
1 3
3L
1 4
4L
(5.92)
1 3
3L
1 4
4L
1 5
5L

1 q0
0
=
2

0
3
(5.93)
5. The nal equilibrium equation are now obtained by imposingon the last equation the
condition that |x=L = i , i.e.
1 (t) + 2 (t)L + 3 (t)L2 = i
(5.94)
which can be achieved by expressing 1 in terms of 2 , 3 and i
Victor Saouma
Draft
518
Victor Saouma
Draft
Chapter 6
INTERPOLATION FUNCTIONS;
NATURAL COORDINATE
SYSTEMS
6.1
Introduction
In the Rayleigh Ritz method we solved the variational problem using a functional approximation for the displacement eld (Chapter ??). This powerful method has its limitation in terms
of the complexity of solvable problems.
1
As stated above, an alternate approach consists in adopting an approximate displacement

eld in terms of the nodal displacements via interpolation functions (or shape functions).
2
6.2
Cartesian Coordinate System
3 For an element (nite or otherwise), we can write an expression for the generalized displacement (translation/rotation), u at any point in terms of all its nodal ones, u.
u=
Ni (X)i = N(x){u}
(6.1)
i=1
where:
1. ui is the (generalized) nodal displacement corresponding to d.o.f i
2. Ni is an interpolation function, or shape function which has the following characteristics:
(a) Ni = 1 at ui
(b) Ni = 0 at uj where i = j.
3. N can be derived on the bases of:
(a) Assumed deformation state dened in terms of polynomial series.
(b) Interpolation function (Lagrangian or Hermitian).
Draft
62
INTERPOLATION FUNCTIONS; NATURAL COORDINATE SYSTEMS
Figure 6.1: Axial Finite Element

4
We shall distinguish between two classes of problems:
C 0 where only displacement continuity must be ensured across elements (truss, torsion, plane
stress/strain, 3D Elasticity).
C 1 where we must ensure continuity of both displacements and their derivatives (such as beams,
plates, and shells).
6.2.1
6.2.1.1
5
C0
Truss element
With reference to Fig. 6.1 we start with:

u = N1 u1 + N2 u2
(6.2)
x = N1 x1 + N2 x2
(6.3)
Since we have 2 d.o.fs, we will assume a linear deformation state

u = a1 x + a2
(6.4)
where u can be either u or , and the B.C.s are given by: u = u1 at x = 0, and u = u2 at
x = L. Thus we have:
u1 = a2
(6.5)
u2 = a1 L + a2
(6.6)
Solving for a1 and a2 in terms of u1 and u2 we obtain:

u2 u1
L
L
= u1
a1 =
(6.7)
a2
(6.8)
Substituting and rearranging those expressions into Eq. 6.4 we obtain

u2 u1
)x + u1
L
L
x
x
u2
= (1 ) u1 +
L
L

u = (
N1
Victor Saouma
(6.9)
(6.10)
N2
Draft
6.2 Cartesian Coordinate System
63
or:
N1 = 1
N2 = Lx
6.2.1.2
9
x
L
(6.11)
Generalization
The previous derivation can be generalized by writing:

a1
u = a1 x + a2 = x 1
a2

[p]
(6.12)
{a}
where [p] corresponds to the polynomial approximation, and {a} is the coecient vector.
10
We next apply the boundary conditions:

0 1
u1
a1
=
u2
a2
L 1

{u}
{a}
[L]
following inversion of [L], this leads to

1 1 1
u1
a1
=
L 0
a2
u2
L

{a}
11
(6.13)
[L]1
(6.14)
{u}
Substituting this last equation into Eq. 6.12, we obtain:

u1
x
x
u = (1 L ) L

u2

1
[p][L]
{u}

(6.15)
[N]
12
Hence, the shape functions [N] can be directly obtained from

[N] = [p][L]1
(6.16)
Note that in some cases L1 is not always possible to obtain, and that in others there may
be considerable algebraic diculties for arbitrary geometries. Hence, we shall introduce later
on Lagrangian and Hermitian interpolation functions.
13
6.2.1.3
14
Constant Strain Triangle Element
Next we consider a triangular element, Fig. 6.2 with bi-linear displacement eld (in both x
Victor Saouma
Draft
64
Figure 6.2: Constant Strain Triangle Element

and y):
u = a1 + a2 x + a3 y
v = a4 + a5 x + a6 y
u = 1 x y

[p]
(6.17)
a1
a2
a3

(6.18)
(6.19)
{a}
15 As before, we rst seek the shape functions, and hence we apply the boundary conditions at
the nodes for the u displacements rst:

1 0 0
a1
u1
u
= 1 x2 0
a
(6.20)
2
2
u3
a3
1 x3 y 3

{u}
16
[L]
{a}
We then multiply the inverse of [L] in Eq. 6.20 by [p] and obtain:
u = N1 u1 + N2 u2 + N3 u3
(6.21)
where
N1 =
N2 =
N3 =
1
(x2 y3 xy3 x2 y + x3 y)
x2 y3
1
(xy3 x3 y)
x2 y3
y
y3
(6.22)
We observe that each of the three shape functions is equal to 1 at the corresponding node, and
equal to 0 at the other two.
17
The same shape functions can be derived for v:

v = N1 v 1 + N2 v 2 + N3 v 3
Victor Saouma
(6.23)
Draft

18
65
Hence, the displacement eld will be given by:
u
v

=
N1 0 N2 0 N3
0 N1 0 N2 0
u1
v1
0
u2
N3
v2
u
3
v3
(6.24)
The element is refereed to as Constant Strain Triangle (CST) because it has a linear displacement eld, and hence a constant strain.
19
6.2.1.4
20
Further Generalization: Lagrangian Interpolation Functions
In our earlier approach, the shape functions were obtained by:

1. Assumption of a polynomial function: u = p{a}
2. Application of the boundary conditions {u} = [L]{a}
3. Inversion of [L]
4. And nally [N] = [p][L]1
By following these operations, we have in eect dened the Lagrangian Interpolation Functions for problems with C 0 interelement continuity (i.e continuity of displacement only).
21
22 The Lagrangian interpolation denes the coecients ([N] in our case) of a polynomial series
representation of a function in terms of values dened at discrete points (nodes in our case).
For points along a line this would yield:
Ni =
23
m+1
j=1,j=i (xxj )
m+1
j=1,j=i (xi xj )
(6.25)
If expanded, the preceding equation would yield:

N1 =
N2 =
Nm+1 =
24
Q
Q
(x x2 )(x x3 ) (x xm+1 )
(x1 x2 )(x1 x3 ) (x1 xm+1 )
(x x1 )(x x3 ) (x xm+1 )
(x2 x1 )(x2 x3 ) (x2 xm+1 )
(x x1 )(x x2 ) (x xm )
(xm+1 x1 )(xm+1 x2 ) (xm+1 xm )
(6.26)
For the axial member, m = 1, x1 = 0, and x2 = L, the above equations will result in:
u=
x
x
x
(x L)
u1 + u2 = (1 ) u1 +
u2
L
L
L
L

N1
(6.27)
N2
which is identical to Eq. 6.10.

Victor Saouma
Draft
66
Figure 6.3: Rectangular Bilinear Element

6.2.1.5
Rectangular Bilinear Element
Next we consider a quadrilateral element, Fig. 6.3 with bi-linear displacement eld (in both
x and y).
25
26 Using the Lagrangian interpolation function of Eq. 6.25, and starting with the u displacement, we perform two interpolations: the rst one along the bottom edge (1-2) and along the
top one (4-3).
27
From Eq. 6.25 with m = 1 we obtain:

u12 =
=
28
x2 x
x1 x
u1 +
u2
x2 x1
x1 x2
ax
x+a
u1 +
u2
2a
2a
(6.28)
Similarly
u43 =
=
x2 x
x1 x
u4 +
u3
x2 x1
x1 x2
ax
x+a
u4 +
u3
2a
2a
(6.29)
Next, we interpolate in the y direction along 1-4 and 2-3 between u12 and u43 . Again, we
use Eq. 6.25 however this time we replace x by y:
29
y1 y
y2 y
u12 +
u43
(6.30)
y2 y1
y1 y2
byx+a
y+bax
y+bx+a
byax
u1 +
u2 +
u4 +
u3
=
2b 2a
2b 2a
2b 2a
2b 2a
(a x)(b y)
(a + x)(b y)
(a + x)(b + y)
(a x)(b + y)
=
u1 +
u2 +
u3 +
u4
4ab
4ab
4ab
4ab

u =
N1
Victor Saouma
N2
N3
N1
Draft
67
30 One can easily check that at each node i the corresponding Ni is equal to 1, and all others
to zero, and that at any point N1 + N2 + N3 + N4 = 1. Hence, the displacement eld will be
given by:
u1

u
N1 0 N2 0 N3 0 N4 0
v2
=
(6.31)
v
u3
0 N1 0 N2 0 N3 0 N4
v3
u
4
v4
6.2.1.6
Solid Rectangular Trilinear Element
31 By extension to the previous derivation, the shape functions of a solid rectangular trilinear
solid element, Fig. 6.4 will be given by:
Figure 6.4: Solid Trilinear Rectangular Element

0 N2 0
0 N3 0
0 N4 0
N1 0
u
0 N2 0
0 N3 0
0 N4
v
= 0 N1 0
0 N2 0
0 N3 0
0
0
0 N1 0
w
where
Ni =
Victor Saouma
(a x)(b y)(c z)
8abc
N4
u1
v1
w1
u2
v2
w2
u3
v3
w3
u4
v4
w4
(6.32)
(6.33)
Draft
68
Figure 6.5: Flexural Finite Element
6.2.2
6.2.2.1
C1
Flexural
With reference to Fig. 6.5. We have 4 d.o.f.s, {u}41 : and hence will need 4 shape functions,
N1 to N4 , and those will be obtained through 4 boundary conditions. Therefore we need to
assume a polynomial approximation for displacements of degree 3.
32
v = a1 x3 + a2 x2 + a3 x + a4
dv
= 3a1 x2 + 2a2 x + a3
=
dx
33
(6.34)
(6.35)
Note that v can be rewritten as:
v = x3
a1
a2
2
x x 1
a3

[p]
a4

(6.36)
{a}
34
We now apply the boundary conditions:

1. v = v 1 at x = 0
2. v = v 2 at x = L
dv at x = 0
3. = 1 = d
x
4. = 2 = dv at x = L
dx
or:
v1
0
0
0
0
1
=
L3 L2
v
3L2 2L
2

{u}
Victor Saouma
[L]
0
1
L
1
1
a1
0
a2
1
a
3
0
a4

{a}
(6.37)
Draft
N1
N2
N3
N4
69
=0
Ni Ni,x
1
0
0
1
0
0
0
0
Function
= (1 + 2 3 3 2 )
= x(1 )2
= (3 2 2 3 )
= x( 2 )
=1
Ni Ni,x
0
0
0
0
1
0
0
1
Table 6.1: Characteristics of Beam Element Shape Functions

35
which when inverted yields:
1
a2
= 3
a
3
a4

{a}
36
2
L
2
L
v1
3L 2L2 3L L2 1
0
L3
0
0
v
2
3
L
0
0
0
2

[L]1
{u}
Combining Eq. 6.38 with Eq. 6.36, we obtain:
2
L
2
L
v1
2 3L L2
1
3L
2L
1
u = x3 x2 x 1 3
L3
0
0
v

L 0
2
[p]
0
0
0
L3
2

[L]1
(6.38)
{u}
v1
(1 + 2 3 3 2 ) x(1 )2 (3 2 2 3 ) ( 2 ) 1

=
N1
N2
N3
N4
v2

2

[p][L]1

{u}
(6.39)
(6.40)
[N]
where = xl .
37
Hence, the shape functions for the exural element are given by:
N1 = (1 + 2 3 3 2 )
N2 = x(1 )2
N3 = (3 2 2 3 )
N4 = x( 2 )
(6.41)
and are shown in Fig 6.6.

38
Table 6.1 illustrates the characteristics of those shape functions
6.2.2.2
C 1 : Hermitian Interpolation Functions
For problems involving the rst derivative of the shape function, that is with C 1 interelement
continuity (i.e continuity of rst derivative or slope) such as for exure, Hermitian interpolation
functions rather than Lagrangian ones should be used.
39
Victor Saouma
Draft
610
Shape Functions for Flexure

(v1; 1; v2; 2)
1.0
0.8
N1
N3
N2
N4
0.6
0.4
0.2
0.0
0.2
0.0
0.2
0.4
0.6
0.8
1.0
(x/L)
Figure 6.6: Shape Functions for Flexure of Uniform Beam Element.

40 Hermitian interpolation functions are piecewise cubic functions which satisfy the conditions
of displacement and slope (C 0 , C 1 ) continuities. They are extensively used in CAD as Bezier
curves.
6.2.3
Characteristics of Shape Functions
1. The basis of derivation of shape functions could be:

(a) A polynomial relation
i. Exact
ii. Approximation
(b) Or other
i. Logarithmic
ii. Trigonometric
2. Shape functions should
(a) be continuous, of the type required by the variational principle.
(b) exhibit rigid body motion (i.e. v = a1 + . . .)
(c) exhibit constant strain.
3. Shape functions should be complete, and meet the same requirements as the coecients
of the Rayleigh Ritz method.
6.3
Natural Coordinate System
Natural coordinates are dimensionless and dened with respect to element length/area/volume
rather than the global coordinate system.
41
42 They are often used in element formulation of simplex elements (where in an n dimensional
space, there are n + 1 vertices and n + 1 surfaces of dimensionality n 1).
Victor Saouma
Draft
6.3 Natural Coordinate System
611
L
L2
L1
P
x1
x1
L
=
2
x2
L
1
= x + x
1 1
2 2
x =x+L
2
Figure 6.7: Natural Coordinate System Along a Straight Line
6.3.1
43
Straight Line
We dene two natural coordinates 1 and 2 , Fig. 6.7

1 =
L1
L
and 2 =
L2
L
(6.42)
since L1 + L2 = L, we have
1 + 2 = 1
44
We note that i = 1 at node i and zero at all other nodes.
45
An arbitrary point P with coordinate x has

x = 1 x1 + 2 x2
(6.43)
46 Thus, for every point x, corresponds a set of (nonindependent) natural coordinates 1 , 2 .

Those equations can be stated as

1
x2 1
1
1
1
1
1 1
(6.44)
and
=
=
2
2
1
x
x
x
x1 x2
L
1
47
Interpolation can be done in natural coordinates

1
= 1 2
2
(6.45)
48 Integration of polynomials in 1 and 2 can be done through (a rigorous proof is developed

below through Eq. 6.60)

k!l!
(6.46)
1k 2l dL = L
(1 + k + l)!
L
Furthermore
Victor Saouma

xdL
=
L (1 x1 + 2 x2 )dL =
L2
2
dL
=
L
=
L 1 2
4!

L
2 (x1
L
12
+ x2 )
(6.47)
Draft
612
L =1
3
A
P 1
A2
A
L =0
2
Side
e
Sid
L =0
L =1
Side 3
L =0
L =1
Figure 6.8: Natural Coordinate System for a Triangle
6.3.2
Triangular Coordinates
49 For a triangular surface, any point P divides the triangle 1-2-3 into three subareas A1 , A2 , A3 ,
Fig. 6.8 thus we dene area coordinates as ratios of areas
L1 =
A1
,
A
L2 =
A21
,
A
L3 =
A3
A
(6.48)
where A is the area of the triangle 1-2-3.

50
Since A = A1 + A2 + A3 then
L1 + L2 + L3 = 1
51
52
The centroid is at L1 = L2 = L3 =
(6.49)
1
3
Again the constraints equations are
1 1 1
L1
1
L2
x
= x1 x2 x3

y1 y2 y3
L3
y

(6.50)
[A]
inverting
x2 y3 x3 y2 y2 y3 x3 x2 1
L1
L
= x3 y1 x1 y3 y3 y1 x1 x3
x

2
L3
x1 y2 x2 y1 y1 y2 x2 x1
y

[A]1

2A23 y23 x32 1
1
2A31 y31 x13
x

2A
2A12 y12 x21
y
(6.51-a)
(6.51-b)
where xij = xi xj
Victor Saouma
Draft

53
613
We observe that the area is equal to the determinant over two.

2A = |A| = x2 x1 y3 y1 x31 y21
(6.52)
and if the labels are reversed such that 1-2-3 is clockwise, then the determinant is negative.
54
A function may be expressed in terms of area coordinates

= (L1 , L2 , L3 )
L1
L2
L3
=
+
+
x
L1 x
L2 x
L3 x
L1
L2
L3
=
+
+
y
L1 y
L2 y
L3 y
and
L1
x
L1
y
y2 y3
2A ;
x3 x2
2A ;
=
=
L2
x
L2
y
y3 y1
2A ;
x1 x3
2A ;
=
=
L3
x
L3
y
=
=
y1 y2
2A ;
x2 x1
2A ;
(6.53-a)
(6.53-b)
(6.53-c)
(6.54)
55 Integration for the stiness matrix will require integration in area coordinate of expressions

of the form A Lk1 Ll2 Lm
3 dA. To accomplish this operation, and with reference to Fig. 6.9,
(Eisenberg and Malvern 1973):
h2
3
3
L =0
2
ds
ds 1
11
00
00
11
00
11
00
11
00
11
00
11
00
11
L =0
2
L =0
3
h1
Figure 6.9: Integration over a Triangle

dA = (ds1 ) csc 3 (ds2 )
Victor Saouma
(6.55-a)
= (h1 dL1 ) csc 3 (h2 dL2 )
(6.55-b)
= 2AdL1 dL2
(6.55-c)
Draft
614
Thus

A
Lk1 Ll2 Lm
3 dA

= 2A
1 1L1
0

Lk1 Ll2 (1
L1 L2 ) dL2 dL1
m
(6.56)
we next substitute L2 = t(1 L1 ) and dL2 = (1 L1 )dt which yields

A
Lk1 Ll2 Lm
3 dA
1
1
k
l+m+1
= 2A L1 (1 L1 )
dL1 tl (1 t)m dt
0
(6.57)
each of the integrals on the right-hand side is of the form of the beta function (Abramowitz
and Stegun 1970)
1
(z)(w)
(6.58)
tz1 (1 t)w1 dt =
B(z, w) =
(z + w)
0
where denotes the gamma function which satises (n + 1) = n! for integers n 0. Thus

(k + 1)(l + 1)(m + 1)
(6.59)
Lk1 Ll2 Lm
3 dA = 2A
(k + l + m + 3)
A
Since k, l, and m are nonnegative integers,

A
Lk1 Ll2 Lm
3 dA = 2A
k!l!m!
(k + l + m + 2)!
(6.60)
56 It should be noted that the above relation is used only if we have straight sided elements and
the element formulation can thus be analytically derived. Alternatively, if we have a curvilinear
side, this formula will not be used, and the element will be numerically integrated within the
context of an isoparametric formulation.
6.3.3
57
Volume Coordinates
Volumes coordinates are a direct extension of triangular coordinates.
An internal point P subdivide the tetrahedron into four subtetrahedra, and thus Li =
for i = 1, 4.
58
59
The constraint equation is

1

x
=
z
L1
1 1 1 1
L
x1 x2 x3 x4
2
y1 y2 y3 y4
L
3
z1 z2 z3 z4
L4

Vi
V ,
(6.61)
[A]
60
Integration is also carried in a similar way,

n
Lk1 Ll2 Lm
3 L4 dA = 6V
V
Victor Saouma
k!l!m!n!
(3 + k + l + m + n)!
(6.62)
Draft
6.3.4
615
Interpolation Functions
61 Triangular elements allow a complete polynomial in Cartesian coordinate to be used for the
eld (e.g. temperature, displacement) quantity. Thus, all terms of a truncated Pascal triangle
are used in the shape functions. (This will not be the case for the bilinear quadrilateral element).
)
62 We seek to determine =
Ni i where i are nodal d.o.f. and Ni = Ni (L1 , L2 , L3 ).
)n
q
r
s
63 We start with =
i=1 ai L1 , L2 , L3 ) where q, r and s range over the n possible combinations
for which q + r + s = p and where p is the order of the polynomial. For example for
Linear element:
= a1 L1 + a2 L2 + a3 L3 = a 1 + a 2 x + a 3 y
(6.63)
Quadratic element:
= a1 L21 + a2 L22 + a3 L23 + a4 L1 L2 + a5 L2 L3 + a6 L3 L1 = a 1 + a 2 x+ a 3 y + a 4 x2 + a 5 xy + a 6 y 2
(6.64)
64
The shape functions for the Linear Triangle are simply the area coordinates,
N1 = L1
N2 = L2
N3 = L3
(6.65)
and each one is equal to unity at one node, zero at the others, and varies linearly.
65 The shape functions for other elements can be obtained from the Lagrangian interpolation
function used earlier Eq. 6.25
-n
i=0,i =k (L Li )
n
(6.66-a)
lk = -n
i=0,i =k (Lk Li )
(L L0 )(L L1 ) (L Lk1 )(L Lk+1 ) (L Ln )

(Lk L0 )(Lk L1 ) (Lk Lk1 )(Lk Lk+1 ) (Lk Ln )
(6.66-b)
66 Hence, denoting a typical node i by three numbers q, r and s corresponding to the position
of coordinates L1i , L2i and L3i , we can write the shape functions in terms of three Lagrangian
interpolation functions
(6.67)
Ni = lqq (L1 )lrr (L2 )lss (L3 )
67
Using this formula, the shape functions for higher order elements are
Linear triangle
Ni = Li
(6.68)
Quadratic triangle
Ni = Li (2Li 1) Corner nodes
Midside nodes, i, j, k are on the same side
Nj = 4Li Lk
Victor Saouma
(6.69)
Draft
616
Cubic triangle
Ni = 12 Li (3Li 1)(3Li 2) Corner nodes
Midside nodes, i, j, k are on the same side
Nj = 92 Li Lk (3Li 1)
Internal node
N10 = 27L1 L2 L3
6.4
(6.70)
Pascals Triangle
68 A schematic interpretation of shape functions in terms of polynomial series terms is given

by Pascals triangle which is shown in Table 6.2.
Constant
Linear
Quadratic
Cubic
Quartic
a11 x4
a1
3
a7 x
a4 x2
a12 x3 y
a2 x
2
a8 x y
a3 y
a5 xy
a13 x2 y 2
a9 xy
a6 y 2
a14 xy 3
a10 y 3
(6.71)
a15 x4
Table 6.2: Interpretation of Shape Functions in Terms of Polynomial Series (1D & 2D)
69
Polynomial terms present in various element formulations is shown in Table 6.3

Element
Terms
Linear
Quadratic
Bi-Linear (triangle)
Bi-Linear (quadrilateral)
Bi-Quadratic (Serendipity)
Bi-Quadratic (Lagrangian)
a1 ,
a1 ,
a1 ,
a1 ,
a1 ,
a1 ,
a2
a2 ,
a2 ,
a2 ,
a2 ,
a2 ,
a4
a3 ,
a3 , a5
a3 , a4 , a5 , a6 , a8 , a9
a3 , a4 , a5 , a6 , a8 , a9 , a13
# of Nodes
(terms)
2
3
3
4
8
9
Table 6.3: Polynomial Terms in Various Element Formulations (1D & 2D)
70
A complete polynomial contains all the terms above a certain line in the Pascal triangle.
71 The more terms are included, the higher the accuracy. For instance u = a1 +a2 x+a3 y +a5 xy
is more accurate than u = a1 +a2 x+a3 y, however the rate of convergence is unchanged (because
the second approximation is not a complete quadratic one).
72 Terms in the approximation which do not improve the rate of convergence are called parasitic
terms.
Victor Saouma
Draft
Chapter 7
FINITE ELEMENT
DISCRETIZATION and
REQUIREMENTS
7.1
Discretization
This section is mostly extracted from (Reich 1993)
7.1.1
Discretization of the Variational Statement for the General TPE Variational Principle
The discretization of Equation 5.8 will be performed on an element domain e using the
procedures described in Chapter 2 of (Zienkiewicz and Taylor 1989);
1
2 The surface of the element subjected to surface tractions t comprises one or more surfaces
of the element boundary . For the present time this discussion will be kept on a very general
level with no mention of the dimensionality of the elements; the number of nodes dening the
elements; or the nature of the constitutive law.
The rst step in the discretization process is to dene the displacements u at a point inside
the element in terms of the shape functions N and the nodal displacements ue for the element
3
(7.1)
u = Nue
4 The virtual displacements u at a point inside the element can also be dened in terms of
the shape functions N and the nodal virtual displacements ue for the element
(7.2)
u = Nue
5
In order to discretize the volume integral in Equation 5.8

=
(Lu)T D(Lu)d
(Lu)T D 0 d +
(Lu)T 0 d
uT bd
uT td = 0
t
(7.3)
Draft
72
FINITE ELEMENT DISCRETIZATION and REQUIREMENTS
dening the virtual strain energy for the element due to the nodal displacements u, the strains
at a point inside the element are expressed in terms of the nodal displacements ue using
Equation 7.1
= Lu = LNue
(7.4)
and the virtual strains at a point inside the element are expressed in terms of the nodal
virtual displacements ue using Equation 7.2
= (Lu) = LNue
6
Dening the discrete strain-displacement operator B as

def
(7.6)
B = LN
the virtual strain energy for an element is written as

T
T
(Lu) D(Lu)d = ue
e
(7.5)
BT DBdue
(7.7)
Dening the element stiness matrix Ke as

BT DBd
Ke =
(7.8)
Equation 7.7 can be rewritten as

(Lu)T D(Lu)d = uTe Ke ue
(7.9)
e
8 In order to discretize the volume integrals in Equation 7.3 dening the virtual strain energy
for the element due to the initial strains 0 and stresses 0 , Equations 7.5 and 7.6, which dene
the virtual strains at a point inside the element in terms of the nodal virtual displacements
ue , are substituted into the integrands

T
T
(Lu) D 0 d = ue
BT D 0 d
(7.10)
e
9
(Lu)T 0 d = uTe
BT 0 d
(7.11)
Dening the initial force vector f0e as

f0e =

B D 0 d
T
BT 0 d
the strain energy due to the initial strains and stresses is

(Lu)T D 0 d
(Lu)T 0 d = uTe f0e
e
Victor Saouma
(7.12)
(7.13)
Draft
7.1 Discretization
73
10 In order to discretize the volume integral dening the work done by the body forces and the
surface integral dening the work done by the surface tractions in Equation 7.3, Equation 7.2
is substituted into the integrands

uT bd = uTe
NT bd
(7.14)
uT td = uTe
t
11
NT td
(7.15)
NT td
(7.16)
Dening the applied force vector fe as

NT bd +
fe =
e
the sum of the internal and external virtual work due to body forces and surface tractions is

T
u bd +
uT td = uTe fe
(7.17)
e
12 Having obtained the discretization of the various integrals dening the variational statement
for the TPE variational principle, it is now possible to dene the discrete system of equations. Substituting Equations 7.7, 7.13, and 7.17 into Equation 7.3 and rearranging terms, the
discretized Principle of Virtual Work is
uTe Ke ue = uTe fe + uTe f0e
(7.18)
Since uTe is an arbitrary (i.e. non-zero) vector appearing on both sides of Equation 7.18,
the discrete system of equations can be simplied into
13
Ke ue = fe + f0e + Pu
(7.19)
as the discrete system of equations for an element.
7.1.2
Discretization of the Variational Statement for the HW Variational

Principle
14 The discretization of the three variational statements dened in Equation 5.65, 5.66, and
5.67 will be performed on an element domain e using the procedures described in Chapter
2 of (Zienkiewicz and Taylor 1989) assembly of the discrete element equations into a discrete
global system of equations is straightforward and will be omitted from this discussion.
The surface of the element subjected to surface tractions t comprises one or more surfaces
of the element boundary e . For the present time this discussion will be kept on a very general
level with no mention of the dimensionality of the elements; the number of nodes dening the
elements; or the nature of constitutive law.
15
The rst step in the discretization process is to dene the displacements u, strains , and
stresses at a point inside the element in terms of the shape functions Nu , N= , and N ,
respectively, and the element nodal displacements ue , strains e , and stresses e
16
Victor Saouma
Draft
74

u = Nu ue
(7.20)
= N = e
(7.21)
= N e
(7.22)
We note that contrarily to the previous case (Eq. 7.1) we now have three discretizations (instead
of just one).
17 The virtual displacements u, virtual strains , and virtual stresses at a point inside
the element can also be dened in terms of the shape functions Nu , N= , and N , respectively,
and the nodal virtual displacements ue , virtual strains e , and virtual stresses e for the
element
18
u = Nu ue
(7.23-a)
= N= e
(7.23-b)
= N e
(7.23-c)
We now need to discretize each one of the corresponding Euler

T
T
(Lu) d u b d
uT t d
t

T [D( 0 ) + 0 ] d

T (L u ) d
equations:
= 0
(7.24-a)
= 0
(7.24-b)
= 0
(7.24-c)
In order to discretize the volume integral in the rst variational statement (i.e. Equ. 7.24-a)
dening the virtual strain energy for the element, Equation 7.23-a is substituted into the virtual
strain-displacement relationship (i.e. Equation 5.6) to dene the virtual strains at a point
inside the element in terms of the nodal virtual displacements ue
19
(L u) = L u = L Nu ue
20
(7.25)
Dening the discrete strain-displacement operator Bu as

Bu = L Nu
(7.26)
and substituting Equation 7.22 into the integrand, the virtual strain energy for an element is
written as

T
T
(Lu) d = ue
BTu N d e
(7.27)
e
21
Dening an element operator matrix Fe as

T
Fe =
BTu N d
(7.28)

(Lu)T d = uTe FTe e
(7.29)
Victor Saouma
Draft
7.1 Discretization
75
22 In order to discretize the volume integral dening the work done by the body forces and the
surface integral dening the work done by the surface tractions in the rst variational statement
(i.e. the rst equation in Equation ??), Equation 7.23-a is substituted into the integrands

T
T
u b d = ue
NTu b d
(7.30)
uT t d = uTe
NTu t d
(7.31)
Dening the applied force vector fe as

fe =

NTu b d +
NTu t d
(7.32)
the sum of the internal and external virtual work is

T
u b d +
uT t d = uTe fe
e
(7.33)
23 Having dened the discretization of the various integrals in the rst variational statement
for the HW variational principle (i.e. Equ. 7.24-a), it is now possible to dene the discrete
system of equations. Substituting Equations 7.29 and 7.33 into the variational statement and
rearranging terms, the discretized Principle of Virtual Work is
uTe FTe e = uTe fe
(7.34)
where the left-hand side is the virtual strain energy and the right-hand side is the internal and
external virtual work. Since ue is an arbitrary (i.e. non-zero) vector appearing on both sides
of Equation 7.34, the discrete system of equations can be simplied into
FTe e = fe
(7.35)
as the discrete system of equations for an element.

24 In order to discretize the second variational statement (i.e. Equ. 7.24-b), Equations 7.21,
7.22, and 7.23-b are substituted into the integrand

T
T
T
T
[D( 0 ) + 0 ] d = e
N= D N= d e e
NT= D 0 d
+ Te
e
25
NT= 0 d Te
NT= N d e = 0
(7.36)
Dening a pair of element operator matrices Ae and Ce as

NT= D N= d
Ae =
(7.37)
NT= N d
Ce =
(7.38)
Victor Saouma
Draft
76
and the initial strain/stress vector ge as

NT= D 0 d
ge =
e
NT= 0 d

T [D( 0 ) + 0 ] d = Te Ae e Te ge Te Ce e = 0
(7.39)
(7.40)
Since the nodal virtual strains are arbitrary they can be eliminated from Equation 7.40
yielding
Ae e Ce e = ge
(7.41)
as the discretized form of the second variational statement.

26 In order to discretize the third variational statement (i.e. Eq. 7.24-c), Equations 7.20, 7.21,
and 7.23-c are substituted into the integrand

T
T
T
T
(L u ) d = e
N Bu d ue e
NT N= d e = 0
(7.42)
Recognizing that
NT Bu d = Fe
(7.43)
NT N= d = CTe
(7.44)

T (L u ) d = Te Fe ue Te CTe e = 0
(7.45)
Since the nodal virtual stresses e are arbitrary they can be eliminated from Equation 7.45
yielding
Fe ue CTe e = 0
(7.46)
as the discretized form of the third variational statement.

27 Having dened the discretized form of all three variational statements, it is now possible to
dene the discrete mixed system of equations for an element. Assembling Equations 7.35, 7.41,
and 7.46 in matrix form adopting the classic arrangement for a constrained system of equations

Ae Ce 0
e ge
CTe
0
Fe
e
0
(7.47)
=
T
0
Fe
0
ue
fe
yields a symmetric system of equations. Although is technically an intermediate variable

in the eld equations indirectly relating to u, e is the primary variable and e to ue are
constraint variables in Equation 7.47.
Victor Saouma
Draft
7.2 General Element Requirements
77
28 Since it would be computationally expensive to solve the system of equations in Eq. 7.47
using direct method, an indirect or iterative procedure (i.e. Gauss-Seidel instead of GaussJordan) is often selected, (Zienkiewicz and Taylor 1989).
Step
Step
Step
Step
1:
2:
3:
4:
uk+1
n
k+1
n
k+1
n
rk+1
n
=
=
=
=
ukn + K1 rkn
CT Fuk+1
n
C1 Ak+1
n
f FT k+1
n
(7.48)
for k = 0, 1, 2, , where k is an iteration index and rk+1

is the residual force vector. It should
n
be noted that this procedure is solved on the structural level, meaning that steps 1 to 3 require
a solution of a system of linear equations.
Step 1, K corresponds to the classical standard displacement stiness matrix, and this step is
used as a pre-conditioner. This implies that at the beginning of the rst iteration, when u0n = 0
and r0n =, step corresponds to the standard displacement-based formulation of the nite element
method. Steps 1, 2, and 3 above require the solution of simultaneous linear equations. Step 3,
however, may be reduced by nodal quadrature and assuming same interpolation functions for
strains and stresses to
(7.49)
i = D i
In this equation, i and i are the stresses at node i, respectively, and D is the stress-strain
constitutive matrix. Then, Step 3 is nothing else but direct computation of nodal stresses from
nodal strains using the constitutive matrix D.
Finally, the uniqueness and the existence of a solution has been addressed by the so-called
Babuska-Brezzi (BB) condition (Babuska 1973, Brezzi 1974).
29
30
Details of the algotithmic implementation will be covered in a later chapter.
7.2
General Element Requirements
A nite element (just a an approximate displacement eld in the Rayleigh-Ritz formulation)

must satisfy two basic requirements
31
Completeness: The FE discretization must at least accommodate constant displacement and

constant strain (or temperature and temperature gradient). This is accomplished by
including in two dimensional problems the following
= a1 + a2 x + a3 y + possibly additional terms
(7.50)
Compatibility or Conformity: The approximation of the eld over element boundaries must
be continuous (C0 or C1 continuity). Most nite elements are conforming, but some are
not.
32 For instance, with respect to Fig.
7.1, element A must be capable of undergoing rigid
body motion without internal strains/stresses, and at node B we should have continuity of
displacement (but not slope for this element).
33 If those two requirements are satised, then convergence is assured. In the FE method
approximate solution are obtained, and the more elements we use, the more accurate is the
Victor Saouma
Draft
78
111
000
000
111
000
111
000
111
000
111
111
000
000
111
000
111
000
111
000
111
B
A
Figure 7.1: Completness and Compatibility

approximate solution. In the limit, for innitely small elements, we require the solution to
be also innitely close to the exact one. Hence, convergence is ensured if completeness and
compatibility requirements are satised.
34
35
Lax equivalence theorem:

Convergence criterion=completeness+ compatibility requirements
Two essential requirements:
Patch test: Completeness can be assessed through the patch test which will be discussed
later.
Zero Strain Energy: For structural problems, there should be zero strain energy when the
element is subjected to a rigid body motion. Recall that in the stiness matrix formulation, the matrix is singular as it embodies not only the force displacement relations,
but the equilibrium equations also. To each of those equations, corresponds a rigid body
mode which can be detected by an eigenvalue analysis (more about this later).
7.3
Discretization Error and Convergence Rate
Approximation will yield an exact solution in the limit as the size h of element approaches
zero.
36
37 In some cases the exact solution is obtained with a nite number of elements (or even with
only one) if the polynomial expansion used in that element can t exactly the correct solution.
(e.g. Truss, beam elements, plane stress element used in a plate under pure axial load).
38
The exact solution can always be expanded in the vicinity of any node i as a polynomial

u
u
(x xi ) +
(y yi ) + ...
(7.51)
u = ui +
x i
y i
If within an element of size h a polynomial of degree p is used, then the Taylor expansion up
to that degree can be accommodated and the error is of the order O(hp+1 ).
Victor Saouma
Draft
7.3 Discretization Error and Convergence Rate
79
Thus for a linear element, the convergence rate is of order O(h2 ), i.e., the error in displacement is reduced to 1/4 of the original error by halving the mesh.
39
40 By a similar argument, it can be shown that the strains (or stresses) which are given by
the mth derivatives of the displacement should converge with an error of O(hp+1m ), and the
strain energy (which is given by the square of the stresses) error will be O(h2(p+1m) )
41 From the above arguments, knowledge of the order of convergence may help in extrapolating
the solutions to the correct one. Hence, for instance, if the displacement converges at O(h2 ),
and we have two approximate solutions u1 and u2 obtained with meshes of sizes h and h/2,
then we can write
O(h2 )
u1 u
=
=4
(7.52)
u2 u
O(h/2)2
where u is the extrapolated value.

42
Convergence to the exact solution can be accelerated by, Fig. 7.2.

p
h
Figure 7.2: h, p and r Convergence
h reducing the size of the elements (or mesh renement).

p increasing the order of the polynomial (same number of elements but higher number of
nodes/dof).
A third form of acceleration is the so called r renement in which the same number of nodes/elements
is retained but the mesh is shifted around to increase its density in zones of high stress gradient.
43 The above convergence procedures can be accelerated within the context of a program which
can accommodate adaptive remeshing techniques.
44
Finally, additional errors may come from round-o within the computer.
Victor Saouma
Draft
710
7.4
Lower Bound Character of Minimum Potential Energy Based

Solutions
A numerical solution that is derived from the principle of minimum total potential energy is
a lower bound solution, because the strain energy is smaller than the exact one (i.e. obtained
from an innite number of elements).
45
This can be readily shown if we consider the displacement ui caused by a load Pi which is
increased from zero to its stipulated value. The work done is Pi2ui and must be equal to the
internal strain energy U . Alternatively, the potential of the applied load is Pi ui , and the exact
potential energy is
Pi uexact
Pi uexact
i
i
Pi uexact
(7.53)
=
exact =
i
2
2
similarly, the approximate value of the potential energy is
46
approx
Pi uapprox
i
=
2
(7.54)
We know that the approximation of is algebraically higher than the exact value (since the
exact value is a minimum), hence
Pi uiapprox
Pi uiexact
<
2
2
exact
(7.55)
approx
or
uapprox
< uexact
i
i
(7.56)
or alternatively, the solution is too sti.

