The Institution of Structural Engineers

Computational engineering | iii

Contents
Foreword 1 ix
Foreword 2 x
Foreword 3 xi
Author biography xii
Acknowledgements xiii
Preface xiv
1 Introduction 1
1.1 Computation and the structural engineer 1
1.2 The computer as Engineering Assistant 1
1.3 The purpose of design and analysis 1
1.4 Garbage in, Garbage out 3
1.5 The engineer of today and tomorrow 4
1.5.1 To code or not to code 5
1.5.2 The need for better design 6
1.6 The computational engineer 6
2 Design 8
2.1 Introduction 8
2.1.1 Learning to be creative 15
2.1.2 Structural art 15
2.2 The design process 16
2.3 Design personalities 17
3 Design parametrically 18
3.1 Introduction 18
3.2 How we work with computers 18
3.3 Parametrics 19
3.3.1 La Sagrada Família 19
3.3.2 Selfridges Birmingham 20
3.3.3 Camp Adventure tower 22
3.3.4 The Gherkin 23
3.3.5 King’s Cross concourse 24
3.3.6 Everyday parametrics 26
3.3.7 Programming and scripting 27
3.4 Design communication 30
3.4.1 BIM, interoperability and digital workflows 31
3.4.2 Standards 35
3.4.3 Beyond 3D CAD: 4D, 5D and 6D 36
3.5 Conclusion 36
iv
| The Institution of Structural Engineers
Computational engineering

4 Analysis basics 37
4.1 Analysis — the key things you need to know 37
4.2 What is a model? 37
4.2.1 Simplification 39
4.2.2 The stool paradox 39
4.3 Software choices 40
4.3.1 Components, section design and detailing 40
4.3.2 Dedicated whole-structure design 40
4.3.3 General structural analysis and design 41
4.3.4 Modelling, documentation and communication 41
4.4 FEA 101 41
4.4.1 What is the purpose of FEA and how does it differ from CAD? 41
4.4.2 Essential aspects of an FEA model 42
4.5 Elements in more detail 43
4.5.1 1D elements 44
4.5.2 Meshed elements 46
4.6 Model types 50
4.6.1 One-dimensional models 50
4.6.2 Two-dimensional models 50
4.6.3 Three-dimensional models 51
4.7 Design the analysis — analyse the design 52
4.7.1 Plan 53
4.7.2 Do 53
4.7.3 Check 53
4.7.4 Act 53
5 Modelling structures 54
5.1 What is a good FEA model? 54
5.1.1 Accurate (as necessary) 54
5.1.2 Realistic (as appropriate) 55
5.1.3 Simple (as possible) 56
5.1.4 Useful (and relevant) 57
5.1.5 Checking 58
5.1.6 Saint-Venant’s principle 58
5.2 Applying good modelling 58
5.2.1 Structural types 58
5.2.2 Meshing 59
5.2.3 Trusses 62
5.2.4 Steel frames 63
5.2.5 Concrete 69
5.2.6 Substructure 77
5.2.7 Timber 78
5.2.8 Masonry 78
5.2.9 Modelling for vibration 79
 The Institution of Structural Engineers
Computational engineering |v

6 Analysis methods 83
6.1 Introduction 83
6.2 Static linear analysis and the stiffness matrix 84
6.2.1 Stiffness matrices 85
6.3 P-delta analysis and the geometric stiffness matrix 91
6.4 Buckling analysis and the geometric stiffness matrix 94
6.4.1 Eigenvector analysis 96
6.4.2 Using buckling load analysis 98
6.5 Dynamic analysis and the mass matrix 100
6.5.1 Introduction 100
6.5.2 Modal dynamic analysis 102
6.5.3 Dynamic response analyses 103
6.6 Nonlinear analysis (with and without matrices) 106
6.6.1 Newton-Raphson method 114
6.6.2 Explicit solvers 116
6.6.3 Meshfree and particle methods 121
6.7 Conclusion 122
7 Resolving problems in FEA models 123
7.1 Finite element approximation 125
7.1.1 Ill-conditioning 126
7.1.2 Warnings and errors 128
7.2 Getting the units wrong 128
7.3 Exercising restraint 129
7.3.1 No restraint 129
7.3.2 Too much restraint 129
7.3.3 Confusing ‘restraints’ with ‘releases’ 134
7.4 Forgetting torsion 135
7.5 Misusing offsets 136
7.5.1 Bad offsets 136
7.5.2 Good offsets 138
7.6 Meshing too coarsely or too finely 139
7.6.1 Other mesh errors 142
7.7 Second-order effects 143
7.7.1 Linear when it should be nonlinear 143
7.7.2 Buckling effects 146
7.7.3 Meshing 147
7.7.4 Linear vs. nonlinear buckling 149
7.7.5 Design 149
7.8 Validate the model 150
7.8.1 Valid models 150
7.8.2 Software validation 151
vi
| The Institution of Structural Engineers
Computational engineering

