Zhuming Bi

Practical Guide
to Digital
Manufacturing
First-Time-Right for Designs of
Products, Machines, Processes
and System Integration
With Contribution by Wen-Jun Chris Zhang
Practical Guide to Digital Manufacturing
Zhuming Bi

Practical Guide to Digital

Manufacturing
First-Time-Right for Designs of Products,
Machines, Processes and System Integration
Zhuming Bi
Department of Civil and Mechanical Engineering
Purdue University Fort Wayne
Fort Wayne, IN, USA

With Contrib. by
Wen-Jun Chris Zhang
Department of Mechanical Engineering
University of Saskatchewan
Saskatoon, SK, Canada

ISBN 978-3-030-70303-5 ISBN 978-3-030-70304-2 (eBook)

https://doi.org/10.1007/978-3-030-70304-2

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Switzerland AG 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

Many monographs and conference proceedings have been published on digital manu-
facturing and virtual manufacturing. Most of these resources focus on the introduc-
tion of new advancement on the theories, methodologies, and digitized tools made
by the authors. Few books are available for readers to understand the thorough theo-
retical foundation and a broad scope of commercial computer-aided tools in digital
manufacturing.
This book is unique and valuable to the readers in manufacturing in the sense
that it will (1) introduce the thorough theoretical foundation, historical development,
and enabling technologies of digital manufacturing; (2) serve as a practical guide
for readers to use digital manufacturing tools in the designs of products, machines,
processes, and integrated systems; (3) cover a broad range of computer-aided tools for
geometric modeling, assembly modeling, motion simulation, finite element analysis,
manufacturing process simulation, machining programming, product data manage-
ment, and product lifecycle management; (4) use many real-world case studies to
illustrate the applications of computer-aided tools to address various engineering
design challenges in digital manufacturing, and (5) discuss the impact of cutting
edge technologies, such as cyber-physical system, Internet of Thing (IoT), cloud
computing, Blockchain Technologies, and Industry 4.0, on the advancement of digital
manufacturing.
This book is written as a textbook for 3rd/4th-year undergraduates in Manufac-
turing Engineering, Mechanical Engineering, Industrial Engineering, and System
Engineering and senior-level students in technical or engineering college, and as
a reference book for Mechanical, Manufacturing, and Industrial Engineers who
have the responsibilities of using computer-aided tools for the designs of products,
machines, manufacturing processes, or system integrations and for the graduates who
conduct researches relevant to manufacturing.

Zhuming Bi, Ph.D.

Professor of Mechanical Engineering
Wen-Jun Chris Zhang, Ph.D., P.Eng.
Professor of Mechanical and Biomedical Engineering

v
Contents

1 Human Civilization, Products, and Manufacturing . . . . . . . . . . . . . . . . 1

1.1 Human Civilization and Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Drivers for Manufacturing Technologies . . . . . . . . . . . . . . . . . . . . . . 6
1.2.1 Pushing Driver—Human Society . . . . . . . . . . . . . . . . . . . . . . 8
1.2.2 Pulling Driver—Consumer Markets . . . . . . . . . . . . . . . . . . . 9
1.2.3 Reciprocating Driver—Relevant Technology
Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3 Formulation of Engineering Design Problems . . . . . . . . . . . . . . . . . 13
1.4 Computers in Engineering Design . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.5 Computers in Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.6 Emerging Digital Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.6.1 Internet of Things (IoT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.6.2 Big Data Analytics (BDA) . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.6.3 Cyber-Physical Systems (CPS) . . . . . . . . . . . . . . . . . . . . . . . . 24
1.6.4 Cloud Computing (CC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.6.5 Blockchain Technology (BCT) . . . . . . . . . . . . . . . . . . . . . . . . 25
1.6.6 Rapid Prototyping (RP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.7 Digital Skills in Modern Manufacturing . . . . . . . . . . . . . . . . . . . . . . 27
1.8 Organization of Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2 Computer-Aided Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.2 Computer-Aided Design (CAD) System . . . . . . . . . . . . . . . . . . . . . . 37
2.3 Properties of Solid Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.3.1 Legality of Solid Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.3.2 Geometric Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.4 Graphic Modeling Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.4.1 Wireframe Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.4.2 Surface Modeling and Boundary Representation
Modeling (B-Rep) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.4.3 Solid Modeling and Space Decomposition . . . . . . . . . . . . . . 49

vii
viii Contents

2.4.4 Feature-Based Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

2.4.5 Knowledge-Based Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.5 Design Parameters, Features, and Intents . . . . . . . . . . . . . . . . . . . . . . 58
2.5.1 Design Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.5.2 Design Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.5.3 Design Intents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
2.6 Modeling Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
2.7 Assembly Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
2.7.1 Terminologies in Assembly Modeling . . . . . . . . . . . . . . . . . . 79
2.7.2 Modeling Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.7.3 New Assembly-Level Features . . . . . . . . . . . . . . . . . . . . . . . . 89
2.7.4 Exploded Views and Bill of Materials . . . . . . . . . . . . . . . . . . 91
2.8 Kinematic and Dynamic Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
2.8.1 Link Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
2.8.2 Joint Types and Degrees of Freedom (DOF) . . . . . . . . . . . . 94
2.8.3 Kinematic Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
2.8.4 Mobility of Mechanical Systems . . . . . . . . . . . . . . . . . . . . . . 96
2.8.5 Motion Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
2.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3 Computer-Aided Engineering (CAE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
3.2 Design Analysis Methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
3.3 Numerical Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
3.4 Finite Element Analysis (FEA) and Modeling Procedure . . . . . . . . 126
3.5 FEA Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
3.5.1 Governing Equations of Engineering Problems . . . . . . . . . . 132
3.5.2 Solution by Direct Formulation . . . . . . . . . . . . . . . . . . . . . . . 134
3.5.3 Principle of Minimum Potential Energy . . . . . . . . . . . . . . . . 139
3.5.4 Weighted Residual Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3.5.5 Types of Differential Equations . . . . . . . . . . . . . . . . . . . . . . . 149
3.5.6 Solutions to Different PDEs . . . . . . . . . . . . . . . . . . . . . . . . . . 152
3.5.7 System Model and Solution . . . . . . . . . . . . . . . . . . . . . . . . . . 163
3.6 CAE Implementation—SolidWorks Simulation . . . . . . . . . . . . . . . . 178
3.6.1 Computational Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
3.6.2 Materials Library . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
3.6.3 Meshing Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
3.6.4 Analysis Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
3.6.5 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
3.6.6 Solvers to FEA Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
3.6.7 Post-processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
3.6.8 Design Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
3.7 CAE Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
3.7.1 Structural Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
Contents ix