Note the similarity with the Rayleigh-Ritz method, the lack of enough terms in our polynomial approximation did also result in a sti solution.
47
7.5
Equilibrium and Compatibiliy in the Solution
48
In an exact solution over each dierential element equilibrium and compatibility prevails.
49
In a nite element discretization, this is not necessarily the case:
Equilibrium of Nodal Forces: Those are automatically satised by denition (ke u f = 0)

at the nodes.
Compatibility at Nodes: Elements connected to one another have the same displacements
(along corresponding dof) at the connecting node.
Interelement Continuity of Stress: is usually not satised, that is equilibrium across the
elements does not necessarily prevail. For example consider Fig. 7.3 if node 4 is the only
one displaced (all others remaining xed), then there will be a discontinuity of xx across
element boundary 2-3.
Victor Saouma
Draft
7.5 Equilibrium and Compatibiliy in the Solution
711
x2
u
2
Figure 7.3: Interelement Continuity of Stress

Interelement Compatibility of Displacements: Most elements have this requirement satised. If an element does not then it is labeled as incompatible or nonconforming,
Fig. 7.4
6
2
y,v
x,u
1
5
Figure 7.4: Interelement Continuity of Strain
Equilibrium Inside Elements: is not always satised. In general satisfaction of equilibrium
at every point demands relations between dof which do not necessarily result from ke u
f =0
Displacement Compatibility Inside an Element: is satised as we only require that the
element displacement be continuous and single-valued.
Victor Saouma
Draft
712
Victor Saouma
Draft
Chapter 8
STRAIGHT SIDED ELEMENTS;

1st GENERATION
8.1
Introduction
1 Having rst introduced the method of virtual displacements, than the interpolation (or shape)
functions [N], which relate internal to external nodal displacements, and nally having applied
the virtual displacement method to nite element in chapter 7, we derive the stiness matrices
of some simple elements.
8.2
Rod Elements
8.2.1
Truss Element
The shape functions of the truss element were derived in Eq. 6.11:
x
N1 = 1
L
x
N2 =
L
The corresponding strain displacement relation [B] is given by:

du
dx
= [ dN1 dN2 ]
dx
dx
= [ L1 L1 ]

xx =
(8.1)
[B]
For the truss element,

the constitutive matrix [D] reduces to the scalar E; Hence, substituting
[B]T [D][B]d and with d = Adx for element with constant cross
into Eq. 1.15 [ke ] =
Draft
STRAIGHT SIDED ELEMENTS; 1st GENERATION
82
sectional area we obtain:

[k ] = A
L
AE
L2
AE
L
L
L1
E L1
1
L
1 1
1
1
0

1 1
1 1
1
L
dx
(8.2-a)

dx
(8.2-b)
We observe that this stiness matrix is identical to the one earlier derived in Eq. ??.
8.2.2
Beam Element
6 For a beam element, for which we have previously derived the shape functions in Eq. 6.41
and the [B] matrix in Eq. 1.4-d, substituting in Eq. 1.15:
l
e
[B]T [D][B] y 2 dA dx
[k ] =
0
(8.3)
y 2 dA = Iz Eq. 1.15 reduces to
and noting that

A
[B]T [D][B]Iz dx
[k ] =
(8.4)
For this simple case, we have: [D] = E, thus:

[ke ] = EIz
[B]T [B] dx
(8.5)
0
8 Using the shape function for the beam element from Eq. 6.41, and noting the change of
integration variable from dx to d, we obtain
[ke ] = EIz
1
0
6
L2 (2 1)
L2 (3 2)
6
L2 (2 + 1)
L2 (3 1)
6
L2 (2
1) L2 (3 2)
6
L2 (2
+ 1) L2 (3 1)
Ld (8.6)

dx
or
[ke ] =
v1
z
V1 12EI
L3
6EI
z
M
1
L2
12EIz
V2
L3
6EI
z
M
2
2
6EIz
L2
4EIz
L
z
6EI
L2
2EIz
L
v2
z
12EI
L3
6EIz
L2
12EIz
L3
z
6EI
L2
6EIz
L2
2EIz
L
z
6EI
L2
4EIz
L
(8.7)
The generation of this stiness matrix with Mathematica is given by:
Victor Saouma
Draft
8.3 Triangular Elements
83
Beam shape functions;

N1[x_,L_]:=1-3 (x/L)^2+2 (x/L)^3;
N2[x_,L_]:=x (1-(x/L))^2;
N3[x_,L_]:=3 (x/L)^2 -2 (x/L)^3;
N4[x_,L_]:=(x-L) (x/L)^2;
Visualize the shape functions;
Plot[{N1[x,1],N3[x,1]},{x,0,1}];
Plot[{N2[x,1],N4[x,1]},{x,0,1}];
Define the B matrix;
BMat= {D[N1[x,L], {x,2}],D[N2[x,L], {x,2}],D[N3[x,L], {x,2}],D[N4[x,L], {x,2}]} // Simplify
Determine the element stiffness matrix in various forms and shapes;
Integrate[EI Outer[Times,BMat,BMat], {x, 0, L}]
% // MatrixForm
% / (EI/L^3) // Simplify
% // MatrixForm
8.3
8.3.1
Triangular Elements
Cartesian Coordinate System (CST)
Having retrieved the stiness matrices of simple one dimensional elements using the principle
of virtual displacement, we next consider two dimensional continuum elements starting with
the triangular element of constant thickness t made out of isotropic linear elastic material. The
element will have two d.o.fs at each node:
10
{u} = u1 u2 u3 v1 v2 v3 T
11
The strain displacement relations is required to determine [B]
12
For the 2D plane elasticity problem, the strain vector {} is given by:
{} = xx yy xy T
(8.8)
(8.9)
hence we can rewrite the strains in terms of the derivatives of the shape functions through the
matrix [B]:

0
xx
u
x
N
0
(8.10)
=
y
v
yy
N
N
xy
y
x

[B]
13 We note that because we have 3 u and 3 v displacements, the size of [B] and [u] are 9(33)6
and 6 1 respectively.
Victor Saouma
Draft
84
14
Dierentiating the shape functions from Eq. 6.22 we obtain:
1
1
0
0
0
0
xx

x
x2
2
N3

yy
x
N1
N2
u1
x
x
xy
u
x
x
x
1
2
3
2
3
2
0
0
0
xx
u
3
x
y
x
y
y
2
2
3
2
3
3
yy
=

v1
2
N1
N2
N3
xy
y
y
y
v
2
x3
1
x3 x2
1
1
xx
0
3
3

yy
x2 y3
x2 y3
y3
x
x2

N3
2

3
{u}
x
N1
N2
xy
N1
N1
N3

x
x
y
y
x

{}
(8.11)
[B]
15 With the constitutive matrix [D] given by Eq. ??, the strain-displacement relation [B] by
Eq. 8.11, we can substitute those two quantities into the general equation for stiness matrix,
Eq. 7.8:
[ke ]
[B]T [D][B]d
6
6
6
6
6
6
6
4
Z
=
x12
|
2
=
6
6
6
6
6
6
4
where =
x3 x2
x2 y 3
3
xx
2 y3
1
y3
x12
1
x2
0
0
0
1
x2
0
0
0
0
x3 x2
x2 y 3
3
xx
2 y3
1
y3
{z
[B]T
y32 + x232
y32 x3 x32
x2 x32
y3 x32
x3 y3 + y3 x32
x2 y3
1
2 ,
1+
2 ,
3
7
2
7
1
7 E
7
4 1
7
7 1 2
0 0
7
5|
{z
1
)
2
[D]
32
x12
56
0
4
}|
x3 x2
x2 y 3
1
x2
0
0
3
xx
2 y3
1
y3
ET
,
2(1 2 )x2 y3
{z
x3 x2
x2 y 3
x12
3
xx
2 y3
1
y3
1
x2
x2 x32
x2 x3
x22
x2 y3
x2 y3
0
y3 x32
y3 x32 + x3 y3
x2 y3
y32 + x232
y32 x3 x32
x2 x32
x3 y3 + y3 x32
x3 y3
x2 y3
y32 x3 x32
y32 + x23
x2 x3
x2 y3
x2 y3
0
x2 x32
x2 x3
x22
x32 = x3 x2 , and y32 = y3 y2 .
16 For the evaluation of the internal stresses {} = [D][B]{u} hence for this particular element
we will have:
u1
u
1
1
0
0
0
0
x2
1
0
x
x2
E
u
x
x
x
3
1
3
2
3
0
0
0
y
=
1
0
x 2 y3
x 2 y3
y3
v1
1 2
x3 x2
x3
1
1
1
xy
)
0 0 1
2
x 2 y3
x 2 y3
y3
x2
x2
v2

{ }
[D]
v3
[B]

{u}
y3
y3
0 x32
= y3
y3
0
x32
x32 x3 x2 y3
Victor Saouma
u1
u2
x3 x2
u
3
x3
x2
(8.12)
v1
y3
0
v3 II; Solid Mechanics

Finite Elements
7 tdxdy
5
| {z }
} dV ol
[B]
y32 x3 x32
y32 + x23
x2 x3
y3 x32 + x3 y3
x3 y3
x2 y3
0
0
3
7
7
7
7
7
7
5
Draft
8.3 Triangular Elements

where =
8.3.2
85
E
(1 2 )x2 y3
Natural Coordinate System
From Eq. 8.12 we note that the element must be positioned in such a way that neither x2
nor y3 is equal to zero.
17
18
A palliative to this problem is to use a formulation based on natural coordinate systems.
8.3.2.1
Linear, T3
19
The 3 noded triangular element has a linear displacemenent eld.
20
For the linear triangle, the shape functions are given by Eq. 6.68, thus
u =
Li ui =
i=1
i=1
Ni ui
(8.13)
Li v i
i=1
Ni v i
i=1
or Ni = Li .
21
The displacement shape functions are given by Equation 6.51-b:
2A23 y23 x32

L1
1
L2
2A31 y31 x13
=
2A
L3
2A12 y12 x21
(8.14)
where xij = xi xj .
22
Derivatives required for strain-displacement relationships are obtained from Eq. 6.54:
L1
x
L1
y
23
24
=
=
y2 y3
2A ;
x3 x2
2A ;
L2
x
L2
y
=
=
Then the ith part of the B matrix becomes:
Ni,x
0
Ni,y
Bi = 0
Ni,y Ni,x
y3 y1
2A ;
x1 x3
2A ;
L3
x
L3
y
=
=
(i = 1, 2, 3)
The stiness matrix is then obtained from:

BT D Bd
K =
y1 y2
2A ;
x2 x1
2A ;
(8.15)
(8.16)
(8.17)
25 This expression is then analytically integrated using Eq. 6.60. In general an isoparametric
numerical integration is preferred (Section 9.2.3).
Victor Saouma
Draft
86
8.3.2.2
26
Quadratic Element (T6)
Quadratic displacement shape functions may be written in natural coordinates as:

u =
Ni ui
v =
i=1
Ni vi
(8.18)
i=1
where the shape functions were obtained in Eq. 6.69

N1 = (2L1 1)L1
N4 = 4L1 L2
N2 = (2L2 1)L2
N5 = 4L2 L3
N3 = (2L3 1)L3
N6 = 4L3 L1
(8.19)
27 It should be noted that alternatively, we could have obtained those shape functions by
assuming
(8.20)
u(L1 , L2 ) = a1 + a2 L1 + a3 L2 + a4 L21 + a5 L1 L2 + a6 L22
and then determine the coecients through satisfaction of the appropriate boundary conditions.
The shape functions and their derivatives with respect to the natural coordinates L1 and L2
are given in Table 8.1 (in terms of L1 , L2 , and L3 ).
28
i
1
2
3
4
5
6
Ni
(2L1 1)L1
(2L2 1)L2
(2L3 1)L3
4L1 L2
4L2 L3
4L1 L3
Ni,L1
4L1 1
0
0
4L2
0
4L3
Ni,L2
0
4L2 1
0
4L1
4L3
0
Ni,L3
0
0
4L3 1
0
4L2
4L1
Table 8.1: Shape Functions and Derivatives for T6 Element

29
The B matrix is obtained from:

0
xx
x
= 0 y
yy
xy
y
x

L
u
v

(8.21)
or
xx (L1 , L2 , L3 )
xx
(L , L , L )
=
yy 1 2 3
yy
xy
xy (L1 , L2 , L3 )
Victor Saouma
(8.22-a)
Draft
8.4 Bilinear Rectangular Element
i=1
2A
where
87
Ni (L1 ,L2 ,L3 )

x
0
Ni (L1 ,L2 ,L3 )
y
Ni (L1 ,L2 ,L3 )

y
Ni (L1 ,L2 ,L3 )
x
L1
x
B=LN
L2
L3
x
x
L1
y
L2
y
L3
y
N1
L1
N1
L2
N1
L3
N2
L1
N2
L2
N2
L3
N3
L1
N3
L2
N3
L3
0

[NL ] =
ui
vi

(8.22-b)
[NL ]

[NL ]
N4
L1
N4
L2
N4
L3
0
L1
y
L1
x
N5
L1
N5
L2
N5
L3
L2
y
L2
x
N6
L1
N6
L2
N6
L3
L3
y
L3
x
[NL ]
[NL ]
(8.23-a)
0
0
4L2
0 4L3
4L1 1
0
4L2 1
0
4L1 4L3
0
=
0
0
4L3 1 0 4L2 4L1
30
The preceding equation can be further simplied

Nx 0
xx
u
yy
0 Ny
=
v
xy
Ny Nx
where
NTx |NTy
"
2A
L1
x (4L1 1)
L2
x (4L2 1)
L3
x (4L3 1)
L2
1
4 L
x L2 + 4 x L1
L3
2
4 L
x L3 + 4 x L2
L3
1
4 x L1 + 4 L
x L3

u
(8.22-c)
v
L1
y (4L1 1)
L2
y (4L2 1)
L3
y (4L3 1)
L2
1
4 L
y L2 + 4 y L1
L3
2
4 L
y L3 + 4 y L2
L3
1
4 y L1 + 4 L
y L3
(8.23-b)
(8.24)
u v = u1 u2 u3 u4 u5 u6 v 1 v 2 v 3 v 4 v 5 v 6
(8.25-a)
(8.25-b)
31 This expression is then analytically integrated using Eq. 6.60. In general an isoparametric
numerical integration is preferred (Section 9.2.3).
8.4
Bilinear Rectangular Element
The structure stiness matrix of the bilinear rectangular shown in Fig. 8.1 is described by
the Mathematica code shown below.
32
Define shape functions;

N1[x_,y_,a_,b_] := (1/4)(1-(x/a))(1-(y/b));
N2[x_,y_,a_,b_] := (1/4)(1+(x/a))(1-(y/b));
Victor Saouma
Draft
88
Figure 8.1: Rectangular Bilinear Element

N3[x_,y_,a_,b_] := (1/4)(1+(x/a))(1+(y/b));
N4[x_,y_,a_,b_] := (1/4)(1-(x/a))(1+(y/b));
Visualize the shape functions;
Plot3D[N1[x,y,1,2], {x, -1, 1}, {y, -2, 2}];
Plot3D[N3[x,y,1,2], {x, -1, 1}, {y, -2, 2}];
Define the differential operators;
diffx[f_] := D[f,x];
diffy[f_] := D[f,y];
Shape Function Matrix
Nmat = {
{N1[x,y,a,b],0,N2[x,y,a,b],0,N3[x,y,a,b],0,N4[x,y,a,b],0},
{0,N1[x,y,a,b],0,N2[x,y,a,b],0,N3[x,y,a,b],0,N4[x,y,a,b]}
}
B Matrix
Bmat = {Map[diffx,Nmat[[1]]],Map[diffy,Nmat[[2]]],
Map[diffy,Nmat[[1]]]+Map[diffx,Nmat[[2]]]}
Define the constitutive matrix
Dmat = E/(1-nu^2) *
{{1,nu,0},{nu,1,0},{0,0,(1-nu)/2}}
Determine the stiffness matrix
Intmat = Transpose[Bmat] . Dmat . Bmat;
Kmat =Simplify[ t *
Integrate[ Integrate[Intmat, {y, -b, b}], {x, -a, a}] ]
TeXForm[Kmat]
33
The resulting stiness matrix is

ke18;14 =
Victor Saouma
Draft
8.5 Element Assessment
e (a2 2 b2 +a2 ) t
6 a b (1+) (1+)
(e t)
8 (1+)
e (a2 +4 b2 +a2 ) t
12 a b (1+) (1+)
(e (1+3 ) t)
8 (1+) (1+)
(e (a2 2 b2 +a2 ) t)
12 a b (1+) (1+)
et
8 (1+)
(e (a2 +b2 +a2 ) t)
6 a b (1+) (1+)
e (1+3 ) t
8 (1+) (1+)
ke18;58
89
(e t)
8 (1+)
(e (2 a2 +b2 b2 ) t)
6 a b (1+) (1+)
e (1+3 ) t
8 (1+) (1+)
(e (a2 b2 +b2 ) t)
6 a b (1+) (1+)
et
8 (1+)
e (2 a2 +b2 b2 ) t
12 a b (1+) (1+)
(e (1+3 ) t)
8 (1+) (1+)
e (4 a2 b2 +b2 ) t
12 a b (1+) (1+)
e (a2 +4 b2 +a2 ) t
12 a b (1+) (1+)
e (1+3 ) t
8 (1+) (1+)
e (a2 2 b2 +a2 ) t
6 a b (1+) (1+)
et
8 (1+)
(e (a2 +b2 +a2 ) t)
6 a b (1+) (1+)
(e (1+3 ) t)
8 (1+) (1+)
(e (a2 2 b2 +a2 ) t)
12 a b (1+) (1+)
(e t)
8 (1+)
(e (1+3 ) t)
8 (1+) (1+)
(e (a2 b2 +b2 ) t)
6 a b (1+) (1+)
et
8 (1+)
(e (2 a2 +b2 b2 ) t)
6 a b (1+) (1+)
e (1+3 ) t
8 (1+) (1+)
e (4 a2 b2 +b2 ) t
12 a b (1+) (1+)
(e t)
8 (1+)
e (2 a2 +b2 b2 ) t
12 a b (1+) (1+)
(e (a2 2 b2 +a2 ) t)
12 a b (1+) (1+)
et
8 (1+)
(e (a2 +b2 +a2 ) t)
6 a b (1+) (1+)
e (1+3 ) t
8 (1+) (1+)
e (a2 2 b2 +a2 ) t
6 a b (1+) (1+)
(e t)
8 (1+)
e (a2 +4 b2 +a2 ) t
12 a b (1+) (1+)
(e (1+3 ) t)
8 (1+) (1+)
et
8 (1+)
e (2 a2 +b2 b2 ) t
12 a b (1+) (1+)
(e (1+3 ) t)
8 (1+) (1+)
e (4 a2 b2 +b2 ) t
12 a b (1+) (1+)
(e t)
8 (1+)
(e (2 a2 +b2 b2 ) t)
6 a b (1+) (1+)
e (1+3 ) t
8 (1+) (1+)
(e (a2 b2 +b2 ) t)
6 a b (1+) (1+)
(e (a2 +b2 +a2 ) t)

6 a b (1+) (1+)
(e (1+3 ) t)
8 (1+) (1+)
(e (a2 2 b2 +a2 ) t)
12 a b (1+) (1+)
(e t)
8 (1+)
e (a2 +4 b2 +a2 ) t
12 a b (1+) (1+)
e (1+3 ) t
8 (1+) (1+)
e (a2 2 b2 +a2 ) t
6 a b (1+) (1+)
et
8 (1+)
e (1+3 ) t
8 (1+) (1+)
e (4 a2 b2 +b2 ) t
12 a b (1+) (1+)
(e t)
8 (1+)
e (2 a2 +b2 b2 ) t
12 a b (1+) (1+)
(e (1+3 ) t)
8 (1+) (1+)
(e (a2 b2 +b2 ) t)
6 a b (1+) (1+)
et
8 (1+)
(e (2 a2 +b2 b2 ) t)
6 a b (1+) (1+)
(8.26-a)
8.5
8.5.1
Element Assessment
CST
34 The CST gives good results in a region of the FE model where there is little strain gradient.
Otherwise, the CST will not work well.
35 A natural extension would be to go linear, i.e. LST. This element would have midside nodes
in addition ot vertex nodes.
36 We should note that for this element the stress is independent of x and y because a linear
displacement relation was assumed resulting in a constant strain and stress (for linear elastic
material).
37 If we consider the element shown in Fig. 8.2 we apply a moment M such that: 1) origin does
not move, and 2) the displacement is u = u at x = l on top of the beam, then from the exact
solution (Timoshenko and Goodier 1970), the correct displacements are
u=
Victor Saouma
u
xy
cl
v=
u
(x2 y 2 )
2cl
(8.27)
Draft
810
Figure 8.2: Linear Triangular Element Subjected to Pure Bending

dierentiating, we recover the anticipated strain eld
xx =
u
y
cl
yy =
u
y
cl
xy = 0
(8.28)
38 Imposing nodal displacements consistent with theory (u1 = u, u2 = u, and u3 = 0) and

similar descriptions for v, and determining the strains from = Bu for the CST element, we
obtain

c
1
2 u
yy = 0
xy =
(8.29)
xx = 0
c
4l
which are clearly wrong. Hence, we conclude the CST element can not properly perform in
bending problems (As is expected from a constant strain element behavior when used in a linear
strain problem).
This deciency will later be addressed by including an additional internal drilling degree
of freedom.
39
40
Zero energy mode can be veried by considering for example the strain xx which is given by
xx =
1
(y3 u1 + y3 u2 )
x2 y3
(8.30)
since u1 = u2 for a rigid body motion, we see that xx = 0, similar conditions can be found for
yy and xy
8.5.2
41
BiLinear Rectangular
Under exure, the top and bottom sides remain straight giving rise to erroneous strains.
42 From Pascalss triangle (Sect. 6.4), the displacement eld for a bilinear rectangular element
is given by
(8.31)
u = a1 + a2 x + a3 y + a5 xy
which is consistent with the element shape functions derived earlier (Eq. 6.31)
43 We note that whereas there is a quadratic term (a5 xy) convergence is governed by the full
linear expansion, and that this last term may cause parasitic eects.
Victor Saouma
Draft
8.5 Element Assessment
811
44 For an element with an inclined boundary given by y = ax + b, insertion of this expression

into Eq. 8.31 yields
(8.32)
u = a1 + a3 b + (a2 + a3 a + a5 b)x + a5 ax2
hence, u varies quadratically with x along the inclined boundary. However, this boundary is
uniquely dened by only two nodal displacements u1 and u2 , thus the quadratic expression is
not uniquely dened and we may have dierent variation along the side from one element to
the adjacent one.
45
Continuity across the element not being satised, the element is non-conforming.
We conclude that the bilinear quadrilateral element must always have its four edges parallel
to the coordinate system.
46
47 This severe restriction will be alleviated in the next lecture by Isoparametric element
formulations.
Renement of this element include the addition of two degrees of freedom to allow a state of
constant curvature
48
u = ni=1 Ni ui + (1 2 )g1 + (1 2 )g2
(8.33-a)
+ (1 )g3 + (1 )g4
(8.33-b)
v =
ni=1 Ni v i
Other renements (mostly in shells) include the inclusion of additional rotational degrees of
freedom termed drilling dof.
49
Victor Saouma
Draft
812
Victor Saouma
Draft
Chapter 9
ISOPARAMETRIC ELEMENTS;
2nd GENERATION
9.1
Introduction
1 We have previously examined simple nite elements, in this lecture we shall distort those
simpler elements into others of more arbitrary shape.
Correspondingly, the natural coordinates will be distorted into new curvilinear sets when
plotted in a cartesian x, y, z space, Fig. 9.1.
2
(-1,1)
(1,1)
(-1,-1)
(1,-1)
3
L 1 =0
L 2 =0
L 3 =0
Figure 9.1: Two-Dimensional Mapping of Some Elements

In the isoparametric formulation, displacements are expressed in terms of natural coordinates, however they must be dierentiated with respect to cartesian coordinates x, y, z. This is
accomplished through a transformation matrix J, and integration can no longer be performed
analytically but must be done numerically.
3
Natural coordinates range from -1 to +1, Fig. 9.2
Nodal displacements at any point inside the element can be written in terms of the nodal
Draft
ISOPARAMETRIC ELEMENTS; 2nd GENERATION
92
y,v
Mapping
x,u
(1,1)
(-1,1)
(-1,-1)
(1,-1)
Parent Element
Actual Element
Figure 9.2: Actual and Parent Elements

known displacements and the shape functions
u
2
= Nue
u = N1 u1 + N2 u2 + = N
..
.
v1
v2
= Nve
v = N1 v 1 + N2 v 2 + = N
..
.
w1
w2
= Nwe
w = N1 w 1 + N2 w2 + = N
..
.
(9.1)
or
u = u v w T = [N]ue
(9.2)
When elements are also distorted, the coordinates of any point can also be expressed in terms
of nodal coordinates
2
= Nx
x = N1 x1 + N2 x2 + = N
..
.
y1
y2
2 y 2 + = N
1 y 1 + N
(9.3)
= Ny
y = N
..
.
z1
z2
= Nz
z = N1 z 1 + N2 z 2 + = N
..
.
or
c = x y z T = [N]c
(9.4)
are functions of , , and . Hence, three possible cases

Shape function matrices [N] and [N]
arise, Fig. 9.3
7
Victor Saouma
Draft
9.2 Element Formulation
93
Geometric
Nodal
Config.
Displacement
Nodal
Config.
Isoparametric
Subparametric
Superparametric
Figure 9.3: Iso, Super, and Sub Parametric Elements

are identical
Isoparamteric: [N] and [N]
Subparamteric: [N] is of higher degree than [N]
Superparamteric: [N] is of lower degree than [N]

8
Sub and Super parametric elements are very seldom used.
9.2
9.2.1
Element Formulation
Bar Element
9 The simplest introduction to isoparamteric elements is through a straight three noded quadratic
elements, Fig. 9.4.
= 1
= 0
= +1
x,u
L
Figure 9.4: Three-Noded Quadratic Bar Element

10 The shape functions for the element can be obtained from the Lagrangian interpolation
function used earlier, Eq. 6.25, and in which we substitute x by . The kth term in a polynomial
Victor Saouma
Draft
94
of order n 1 would be
Nkn
-n
( i )
-ni=1,i =k
i=1,i =k (k i )
(9.5-a)
( 1 )( 2 ) ( k1 )( k+1 ) ( n )
(k 1 )(k 2 ) (k k1 )(k k+1 ) (k n )
(9.5-b)
For a three noded quadratic element 1 = 1, 2 = +1, and 3 = 0. Substituting, we obtain

the three shape functions
N1 () =
N2 () =
N3 () =
(2 )(3 )
(1 2 )(1 3 )
(1 )(3 )
(2 1 )(2 3 )
(1 )(2 )
(3 1 )(3 2 )
=
=
=
(1)(0)
(11)(10)
(+1)(0)
(1+1)(10)
(+1)(1)
(0+1)(01)
=
=
1 2
2 (
1 2
2 (
)
+ )
(9.6)
= 1 2
Hence,
x() = N x1 x2 x3 T
and
u() = N u1 u2 u3 T
(9.7)
where
N =
1 2
2 (
1 2
2 (
+ ) 1 2
(9.8)
The strain displacement relation is given by Eq. 7.4, = Lu = LNue = Bue , and the
dierential operator L is equal to d . For this one dimensional case, this reduces to
dx
u1
d
du
=
N
u
(9.9)
x =
2
dx
dx
u3
L
11
12 We invoke the chain rule since the shape functions are expressed in terms of natural coordinates:
dN d
dN
=
(9.10)
B=
dx
d dx
The rst term may be readily available from the shape functions, Eq. 9.8, however the second
one is not.
d is not available, we may determine its inverse dx , from Eq. 9.7, which we shall
13 Whereas,
dx
d
denote by J or Jacobian. The Jacobian operator J is a scale factor which relates cartesian to
natural coordinates dx = Jd.
x1
x1
d
dx
x2
x2
=
N
(9.11)
= 1 (2 1) 12 (2 + 1) 2
J() =
2
d
d

x3
x3
dN
d

dx
d
14
We can rewrite Eq. 9.10 as

B=
Victor Saouma
d dN
dN
=
(9.12)
dx
dx d

J 1
Draft
95
and the B matrix is thus obtained by substituting into Eq. 9.9

B() =
15
1
1 d
N =
J d
J
1
2 (2
1)
1
2 (2
+ 1) 2
The dierential area is

d = Adx = AJd
16
(9.13)
(9.14)
Substituting, the element stiness matrix is nally obtained from Eq. 7.8

L
T
Ke () =
+1
B ()AEB()dx =
BT ()AEB()J()d
(9.15)
17 We observe that B, in general, contains terms in both the numerator and denominator, and
hence the expression can not be analytically inverted. Furthermore, the limits of integration
are now from -1 to +1, and we shall see later on how to numerically integrate it.
18 A simple Mathematica code to generate the stiness matrix of three noded (quadratic)
element:
x1=0;x2=L;x3=L/2;
N1[x_,L_]:=(x-x2) * (x-x3)/( (x1-x2) * (x1-x3));
N2[x_,L_]:=(x-x1) * (x-x3)/( (x2-x1) * (x2-x3));
N3[x_,L_]:=(x-x2) * (x-x1)/( (x3-x2) * (x3-x1));
BMat={D[N1[x,L],x], D[N2[x,L],x], D[N3[x,L],x]};
Integrate[EA Outer[Times, BMat, BMat], {x,0,L}];
MatrixForm[%]
9.2.2
9.2.2.1
Quadrilaterals
Linear Element (Q4)
19 We have previously derived the stiness matrix of a rectangular element (aligned with the
coordinate axis), this formulation will generalize it to an arbitrary quadrilateral shape.
20
Fig. 9.5 illustrates the element
= 1
= 1/2
= 1
= 1/2
= 1/2
= 1
4
1
4
1
3
1
= 1/2
y,v
= 1
1
1
x,u
Figure 9.5: Four Noded Isoparametric Element
Victor Saouma
Draft
96
21
For the two-dimensional case

u(, ) =
Nij uk =
Nj ()Ni ()uk
(9.16)
i=1 j=1
where k = (i 1)m + j. For a bilinear element, n = m = 2, this can be rewritten as

u1 u3
N1 ()
u(, ) = N1 () N2 ()
= NT uN
u2 u4
N2 ()
= N1 ()N1 ()u1 + N2 ()N1 ()u2 + N1 ()N2 ()u4 + N2 ()N2 ()u3
= N1 (, )u1 + N2 (, )u2 + N3 (, )u3 + N4 (, )u4
4
Ni ui
=
(9.17-a)
(9.17-b)
i=1
22
Applying the Lagrangian interpolation equation, Eq. 9.5-a we obtain

N1 () =
N2 () =
N1 () =
N2 () =
( 1)
1
( 2 )
=
= (1 )
(1 2 )
(1 1)
2
( + 1)
1
( 1 )
=
= (1 + )
(2 1 )
(1 + 1)
2
( 1)
1
( 2 )
=
= (1 )
(1 2 )
(1 1)
2
( + 1)
1
( 1 )
=
= (1 + )
(2 1 )
(1 + 1)
2
(9.18-a)
(9.18-b)
(9.18-c)
(9.18-d)
Substituting into Eq. 9.17-a

N1 (, ) =
N3 (, ) =
1
4 (1
1
4 (1
)(1 ); N2 (, ) =
+ )(1 + ); N4 (, ) =
1
4 (1
1
4 (1
+ )(1 );
)(1 + );
(9.19)
It should be noted that for this simple case, the shape functions could have been determined
by mere inspction.
23
24
Coordinates and displacements are given by

)
)
x = ) Ni (, )xi ; y = ) Ni (, )y i
u =
Ni (, )ui ; v =
Ni (, )v i
The strain displacement relation is given by Eq. 7.4

xx
x
0
yy
=
{} =
y
xy
y
x

L
(9.20)

u
v

(9.21)
However the displacements can be obtained from Eq. 9.20

u
v
| {z }
u

=
Victor Saouma
N1 (, )
0
0
N1 (, )
N2 (, )
0
8 u 9
1 >
>
>
>
>
v1 >
>
>
>
>
>
>
>
u2 >
>
>
>
< v >
=
0
N3 (, )
0
N4 (, )
0
2
(9.22)
N2 (, )
0
N3 (, )
0
N4 (, )
u
>
>
3
>
>
>
>
{z
}>
v3 >
>
>
>
>
>
>
>
N
>
>
: u4 >
;
v4
Finite Elements II; Solid
Mechanics
| {z }
u
Draft

25
97
Combining Eq. 9.21 and 9.22 yields

= LNu = Bu
Ni (,)
4
xx (, )
x
0
yy (, )
=
N
i (,)
i=1
xy (, )
y

Ni (,)
y
Ni (,)
x
B=LN
=
N2,x
0
N2,y
0
N2,y
N2,x

ui
vi

(9.23-a)
N3,x
0
N3,y
0
N3,y
N3,x
N4,x
0
N4,y
0
N4,y
N4,x
0
N1,y
N1,x
N1,x
0
N1,y
Ni (,)
Ni (,)
Ni (,)
i=1
Ni (,)
u1
v1
u2
v2
u3
v3
u4
v4
$
(9.23-b)

[J]1
26 Considering the local set of coordinates , and the corresponding global one x, y, the chain
rules would give
#
$

Ni
Ni
Ni
x
Ni
y
=

J

= [J]
Ni
x
Ni
y
(9.24-a)
Ni
Ni
(9.24-b)
This last equation is the key to get all the components which will go inside the B matrix.
27
Expanding the denition of the Jacobian

#
$
N (,)
i
Ni (,)
=

#
=
1
4

J
N1
N1
N2
N2
Ni
x
Ni
y
N3
N3
i=1
N4
N4
Ni
xi
Ni
xi
x1
x2
x3
x4
y1
y2
y3
y4
Ni
x
Ni
y
(9.25-a)
Ni
x
Ni
y
(1 ) (1 ) (1 + ) (1 + )
(1 ) (1 + ) (1 + ) (1 )
$
Ni
y i
Ni
y i
(9.25-b)
x1
x2
x3
x4
y1
y2
y3
y4
Ni
x
Ni
y

(9.25-c)
Victor Saouma
Draft
98
28
Back to the Jacobian

#
$
[J]1 =
def
29
1
=
detJ
=
detJ
4
i=1
Ni
y i
i
N
xi
i
N
y i
Ni
xi
$
(9.26)
From calculus, if and are some arbitrary curvilinear coordinates, Fig. 9.6, then
y
y
d
ds
y
d
de
dr

x
d
x
d
Figure 9.6: Dierential Element in Curvilinear Coordinate System

dr =

d
and
ds =
(9.27)
are vectors directed tangentially to = constant, and = constant respectively.

30
From vector algebra, the cross product of two vectors lying in the x-y plane, Fig. 9.7 is
C=A x B
|B| sin
Figure 9.7: Cross Product of Two Vectors

C = AB
= |A||B| sin k
Victor Saouma
(9.28-a)
(9.28-b)
Draft
99

i
j
k

= Ax Ay 0 = (Ax By Bx Ay ) k

Bx By 0
Area
|C| = |A||B| sin
(9.28-c)
(9.28-d)
hence, the dierential area dxdy is then equal to the length of the vector resulting from the
cross product of drds and is equal to
#
d(area) = dxdy = det
31
dd

(9.29)
Finally determine the element stiness matrix from

[k]88 =

[B]T83 [D]3x3 [B]38 tdxdy
= [k] =
1
[B]T [D][B]t|J|dd
(9.30)
The evaluation of the element stiness matrix involves dA. If we consider an innitesimal
element, of length dr and ds, at the vertex of an element, it has the boundaries of the element
as its sides. Then, from Eq. 9.28-d
32
dA = dx.dy. sin
(9.31)
however, from Eq. 9.29 we have dA = detJd.d, thus

detJ =
dx.dy
sin
d.d
(9.32)
Thus we observe that if is small or close to 180o , then det J will be very small, if the angle is
greater than 180 o , the determinant is negative (implying a negative stiness which will usually
trigger an error/stop in a FE analysis).
33
In general it is recommended that 30o < < 150o .
34 The inverse of the jacobian exists as long as the element is not much distorted or folds back
upon itself, Fig. ?? in those cases there is no unique relation between the coordinates.
Figure 9.8: *Elements with Possible Singular Jacobians

35 It can be easily shown that for parallelograms, the Jacobian is constant, whereas for nonparallelograms it is not.
In general J is an indicator of the amount of element distorsion with respect to a 2x2 square
one. Some times it is constant, others it varies within the element.
36
Victor Saouma
Draft
910
9.2.2.1.1 Example: Jacobian Operators, (Bathe 1996) Determine the Jacobian operators J for the following 2 dimensional elements. (Bathe 1982).
y
2
1
1
x
4
3
3/4
3
4
2
x 1
60
The coordinates are given by Eq. 9.4, the shape functions by Eq. 9.19, and the Jacobian
by Eq. 9.24-b.
Element 1:
x =
y =
[J] =
1
1
(1 )(1 )x3 + (1 + )(1 )x4
4
4
1
1
+ (1 + )(1 + )x1 + (1 )(1 + )x2
4
4
1
1
(1 )(1 )(3) + (1 + )(1 )(3)
4
4
1
1
+ (1 + )(1 + )(3) + (1 )(1 + )(3)
4
4
1
1
(1 )(1 )y 3 + (1 + )(1 )y 4
4
4
1
1
+ (1 + )(1 + )y 1 + (1 )(1 + )y 2
4
4
1
1
(1 )(1 )(2) + (1 + )(1 )(2)
4
4
1
1
+ (1 + )(1 + )(2) + (1 )(1 + )(2)
4
4

3 0
0 2
(9.33-a)
(9.33-b)
(9.33-c)
(9.33-d)
(9.33-e)
We note that A = 24 = det[J](2 2) = 6 4

Element 2:
1
1
(1 )(1 )((3 + 1/(2 3))) + (1 + )(1 )(3 1/2 3))
4
4
1
1
+ (1 + )(1 + )(3 + 1/2 3)) + (1 )(1 + )((3 1/2 3)))
4
4
1
1
(1 )(1 )(2) + (1 + )(1 )(2)
y =
4
4
x =
Victor Saouma
(9.34-a)
Draft
[J] =
911
1
1
+ (1 + )(1 + )(2) + (1 )(1 + )(2)
4
$
#4
3
0
1
2 3
(9.34-b)
(9.34-c)
1
2
Element 3:
1
1
(1 )(1 )(1) + (1 + )(1)
4
4
1
1
+ (1 + )(1 + )(1) + (1 )(1)
4
4
1
1
y =
(1 )(1 )(3/4) + (1 + )(1 )(3/4)
4
4
1
1
+ (1 + )(1 + )(5/4) + (1 )(1 + )(1/4)
4
4

1 4 (1 + )
[J] =
4 0 (3 + )
x =
9.2.2.2
37
(9.35-a)
(9.35-b)
(9.35-c)
Quadratic Element
For a quadratic quadrilateral element, there are two possibilities, Fig. 9.9.
9
6
1
5
2
5
2
Figure 9.9: Serendipity and Lagrangian Quadratic Quadrilaterals

38 The Pascal triangle, Fig.
9.10, will be later used to justify the choice of terms in the
displacement eld of isoparameteric elements.
9.2.2.2.1
39
Serendipity Element (Q8)
Based on Pascals triangle, the displacement eld is given by

u = a1 + a2 x + a3 y + a4 x2 + a5 xy + a6 y 2 + a8 x2 y + a9 xy 2

0
1
2
3
(9.36)
In this formulation, the x2 y 2 term is missing and the 8 terms in the assumed polynomial
expansion correspond the 8 nodes (4 corner and 4 midside).
40
Victor Saouma
Draft
912
Constant term
Linear terms
2
Quadratic terms
3
Cubic terms
Quartic terms
Linear element
Quadratic element
3
Cubic element
Serendipity Quadrilateral Elements

1
Constant term
Quadratic terms
Cubic terms
Linear terms
3
3
Quartic terms
2
2 2
3 2
Quintic terms
Linear element
Quadratic element
3
Cubic element
3
2 3
3 3
Sixtic terms
Lagrangian Quadrilateral Elements

1
Constant term
Quadratic terms
Cubic term
Linear terms
Linear elements
Quadratic element
3
Cubic element
Triangle Elements
Figure 9.10: Pascal Triangles for Quadrilateral and Triangle Elements
Victor Saouma
Draft
913
y,v
1
6
1
5
5
1
x,u
2
1
Figure 9.11: Serendipity Isoparametric Quadratic Finite Element: Global and Parent Element
41
The shape functions may be obtained by mere inspection (i.e. serependitiously),

Ni =
Ni =
Ni =
1
4
1
2
1
2
(1
% + 2i&) (1 + i ) (i + i 1) i = 1, 2, 3, 4
i = 5, 7
1 (1
% + i&)
(1 + i ) 1 + 2
i = 6, 8
(9.37)
and are tabulated in Table 9.1.

i
Ni
1
2
3
4
5
6
7
8
1
4 (1 )(1 )( 1)
1
4 (1 + )(1 )( 1)
1
4 (1 + )(1 + )( + 1)
1
4 (1 )(1 + )( 1)
1
2
2 (1 )(1 )
1
2
2 (1 + )(1 )
1
2
2 (1 )(1 + )
1
2
2 (1 )(1 )
Ni,
+ )(1 )
)(1 )
+ )(1 + )
)(1 + )
(1 )
1
2
2 (1 )
(1 + )
12 (1 2 )
1
4 (2
1
4 (2
1
4 (2
1
4 (2
Ni,
)(2 + )
+ )(2 )
+ )(2 + )
)(2 )
12 (1 2 )
(1 + )
1
2
2 (1 )
(1 )
1
4 (1
1
4 (1
1
4 (1
1
4 (1
Table 9.1: Shape Functions, and Natural Derivatives for Q8 Element
42
The shape functions for the corner and midisde nodes are shown in Fig. 9.12.
9.2.2.2.2
Lagrangian element (Q9)
43 If we were to follow a similar procedure to the one adopted to extract the bilinear shape
fuctions, we would obtain 9 shape functions, which must in turn correspond to 9 (rather than
Victor Saouma
Draft
914
1
0.8
0.5
0.6
0
0.4
0.5
0.2
1
1
0
1
1
0.5
0.5
0.5
0
0
0.5
0.5
0.5
1
0.5
0
0.5
1
1
0
0.1
0.2
0.8
0.8
0.6
0.6
0.6
0.4
0.4
0.4
0.2
0.2
0
0.2
0.2
0.4
0.7
0.2
0.3
0.8
0.4
0.4
0.6
0.6
0.2
0.4
0.8
0.8
1
1
0.5
0.9
0.8
0.6
0.8
0.6
0.4
0.2
0.2
0.4
0.6
0.8
1
1
0.8
0.6
0.4
0.2
0.2
0.4
0.6
0.8
Figure 9.12: Shape Functions for 8 Noded Quadrilateral Element

8) nodes.
44
In this element, the displacement eld is given by

u = a1 + a2 x + a3 y + a4 x2 + a5 xy + a6 y 2 + a8 x2 y + a9 xy 2 + a12 x2 y 2

0
1
2
3
4

(9.38)
Serendipity
45 All the quadratic terms are present, hence there are 9 terms in the polynomial expansion,
and the 9th node will correspond to an internal node.
46 The shape functions in this case can be directly obtained from the Lagrangian interpolation
functions, yielding
Ni
Ni
Ni
N9
= 14 (1
% + i&) (1 + i ) (i + i 1)
= 12 1 2 (1
% + 2i&)
1
= 2 (1 + i ) 1
= (1 2 )(1 2 )
i = 1, 2, 3, 4
i = 5, 7
i = 6, 8
i=9
47
The last shape function is often called bubble function
48
Those shape function dier slightly from those of the serendipity element.
Victor Saouma
(9.39)
Draft
915
49
Q9 elements perform much better than the Q8 if edges are not parallel or slightly curved.
50
The shape functions for the corner and midisde nodes are shown in Fig. 9.13.
1
1
0.5
0.8
0.6
0
0.4
0.2
0.5
1
1
0.2
1
0.5
0.5
0.5
1
0.5
0.5
0.5
1
0.8
0.8
0.6
0.6
0.4
0.6
0.4
0.6
0.2
0.5
0.5
0.2
0.2
0
0.4
0.2
0.2
0.8
0.4
0.4
0.4
0.6
0.8
0.8
1
1
0.8
0.8
0.2
0.6
0.6
0.2
0.4
0.6
0.4
0.2
0.2
0.4
0.6
0.8
1
1
0
0.8
0.6
0.4
0.2
0.2
0.4
0.6
0.8
0.1
0.8
0.4
0.6
0.6
0.2
0.8
0.4
0.8
0.6
0.2
0.4
0.2
0.2
0.9
0.4
0
1
0.7
0.5
0.5
0
0
0.5
0.5
1
0.5
0.6
0.3
0.8
1
1
0.8
0.6
0.4
0.2
0.2
0.4
0.6
0.8
Figure 9.13: Shape Functions for 9 Noded Quadrilateral Element
9.2.2.2.3 Variable (Hierarchical) Element 51 Based on the above, we can generalize

the formulation to one of a quadrilateral element with variable number of nodes.
52 This element may have dierent order of variation along dierent edges, and is quite useful
to facilitate the grading of a nite element mesh.
53
In its simplest formulation, it has four nodes, and has a linear variation along all sides, and
Victor Saouma
Draft
916
in the most general case it is a full quadratic element.