7.9 Verify the model 152

7.9.1 Envelope results 154
7.9.2 Software errors 154
7.9.3 Checklist for model-checking 155
7.10 Conclusion 155
7.11 Further reading 156
8 Optimisation 157
8.1 Why optimise? 157
8.2 What is optimisation? 158
8.3 What are we trying to optimise? 160
8.4 Optimisation types 161
8.4.1 The difference between civil and mechanical engineering optimisation 161
8.5 Topology optimisation 162
8.5.1 Maxwell’s Load Paths and the structural volume 162
8.5.2 Layout optimisation 167
8.5.3 Evolutionary topology optimisation 167
8.5.4 Ground structure 170
8.6 Shape optimisation 170
8.6.1 Nodal adjustment 170
8.6.2 Graphic statics and reciprocal force diagrams 171
8.6.3 Form-finding 179
8.7 Size optimisation 180
8.7.1 Optimality criteria 180
9 Optimisation methods 183
9.1 Introduction 183
9.2 The design space 183
9.2.1 Exploring the design space 184
9.2.2 Design space example 184
9.3 Optimisation methods 187
9.3.1 Deterministic methods 187
9.3.2 Stochastic methods 193
9.4 Optimising optimisation 198
9.4.1 Algorithms, heuristics and metaheuristics 200
9.4.2 Exploration vs. exploitation 201
9.4.3 Multi-objective optimisation and the Pareto Front 201
9.4.4 The Stopping Problem 202
9.4.5 Critical optimisation questions 203
9.5 Conclusion 203
10 Artificial Intelligence and Machine Learning 205
10.1 Introduction 205
10.2 What is Artificial Intelligence? 206
10.2.1 General AI 206
 The Institution of Structural Engineers
Computational engineering | vii

10.2.2 Narrow AI 207

10.3 AI methods 207
10.3.1 Predicate Calculus and automated reasoning 207
10.3.2 Machine Learning 210
10.4 The problems and limitations of AI 218
10.5 How might structural engineers use AI? 222
10.6 Conclusion 223
11 Engineering the future 224
11.1 Change 224
11.2 Digital fabrication 224
11.3 Mobile computing 225
11.4 Cloud computing 226
11.5 Quantum computing 228
11.5.1 Why might quantum computers be useful? 228
11.5.2 Classical computers 228
11.5.3 Quantum computers 229
11.6 Optimisation for zero-carbon construction 233
11.7 What is the future role of the engineer? 234
Author’s note 235
Appendix A: Section properties 237
Appendix B: Strain Energy 239
Strain Energy Density 239
Appendix C: Finite Element Analysis (FEA) 240
Matrices 240
Example bar matrices 240
Matrix determinates 241
Matrix inversion 241
Inversion example 242
Finding eigenvectors and eigenvalues 243
Example eigensolution 243
Appendix D: Buckling 245
Euler buckling 245
Buckling amplification 245
Appendix E: Dynamics 247
Vibration of a string 247
Vibration of a beam 247
Vibration of a building 247
Footfall/human-induced vibration 248
Appendix F: Linear structures 251
Appendix G: Cable structures 255
viii
| The Institution of Structural Engineers
Computational engineering

Appendix H: Arches and domes 256

Corbelled arch 256
Parabola 256
Catenary 256
Analysing arches 257
Appendix I: Brohn diagrams 258
Appendix J: Braess’ Paradox 259
Appendix K: Quantum computing 262
Vector notation 262
Logic gates 263
Entanglement 266
References 267