3.7.2 Modal Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

3.7.3 Heat Transfer Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
3.7.4 Transient Heat Transfer Problems . . . . . . . . . . . . . . . . . . . . . 201
3.7.5 Fluid Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
3.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
4 Computer-Aided Manufacturing (CAM) . . . . . . . . . . . . . . . . . . . . . . . . . 223
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
4.2 Material Characteristic, Structures, and Properties . . . . . . . . . . . . . . 224
4.3 Composite Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
4.4 Geometric Dimensioning and Tolerancing (GD&T) . . . . . . . . . . . . 241
4.4.1 Datum Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
4.4.2 Geometric Tolerancing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
4.4.3 Basic Concepts of Dimensioning . . . . . . . . . . . . . . . . . . . . . . 244
4.4.4 Engineering Fits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
4.4.5 Computer-Aided GD&T (DimXpert) . . . . . . . . . . . . . . . . . . 251
4.5 Manufacturing Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
4.5.1 Shaping Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
4.5.2 Design and Planning of Manufacturing Process . . . . . . . . . . 258
4.6 Simulation of Manufacturing Processes—Mold Filling
Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
4.6.1 Injection Molding and Machines . . . . . . . . . . . . . . . . . . . . . . 259
4.6.2 Moldability and Design Factors . . . . . . . . . . . . . . . . . . . . . . . 261
4.6.3 Design of Injection Molding Systems . . . . . . . . . . . . . . . . . . 262
4.6.4 Design Variables and Considerations . . . . . . . . . . . . . . . . . . . 263
4.6.5 Defects of Molded Parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
4.6.6 Mold Flow Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
4.7 Designs of Tools, Dies, and Molds (TDM) . . . . . . . . . . . . . . . . . . . . 272
4.7.1 Design Criteria of TDM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
4.7.2 Computer-Aided Mold Design . . . . . . . . . . . . . . . . . . . . . . . . 275
4.7.3 Shrinkage Compensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
4.7.4 Draft Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
4.7.5 Parting Lines and Shut-Off Surfaces . . . . . . . . . . . . . . . . . . . 279
4.7.6 Parting Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
4.7.7 Mold Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
4.7.8 Mold Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
4.8 Computer-Aided Fixture Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
4.8.1 Functional Requirements (FRs) . . . . . . . . . . . . . . . . . . . . . . . 284
4.8.2 Axioms for Geometric Control . . . . . . . . . . . . . . . . . . . . . . . . 286
4.8.3 Axioms for Dimensional Control . . . . . . . . . . . . . . . . . . . . . . 288
4.8.4 Axioms for Mechanical Control . . . . . . . . . . . . . . . . . . . . . . . 288
4.8.5 Form Closure and Force Closure . . . . . . . . . . . . . . . . . . . . . . 289
4.8.6 Fixture Design in Manufacturing Processes . . . . . . . . . . . . . 290
4.8.7 Computer-Aided Fixture Design (CAFD) . . . . . . . . . . . . . . . 292
x Contents

4.9 Computer-Aided Machining Programming . . . . . . . . . . . . . . . . . . . . 295

4.9.1 Procedure of Machining Programming . . . . . . . . . . . . . . . . . 297
4.9.2 Standards of Motion Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
4.9.3 Default Coordinate Systems and Planes . . . . . . . . . . . . . . . . 299
4.9.4 Machine, Part, and Tool References . . . . . . . . . . . . . . . . . . . . 302
4.9.5 Absolute and Incremental Coordinates . . . . . . . . . . . . . . . . . 303
4.9.6 Types of Motion Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
4.9.7 Programming of Machining Processes . . . . . . . . . . . . . . . . . 306
4.9.8 Automatically Programmed Tools (APT) . . . . . . . . . . . . . . . 307
4.9.9 Computer-Aided Machining Programming . . . . . . . . . . . . . 307
4.10 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
5 Computer Integrated Manufacturing (CIM) . . . . . . . . . . . . . . . . . . . . . . 321
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
5.1.1 Continuous and Discrete Manufacturing Systems . . . . . . . . 322
5.1.2 Variety, Quantity, and Quality . . . . . . . . . . . . . . . . . . . . . . . . . 325
5.1.3 Decoupled Points in Production . . . . . . . . . . . . . . . . . . . . . . . 326
5.2 Manufacturing System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . 328
5.3 Production Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
5.3.1 Machine Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
5.3.2 Material Handling Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
5.3.3 Fixtures, Molds, Dies, and Tools . . . . . . . . . . . . . . . . . . . . . . 335
5.3.4 Facilities for Other Manufacturing Operations . . . . . . . . . . . 336
5.3.5 Manufacturing System Layouts . . . . . . . . . . . . . . . . . . . . . . . 336
5.4 Cellular Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
5.4.1 Design of Cellular Manufacturing System . . . . . . . . . . . . . . 338
5.4.2 Group Technology (GT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
5.4.3 Production Flow Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
5.4.4 Cellular Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
5.5 Discrete Event Dynamic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
5.6 Simulation of Discrete Event Dynamic Systems . . . . . . . . . . . . . . . . 360
5.6.1 Modelling Paradigms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
5.6.2 Object Types and Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
5.6.3 Types of Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
5.6.4 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
5.7 Computer Integrated Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . 367
5.8 Computer-Aided System Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . 373
5.8.1 Sustainability of Manufacturing Systems . . . . . . . . . . . . . . . 374
5.8.2 Main Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
5.8.3 SolidWorks Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
5.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
Contents xi