The shape functions may than be obtained from Table 9.2. Note that these shape functions
are for the hierarchical element in which the corner nodes are numbered rst, and midside ones
after, Fig. 9.14.
54
Figure 9.14: Nodal Numbering for Isoparameteric Elements
N1
N2
N3
N4
N5
N6
N7
N8
N9
1
4 (1 )(1 )
1
4 (1 + )(1 )
1
4 (1 + )(1 + )
1
4 (1 )(1 + )
1
2
2 (1 )(1 )
1
2
2 (1 + )(1 )
1
2
2 (1 )(1 + )
1
2
2 (1 )(1 )
(1 2 )(1 2 )
i=5
12 N5
12 N5
Only if node i is present

i=6
i=7
i=8
12 N8
1
2 N6
12 N6 12 N7
12 N7 12 N8
i=9
1
4 N9
1
4 N9
1
4 N9
1
4 N9
12 N9
12 N9
12 N9
12 N9
Table 9.2: Shape Functions for Variable Node Elements
9.2.3
Triangular Elements
For the six noded triangle, the partial derivatives of a variable with respect to x and y can
be expressed as, (Felippa 2000)
#
$ )6 i Ni

)i=1 L1
L2
L3
L1
6
Ni
x
x
x
x
= L
(9.40)
L
L
i=1 i L2
1
2
3
)
6
y
y
y
Ni
y
i=1 i L3
55
56
Transposing both sides
)6
Ni
i=1 i L1
Victor Saouma
)6
Ni
i=1 i L2
)6
Ni
i=1 i L3
L1
x
L2
x
L3
x
L1
y
L2
y
L3
y
=
(9.41)
Draft
9.3 Numerical Integration

57
917
We now make 1, x, and y:

)6
)6 Ni
Ni
i=1
i=1 L2
L
1
)6
Ni )6
i
x
xi N
i
L1
L2
)i=1
)i=1
6
6
Ni
Ni
i=1 yi L1
i=1 yi L2
)6 Ni
3
)6 i=1 L
Ni
x
i
L3
)i=1
6
Ni
i=1 yi L3
L1
x
L2
x
L3
x
L1
y
L2
y
L3
y
1
x
x
x
y
x
1
y
x
y
y
y
(9.42)
y
y
1
1
x
But x
x = y = 1, and x = y = y = x = 0 since x and y are independent coordinates.
Furthermore all entries in the rst row are equal to a constant (3 for the T6 element), and since
the corresponding right hand side, this row can be normalized, yielding the Jacobian matrix
for this element
L

L1
1
1
1
1
0 0
x
y
)
)
)
6
6
6
N
N
N L
L
(9.43)
)i=1 xi L1i )i=1 xi L2i )i=1 xi L3i x2 y2 = 1 0
6
6
6
Ni
Ni
Ni
L3
L3
0
1
y
y
y
i=1 i L1
i=1 i L2
i=1 i L3
x
y

58
59
Substituting from Table 8.1 we obtain

2
1
4 x1 (4L1 1) + 4x4 L2 + 4x6 L3
y1 (4L1 1) + 4y4 L2 + 4y6 L3
1
x2 (4L2 1) + 4x5 L3 + 4x4 L1
y2 (4L2 1) + 4y5 L3 + 4y4 L1
2
6
4
L1
x
L2
x
L3
x
L1
y
L2
y
L3
y
{z
J
3
0
7=4 1
5
0
1
x3 (4L3 1) + 4x6 L1 + 4x5 L2 5
y3 (4L3 1) + 4y6 L1 + 4y5 L2
(9.44)
0
0 5
1
60 Next we invert the matrix and solve for the six triangular coordinates partials and substitute
in Eq. 9.40 which in turn will enable us to determine the B matrix in Eq. 9.23-a
Ni (L1 ,L2 ,L3 )
0

6
xx (L1 , L2 , L3 )
x
ui
Ni (L1 ,L2 ,L3 )

0
yy (L1 , L2 , L3 )
=
(9.45)
y
vi
N
N
(L
,L
,L
)
(L
,L
,L
)
1
2
3
1
2
3
i
i
i=1
xy (L1 , L2 , L3 )

y
x
u

B=LN
9.3
Numerical Integration
Understanding numerical integration is not only essential for a proper integration of the
isoparameteric family of elements, but also helpful in understanding the Weighted Residual
methods (Chapter 11),
61
A crucial aspect of isoparametric element formulation is the numerical integration which can
be expressed as

62
F()d
63
or
F(, )dd
(9.46)
In practice we perform the integration numerically using

Wi F(i ) + Rn or
F(, )dd =
Wij F(i , j ) + Rn
F()d =
i
Victor Saouma
(9.47)
Draft
918
where the summations extend over all i and j, and Wi , Wij are weighting factors, and F(i )
and F(i , j ) are the matrices evaluated at the points specied in the arguments.
The matrices Rn are error matrices, which are in general not computed.
64
As shown above, F = BT .D.B for nite element stiness matrix evaluation, and each element
is integrated individually.
b
66 The integration of
is essentially based on passing a polynomial P () through given
a F ()d
b
values of F () and then use a P ()d as an approximation.
65
F ()d
(9.48)
P ()d
a
Using P () = F () at n points, and recalling the properties of Lagrangian interpolation

functions, we obtain
67
P () = l1 ()F (1 ) + l2 ()F (2 ) + + ln ()F (n )

n
li ()F (i )
=
(9.49-a)
(9.49-b)
i=1
and note that at = i li = 1, while all other li = 0.
9.3.1
Newton-Cotes
68 In Newton-Cotes integration, it is assumed that the sampling points are equally spaced,
Fig. 9.15, thus we dene
P( )
F( )
-1
Figure 9.15: Newton-Cotes Numerical integration

P ()d =
or
Approximation
Weights
b
a
li ()dF (i )
(9.50)
)n
(n)
P ()d =
Wi F (i )
i=1
b
(n)
(n)
Wi
= a li ()d = (b a)Ci
(9.51)
a i=1
li ()F (i )d =
i=1
(n)
where Ci are the weights of the Newton-Cotes quadrature for numerical integration
with n equally spaced sampling points.
Victor Saouma
Draft

n
2
3
4
5
(n)
C0
1
2
1
6
1
8
7
90
919
(n)
C1
1
2
4
6
3
8
32
90
(n)
C2
1
6
3
8
12
90
(n)
C3
1
8
32
90
(n)
C4
7
90
Error
101 (b a)3 F II ()
103 (b a)5 F IV ()
103 (b a)5 F IV ()
106 (b a)7 F V I ()
Table 9.3: Weights for Newton-Cotes Quadrature Formulas

69
Newton-Cotes constants, and corresponding reminder are shown in Table 9.3, (Bathe 1982).
70 It can be shown that this method permits exact integration of polynomial of order n 1,
and that if n is odd, then we can exactly integrate polynomials of order n. Hence we use in
general odd values of n,
711 For n = 2 over [1, 1], we select equally spaced points at 1 = 1 and 2 = 1 to evaluate
1 P ()d
P () =
li ()F (i )
(9.52-a)
i=1
l1 () =
l2 () =
(2)
(2)
W1
W2

1
1
F ()d
2
1
= (1 )
1 2
2
1
1
= (1 + )
2 1
2

1
1 1
l1 ()d =
(1 )d = 1
2 1
1

1
1 1
l2 ()d =
(1 + )d = 1
2 1
1
1
2
(2)
P ()d =
Wi F (i ) = F (1) + F (1)
1
(9.52-b)
(9.52-c)
(9.52-d)
(9.52-e)
(9.52-f)
i=1
which is the trapezoidal rule

For n = 3 over [1, 1], we select equally spaced points at 1 = 1 2 = 0, and 3 = 1, to
1
evaluate 1 P ()d
72
P () =
li ()F (i )
(9.53-a)
i=1
1
( 2 )( 3 )
= ( 1)
(1 2 )(1 3 )
2
( 1 )( 3 )
= (1 + )( 1)
l2 () =
(2 1 )(2 3 )
1
( 1 )( 2 )
= (1 + )
l3 () =
(3 1 )(3 2 )
2
1
1
1
1
(3)
=
l1 ()d =
( 1)d =
W1
2 1
3
1
l1 () =
Victor Saouma
(9.53-b)
(9.53-c)
(9.53-d)
(9.53-e)
Draft
920
(3)
W2
1
1
1
1
(3)
W3

l2 ()d =
1
2
l3 ()d =
P ()d =
1
(1 + )d =
1
i=1
4
3
(1 + )( 1)d =
1
F ()d
(9.53-f)
1
3
(9.53-g)
1
(3)
Wi F (i ) = [F (1) + 4F (0) + F (1)] (9.53-h)
3
which is Simpsons rule
9.3.2
Gauss-Legendre Quadrature
73 In Gauss-Legendre quadrature, the points are not xed and equally spaced, but are selected
to achieve best accuracy.
74
Again we start with

F ()d
P ()d =
1
(n)
(n)
(9.54)
Wi F (i )
i=1
and i are unknowns to be yet determined. Thus, we
however in this formulation both Wi

have a total of 2n unknowns.
At the integration points P (i ) = F (i ), however at intermediary points the dierence can

be expressed as
(9.55)
F () = P () + ()(0 + 1 + 2 2 + )

0 at =i ;i=1,2,,n
75
since we want the left side to be exactly equal to P () at the integration points, we dene
() = ( 1 )( 2 )...( n )
(9.56)
as a polynomial of order n, to be equal to zero at the integration points i .

76 i should be appropriately selected in orer to eliminate the gap between F () and P () at
intermediary points.
77
Integrating
F ()d =
P () +
j=0
() j d
(9.57)
We split the last term
j=0
78
() d =
1
n1
j=0
() d +
1
j=n
() j d
(9.58)
Truncating the last terms of the expansion

1
1
Victor Saouma

F ()d
P () +
1
n1
j=0
() j d
(9.59)
Draft
921
we observe that the rst integral on the right-hand side involves a polynomial of order n 1,
and the second integral a polynomial of order 2n 1. Thus we set
1
() j d = 0
j = 0, 1, , n 1
(9.60)
1
which would result in a set of n simultaneous equations of order n in terms of the unknowns
i , i = 0, 1, , n 1.
79
Back to Eq. 9.59

1
1

F ()d
P () =
1
Approximation
Weights
Gauss Points

F (i )
i=1
or
80
1
1 F ()d
(n)
Wi
1
j
1 () d
li ()d =
(n)
Wi F (i )
(9.61)
i=1
)n
(n)
W F (i )
1i=1 i
= 1 li ()d
= 0
j = 0, 1, , n 1
(n)
(9.62)
are given in Table 9.4, (Bathe 1982).

(n)
i
0
3
1/
1/
0
03
3/5 0
3/5
Integration points i and weight coecients Wi

n
1
2
3
Wi
2
1 1
5/9 8/9 5/9
Error
1 (2)
()
3F
1
(4)
()
135 F
1
(6) ()
F
15,750
Table 9.4: Integration Points and Weights for Gauss-Quadrature Formulaes Over the Interval
[1, 1]
9.3.2.1
Legendre Polynomial
81 The solutions (Gauss integration points) are equal to the roots of a Legendre polynomial
dened by
L0 () = 1
L1 () =
(9.63)
2k1
k1
2kn
Lk () =
k Lk1 () k Lk2 ()
and the n Gauss integration points are determined by solving Ln () = 0 for its roots i ,
i = 0, 1, , n 1.
82
The weighting functions are then given by

(n)
Wi
Victor Saouma
2(1 i2 )
[nLn1 (i )]2
(9.64)
Draft
922
9.3.2.2
Gauss-Legendre Quadrature for n = 2
83 First we seek the integration points for n = 2, () = ( 1 )( 2 ), and the resulting

equations are
1
( 1 )( 2 ) 0 d = 0
(9.65-a)
1
1
( 1 )( 2 ) 1 d = 0
(9.65-b)
(9.65-c)
Upon integration, we obtain

1 2 =
hence
1
3
1
1 =
3
and 1 + 2 = 0
(9.66)
1
and 2 =
3
(9.67)
The weight coecients are

=
(2)
W1
(2)
W2
9.3.3
(n)
Wi
1
1
1
1
=
1
li ()d
(9.68-a)
22
2
d =
= 1.0
1 2
1 2
(9.68-b)
21
1
d =
= 1.0
2 1
1 2
(9.68-c)
Rectangular and Prism Regions
84 Numerical integration of F (, ) over a rectangular region 1 1, and 1 1,

is accomplished by selecting m and n (not to be confused with the order of the polynomial)
integration points in the and directions
1
1
1
1

F (.)dd
P (, )dd =
1
(m)
Wi
(n)
Wj F (i , j )
(9.69)
i=1 j=1
and the total number of integration points will thus be m n, Fig. 9.16.
9.3.4
Triangular Regions
85 For the numerical integration over a triangle, the Gauss points are shown in Fig. 9.17, and
the corresponding triangular coordinates are given by Table 9.5.
Victor Saouma
Draft
= 1/ 3
923
= +1/ 3
= + 0.6
= 0.6
= +1/ 3
= + 0.6
6
3
5
4
7
= 1/ 3
= 0.6
Figure 9.16: Gauss-Legendre Integration Over a Surface
b
a
b
a
c
d
Quadratic
Linear
Cubic
Figure 9.17: Numerical Integration Over a Triangle
Order
Linear
Quadratic
Error
R = O(h2 )
R = O(h3 )
Cubic
R = O(h4 )
Points
a
a
b
c
a
b
c
d
Triang. Coord.
1 1 1
3, 3, 3
1 1
2, 2, 0
0, 12 , 12 ,
1
1
2 , 0, 2
1 1 1
3, 3, 3
0.6, 0.2, 0.2
0.2, 0.6, 0.2
0.2, 0.2, 0.6
Weights
1
1
3
1
3
1
3
27
48
25
48
25
48
25
48
Table 9.5: Coordinates and Weights for Numerical Integration over a Triangle
Victor Saouma
Draft
924
9.4
86
Stress Recovery; Nodal Stresses
Stresses are evaluated from

= DBu
(9.70)
in general, it is desirable to have them evaluated at the elements nodal points. However, it
should be kept in mind that stresses computed at a given nodes from dierent elements need
not be the same (since stresses are not required to be continuous in displacement based nite
element formulations).
87
Hence some sort of stress averaging at nodal points may be desirable.
88 In the isoparametric formulation, nodal stresses are very poor, and best results are obtained
at the Gauss points.
89
To evaluate nodal stresses, two approaches:

1. Evaluate directly at the nodes ( = = 1)
2. Evaluate the stresses at the Gauss points and then extrapolate.
The second approach yields far better results.

Extrapolation from Gauss points will be illustrated through Fig. 9.18 for the 4 noded isoparametric quadrilateral. We specify an internal element with its own nodes and natural coordi90
Figure 9.18: Extrapolation from 4-Node Quad Gauss Points (Felippa 1999)
nates and which are related to and through Table 9.6 or
= 3
= 3
(9.71)
hence any scalar quantity (such as xx ) whose values i is known at the Gauss element corners
can be interpolated through the usual bilinear shape functions now expressed in terms of and

2
( , ) = N1e N2e N3e N4e
(9.72)

4
Victor Saouma
Draft
9.5 Nodal Equivalent Loads

Corner
Node
1
2
3
4
925
1
+1
+1
1
1
1
+1
+1
3
+3
+3
3
3
3
+3
+ 3
Gauss
Node
1
2
3
4
1/3
+1/3
+1/3
1/ 3
1/3
1/3
+1/3
+1/ 3
1
+1
+1
1
1
1
+1
+1
Table 9.6: Natural Coordinates of Bilinear Quadrilateral Nodes

where
N1e
N2e
N3e
N4e
=
=
=
=
1
(1 )(1 )
4
1
(1 + )(1 )
4
1
(1 + )(1 + )
4
1
(1 )(1 + )
4
(9.73-a)
(9.73-b)
(9.73-c)
(9.73-d)
Similarly, we can extrapolate to the cornersof the element. For corner 1, for instance, we
replace and in the preceding equations by 3. Doing that for the four corners, we obtain

12 1 12 3
12
1 + 12 3

1
1
1
1
1
2
2 1 + 2 3
2 1 2 3
2
=
(9.74)
1 1 3
3
3
12 1 + 12 3
12
4
4
12
1 12 3
12
1 + 12 3
As expected, the sum of each row is equal to one, and for stresses we replace by xx , yy ,
and xy
As we know, dierent nodal stresses will be obtained from adjacent elements. To obtain a
single value we can either take
91
1. Unweighted average of all the nodal stresses.

2. Weighted average of nodal stresses based on the relative sizes of the elements as determined
from their area through det(J).
9.5
Nodal Equivalent Loads
In the nite element formulation, all loads must be replaced by an energy equivalent nodal
load.
92
93
We shall consider the following cases: nodal load, gravity, tractions, and thermal.
9.5.1
Gravity Load
94 Gravity forces are equivalent to a body force/unit volume acting within the solid in the
direction of the gravity axis, Fig. 9.19,
Victor Saouma
Draft
926
Volume d
6
5
Direction of Gravity
Figure 9.19: Gravity Loading

(which need not be coincident with either of the coordinate axis).
dGx = gd sin
(9.75-a)
= gd cos
dGy
(9.75-b)
where g is the gravitational acceleration and is the mass density.

95
Recalling from Eq. 7.16 that

NT bd
fe =
(9.76)
we obtain
Pxi
Pyi

Ni g
=
e
sin
cos

d
(9.77)
or the energy equivalent nodal forces for node i are

96
Pxi
Pyi

=
ngaus
ngaus
j=1
k=1

gt
sin
cos

Ni (j k )Wj Wk detJ(j , k )
(9.78)
The angle is to be measured counter-clockwise from the positive y axis.
9.5.2
Traction Load
Any element edge can have a distributed load per unit length in a normal and tangential
direction prescribed.
97
98 The variation of the distributed load is polynomial and its order can not exceed the order of
the element.
99
For the sake of consistency, loaded nodes are listed also counterclockwise.
100 First we determine the components of the distributed loads in the x and y directions by
considering the forces acting on an incremental length dS of the loaded edge, Fig. 9.20:
dPx = (pt dS cos pn dS sin ) = (pt dx pn dy)

dPy = (pn dS cos pt dS sin ) = (pn dx pt dy)
Victor Saouma
(9.79)
Draft
9.5 Nodal Equivalent Loads
927
6
dy
Pt
dx
Pn
Pn
Figure 9.20: Traction Load in Isoparametric Elements

101
But since integration is to be carried on in terms of natural coordinates:

dx =
x
d
dy =
y
d
(9.80)
Substituting

x
y
pn
d

x
y
+ pt
d
=
pn
dPx =
dPy
(9.81-a)
pt
(9.81-b)
(9.81-c)
102
From Eq. 7.16 we have

NT td
fe =
(9.82)
or

x
y
pn
d
=
Ni pt
t

x
y
+ pt
d
=
Ni pn
t

Pxi
Pyi
(9.83-a)
(9.83-b)
The integration is again carried on numerically (along the edge) and the energy equivalent
nodal forces for node i are
Pxi =
Pyi =
ngaus
j=1
ngaus
j=1
Victor Saouma

Ni

Ni
x
y
pn
pt
x
y
+ pt
pn

Wj
(9.84)
Wj
(9.85)
Draft
928
where the integration is carried on numerically along the edge. and note that
taken straight out of the Jacobian matrix.
103
Note that since integration is to be carried along the edge, we have used .
104
For adjacent elements, Fig. 9.21.
and
are
Figure 9.21: Traction Load in Contiguous Isoparametric Elements
9.5.3
105
Initial Strains/Stresses; Thermal Load
We distinguish between two cases:
Plane Stress which is the simplest

0xx = T
(9.86-a)
0yy
0
xy
= T
(9.86-b)
= 0
(9.86-c)
Plane Strain we have

0
zz
+ T
E
0
= zz + T
E
= 0
0xx =
(9.87-a)
0yy
(9.87-b)
0
xy
0zz =
0
zz
+ T = 0
(9.87-c)
(9.87-d)
0 , we obtain
Using the last expression to eliminate zz
Victor Saouma
0xx = (1 + )T
(9.88-a)
0yy
0
xy
0
zz
= (1 + )T
(9.88-b)
= 0
(9.88-c)
= ET
(9.88-d)
Draft
9.6 Computer Implementation

106
929
Those expressions are then substituted into Eq. 7.12

T
B D 0 d
BT 0 d
f0e =
e
(9.89)
and integrated numerically.
9.6
9.6.1
Computer Implementation
Algorithm
107 The computer implementation of a numerically integrated isoparametric element is summarized as follows.
But rst, it is assumed that this operation is to be performed in a function called stiff and
it takes as input arguments elcod, young, poiss, type, ndime, ndofn, ngaus. In turn it
will compute the stiness matrix KELEM of element ielem.
1. Retrieve element geometry and material properties for the current element
2. Zero the stiness matrix
3. Call function dmat to set the constitutive matrix De of the element
4. Enter (nested) loop covering all integration points
(a) Look up the sampling position of the current point (p , q ) (s, t) and their weights
(weigp)
(b) Call shape function routine sfr given p , q which will return the shape function Nie
N e
N e
(sfr) and the derivatives i and i (deriv) at the point p , q
Nie
Nie
and at point p , q will return
Nie
Nie
e
x and y (cartd), the Jacobian matrix J
determinant det Je (djac) and the x and y
(c) Call another subroutine (jacob), given Nie ,

cartesian shape function derivatives
(jacm), its inverse Je1 (jaci), its
coordinates all at the point p , q
(d) Call strain matrix (bmatps) routine, given Nie ,

strain matrix Bei (bmat)
Nie Nie
x , y ,
at p , q will return the
(e) Call a routine (dbmat) to evaluate De Bei (dbmat)

(f) Evaluate Bei T De Bej detJe integration weights and assemble them into the element
stiness matrix Keij
(g) Assemble De Be (smat into a stress array for later evaluation of stresses from the
nodal displacements.
5. Write Stiness matrix
Suggested list of variables:
Victor Saouma
Draft
930
idime, ndime
idofn,ndofn
ielem,nelem
igaus,jgaus,ngaus
inode, nnode
kgasp,ngasp
type
poiss
young
elcod(ndime,nnode)
s
t
gpcod(ndime,ngasp)
posgp(mgaus)
weigp(mgaus)
shape(nnode)
deriv(ndime,nnode)
cartd(ndime,nnode)
djacb
jacm(ndime,ndime)
jaci(ndime,ndime)
bmat(nstre,nevab)
Index, Number of dimensions (2 for 2D)

Index, Number of degree of freedom per node
Index, number of elements
Index, Index, Number of Gauss rule adopted
Index, number of nodes per element
Kounter, number of Gauss points used
1 for plane stress; 2 for plane strain
Poissons ratio
Youngs modulus
Local array
of the element currently under consid of nodal cartesian coordinates

x(1 , 1 ) x(8 , 8 )
eration
y(1 , 1 ) y(8 , 8 )
coordinate of sampling point
coordinate of sampling point
Local array of cartesian
coordinates of the Gauss pointsfor element currently

x(G1 , G1 x(G5 , G5 )
under consideration
y(G1 , G1 y(G5 , G5 )
coordinates of Gauss point
Weight factor for Gauss point
Shape function associated with each node of current element at (p , p )
"
#
N1 (p , p )
..
.N8 (p , p )
Shape
at sampling
point (p , p ) within the element
" N function derivative
#
N8
1
(
,
(
,
)
p
p
p
p
N1
8
(p , p ) N
(p , p )
Cartesian shape function derivatives associated with the nodes of the

element sampled at any point (p , p ) within the element
current

N1
8
(
,
p ) N
(p , p )
p
x
x
N1
8
(p , p ) N
(p , p )
y
y
Determinant of the Jacobian matrix sampled at any point (p , p ) within the element
Jacobian matrix at sampling point
Inverse of Jacobian matrix at sampling point
Element strain matrix at any point within the element B = [ B1 B2 B8 ]
2 N
3
i
0
x
6
Ni 7
where Bi = 4 0
y 5
Ni
y
dbmat(istre,ievab)
9.6.2
9.6.2.1
Ni
x
stores DB
MATLAB Code
stiff.m
function KELEM = stiff(ngaus,posgp,weigp,type,nelem,lnods,coord,nnode)

%----------------------------------------------------------------------------------% The purpose of this function is calculate element stiffness matricies for
% bilinear and biquadratic isoparemtric elements using Gaussian integration.
% Functions called by this function are: dmat,sfr,jacob,bmatps
%----------------------------------------------------------------------------------%
VARIABLES
%----------------------------------------------------------------------------------% nelem
Global variable NELEM
% nnode
Global variable NNODES
% posgp
Global variable POSGP
% weigp
Global variable WEIGP
% ngaus
Global variable NGAUS
% type
Global variable TYPE
Victor Saouma
Draft
931
% lnods
Global variable LNODS
% coord
Global variable COORD
%
% stifsize
Number of columns in the element stiffness matrix
% nrowcount
Position indicator for element stiffness matrix
% elmt
Current element for formation of stiffness matrix
% young
Modulus of elasticity for current element
% poiss
Poissons ration for current element
% D
Constitutive matix
% elcod
Matrix of element coordinates
% row
Counter
% kelem
Element stiffness matrix
% KELEM
Element stiffness matricies for all elements returned by function
% s
Current integration position
% t
Current integration position
% shape
Shape function at current point
% deriv
Derivative of shape function at current point
% cartd
Cartesian shape function derivatives
% jacm
Jacobian matrix
% jaci
Jacobian matrix inverse
% djac
Determinant of Jacobian matrix
% xy
x and y coordinates at the current point in the element
% bmat
Strain matrix [B]
% dbmat
Strain matrix * constitutive matrix [B]*[D]
%----------------------------------------------------------------------------------tic
fprintf(Calculating ELEMENT STIFFNESS matricies\n)
stifsize = 2*nnode;
nrowcount = stifsize-1;
for ielem = 1:nelem
%----------------------------------------% Extract material constants from lnods
%----------------------------------------elmt = lnods(ielem,1);
young = lnods(ielem,2);
poiss = lnods(ielem,3);
%----------------------------------------% Extract element coordinates
%
% elcod = [ X1 X2 X3 . . . Xn]
%
[ Y1 Y2 Y3 . . . Yn]
%----------------------------------------elcod = zeros(2,nnode);
for inode = 1:nnode
row = find(coord(:,1:1)==lnods(ielem,inode+3));
elcod(:,inode:inode) = coord(row:row,2:3);
end
%----------------------------------------% Constitutive matrix
%----------------------------------------D = dmat(young,poiss,type);
%----------------------------------------% Element stiffness matrix - element ielem
%----------------------------------------kelem = zeros(stifsize);
for igaus = 1:ngaus
Victor Saouma
Draft
932
for jgaus = 1:ngaus

s = posgp(igaus);
t = posgp(jgaus);
W = weigp(igaus)*weigp(jgaus);
[shape,deriv] = sfr(s,t,nnode);
[cartd,jacm,jaci,djac,xy] = jacob(shape,deriv,elcod);
[bmat,dbmat] = bmatps(shape,cartd,D);
kelem = kelem + bmat*dbmat*djac;
end
end
%---------------------------------------------------------------------% Store element stiffness matricies in as a stack in a single matrix :
%
kelem(1)
% KELEM =
:
%
kelem(nelem)
%---------------------------------------------------------------------startrow = stifsize*ielem - nrowcount;
endrow
= ielem*stifsize;
KELEM(startrow:endrow,:) = kelem;
end
t = toc;
fprintf(1,Elapsed time for this operation =%3.4fsec\n\n,t)
9.6.2.2
dmat.m
function D = dmat(young,poiss,type)
%----------------------------------------------------------------------------------% This function calculates the constitutive matrix for an element
%----------------------------------------------------------------------------------%
VARIABLES
%----------------------------------------------------------------------------------% young
Youngs modulus/Modulus of elasticity
% poiss
Poissons ratio
% E
Modulus of elasticity
% type
type of problem - plain stress = 1, plain strain = 2
% D
Constitutive matrix returned by function
%----------------------------------------------------------------------------------tic
%fprintf(Calculating element constitutive matrix\n)
E = young;
v = poiss;
%-------------------------% Plain stress
%-------------------------if type == 1.0
D = (E/(1-v^2))*[
1
v
0
%-------------------------% Plain strain
%-------------------------else
Victor Saouma
v
1
0
0;
0;
(1-v)/2];
Draft

D = E/((1+v)*(1-2*v))*[ 1-v
v
0
933
v
1-v
0
0
;
0
;
(1-2*v)/2];
end
t = toc;
%fprintf(1,Elapsed time for this operation =%3.4fsec\n\n,t)
9.6.2.3
sfr.m
function [shape,deriv] = sfr(s,t,nnode)

%----------------------------------------------------------------------------------% This function calculates the shape function and derivative for the current node
%----------------------------------------------------------------------------------%
VARIABLES
%----------------------------------------------------------------------------------% shape
Shape function returned by function
% deriv
Derivative of shape function returned by function
% nnode
Number of nodes per element
% s
Natural coordinate (xi) of sampling point - horizontal
% t
Natural coordinate (eta) of sampling point - vertical
%----------------------------------------------------------------------------------tic
%fprintf(Calculating shape functions and derivatives\n)
%------------------% Q9 Element
%------------------if nnode == 9
N9 = (1-s^2)*(1-t^2);
N8 = .5*(1-s)*(1-t^2) -.5*N9;
N7 = .5*(1-s^2)*(1+t) -.5*N9;
N6 = .5*(1+s)*(1-t^2) -.5*N9;
N5 = .5*(1-s^2)*(1-t) -.5*N9;
N4 = .25*(1-s)*(1+t) -.5*N7 - .5*N8 - .25*N9;
N3 = .25*(1+s)*(1+t) -.5*N6 - .5*N7 - .25*N9;
N2 = .25*(1+s)*(1-t) -.5*N5 - .5*N6 - .25*N9;
N1 = .25*(1-s)*(1-t) -.5*N5 - .5*N8 - .25*N9;
shape = [N1 N2 N3 N4 N5 N6 N7 N8 N9];
dN9ds
dN9dt
dN8ds
dN8dt
dN7ds
dN7dt
dN6ds
dN6dt
dN5ds
dN5dt
dN4ds
dN4dt
dN3ds
dN3dt
dN2ds
dN2dt
dN1ds
dN1dt
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
-2*s*(1-t^2);
-2*t*(1-s^2);
-.5*(1-t^2) -.5*dN9ds;
-t*(1-s)
-.5*dN9dt;
-s*(1+t)
-.5*dN9ds;
.5*(1-s^2) -.5*dN9dt;
.5*(1-t^2) -.5*dN9ds;
-t*(1+s)
-.5*dN9dt;
-s*(1-t)
-.5*dN9ds;
-.5*(1-s^2) -.5*dN9dt;
-.25*(1+t) -.5*dN7ds .25*(1-s)
-.5*dN7dt .25*(1+t)
-.5*dN6ds .25*(1+s)
-.5*dN6dt .25*(1-t)
-.5*dN5ds -.25*(1+s) -.5*dN5dt -.25*(1-t) -.5*dN5ds -.25*(1-s) -.5*dN5dt -
Victor Saouma
.5*dN8ds
.5*dN8dt
.5*dN7ds
.5*dN7dt
.5*dN6ds
.5*dN6dt
.5*dN8ds
.5*dN8dt
.25*dN9ds;
.25*dN9dt;
.25*dN9ds;
.25*dN9dt;
.25*dN9ds;
.25*dN9dt;
.25*dN9ds;
.25*dN9dt;
Draft
934
deriv = [dN1ds dN2ds dN3ds dN4ds dN5ds dN6ds dN7ds dN8ds dN9ds;
dN1dt dN2dt dN3dt dN4dt dN5dt dN6dt dN7dt dN8dt dN9dt];
%------------------% Q8 Element
%------------------elseif nnode == 8
N8 = .5*(1-s)*(1-t^2);
N7 = .5*(1-s^2)*(1+t);
N6 = .5*(1+s)*(1-t^2);
N5 = .5*(1-s^2)*(1-t);
N4 = .25*(1-s)*(1+t) -.5*N7
N3 = .25*(1+s)*(1+t) -.5*N6
N2 = .25*(1+s)*(1-t) -.5*N5
N1 = .25*(1-s)*(1-t) -.5*N5
.5*N8;
.5*N7;
.5*N6;
.5*N8;
shape = [N1 N2 N3 N4 N5 N6 N7 N8];

dN8ds
dN8dt
dN7ds
dN7dt
dN6ds
dN6dt
dN5ds
dN5dt
dN4ds
dN4dt
dN3ds
dN3dt
dN2ds
dN2dt
dN1ds
dN1dt
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
-.5*(1-t^2);
-t*(1-s);
-s*(1+t);
.5*(1-s^2);
.5*(1-t^2);
-t*(1+s);
-s*(1-t);
-.5*(1-s^2);
-.25*(1+t) -.5*dN7ds
.25*(1-s)
-.5*dN7dt
.25*(1+t)
-.5*dN6ds
.25*(1+s)
-.5*dN6dt
.25*(1-t)
-.5*dN5ds
-.25*(1+s) -.5*dN5dt
-.25*(1-t) -.5*dN5ds
-.25*(1-s) -.5*dN5dt
.5*dN8ds;
.5*dN8dt;
.5*dN7ds;
.5*dN7dt;
.5*dN6ds;
.5*dN6dt;
.5*dN8ds;
.5*dN8dt;
deriv = [dN1ds dN2ds dN3ds dN4ds dN5ds dN6ds dN7ds dN8ds;

dN1dt dN2dt dN3dt dN4dt dN5dt dN6dt dN7dt dN8dt];
%------------------% Q4 Element
%------------------else
N4 = .25*(1-s)*(1+t);
N3 = .25*(1+s)*(1+t);
N2 = .25*(1+s)*(1-t);
N1 = .25*(1-s)*(1-t);
shape = [N1 N2 N3 N4];
dN4ds
dN4dt
dN3ds
dN3dt
dN2ds
dN2dt
dN1ds
dN1dt
=
=
=
=
=
=
=
=
-.25*(1+t);
.25*(1-s);
.25*(1+t);
.25*(1+s);
.25*(1-t);
-.25*(1+s);
-.25*(1-t);
-.25*(1-s);
deriv = [dN1ds dN2ds dN3ds dN4ds;
Victor Saouma
Draft
935
dN1dt dN2dt dN3dt dN4dt];

end
t = toc;
9.6.2.4
jacob.m
function [cartd,jacm,jaci,djac,xy] = jacob(shape,deriv,elcod)

%----------------------------------------------------------------------------------% This function calculates the cartesian shape function derivatives
% the Jacobian matrix, Jacobian inverse and Jacobian determinant
%----------------------------------------------------------------------------------%
VARIABLES
%----------------------------------------------------------------------------------% shape
% deriv
Derivative of shape function at current point
% cartd
Cartesian shape function derivatives returned by function
% jacm
Jacobian matrix returned by function
% jaci
Jacobian matrix inverse returned by function
% djac
Determinant of Jacobian matrix returned by function
% xy
x and y coordinates at the current point in the element
%----------------------------------------------------------------------------------tic
%fprintf(Calculating Jacobian matrix\n)
%---------------------------------------------------------------% The cartesian shape function derivatives {cartd} are given by:
%
%
{dN/dx}
-1 {dN/ds}
% {cartd} = {
} = [J] {
}
%
{dN/dy}
{dN/dt}
% Start by calculating Jacobian [J] = jacm:
%
T
T
%
[dX/ds dY/ds]
[{dN/ds}*{x} {dN/ds}*{y}]
% [J] = [
] = [
T
T
]
%
[dX/dt dY/dt]
[{dN/dt}*{x} {dN/dt}*{y}]
%---------------------------------------------------------------jacm = deriv*elcod;
jaci = inv(jacm);
djac = det(jacm);
cartd = jaci*deriv;
xy = elcod*shape;
t = toc;
9.6.2.5
bmatps.m
function [bmat,dbmat] = bmatps(shape,cartd,D)

%----------------------------------------------------------------------------------% This function calculates the strain matrix B
%----------------------------------------------------------------------------------%
VARIABLES
%----------------------------------------------------------------------------------% shape
% cartd
Cartesian shape function derivatives
% bmat
Strain matrix returned by function
% dbmat
Strain matrix * constitutive matrix D
Victor Saouma
Draft
936
%----------------------------------------------------------------------------------tic
%fprintf(Calculating strain matrix [B]\n)
numcols = 2*length(cartd);
bmat = zeros(3,numcols);
cartdcol = 0;
for ibmatcol = 1:2:numcols
cartdcol = cartdcol+1;
bmat(:,ibmatcol:ibmatcol+1) = [cartd(1,cartdcol)
0
;
0
cartd(2,cartdcol);
cartd(2,cartdcol)
cartd(1,cartdcol)];
end
dbmat = D*bmat;
t = toc;
9.6.3
108
Plott of Shape Functions
Matlab code to generate shape function plots
% Shape FUnctions
X=-1:1/20:1;
Y=X;
YT=Y;
XT=X;
N9=(1-YT.*YT)*(1-X.*X);
N8=0.5*(1-YT.*YT)*(1-X);
N7=0.5*(1-XT.*XT)*(1+Y);
N6=0.5*(1-YT.*YT)*(1+X);
N5=0.5*(1-XT.*XT)*(1-Y);
N4=0.25*(1-XT)*(1+Y)-0.5*(N7+N8);
N3=0.25*(1+XT)*(1+Y)-0.5*(N6+N7);
N2=0.25*(1+XT)*(1-Y)-0.5*(N5+N6);
N1=0.25*(1-XT)*(1-Y)-0.5*(N8+N5);
meshc(X,Y,N1)
print -deps2 shap8-1.eps
c=contour(X,Y,N1);
clabel(c)
print -deps2 shap8-1-c.eps
meshc(X,Y,N8)
c=contour(X,Y,N8);
clabel(c)
N1=N1-0.25*N9;
meshc(X,Y,N1)
c=contour(X,Y,N1);
clabel(c)
N8=N8-0.5*N9;
meshc(X,Y,N8)
c=contour(X,Y,N8);
Victor Saouma
Draft
937
clabel(c)
meshc(X,Y,N9)
c=contour(X,Y,N9);
clabel(c)
Victor Saouma
Draft
938
Victor Saouma
Draft
Chapter 10
ELEMENT FORMULATION and

STRAIN RECOVERY in HW
FORMULATION
Taken from (
Cervenka, J. 1994)
10.1
Element Formulation
ervenka, J. 1994)
Taken from (C
It is necessary to select appropriate interpolation functions for all three elastic elds (i.e. u,
and ). The choice of these shape functions must be such that the BB condition is satised
(Appendix ??). In this work, the same interpolation functions are used for all three elds (i.e.
displacements, strains and stresses), which implies that there is a full number of unknowns in
each node.
dim(un ) = N dim R,
dim( n ) = dim( n ) = N dim()
(10.1)
where N denotes the number of nodes, dim is the problem dimension and R is the number
of rigid body modes. In Section 10.3, it will be shown that this formulation guarantees the
satisfaction of the BB condition.
The polynomial orders of the eld approximations are given in Table 10.1.
Table 10.1: Polynomial orders of the shape
eld
2D
T3
T6
T4
displacement u linear quadratic linear
strain
linear quadratic linear
stress
linear quadratic linear
functions.
3D
T10
quadratic
quadratic
quadratic
In general case a variable x is interpolated over a nite element using the expression:
x=
Nen
i xi
(10.2)
Draft
102
ELEMENT FORMULATION and STRAIN RECOVERY in HW

FORMULATION
where Nen is the number of element nodes, i is an interpolation function associated with node
i, and xi is the value of variable x at element node i. For the linear triangular element (T3)
the interpolation functions are:
i = li ,
i = 1, 2, 3
(10.3)
and for the six noded triangular element (T6) with three corner nodes and three mid-side nodes
the interpolation functions are:
i = li (2li 1)
i = 1, 2, 3,
4 = 4l1 l2 ,
5 = 4l2 l3 ,
6 = 4l3 l1 ,
(10.4)
where element nodes 1 to 3 indicate the element corner nodes and 4 to 6 are the mid-side nodes.
Symbols li denote the natural area coordinates of the element, which are related to the element
natural coordinates and by relations:
l1 = ,
l3 = 1
l2 = ,
(10.5)
For three-dimensional nite elements, the interpolation functions are similar. For the linear
tetrahedron T4 element they are:
i = li ,
i = 1, 2, 3, 4
(10.6)
and for the T10 element with four corner nodes and six mid-edge nodes they are dened
analogically to the six noded triangular T6 element as:
i = 1, 2, 3, 4, 5 = 4l1 l2 , 6 = 4l2 l3 ,
i = li (2li 1),
8 = 4l1 l4 ,
9 = 4l2 l4 , 10 = 4l3 l4
7 = 4l3 l1 ,
(10.7)
Similarly to the two-dimensional elements, symbols li denotes the volumetric natural coordinates, which are again related to the element natural coordinates by relations:
l1 = ,
10.2
l2 = ,
l3 = ,
l4 = 1
(10.8)
Strain Recovery
ervenka, J. 1994)
Taken from (C
In Equation ??, Step 2 is essentially an expression for the computation of nodal strains from
nodal displacements. It involves the inversion of a symmetric matrix C T , or in other words
the solution of a system of linear simultaneous equations. Three strain recovery techniques,
described below, represent an attempt to avoid the direct solution of a large linear system of
equations in this step, as its assembly and factorization is computationally expensive.
Three algorithms are discussed and compared: (1) C-lumping (CL), (2) Strain smoothing
(SS) and (3) C-splitting (CS). The CL as the simplest algorithm, in which a lumped form of
matrix C is constructed, and the inversion of the resulting diagonal matrix is trivial. The other
two algorithms are dierent iterative techniques to solve the system of equations. Between
them, the CS method is tailored for the fastest convergence for linear elements.
In the sequel, the MIM iteration index k is omitted as it focuses on faster solution techniques
for the Step 2 only.
Victor Saouma
Draft
10.2 Strain Recovery
10.2.1
103
C-lumping.
The inversion of C for the C-lumping (CL) technique is simplied by forming a diagonalized
C matrix. This lumped C matrix 1 is evaluated by the following expression:

(I )dV
(10.9)
CL =
V
where I is the identity matrix and is the shape function matrix. Since identical shape functions are used for all three primary elds, the subscript at is no longer necessary. Replacing
C by C L Step 2 of Equation ?? reduces to:
n = C 1
L Eun
At the element level, the lumped C matrix
0
C eL =
0
0
(10.10)
for the four node linear tetrahedron (T4) is:
0 0 0
5 0 0
(10.11)
0 5 0
0 0 5
where is a constant based on the element volume.