6 Digital Manufacturing (DM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389
6.2 Functional Requirements (FRs) of Digital Twin . . . . . . . . . . . . . . . . 390
6.2.1 Data Availability, Accessibility, and Transparency . . . . . . . 391
6.2.2 Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
6.2.3 Coordination, Collaboration, and Cooperation . . . . . . . . . . . 393
6.2.4 Decentralization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
6.2.5 Reconfigurability, Modularity, and Composability . . . . . . . 393
6.2.6 Resiliency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
6.2.7 Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
6.3 System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
6.4 Example of Digital Engineering—Reverse Engineering . . . . . . . . . 398
6.4.1 Forward Engineering (FE) and Reverse
Engineering (RE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
6.4.2 Procedure of Reverse Engineering . . . . . . . . . . . . . . . . . . . . . 400
6.4.3 Reverse Engineered Modeling . . . . . . . . . . . . . . . . . . . . . . . . 401
6.4.4 Computer-Aided Reverse Engineering (CARE) . . . . . . . . . . 402
6.5 Example of Digital Engineering—Direct Manufacturing . . . . . . . . 406
6.5.1 Prepare Digital Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
6.5.2 Prepare STL Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
6.5.3 Slicing Algorithms and Visualization . . . . . . . . . . . . . . . . . . 412
6.5.4 Setup Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412
6.5.5 Building Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
6.5.6 Post-processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
6.5.7 Verification and Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . 414
6.6 Studies in Customizing Digital Manufacturing Technologies . . . . . 414
6.6.1 Ubiquitous Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414
6.6.2 Holistic Multi-sensing Solution for Real-Time
Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
6.6.3 Methods of Coping with Big Data . . . . . . . . . . . . . . . . . . . . . 416
6.6.4 Methods of Data Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417
6.6.5 Data Visualization Methods for Human-System
Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
6.6.6 Data-Driven Decision-Making Units . . . . . . . . . . . . . . . . . . . 418
6.6.7 Methods for Workflow Compositions . . . . . . . . . . . . . . . . . . 419
6.6.8 Standardization of Specifications . . . . . . . . . . . . . . . . . . . . . . 419
6.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
Chapter 1
Human Civilization, Products,
and Manufacturing

Abstract The importance of manufacturing to the human civilization is discussed;

main drivers to advance manufacturing technologies are discussed and classified.
The technological evolution over the human history is examined to gain a good
understanding of the needs, limitations, opportunities, and challenges of modern
manufacturing technologies. Critical emerging technologies including Internet of
things (IoT), Big Data Analytics (BDA), Cloud Computing (CC), Blockchain Tech-
nology (BCT), and Rapid Technologies (RP) are briefly introduced. The roles of
computer-aided technologies in digital manufacturing are discussed. It is found that
engineers should master digital skills to perform engineering practices in modern
manufacturing. The organization of the book is presented as the conclusion of this
chapter.

Keywords Human civilization · Human Development Index (HDI) · Gross

National Income (GNI) · Primary industry · Secondary industry · Territory
industry · Computer-aided technologies · Product Life Cycle (PLC) · Internet of
Things (IoT) · Digital Manufacturing (DM) · Big Data Analytics (BDA) · Cloud
Computing (CC) · Blockchain technology (BCT) · Rapid Prototyping (RP)

1.1 Human Civilization and Products

Human civilization reflects the advancement of human society in the forms of govern-
ment, culture, industry, and common social norms (Sullivan 2020). Human civiliza-
tion can be measured by HDI, which is a comprehensive measure of the averaged
achievement of human development at the dimensions of quality of life, education
and knowledge, and standard of living. Accordingly, HDI in these dimensions is
quantified by three indices, i.e., life expectation, average years of schooling, and
GNI per capita. The scores of these indicators are composed and normalized as an
HDI (1) to reflect the advancement of human civilization at specific time and (2) to
evaluate the advancement trend of human civilization within a specified period. For
example, Table 1.1 shows the historical relations of life expectation, average years
of schooling, and GNI and HDI from 1990 to 2018 by data (UNDP UNPD 2020a,

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1

Z. Bi, Practical Guide to Digital Manufacturing,
https://doi.org/10.1007/978-3-030-70304-2_1
2 1 Human Civilization, Products, and Manufacturing

Table 1.1 The historical relation of GNI per capital and HDI by data (1990–2018) (UNDP 2020a,
b)
Life expectancy Expected years Mean years of
Year GNI per capita HDI value
at birth of schooling schooling
1990 75.2 15.4 12.3 37154 0.860
1995 76.1 16.0 12.7 39449 0.878
2000 76.8 15.1 12.7 45974 0.881
2005 77.7 15.7 12.8 49980 0.896
2010 78.7 16.2 15.4 50297 0.911
2015 78.9 16.2 13.3 54039 0.917
2016 78.9 16.3 13.4 54443 0.919
2017 78.9 16.3 13.4 55351 0.919
2018 78.9 16.3 13.4 56240 0.920

b), and it shows that the human civilization has been improved monotonously with
the increases of three indices in our human society.
Three constitutive components of HDI are correlated to each other since all of
them are determined by the availabilities of products and services. For example, a
further look into the historical data in Table 1.1 shows the relation of HDI and GNI
and that of GNI and the life expectance at birth in Figs. 1.1 and 1.2, respectively.
While both of GNI and the life expectance at birth have contributed to the continuous
improvement of HDI, the relations of GNI and life expectance at birth are very
complex. In general, an increase of GNI leads to an increase of life expectance at
birth. The greater a GNI is, the more income people can allocate to obtain products
and services for more education and prolonged life spans, and the lower likelihood
of diseases or premature deaths occur to people. Sufficient evidences have shown
that in a nation, people at a lower level of incomes are less healthy than those at a

60000 0.93

0.92
55000
HDI 0.91
Gross naƟonal income (GNI)
per capƟal in dollars)

0.9
50000

GNI per 0.89

HDI

45000
capital
0.88

0.87
40000
0.86

35000 0.85
1990 1995 2000 2005 2010 2015 2020
Year

Fig. 1.1 The historical relation of GNI per capital and HDI from 1990 to 2018
1.1 Human Civilization and Products 3

60000 79

Life expectancy 78.5

55000 at birth
78
Gross naƟonal income (GNI)

Life expectancy at birth

per capƟal in dollars)

77.5
50000

GNI per 77

45000
capital
76.5

76
40000
75.5

35000 75
1990 1995 2000 2005 2010 2015 2020
Year

Fig. 1.2 Gross national income (GNI) per capital and life expectancy

higher level of incomes. The relationship of incomes and expected life experience is
monotonously ascending; they are connected step-wise at every level of the economic
ladder. For example, GNI can be interpreted into family incomes, which have been
found to have the direct impact on the probabilities of occurrences of major diseases
and illnesses that threaten humans’ lives.
The survey by Woolf et al. (2015) gave the results in Figs. 1.3 and 1.4. It has
showed that the reported rates of disease were higher for low-income Americans and
that were accompanied by the higher rates of risk factors. For example, the obesity
rates for rich people and poor people were 21.2% and 31.9%, respectively. It was
found that such a difference was caused by the standards of living: people earned
<$35,000 a year had three-time higher chance of smoking than those earned more
than $100,000 a year. Only 36.1% of poor people had recommended levels of aerobic
exercise in comparison with 60.1% of rich people. Similarly, Fig. 1.4 showed that
incomes were associated with mental health: in comparison with people with an
average family income for more than $100,000 a year, the people in a family with
less $35,000 a year were four times or five times more likely nervous and sad. Similar
trends occurred to somatic complaints such as pains and other ailments related to
stress and depression).
The dependence of three indices for HDI is illustrated in Fig. 1.5. All aspects of
GNI, education and knowledge, and standards of living are built on the availabilities
of products. Products can be (1) goods for tangible things such as cars, phones,
and computers and (2) services for intangible things such as education, banking,
personal care, and entertainments (McCloy 1984). Products can be made in the ways
of (1) direct production where people provide own goods and services or (2) indirect
production where people cooperate with each other to produce the required goods and
services. On the one hand, direct production is inefficient and only applicable to some
4 1 Human Civilization, Products, and Manufacturing

Fig. 1.3 The impact of family income on prevalence of physical diseases and illnesses (Woolf et al.
2015)

Fig. 1.4 The impact of family income on prevalence of mental illnesses (Woolf et al. 2015)
1.1 Human Civilization and Products 5

Human Development
Index (HDI)

Main Life Expectancy Education Gross National

Indices Index Index Index (GNI) per capital

Main Education and Standard of Living

Dimensions Quality of Life Knowledge

Products Goods Services

Industry

Primary Secondary Tertiary

Industry Industry Industry

Fig. 1.5 Dependence of human civilization on primary, secondary, and tertiary industries

basic living needs. On the other hand, an indirection production led to the division
of labor for the specializations in the modern society. Humans’ wealth came from
the primary, secondary, and tertiary industries with these specializations (McCloy
1984).
The standard of living is reflected by the value of goods and services that people
can access and use; while GNI per capital determines what people can afford to
buy. Alternatively, gross domestic product (GDP) is proposed as a measure of the
market value of all of the products in a given time period. Furthermore, the products
are made in three industries, i.e., primary industry, secondary industry, and tertiary
industry. Note that an industry refers to a sector that applies the technologies to
produce goods or services through a series of economic activities. In primary industry,
natural resources are exploited in three different ways: (1) extrusion such as mining,
quarry, and fishing to extract raw materials (such as oil, gas, metallic ores, diamonds,
and sands) from the nature directly, (2) cultivation such as agriculture and forestry
to obtain basic living supplies (fruits, vegetables, construction materials, timber,
etc.), and (3) animal husbandry such as fish forming and cattle rearing for meat,
milk, and furs. In secondary industry, the products from the primary industry are
6 1 Human Civilization, Products, and Manufacturing

used as raw materials and converted into manufactured goods. Manufacturing goods
are classified into (1) capital goods such as machinery and tooling which are used
by other manufacturers and (2) consumer goods such as appliance and furniture
which are used by end users. The secondary industry also include the businesses
in utilities and constructions where manufactured goods are assembled as houses,
factories, and plants. The tertiary industry provides services as products, and the
services can be classified into (1) commercial services such as delivering goods
to customers and providing financial supports which are crucial to the businesses in
primary or secondary industries (2) personal services such as taxing, mail-delivering,
and healthcare which are relevant to the quality and standards of living. Note that
the majority of services in the tertiary industry rely on the wealth produced in the
primary and secondary industries; therefore, it is worth to have a close look on the
three industry sectors for their roles of wealth creation.
HDI is determined by the available wealth to earn incomes, gain knowledge,
and improve standards of living, and it is interesting to see that the wealth of our
human society is not just the accumulated commodities or resource reserves that
we possess, the wealth includes the productive knowledge of humans to transform
existing resources into desirable goods or services. Ruby (2003) argued that the
wealth was an embodiment of physical and human capital, and it was used to generate
the incomes to pay for desired goods and services X i (i = 1, 2, …. n),