CL is by far the simplest strain recovery method since no iterations are required to compute
the nodal strains. However, the numerical experiments reported in Section ?? indicate that
the displacement solution converges to an erroneous value. Hence, the C-lumping technique is
kinematically inconsistent.
10.2.2
Strain smoothing.
Strain Smoothing (Zienkiewicz, Vilotte, Toyoshima and Nakazawa 1985) (SS) is an indirect
procedure within Step 2 that avoids the direct decomposition of the C matrix. Nodal strains
are iteratively evaluated until the ratio of the Euclidean norms of strain correction to total
strains satises a prescribed limit.
This technique exploits the diagonal matrix C L previously described and the consistent
matrix C dened below. Iteratively the nodal strains are evaluated by:
j
= jn + C 1
j+1
n
L (Eun C n ).
(10.12)
where j = 0, 1, 2 . . . is the strain-iteration count. Note that this represents an internal iteration,
not to be confused with the MIM iteration of (??). The iteration process involves the nodal
strains in the whole mesh since Step 2 is equivalent to the least square t of the nodal based
strain eld to the strain eld derived from the displacement eld (Zienkiewicz and Taylor 1989).
For a four noded linear tetrahedral element (T4), the consistent matrix C is given at the
element level by:
2 1 1 1

1 2 1 1
e
T dV =
(10.13)
C =
1 1 2 1
V
e
1
Note: The diagonalized matrix C L and the consistent matrix
in Section 10.2.2 and 10.2.3 are
1 1 C 1described
2
determined in an analogous way as the standard lumped and consistent mass matrices in dynamics. The only
dierence is the exclusion of the weight density of the material which is replaced by unity.
Victor Saouma
Draft
104

FORMULATION
where is again a constant based on the element volume.

The correction of nodal strain jn during one iteration is:
jn = A() nj1 ,
jn = j+1
n
where:
(j 1)
A() = I C 1
L C
(10.14)
(10.15)
where A() is a xed amplication matrix having a spectral radius = 45 . The spectral radius
is dened as the largest eigenvalue of amplication matrix A(). Since Equation 10.15 involves
a product of C and inverse of C L , the constants are cancelled out. By Banachs xed point
theorem (Haser and Sullivan 1991) it is necessary for the spectral radius to be less than 1
to ensure convergence of the iterative process given by Equation 10.12. Thus, this value of the
spectral radius indicates an error decay of 15 .
10.2.3
C-splitting.
A new iterative process was recently developed by (Cervenka,

Keating and Felippa 1993) to solve
Step 2. This new technique guarantees faster convergence for linear triangular and tetrahedral
elements (T3 and T4). This technique is referred to as C-splitting (CS). This method splits
the consistent matrix C of Equation 10.13 into two matrices. One matrix is diagonalized and
the second is formed such that their algebraic sum is equivalent to the original C matrix:
C = CD + CR
(10.16)
where:
C D = diag(C)
CR = C CD
(10.17)
is a splitting coecient controlling the splitting of the matrix C. Using this method Step
2 in Equation ?? is modied to:
j
= C 1
j+1
n
D (Eun C R n ).
(10.18)
For the C-splitting method, the per-iteration strain correction is:

jn = A() nj1 ,
jn = j+1
n
(j 1)
(10.19)
where the amplication matrix A() is given by:

A() = C 1
D C R,
(10.20)
It is possible to select the coecient such that the spectral radius of the amplication matrix
is minimal. For four-node tetrahedron elements using = 32 , which as shown below minimizes
the spectral radius of the amplication matrix, C splits at the element level into the following
matrices:
3 0 0 0
0 3 0 0
(10.21)
C eD =
0 0 3 0
0 0 0 3
and
C eR
Victor Saouma
1 1
1
1
1 1 1
1
=
(10.22)
1
1 1 1
1
1
1 1
Draft
10.3 Uniqueness and Existence of a Solution
105
Table 10.2: Table of coecients and spectral radii for CS technique.

Element type
splitting coef. spectral radius
three node linear triangle T3

5/4
3/5
four node
quadrilateral Q4
5/4
4/5
four node
tetrahedral element T4
3/2
2/3
eight node brick element B8
7/4
0.929
Then the spectral radius is equal to 23 . Thus, CS has an error decay rate of 13 allowing
for a faster convergence than the SS method. For example, 10 steps of CS can be expected to
10
reduce the initial strain errors by ( 23 ) 0.0173 whereas 10 steps of SS would reduce those
10
errors by only ( 45 ) 0.1074.
This technique was investigated also for other low order element types. The best splitting
coecients and resulting spectral radii of the operator A() (Eq. 10.20) for other element
types are summarized in Table 10.2. The coecients for four node quadrilateral and eight
node brick element are however valid only for elements with parallel or almost parallel sides.
Therefore, the (SS) technique would be probably more reliable for these element types.
10.3
Uniqueness and Existence of a Solution
The BB condition for uniqueness and existence of a solution of the three-eld variational principle is stated in Appendix ??. This condition was derived by (Babuska 1971), (Babuska 1973)
and (Brezzi 1974). Xue and Atluri (1985) extended the condition to a general three-eld problem, and derived its discrete form. The continuous and discrete forms of the BB conditions are
again described in Appendix ??, and it is shown that they are equivalent to the following three
conditions:
rank(E) = nu n
(10.23)
rank(C) = n n= + nu
A is positive denite
where matrices E, C and A are derived in Appendix ?? and are given by the following integrals:

T
BdV, C =
dV, A =
DT dV
(10.24)
E=
V
The third condition is satised as long as the material does not exhibit softening. This is always
guaranteed in the discrete crack approach, since softening is modeled only along the interface
elements which are not included in the mixed iterative solution.
If identical shape functions are used for all three elds, then the number of unknowns for
each eld is given by Equation 10.1, and the inequalities in the rst and second condition of
Equation 10.23 are always satised.
n
=
N dim()
nu = N dim R
n = N dim() n= + nu = N dim() + N dim R
Victor Saouma
(10.25)
Draft
106

FORMULATION
We note that the rank condition of matrix C is also satised, as it is analogous to the
consistent mass matrix of isoparametric elements, which has always full rank, as can be seen
from Equation 10.13.
More complex is the verication of the rank condition of matrix E. In Equation 10.24, E
is given by the integral:

E=
B dV
(10.26)
where matrix B is the matrix relating strains at a certain point to the nodal displacements
and is the matrix of shape functions relating strains or stresses at a certain point to their
nodal counterparts. Matrix E will have a rank equal to nu if the following two conditions are
satised.
x V
: B(x)un = 0
un = 0,
(10.27)
T
(x) = (x) n : (x) is unique
n ,
The rst condition is equivalent to the requirement that a nonzero vector of nodal displacements
must cause non-zero strain eld. It should be noted that the rigid body modes are excluded
from vector un . They would be the only modes allowed to produce a zero strain eld. In
this case, matrix B corresponds to that of a standard isoparametric triangular or tetrahedral
element, and will therefore satisfy this condition.
The second condition is also satised, since the shape functions in matrix are those of a
standard isoparametric element, and the uniqueness of the interpolation is guaranteed.
For higher order triangular and tetrahedral elements (i.e. 6 noded triangle and 10 noded
tetrahedron), it would seem preferable to select interpolation functions for strains and stresses,
which are one order lower than those for the displacements. This would correspond to the
mathematical relation between strains and displacements, since the strains are determined by
dierentiation of the displacement eld. For this formulation, there would be unknown displacements, strains and stresses at each element corner node, but only displacement unknowns
at the midside element nodes. The strains and stresses would be interpolated using linear shape
functions and displacement using quadratic shape functions. Now, it is possible to show that
this formulation would not guarantee the satisfaction of the inequality in the rst condition of
Equation 10.23.
We consider a patch of two six-noded triangular element as shown in Figure 10.1. From
F
2F
5
20
5
F
8 and 9 noded isoparametric quadrilaterals; Pure Axial and Flexure
8 noded Isoparametric constant axial stress
Figure 10.1: Patch test.
the previous assumptions of linear stress and strain interpolation and quadratic displacement
interpolation, the unknown stresses and strains are only at the element corner nodes, while all
nodes have unknown displacements. The total number of stress and displacement unknowns is
Victor Saouma
Draft
10.3 Uniqueness and Existence of a Solution
107
then given by:

nu = 9 2 3 = 15 > n = 4 3 = 12
(10.28)
and clearly the important inequality nu n is not satised and the existence and uniqueness
of a solution cannot be guaranteed. This should be contrasted by the previous formulation, in
which the same interpolation functions are used for all three elds, and the satisfaction of the
BB is guaranteed by Equation 10.25 and 10.27.
Victor Saouma
Draft
108
Victor Saouma

FORMULATION
Draft
Chapter 11
WEIGHTED RESIDUAL
METHODS
Adapted from (Ottosen and Petersson 1992)
11.1
Introduction
1 Thus far, the nite element method was derived from a variational principle through the
Rayleigh-Ritz method.
2 In Elasto-Statics, the variational principle used is the minimization of the total potential
energy, and the resulting weak formulation is equivalent to the equation of equilibrium (strong
form).
3 In many branches of physical sciences, a variational principle may not be available, and all
what is available is a dierential equation and its associated boundary conditions.
It should be noted that a variational formulation is not possible when the order of the dierential equation is odd.
4
It is for those problems that the weighted residual methods are employed.
The resulting formulation is an integral one, and is thus weak.
11.2
General Formulation
11.2.1
Dierential Operators
Considering a general dierential equation

Governing Dierential Equation L + g = 0;
Boundary conditions
Bi () = qi ; i
(11.1)
where L is a dierential operator, (assuming a 1D problem) = (x) is an unknown function

which represents the state variable to be determined (i.e. displacement, temperature, head,
etc...), g(x) is a known forcing function.
Draft
112
8
WEIGHTED RESIDUAL METHODS
We assume that the dierential operator to be symmetric and positive denite,

(L[u])vd =
(L[v])ud
Symmetric

(L[u])ud > 0
Positive Denite
(11.2)
where u and v are any functions which satisfy essential and natural boundary conditions.
11.2.1.1
Application to 1D Axial Member
9 As an example, let us consider the steady state response of an axial bar. The governing
dierential equation and boundary conditions are
Governing Dierential Equation

Essential Boundary conditions
Natural Boundary conditions
Hence
EA xu2 = 0
u|x=0 = 0

EA u
x x=l = P
(11.3)
= u;
g = 0;
L[] = EA x
2;
q1 = 0; B2 = EA x
; q2 = P ;
B1 = 1;
(11.4)
To determine whether the operator L is symmetric and positive denite, we consider P = 0,

and integrating by part twice1 we obtain
10

l
2u
u l
u v
dx
EA 2 vdx = EA v +
EA
x
x 0
x x
0

l
v l
2v
u l
EA 2 udx
= EA v + EAu
x
x
x
0
(11.5-a)
(11.5-b)
But since the boundary conditions are u = v = 0 at x = 0 and u/x = v/x = 0 at x = l,

then we only have
l
l
2u
2v
EA 2 vdx =
EA 2 udx
(11.6)
x
x
0
0
hence the operator is symmetric.
We can also conclude that the operator is positive denite from Eq. 11.5-a

l
0
11.2.2
11
2u
EA 2 udx =
x
EA
0
u
x
2
dx
(11.7)
Residual Formulation
We now alter Eq. 11.1, and for a 1D problem rewrite it as

w(L + g)dx = 0
Z
1
From Eq.1.39:
a
Victor Saouma
u(x)v (x)dx = u(x)v(x)|ba
(11.8)
v(x)u (x)dx
Draft
11.2 General Formulation
113
where w is an arbitrary weight function2 . We note that through this equation, the governing
dierential equation is satised in a weak sense.
12 We now seek an approximate numerical solution for the unknown function (x), hence, in
general we may assume
(x)app = a1 1 (x) + a2 2 (x) + + an n (x)

= a
(11.9-a)
where ai are the unknown parameters, and (x) are prespecied trial functions3 of x.
13
Substituting the previous equation into Eq. 11.8

w(Lapp + g)dx = 0
(11.10)
Since app will in general not satisfy the dierential equation exactly, then
Lapp + g = R
b
wR = 0
(11.11)
(11.12)
where R(x) is a measure of the error, and is called the Residual; We thus impose the requirement that weights and residuals be orthogonal functions4 .
14 The residual clearly depends on the coecients ai , thus our objective is to select the weight
w(x) in such a way that the weighted residuals w(x)R(x) over a b is zero.
15
Arbitrary weight functions are selected rst

w = c1 W1 (x) + c2 W2 (x) + + cn Wn (x)
c1
c2
= W1 (x) W2 (x) Wn (x)

cn
(11.13-a)
and Wi (x) are predened functions of x, ci are parameters. Hence, we can rewrite Eq. 11.8 as

WT Rdx = 0
(11.14)
But since c does not depend on the coordinates, then

WT Rdx = 0
(11.15)
2
Note similarity with the derivation of the principle of virtual work from the equilibrium equation and boundary condition.
3
Later on in the Galerkin Method, which is a special form of Weighted Residuals, will correspond to the
to the shape functions N, and a to the nodal displacements u.
displacement,
4
T
R Two vectors a and b are orthogonals if a b = 0, similarly, two functions u(x) and v(x) are orthogonals if
v(x)u(x) = 0.
Victor Saouma
Draft
114
This equation really represents a system of n equations.

16
Substituting Eq. 11.8, and since a does not depend on the coordinates
(11.16)
R = L(a) + g = L()a + g
Hence, taking the weighted average

W L()dx a =
a

T

or
Ke
Wgdx

(11.17)
W1 L(1 )dx
W1 L(2 )dx
W1 L(n )dx
W1 gdx
a
a
a
ab

b
b
b
W
L(
)dx
W
L(
)dx
W
L(
)dx
W2 gdx
2
1
2
2
2
n
Ke = a
;f =
a
a
a
..
..
..
..
..
.
.
.
.
.

b
b
b
Wn L(1 )dx
Wn L(2 )dx
Wn L(n )dx
Wn gdx

(11.18)
Ke is thus an n by n matrix, and from the preceeding equation we can solve for thee unknown
coecients a.
17
11.3
Weighted Residual Methods
18 The previous formulation was very general, and dierent techniques will vary according to
the selected weights, i.e. our choice for W.
11.3.1
Point Collocation Method
In this method, the weight function w is based on Diracs delta function (x xi ) and
thus the residual is set to be exactly equal to zero at n points. This is achieved by adopting for
the weight function w the Diracs delta function (x xi ) where

if x = xi
(11.19-a)
(x xi ) =
0 otherwise

(x xi )dx = 1
(11.19-b)
19
11.3.2
Subdomain Collocation Method
20 We subdivide the region of interest into n subregions, and within each one of them the
integral of the residual is set to zero.
Victor Saouma
Draft
11.3 Weighted Residual Methods
11.3.3
21
Least-Squares Method
The integral of the square of the residuals are minimized with respect to ai .

I =
R2 (x, y, z, a1 , , an )dx
I
ai
11.3.4
22
115
= 0;
(11.20-a)
i = 1, 2, , n
(11.20-b)
Galerkin Method
In the Galerkin method, we select

W=
(11.21)
i.e. the weight functions are equal to the trial functions. Since the trial functions are orthogonal
to the residuals (Eq. 11.8), then Eq. 11.15 becomes
b
T Rdx = 0
(11.22)
a
and
1 L(1 )dx
1 L(2 )dx
1 L(n )dx
1 gdx
a
a
a
ab

b
b
b
L(
)dx
L(
)dx
L(
)dx
2 gdx
2
1
2
2
2
n
Ke = a
;f =
a
a
a
..
..
..
..
..
.
.
.
.
.

b
b
b
n L(1 )dx
n L(2 )dx
n L(n )dx
n gdx

(11.23)
Example 11-1: String Vibration

Formulate the nite element matrices by the Galerkin method for an an element of a string
subjected to a tension T (x), and to an external transverse force per unit length be p(x, t). The
mass per unit length is denoted by . Assume small displacements.
Solve for the fundamental frequency of vibration.
Solution:
1. We rst write Newtons equation of motion in the transverse direction (sigf y = ma) for
an element of the string

w(x, t) 2 w(x, t)
2 w(x, t)
w(x, t)
T (x)
dx
+
= dx
dx
+ p(x, t)dx T (x)
T (x) +
2
x
x
x
x
t2
(11.24)
Victor Saouma
Draft
116

w(x, t)
2 w(x, t)
T (x)
+ p(x, t =
x
x
t2
which reduces to
(11.25)
w
T(x)+
p(x,t)
w(x,t)
x
T(x) dx
x
w(x,t) + w(x,t) dx
x
x2
w(x,t)
T(x)
x+dx
2. In the case of free vibration p(x, t) = 0, and for constant tension T =cst, we obtain
T w,xx w,tt = 0
(11.26)
which is the wave equation.

3. We next apply the Galerkin method
L
NT [T w,xx w]
dx = 0
(11.27)
where w = Nw
4. Integrate by parts

NT,x T w,x dx
L
0
!
"L
NT wdx
+ NT T w,x 0 = 0
(11.28)
The last term is equal to zero because the shape functions at the corresponding nodes
are equal to zero. By analogy to the virtual displacement method, N corresponds to the
virtual displacement which must be equal to zero at the supports in order to satisfy the
essential boundary conditions.
5. Substituting w = Nw and w,x = N,x w we obtain
L
L
T
=0
N,x T N,x dx w +
NT Ndx w
0
0

m
where
N=
Lx
L
x
L
6. Substituting
k =
m =
Victor Saouma
(11.29)
and N,x =
1
1 1
L

T
1 1
L 1 1

L 2 1
1 2
6
(11.30)
(11.31)
(11.32)
Draft
117
7. If we consider a two elements assemblage, and suppress the xed d.o.f (1 and 3), then
L
T
2 w 2 + 4 w2 = 0
L
6
(11.33)
8. For free vibration

w 2 = A sin t
w2 = 2 A sin t
9. Substituting

2

L
T
T
4 2 A sin t = 0 2 = 3 L
2
L
6
Note that the exact answer is w2 =
11.4
23
(11.34)
(11.35)
(11.36)
2 T
4L2
Applications of the Galerkin Method to 3D Elasticity Problems
We will apply the Galerkin method to two classes of problems.
We start with elasticity problems, to show that the Galerkin Method would result in exactly
the same formulation as the one obtained from the variational principle.
24
25 Following this validation of the method, we shall consider (in the next chapter) a class of
problems which does not have a variational principle, and thus the Galerkin method is the only
applicable technique. The second application will be the heat equation.
11.4.1
Derivation of the Weak Form
26 We will now apply the Galerkin method to the equation of elasticity and show that we will
retrieve the principle of virtual work which was derived from a variational principle. Hence, in
this particular case the two methods are indeed (as anticipated) equivalent.
27
Starting with the equilibrium equation (Eq. ??)

LT + b = 0
where = Lu and
LT = 0
0
Victor Saouma
0
0
xx
yy
zz
z ; =
xy
xz
yz
(11.37)
bx
b
;
; b=
(11.38)
Draft
118
28
Expanding the equations of equilibrium we have

xx
x
yx
x
zx
x
+
+
+
xy
y
yy
y
zy
y
+ zxz + bx = 0
+ zyz + by = 0
+
zz
z
(11.39)
+ bz = 0
We multiply the rst equation in Eq. 11.39 by the arbitrary function wx and we integrate
over the volume5 :

xy
xx
xz
d +
d +
d +
wx
wx
wx
wx bx d = 0
(11.40)
x
y
z
29
30
31
We now apply Green-Gauss theorem (a form of the divergence theorem6 )

wx
wx
xx d +
xy d
wx xx nx d
wx xy ny d
x
y

wx
xz d +
wx xz nz d
wx bx d = 0
+
Recalling that the tractions are given by Eq. ??

tx = xx nx + xy ny + xz nz
ty = yx nx + yy ny + yz nz
tz = zx nx + zy ny + zz nz
then the preceding equation reduces to

wx
wx
wx
xx +
xy +
xz d +
wx tx d
x
y
z
32
(11.41-a)
(11.42)
wx bx d = 0
(11.43)
Similarly, we obtain for the other two equations of equilibrium

wy
wy
wy
yx +
yy +
yz d +
wy ty d
x
y
z

wz
wz
wz
zx +
zy +
zz d +
wz tz d
x
y
z
wy by d = 0 (11.44-a)
wz bz d = 0 (11.44-b)
Summing those three equations

(wx tx + wy ty + wz tz )d +
(wx bx + wy by + wz bz )d

WT t
wT b

(11.45)
wy
wz
wx
xx +
yy +
zz +
5
x
y
z
Note similarity
the
derivationRof the principle

with
R w
R of virtualwork.
6
T
w
wIfy we w
x
y vd =
x ndw
z (r)T vd.
z vT = [ 0 0 ], we obtain
From Eq. w
1.43,
r
v
set

R
R
+ R
xy +
+
+
yz d = 0
xz +

d = y
nx dx
x d
z
x
z
y
x
33
Victor Saouma
Draft
119
We next focus on the third term in the preceding equation. For w = wx wy wz T , then
34
(Lw)T =
wy
y ;
wx
x ;
wz
z ;
wx
y
wy
x ;
wx
z
wz
x ;
wy
z
wz
y ;
(11.46)
and
(Lw)T =
wx
x' xx
x
+ w
y
+
+
wy
z
+ w
y
z zz
(yy
%
wy
x
xy + w
x
z
wz
x
&
'
xz +
Hence, substituting Eq. 11.47 into Eq. 11.45 we obtain

T
T
(Lw) d =
w td +
35
wy
z
wT bd
wz
y
(
yz
(11.47)
(11.48)
But since the L operator is symmetric, from Eq. 11.2, the above equation can also be written
36
as
w (L )d =
T
wT bd
w td +
(11.49)
which is the weak form of the dierential equation of equilibrium subjected to the traction
(natural) boundary conditions.
37 The weight vector w is completely arbitrary. In the principle of virtual displacement, it
would correspond to the virtual displacement, compare this last equation with Eq. 5.11.

(11.50)
+ uT {G [D( 0 ) + 0 ] t}d = 0

11.4.2
38
FE Discretization
Given that the displacement eld is written as

(11.51)
u = Nu
where N are the shape functions, and u are the nodal known displacements. Note that this is
the application of Eq. 11.9-a
39 The Galerkin method uses the shape functions for the weight (functions Wi ) Hence, Eq.
11.13-a will be
wT = cT NT
w = Nc;
(11.52-a)
and
Lw = LNc = Bc
(11.53)
40 Inserting the last two equations into the weak form, and recalling that the L operator is
symmetric, Eq. 11.49, yields

T
T
T
T
B d
N td
N bd = 0
(11.54)
c
Victor Saouma
Draft
1110
(Note similarity with Eq. 11.14). Since c is an arbitrary matrix, we conclude that

BT d =
NT td +
NT bd
(11.55)
This equation is applicable to both linear and nonlinear systems. The right hand side represent
the eect of the load.
41
Next we introduce a constitutive model, Eq. 3.90 which is rewritten in vector form
= D( 0 )
42
43
(11.56)
The stress and strain are given by

= Lu = LNu = Bu
(11.57-a)
= DBu D0
(11.57-b)
Inserting those two equations into Eq. 11.55, yields

B DBd u =

Ke
N td +

f1
N bd +

fb
BT D0 d

(11.58)
f0
Thus, using the Galerkin approach, we have recovered the same stiness formulation as the
one previously obtained from a Variational Principle.
44
45 Whereas, there may not be any clearer advantage in using the Galerkin approach when a
variational principle is available, this may be the only formulation possible in the absence of
the second.
Victor Saouma
Draft
Chapter 12
FINITE ELEMENT
DISCRETIZATION OF THE
FIELD EQUATION
Adapted from Ottosen (1992)
1 Having proven the validity of the Galerkin method by its ability to recover the variational
based formulation, we now turn our attention to a problem which (strictly speaking) could only
be handled by such an approach.
2 As a vehicle for such an application, we will consider the eld equation which was discussed
in Chapter 3.3.
12.1
3
Derivation of the Weak Form
Our starting point will be the eld equation, Eq. 3.113

div (D) Q = c
(12.1)
Recalling (Eq. 3.98) that q = D, and that for steady state problems the right hand side
reduces to zero, and nally (for the mere sake of clarity) substituting by T , this equation
reduces to
(12.2)
div q Q = 0

4
L+g
5
The boundary conditions are given by

Essential, Temperature
T T = g
Natural, Flux
q qn = qt n = qf
Natural, Convection Flux c qc = h(T T )
(12.3)
Note that in a n, the convection heat loss can be considered as a negative heat source.
Draft
122
7
FINITE ELEMENT DISCRETIZATION OF THE FIELD EQUATION
We now apply Eq. 11.8

w(div q Q)d = 0
(12.4)
for 2D problems, we replace d by tdA

wdiv qtdA
wQtdA = 0
Applying Green-Gauss Theorem1 , Eq. ??, the rst term becomes

,

wdiv qtdA =
wtqT nd
(w)T qtdA
A
(12.5)
(12.6)
Thus Eq. 12.5 becomes

wtq nd
(w) qtdA =
A
wQtdA
(12.7)
which is the weak form.

10
Applying the boundary conditions, = T + q

(w)T qtdA =
A

wqn td
wqf td +
q
wQtdA
(12.8)
where along q the ux (which may include convection) is known but the temperature is not, and
along T , the temperature is known but ux is not. This equation is valid for any constitutive
relation and is analogous to Eq. 11.49.
11
Applying Fourrier Law, we obtain

(w) DT tdA =
wqf td
wqn td +
wQtdA
(12.9)
This is the weak form of our problem.

12
By extension, the weak form in 3D would be

,
,
wqf dS
(w)T DT dV =
V
12.2
13
Sq

wqn dS +
ST
wQdV
(12.10)
FE Discretization
The temperature eld is expressed in terms of the nodal ones

(12.11)
T = NT
where N are the shape functions, and T are the nodal known temperatures.
1
RR
div qdA =
Victor Saouma
qT nd
RR
(r)T qdA
Draft
12.2 FE Discretization
14
123
The ux is given by q = DT where

T =
T
T
y
T
x
(12.12)
furthermore, the ux qn at the boundary is given by qn = qT n.

15
Substituting
T = NT = BT
#
where
B=
(12.13)
Ni
x
Ni
y
(12.14)
Note, this denition of the B matrix for scalar eld problems, should not be confused with
the one for vector problems, Eq. ??
N
i
i
0
0
x Ni
Ni
1
(12.15)
B= 0
0
y = J

16
Ni
y
12.2.1
17
Ni
Ni
No Convection
Substituting the previous set of equations, Eq. 12.9 reduces to

T
(w) DBtdA T =
wqf td
wqn td +
A
18
Ni
x
wQtdA
(12.16)
In accordance with the Galerkin method, the weight function w is

w = Nc
(12.17)
Since w is arbitrary, the matrix c is also arbitrary, thus

w = Nc = Bc
(12.18)
w = cT NT
(12.19)
and since w is a scalar
19
Inserting those equations into Eq. 12.16

#

T
T
B DBtdA T +
NT qf td +
c
A
20

N qn td
T
N QtdA = 0
(12.20)
But since this expression should hold for any arbitrary cT matrix, we conclude that
Ke T
Ke
fbe
e
fQ
e
= fbe + fQ

=
BT DBtdA
A

T
=
N qf td
q

=
NT QtdA
(12.21)
(12.22)
NT qn td
(12.23)
(12.24)
Victor Saouma
Draft
124
12.2.2
21
Convection
Next we consider the eet of the convection term, and thus

= q + T + c
(12.25)
The presence of the convection term aects only fb in Eq. 12.23 which now becomes

fb =
NT qf td
NT qn td
NT qc d
q
(12.26)
and based on Newtons convection boundary condition (Eq. 3.97)

qc = h(T T )
(12.27)
qc = hNT hT
(12.28)
or
22
Substituting into Eq. 12.26 yields

NT qf td
NT qn td
fb =
q
23

hNT Nd T + T
NT hd
(12.29)
Using this relation, Eq. 12.23 becomes

(Ke + Kec )T
Ke
Kec
fbe
e
fQ
e
= fbe + fQ

=
BT DBtdA
A
=
hNT Nd
c

T
=
N qf d
q

=
NT QtdA
(12.30)
(12.31)
(12.32)

T
N qn d + T
NT hd
(12.33)
(12.34)
It should be noted that convection occurs around the perimeter of an element, and conduction
occurs inside its volume.
24 Note the analogy between this load case and P- eects in structural engineering. As will
be seen in chapter 14, the geometric stiness matrix (KG ), which accounts for large deformation,
will be very similar to Kc .
12.3
Illustrative Examples
Example 12-1: Composite Wall
Victor Saouma
Draft
12.3 Illustrative Examples
125
Let us consider a composite wall made up of three materials, (k1 = 20 W/mo C, k2 = 30

W/mo C, k3 = 50 W/mo C),
T0 = 20 c
K1
K2
K3
0.3 m
0.15 m 0.15 m
inf
4
20
1
0
0
1
0
1
0
1
0
1
0
1
0
1
Natural B.C.
11
00
00
11
00
11
00
11
00
11
00
11
00
11
Essential B.C.
The outer temperature is T0 = 20o C, convection heat transfer takes place on the inner surface
of the wall with the uid temperature T = 800o C, the lm coecient (or convection term)
h = 25 W/m2 .o C. We seek the temperature distribution across the wall.
Solution:
1. Selecting linear elements, we have
T () = NT =
1
1+
T1 +
T2
2
2

N1
=
d =
dx =
dT
dx
=
=
B =
(12.35-a)
N2
2
2
(x x1 ) 1 = (x x1 ) 1
x2 x1
L
2
dx
L
L
d
2
dT d
2 dN
=
.T
d dx
L d
2! 1 1 "
2 2 T = BT
L
"
1!
1, 1
L
(12.35-b)
(12.35-c)
(12.35-d)
(12.35-e)
(12.35-f)
2. The conductivity matrix for the one dimensional element is

D=k
(12.36)
3. Substituting into Eq. 12.34, we obtain

Victor Saouma
Draft
126
Element Conductivity Matrix:

L
e
T
B DBtdA =
BT DBdx
K =
0
A

kL 1 T
B Bd
=
2 1

"
kL 1 2 1
1 !
1, 1 d
=
1
2 L
1

k
1 1
=
L 1 1
K1 =
20
0.3
K2 =
30
0.15
K3 =
50
0.15
1
2
1

1 1
66.67
=
1 1
66.67
(12.37-b)
3

200.00
200.00
3
4
3

1 1
333.34
=
1 1
333.34
4

333.34
333.34
1
66.67
66.67
2
66.67
66.67 + 200.00
200.00
1
66.67
66.67
0.
0.
2
66.67
266.67
200.00
0.
2

66.67
66.67
2
3
2

1 1
200.00
=
1 1
200.00
K =
(12.37-a)
(12.37-c)
(12.37-d)
200.00
200.00 + 333.34
333.34
3
0.
200.00
533.34
333.34
(12.37-e)
333.34
333.34
0.
0.
333.34
333.34
(12.37-f)
Element Convection Matrix: Since convection occurs only at one point, 1, we just
add the convection term h = 25 to K1,1
1
66.67 + 25
66.67
0.
0.
K =
1
91.67
66.67
0.
0.
Victor Saouma
2
66.67
266.67
200.00
0.
2
66.67
266.67
200.00
0.
3
0.
200.00
533.34
333.34
3
0.
200.00
533.34
333.34
0.
0.
333.34
333.34
0.
0.
333.34
333.34
(12.38-a)
Draft
127
Element heat rate vector: Since there is no heat generation

Q, fQ =

0, and we do
NT qf d
NT qn d +
not have a distributed convection (hence fb =
q
T

T
N hd = 0) and we only have a point convection load
T
c

(800)(25)
P1
0
P2
=
0
P

3
P4 ?
P4 ?
(12.39)
where P4 corresponds to the point heat ux transmitted at node 4.

Solve
0.
91.67
66.67
0.
T1 ?
T2 ?
66.67 266.67 200.00
0.
0.
200.00 533.34 333.34
T3 ?
T4 = 20
0.
0.
333.34 333.34
(800)(25)
0
=
0
P4 ?
(12.40)
In this case, one of the essential boundary conditions is non-zero, hence following a
procedure similar to the one outlined in Eq. this reduces to
0
81.67
66.67
0.
20, 000.
T1 ?
66.67 266.67 200.00
=
0
20.
(12.41-a)
0
T2 ?
0
333.34
T3 ?
0.
200.00 533.34
20, 000.
=
0
(12.41-b)
6, 666.80

T1 304.6
o
T2
119.0
C
=

T3
57.1
and the solution is
(12.42)
Verication We can check the numerical solution by verifying that

q=
T1 T2
x1 x2
k1
=q=
T2 T3
x2 x3
k2
=q=
T3 T4
x3 x4
k3
(12.43)
Example 12-2: Heat Transfer across a Fin

A metallic n with thermal conductivity k = 360 W/m.o C, 0.1 cm thick and 10 cm long
extends from a plane whose temperature is 235 o C.
Victor Saouma
Draft
128
Tinf= 20o C
2
h = 9 W/m
oc
0.1 cm
10 cm
1
Tinf= 20o C
We seek to determine the temperature distribution and amount of heat transfered from the n
to the air at 20o C where the lm coecient h=9 W/mo C.
Using a three element discretization, and assuming the tip of the n to be isolated.
Solution:
25
The solution described below, considers a unit volume of the n.

Element conductivity matrix: Since we are considering a unit volume

K
=
=
=
=
K1 =
Victor Saouma
BT DBdx
0

kL 1 T
B Bd
2 1

"
kL 1
1 !
1, 1 d
1
2 1

k
1 1
L 1 1
360
0.033
K2 =
360
0.033
K3 =
360
0.033
1
2

1 1
1 1
2
3

1 1
1 1
3
4

1 1
1 1
(12.44-a)
(12.44-b)
(12.44-c)
(12.44-d)
Draft
129
1
2
1
1
360
1 1 + 1
1
0.033
K =
1
1
1
1+1
1
1
2
3
4
1 1
360
1 2 1
1 2 1
0.033
1 1
(12.44-e)
(12.44-f)
Convection Matrix The shape functions for the linear element are given by Eq. 12.35-a.
1 1+
N =
2
2
N1
1
2
1+
2

NT N =
'
=

N N =
1
1
3
N2

1
2
1
2
(2
1+ 1
2
2
2 1
1 2
(12.45-a)
1+
2
1+ 1
2
2
'
(2
1+
2
(12.45-b)
(12.45-c)
(12.45-d)
With respect to Fig. 12.1 heat transfer by convection occurs across the perimeter of the
co
nv
ec
tio
he
at
lo
ss
q
dx
Figure 12.1: Heat Flow in a Thin Rectangular Fin

n (P = 2(B + t)); however, as noted above, we are considering a unit volume of the n
Victor Saouma
Draft
1210
(AC = Bt), hence we rewrite Eq. 12.32 as

Kec
=
"
but, from Eq. 12.35-d dx =
L
2 d,

P L
hNT Ndx
Ac 0

2h L T
N Ndx
t 0
(12.46-a)
(12.46-b)
thus
hL
Kc =
t
NT Nd
(12.47)
Inserting the term from Eq. 12.45-d we nally obtain

hL 2 1
Kc =
3t 1 2
(12.48)
As for the conduction stiness matrix, the assembled convection stiness matrix will
26
be
1
2
2
(9)(3.33 10 ) 1
Kc =
3(103 )
2 3
1
4 1
1 4
1
1
2
Element heat rate vector: Can be determined from

T
T
N qf td
N qn td + T
fb =
q
NT htd
(12.49)
(12.50)
however in this problem, only the third term is non-zero.

Following a similar procedure as above,

e
fb = T

= T h
=
=
NT htd
(12.51-a)
NT d
(12.51-b)

2T hL
2t
c

T hL 1
1
t
1
2
1+
2
Assembling the load vector for this problem would yield
1
(20)(9)(3.33 102 ) 2
fb =
2
103
Victor Saouma
(12.51-c)
(12.51-d)
(12.52)
Draft

Solution of the system
360
1
0.033
1211
becomes
2
3
4
1
(9)(3.33 102 )
2 1
+
1 2 1
3(103 )
1 1
(20)(9)(3.33 102 ) 2
2
103
1
2
3
10, 800. 10, 800.
0.

10, 800. 21, 600. 10, 800.

0.
10, 800. 21, 600.
0.
0.
10, 800.
2
1
2 3 4

1
4 1
1 4 1
1 2
T1
T2 ?
T3 ?
T4 ?
(12.53-a)
0.
0.
+
10, 800.
10, 800.
P1 ?
12, 000.
12, 000.
6, 000.
0.
0.
11, 000. 10, 700.
10, 700. 22, 000. 10, 700.
0.
10, 700. 22, 000. 10, 700.

0.
0.
10, 700 11, 000.
0.
200.
100.
0.
0.
2
100.
400.
100.
0.
3
0.
100.
400.
100.
4

0.
T1 = 235
T2 ?
0.
T
100.
3?
T4 ?
200.
(12.53-b)

235.
P
?
T2 ? 12, 000.
T3 ? = 12, 000.
T4 ?
6, 000.
As in the previous example, we rewrite this last equation as
22, 000. 10, 700.