X i = f (L i , K i , Mi ) for all i = 1, 2, . . . n (1.1)

where
X i is the value of the ith of good or service;
L i represents the quantity and ability of available labor;
K i represents capital, machinery, transportation devices, and production infras-
tructure;
M i represents available materials and other resources for production.
Since all of L i , K i, and M i are manufacturing resources, the function f (·) in Eq. (1.1)
represents a value-adding transformation in manufacturing. The amount of added
value relates to the level of existing manufacturing technologies and know-hows
used to convert inputs to outputs with better productivity. Therefore, other than a
weighted average of the rate of population growth and the capital-accumulating rate,
human civilization depends greatly on the growth rate of manufacturing technologies
(Ruby 2003).

1.2 Drivers for Manufacturing Technologies

Manufacturing is to transfer raw materials into final products for users through a
series of value-added and non-value-added processes. The transformation requires
capital, people, machinery, tooling, processes, and systems to perform manufacturing
1.2 Drivers for Manufacturing Technologies 7

businesses. Manufacturing contributes greatly to the value of national economies

in the forms of assets, wealth, and strategic capabilities for defense and security,
arts, literature, or other culture artifacts (NSTC 2008). Manufacturing has shown its
importance to the national policies at the aspects of (1) providing employments, (2)
attracting foreign direct investment (FDI), and (3) improving productivity and other
economic measures (López-Gómez et al. 2017).
Manufacturing technology provides the tools to make products. Manufacturing
technology covers machine tools (such as lathes and drill benches), material removals
(such as turning and drilling) and material forming (such as pressing, stamping, and
shearing), additive processing (such as 3D printing and laser sintering), workholding
(such as clamps, blocks, and chucks)), tooling (such as drills, taps, and grinding
wheels), materials handling (such as conveyors and automated guided vehicles),
automated systems (such as flexible manufacturing systems and transfer machines),
and software tools (such as computer-aided design, computer numerical controls,
and computer-integrated manufacturing).
Figure 1.6 shows the main drivers to advance manufacturing technologies. These
technological drivers are classified from the perspective of the cycle of continuous
improvement (CI) of technological development. On the one hand, manufacturing
technologies are pushed by the needs of the human society for further development
and civilization. On the other hand, manufacturing technologies are pulled by the
users’ demands to make more and better products. In addition, for the transformation
from raw materials to final products, manufacturing technologies are evolved by the
integration of other relevant technologies in manufacturing businesses. Therefore,
manufacturing technologies are gradually and iteratively advanced through CI cycles.

Pushing drivers Pulling drivers

from human Manufacturing
Technologies by consumers’
society market

Manufacturing
business
environment

Reciprocating drivers by enabling

technologies in materials, machinery,
processes, design, tooling, assembling,
controlling, decision-making supports,
product lifecycle supports, and
information technologies….

Fig. 1.6 Drivers of manufacturing technologies

8 1 Human Civilization, Products, and Manufacturing

1.2.1 Pushing Driver—Human Society

Manufacturing technologies are developed to meet the needs of the human society.
Therefore, the evolution of human society is the primary driver to advance manu-
facturing technologies. Numerous researchers have discussed the evolving trends of
the human society, and the identified main trends include (1) the globalization of the
manufacturing business environment, which are featured by offshoring, outsourcing,
and the growth of manufacturing capabilities, (2) the awareness of sustainability,
which can be measured by carbon footprints, the efficiency of manufacturing oper-
ations, and quality of life and consumption, (3) the change of demographics such as
the growing size of middle-class population and ageing workforce, (4) the urban-
ization of mobility, housing, infrastructure, and factories, (5) the threats to global
stability from natural disasters and terrorism threats, (6) shortening of product life-
cycles due to the advances of materials, processes, and technologies, (7) the change
of customers’ expectations on products such as personalization and fast rate of tech-
nology adoption. Accordingly, Fig. 1.7 shows six drivers to advance manufacturing
technologies (López-Gómez et al. 2017). These drivers include the ageing of work-
force in developing countries, change of manufacturing skillsets, personalized prod-
ucts, increasing demands for products in cities, growing interest in technologies
strategies, and reshoring manufacturing activities globally.

Ageing of workforce in
developed countries: population
ageing is challenging the
lifestyles and consumption
patterns as well as shrink the size
of available workforce for
Reshoring manufacturing manufacturing sector Changing manufacturing
activities: reshoring has gained a skills: The evolution of
great deal of attention since the manufacturing technologies is
developed nations strategically transforming the requirements of
expand the domestic skill sets; together with ageing
manufacturing base and foster of population, they lead to the
high-wage job creation, shortage of qualified labour in
innovation and exports. manufacturing.
Development of
Manufacturing
Technologies
Growing interest in Personalized needs:
technology strategies: adaptability to meet
Developing and developed personalized needs is
counties are embarking on the becoming critical to market
designs of detailed industrial and value capture for
and technology strategies to companies around the world
maximize the economic Increasing demand for goods
benefits of manufacturing and services in cities: urban
infrastructure shows a growing
demand of urban mobility,
energy, housing and
telecommunication solutions,
which is a prominent driver of
stimulating manufacturing

Fig. 1.7 Pushing driver of manufacturing technologies from human society

1.2 Drivers for Manufacturing Technologies 9

1.2.2 Pulling Driver—Consumer Markets

The added value of manufacturing operations is redeemed when products are deliv-
ered to meet customers’ needs, and the customers’ needs determine the functions,
types, and volumes of products that manufacturers should make for. From this aspect,
manufacturing technologies must be continuously advanced; since the demanded
products become more and more complicated. Figure 1.8 shows that the customers’
needs are continuously changing in terms of spectrum, versatility, variety, life span,
volume, lead time, integration, suppliers, and life supports of products. The average
volume of a specific product is gradually reduced to the extreme case of size-one,
the average life span and lead time is continuously reduced. The measures of other
aspects are increased.
The spectrum of products has been greatly expanded over time. At small scale,
nano-machines were designed to operate in human in vivo environment for diagnosis,
inspection, and medical treatment as shown in Fig. 1.9a (Amato 2012); at large
scale, massive machines were expected to transport people to deep space as shown
in Fig. 1.9b (Wikipedia 2021a). Manufacturing technologies must be advanced to
explore new materials, and to design, fabricate, and assemble extreme small or large
and complex products.
Taking an example of the iPhone products in Fig. 1.10, generation after gener-
ation, products become more and more versatile since new features will be added
and the capacities will be expanded in each successive generation to make the prod-
ucts better. In each generation of products, customers have personalized choices of
available features, and more product features imply more variations of customized
products. For example, Fig. 1.10 shows the adoption of new features in iPhone
series at the aspects of screen technology, 3D touch, networking support, cameras,
video records, live photos, identification, charging, and water-resistance. Moreover,
product varieties have been increased gradually due to an increase of product features.
An iPhone camera module has over 200 separate parts; the Boeing with 5400
supplier factories including secondary supplies; more than 750 million components

Spectrum

Volume

Life supports

Fig. 1.8 Pulling driver of manufacturing technologies from consumer market

10 1 Human Civilization, Products, and Manufacturing

(a) Nano-machines in medical applications (b) Humans mission to Mars

Fig. 1.9 Examples in expanded product spectrum

IPhone 1 IPhone 2 IPhone 3 IPhone 4 IPhone 4S IPhone 5 IPhone 5C IPhone 5S IPhone 6

2007 2008 2009 2010 2011 2012 2013 2013 2014

IPhone 6Plus IPhone 6S IPhone 6SPlus IPhone 6SE IPhone 7 IPhone 7Plus IPhone 8 IPhone 8Plus IPhone X
2014 2015 2015 2015 2016 2016 2017 2017 2018

IPhone XS IPhone XR IPhone XR Max IPhone 11 IPhone 11Pro IPhone 11 MaxPro

2018 2018 2018 2019 2019 2019

Fig. 1.10 Example of versatility and variety versus generation of products

1.2 Drivers for Manufacturing Technologies 11

Fig. 1.11 Examples of complex products

and assemblies were acquired by Boeing in 2012, around 500,000 people are hired
by these suppliers. The program scales of the control systems for a Lockhead F-
22 raptor, Boeing 787 Dreamliner, Airbus A380 and 2015 Ford F-150 were 2, 7,
100, and 150 million lines of code, respectively. Figure 1.11 shows some product
examples, whole complexities can be defined mainly based on the number of parts
and components in products (Digital Engineering 2016).
The complexity of products and processes determines the complexity of manu-
facturing technologies in making these products, and the five main factors to affect
the complexity of manufacturing technologies are (1) functions of products, (2)
number of complexities of manufacturing processes, (3) correlations of products and
processes, (4) the interactions in enterprise information systems, and (5) the number
of involved national and international standards, regulations, and specifications. As
shown in Fig. 1.12, the product complexities by the five aforementioned factors have
been continuously increased from very simple to very complex years over years.
Taking an example of computers, based on today’s standards, early computers were
so simple in terms of its capabilities of computing and storage. A new-generation
central processing unit (CPU) was developed at an average of every 18 months; it has
the double complexity for quadruple capabilities at roughly half price in compar-
ison to previous CPU. Other products have similar changing trends of complexities
and lead times: products become old-fashioned, their attractions to customers are
gradually vanished, and the market’s demands decline in less and less time (Prasad
1997).
Correspondingly, average life span of products becomes shorter and shorter due to
the early availability of new products with more functions and better performances.
Taking the change of computer life spans in Fig. 1.13 as an example, according
to the study at Arizona State University (ASU), the average life spans of personal
12 1 Human Civilization, Products, and Manufacturing

> 5 years

4-5 years
Time to Decreased
market lead time
3-4 years
Time to market

Goods and
services
Very complex
Product complexity
Complex

Product Increased
complexity Simple complexity
Very
simple
1980 1985 1990 1995 2000

Fig. 1.12 Shortened lead time

Fig. 1.13 Example of 12

shortened product
lifespan—PCs at Arizona 10
State University (ASU)
8
Lifespan (Years)

(Babbitt et al. 2009)

6

4 Mean
LowerDeviaƟon
2
UpperDeviaƟon
0

Calendar year

computers (PC) had been reduced greatly from 12 years in 1985 to <3 years in 2000
(Babbitt et al. 2009); the increase of computer ownership at a rapid rate was the main
cause of such a change.