0.
12, 000.
10, 700
T2 ?
10, 700. 22, 000. 10, 700.
=
12, 000.
235.
0.
T ?
3
6, 000.
0.
T4 ?
0.
10, 700. 11, 000.
2.514 106
12, 000.
(12.54-a)
=
6, 000.
Solving, we obtain
T2 208.77
o
T
C
(12.55)
=
194.25
3
T4
189.50
)
Heat loss in the n can be computed from H = e H e where H e = h(Tavg T )As sand the
surface area of each element is As = 2(1 0.0333)m2
Victor Saouma
(12.53-c)
Draft
1212
12.4
Comparison Between Vector and Scalar Formulations
25 We now complement Table 3.2, by contrasting major governing equations in both scalar and
vector problems, Table 12.1.
Scalar
Vector
Conduction
Solid Mechanics
I Dierential Equation
div q Q
qn
q
T
qn
=
T
rT
q
=
=
on
on
LT + b
"
qT n
DrT
T
q c
w (L )d =
t

u
t
(Ke + Kec )T
Kec
fbe
e
fQ
NT bd
III Discretized Solution

#
B DBtdA
Zc
Discretized Weak Form
Ni
x
Ni
y
NT td
e
fbe + fQ
+P
Z
n
D"
u
t
wT bd
NT
BT
"
Balance/Equilibrium
State variable
kinematic
Flux vector
Flux at boundary
Constitutive law
Essential BC
Natural BC
Lu
=
=
BT d
w td
+
T
T
IIZ Weak Formulation

Z
=
u
=

=
=
on
on
u
"
=
=
Nu
Bu
2
6
4
Ke ue
Ke
Ni
x
Ni
y
Ni
x
0
Ni
y
3
7
5
fe + f0e + P
B DBd
hNT Ntd
NT qf td
T
A
State variable gradient

Balance/Equilibrium
Stiness matrix
Convection Stiness Matrix
fe
fe
NT qn td
NT QtdA
State variables
Flux
R
t
NT td
Z
Z
NT bd
Boundary forces
Body forces
BT D0 d
Initial strain
BT 0 d
Initial stress
Table 12.1: Comparison of Scalar and Vector Field Problems, Revisited
Victor Saouma
Draft
Chapter 13
TOPICS in STRUCTURAL
MECHANICS
13.1
Condensation/Substructuring
1 Condensation is the process of reducing the number of d.o.f. by substitution. At an element

level, we may want to condense an internal node (such as in the Lagrangian element), at the
global level, this is closely associated with substructuring.
2
Let us consider kd = p and let d be partitioned into

dr
{d} =
dc
(13.1)
where the subscripts r and c refer to retained and condensed d.o.f. respectively.
3
For substructuring, subscripts r and c refer to interface and interior d.o.f. Hence,

dr
pr
krr krc
=
kcr kcc
dc
pc
or
k dr = p
(13.2)
(13.3)
where
k = krr krc k1
cc kcr

p r
krc k1
cc pc
(13.4-a)
(13.4-b)
4 Noting the common form of the above two equations, we deduce that condensation may
also be viewed as a transformation from a d, p, k system to a dr , p , k through the following
transformation
k = Tc kc

Tc p
(13.5-a)
(13.5-b)
Draft
132
TOPICS in STRUCTURAL MECHANICS

where
c =
k1
cc kcr
I

(13.6)
k is a r by r matrix, whereae k was r+c by r+c (yet we still have to separately decompose
kcc which is c by c).
5
13.2
Element Evaluation
13.2.1
Patch Test
6 The patch test is a check which ascertains whether a patch of innitesimally small elements
subjected to constant strain reproduces exacly the constitutive behavior of the material through
correct stresses.
7 It has been argued that an element which passes the patch test satises the two essential
conditions for convergence.
First we assemble a small number of elements into a patch in such a way that there is at
least one internal node shared by two or more elements, and that one or more interelement
boundaries exist. Enough support must be provided to prevent rigid body motion, Fig. 13.1.
8
2F
5
20
5
F
8 and 9 noded isoparametric quadrilaterals; Pure Axial and Flexure
8 noded Isoparametric constant axial stress
Figure 13.1: Patch Tests

Boundary nodes are then loaded by a consistently derived nodal loads appropriate to a state
of constant stress. This can be achieved by applying nodal load or nodal displacement. We
then verify that a state of constant stress is obtained inside the element. This test must be
repeated for each of the appropriate strains (xx , yy and xy ).
9
10
The patch test may pass by an individual element, but fail by an assemblage of such elements.
11 Failure of the patch test is also an indication of lack of stability. That is we have zero-energy
deformation (as detected by the eigenvalue test) is present.
13.2.2
Eigenvalue Test
12 The stiness matrix is, by denition, singular due to the fact that equilibrium relations are
embedded in its formulation, or alternatively, the assumed displacement eld on which it is
based is supposed to provide for rigid body motions (translations and/or rotations).
Victor Saouma
Draft
13.2 Element Evaluation

13
133
The augmented stiness matrix may be expressed as:

[d]1 [B]T
[d]1
[k] =
[B][d]1 [B][d]1 [B]T
(13.7)
where B is the statics (or equilibrium) matrix, relating external nodal forces to internal forces;
d is a exibility matrix, and d1 is its inverse or reduced stiness matrix. We note that the
stiness matrix is obviously singular, since the second row is linearly dependent on the rst
one (through [B]) and thus, its determinent is equal to zero.
14
The reduced stiness matrix, which is the inverse of a exibility matrix, is not.
15
The internal strain energy stored in an element can be determined from

U
=
=
1
u {F}
2
1
u[K] {u}
2
(13.8)
16 If we consider a system where the load {F} applied to each node is proportional to the
element nodal displacement {u} by a factor , we have:
[K] {u} = {u}
(13.9)
([K] [I]) {u} = 0
(13.10)
or
17 This is by denition an eigenproblem. There will be as many eigenvalues i as there are

degrees of freedom (or rows in [K]). To each eigenvalue i corresponds an eigenvector {u}i .
18
If the eigenvectors are normalized such that:

ui {u}i = 1
(13.11)
ui [K] {u}i = i
(13.12)
then Eq. 13.9 becomes:

and this last equation, when compared to Eq. 13.8, shows that the eigenvalue i is equal to
twice the internal strain energy stored in an element undergoing a (normalized) deformation
dened by {u}i .
19
Hence, we observe that:

1. In a rigid body motion, all nodes displace by the same amount, and there are no internal
strains. Hence in a rigid body motion the strain energy U (and thus corresponding )
must be equal to zero.
2. There should be as many zero eigenvalues as there are possible independent rigid body
motions (i.e. number of equations of equilibrium). For a two dimensional Lagrangian
element, there should be three zero eigenvalues, corresponding to two translations and
one rotation.
3. Too few zero eigenvalues is an indication of an element lacking the capability of rigid body
motion without strain.
Victor Saouma
Draft
134
4. Too many zero eigenvalues is an indication of undesirable mechanism (or failure).

5. Eigenvalues should not change when the element is rotated.
6. Similar modes (such as exure in two orthogonal directions) will have identical eigenvalues
(for isotropic material).
7. When comparing the stiness matrices of two identical elements but based on dierent
formulations, the one with the lowest strain energy (tr[K] = i ) is best.
20
Hence, the element stiness matrix will have:
Order: corresponds to the number of degrees of freedom (i.e size of the matrix).
Rank: corresponds to the total number of linearly independent equations which is equal to the
order minus the number of rigid body motions.
Rank Deciency: would be equal to the total number of zero eigenvalues minus the rank.
21
For a square bilinear element the non-zero eigenvectors are shown in Fig. 13.2.
Figure 13.2: Eigenvectors Corresponding to a) Non-Zero and b) Zero Eigenvalues for a Square
Bilinear element
13.2.3
13.2.3.1
Order of Integration
Full Integration
For numerically integrated elements, a question of paramount importance is the order of

the numerical integration. If the order of integration is too
22
High: it would certainly result in an exact formulation, which however may be computationally
expensive.
Low: Then we may have rank deciency.
Hence, we ought to properly select the order of numerical quadrature.
23
We recall that
Victor Saouma
Draft
13.2 Element Evaluation
135
1. We seek to integrate F = BT DBdetJ.

2. The Jacobian is constant only for parallelograms.
3. In Gaussian Quadrature, integration with n points correctly integrates a polynomial of
order 2n 1 (whereas for Newton-Cotes it was n 1).
Hence, for a bilinear parallelogram element, F is a polynomial function, of order 2, in terms of
the natural coordinates and which can be exactly integrated by n = 2 points (i.e 2 2).
24 By similar arguments, a quadratic quadrilateral should be integrated by three points (F is
now of order 4, and thus 2 3 1 = 5).
Those orders of approximations are only valid for parallelograms, as the edge becomes inclined, then the error increases.
25
13.2.3.2
Reduced Integration
26 Reduced integration, i.e. using fewer Gauss points than required for an exact integration,
will cause the element to soften because higher order polynomial terms may vanish at the points
of lower order rule.
27 In an elastic body, the FE solution permits only those displacements which can be represented
by a linear combination of the shape functions. In general, the exact displacement eld will
dier from the assumed displacement, in other words the shape functions prevent the structure
from deforming the way it wants1 .
Given that the stiness formulation is too sti, (Ref. Sect. ??), the softening introduced by
a reduced numerical integration is benecial.
28
29
Recommended orders of integration for quadrilaterla elements are shown in Table 13.1.
Element
4-node
4-node distorted
8-node
8-node distorted
9-node
9-node distorted
Reliable Integration
2x2
2x2
3x3
3x3
3x3
3x3
Reduced Integration
2x2
2x2
2x2
2x2
Table 13.1: Full and Reduced Numerical Integrations for Quadrilateral Elements
.
30 If the element is articially softened, then there will be spurious zero energy modes, and
the stiness matrix will be rank decient, i.e. its rank is less than the number of element
d.o.f. minus the number of rigid body modes.
31
The independent displacement modes of a bilinear element are shown in Fig. 13.3. First
1
This is analogous to trucating a Fourrier series, which otherwise could have been an exact substitute to a
set of physical observations.
Victor Saouma
Draft
136

4
3
2
Figure 13.3: Independent Displacement Modes for a Bilinear Element

three modes are rigid body motion, the next three are constant strains, and the last two are
bending modes.
32 Since the element does not provide any resistance to all spurious zero-energy modes (which
can nevertheless be used in linear combination with the other modes to describe the displacement eld), then the eect of reduced integration is to soften the stiness of the FE model.
33 Should we adopt a reduced integration scheme, we will substantially reduce the computational eort, however we will forfeit some very special properties of the displacement method,
namely: a) Boundedness and b) monotonic convergence toward an exact solution.
34 An illustration of problems arising with reduced integration is illustrated for the following
elements
Q4 For a single point integration, the Gauss point is at the center. At the center of an element
under bending xx = yy = xy = 0, accordingly U = 0 for the corresponding modes 4
and 5 in Fig. 13.2, those spurious modes will disappear if n = 2.
Q9 Displacement is given by u = 3 2 2 2 2 and v = 0. If the element is integrated by
2x2 points, at = = 13 it can be shown that in those points the strains are zero and
the corresponding mechanisms, or hourglass are shown in Fig. 13.4.
Eight- or nine-node elements

Nine-node element only
Nine-node element only

Eight- or nine-node elements
Figure 13.4: Hourglas Modes in Under-Integrated Quadratic Element

Q8
13.2.3.3
Selective Reduced Integration
35 In selective integration, we integrate the dierent strain terms with dierent orders of integration.
Victor Saouma
Draft
13.3 Parasitic Shear/Incompatible Elements
137
Figure 13.5: Rectangular Bilinear Element Subjected to Bending; Bilinear Element and Correct
Geometry
36
More about this technique in subsequent chapter.
13.3
Parasitic Shear/Incompatible Elements
13.3.1
Q4, The Problem
37
The shape functions for the bilinear element are given by Eq. 9.19
N1 (, ) =
N3 (, ) =
1
4 (1
1
4 (1
)(1 ); N2 (, ) =
+ )(1 + ); N4 (, ) =
1
4 (1
1
4 (1
+ )(1 );
)(1 + );
(13.13)
where = x/a and = y/b.

38 Imposing the displacement eld eld of Fig. 13.5 the bilinear element and correct displacements and strains are given by Table 13.2.
u
v
xx
yy
xy
Bilinear
Element
u
0
ua
0
ub
Exact
Formulation
u
au
bu
2
(1 ) 2b + (1 2 ) 2a
ua
ua
0
Table 13.2: Bilinsear and Exact Displacements/Strains

From this table we note that parasitic shear contributes to the strain energy stored in the
element (which is equal to PT .d). Hence for a given imposed displacement, the forces would
be larger and M1 > M2 .
39
40
Computing the ratio of strain energies of the two elements, we obtain

1
1 ' a (2
1
M1
+
=
M2
1+ 1+ 2 b
(13.14)
hence, as a/b increases, the element locks (a term which will be explained in later chapters).
41
Hence, the unwanted shear strain that produces M1 > M2 is called parasitic shear.
Victor Saouma
Draft
138
Figure 13.6: Displacements Associated with Incompatible Modes for the Q6 Element
13.3.2
Q6, The Solution
42 Comparing the displacement elds in Table 13.2 we observe that the bilinear element errs
from the exact solution by omitting the displacement mode associated with (1 2 ) and (1 2 ).
43
We can remedy to this situation by adding the missing modes as internal nodes
u =
Ni ui + (1 2 )a1 + (1 2 )a2
v =
Ni v i + (1 2 )a3 + (1 2 )a4
(13.15-a)
(13.15-b)
where the Ni are given by Eq. 13.13, and ai are nodeless d.o.f. The element now has 6 degrees
of freedom, and hence the element is referred to as Q6.
The B matrix would have to be correspondingly adjusted and following its formulation, the
additional d.o.f. ai are eliminated through condensation (Sect. 13.1).
44
45 The Q6 element is incompatible (or nonconforming) due to the square distribution of u or

v along the edge which is dened by only two nodes, Fig. 13.6.
The softening in this case counters the inherent overstiness of the standard assumed displacement formulation.
46
13.3.3
47
QM6, Further Enhancements
It can be shown that the Q6 element passes the patch test only if it is rectangular.
To remedy to this deciency, we consider that portion of the B matrix associated with the
nodeless d.o.f. and label it Ba , and then the corresponding consistent nodal load vector is
e
e
BTa 0 d
(13.16)
pa =
48
where 0 represent the initial stresses caused by the standard element formulation associated
with dr on the nodal d.o.f.
49 Because the standard isoparametric element passes the patch test, when 0 is constant,
without this additional load associated with the ai , then pea should also be zero for constant 0
Victor Saouma
Draft
13.4 Rotational D.O.F.
139
Figure 13.7: Side Displacements Induced by Drilling d.o.f.

or
1
1
1
1
BTa tdetJdd = 0
(13.17)
For general (non-rectangular elements) this equation is not satised and the patch test fails.
50 This can be articaily remedied in forming Ba and integrating by using J0 where the jacobian is evaluated at = = 0 rather than at the Gauss quadrature point, and we use the
corresponding t0 for the thickness.
51 When such an operation is performed, the resulting element is called the QM6, and it works
almost as well as the quadratic element if rectangular, and is considerably more accurate than
the bilinear element.
52 Stresses are computed with all the d.o.f., including the ai . Better results are achieved if
element nodal loads pe that are associated with incompatible modes are set to zero during
recovery of the associated d.o.f. ai (following condensation).
13.4
Rotational D.O.F.
53 Drilling d.o.f. are rotations at corner nodes. In the context of plane elements, they are
associated with parabolic displaced shapes of the edges, Fig. 13.7. They are of importance
in shells dened as an assemblage of at elements. Hence, the element can model exure and
membrane action.
Assuming rotation i and j at nodes i and j of an element side of length L, the end moments
of an immaginary beam will be
54
Mi =
Victor Saouma
2EI
(2 i + j )
L
(13.18-a)
Draft
1310
Mj
2EI
( i + 2 j )
L
(13.18-b)
Hence, from M = Mi + (Mj Mi ) Lx and the vitual force method we can solve for the midside
edge-normal displacement
L
(13.19)
= ( j i )
8
55
Hence the edge displacement is caused by two parts: translation and parabolic distorsion:

s
Ls
L s ui
cos
uj
u
s(j i )
(13.20)
+
+
=
sin
vi
vj
v
L
L
2L
and similar expressions can be written for other edges.
13.5
Constraints
Victor Saouma
Draft
Chapter 14
GEOMETRIC NONLINEARITY
14.1
1
Introduction
With refernce to Fig. 14.1, we distinguish dierent levels of analysis:
Load
First order
elastic analysis
Stiffening
Bifurcation
Elastic Critical Load

Elastic Stability Limit
Bifurcation
Softening
Bifurcation
Second order
elastic analysis
Inelastic Critical Load

Plastic Limit Load
First order inelastic analysis
Inelastic Stability Limit

Second order
inelastic analysis
Displacement
Figure 14.1: Level of Analysis
First Level elastic which excludes anly nonlinearities. This is usually acceptable for service
loads.
Elastic Critical load is usually determined from an eigenvalue analysis resulting in the
buckling load.
Secon-order elastic accounts for the eects of nite deformation and displacements, equilibrium equations are written in terms of the geometry of the deformed shape, does not
account for material non-linearilties, may be able to detect bifurcation and or increased
stiness (when a member is subjected to a tensile axial load).
Draft
142
First-order inelastic equilibrium equations written in terms of the geometry of the undeformed structure, accounts for material non-linearity.
Second-order inelastic equations of equilibrium written in terms of the geometry of the
deformed shape, can account for both geometric and material nonlinearities. Most suitable
to determine failure or ultimate loads.
2 This chapter will focus on elastic critical load determination. We will begin by reviewing the
derivation of some of the fundamental equations in stability analysis. We will examine both
the strong form, and the weak formulation.
14.1.1
3
Strong Form
Column buckling theory originated with Leonhard Euler in 1744.
An initially straight member is concentrically loaded, and all bers remain elastic until buckling occur. For buckling to occur, it must be assumed that the column is slightly bent as shown
in Fig. 14.2. Note, in reality no column is either perfectly straight, and in all cases a minor
4
x and y are
principal axes
Slightly bent position

L
Figure 14.2: Euler Column

imperfection is present.
14.1.1.1
Lower Order Dierential Equation
5 At any location x along the column, the imperfection in the column compounded by the
concentric load P , gives rise to a moment
Mz = P y
(14.1)
Note that the value of y is irrelevant.

6
Recalling that
Mz
d2 y
=
dx2
EI
upon substitution, we obtain the following dierential equation
P
d2 y
y=0
2
dx
EI
Victor Saouma
(14.2)
(14.3)
Draft
14.1 Introduction
7
Letting k2 =
P
EI ,
143
the solution to this second-order linear dierential equation is

y = A sin kx B cos kx
(14.4)
The two constants are determined by applying the boundary conditions

1. y = 0 at x = 0, thus B = 0
2. y = 0 at x = L, thus
A sin kL = 0
(14.5)
This last equation can be satised if: 1) A = 0, that is there is no deection; 2) kL = 0, that
is no applied load; or 3)
kL = n
(14.6)
%
&
2
P
= n
or
Thus buckling will occur if EI
L
9
P =
n2 2 EI
L2
10 The fundamental buckling mode, i.e. a single curvature deection, will occur for n = 1; Thus
Euler critical load for a pinned column is
2 EI
L2
(14.7)
2 E
cr = % &2
L
(14.8)
Pcr =
11
The corresponding critical stress is
where I = Ar 2 .
12
Note that buckling will take place with respect to the weakest of the two axis.
14.1.1.2
Higher Order Dierential Equation
In the preceding approach, the buckling loads were obtained for a column with specied
boundary conditons. A second order dierential equation, valid specically for the member
being analyzed was used.
13
In the next approach, we derive a single fourth order equation which will be applicable to
any column regardelss of the boundary conditions.
14
15 Considering a beam-column subjected to axial and shear forces as well as a moment, Fig.
dv
between the
14.3, taking the moment about i for the beam segment and assuming the angle dx
axis of the beam and the horizontal axis is small, leads to

(dx)2
dV
dv
dM
dx + w
+ V +
dx P
dx = 0
(14.9)
M M+
dx
2
dx
dx
Victor Saouma
Draft
144
w
w(x)
1111
0000
0000
1111
0
1
0
1
0
1
0 y, u
1
1111
0000
1111
0000
dx
V+ V dx
x
P
v
x
P
dx
P
M+ M dx
x
i
j
Figure 14.3: Simply Supported Beam Column; Dierential Segment; Eect of Axial Force P
Neglecting the terms in dx2 which are small, and then dierentiating each term with respect
to x, we obtain
d2 v
dV
d2 M
=0
(14.10)
dx2
dx
dx2
16
17
However, considering equilibrium in the y direction gives

dV
= w
dx
18
(14.11)
From beam theory, neglecting axial and shear deformations, we have

M = EI
d2 v
dx2
(14.12)
19 Substituting Eq. 14.11 and 14.12 into 14.10, and assuming a beam of uniform cross section,
we obtain
d4 v
d2 v
(14.13)
EI 4 P 2 = w
dx
dx
P
Introdcing k2 = EI
, the general solution of this fourth order dierential equation to any set
of boundary conditions is
20
v = C1 sin kx + C2 cos kx + C3 x + C4
21
(14.14)
If we consider again the stability of a hinged-hinged column, the boundary conditions are
v = 0, v,xx = 0 at x = 0
v = 0, v,xx = 0 at x = L
(14.15)
substitution of the two conditions at x = 0 leads to C2 = C4 = 0. From the remaining

conditions, we obtain
C1 sin kL + C3 L = 0
(14.16-a)
C1 k sin kl = 0
(14.16-b)
these relations are satised either if C1 = C3 = 0 or if sin kl = C3 = 0. The rst alternative

leads to the trivial solution of equilibrium at all loads, and the second to kL = n for n =
1, 2, 3 . For n = 1, the critical load is
Pcr =
Victor Saouma
2 EI
L2
(14.17)
Draft
14.1 Introduction
145
which was derived earlier using the lower order dierential equation.
Next we consider a column with one end xed (at x = 0), and one end hinged (at x = L).
The boundary conditions are
22
v = 0, v,xx = 0 at x = 0
v = 0, v,x = 0 at x = L
(14.18)
These boundary conditions will yield C2 = C4 = 0, and

sin kL kL cos kL = 0
(14.19)
But since cos kL can not possibly be equal to zero, the preceding equation can be reduced to
tan kL = kL
(14.20)
which is a transcendental algebraic equation and can only be solved numerically. We are
essentially looking at
the intersection of y = x and y = tan x, Fig. 14.4 and the smallest
10.0
8.0
6.0
4.0
2.0
0.0
-2.0
-4.0
-6.0
-8.0
-10.0
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
Figure 14.4: Solution of the Tanscendental Equation for the Buckling Load of a Fixed-Hinged
Column
positive root is kL = 4.4934, since k2 =
Pcr =
P
EI ,
the smallest critical load is
(4.4934)2
2
EI =
EI
2
L
(0.699L)2
(14.21)
Note that if we were to solve for x such that v,xx = 0 (i.e. an inection point), then x = 0.699L.
We observe that in using the higher order dierential equation, we can account for both
natural and essential boundary conditions.
23
Victor Saouma
Draft
146
14.1.2
14.1.2.1
Weak Form
Strain Energy
24 Considering a uniform section prismatic element, Fig. ??, subjected to axial and exural
deformation (no shear), the Lagrangian nite strain-displacement relation is given by 3.26
1
2
2
xx = u,x + (u2,x + v,x
+ w,x
)
2
(14.22)
thus, the total strain would be

xx =
2

1 dv 2
d v
du
+
y
dx
dx2
2 dx

Axial Flexure Large Deformation
(14.23)
25 We note that the rst and second terms are the familiar components of axial and exural
strains respectively, and the third one (which is nonlinear) is obtained from large-deection
strain-displacement.
26
27
28
The Strain energy of the element is given by

1
e
E2 d
U =
2 xx
(14.24)
Substituting Eq. 14.23 into U e we obtain

2
2 2
# 2
d v
1
1 dv 4
du
du
e
2 d v
+y
+
2y
U =
2 L A
dx
dx2
4 dx
dx
dx2
2 2 2 $
dv
du
dv
d v
+
EdAdx
y
dx2
dx
dx
dx
Noting that
dA = A;
y 2 dA = I
ydA = 0;
A
4
% dv &4
dx
(14.26)
for y measured from the centroid, U e reduces to

#
4
2 2
2 $

2
dv
dv
1
du
d
A
du
v
E A
+I
+
+A
dx
Ue =
2
2 L
dx
dx
4 dx
dx
dx
We discard the highest order term
linear instability formulation.
(14.25-a)
(14.27)
in order to transform the above equation into a
29 Under the assumption of an independent prebuckling analysis for axial loading, the axial
load Px is
du
(14.28)
Px = EA
dx
Victor Saouma
Draft
14.1 Introduction
147
Thus Eq. 14.27 reduces to

1
U =
2

EA
du
dx
2

+ EI
d2 v
dx2
2

+ Px
dv
dx
2 $
dx
(14.29)
We can thus decouple the strain energy into two components, one associated with axial and
the other with exural deformations
30
U e = Uae + Ufe
2

1
du
e
EA
dx
Ua =
2 L
dx
#
2 $
2 2

dv
d
1
v
EI
dx
Ufe =
+ Px
2
2 L
dx
dx
14.1.2.2
(14.30-a)
(14.30-b)
(14.30-c)
Euler Equation
Recall, from Eq. 2.74 that for a functional in terms of two eld variables (u and v) with
higher order derivatives of the form

(14.31)
=
F (x, y, u, v, u,x , u,y , v,x , v,y , , v,yy )dxdy
31
there would be as many Euler equations as dependent eld variables, Eq. 2.75
F
F
F
2 F
2
2 F
F
u x u,x y u,y + x2 u,xx + xy u,xy + y 2 u,yy = 0
32
F
v
F
x
v,x
F
y v,y
F
+ x
2 v,xx +
F
2
xy v,xy
2 F
y 2 v,yy
(14.32)
= 0
For the problem at hand, those two equations reduce to

"
1!
e
2
2
EIv,xx
+ Px v,x
dx
Uf =
L 2

(14.33)
and the corresponding Euler equation will be
2 F
F
+ 2
=0
x v,x x v,xx
(14.34)
The terms of the Euler Equation are given by

F
v,x
F
v,xx
= Px v,x
(14.35-a)
= EIv,xx
(14.35-b)
Substituting into the Euler equation, and assuming constant Px , and EI, we obtain
EI
d4 v
d2 v
P
=0
x
dx4
dx2
(14.36)
which is identical to Eq. 14.13

Victor Saouma
Draft
148
14.2
Finite Element Discretization
33 Assuming a functional representation of the transverse displacements in terms of the four

joint displacements
v = Nu
dv
= N,x u
dx
d2 v
= N,xx u
dx2
34
(14.37-a)
(14.37-b)
(14.37-c)
Substituting this last equation into Eq. 14.30-c, the element potential energy is given by
e = Ufe + W e
1
1
ue [ke ] {ue } + ue [kg ] {ue } u {P}
=
2
2

where
(14.38-a)
(14.38-b)

EI {N,xx } N,xx dx
[ke ] =
(14.39)
and

[kg ] = P

{N,x } N,x dx
(14.40)
where [ke ] is the conventional element exural stiness matrix.

[kg ] introduces the considerations related to elastic instability. We note that its terms solely
depend on geometric parameters (length), therefore this matrix is often referred to as the
geometric stiness matrix.
35
Using the shape functions for exural elements, Eq. 6.41, and substituting into Eq. 14.39
and Eq. 14.40 we obtain
36
u1
EA
L
ke = EA
L
0
0
v1
0
1
0
12EI
L3
6EI
L2
6EI
L2
4EI
L
12EI
L3
6EI
L2
6EI
L2
2EI
L
u2
EA
L
0
0
EA
L
0
0
v2
0
12EI
L3
6EI
L2
0
12EI
L3
6EI
L2
0
6EI
L2
2EI
L
6EI
L2
(14.41)
4EI
L
which is the same element stiness matrix derived earlier in Eq. 8.7.
37
The geometric stiness matrix is given by

u1
0
0
P
0
kg =
L 0
0
0
Victor Saouma
v1
0
1
0
6
5
L
10
L
10
2 2
15 L
0
65
L
10
0
L
10
2
L30
u2 v2
2
0
0
0
L
0 65
10
2
L
0 10 L30
(14.42)
0
0
0
6
L
0
10
5 Elements
Finite
II; Solid Mechanics
2 2
L
0 10 15 L
Draft
14.3 Elastic Instability; Bifurcation Analysis

38
149
The equilibrium relation is thus

ku = P
(14.43)
where the element stiness matrix is expressed in terms of both the elastic and geometric
components)
39
k = ke + kg
(14.44)
K = Ke + Kg
(14.45)
In a global formulation,we would have
we note that the structure becomes stier for tensile load P applied through Kg , and weaker
in compression.
We assume that conservative loading is applied, that is the direction of the load does not
follow the deected direction of the member upon which it acts.
40
14.3
Elastic Instability; Bifurcation Analysis
41 In elastic instability, the intensity of the axial load system to cause buckling is yet unknown,
the incremental stiness matrix must rst be numerically evaluated using an arbitrary chosen
load intensity (since Kg is itself a function of P ).
42 For buckling to occur, the intensity of the axial load system must be times the initially
arbitrarily chosen intensity of the force. Note that for a structure, the initial distribution of P
must be obtained from a linear elastic analysis. Hence, the buckling load, P is given by
P = P
(14.46)
43 Since the geometric stiness matrix is proportional to the internal forces at the start, it
follows that
(14.47)
Kg = Kg
where Kg corresponds to the geometric stiness matrix for unit values of the applied loading
( = 1).
44
The elastic stiness matrix Ke remains a constant, hence we can write
(Ke + Kg )u P = 0
45
The displacements are in turn given by
u = (Ke + Kg )1 P
and for the displacements to tend toward innity (i.e buckling/bifurcation/instability), then
|Ke + Kg | = 0
Victor Saouma
(14.48)
Draft
1410
which can also be expressed as

|K1
g Ke + I| = 0
(14.49)
46 Alternatively, it can simply be argued that there is no unique solution (bifurcation condition)
to u.
The lowest value of , crit will give the buckling load for the structure and the buckling
loads will be given by
47
Pcrit = crit P
48
(14.50)
The corresponding deformed shape is directly obtained from the corresponding eigenvector.
Example 14-1: Column Stability

Determine the buckling load of the following column.
P
011
1010 2
10
1010
1010
10
1010
6 1
0 5
100
4 1
10
1010
1010
010 8
9 1
10 000
111
10107
1
0
0
1
0
1
11
00
0110
10
1010
1010
1010
1010
10
1010
1010
1010
Solution:
The following elastic stiness matrices are obtained
k1e =
Victor Saouma
1
EA
L
EA
L
0
0
2
0
3
0
12EI
L3
6EI
L2
6EI
L2
4EI
L
12EI
L3
6EI
L2
6EI
L2
2EI
L
4
EA
L
0
0
EA
L
0
0
5
0
6
0
12EI
L3
6EI
L2
0
6EI
L2
2EI
L
12EI
L3
6EI
L2
6EI
L2
(14.51-a)
4EI
L
Draft
k2e =
4
EA
L
EA
L
0
0
5
0
6
0
12EI
L3
6EI
L2
6EI
L2
4EI
L
12EI
L3
6EI
L2
6EI
L2
1411
7
2EI
L
EA
L
0
0
EA
L
0
0
8
0
9
0
12EI
L3
6EI
L2
0
6EI
L2
2EI
L
12EI
L3
6EI
L2
6EI
L2
(14.51-b)
4EI
L
Similarly, the geometric stiness matrices are given by

1 2
0 0
0 6
5
L
P
0
10
L 0 0
0 65
L
0 10
k1g =
4 5
0 65
L
0 10
P
0 0
L
0 65
L
0 10
k2g =
3
0
L
10
2 2
15 L
0
L
10
2
L30
6
L
10
2 2
15 L
0
L
10
2
L30
4
5
0
0
0 65
L
0 10
0
0
6
0
5
L
0 10
7
8
0 65
L
0 10
0
0
6
0
5
L
0 10
0
L
10
2
L30
0
L
10
2 2
15 L
9
L
10
2
L30
(14.52-a)
0
L
10
2 2
15 L
(14.52-b)
The structures stiness matrices Ke and Kg can now be assembled from the element stinesses. Eliminating rows and columns 2, 7, 8, 9 corresponding to zero displacements in the
column, we obtain
1
4
3
5
6
AL2
AL2
I
0
0
0
I
AL2 2 AL2
0
0
0
I
I
EI
2
(14.53)
Ke = 3 0
6L 2L2
0
4L
L
0
0
6L 24
0
0
8L2
0
0
2L2
and
1
0
0
P
0
Kg =
L
0
0
Victor Saouma
4
0
0
0
0
0
3
0
0
2 2
15 L
L
10
L2
30
5
0
0
L
10
12
5
0
0
L2
30
0
(14.54)
4 2
15 L
Draft
1412
noting that in this case Kg = Kg for P = 1, the determinant |Ke + Kg | = 0 leads to

1
AL2
1 I

2
4 AL
I

3 0

5 0
6 0
introducing =
AL2
I
4
2
AL
I
2
2 AL
I
0
0
0
3
0
0
2 L4
2
4L 15
EI
1 L3
6L + 10
EI
1 L4
2L2 + 30
EI
and =
L2
EI ,
5
0
0
1 L3
6L + 10
EI
L2
24 12
5 EI
0
6
0
0
1 L4
2
2L + 30
EI
0
4 L4
8L2 15
EI

=0

(14.55)
the determinant becomes
1
4
3
5
6

1
0
0
0

4 2
0
0
0

%
&

6L
+ 10&
3 0
0 2 2 15
2 + 30 = 0
%

5 0
0
6L + 10
12 2 5
% 0 &

0
4 2 15
6 0
0
2 + 30
(14.56)
Expanding the determinant, we obtain the cubic equation in

33 2202 + 3, 840 14, 400 = 0
(14.57)
and the lowest root of this equation is = 5.1772 .

We note that from Eq. 14.21, the exact solution for a column of length L was
Pcr =
(4.4934)2
(4.4934)2
EI
=
EI = 5.0477 EI
L2
l2
(2L)2
(14.58)
and thus, the numerical value is about 2.6 percent higher than the exact one. The mathematica
code for this operation is:
(* Define elastic stiffness matrices
ke[e_,a_,l_,i_]:={
{e a/l , 0
, 0
,
{0
, 12 e i/l^3 , 6 e i/l^2 ,
{0
, 6 e i/l^2
, 4 e i/l
,
{-e a/l , 0
, 0
,
{ 0
, -12 e i/l^3 , -6 e i/l^2 ,
{ 0
, 6e i/l^2
, 2 e i/l
,
}
ke1=N[ke[e,a,l,i]]
ke2=N[ke[e,a,l,i]]
*)
-e a/l
0
0
e a/l
0
0
,
,
,
,
,
,
0
-12 e i/l^3
-6 e i/l^2
0
12 e i/l^3
-6 e i/l^2
,
,
,
,
,
,
0
6 e i/l^2
2 e i/l
0
-6 e i/l^2
4 e i/l
},
},
},
},
},
}
(* Assemble structure elastic stiffness matrices *)

ke={
{ke1[[3,3]], ke1[[3,5]]
, ke1[[3,6]]
},
{ ke1[[5,3]], ke1[[5,5]]+ke2[[2,2]], ke1[[5,6]]+ke2[[2,3]]},
{ ke1[[6,3]], ke1[[6,5]]+ke2[[3,2]], ke1[[6,6]]+ke2[[3,3]]}
}
WriteString["mat.out",MatrixForm[ke1]]
WriteString["mat.out",MatrixForm[ke]]
(* Define geometric stiffness matrices *)
kg[p_,l_]:=p/l{
Victor Saouma
Draft
1413
{0 , 0
, 0
, 0 , 0
, 0
},
{0 , 6/5 , l/10
, 0 , - 6/5 , l/10
},
{0 , l/10 , 2 l^2/15 , 0 , - l/10 , - l^2/30 },
{0 , 0
, 0
, 0 , 0
, 0
},
{0 , -6/5 , - l/10
, 0 , 6/5
, - l/10
},
{0 , l/10 , - l^2/30 , 0 , - l/10 , 2 l^2/15 }
}
kg1=kg[p,l]
kg2=kg[p,l]
(* Assemble structure geometric stiffness matrices *)
kg={
{kg1[[3,3]], kg1[[3,5]]
, kg1[[3,6]]
},
{ kg1[[5,3]], kg1[[5,5]]+kg2[[2,2]], kg1[[5,6]]+kg2[[2,3]]},
{ kg1[[6,3]], kg1[[6,5]]+kg2[[3,2]], kg1[[6,6]]+kg2[[3,3]]}
}
(* Determine critical loads in terms of p (note p=1) *)
p=1
keigen= l^2 (Inverse[kg] . ke)/( e i)
pcrit=N[Eigenvalues[keigen]]
(* Alternatively*)
knew =ke - x kg
pcrit2=NSolve[Det[knew]==0,x]
Example 14-2: Frame Stability

Determine the buckling load for the following frame. Neglect axial deformation.
2
1
0
0
1
0
1
0
1
11
00
u1
3
u1
I=200
I=50
10
I=100
1111
0000
0000
1111
15
11111111111111111111
00000000000000000000
Solution:
The element stiness matrices are given by
u1
20
1, 208
k1e =
Victor Saouma
2
1, 208
96, 667
(14.59-a)
Draft
1414
u1
0.01
0.10
= P

k1g
k2e =
128, 890
64, 440
u1
47
1, 678
k3e =
3
1, 678
80, 556
u1
0.01667
0.1
= P

k3g
2
0.10
16.00
(14.59-b)
64, 440

128, 890
3
0.1
9.6
(14.59-c)
(14.59-d)
(14.59-e)
The global equilibrium relation can now be written as

(Ke P Kg ) = 0
u1

(66.75) P (0.026666)

(1, 208.33) P (0.1)

(1, 678.24) P (0.1)
2
(1, 208.33) P (0.1)
(225, 556.) P (16.)
(64, 444.) P (0)
(14.60)
3

(1, 678.24) P (0.1)
(64, 444.4) P (0) = 0
(209, 444.) P (9.6)
(14.61)
The smallest buckling load amplication factor is thus equal to 2, 017 kips.
(* Initialize constants *)
a1=0
a2=0
a3=0
i1=100
i2=200
i3=50
l1=10 12
l2=15 12
l3=6 12
e1=29000
e2=e1
e3=e1
(* Define elastic stiffness matrices
ke[e_,a_,l_,i_]:={
{e a/l , 0
, 0
,
{0
, 12 e i/l^3 , 6 e i/l^2 ,
{0
, 6 e i/l^2
, 4 e i/l
,
{-e a/l , 0
, 0
,
{ 0
, -12 e i/l^3 , -6 e i/l^2 ,
Victor Saouma
*)
-e a/l
0
0
e a/l
0
,
,
,
,
,
0
-12 e i/l^3
-6 e i/l^2
0
12 e i/l^3
,
,
,
,
,
0
6 e i/l^2
2 e i/l
0
-6 e i/l^2
},
},
},
},
},
Draft
1415
{ 0
, 6e i/l^2
, 2 e i/l
, 0
, -6 e i/l^2 , 4 e i/l
}
}
ke1=ke[e1,a1,l1,i1]
ke2=ke[e2,a2,l2,i2]
ke3=ke[e3,a3,l3,i3]
kg[l_,p_]:=p/l{
{0 , 0
, 0
, 0 , 0
, 0
},
{0 , 6/5 , l/10
, 0 , - 6/5 , l/10
},
{0 , l/10 , 2 l^2/15 , 0 , - l/10 , - l^2/30 },
{0 , 0
, 0
, 0 , 0
, 0
},
{0 , -6/5 , - l/10
, 0 , 6/5
, - l/10
},
{0 , l/10 , - l^2/30 , 0 , - l/10 , 2 l^2/15 }
}
kg1=kg[l1,1]
kg3=kg[l3,1]
(* Assemble structure elastic and geometric stiffness matrices *)
ke={
{ ke1[[2,2]]+ke3[[2,2]] , ke1[[2,3]]
, ke3[[2,3]]
},
{ ke1[[3,2]]
, ke1[[3,3]]+ke2[[3,3]] , ke2[[3,6]]
},
{ ke3[[3,2]]
, ke2[[6,3]]
, ke2[[6,6]]+ke3[[3,3]] }
}
kg={
{ kg1[[2,2]]+kg3[[2,2]] , kg1[[2,3]] , kg3[[2,3]] },
{ kg1[[3,2]]
, kg1[[3,3]] , 0
},
{ kg3[[3,2]]
, 0
, kg3[[3,3]] }
}
(* Determine critical loads in terms of p (note p=1) *)
p=1
keigen=Inverse[kg] . ke
pcrit=N[Eigenvalues[keigen]]
modshap=N[Eigensystems[keigen]]
14.4
Second-Order Elastic Analysis; Geometric Non-Linearity
49 From Eq. 14.44 it is evident that since kg depends on the magnitude of Px , which itself may
be an unknown in a framework, then we do have a geometrically non-linear problem.
50
We rewrite Eq. 14.45 as

[Ke + Kg ] u = P
(14.62)
but since Kg depends on the axial laod P , the preoblem is nonlinear.