1.2.3 Reciprocating Driver—Relevant Technology

Development

On the one hand, manufacturing technologies are the results of applying funda-
mental science and technology in solving problems in designing, making, and using
1.2 Drivers for Manufacturing Technologies 13

Manufacturing technologies

Mass production

Sensors, automation

Fundamental science
and technology

Fig. 1.14 Reciprocating drivers of manufacturing technologies by relevant technologies

products; on the other hand, new challenges of designing and making better prod-
ucts stimulate the further advancement of fundamental science and technology.
From this perspective, technology development is a reciprocating driver to manufac-
turing technologies (Jiang et al. 2014). The impact on relevant technology develop-
ment on manufacturing technologies has been clearly evidenced in the evolution of
manufacturing technologies.
As shown in Fig. 1.14, every major shift of the manufacturing paradigm has
been driven by the new development of relevant technologies (Cheng et al. 2000).
For example, sensors and automation made the closed-loop automation in mass
production possible. The emerging information technologies (IT) such as Internet of
Things (IoT), cyber-physical system (CPS), human cyber-physical system (HCPS),
and blockchain technology have laid the foundation to practice digital manufacturing
paradigm (Bi et al. 2020). On the other hand, the demands for green, miniaturized,
highly valued products have stimulated the studies in many emerging fields such
as nanoscience, environmental science, big data analytics, machine learning, and
cybersecurity.

1.3 Formulation of Engineering Design Problems

Figures 1.15 and 1.16 show that engineering designs are involved in all stages of
product lifecycles (Wikipedia 2021b) and all aspects of corresponding manufacturing
processes (Wang 2006), respectively.
A manufacturing system aims to make products, and numerous decision-making
activities are involved in at different stages of product lifecycles. Therefore, any
14 1 Human Civilization, Products, and Manufacturing

Design parts, Design manufacturing

components and processes and tools for
assemblies fabrications

Make parts, and assembly

Acquire raw Engineering designs, parts and components into
materials and planning and products
supplies controlling, and
other decision-
making supports
Design for recycling, Manage logistics,
reusing, redesigning, inventories, and
and disposals supply chains

Design testing, inspection,

quality assurances, and
maintenance planes

Fig. 1.15 Engineering designs in product lifecycle

Product data Quality

management control
Assembly Identify
processes customers’
needs

Manufacturing Engineering Acquisition of

processes raw materials
Design
Storages and Marketing
transportation and sales
Materials Cost
handling Marketing analysis
and sales

Fig. 1.16 Engineering designs in all aspects of manufacturing operations

decision-making activity in a manufacturing system can be formulated as an engi-

neering problem, which consists of inputs (I), outputs (O), purposes (P), variables
(V), and constraints (C). As shown in Fig. 1.17, to formulate an engineering problem,
the scope of the problem should be defined to identify the set of aforementioned
1.3 Formulation of Engineering Design Problems 15

Environment (E) Constraints (C)

Inputs (I) Outputs (O)

Design
Analysis

Variables (V) Design

Synthesis Goals (G)

Reverse engineering
Experiments
Human Experience

Computer Aided Tools

Trial and Error

Mockup

Prototyping

Fig. 1.17 Formulating an engineering problem

elements. Further, the relations of the system elements must be modeled and analyzed
so that different solutions can be compared to each other to find an optimal solution
to the formulated problem. With the trend of ever-increasing complexity, scale, and
dynamics, various engineering approaches, such as computer-aided technique, proto-
typing, and reverse engineering are needed to assist humans in performing design
analysis and synthesis in seeking engineering solutions.
Once an engineering problem is formulated, a general procedure shown in
Fig. 1.18 can be followed to seek an optimal solution to the defined problem. The
procedure consists of five design phases. At phase II, a design space is defined to
include all of the possible solutions to be evaluated, and the relations of system
elements with outputs (O) and constraints (C) are modeled. At phase III, possible
solutions are continuously analyzed and evaluated against the expected design
criteria; they are compared each other to identify an optimal solution through design
synthesis. At phase IV, the detailed design is performed for the optimized solution,
and the design is verified and validated to ensure no violence to the given constraints;
if there is, then an iterative process is needed from phase II to IV until a valid solu-
tion is identified. At phase V, the physical solution is implemented as a practical
solution to the stated engineering problem. It is clear that the information about an
engineering solution is accumulated, while the design constraints must be satisfied.
16 1 Human Civilization, Products, and Manufacturing

Formulate problems to identify

Identify new inputs (I), variables (V), outputs
customers’ needs (O), purposes (P), environmental
factors (E), constraints (C), and Phase I
available design tools.

(1) Define design space with feasible

alternatives Perform conceptual and
(2) Establish relational models, quantify preliminary designs Phase II
the outputs and purpose

(1) Analyze design alternatives based on the

required outputs and constraints Perform design analysis
(2) Evaluate and compare design and synthesis Phase III
alternatives according to the purposes to
optimize solutions

(1) Implement detailed design

(2) Verify and validate design solution Design details, verify, valid
(3) Conduct iterative process when it is the solution, and generate Phase IV
needed. implementation plan

GOAL: FIRST-TIME-RIGHT RESULT

Physical
implementation for Document the solution, and
implement the solution in the Phase V
customers’
satisfaction physical world

Fig. 1.18 General procedure of solving an engineering problem

1.4 Computers in Engineering Design

Regardless of the complexity level of a product or a manufacturing system, the

engineering process follows similar process as discussed in Fig. 1.18. However,
advanced techniques and tools are needed when the scale and complexity of an engi-
neering problem is beyond of the capabilities of manual decision-makings. From
this perspective, more and more human’s roles are being replaced or assisted by
computer-aided technologies such as computer-aided design (CAD), computer-aided
engineering (CAE), and computer-aided manufacturing (CAM). Here, the charac-
teristics of engineering design problems are discussed to understand the needs of
using computer-aided techniques in engineering designs (Bi 2018). An engineering
problem can be one of six typical types as shown in Table 1.2.
The typical engineering problems in Table 1.2 can be classified based on the levels
of creativity of the corresponding engineering solutions. The level of creativity relies
1.4 Computers in Engineering Design 17