51 A simple way to solve this nonlinear equation is to use a step-by-step incremental procedure. The linearized incremental formulation can be obtained by applying an incremental
operator
* +
P i = [Ke + Kg ]i1 {u}i (14.63)
* +
{u}i = [Ke + Kg ]1
i1 P i (14.64)
Victor Saouma
Draft
1416
Example 14-3: Eect of Axial Load on Flexural Deformation

Determine the midspan displacement and member end forces for the beam-column shown
below in terms of Px ; The concentrated force is 50 kN applied at midspan, E=2 109 kN/m2
and I=2 103 m4 .
50 kN
1
0
0
1
0
1
0
1
0
1
0
80,000 kN1
0
1
1
0
0
1
0
1
0
1
0000000000000000
1
0
6
m
6
m
111111111111111000000000000000
1111111111111111
80,000 kN
Solution:
Using two elements for the beam column, the only degrees of freedom are the deection and
rotation at midspan (we neglect the axial deformation).
The element stiness and geometric matrices are given by
0
0
0
0
0 222, 222.
0 666, 666.
1
[Ke ] =
0
0
0 222, 222.
0 666, 666.
0
0
666, 666.
2, 666, 666
0
666, 666.
1, 333, 333
0
0
0
0.
0
0
0.
v1
0
222, 222.
666, 666.
0
222, 222.
666, 666.
0
666, 666.
1, 333, 333
666, 666.
2, 666, 666
0
v1
0
0
0 222, 222.
0 666, 666.
2
[Ke ] =
0
0
0 222, 222.
0 666, 666.
2
0
666, 666.
2, 666, 666
0
666, 666.
1, 333, 333
0
0
0
0.
0
0
0.
0
0
222, 222.
666, 666.
0
222, 222.
666, 666.
0
666, 666.
1, 333, 333
666, 666.
(14.65)
(14.66)
2, 666, 666
0
0
0
0
0 16, 000
0 8, 000
1
[Kg ] =
0
0
0 16, 000
0 8, 000
0
0
8, 000
64, 000
0
8, 000
16, 000
0
v1
0
0
0
16, 000
0
8, 000
0
0
0 16, 000
0.
8, 000
0
v1
0
0
0 16, 000
0 8, 000
2
[Kg ] =
0
0
0 16, 000
0 8, 000
Victor Saouma
2
0
8, 000
64, 000
0
8, 000
16, 000
0
0
0
0
0
0
0
16, 000
8, 000
0
8, 000
16, 000
(14.68)
0
0
0
0 16, 000
8, 000
0.
8, 000
64, 000
0
8, 000
16, 000
8, 000
64, 000
(14.67)
Draft
1417
Assembling the stiness and geometric matrices we get

[K] =
v1
412, 444.
0.
2

0.
5, 205, 330
(14.69)
and the displacements would be

and the
Plf t
Vlf t
Mlf t
Prgt
Vrgt
Mrgt
v1
2

=
0.00012123
0

(14.70)
member end forces for element 1 are given by
ulf t
lf
t
"
! 1
lf t
1
= [Ke ] + [Kg ]
urgt
vrgt
rgt
0
0
0
0
0 206, 222.
0 658, 667.
=
0
0
0 206, 222
0 658, 667.
25.
79.8491
=
0.
25.
79.8491
0
0
658, 667.
260, 2670
0
658, 667.
1, 349, 330
0
v1
0
0
0 206, 222.
0 658, 667.
0
0
0 206, 222.
0 658, 667.
,0
0
658, 667.
0
1, 349, 330
0
0
0
658, 667.
0.00012123
2, 602, 670
0
(14.71-a)
Note that had we not accounted for the axial forces, then

0.0001125
v1
=
2
0
0
Plf t
V
25.
lf
t
Mlf t
75.
=
Prgt
0.
Vrgt
25.
Mrgt
75.
(14.72-a)
(14.72-b)
Alternatively, if instead of having a compressive force, we had a tensile force, then

0.000104944
v1
=
(14.73-a)
2
0
Victor Saouma
Draft
1418
Plf t
Vlf t
Mlf t
Prgt
rgt
Mrgt
0
25.
70.8022
0.
25.
70.8022
(14.73-b)
We observe that the compressive force increased the displacements and the end moments,
whereas a tensile one stiens the structure by reducing them.
The Mathematica to solve this problem follows
OpenWrite["mat.out"]
a1=0
a2=0
e=2 10^9
i=2 10^(-3)
i1=i
i2=i1
l=6
l1=l
l2=6
p=-80000 (* negative compression *)
load={-50,0}
(* Define elastic stiffness matrices *)
ke[e_,a_,l_,i_]:={
{e a/l , 0
, 0
, -e a/l , 0
{0
, 12 e i/l^3 , 6 e i/l^2 , 0
, -12 e i/l^3
{0
, 6 e i/l^2
, 4 e i/l
, 0
, -6 e i/l^2
{-e a/l , 0
, 0
, e a/l , 0
{ 0
, -12 e i/l^3 , -6 e i/l^2 , 0
, 12 e i/l^3
{ 0
, 6e i/l^2
, 2 e i/l
, 0
, -6 e i/l^2
}
ke1=N[ke[e,a1,l1,i1]]
ke2=N[ke[e,a2,l2,i2]]
(* Assemble structure elastic stiffness matrices *)
ke=N[{
{ ke1[[5,5]]+ke2[[2,2]], ke1[[5,6]]+ke2[[2,3]]},
{ ke1[[6,5]]+ke2[[3,2]], ke1[[6,6]]+ke2[[3,3]]}
}]
WriteString["mat.out",MatrixForm[ke]]
kg[p_,l_]:=p/l {
{0 , 0
, 0
, 0 , 0
, 0
},
{0 , 6/5 , l/10
, 0 , - 6/5 , l/10
},
{0 , l/10 , 2 l^2/15 , 0 , - l/10 , - l^2/30 },
{0 , 0
, 0
, 0 , 0
, 0
},
{0 , -6/5 , - l/10
, 0 , 6/5
, - l/10
},
{0 , l/10 , - l^2/30 , 0 , - l/10 , 2 l^2/15 }
}
kg1=N[kg[p,l1]]
kg2=N[kg[p,l2]]
(* Assemble structure geometric stiffness matrices *)
kg=N[{
{ kg1[[5,5]]+kg2[[2,2]], kg1[[5,6]]+kg2[[2,3]]},
{ kg1[[6,5]]+kg2[[3,2]], kg1[[6,6]]+kg2[[3,3]]}
Victor Saouma
,
,
,
,
,
,
0
6 e i/l^2
2 e i/l
0
-6 e i/l^2
4 e i/l
},
},
},
},
},
}
Draft
1419
}]
(* Determine critical loads and normalize wrt p *)
keigen=Inverse[kg] . ke
pcrit=N[Eigenvalues[keigen] p]
(* Note that this gives lowest pcrit=1.11 10^6, exact value is 1.095 10^6 *)
(* Add elastic to geometric structure stiffness matrices *)
k=ke+kg
(* Invert stiffness matrix and solve for displacements *)
km1=Inverse[k]
dis=N[km1 . load]
(* Displacements of element 1*)
dis1={0, 0, 0, 0, dis[[1]], dis[[2]]}
k1=ke1+kg1
(* Member end forces for element 1 with axial forces *)
endfrc1=N[k1 . dis1]
(* Member end forces for element 1 without axial forces
knopm1=Inverse[ke]
disnop=N[knopm1 . load]
disnop1={0, 0, 0, 0, disnop[[1]], disnop[[2]]}
(* Displacements of element 1*)
endfrcnop1=N[ke1 . disnop1]
*)
Example 14-4: Bifurcation

Analyse the stability of the following structure. Compare the axial force caused by the
coupled membrane/exural eects with the case where there is no interaction.
1,000
1
0
0
1
0
1
0000000000
1111111111
12
0000000000
1111111111
0
1
0000000000
1111111111
0000000000
1111111111
12
0
1
0000000000
1111111111
0000000000
1111111111
0000000000
00000000001111111111
1111111111
/8
/8
0
1
0
1
0
1
0
1
Solution:
In the following solution, we will rst determine the axial forces based on the elastic stiness
matrix only. Then, on the basis of those axial forces, we shall determine the geometric stiness
matrix, and solve for the displacements. Because of the non-linearity of the problem, we may
have to iterate in order to reach convergence. Following each analysis, we shall recompute the
geometric stiness matrix on the basis of the axial loads detemined from the previous iteration.
Note that convergence will be reached only for stable problems. If the method fails to
converge, it implies possible biurcation which could be caused by elastic displacements approaching L sin , due to either being too small, or E being too small (i.e not sti enough).
Victor Saouma
Draft
1420
NEEDS SOME CORRECTION

a1 = 1
a2 = 1
i1 = 1 1^3/12
i2 = i1
l1 = 12
l2 = 12
e1 = 200000
e2 = e1
e3 = e1
theta1 =N[Pi/8]
theta2 = Pi-theta1
load ={0, -1000, 0}
normold = 0
epsilon = 0.01
puncpl = load[[2]] / (Sin[theta1] 2)
(*
Define elastic stiffness matrices
*)
ke[e_,a_,l_,i_] := {
{e a/l , 0
, 0
, -e a/l , 0
{0
, 12 e i/l^3 , 6 e i/l^2 , 0
, -12 e i/l^3
{0
, 6 e i/l^2
, 4 e i/l
, 0
, -6 e i/l^2
{-e a/l , 0
, 0
, e a/l , 0
{ 0
, -12 e i/l^3 , -6 e i/l^2 , 0
, 12 e i/l^3
{ 0
, 6e i/l^2
, 2 e i/l
, 0
, -6 e i/l^2
}
(*
Define geometric stiffness matrix
*)
kg[l_,p_] := p/l {
{0 , 0
, 0
, 0 , 0
, 0
},
{0 , 6/5 , l/10
, 0 , - 6/5 , l/10
},
{0 , l/10 , 2 l^2/15 , 0 , - l/10 , - l^2/30 },
{0 , 0
, 0
, 0 , 0
, 0
},
{0 , -6/5 , - l/10
, 0 , 6/5
, - l/10
},
{0 , l/10 , - l^2/30 , 0 , - l/10 , 2 l^2/15 }
}
(*
Define Transformation matrix and its transpose
*)
gam[theta_] := {
{ Cos[theta] , Sin[theta], 0 , 0
, 0
,
{ -Sin[theta], Cos[theta], 0 , 0
, 0
,
{ 0
, 0
, 1 , 0
, 0
,
{ 0
, 0
, 0 , Cos[theta] , Sin[theta] ,
{ 0
, 0
, 0 , -Sin[theta] , Cos[theta] ,
{ 0
, 0
, 0 , 0
, 0
,
}
gamt[theta_] := {
{ Cos[theta] , -Sin[theta], 0 , 0
, 0
{ Sin[theta] , Cos[theta] , 0 , 0
, 0
{ 0
, 0
, 1 , 0
, 0
{ 0
, 0
, 0 , Cos[theta] , -Sin[theta]
{ 0
, 0
, 0 , Sin[theta] , Cos[theta]
{ 0
, 0
, 0 , 0
, 0
}
Victor Saouma
,
,
,
,
,
,
0
6 e i/l^2
2 e i/l
0
-6 e i/l^2
4 e i/l
0
0
0
0
0
1
},
},
},
},
},
}
,
,
,
,
,
,
0
0
0
0
0
1
},
},
},
},
},
}
},
},
},
},
},
}
Draft

(*
Define functions for local displacments and loads
*)
u[theta_,v1_,v2_] := Cos[theta] v1 + Sin[theta] v2
(*
Transformation and transpose matrices
*)
gam1 = gam[theta1]
gam2 = gam[theta2]
gam1t = gamt[theta1]
gam2t = gamt[theta2]
(*
Element elastic stiffness matrices
*)
ke1 = ke[e1, a1, l1, i1]
ke2 = ke[e2, a2, l2, i2]
Ke1 = gam1t . ke1 . gam1
Ke2 = gam2t . ke2 . gam2
(*
Structures global stiffness matrix
*)
Ke={
{ Ke1[[4,4]] + Ke2[[1,1]] , Ke1[[4,5]] + Ke2[[1,2]] , Ke1[[4,6]] +
{ Ke1[[5,4]] + Ke2[[2,1]] , Ke1[[5,5]] + Ke2[[2,2]] , Ke1[[5,6]] +
{ Ke1[[6,4]] + Ke2[[3,1]] , Ke1[[6,5]] + Ke2[[3,2]] , Ke1[[6,6]] +
}
(*
======= uncoupled analysis ==========
*)
dise=Inverse[Ke].load
u[theta_,diseg1_,diseg2_] := Cos[theta] diseg1 + Sin[theta] diseg2
uu1 = u[ theta1, dise[[1]], dise[[2]] ]
uu2 = u[ theta2, dise[[1]], dise[[2]] ]
up1 = a1 e1 uu1/l1
up2 = a2 e2 uu2/l2
(*
========== Coupled Nonlinear Analysis ==============
Start Iteration
*)
diseg = N[dise]
For[ iter = 1 , iter <= 100, ++iter,
(* displacements in local coordinates *)
disloc={ 0,0,0,
u[ theta1, diseg[[1]], diseg[[2]] ],
u[ theta2, diseg[[1]], diseg[[2]] ],
0};
(* local force *)
ploc = ke1 . disloc;
p1 = ploc[[4]];
p2 = p1;
kg1 = kg[ l1 , p1 ];
kg2 = kg[ l2 , p2 ];
Kg1 = gam1t . kg1 . gam1;
Kg2 = gam2t . kg2 . gam2;
Kg={
{ Kg1[[4,4]] + Kg2[[1,1]] , Kg1[[4,5]] + Kg2[[1,2]] , Kg1[[4,6]] +
{ Kg1[[5,4]] + Kg2[[2,1]] , Kg1[[5,5]] + Kg2[[2,2]] , Kg1[[5,6]] +
{ Kg1[[6,4]] + Kg2[[3,1]] , Kg1[[6,5]] + Kg2[[3,2]] , Kg1[[6,6]] +
};
Victor Saouma
1421
Ke2[[1,3]] },
Ke2[[2,3]] },
Ke2[[3,3]] }
Kg2[[1,3]] },
Kg2[[2,3]] },
Kg2[[3,3]] }
Draft
1422
(*
Solve
*)
Ks = Ke + Kg;
diseg = Inverse[Ks] . load;
normnew = Sqrt[ diseg . diseg ];
ratio = ( normnew-normold ) / normnew;
Print["Iteration ",N[iter],"; u1 ",N[u1],"; p1 ",N[p1]," ratio ",N[ratio]];
normold = normnew;
If[ Abs[ ratio ] < epsilon, Break[] ]
]
Print[" p1 ",N[p1]," up1 ",N[up1]," p1/up1 ",N[p1/up1]," ratio ",N[ratio]]
14.5
Summary
Victor Saouma
Draft
14.5 Summary
1423
STRONG FORM
WEAK FORM
2nd Order D.E.

2 B.C.
4th Order D.E.

4 B.C.
x =
du
dx
d2 y
dx2
=0
v = A sin kx B cos kx
P
EI
U=
1
2
'
(
2
d v
dx2
d4 v
EI dx
4
d2 v
dx2
P
=w
v = C1 sin kx + C2 cos kx + C3 x + C4
1
2
% dv &2
dx
2
E d
K=
Kg
Ke
P = (Ke + Kg )u
|Ke + Kg | = 0
Figure 14.5: Summary of Stability Solutions
Victor Saouma
Draft
1424
Victor Saouma
Draft
Chapter 15
PLATES
Adapted from Pilkey & Wunderlich
Draft
1 This chapter will cover the transverse deformation of plates. The approach followed will be
consistent with the nite element formulation, Fig. 15.1.
Fundamental Relations
Kinematics
Constitutive
Equilibrium
Dierential Equation
Variational Formulation
Finite Element Discretization
Figure 15.1: Finite Element Formulation
15.1
15.1.1
Equilibrium
Considering an arbitrary plate, the stress are given by, Fig. 15.2, resultants per unit width
Draft
152
PLATES
Figure 15.2: Stresses in a Plate

are given by
Nxx
2
dz
Nyy
Membrane Force N =
2t
Nxy
Mxx
2
zdz
Bending Moments M =
Myy
2t
Mxy

Transverse Shear Forces
V =
t
2
2t
dz

=

=

=

=

=

=
Vx =
Vy =
t
2
t
2
2t
t
2
2t
t
2
2t
xx dz
yy dz
xy dz
t
2
2t
t
2
2t
t
2
2t
2t
t
2
2t
xx zdz
(15.1-a)
yy zdz
xy zdz
xz dz
yz dz
Note that in plate theory, we ignore the eect of the membrane forces, those in turn will be
accounted for in shells.
3
4 The equation of equilibrium is derived by considering an innitesimal element tdx dy subjected to an applied transverse load pz . We would have to consider three equations of equilibrium, Fig. 15.3:
Summation of Forces in the z direction
or
Victor Saouma
Vy
Vx
dxdy +
dxdy + pz dxdy = 0
x
y
(15.2)
Vx Vy
+
+ pz = 0
x
y
(15.3)
Draft
Draft
15.1 Fundamental Relations
153
x
-
Vy
;
;
6 Mxy
;
;
Myy
;
;
z ?V
-;
x
6 Mxx
pz

;;
;
-?

;
Mxy
Mxx + @M@x dx ;

Mxy + @M@x dx

?
;
Myy + @M@y dy ;;
Vx + @V@x dx
?Vy + @V@y dy
;
Mxy + @M@y dy
xx
yy
xy
xy
Figure 15.3: Free Body Diagram of an Innitesimal Plate Element

Summation of Moments about the x axis
or
Myy
Mxy
dxdy +
dxdy Vy dxdy = 0
x
y
(15.4)
Myy
Mxy
+
Vy = 0
x
y
(15.5)
Summation of Moments about the y axis

Myx Mxx
+
Vx = 0
y
x
(15.6)
Since Mxy = Myx , those equations can be expressed in matrix form as
0
0
1
0
0
1
0 x
y

LT
Mxx
Myy
Mxy
Vx
Vy

M

0 0
+
0
=
0
p
0
(15.7)
Note that the left matrix corresponds to LT where the 1 term has been substituted by 1
Victor Saouma
Draft
154
15.1.2
PLATES
Kinematic Relations
5 From Fig. 15.4 we have ve displacements u, v, w, xx and yy . However, two of the three
displacements, u, v can be expressed in terms of the xx and yy which are the rotations of the
Figure 15.4: Displacements in a Plate

plate middle surface. The third displacement being the transverse one w.
We will use a notation consistent with the traditional one adopted for plate bending, rather
than one consistent with the coordinate directions as used in nite element.
6
7 The middle surface will be assumed to remain unstrained, and that plane sections remain
plane. When shear deformations will later be taken into account, then the plane will not
necessarily remain normal to the middle surface.
8
Based on the plane section remaining plane

z
u
= 0
v
w
0
However, from Eq. ?? we have = Lu or

x
xx
yy
zz
xy
xz
yz
0

assumption, we have,
0 0 xx
yy
z 0
w
0 1

L
0
0
0

x
u
v
w

(15.8)
(15.9)
10 Since w corresponds to the transverse deection of the middle surface, and does not vary
with z, then zz = 0.
Victor Saouma
Draft
15.1 Fundamental Relations

11
155
Substituting Eq. 15.8, we obtain

zxx,x
xx
zyy,y
yy
zz
0
=
+ yy,x )
z(
xy
xx,y

(xx + w,x )
xz
yz
(yy + w,y )
zxx
zyy
0
=
2z
xy
xz
yz
(15.10)
Note that this equation assumes that the displacement w is small compared to the thickness of
the plate, and the rotation is small. Because the rotation is small, its square is negligible with
respect to unity, and hence the curvature
=
(1
x
+ 2 )3/2
(15.11)
is equal to the rate of change of rotation.

12
Hence, the kinematic relation for the transverse displacement of a plate is

0
0
x
xx
0
yy

xx
y

0
2xy
yy
= y
1
0 x
xz

yz
0
1 y
u

15.1.3
13
(15.12)
Constitutive Relations
For three dimensional continuum, the strain-stress relation is
xx
yy
0
1
zz
1
=
xy
2(1 + )
0
0
0
2(1
+
)
0
0
xz
yz
0
0
2(1 + )
But From Eq. 15.10 zz = 0, and we can neglect zz which is

stresses. Hence, inverting the previous equation yields
1
0
xx
0
0
yy
E
0 0 1
xy
=
2
1 2
1
0
xz
0
1
yz
0
2

D
Note that the shear modulus is =
Victor Saouma
xx
yy
zz
xy
xz
yz
(15.13)
much smaller than the other

xx
yy
xy
xz
yz
(15.14)
E
2(1+) .
Draft
156
PLATES
Figure 15.5: Positive Moments and Rotations

We now seek to write moments in terms of the curvatures, Fig. 15.5, introducing the stresses
from Eq. 15.14 into Eq. 15.1-a, and using Eq. 15.10 we derive the rst term
14

Mxx =
=
t
2
E
E
xx zdz =
(xx + yy )zdz =
(zxx + zyy )zdz
t
t 1 2
t 1 2
2
2
2
t
2
Et3
E
2
(xx + yy )
(
+
)
z
dz
=
(15.15-a)
xx
yy
1 2
12(1 2 )
t
t
2
t
2
15 Following a similar procedure for the other two

curvature relation
1
Mxx
Et3

Myy
=
12(1 2 )
Mxy
0

terms, we obtain the following moment

1
0
0
0
1
2

xx
yy
2xy

(15.16)
Et
The 12(1
2 ) term is referred to as the exural rigidity and is analogous to the exural
stiness EI of a beam (if the plate has unit width, and = 0, then EI = Et3 /12).
16
15.2
Plate Theories
15.2.1
Reissner-Mindlin
This theory, is primarily applicable to thick plates in which shear deformations can not be
neglected.
17
15.2.1.1
In this formulation, shear deformation is accounted for and is the extension of the Timoshenkos beam theory.
18
Victor Saouma
Draft
15.2 Plate Theories
157
Kinematic: The kinematic relation of Eq. 15.12 still holds and is written in slightly modied
form

0
xx
x xx
(15.17-a)
yy
= 0 y
yy
2xy

y
x

Lf
w

xz
x 1 0
=
(15.17-b)
yz
xx
y 0 1

yy

Ls
u
Equilibrium: The equilibrium equation, Eq. 15.7 also
Mxx
0
1
0
Myy
x
y
Mxy
0 y x 0 1
0
0
0 x y
Vx

Vy

LT = LT
LTs
f
M
still holds
0 0
+
0
=
0
p
0
Constitutive: the constitutive relation of Eq. 15.16 must be augmented to account for the
relationships between xz and yz (which are no longer negligible in this case) with Vx
and Vy respectively. This is obtained by equating the internal work with the external one

xz xz dzdA =
Vx xz dA
(15.18)
A
xz
,
and from Eq. 15.40, we obtain

Recalling that xz =

t/2

' z (2 2
2
xz
1 3 Vx
dzdA =
1 2
xz xz dzdA =
dzdA
2 t
t
A z
A z
A t/2

8 3 Vx 2 t
6 Vx2
6 Vx
dA =
dA =
dA(15.19-a)
Vx
=
15
2
t
5
t
5 t
A
A
A
Hence, comparing with Eq. 15.18
matrix becomes

1
Mxx
Myy

0
Mxy
=
Vx
Vy

M
where K =
19
Et3
12(1 2 )
we conclude that xz =
1
0
0
2

1 0
0 1

0
0
and the nal constitutive

xx
yy
2xy
xz
yz
(15.20)
and = 56 t
The constitutive matrix can be symbolically written as

Df 0
D=
0 Ds
Victor Saouma
6 Vx
5 t ,
(15.21)
Draft
158
15.2.1.2
19
Dierential Equation
The dierential equation for the Mindlin plate will not be derived in this course.
15.2.1.3
20
PLATES
The internal virtual work is expressed through two distinct terms

T
Wi =
MdA +
T VdA
A
A

T
(Lf ) Df Lf dA +
(Ls u)T Ds Ls udA
=
A
A
T T
Lf Df Lf dA +
uT LTs Ds Ls udA
=

A
A
Substituting
"
Z
W
xx
{z
yy

1
+ y
x x
2 y
+
y x
x 2 y
|
2
Z
w
A
xx
{z
yy 4
uT
{z
1
+ y
x y
2 x

1
+
y y
x 2 x
Ke
f
{z
5t
Internal, Bending

x x

y y
{z
1
0
Ke
s
{z
38
<
0 5
:
1
}|
w
+
A
xx
{z
uT
yy
8
9
< pz =
0
}: 0 ;
| {z }
pv
(15.22-c)
w
xx
{z
uT
}|
xx
yy
{z

dA
}
}
{z
u
Z
dA +
#
w =
xx
dA
;
yy
Internal, Shear
(15.22-b)
Kes
kef
21
(15.22-a)
yy
8
9
< ps =
0
dA
}: 0 ;
| {z }
{z
(15.23-a)
External
After substitution, and accounting for the external work, we obtain

6ks (1 )
T e
W = K
T Kef dA +
u
K
udA
s
t2
A
A

uT pV dA +
uT pd = 0
+
A
(15.24)
Kef
= LTf Df Lf
(15.25)
Kes
LTs Ds Ls
(15.26)
where ks = 5/6 is the shear correction factor

Z
W =
2
6
4
w xx
A
5t
5t
+ y
x 6 x
6 y
5t
y 6
5t
6 x
Victor Saouma
yy
5t
y 6
K(1)
K x + y
+ 5t
x
2
y
6
K(1)
K x + x
y
2
y
5t
6 x
K(1)
K y + y
x
2
x
K(1)
K y + x
+ 5t
y
2
x
6
38
9
< w =
7
5 : xx ; dA
yy
Draft
15.2 Plate Theories
159
w
A
15.2.2
xx
yy
8
9
< pz =
:
0
dA +
;
0
w
yy
xx
8
9
< ps =
:
0
dA = 0
;
0
(15.27-a)
Kirchho
This theory, is primarily applicable to thin plates in which shear deformations can be neglected.
22
15.2.2.1
In this formulation, shear deformations are neglected, and formulation is analogous to the
conventional Euler-Bernouilli beam theory.
23
Kinematic Relations Since shear deformations are neglected, xz = yz = 0 and thus the
last two relation of Eq. 15.10 reduce to
xx = w,x
and yy = w,y
(15.28)
and the rst three strains become

Oxx = z
or
2w
;
x2
xx
yy
2xy

Oyy = z
2w
;
y 2
x
2
2
y2
=

2
x y

L
Constitutive Relation The constitutive relation of
1
Mxx
Et3

Myy
=
12(1 2 )
Mxy
0

M
xy = 2z
2w
;
xy
(15.29)
+
w

(15.30)
Eq. 15.16 still applies
0
xx
1
0
yy
1
2xy
0
2

(15.31)
Equilibrium The equilibrium equation, as expressed by Eq. 15.7 still holds. If we were to
substitute the second and third relations into the rst one, then we would obtain the
following equilibrium relation in terms of the moments
2 Myy
2 Mxy
2 Mxx
+
+
2
+ pz = 0
x2
xy
y 2
(Note the similarity with the corresponding equations for beam exure
Victor Saouma
(15.32)
d2 M
dx2
Vx = 0)
Draft
1510
15.2.2.2
PLATES
Dierential Equation
24 If we combine the kinematic with the constitutive relation, 15.30 and 15.31 respectively we
obtain
2
2
+ y
2
Mxx
x2
3
*
+
Et
2
2
Myy
=
(15.33)
y2 + x
w
2
2
12(1 )
Mxy
(1 )
x y
Finally, we substitute the equilibrium equation, 15.32, into the previous one,
25
4w
4w
4w
+2 2 2 +
=
4
x
x y
y 4
pz
Et3
12(1 2 )
(15.34)
or
4 w =
pz
Et3
12(1 2 )
(15.35)
Note the similarity with the corresponding equation for beams

2w
4
pz
2
EI
=
p
or
=
z
2
2
4
x
x
x
EI
15.2.2.3
(15.36)
Stresses
26 Combining the stress-strain relation of Eq. 15.14, with Eq. 15.29, and 15.33, the stresses
could be expressed in terms of the moments
xx =
Mxx
t3
12
z;
yy =
Myy
t3
12
z;
xy =
Mxy
t3
12
again we note the analogy with the exural stress expression in beams =
dimensional equilibrium equation, Eq. 3.57:
xz
xx xy
+
+
x
y
z
yz
yx yy
+
+
x
y
z
zz
zx zy
+
+
x
y
z
integrating the rst equation yields

z
xx xy
+
dz =
xz =
x
y
t/2

12 t/2
zVx dz
= 3
t
z
Finally
xz
Victor Saouma
t/2
z
(15.37)
z
My
I .
Using the three
= 0
= 0
(15.38-a)
= 0
12z
t3
Mxx Mxy
+
x
y

dz
(15.39-a)
#
#
2 $
2 $
2z
2z
3Vx
3Vy
1
1
=
and yz =
2t
t
2t
t
(15.40)
Draft
15.2 Plate Theories
1511
Thus, the shear stress distribution across the plate thickness is parabolic (though they tend to
be very small compared to xy , and the peak shear stresses occur at the middle surface (z = 0)
where
3 Vx
3 Vy
xz |max =
and yz |max =
(15.41)
2 t
2 t
Finally, it can be shown that zz varies cubically.
15.2.2.4
27 Prior to the nite element discretization, we seek to obtain from the previously derived
relations a variational formulation of the problem.
28
The internal virtual work is given by

Wi =
T dA =
A
T MdA
(15.42)
where M = D is obtained from Eq. 15.31, and = Lu is obtained from Eq. 15.30.
29
Accounting for the external virtual work, we obtain

T
T
T
W = Wi + We =
u (L DL)udA +
u pv dA +
A
A

uT pd = 0
p

(15.43)
where pv contains the applied transverse loading pz , and p the edge loading ps .
30
Substituting for the LT DL (Eq. 15.30) terms we obtain
Ke = LT DL = x
2
= K

2
x2
y
2
2
2
+ 2
2
x
y

2 x
y
1
1

12(1 2 )
0 0

+ 2(1 )
x y
0
0
Et3

x y
2
+ 2
y
1
2
x22
y
2
2
x y
2
2
+ 2
2
y
x

(15.44-a)
and nally, noting that u = w, the virtual work becomes

W = K
w (Lwpz ) dA +
A
p wps d = 0
(15.45)
where K = Et3 /12(1 2 ).
15.2.3
Summary
31 Table 15.1 summarizes some of the major equations governing plate bending, and contrasts
them with the equivalent elasticity ones.
Victor Saouma
Victor Saouma
2
6
6
4
{z
xx
yy
zz
xy
yz
zx
0
0
0
D
{z
1
0
0
0
{z
0
0
0
0
0
8
3>
>
>
>
>
<
7
7
5>
>
>
>
>
}:
|
0
0
0
0
0
0
{z
0
0
0
0
0
xx
yy
zz
xy
xz
yz
"
38
xx
>
>
>
7
>
7>
< yy
7
zz
7
xy
7
>
>
5>
>
>
: yz
zx
} | {z
>
>
>
>
>
;
}
9
>
>
>
>
>
=
8
9
< bx =
by
+
>
: bz ; = 0
>
>
>
>
; | {z }
b
}
9
>
>
>
>
>
=
3
7
8
9
7
7
< ux =
7
7
u
y
7
: uz ;
7
7
5 | {z }
; = 12
= (1+)(12) ; = 1
2(1)
0
0
0
E(1)
2
6
6
6
= 6
6
>
>
>
4
>
>
;
} |
9
>
>
>
>
>
=
LT
{z
Elasticity
{z
2
6
6
4
xx
yy
2xy
{z
Mxx
Myy
Mxy
Vx
Vy
W = K
Z
A
0
0
{z
{z
Z
)
{z
1
2
0
0
Et3
;
12(1 2 )
1
0
{z
1
0
38
>
>
>
<
7
7
5>
>
>
:
}|
38
>
>
>
<
7
7
5>
>
>
:
}|
}|

{z
3
5
"
1
0
38
<
5
:
}|
0
1
0
1
9
=
;
}
u pV dA +
{z
0
0
0
0
0
0
u pd = 0
9
>
>
>
=
>
>
>
;
}
9
=
;
9
=
;
w
xx
yy
9
=
:
;
} | {z }
#8
<
xx
yy
2xy
xz
yz
9 8
= <
;=:
9 8
= <
;=:
38
>
>
7
>
<
7
7
7
>
5>
>
:
}|
{z
Ls
1
0
p wps d = 0
{z
0
0
pz
0
0
pz
xx
yy
2xy
9
>
8
>
>
= <
+
:
>
>
>
;
}
u Ks udA +
9
8
>
>
>
= <
+
> :
>
>
;
}
9
>
>
>
=

w
>
>
|
{z
>
; u }
}
= 5
t
6
1
2
0
0
{z
Mxx
Myy
Mxy
Vx
Vy
Mxx
Myy
Mxy
Vx
Vy
{z
xz
yz
{z
2
2
y

2 x
y
2
2
x
w (KLwpz ) dA +
K =
9
2
=
=K4
;
} |
x
{z
A
6ks (1
t2
W =
xx
yy
8
9 >
>
= >
<
;=>
>
>
} :
|
3

7
5
LT
s
2
K4
Mxx
Myy
Mxy
LT
f
{z
LT
9 2
>
>
>
= 6
6
=6
6
>
>
>
; 4
} |
8
<
:
|
{z
xx
yy
2xy
{z
Lf
8
<
:
|
Plate Theory (Kirchho/Mindlin)
Kf dA +
T
Virtual Work
8
>
>
>
<
>
>
>
:
|
Constitutive
2
6
6
4
LT =
9 2
= 6
;=4
} |
Equilibrium
8
<
:
|
Kinematic
1512
Table 15.1: Comparison of Governing Equations in Elasticity and Plate Bending
td = 0
= (Lu)T Dd uT bd uT
t
>
>
>
>
>
:
|
8
>
>
>
>
>
<
>
>
>
>
>
:
|
8
>
>
>
>
>
<
2
9
xx >
6
>
>
6
yy >
>
= 6
6
zz
=6
6
xy >
>
6
>
>
xz >
; 6
4
yz
{z }
|
"
Draft
PLATES
Draft
15.3 Finite Element Formulations
15.3
1513
Finite Element Formulations
32 Numerous elements have been proposed for plate bending. An ideal plate element should
have the following properties:
1. Formulation should be based on continuum mechanics and plate theory, the nodal d.o.f.
are the transverse displacement w, and sectional rotations xx and yy .
2. The element should be numerically correct, and convergent. The element stiness
matrix must contain the three rigid-body modes and no rank deciency (spurious zeroenergy modes).
3. The element should not lock in thin plate analyses.
4. The predictive capability of the element should be insensitive to element geometric distortions.
Since plate structure is a special case of shell structures, and plates under large displacement
exhibit shell characteristics (through membrane actions), similar formulations can be employed
for both plates and shells.
33
15.3.1
15.3.1.1
Rectangular Element
Formulation
The stiness matrix, including shear deformation eects, for a rectangular element, Fig. ??
will be developed.
34
Figure 15.6: Rectangular Plate Element

The selected degrees of freedom are w, xx and yy , thus
0
N1 N2 N3 N4
w
xx
0
N1 N2 N3 N4
=
yy
0
0

w
Nw 0
=
0 N
35
0
0
N1 N2 N3 N4
xx
yy
where
N1 = (1 )(1 );
Victor Saouma
N2 = (1 );
N3 = ;
N4 = (1 )
(15.47)
Draft
1514
PLATES
Note that the natural coordinate system selected has its origin at node 1, and 0 , 1. Thus
through these shape functions, we will have a rst order (C0 ) element, for which displacement
and rotations will be continuous along the boundaries (since xx and yy are also variables).
36 The stiness matrix, for a Mindlin plate, will be derived on the basis of the principle of
virtual work from Eq. 15.24
Wi = K
6ks (1
Kef dA +
2

Flexure
uT Kes udA

A
Shear
(15.48)
where Kef = LTf Df Lf , and Kes = LTs Ds Ls .

The rst integral will lead to the stiness matrix corresponding to bending, Kf , and the
second one to the stiness matrix corresponding to shear deformations Ks .
37
38
We recall that LN = B, and that x = a and y = b, hence

1
=
and
=
x
a
y
b
(15.49)
thus substituting into Eq. 15.17-a

2
6
4
|
1
a
0
1
b
{z
Lf
1
b
1
a
7
5
}

|
(1 )(1 )
0
(1 )
0
(1 )
0
{z
0
(1 )(1 )
0
(1 )
0
(1 )
1
a
= 0
1
b
1
a
0
b
a
0
1
b
0
0
b
1
b
1
a
1
a
1
b
a

}
(15.50)
(15.51)

2
t
We note that Lf = Lu (used in elasticity) for in plane deformation, and Df = 12
D, thus
the stiness matrix for bending can be formed directly from the stiness matrix for in-plane
deformation.
39
40
Upon substitution and integration we obtain

Kef
t2
Et
24(1 2 ) 12
2 w1 w2 w3 w4 x1
6 0
0
6
0
6
0
6
A
6
6
6
6
6
6
6
6
6
6
4
x2
x3
x4
y1
y2
0
C
A
A /2
B
A
B
A /2
C
A
2
3
2
3
A
3
2
3
2
B
A
y3
y4
7
7
7
7
2
3
7
7
3
2
7
2
3 7
3
2 7
7
A /2
C 7
C
A /2 7
7
A
B 7
5
A
(15.52)
Victor Saouma
Draft
1515
where = b/a, = a/b, 1 = 1 , 2 = 3(1 + ), 3 = 3(1 3), A = 8 + 41 ,

B = 4 21 , C = 8 + 21 , and A , B , C are obtained by interchanging
and in the expressions of A , B , C respectively.
41
Considering the second integral in Eq. 15.48, and substituting from Eq. 15.21
Kes = LTs Ds Ls

1 0
Ds =
0 1
5
t
=
6
we obtain
(15.53-a)
(15.53-b)
(15.53-c)
Kes udA
= K

(LTs u)T ILs udA
(15.54-a)
(LTs Nu)T Ls NudA
(15.54-b)
BTs Bs dAu
(15.54-c)
= K
= KuT
A
Recalling that Bs = Ls N, and substituting from Eq. 15.17-b for Ls we obtain

BTs Bs dA
Ks =

B11 B12
0
Bs =
0 B23
B21
(15.55-a)
(15.55-b)
Note that the rst row and second row in the second equations correspond to xz and yz
respectively.
42
Upon substitution,
1
B11 =
B12 =
B21 =
B12 =
!
1
!
(1)
a
(1)
a
(1 )(1 ) (1 ) 1
2
(1)
(1)
b
b
b
b
(15.56-a)
"
(1 )(1 ) (1 ) (1 )
(15.56-b)
"
(15.56-c)
(15.56-d)
43 In order to facilitate the integration of Ks , a one point Gauss integration = = 0.5 is

used. Upon substitution

1
1
1
1
1
1
1
1
2a
0
2a
2a 2a
4
4
4
4
s =
(15.57)
B
1
1
1
1
1
1
1
1
2b
0
2b
2b
2b
4
4
4
4
Finally Ks is obtained from
TB

B
s s dA
Ks =
(15.58)
Victor Saouma
Draft
1516
PLATES
which gives
Ks
1
24
2
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
4
w1
6
w2
6
6
w3
6
6
6
w4
6
6
6
6
x1
3b
3b
3b
3b
1.5ab
x2
3b
3b
3b
3b
1.5ab
1.5ab
x3
3b
3b
3b
3b
1.5ab
1.5ab
1.5ab
x4
3b
3b
3b
3b
1.5ab
1.5ab
1.5ab
1.5ab
y1
3a
3a
3a
3a
y2
3a
3a
3a
3a
y3
3a
3a
3a
3a
y4
3a
3a
3a
3a
7
7
7
7
7
7
7
7
7
7
7
7
7
7
1.5ab 1.5ab 1.5ab 1.5ab 7
7
1.5ab 1.5ab 1.5ab 7
7
1.5ab 1.5ab 5
1.5ab
(15.59-a)
where = ab , = ab , = + , and = .
44
The nal element stiness matrix is

!
"
Ke = K Kef + Kes
(15.60)
and in terms of the global stiness matrix we will have

[Kf + Ks ] u =
15.3.1.2
P
K
(15.61)
Shear Locking
Through inspection of the previous equation, and noting that = 6ks (1)
, we observe that
t2
for very thin plates is very large, hence as t 0, . Hence, unrealistic u = 0 can be
obtained independently of the load. This phenomena is called locking (of the displacements).
45
46
Locking occurs also in thin beam elements when shear deformation eects are accounted for.
To alleviate the locking problems is to make Ks singular so that Ks is nite. This can
be accomplished through reduced numerical integration for Ks , Table 15.2, which in turn will
reduce the rank of the matrix.
47
Bilinear
Quadratic
Serendipity
4 nodes, 12 d.o.f.
9 nodes, 27 d.o.f.
8 nodes, 24 d.o.f.
Integration Rule
Full
Reduced Selective
e
e
Kf Ks Kef Kes Kef Kes
2x2 2x2 1x1 1x1 2x2 1x1
3x3 3x3 2x2 2x2 3x3 2x2
3x3 3x3 2x2 2x2 3x3 2x2
Table 15.2: Integration Rules for Mindlin Plate Elements
Victor Saouma
Draft
1517
48 The same phenomena occurs for beams. This is illustrated by evaluating the deection of a
cantilever beam, due to both exural and shear deformation, under a point load

PL
P L3
3EI
P L3
+
=
1+
(15.62)
w=
3EI
ks A
3EI
ks AL2
or
w=
P L3
3EI

1+

;
ks AL2
= ks
EI
E
2
2
L
L
=O
r
h2
(15.63)
Hence, as the slenderness ratio increases, shear deformation decreases. Furthermore, Euler
theory is recovered if the shear stiness , however shear locking can occur for small .
49
Shear locking eects are aggravated by large element distortions.
15.3.2
15.3.2.1
Nonconforming Kirchho Triangular Element

Formulation
Figure 15.7: Triangular Plate Element in Natural Coordinate System

50 Since rectangular elements can hardly accommodate general boundary conditions, and since
quadrilateral elements can more readily be obtained through an assemblage of triangular ele-
Victor Saouma
Draft
1518
PLATES
ments rather than through the distortions of a rectangular element, we consider next the three
noded, nine d.o.f. Kirchho element.
51 The element is shown in Fig. 15.7, has the following 3 dof at each node: displacement w, and
w
slopes xx = w
x and yy = y . Because the complete cubic polynomial expansion contains
10 terms (1, , , 2 , , 2 , 3 , 2 , 2 , 3 ), yet we have only nine dofs, thus we opt instead to
a natural coordinate system and we express the displacement as
w = w1 + w2
(15.64)
where w1 corresponds to the rigid body part, and w2 the displacements incurred when the
triangle is simply supported at the nodes.
52
w1 has three rigid body motions, Fig. 15.7
1. Translation: in the z direction with w1 = cst.

2. Rotation about side 1: of the triangle. This results in a rigid body translation in the
z direction of line ab which is parallel to side 1 is w1 = h1 . If we consider a triangle
with vertex b, than this point would have L1 = A1 /A = h1 /h, thus w1 = h1 = hL1 or
w1 = (cst)L1 .
3. Rotation about side 2: where w1 = (cst)L2 .
Thus for rigid body motions we can write:
1 L1 + w
2 L2 + w
3 L3
w1 = w
(15.65)
53 Since the element has 9 dof, a cubic polynomial will be used for shape functions; i.e. the
shape function can be a linear combination of the cubic terms
L21 L2 , L22 L3 , L23 L1 , L22 L1 , L23 L2 , L21 L3 , L1 L2 L3

Note that if we were to substitute into

x2 y3 x3 y2 y23 x32 1
L1
L
x
= x3 y1 x1 y3 y31 x13

2
L3
x1 y2 x2 y1 y12 x21
y
(15.66)
(15.67)
where xij = xi xj , then we will have a cubic polynomial in terms of x and y.

54
55
Since w1 took already care of the rigid body movements, the cubic terms can be used for w2 .
Zienkiewicz found it convenient to form the terms of w2 as follows

1
1
2
2
4 L3 L1 + L1 L2 L3 + w
5 L3 L2 + L1 L2 L3
w2 = w
2
2

1
1
2
2
7 L1 L2 + L1 L2 L3
+w
6 L1 L3 + L1 L2 L3 + w
2
2

1
1
2
2
9 L2 L1 + L1 L2 L3
w
8 L2 L3 + L1 L2 L3 + w
2
2
Victor Saouma
(15.68-a)
Draft

56
1519
Hence

w = w1 + w2 = Nu w
(15.69)
are dened in Eq. 15.65 and 15.68-a.

where Nu and w
57 The unknown quantities w
i can in turn be expressed in terms of the nodal unknown disw
placements by substituting w, xx = w
x and yy = y in Eq. 15.69.
58
The nodal displacement are hence expressed as

uw

u=N
(15.70)
1 u = Gu and
=N
where u = w1 x1 y1 w3 x3 y3 thus w
u
w = Nu Gu = Nu
(15.71)
with w = N1 u1 + N2 u2 + N3 u3 where u1 = wi xxi yyi T

59
Upon substitution, it can be shown that

T
g 21 L2 + L21 L3 L1 L2 L1 L23
L1 + L
b3 (L21 L
g 2 + 12 L1 L2 L3 ) + b2 (L21 L
g 3 + 12 L1 L2 L3 )
N1 =
g 2 + 12 L1 L2 L3 ) c2 (L21 L
g 3 + 12 L1 L2 L3 )
c3 (L21 L
(15.72)
where b2 = x1 x3 , b3 = x2 x1 , c2 = y3 y1 , and c3 = y1 y2 . The other two shape functions

can be obtained by cyclic permutation of sux 1-2-3.
60
Finally, the stiness matrix is derived from Eq. 15.45

e
w (K wpz ) dA +
p wps d = 0
W =
A
(15.73)
where
Ke = LT DL
(15.74)
Thus

Ke =
NT LT DLNdA (15.75)

NT pz dA
(15.76)
These matrices are evaluated using

Lk1 Ll2 Lm
3 dA = 2A
A
15.3.2.2
k!l!m!
(2 + k + l + m)!
(15.77)
Nonconformity
Considering Fig. 15.8

it can be shown that the displacement w along sides 2-3 (where L1 = 0) are cubic and is the
same for element 1 as for element 2 (since they share the same nodal d.o.f.). The four terms of
the cubic polynomial can be uniquely determined from w2 , w3 , xx2 and xx3 at nodes 2 and 3.
Hence, both w and xx = w
x are continuous along side 2-3 for both elements.
61
Victor Saouma
Draft
1520
PLATES
1
1
2
x
3
2
Figure 15.8: Edges of Adjacent Triangular Elements

However, for yy = w
y things are quite dierent. If we set L1 = 0 and take the derivative
of w in Eq. 15.69, then we have
62
yy =
w
= 0 + 1 x + 2 x2
y
(15.78)
but since we only have two d.o.f. (yy at nodes 2 and 3), then the parameters can not be
uniquely expressed in terms of these two d.o.f. Hence, we will have discontinuity of normal
slope, or kinks.
63
Elements with element boundary discontinuity are referred to as nonconforming elements.
64 This nonconforming element, nevertheless is a viable one as the 3 rigid body modes, and
constant strains are present. In addition, it passes the patch test.
Despite its nonconformity, this element is quite appealing because the deformation is expressed only in terms of the middle surface deection w.
65
15.3.3
Discrete Kirchho Triangle
66 From the previous formulation, it is apparent that it is dicult to formulate a compatible

triangular element with 9 d.o.f. using a single polynomial approximation for w.
A viable approach consists in starting with the thick plate, or Reissner-Mindlin plate theory,
by separating the d.o.f. for translation and rotations and making them independent one from
the other. These in turn are made continuous at the inter-element boundaries, thus ensuring
C0 continuity. Since the plate is very thin, the term corresponding to shear deformation in the
virtual work will be neglected. The shape functions will be designed to ensure compatibility,
thus the element will be conforming. Finally, the Kirchho plate theory assumptions will be
only introduced at discrete points along the element boundaries.
67
68
The starting point will be Eq. 15.22-c, Fig. 15.9

T LTf Df Lf dA +
Wi =
A
uT LTs Ds Ls udA
(15.79)
where the rst term corresponds to exure, and the second one to the eect of shear which will
be neglected.
Victor Saouma
Draft
1521
3
6
5
y
1
4
2
x
Figure 15.9: Discrete Kirchho Triangular Element

69 Since for bending, only the rotations appear in the formulation, the rotations can be approximated by
6
6
Ni xi ;
yy =
Ni yi
(15.80)
xx =
i=1
70
i=1
The shape functions Ni for the natural coordinates with quadratic expansions are
N1 = L1 (2L1 1) = (1 L2 L3 )(1 2L2 2L3 )
N2 = L2 (2L2 1)
N3 = L3 (2L3 1)
N4 = 4L2 L3
N5 = 4L1 L3 = 4(1 L2 L3 )L3
N6 = 4L1 L2 = 4(1 L2 L3 )L2
(15.81-a)
For this element, so far we have a total of 12 dof, and we seek to formulate it in terms of only
i
i
9 dof, 3 at each corner node corresponding to wi , xx and yy thus we need to impose certain
constraints for this reduction.
71
72
Hence, Kirchho assumptions will be selectively applied:
Victor Saouma
Draft
1522
PLATES
1. Zero transverse shear strains at the corners: or from Eq. 15.17-b
w

1
0
w,x
0
xx
xz
x
=
=
=
yz
yy
w,y
0
xx
y 0 1
yy
(15.82)
2. Zero transverse shear strain at the midside nodes: or

k
tk = w,s
(15.83)
where k is a superscript denoting the midside nodes only, tk is a rotation about the normal
to the boundary at node k, and wsk is the derivative of w with respect to the s direction
along the boundary.
3. Linear variation of the tangential slope: , i.e the rotation about the s direction at the
middle nodes is the average of the rotations at the end nodes
1
nk = (ni + nj )
2
(15.84)
where the corner nodes i j are 2-3, 3-1, and 1-2. k corresponds to the midside nodes.
73
We let the displacement of w along the boundary of length lij to be cubic

w = a0 + a1 s + a2 s2 + a3 s3 ;
w = a1 + 2a2 s + 3a3 s
(15.85)
and thus w = w,s varies quadratically and the three coecients can be obtained by matching
w,s = n at the three points along the edge. Applying the boundary conditions at s = 0 and
s = lij in terms of both w and w,j , we have 4 equations to determine the four coecients,
wi = 0
wi
wj = a0 +
wj
2
a1 lij + a2 lij
= a1 + a2 lij +
ats = 0 node i
(15.86-a)
= a1
ats = 0
(15.86-b)
3
a3 lij
2
a3 lij
ats = lij node j
(15.86-c)
ats = lij
(15.86-d)
yielding
k
=
w,s
3 i 1 i
3 j 1 j
w w,s +
w w,s
2lij
4
2lij
4
(15.87)
74 Since the rst term of Eq. 15.22-c, does not involve w, there is no need to dene an interpolation function for w inside the element.
75
Relations between xx , yy and nodal variables n and t are given by

cos ij sin ij
n
cos ij sin ij
w,n
xx
=
=
yy
sin ij cos ij
t
sin ij cos ij
w,s
and
Victor Saouma
w,n
w,s

=
cos ij sin ij
sin ij
cos ij

xx
yy
(15.88)

(15.89)
Draft
1523
76 Substituting the last three equation into Eq. 15.80 at each node to obtain xi and yi to nd
the shape functions results in
xx = Nx vi ;
yy = Ny vi
or
(15.90)

i
xx
Nx1 Nx2 Nx3 Nx4 Nx5 Nx6 Nx7 Nx8 Nx9
=
i
Ny1 Ny2 Ny3 Ny4 Ny5 Ny6 Ny7 Ny8 Ny9
yy

where
Nx1
Nx2
Nx3
Nx4
Nx5
Nx6
Nx1
Nx2
Nx3
x
and ak = l2ij , bk =
ij
=
=
=
=
=
=
=
=
=
1.5(a6 N6 a5 N5 ),
N1 b5 N5 b6 N6 ,
c5 N5 c6 N6 ,
1.5(a6 N4 a5 N6 ),
N2 b5 N6 b6 N4 ,
c6 N6 c4 N4 ,
1.5(a6 N5 a5 N4 ),
N3 b5 N4 b5 N5 ,
c4 N4 c5 N5 ,
1 2
x 1 y2
4 ij 2 ij
2
lij
, ck =
Ny1
Ny2
Ny3
Ny4
Ny5
Ny6
Ny1
Ny2
Ny3
3 xij yij
,
2
4 lij
=
=
=
=
=
=
=
=
=
w1
1
x
1
y
w2
2
x
2
y
w3
3
x
3
y

v
(15.91)
1.5(d6 N6 d5 N5 )
Nx3
N1 e5 N5 e6 N6
1.5(d4 N4 d6 N6 )
Ny3
N2 e5 N6 e4 N4
1.5(d5 N5 d4 N4 )
Nx3
N3 e4 N4 e5 N5
(15.92)
1 2
y 1 x2
4 ij 2 ij
2
lij
and k = 4, 5, 6
dk = l2ij , ek =
ij
corresponds to sides ij=23, 31, and 12 respectively.

With the shape functions thus dened, we can nally determine the stiness matrix from
the principle of virtual work (Eq. 15.22-c). Since shear deformations are neglected, we use only
the rst part.

T LTf Df Lf dA
(15.93-a)
Wi =

A
77
= vi
= vi
Kef
NT LTf Df Lf NdAvi
BT Df BdA vi

A
(15.93-b)
(15.93-c)
Ke
(15.93-d)
Thus the stiness matrix can be evaluated
Victor Saouma
Draft
1524
15.4
78
PLATES
Summary
Major conclusions
1. Two theories: thin plates (Kirchho/Euler-Bernouilli), and thick plates (Mindlin-Reissner/Timoshenko),
without and with shear deformations.
2. Strong analogy between plates and beams.
3. If membrane actions are accounted for, lead to formulation of shell elements.
4. M , .
5. If thick plate theory is applied to thin plates, will have shear locking problems.
6. Shear locking is alleviated by reduced integration of the shear components.
7. For thin plate theory, will have non-conforming elements if only 3 d.o.f. are adopted at
each node.
8. The DKT element alleviated this problem.
9. Reduced integration will alleviate problems arising from mesh distortions.
10. Shear locking is primarily an issue for 3 node and 4 node elements. Higher order expansions perform much better, whereby Lagrangian 6 and 16 node elements perform
signicantly better than their Serendipity counterparts of 8 node and 12 node elements.
11. The usual 3 DOF/node Kirchho elements do not maintain normal slope continuity.
Victor Saouma
Draft
Chapter 16
MATERIAL NONLINEARITIES
16.1
Introduction
16.1.1
Linearization
We dene a constitutive operator as
= ()
(16.1)
denotes the constitutive operator (analogous to the L).

where
2
, the corresponding stress will be

= (
).
Given a strain state
can be expanded into a Taylor series with respect to
The constitutive operator

+ ) = ()
+
+
(
=
(16.2)
Neglecting quadratic and higher order terms leads to a linearized constitutive law
+ D()
()
(16.3)
which approximates Eq. 16.1 for strains in the neighborhood of , and

D
(16.4)
is the tangent stiness matrix which is a function of the current strain.

4
We rewrite Eq. 7.8, 7.12 and 7.16 in terms of the newly dened constitutive operator

T
T
T
T
B (Bu)d
=
B D 0 d
B 0 d +
N bd +
NT td
(16.5)
e
e
t
e
t

f
int
f 0e

f
ext
fe
Draft
162
or
f
int
(u) = f
ext
(16.6)
5 We now develop a linearized expression for the internal forces. Given u as nodal displacements
= (Bu).
Then, the Taylor expansion
yielding strain eld = Bu, and the stress eld = ()
of the internal forces around u yields

int
f
int
int
f (u + u) = f (u) +
u +
(16.7)

u
u=u
We again neglect the quadratic and higher order terms, leading to

f
int
(u + u) f
int
(u) + KT (u)u
(16.8)
where
int
f
KT
u
(16.9)
is the tangent stiness matrix of the structure

6
Dierentiating Eq. 16.5

int
f
=
KT =
u
u
B d
=
T
d =
B
u
d =
B
u
BT DBd
(16.10)
which is the well known formula for the stiness matrix, however Del is now replaced by the
tangent moduli D
16.1.2
Solution Strategies
7 Before we discuss solution strategies, it may be helpful to point out the parallelism which
exists between (numerical) solution strategies, and (experimental) testing methods. Modern
testing equipment can be programmed to apply a pre-determined rate of load (as measured by
a load cell), of displacement (as measured by an internal displacement transducer), or of strain
(or relative displacement such as crack mouth opening displacement) measured by a strain/clip
gage or other instruments, Fig. 16.1
Load Control: the cross-head applies an increasing load irrespective of the specimen deformation or response. For all materials, when the tensile strength is reached, there is a sudden
and abrupt brittle failure. The strain energy accumulated in the specimen is suddenly
released once the ultimate load of the specimen is reached, thus the sudden failure can be
explosive.
Displacement/Stroke Control: the cross-head applies an increasing displacement to the
specimen. For softening material there will be a post-peak response with a gradual decrease in stress accompanying an increase in displacement. In this case, there is a gradual
release of strain energy which is then transferred to surface energy during crack formation.
Victor Saouma
Draft
16.1 Introduction
163
P
11
00
11
00
11
00
11
00
,
11
00
111
000
11
00
11
00
111
000
P, , CMOD
11
00
t
11
00
11
00
11
00
111
000
CMOD
Actual
Programmed
Figure 16.1: Test Controls

Strain Control: is analogous to displacement control, except that the feedback is provided by
(strategically positioned) strain gage or a clip gage or an arbitrary specimen deformation
(not necessarily corresponding to the loading direction). To accomplish this test a clip
gage or a strain gage has to provide the feedback signal to the testing equipment in order
to accordingly adjust the stroke.
Similarly, the objective of a nonlinear nite element analysis is to trace the (nonlinear)
response of a structure subjected a given load history. This is best done in an incrementaliterative procedure where the load (or the displacement) is applied through several increments,
and within each increment we seek to satisfy equilibrium through an iterative procedure (caused
by the nonlinearity of the problem).
The incremental analysis can be performed under

1. Load control; Load is incrementally applied on the structure.
2. Direct displacement control; An imposed displacement is applied.
3. Indirect displacement control (such as relative displacements between two degrees of freedom)
4. Arc-Length control
10
Alternatively, iterative techniques include

1. Newton-Raphson
2. Modied Newton-Raphson
3. Initial Stiness
4. Secant Newton
11
Finally, an essential ingredient of an incremental-iterative solution strategy are

1. Convergence Criteria
2. Convergence Accelerators (such as line-search or step-size adjustments).
Victor Saouma
Draft
164
16.2
Load Control
16.2.1
Newton-Raphson
For the sake of discussion, we will assume in the following sections that the incremental
ext
analysis is under load control, with increments of loads f .
12
13 At the end of each load increment, internal forces must be in equilibrium with the external
ones. Hence, we dene the vector of residual forces R as
Rn+1 R(un+1 ) = f
where f
int
int
(un+1 ) f
ext
=0
(16.11)
is the vector of internal forces, also commonly known as reaction vector.

int
For equilibrium to be satised, the vector of reactions internal forces f must be equal to
ext
the one of external ones f . This is automatically satised in linear elastic analysis, but not
necessarily so in nonlinear analyses.
14
15
We start the analysis from an equilibrium conguration, at the end of increment n such that
u = un ,
Rn = 0
(16.12)
ext
and apply an increment of load f n such that

ext
ext
ext
f n+1 = f n + f n
(16.13)
and we seek to determine the corresponding change in displacement

un+1 = un + un
(16.14)
ext
We will keep f n reasonably small to capture the full nonlinear response.

16.2.1.1
Newton-Raphson/Tangent Stiness Method
This is the most rapidly convergent process (albeit computationally expensive) of non-linear
problems.
16
At the beginning of each step n + 1, we start from the displacement un that were computed
int
ext
in the previous step through equilibrium Rn 0 or f n f n . The external forces are
ext
ext
ext
ext
now increased from f n to f n+1 = f n+1 + f , and we seek to determine the corresponding
17
int
ext
displacements un+1 through equilibrium Rn+1 0 or f n+1 f n+1 .

18 Within the current step (identied through the subscript n), we will be iterating (through
superscript k) in order to achieve equilibrium.
As initial guess for u0n+1 we take it to be un and based on the linearization around this initial
state we have
ext
f int (u0n+1 ) + KT (u0n+1 )u1n+1 = f n+1
(16.15)
19
where u1n+1 is the rst approximation for the unknown displacement increment un+1 =
un+1 un .
Victor Saouma
Draft
16.2 Load Control
165
f ext
2
Rn+1
f ext
n+1
3
Rn+1
n+1
1
Rn+1
fnext
n+1
f int,
2
1
T
n+1
f int,
1
f ext
n
un2
un1
un1
un
un2
1
un+1
2
un+1
3
un+1
Figure 16.2: Newton-Raphson Method

20
Alternatively, we begin from a linearization of Eq. 16.11, Fig. 16.2

i

R
i+1
i
R(un+1 ) R(un+1 ) +
ui = 0
u n+1 n
(16.16)
where i is a counter starting from u1n+1 = un .

21
Observing that
int
f
R
=
= KT
u
u
assuming that f
ext
(16.17)
is constant, and KT is the tangent stiness matrix. Thus, Eq. 16.16 yields
i
KiT uin = Rn+1
(16.18)
uin = (KiT )1 Rn+1
(16.19)
or
i
22
Thus, a series of successive approximations yields

i
i
i
ui+1
n+1 = un + un = un+1 + un
with
uin =
ukn
(16.20)
(16.21)
ki
very rapidly.
23
It should be noted that each iteration involves three computationally expensive steps:
Victor Saouma
Draft
166
1. Evaluation of internal forces f
int
(or reactions)
2. Evaluation of the global tangent stiness matrix KT

3. Solution of a system of linear equations
16.2.1.2
Modied Newton-Raphson
This method is essentially the same as the Newton-Raphson however in Eq. 16.23 (KiT ) is
replaced by KT which is the tangent stiness matrix of the rst iteration of either 1) the rst
increment KT = K1T,0 , Fig. 16.4, or 2) current increment, Fig. 16.3 KT = K1T,n Fig. 16.3
24
f ext
2
Rn+1
f ext
n+1
3
4
Rn+1
Rn+1
n+1
1
T
n+1
f int,
2
1
Rn+1
fnext
n+1
f int,
3
1
K
T
f ext
n
un2
n+1
f int,
1
un1
un2
un3
un1
un3
un
1
un+1
2
un+1
3
un+1
Figure 16.3: Modied Newton-Raphson Method, Initial Tangent in Increment
uin = (KT )1 Rn+1

i
(16.22)
In general the cpu time required for the extra iterations required by this method is less than
the one saved by the assembly and decomposition of the stiness matrix for each iteration.
25
26 It should be mentioned that the tangent stiness matrix does not necessarily have to be the
true tangent stiness matrix; an approximation of the true tangent stiness matrix or even the
initial stiness matrix will generally produce satisfactory results, albeit at the cost of additional
iterations.
16.2.1.3
Secant Newton
27 This method is a compromise between the rst two. First we seek two displacements by two
cycles of modied Newton-Raphson, then a secant to the curve is established between those
Victor Saouma
Draft
16.2 Load Control
167
f ext
2
Rn+1
f ext
n+1
3
4
Rn+1
Rn+1
n+1
0
T
n+1
f int,
3
0
T
1
Rn+1
fnext
n+1
f int,
2
n+1
f int,
1
f ext
n
un2
un2
un1
un3
un1
un3
un
1
un+1
2
un+1
3
un+1
Figure 16.4: Modied Newton-Raphson Method, Initial Problem Tangent

two points, and a step taken along it, Fig. 16.5.
uin = (KT )1 Rn+1
i
(16.23)
28 Subsequently, each step will be taken along a secant connecting the previous two points.
Hence, starting with
1
(16.24)
u1n = K1
T Rn+1
the secant slope can be determined

(K2S )1 =
and then
u1n
1
(16.25)
(Rn+1 Rn+1 )
u2n = (K2S )1 Rn+1

2
(16.26)
uin = (KiS )1 Rn+1

i
29
This process can be generalized to (KiS )1 =
16.2.2
uin
i1
(Rn+1 Rn+1 )
(16.27)
(16.28)
Acceleration of Convergence, Line Search Method

30 The line search is an iterative technique for automatically under- or over-relaxing the displacement corrections uj so as to accelerate the convergence of nonlinear solution algorithms.
Victor Saouma
Draft
168
f ext
2
Rn+1
f ext
n+1
3
Rn+1
4
Rn+1
n+1
n+1
f int,
3
K0
1
Rn+1
fnext
n+1
f int,
2
T
n+1
f int,
1
f ext
n
un2
un2
un1
un3
un1
un3
un
3
un+1
2
un+1
1
un+1
Figure 16.5: Incremental Secant, Quasi-Newton Method

31 The amount of under- or over-relaxation is determined by enforcing an orthogonality condij+1
tion between the displacement corrections uj and the residual loads R , which amounts to
forcing the iterative change in energy to be zero.
The displacement corrections are multiplied by a scalar value sk dening the amount of
under- or over-relaxation such that the total displacements uj+1,k are dened as
32
uj+1,k = uj + sk uj
(16.29)
For k = 0 and k = 1, the values of sk are 0.0 and 1.0, respectively. Therefore, uj+1,0 = uj
and uj+1,1 = uj+1 . The orthogonality condition is quantied by a scalar value gk representing
the iterative change in energy, which is dened as
33
j+1,k
gk = uj R
where
j+1,k
=f
ext
int
(16.30)
(uj+1,k )
(16.31)
are the residual loads at the end of solution iteration j and line search iteration k.
gk can be expressed as a function of sk (see Figure 16.6) and the object of the line search is
to nd sk such that gk is zero. An estimate of sk+1 such that gk+1 is zero can be computed
using a simple extrapolation procedure based on similar triangles
34
sk
sk+1
=
g0
g0 gk
On rearranging terms, sk+1 is dened as
s
Victor Saouma
k+1

= s
g0
g0 gk
(16.32)

(16.33)
Draft
16.2 Load Control
169
As a preventative measure, sk+1 is assigned a value of 5.0 for all sk+1 > 5.0 so that unreint
strained over-relaxation is inhibited. Once sk+1 is estimated, uj+1,k+1 , f j+1,k+1 , and Rj+1,k+1
are computed for the next line search iteration, Fig. 16.6.
35
g
g0
g1
s1
s2
Figure 16.6: Schematic of Line Search, (Reich 1993)

36
The line search terminates after three iterations or when

| g0 |
0.8
| gk |
(16.34)
and g0 gk 0.001 | g0 |. Smaller tolerances may be used to determine if the line search has
converged, Zienkiewicz and Taylor (1989) prefer to use 0.6, but Criseld, M.A. (1979) concluded
that there was little advantange to be gained by doing such.
37
The owchart illustrating the Line Search algorithm is shown in Fig. 16.7.
Initialize s1
Compute R
DO 30 k=1,3
Do 10 j=0, Niter
Compute gk and sk+1
j
j+1 0
Compute u, u,
j+1
j+1
Compute and
j+1
j+1
Compute fint and R
No
Compute u
j+1,k+1
j+1,k+1
Compute
Converged
No
Yes
Converged
j+1,k+1
and
Yes
j+1,k+1
Compute fint
and R
j+1,k+1
10
30
Figure 16.7: Flowchart for Line Search Algorithm, (Reich 1993)
Victor Saouma
Draft
1610
16.2.3
Convergence Criteria
38 In all preceding methods, iterations are performed until one or all of a variety of convergence
criteria are satised. Relative convergence criteria are optionally enforced on the displacements,
loads, and/or incremental energy to dene the termination conditions.
The relative displacement criteria is dened in terms of the displacement corrections uj and
the updated incremental displacements uj+1 as
39
Ou =
uj 2
uj+1 2
(16.35)
where . . . 2 is the Euclidean norm.

40
The Euclidean norm, which is also known as the L2 norm, of a vector v is dened as
#N
v 2 =
$1/2
vi2
(16.36)
i=1
where N is the size of v.

j+1
The relative load criteria is dened in terms of the updated residual loads R
int
reactions f j+1 as either
41
Or =
or
Or =
j+1
2
int
f j+1
2
R
j+1
int
f j+1
R

where . . . is the innity norm. The innity norm of a vector v is dened as

3N
4
|vi |
v = max
and the
(16.37)
(16.38)
(16.39)
i=1
where N is the size of v.

The relative incremental energy criteria is dened in terms of displacement corrections uj ,
j+1
the updated residual loads R , the updated incremental displacements uj+1 , and the upint
dated reactions f j+1 as
42
j+1
OW =
uj R
int
uj+1 f j+1
(16.40)
where the numerator is the change in the incremental energy for iteration j and denominator
is the incremental energy.
16.3
Direct Displacement Control
Adapted from (Jirasek and Ba

zant 2001)
Victor Saouma
Draft
16.3 Direct Displacement Control
1611
43 Independently of the choice of iterative algorithm, any solution strategy using load control
fails if the prescribed external loads cannot be maintained in equilibrium by the internal forces.
This would typically occur if the load is monotonically increased until the load-carrying capacity
of the structure is exhausted, Fig. 16.8
f ext
Figure 16.8: Divergence of Load-Controled Algorithms

In most engineering analyses, it is simply required to determine the maximum load carrying
capacity, and the corresponding displacements. As such, divergence of the iterative process is
often taken as an indicator of structural failure, and the last converged step provides information
on the state prior to collapse.
44
45 However, nite element simulations of complex engineering problems can diverge for a number of other reasons, many of which are purely numerical and have nothing to do with the real
structural failure.
46 If the load-displacement diagram is to be followed beyond the peak, i.e post-peak response
is required, then alternative solution strategy to the load-control one must be devised.
Post-peak response may be of interest not only in problems in structures with imposed
displacements (such as initial settlements), but also to assess the ductility of the structure
(specially when cracks are present).
47
To outline the displacement controlled algorithm, we divide the displacements into two
groups: one with unknown displacements at nodes that are left free, and the second with
prescribed displacements at nodes that are controlled. Accordingly, we partition the displacement vector into {uf , up }T and the internal and external force vectors into {f int,f , f int,p }T and
{f ext,f , f ext,p }T , respectively. External forces f ext,f (corresponding to the unknown displacements uf ) are prescribed, and for simplicity we will assume that they are equal to zero. All
external forces acting on the structure are represented by reactions f ext,p at the supports with
prescribed displacements up .
48
49
Hence, the equilibrium equations are partitioned as

f int,f (uf , up ) = 0
(16.41-a)
f int,p (uf , up ) = f ext,p
(16.41-b)
50 For given up , the unknown displacements uf can be computed by solving Eq. 16.41-a. After
that, the reactions f ext,p are obtained by simple evaluation of the left-hand side in (16.41-b).
Victor Saouma
Draft
1612
51
(n1)
In a typical incremental step number n, we start from the converged displacements uf

(n1)
and up
from the previous step, and we replace Eq. 16.41-a by the linearized equations
(n1)
ff
where K11
f int,f
uf
(n1)
+ K11
and K12
f int,f
up
(n,1)
uf
(n1)
+ K12
u(n,1)
=0
p
(16.42)
are blocks of the global tangent stiness matrix
f int,f

K
K
up
11
12
=
K21 K22
f int,p
up
f int,f
f int uf
=
K
f int,p
u
uf
(16.43)
The increment of the prescribed displacements up is known in advance, and so we set

(n,1)
(n)
(n)
(n1)
= up = up up
and rewrite (16.42) as
up
52
(n1)
K11
(n,1)
Having solved for uf

(n,1)
up
(n1)
up
(n,1)
up
(n,1)
uf
(n1)
(n1)
= f int,f K12
u(n)
p
(n,1)
, we construct the rst approximation uf
(n1)
= uf
(n,1)
+ uf
and
(n)
up .
(n,1)
(n,1)
Equations (16.41-a) are then linearized around (uf , up ), corrections of displacements

uf are computed, and the procedure is repeated until the convergence criteria are satised.
53
54
The iterative process can be described by recursive formulas

(n,i1)
K11
(n,i1)
(n,i)
= f int
(n,i)
= uf
uf
uf
(n,i1)
(n,i1)
K12
(n,i)
up
(n,i)
+ uf
i = 1, 2, 3, . . .
where
(n,0)
(n1)
(16.44-a)
u(n,0)
= u(n1)
p
p
(16.44-b)
uf
= uf
(n1)
= u(n)
u(n,1)
p
p up
(16.44-c)
u(n,i)
p
(16.44-d)
= 0
for i = 2, 3, . . .
55 Note that, starting from the second iteration, the correction up is zero, and so the term with
K12 on the right-hand side of (16.3) vanishes. This term is present only in the rst iteration. It
(n,0)
(n)
(n,0)
(n1)
= up instead of up
= up
, and
might seem that one could start immediately from up
then the correction up would be zero already in the rst iteration and the matrix K12 would
never have to be evaluated. However, this is in general not a good idea because such an initial
approximation would be too far from the equilibrium path and the process might diverge.
Victor Saouma
Draft
16.4 Indirect Displacement Control
16.4
1613
Indirect Displacement Control
56 Direct displacement control can be applied only on structures loaded only at one point, or
when the load is transmitted by a sti platen so that all points on the loaded surface exhibit
the same displacements.
57 However, this is not always the case. As an example, consider a dam loaded by hydrostatic
pressure due to reservoir overow; see Fig. 16.9. Here, the load is applied along a large portion of
Figure 16.9: Hydrostatically Loaded Gravity Dam

the boundary, and the shape of the corresponding displacement prole is not known in advance.
Another case in which direct displacement control fails is very brittle failure characterized by
a load-displacement diagram with a snapback, Fig. 16.10.
l
l
u
Load Control
u
Displacement Control
u
Arc-Length Control
Figure 16.10: Load-Displacement Diagrams with Snapback

58 Advanced incrementation control techniques abandon the assumption that the values of
external loads and/or displacements at supports after each incremental step are prescribed in
advance. Instead, the loading program is parameterized by a scalar load multiplier.
16.4.1
Partitioning of the Displacement Corrections
Victor Saouma
Draft
1614
Restricting the applied loading to be proportional, a scalar load parameter can be used to
ext
scale an arbitrary set of applied loads f .
59
ext
The applied loads at the start of increment i are dened as the scalar-vector product i f ,
where i is the load parameter at the start of increment i. i is zero at the start of the rst
increment.
60
The applied incremental loads for increment i are dened as the scalar-vector product
ext
i f , where i is the incremental load parameter for increment i. The updated load
parameter i+1 at the end of increment i is
61
i+1 = i + i
(16.45)
62 The incremental displacements due to the applied incremental loads are obtained using the
standard modied-Newton algorithm, as described in Zienkiewicz & Taylor (1991). The incremental displacements uj+1 at the end of iteration j for a generic increment are dened
as
(16.46)
Duj+1 = uj + uj
where uj are the incremental displacements at the start of iteration j and uj are the incremental displacement corrections for iteration j.
The incremental load parameter j+1 at the end of iteration j is dened in an analogous
manner as
(16.47)
j+1 = j + j
63
where j is the incremental load parameter at the start of iteration j and j is the incremental
load parameter correction for iteration j. At the start of the rst iteration uj and j are
identically zero.
64
Incremental displacement corrections are determined by solving

K uj = (f
ext
+ j f
ext
+ j f
ext
f int )
(16.48)
where K is the global stiness matrix and

j
f int
elem
e=1

BT D ( + j )
(16.49)
are the reactions for the state of stress at the start of iteration j.
j
65
Dening the residual forces R at the start of iteration j as

j
R = f
ext
+ j f
ext
f int
(16.50)
Equation 16.48 can be written more simply as

uj = K1 ( j f
ext
+ R )
(16.51)
ext
The matrix-vector product K1 f

is invariant for the increment and, therefore, can be
treated as a vector constant uT , which Criseld (1981) referred to as the tangent displacements
66
uT = K1 f
Victor Saouma
ext
(16.52)
Draft
1615
The matrix-vector product K1 R denes the displacement corrections ujr due to the residual forces
j
(16.53)
ujr = K1 R
67
but they are obviously not invariant for the increment. The displacement corrections for iteration j are then dened as
uj = j uT + ujr
(16.54)
68 Figure 16.11 shows a owchart for an incremental nonlinear nite element program based on
the modied-Newton algorithm with indirect displacement control capabilities. The numbers
in the boxes in Figure 16.11 correspond to those appearing in Figure ??.
16.4.2
Arc-Length
Adapted from (Jirasek and Ba

zant 2001)
69 The basic idea of a exible incrementation control technique is that the step size is specied
by a constraint equation that involves the unknown displacements as well as the load multiplier.
The original motivation was provided by the requirement that the size of the step measured as
the geometric distance between the initial and nal state in the load-displacement space should
be equal to a prescribed constant, Fig. 16.10.
70 Despite the apparent simplicity of the condition of a constant arc length, it must be used with
caution. First of all, it is important to realize that forces and displacements have completely
dierent units, and so the purely geometrical measure of length in the load-displacement space
does not make a good sense. It is necessary to introduce at least one scaling factor, denoted as c,
that multiplies the load parameter and converts it into a quantity with the physical dimension
of displacement. The length of a step during which the load parameter changes by and the
displacements change by u is then dened as
5
(16.55)
l = uT u + (c )2
By adjusting the scaling factor we can amplify or suppress the relative contribution of loads
and displacements. One reasonable choice is derived from the condition that the0
contributions
should be equal as long as the response remains linear elastic, which leads to c = uTe ue where
ue is the solution of Ke ue = f .
71
In some cases, e.g., for frame, plate, and shell models that use both translational and rotational degrees of freedom, the components of the generalized displacement vector u do not have
the same physical dimension. It is then necessary to apply scaling also to the vector u.
72
Consider an incremental solution process controled by the arc-length method. In a typical

step number n, we start with displacements u(n1) and load parameter (n1) computed in
the previous step, and we search for displacements u(n) and load parameter (n) . The state
at the end of the step must satisfy the equations of equilibrium between the internal forces
f int (u(n) ) and external forces f ext ((n) ). Compared to the load control or direct displacement
control, the load parameter is an additional unknown. The corresponding additional equation
is provided by the constraint that xes the size of the step. For example, we can require that
We
the length of the step evaluated from formula (16.55) be equal to a prescribed value, l.
could treat the problem as a system of Ndf + 1 nonlinear equations, where Ndf is the number
73
Victor Saouma
Draft
1616
Figure 16.11: Flowchart for an incremental nonlinear nite element program with indirect
displacement control
Victor Saouma
Draft
1617
Figure 16.12: Two points on the load-displacement curve satisfying the arc-length constraint
of unknown displacement components (degrees of freedom), and solve it by Newton-Raphson
iteration. However, a more elegant and computationally more ecient procedure treats the
equilibrium equations and the constraint equation to a certain extent separately. Assume for
simplicity that the loading program is described by (??). The linearized equations of equilibrium
in the i-th iteration read
(n,i1)
K(n,i1) u(n,i) = f 0 + (n,i1) f f int
+ (n,i) f
(16.56)
where u(n,i) is the unknown displacement correction, and (n,i) is the unknown correction of
the load parameter. The rst three terms on the right-hand side are known, and the last term
is an unknown scalar multiple of a given vector f . We can therefore separately solve equations
(n,i1)
K(n,i1) u0 = f 0 + (n,i1) f f int

K
(n,i1)
uf
= f
(16.57-a)
(16.57-b)
and then express the displacement correction as

u(n,i) = u0 + (n,i) uf
(16.58)
When this expression is substituted into the constraint condition,

2
(u(n,i1) + u(n,i) )T (u(n,i1) + u(n,i) ) + c2 ((n,i1) + (n,i) )2 = (l)
(16.59)
we obtain a quadratic equation for a single unknown, (n,i) . This equation usually has two real
roots, corresponding to the two points of the equilibrium path that have the prescribed distance
from point (u(n1) , (n1) ); see Fig. 16.12. The correct root is selected depending on the sense in
which we march on the equilibrium path (Criseld, M.A. 1981), and the displacement correction
is determined from (16.58). After standard updates of the displacement vector and the load
parameter, the iteration cycle is repeated until the convergence criteria are satised.
16.4.3
Relative Displacement Criterion

74 The standard arc-length control performs well if the entire structure or its large portion
participates in the failure mechanism. In cases when the failure pattern is highly localized,
robustness of the technique may deteriorate. The remedy is to adapt the constraint equation
to the particular problem and control the incrementation process by a few carefully selected
displacement components.
75 Motivation is again provided by the physical background. If the load-displacement diagram
of a brittle structure exhibits snapback, direct displacement control applied in an experiment
leads to sudden catastrophic failure. When the displacement imposed by the loading device
reaches a critical value, fracture starts propagating even though the imposed displacement at
the load point is kept xed. However, opening of the crack monotonically increases during the
entire failure process, and so it can be used as a control variable. If the experimental setup
is arranged such that the applied force is continuously adjusted depending on the currently
measured value of the crack opening, the response can be traced in a stable manner even after
Victor Saouma
Draft
1618
the point at which the load-displacement diagram snaps back. The same idea can be exploited
by a numerical simulation. It suces to select a suitable linear combination of displacement
components that increases monotonically during the entire failure process, and to use this
combination as the control variable.
de Borst (1985,1986) concluded that arc-length methods (Riks 1979, Ramm 1981, Criseld
1981), which were the original IDC methods, were not satisfactory for analyses involving cracking accompanied by softening.
76
The main problem with the arc-length methods, when used in this context, was that the
constraint involved all displacement components equally when, in fact, only a few displacement
components were dominant. The dominant displacement components were typically those for
nodes at or near the crack mouth. This being the case, de Borst proposed using a transformed
relative displacement component between two nodes as the constraint.
77
78 The transformed relative displacement component can dene the crack mouth opening displacement (CM OD), crack mouth sliding displacement (CM SD), or some arbitrary displacement u between two points on a structure. The arbitrary displacement u may correspond
to a relative displacement measured during an experiment such as the relative vertical displacement between a point on the neutral axis of a 3-point bend beam over a support and the bottom
of the beam at mid-span.
As it is the most general case, the relative displacement criterion will be described in terms
of the arbitrary relative displacement u. A pair of nodes, m and n, are selected to dene u,
with their total displacements being (u)m and (u)n , respectively. The direction associated with
u is dened by a unit vector v. u is thereby dened as
79
u = vT [(u)n (u)m ]
(16.60)
If m and n are nodes on opposite sides of a discrete crack u CM OD if v is normal to

the crack surface and u CM SD if v is tangent to the crack surface. The value for u is
prescribed for an increment and the applied loads are scaled such that the total displacements
at the end of each iteration reect that value.
80
81
Recalling that the total displacements uj+1 at the end of iteration j are dened as
uj+1 = uj + j uT + ujr
(16.61)
the load parameter correction j for iteration j is

j =
16.4.4
1
2
!
"
u vT (uj )n (uj )m vT (ujr )n (ujr )m
vT [(uT )n (uT )m ]
(16.62)
IDC Methods with Approximate Line Searches
82 Employing a procedure proposed by Criseld (1983) for use with the arc-length method,
the convergence of the solution alogrithm can be accelerated by performing approximate line
searches; approximate line searches under xed (i.e. non-scalable) loads are described in Section
??.
Victor Saouma
Draft
1619
83 This procedure requires an extra iterative loop at the beginning of the line search loop in
which a combination of j and sk+1 satisifying the constraint conditions (i.e. Equations ??
and 16.60) is computed. As j is initially computed for s1 = 1.0, any change in sk requires
a corresponding change in j for the IDC constraint to remain satised. Consequently, an
iterative loop, in which j is recomputed based on the estimated value of sk+1 , is required to
obtain a compatible combination of j and sk+1 .
After recomputing j , the values of g0 and gk are also recomputed using Equation 16.30
j+1,k
j+1,k
caused by the new value of j . f int is not
to reect the change in the residual loads R
j+1,k
, which is strictly not correct,
updated to reect the changes in sk+1 when recomputing R
but it does signicantly reduce the number of computations without causing any diculties
(Criseld 1983).
84
Finally, from the new values of g0 and gk , sk+1 is re-estimated using Equation 16.33. The
loop is terminated when
k+1 |
| sk+1
new s
0.05
(16.63)
| sk+1
new |
85
which generally requires only a few iterations. A ow chart of this procedure is shown in Figure
16.13.
86
Since the total displacements uj+1,k are now dened as

uj+1,k = uj + sk (ujr + j uT )
(16.64)
reecting the introduction of the relaxation parameter sk , the IDC constraint equations must
be modied accordingly. j for the stress criterion is now dened as
1
2
ft (j )n + sk (jr )n (n)n
(16.65)
j = min
sk (T )n (n)n
and j for the relative displacement criterion is now dened as
1
2
!
"
u vT (uj )n (uj )m sk vT (ujr )n (ujr )m
j =
sk vT [(uT )n (uT )m ]
(16.66)
It is these general forms of the constraint equations that are implemented in MERLIN.
Victor Saouma
Draft
1620
Figure 16.13: Flow chart for line search with IDC methods
Victor Saouma
Draft
Bibliography
Abramowitz, M. and Stegun, I.: 1970, Handbook of mathematical functions, Technical report,
National Bureau of Standard. Applied Mathematics Series, No. 55.
Babuska, I.: 1971, Error-bounds for nite element methods, Numer. Math 20(3), 179192.
Babuska, I.: 1973, The nite element method with lagrange multipliers, Numer. Math.
20(3), 179182.
Bathe, K.: 1996, Finite Element Procedures, Prentice-Hall Inc.
Brezzi, F.: 1974, On the existence uniqueness and approximation of saddle-point problems
arising from lagrangian multipliers, RAIRO 8-R2 pp. 129151.
Cervenka,
V., Keating, S. and Felippa, C.: 1993, A comparison of strain recovery techniques for
the mixed iterative method, Communications in Applied Numerical Methods 9, 925932.
Criseld, M.A.: 1979, A faster modied newton-raphson iteration, Computer Methods in Applied Mechanics and Engineering 20(3), 267278.
Criseld, M.A.: 1981, A fast incremental/iterative solution procedure that handles snap
through, Computers and Structures 13(1-3), 5562.
Delaunay, B.: 1934, Sur la Sphère Vide, Izv. Akad. Nauk SSSR, Otdelenie Matematicheskii i
Estestvennyka Nauk 7, 793800.
Eisenberg, M. A. and Malvern, L.: 1973, On nite element integration in natural co-ordinates,
International Journal for Numerical Methods in Engineering 7, 574575.
Felippa, C.: 1999, Finite element lecture notes, Technical report, Dept. of Aerospace Engineering, University of Colorado, Boulder.
Felippa, C.: 2000, Lecture notes in advanced nite element methods (asen 5367),
Technical report, Dept. of Aerospace Engineering, University of Colorado, Boulder.
http://caswww.colorado.edu/courses.d/AFEM.d/Home.html.
Haser, N. and Sullivan, J.: 1991, Real Analysis, Dover Publications, New-York.
Jirasek, M. and Bazant, Z.: 2001, Inelastic Analysis of Structures, John Wiley, Chichester.
Kardestuncer, H. (ed.): 1987, Finite Element Handbook, McGraw-Hill.
Okabe, A., Boots, B., Sugihara, K. and Chiu, S.: 2000, Spatial Tessellations; Concepts and
Applications of Voronoi Diagrams, John Wiley & Sons.
Draft
2
BIBLIOGRAPHY
Ottosen, N. and Petersson, H.: 1992, Introduction to the Finite Element Method, Prentice-Hall.
Reich, R.: 1993, On the Marriage of Mixed Finite Element Methods and Fracture Mechanics:
An Application to Concrete Dams, PhD thesis, University of Colorado, Boulder.
Schey, H.: 1973, Div Grad Curl and all That; An Informal Text on Vector Calculus, W.W.
Norton.
Terzaghi, K. and Peck, R.: 1967, Soil Mechanics in Engineering Practice, 2nd edition, John
Wiley & Sons, New York, NY.
Timoshenko, S. and Goodier, J.: 1970, Theory of Elasticity, McGraw Hill.
Cervenka,
J.: 1994, Discrete Crack Modeling in Concrete Structures, PhD thesis, University of
Colorado, Boulder.
Voronoi, G.: 1907, Nouvelles Applications des Paramètres Continus à la Theorie des Formes
Quadratiques, J. Reine Angew. Math 133, 97178.
Xue, W. and Atluri, N.: 1985, Existance and stability, and discrete bb and rank conditions
for general mixed-hybrid nite elements in elasticity, Hybrid and Mixed Finite Element
Methods pp. 91112.
Zienkiewicz, O. C. and Taylor, R. L.: 1989, The Finite Element Method, Vol. 1, Basic Formulation and Linear Problems, 4th ed., McGraw-Hill, London.
Zienkiewicz, O. C., Vilotte, J. P., Toyoshima, S. and Nakazawa, S.: 1985, Iterative method for
constrained and mixed approximations. an inexpensive improvement of fem performance,
Computer Methods in Applied Mechanics and Engineering 51(13), 329.
Victor Saouma
Draft
Appendix A
VECTOR OPERATIONS
1 This appendix covers some elements of Vector Calculus essential to understand some of the
derivations. An excellent reference is (Schey 1973).
A.1
Vector Dierentiation
A eld is a function dened over a continuous region. This includes, Scalar Field g(x),
Vector Field v(x), or Tensor Field T(x).
2
We rst introduce the dierential vector operator Nabla denoted by
i+
j+
k
x
y
z
(A.1)
4 We also note that there are as many ways to dierentiate a vector eld as there are ways of
multiplying vectors, the analogy being given by Table A.1.
Multiplication
uv
dot
uv cross
u v tensor
Dierentiation
v
divergence
v curl
v
gradient
Tensor Order
Table A.1: Similarities Between Multiplication and Dierentiation Operators
A.1.1
5
Derivative WRT to a Scalar
The derivative of a vector p(u) with respect to a scalar u, Fig. A.1 is dened by
p(u + u) p(u)
dp
lim
du u0
u
(A.2)
If p(u) is a position vector p(u) = x(u)i + y(u)j + z(u)k, then

dx
dy
dz
dp
=
i+
j+
k
du
du
du
du
(A.3)
Draft
A2
VECTOR OPERATIONS
p=p (u+ u)- p(u)
C
u)
+
u
(
p
p (u)
Figure A.1: Dierentiation of position vector p

is a vector along the tangent to the curve.
7
If u is the time t, then
dp
dt
is the velocity
In dierential geometry, if we consider a curve C dened by the function p(u) then dp

du is a
vector tangent ot C, and if u is the curvilinear coordinate s measured from any point along the
curve, then dp
ds is a unit tangent vector to C T, Fig. A.2. and we have the following relations
8
Figure A.2: Curvature of a Curve

dp
= T
ds
dT
= N
ds
B = TN
(A.4)
(A.5)
(A.6)
we also note that p dp

ds
Victor Saouma
curvature
(A.7)
1
Radius of Curvature (A.8)
=

= 0 if dp
ds = 0.
Draft
A.1 Vector Dierentiation
A3
Example A-1: Tangent to a Curve

Determine the unit vector tangent to the curve: x = t2 + 1, y = 4t 3, z = 2t2 6t for
t = 2.
Solution:
"
dp
d ! 2
=
(t + 1)i + (4t 3)j + (2t2 6t)k = 2ti + 4j + (4t 6)k
dt
dt
0
dp
=
(2t)2 + (4)2 + (4t 6)2
dt
T =
=
A.1.2
2ti + 4j + (4t 6)k
(2t)2 + (4)2 + (4t 6)2

2
1
2
4i + 4j + 2k
0
= i + j + k for t = 2
3
3
3
(4)2 + (4)2 + (2)2
(1.9-a)
(1.9-b)
(1.9-c)
(1.9-d)
Divergence
The divergence of a vector eld of a body B with boundary , Fig. A.3 is dened by
considering that each point of the surface has a normal n, and that the body is surrounded by
a vector eld v(x). The volume of the body is v(B).
9
v(x)
Figure A.3: Vector Field Crossing a Solid Region

10
The divergence of the vector eld is thus dened as

1
div v(x) lim
v(B)0 v(B)

vndA
(A.10)
where v.n is often referred as the ux and represents the total volume of uid that passes
through dA in unit time, Fig. A.4 This volume is then equal to the base of the cylinder dA
times the height of the cylinder vn. We note that the streamlines which are tangent to the
boundary do not let any uid out, while those normal to it let it out most eciently.
11
The divergence thus measure the rate of change of a vector eld.
Victor Saouma
Draft
A4
VECTOR OPERATIONS
n
dA
v
v.n
Figure A.4: Flux Through Area dA

The denition is clearly independent of the shape of the solid region, however we can gain
an insight into the divergence by considering a rectangular parallelepiped with sides x1 , x2 ,
and x3 , and with normal vectors pointing in the directions of the coordinate axies, Fig. A.5.
If we also consider the corner closest to the origin as located at x, then the contribution (from
12
x3
-e
e3
-e
x3
e2
e1
x2
-e
x2
x1
x1
Figure A.5: Innitesimal Element for the Evaluation of the Divergence
Eq. A.10) of the two surfaces with normal vectors e1 and e1 is

1
[v(x + x1 e1 )e1 + v(x)(e1 )]dx2 dx3
lim
x1 ,x2 ,x3 0 x1 x2 x3
x2 x3
or
1
x1 ,x2 ,x3 0 x2 x3
lim
v(x + x1 e1 ) v(x)
e1 dx2 dx3 =
x1
x2 x3
(A.11)
v
e
(1.12-a)
1
x1 0 x1
v
e1
(1.12-b)
=
x1
lim
hence, we can generalize

div v(x) =
13
v(x)
ei
xi
(1.13)
or alternatively
Victor Saouma
Draft
A5
e1 +
e2 +
e3 )(v1 e1 + v2 e2 + v3 e3 ) (1.14)
x1
x2
x3
v2
v3
vi
v1
+
+
=
= i vi = vi,i
(1.15)
x1 x2 x3
xi
div v = v = (
=
14
The divergence of a vector is a scalar.
15
We note that the Laplacian Operator is dened as

2 F F = F,ii
(1.16)
Example A-2: Divergence

Determine the divergence of the vector A = x2 zi 2y 3 z 2 j + xy 2 zk at point (1, 1, 1).
Solution:

i+
j+
k (x2 zi 2y 3 z 2 j + xy 2 zk)
v =
x
y
z
x2 z 2y 3 z 2 xy 2 z
+
+
=
x
y
z
= 2xz 6y 2 z 2 + xy 2

= 2(1)(1) 6(1) (1) + (1)(1) = 3 at (1, 1, 1)

2
16
(1.17-a)
(1.17-b)
(1.17-c)
(1.17-d)
By analogy to Eq. A.10, the divergence of a second-order tensor eld T is

1
T = div T(x) lim
v(B)0 v(B)

TndA
(1.18)
which is the vector eld

T =
A.1.3
Gradient
A.1.4
Scalar
Tpq
eq
xp
(1.19)
The gradient of a scalar eld g(x) is a vector eld g(x) such that for any unit vector v,
the directional derivative dg/ds in the direction of v is given by
17
dg
= gv
ds
Victor Saouma
(1.20)
Draft
A6
VECTOR OPERATIONS
where v = dp
ds We note that the denition made no reference to any coordinate system. The
gradient is thus a vector invariant.
To nd the components in any rectangular Cartesian coordinate system we use
18
v =
dg
ds
dxi
dp
=
ei
ds
ds
g dxi
xi ds
(1.21-a)
(1.21-b)
which can be substituted and will yield

g =
or
g
ei
xi
i+
j+
k
x
y
z
i+
j+
k
x
y
z
(1.22)

=
(1.23-a)
(1.23-b)
and note that it denes a vector eld.

19 The physical signicance of the gradient of a scalar eld is that it points in the direction in
which the eld is changing most rapidly (for a three dimensional surface, the gradient is pointing
along the normal to the plane tangent to the surface). The length of the vector ||g(x)|| is
perpendicular to the contour lines.
20
g(x)n gives the rate of change of the scalar eld in the direction of n.
Example A-3: Gradient of a Scalar

Determine the gradient of = x2 yz+4xz 2 at point (1, 2, 1) along the direction 2ij2k.
Solution:
= (x2 yz + 4xz 2 ) = (2xyz + 4z 2 )i + (x2 zj + (x2 y + 8xz)k
= 8i j 10k at (1, 2, 1)
1
2
2
2i j 2k
= i j k
n = 0
2
2
2
3
3
3
(2) + (1) + (2)

1
2
16 1 20
37
2
i j k =
+ +
=
n = (8i j 10k)
3
3
3
3
3
3
3
(1.24-a)
(1.24-b)
(1.24-c)
(1.24-d)
Since this last value is positive, increases along that direction.
Example A-4: Stress Vector normal to the Tangent of a Cylinder

Victor Saouma
Draft
A7
The stress tensor throughout a continuum is
3x1 x2
= 5x22
0
given with respect to Cartesian axes as
5x22 0
(1.25)
0 2x23
2x3 0
Determine the stress vector (or traction) at the point P (2, 1, 3) of the plane that is tangent
to the cylindrical surface x22 + x23 = 4 at P , Fig. A.6.
x3
n
x2
P
3
1
Figure A.6: Radial Stress vector in a Cylinder

Solution:
At point P , the stress tensor is given by
6
5
0
0
2 3
= 5
0
0 2 3
(1.26)
The unit normal to the surface at P is given from
At point P ,
(x22 + x23 4) = 2x2 22 + 2x3 e3
(1.27)
(x22 + x23 4) = 222 + 2 3e3
(1.28)
3
1
e3
n = e1 +
2
2
determined from

5
0
0 5/2
1/2
3
0
2 3
=

3/2
3
0
2 3
and thus the unit normal at P is
Thus the traction vector will be
6
= 5
0
or tn = 52 e1 + 3e2 + 3e3
Victor Saouma
(1.29)
(1.30)
Draft
A8
21
VECTOR OPERATIONS
We can also dene the gradient of a vector eld as
[x v] =
[vx ] =
vx
x
vx
y
vx
z
vx
x
vy
x
vz
x
x
vy
y
vy
z
vx
y
vy
y
vz
y
vz
x
vz
y
vz
z
vx
z
vy
z
vz
z
(1.31)
(1.32)
that is [v]ij gives the rate of change of the ith component of v with respect to the jth
coordinate axis.
Note the diference between vx and x v. In matrix representation, one is the transpose
of the other.
22
23
The gradient of a vector is a tensor of order 2.
24 We can interpret the gradient of a vector geometrically, Fig. A.7. If we consider two points
a and b that are near to each other (i.e s is very small), and let the unit vector m points in
the direction from a to b. The value of the vector eld at a is v(x) and the value of the vector
eld at b is v(x + sm). Since the vector eld changes with position in the domain, those two
vectors are dierent both in length and orientation. If we now transport a copy of v(x) and
place it at b, then we compare the dierences between those two vectors. The vector connecting
the heads of v(x) and v(x + sm) is v(x + sm) v(x), the change in vector. Thus, if we
divide this change by s, then we get the rate of change as we move in the specied direction.
Finally, taking the limit as s goes to zero, we obtain
lim
s0
v(x + sm) v(x)

Dv(x)m
s
(1.33)
v(x+ s m ) -v(x)
v(x+ s m )
x3
v(x)
a sm b
x2
x1
Figure A.7: Gradient of a Vector
The quantity Dv(x)m is called the directional derivative because it gives the rate of
change of the vector eld as we move in the direction m.
Victor Saouma
Draft
A.2 Vector Integrals
A.2
A.2.1
25
A9
Vector Integrals
Integral of a Vector
The integral of a vector R(u) = R1 (u)e1 + R2 (u)e2 + R3 (u)e3 is dened as

R(u)du = e1 R1 (u)du + e2 R2 (u)du + e3 R3 (u)du
d
(S(u)), then
if a vector S(u) exists such that R(u) = du

d
(S(u)) du = S(u) + c
R(u)du =
du
A.2.2
(1.34)
(1.35)
Line Integral
Given r(u) = x(u)e1 + y(u)e2 + z(u)e3 where r(u) is a position vector dening a curve
C connecting point P1 to P2 where u = u1 and u = u2 respectively, anf given A(x, y, z) =
A1 e1 + A2 e2 + A3 e3 being a vectorial function dened and continuous along C, then the integral
of the tangential component of A along C from P1 to P2 is given by

P2
Adr = Adr = A1 dx + A2 dy + A3 dz
(1.36)
26
P1
If A were a force, then this integral would represent the corresponding work.
27
If the contour is closed, then we dene the contour integral as

,
Adr = A1 dx + A2 dy + A3 dz
C
28
(1.37)
It can be shown that if A = then

P2
Adr
P1
Adr = 0
is independent of the path C connecting P1 to P2
(1.38-a)
along a closed contour line
(1.38-b)
A.2.3
29
Integration by Parts
The integration by part formula is

a
Victor Saouma
u(x)v (x)dx = u(x)v(x)|ba
v(x)u (x)dx
(1.39)
Draft
A10
A.2.4
30
VECTOR OPERATIONS
Gauss; Divergence Theorem
In the most general case we have

F =
(1.40)
31 The divergence theorem (also known as Ostrogradskis Theorem) comes repeatedly in solid
mechanics and can be stated as follows:

(1.41)
vd =
v.nd or
vi,i d =
vi ni d
That is the vector divergence over a volume is equal to the vector ux over a surface.
32
For 2D-1D transformations, we have

,

qdA =
A
qT nds
(1.42)
33
This theorem is sometime refered to as Greens theorem in space.
34
Green-Gauss theorem

vT nd
vd =
()T vd
If we select vT = [ 0 0 ], we obtain

d =
d
nxd
x
A.2.5
35
(1.43)
(1.44)
Stokes Theorem
Stokes theorem states that

,
Adr =
(A)ndS =
(A)dS
C
(1.45)
where S is an open surface with two faces conned by C
A.2.6
36
Green; Gradient Theorem
Greens theorem in plane is a special case of Stokes theorem.

,
(Rdx + Sdy) =
Example A-5:
Victor Saouma
S R
x
y

dxdy
(1.46)
Physical Interpretation of the Divergence Theorem

Draft
A.2 Vector Integrals
A11
Provide a physical interpretation of the Divergence Theorem.

Solution:
A uid has a velocity eld v(x, y, z) and we rst seek to determine the net inow per unit
time per unit volume in a parallelepiped centered at P (x, y, z) with dimensions x, y, z,
Fig. A.8-a.
Z
C
V
P(X,Y,Z) H
X
B
a)
S
dV=dxdydz
n
Vt
dS
dS
b)
c)
Figure A.8: Physical Interpretation of the Divergence Theorem
vx |x,y,z

vx xx/2,y,z
vx
(1.47-a)
1 vx
x
2 x

1 vx
x
vx x+x/2,y,z vx +
2 x
The net inow per unit time across the x planes is

1 vx
x yz vx
vx +
Vx =
2 x
vx
xyz
=
x
Similarly
vx
AFED
GHCB

1 vx
x yz
2 x
vy
xyz
y
vz
xyz
Vz =
z
Hence, the total increase per unit volume and unit time will be given by
(
'
vy
vx
vz
+
+
xyz
x
y
z
= div v = v
xyz
Vy =
Victor Saouma
(1.47-b)
(1.47-c)
(1.48-a)
(1.48-b)
(1.49-a)
(1.49-b)
(1.50)
Draft
A12
VECTOR OPERATIONS
Furthermore, if we consider the total of uid crossing dS during t, Fig. A.8-b, it will be given
by (vt)ndS = vndSt or the volume of uid crossing dS per unit time is vndS.
Thus for an arbitrary
volume, Fig. A.8-c, the total
amount of uid crossing a closed surface

vdV (Eq. 1.50), thus
vndS. But this is equal to
S per unit time is

S
vdV
vndS =
S
(1.51)
which is the divergence theorem.
Victor Saouma
Draft
Appendix B
CASE-STUDY: FRACTURING of
A DAM DUE TO THERMAL
LOAD
B.1
INTRODUCTION
The dam is a 60 m high arch gravity dam built in the early fties. In its original conguration,
the crest supported a 5 m wide road bridge. There is no indication of any malfunction until the
ood of 1981 occurred. At that time, the bridge abutment were clogered by trees and debris,
preventing the eective overtopping of the water.
Following that incident, it was decided to dismantle the bridge from the dam crest, and a
new one, parallel to the crest chord, was built.
Soon after, a crack was observed inside the entire upper gallery, on the downstream face.
Crack mouth opening displacement were recorded, and crack maps drawn. Unfortunately,
readings were not taken systematically at the same pool elevation (which varies by +/- 10 m),
and no clear trend can be observed. However, it should be noted that the crack opening is
stable and there is no indication of increased values. Finally, no observation were conducted on
the downstream face to detect the presence of any daylighting crack.
B.1.1
Elastic and Thermal Properties
Units adopted:
Mass
Length
Temperature
Time
Kg
m
oK
day
Elastic properties of the concrete are given by Table B.1.

For the thermal properties, we must account for the dierent mixes used. The core and
lower part of the dam had a mix with 180 Kg/m3 of cement, while the dam face and upper
part had a mix of 280 Kg/m3 of cement. The respective thermal properties are shown in Table
B.2.
A key component of the Merlin program (used in this investigation) is the interface element
placed along the crack in a non-linear analysis. The formulation of the interface element is
Draft
B2
CASE-STUDY: FRACTURING of A DAM DUE TO THERMAL LOAD
hair
hwat
1
2,400
36 109
0.2
1 105
34
100
m
Kg/m3
Pa
m/m/o C
W/m2 o C
W/m2 o C
Thickness
Mass density
Youngs modulus
Poissons ratio
Coecient of thermal expansion
Coecient for heat transfer by convection, air
Coecient for heat transfer by convection, water
Table B.1: Concrete Material Properties

Cement Content
mz [Kg/m3 ]
180
280
Specic heat
cb [J/Kg.K]
1,000
1,040
Thermal conductivity
k [J/sec.m.K] J/day.m.K
2.7
233,280
2.7
233,280
Table B.2: Thermal Properties of the concrete

briey described in the Appendix, and the selected properties are shown in Table B.3.
fc
ft
Kt
Kn
F
D
GIF
GIIF
uDmax
s1
w1
c1
cw1
25 106
3.75 106
360 109
360 109
40o
20o
400
4,000
0.3
1 102
0.94 106
105
1.21 106
6.2 104
Pa
Pa
Pa
Pa
From Report
From Whittman for Dam Concrete
10 times E
10 times E
N/m
N/m
From Whittman for Dam Concrete

10 times GIF
m
Pa
m
Pa
m
ft /4
0.75 GIF /ft
c/4
0.75 GII F /c
Table B.3: Interface Element Material Properties
B.1.2
Loads
Loads adopted in the analysis are shown in Table B.4.
B.2
ANALYSIS II; Thermal Shock
In this second analysis, we explore a dierent potential cracking scenario: Thermal Shock
during construction. That is we speculate that during construction, possibly in winter, the heat
Victor Saouma
Draft
B.2 ANALYSIS II; Thermal Shock

Pool Level
Gravity
Bridge
B3
786 m
Dam only, not on bridge abutments
62,640 N/m2 (estimated weight of(concrete slab
'
Thermal (After Construction)

Thermal (During Construction)
(6)m(5.8)m(.3)m(2400)Kg/m3 (10)N/Kg
(0.8)m(5)m
Water 5o C, Dam 20o C, Air 20o C
Gallery and outside: 0 o C
Table B.4: Loads applied on the Dam

of hydration caused by the 280 Kg/m3 concrete caused a thermal gradient large enough in the
downstream face to induce cracking.
Hence, our rst step was to create a boundary le which accounts for the dierent concrete
mixes used in in the dam, Fig. B.1
B.2.1
Thermal Analysis
Then, we need to undertake a transient thermal analysis, and the heat of hydration must be
known. Ideally, a staged construction simulation should be undertaken, as this would result in
a good estimate for the heat of hydration.
In this analysis, and as a rst order approximation, we made a number of simplifying
assumptions, the rst one is to consider the heat of hydration for Portland Cement type I
(ASTM) at only 20o C (whereas in actuality, it is well known that this quantity is a function of
the temperature, and such a model would be clearly nonlinear).
In the literature (Neville), heat of hydration is often tabulated or graphed for few time
increments. We adopt such set of values in Table B.5. Bold faced values constitute our starting
point, and we deduce all other quantities (i.e Heat/mass/time in J/Kg/day).
In our incremental analysis, we will be focusing on the rst 50-60 days, thus we must
interpolate a set of heat of hydration from these selected points. This is accomplished through
interpolation in Figure B.2, and extracted values are shown in Table B.6.
We note that these set of values were later multiplied by 2.400 Kg/m3 as Merlin needs the
heat per volume as opposed to heat per mass.
Clearly, a number of simplifying assumptions were made so far, and in addtion to further
simplify the analysis we assume:
1. Consider only zones 6, 7, 8 and 9.
2. Assume that at the interfaces 6-3 and 8-5 the temperature is 25o C, whereas 7-4 it is 20o C
3. Assume that the outside temperature, and the one of the gallery is equal to 0o C
4. Do not perform an incremental analysis (i.e. simulating staged construction)
5. Perform a transient analysis with time increment equal to one day. Heat of hydration
values are taken from day 10 to 60.
6. Assume uniform heat of hydration within the considered zones.
7. Dierentiate between the two concrete mixes.
Victor Saouma
Draft
B4
35 34 33 56
32
58 57
55
59
36
60
37
31
54
61
19
51 18
52
30
24
2953 30
49
34
35
6 31 29
36
46
15 28
14
18
3
17 23
19
13
33
50 48 8
32
22
47
16 26
21
16
20
14 12
13
15
17
25
27
40
39
26
24
11 12
38
27
28
10
41
25 24 42
23 43
22
37
44
11
10
20 45
21
7
8
6
9
9
1
2
2
Figure B.1: Boundary Description of Dam for Transient Thermal Analysis
Victor Saouma
Draft
Age
[days]
Hc
[J/g]
0
1.5
3
5
7
17.5
28
59
90
227.5
365
1,368.5
2,372
B5
Hc
[J/g.day]
Hb = Hc mz /b
[J/Kg.day]
mz
180
280
85.000
6,375
9,917
20.000
1,500
2,333
3.09524
232.14
361.11
0.48387
36.29
56.45
0.10909
8.18
12.73
0.01495
1.12
1.74
255
335
400
430
460
490
Table B.5: Heat of Hydration From the Literature
Days
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
M-180
510
480
440
410
380
340
310
280
250
220
200
180
165
155
140
130
120
119
110
105
M-280
800
745
690
640
580
530
480
440
380
340
300
275
245
230
220
200
190
180
178
165
Days
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
M-180
100
97
94
90
87
85
80
77
75
72
70
67
65
63
61
59
57
55
53
52
M-280
160
154
147
142
135
128
122
115
110
105
100
95
92
87
85
82
80
77
75
73
Days
50
51
52
53
54
55
56
57
58
59
M-180
50
48
47
45
45
43
42
40
38
37
M-280
70
68
66
65
63
62
60
59
59
57
Table B.6: Heat of Hydration Adopted in the Simulation; Days and J/Kg/Day
Victor Saouma
Draft
B6
Figure B.2: Heat of Hydration Interpolations
Victor Saouma
Draft
B7
Contour Plot,
Temperatures, Temperatures
3.00e+1
2.70e+1
2.40e+1
2.10e+1
1.80e+1
1.50e+1
1.20e+1
9.00e+0
6.00e+0
3.00e+0
Y
X
-2.25e-3
Figure B.3: Temperature Distribution in the Transient Thermal Analysis at Day 8

When such an analysis was performed, it was observed that the concrete temperatures were
very small (about 1-2o C). This was attributed to the lack of incremental analysis, of having one
single large lift being investigated, and with an initial temperature of zero, not enough heat
was released to warm the temperature. Clearly this is erroneous.
To qualitatively alleviate for this limitation (caused by lack of incremental analysis), and
keeping in mind that we are primarily interested at this point in determining qualitatively a
mechanism which may cause cracking, the specic heat was reduced by a factor of 10 (no other
reduction factors were attempted).
With such a set of values, the largest temperature gradient was observed to occur at day 8
with the temperature distribution shown in Fig. B.3. We observe the thermal gradient along
the crack line.
B.2.2
Stress Analysis
With the nodal temperature extracted from the previous analysis, a stress analysis was next
undertaken. In this analysis, the only loads considered were thermal and gravity (together in
one increment).
We observe from Fig. B.4 that in this case, the crack did open and the magnitude of the
maximum crack opening displacement is about 0.8 mm. The crest displacements are: -4.1 mm
(horizontal) and 1.6 mm (vertical)
Normal and shear stresses, as well as crack opening and sliding displacements are shown in
Table B.7 and B.8.
Victor Saouma
Draft
B8
Contour Plot,
Principal Stresses, Maximum
1.19e+7
1.05e+7
9.03e+6
7.58e+6
6.14e+6
4.69e+6
3.24e+6
1.80e+6
3.53e+5
-1.09e+6
Y
-2.54e+6
X
Z
Figure B.4: Maximum Principal Stresses and Deformed Mesh at Day 8
Crack Lig.
0.000E+00
4.168E-01
6.947E-01
1.228E+00
1.317E+00
1.690E+00
1.873E+00
2.205E+00
2.537E+00
2.664E+00
2.996E+00
3.059E+00
3.455E+00
3.455E+00
3.959E+00
4.296E+00
4.633E+00
4.969E+00
n
-1.136E+06
-8.677E+05
-9.974E+05
-1.112E+06
-1.123E+06
-7.825E+05
1.907E+05
5.690E+05
8.226E+05
7.725E+05
7.024E+05
6.063E+05
5.663E+05
4.615E+05
3.259E+05
1.699E+05
3.613E+04
3.750E-04
t
-1.341E+06
-1.264E+06
-9.699E+05
-7.180E+05
-7.853E+05
-1.284E+06
-6.107E+05
-5.777E+05
-5.414E+05
-4.123E+05
-3.491E+05
-2.992E+05
-2.918E+05
-2.169E+05
-1.364E+05
-6.244E+04
-1.220E+04
-1.098E-04
Table B.7: Sresses Along the Interface Element; m] and [Pa]
Victor Saouma
Draft
B.3 CONCLUSIONS
Crack Lig.
0.00000E+00
2.77864E-01
5.55728E-01
1.08911E+00
1.17780E+00
1.55140E+00
1.73377E+00
2.06576E+00
2.39775E+00
2.52472E+00
2.85671E+00
2.92020E+00
3.31568E+00
3.31568E+00
3.65225E+00
3.98882E+00
4.32540E+00
4.66197E+00
B9
Crack Opening
-3.73773E-06
-1.99149E-06
-2.75825E-06
-3.21438E-06
-3.30575E-06
-2.63906E-06
4.83795E-06
3.44441E-05
1.05330E-04
1.39448E-04
2.40532E-04
2.63296E-04
3.22496E-04
3.95863E-04
4.98598E-04
6.05647E-04
7.18990E-04
8.19540E-04
Crack Sliding
-3.78857E-06
-3.59965E-06
-2.88216E-06
-1.65068E-06
-1.47816E-06
-4.04809E-06
-1.74152E-05
-3.95564E-05
-6.29998E-05
-7.66532E-05
-1.16057E-04
-1.25226E-04
-1.71896E-04
-1.67985E-04
-2.02377E-04
-2.19765E-04
-2.29582E-04
-2.30053E-04
Table B.8: Crack Opening and Sliding Displacements; [m]
B.2.3
Data Files
The data les

t10.bd
t12.inp
t13.dat
t14.inp
B.3
used in this analysis are given below.

Master le contains all the various options
Thermal analysis
File containing the nodal temperature at day 8
Stress analysis
CONCLUSIONS
The following conclusions may be drawn from this preliminary analysis

1. There are two potential reasons for which cracking might have occurred:
(a) Thermal gradient between cold water and warm downstream air temperature aggravated by the removal of the bridge which was acting as a corbel on a cathedral
buttress (i.e. its dead weight was osetting exural tensile stresses). The analysis
conducted to investigate this scenario was quite reliable quantitatively.
(b) Thermal shock during construction. The analysis conducted toward such a cause is
mostly qualitative. A more rigorous incremental transient thermal analysis would be
needed to further support this theory. However, this analysis may be quite complex
and may be outside the scope of this investigation.
2. Dierential settlement was ruled out as there was no indication of foundation measurable
excessive deformation.
Victor Saouma
Draft
B10
3. A three dimensional analysis has been initiated (hydrostatic load), but due to limitation
of the authors PC, it was not extended to the thermal analysis. It would be interesting
to contrast results of such an analysis with its 2D counterpart.
4. Through these investigations, some of the possibilities oered by modern computational
tools which address cracking were exercised.
5. It would be interesting to determine if indeed upstream cracks are present in the eld (as
this investigation predicts).
Victor Saouma
Draft
Appendix C
MISC.
C.1
Units & Conversion Factors

length, m (meter)
Force, N (Newton)
Mass, Kg (kilogram)
Density, Kg/m3
Temperature, T
Acceleration, m/s2
Stiness, N/m
Stress, Pa = N/m2
Work, energy, N-m=Joule
Power, J/s=W
Convection coecient, h
Heat, J
Heat Source/Sink, Q
Heat ux (q)
Specic heat, c
Thermal conductivity, k
1 inch = 0.0254 m; 1 m = 39.37 inch

1 lb = 4.4482 N; 1 N = 0.22481 lb
1 lbm = 0.45359 Kg; 1 Kg=2.2046 lb
1 lbm/ft3 = 16.018 Kg/m3 ;1 Kg/m3 =0.062428 lbm/ft3
T o F=[(9/5)To C+32]
1 in/s2 = 0.0254 m/s2 ;
1 lb/in = 175.1 N/m
1 psi = 6,894.8 Pa; 1 MPa = 145.04 psi
1 ft-lbf= 1.3558 J; 1 J = 0.73756 ft- lbf
Heat Transfer
1 Btu/h.ft2 .o F = 5.6783 W/m2 .o C
1 Btu=1055.06 J; 1 Btu = 778.17 ft-lb
W/m3 =
1 Btu/h.ft2 = 3.1546 W/m2
1 Btu/o F = 1,899.108 J/o C
1 Btu/h.ft.o F = 1.7307 W/m.o C
Seepage Flow
permeability, k
Stress intensity factor, K
Fracture energy GF
Fracture Mechanics
1 MPa m=1.099 ksi in

1 lb/in =.0057 N/m;
Draft
C2
C.2
MISC.
Metric Prexes and Multipliers

Prex
tera
giga
mega
kilo
hecto
deca
deci
centi
milli
micro
nano
pico
Victor Saouma
Abbreviation
T
G
M
k
h
da
d
c
m
n
p
Multiplier
1012
109
106
103
102
10
101
102
103
106
109
1012

Finite Element II, Solid Mechanics - Victor E.saouma - 2001

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Finite Element II, Solid Mechanics - Victor E.saouma - 2001

Hochgeladen von

Copyright:

Verfügbare Formate

Draft

Lecture Notes in:

Dept. of Civil Environmental and Architectural Engineering

Finite Elements II; Solid Mechanics

Finite Elements II; Solid Mechanics

5 VARIATIONAL and RAYLEIGH-RITZ METHODS

6 INTERPOLATION FUNCTIONS; NATURAL COORDINATE SYSTEMS 61

Finite Elements II; Solid Mechanics

7.1.2 Discretization of the Variational Statement for the HW Variational Principle73

8 STRAIGHT SIDED ELEMENTS; 1st GENERATION

9 ISOPARAMETRIC ELEMENTS; 2nd GENERATION

Finite Elements II; Solid Mechanics

10 ELEMENT FORMULATION and STRAIN RECOVERY in HW FORMULATION

12 FINITE ELEMENT DISCRETIZATION OF THE FIELD

13 TOPICS in STRUCTURAL MECHANICS

Finite Elements II; Solid Mechanics

13.2.3.1 Full Integration . . . . . . . .

Finite Elements II; Solid Mechanics

B CASE-STUDY: FRACTURING of A DAM DUE TO

Finite Elements II; Solid Mechanics

B.2.2 Stress Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B7

Finite Elements II; Solid Mechanics

Summary of Variational Methods . . . . . . . . .

Finite Element Process, (Bathe 1996) . . . . . . . . . . . . . . . . . . . . . . . . 22

Stresses as Tensor Components . . . . . . . . . . . .

Voronoi and Delaunay Tessellation . . . . . . . . . . . . . . . . . . . . . . . .

Tonti Diagram for the Total Potential Energy, (Cervenka,

Tonti Diagram for Hu-Washizu, (Cervenka,

Uniformly Loaded Simply Supported Beam Analysed by the Rayleigh-Ritz Method515

Axial Finite Element . . . . . . . . . . . . . . . . . . . .

Completness and Compatibility .

Rectangular Bilinear Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Two-Dimensional Mapping of Some Elements . . . . . . . . . . . . . . . . . . . . 91

10.1 Patch test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

13.1 Patch Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

Finite Elements II; Solid Mechanics

13.3 Independent Displacement Modes for a Bilinear Element . . . . . . . . . .

16.1 Test Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Finite Element Formulation . . . . . . . . . . . . . . . .

Dierentiation of position vector p . . . . . . . . . . . . .

B.1 Boundary Description of Dam for Transient Thermal Analysis . . . . . . . . . . . B4

Finite Elements II; Solid Mechanics

B.2 Heat of Hydration Interpolations . . . . . . . . . . . . . . . . . . . . . . . . . . . B6

Finite Elements II; Solid Mechanics

Summary of Variational Terms Associated with One Dimensional Elements . . . 14

Selected Examples of Diusion Problems . . . . . . . . . . . . . . . . . . . . . . . 319

Functionals in Linear Elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Characteristics of Beam Element Shape Functions . . . . . . . . . . . . . . . . . 69

Shape Functions and Derivatives for T6 Element . . . . . . . . . . . . . . . . . . 86

Shape Functions, and Natural Derivatives for Q8 Element . . . . . . . . . . . .

10.1 Polynomial orders of the shape functions. . . . . . . . . . . . . . . . . . . . . . . 101

Concrete Material Properties . . . . .

B.6 Heat of Hydration Adopted in the Simulation; Days and J/Kg/Day . . . . . . . . B5

Finite Elements II; Solid Mechanics

Vector of coecients in assumed displacement eld

Element exibility matrix

surface tractions, t are specied

Finite Elements II; Solid Mechanics

Neighbour function to u(x)

Matrix relating nodal dof to structure dof

Finite Elements II; Solid Mechanics

Finite Elements II; Solid Mechanics

Introduction; Course objective; Overview; Notation.