Table 1.2 Classification of engineering designs

Type Description Example

Routine design To find a design solution by following the Follow ASTM fastener standards
standards and guidance that specifies the steps to select bolts, screws, or nuts for a
and calculations. fastener.
Redesign To update an existing design when some Modify a robot program when the
functional requirements have been changed in working points on a path have been
its application. changed in a new task.
Selection design To select one of the pre-existing solutions for Select an electronic device such as
a given application. a computer at an office.
Parametric design To optimize a set of discrete, continuous, or Design a four-bar mechanism to
mixed design variables in a given conceptual operate a door subjected to the
structure. specified weight, operating range,
space, and cost.
Integrated design To build a product or system using existing Design a workcell for the
components and modules. identified machines and products
in cellular manufacturing (CE).
Original design To conceptualize a design from scratch to Propose a new solution to
meet a set of newly identified functional guarantee the security, trustiness,
requirements. and responsiveness of data sharing
in a large-scale distributed system.

to the characteristics of a ‘solution space’ and ‘design variables’ in a formulated engi-

neering problem, and Table 1.3 shows a classification of engineering problems based
on the levels of creativity. An engineering design can be a ‘routine’, ‘innovative’, or
‘creative’ design.
Both humans and computers are important in seeking the solutions of engineering
problems, but they are competent to different levels and scopes of decision-making
tasks. The differences of humans and computers have been widely discussed, and
Table 1.4 gives a comparison of their strengths and weaknesses in engineering
designs.
As shown in Fig. 1.19, in engineering practice, both humans and computers
play their roles in accomplishing certain activities. Since computers and humans
are good at different tasks as shown in Table 1.4, the strengths of computers and
humans can be synergized by using computer-aided techniques in solving various
engineering problems. Especially, computer-aided techniques can be utilized to auto-
mate design processes and improve effectiveness of engineering designs. Taking an

Table 1.3 Difference of routine, innovative, and creative designs

Level of creativity Routine design Innovative design Creative design
Solution structure Known Known Unknown
Space
Search procedure Known Unknown Unknown
Design Types Fixed Fixed Changed
Variables
Ranges Fixed Changed Changed
18 1 Human Civilization, Products, and Manufacturing

Table 1.4 Differences of humans and computers (Bi 2018)

Human Designers Computers
Identifying design needs, Fast speed, reliable, endurance, and
Strengths
brainstorming to think solutions ‘out of box’, consistent,
engineering intuition and big knowledge Capable of exploring a large number of
base, options,
selecting design variations, Carry out long, complex, and laborious
the flexibility to deal with changes calculations,
qualitative reasoning, Store and efficiently search large
psychologically, human decision is more databases, and
trusted than artificial intelligence, Provide information on design
predict trends, patterns, or anomalies, and methodologies, heuristic data, and
learn from experience stored expertise.
easily tired and bored difficult to synthesize new rules
Weaknesses
cannot micro-manage limited knowledge base
biased and inconsistent no common sense
prone to make errors
not good at quantified reasoning
incapable of utilizing the data presented in
awkward manner

Engineering Design

SelecƟon Parametric Integrated

RouƟne Design Redesign Original Design
Design Design Design
To find a design To update an To select one of To optimize a set To build a product To conceptualize a
solution by existing design the pre-existing of discrete, or system using design from scratch
following the when some solutions for a continuous or existing to meet a set of
standards and functional given application. mixed design components and newly identified
guidance that requirements have variables in a modules. functional
specifies the steps been changed in its given conceptual requirements.
and calculations. application. structure.

Human designers for creaƟve design

Computer aided tools for rouƟne design

Level of Design Effectiveness

Level of Design Automation

Fig. 1.19 Human designers and computers in engineering designs (Bi 2018)
1.4 Computers in Engineering Design 19

example, Bi et al. (2006, 2010) automated the process of the configuration design for
a modular robotic system, which was able to evaluate 104 –105 robotic configurations
automatically in finding the best configuration for the specified task.

1.5 Computers in Manufacturing

Due to the complexity of modern manufacturing systems, more and more manufac-
turing businesses rely on computers, and computer-aided techniques were widely
applied to expand the scope of automation in manufacturing systems. The capa-
bility, the productivity, and the efficiency of a manufacturing system can be greatly
improved by maximizing the degree of automation.
The dependence of the performance of a manufacturing system on the automa-
tion was explored by many researchers. For example, Williams (2000) analyzed the
importance of manufacturing operations from the perspective of human-machine
relations. As shown in Fig. 1.20, manufacturing activities are classified into two
groups, i.e., the activities involved in the information flow, and the activities in the
material flow. In the material flow, computers are used to mechanize and automate
manufacturing operations; intelligent machines such as computer numerical control
(CNC) machines and automated guided vehicles (AGVs) are widely used to replace
human operations. In the information flow, computers are used to support decision-
making activities; computer-aided techniques assist humans in making decisions at
all phases of manufacturing systems’ lifecycles from design, construction, operation,
to the dissolve phase. While humans are still important in sustaining the agility and
flexibility of a manufacturing system, it becomes more and more critical that humans
and computers collaborate harmoniously to expand manufacturing businesses in a
highly competitive environment.
A broad adoption of computer-aided techniques helps to maximize the scope of
the automation. In particular, computer-aided techniques are taking over more and
more human’s efforts in decision-making supports. As a matter of fact, the advance-
ment of manufacturing technologies was measured by the capabilities of computer-
aided techniques in tackling with the growing scale, complexity, and responsive-
ness of manufacturing systems. Figure 1.21 shows the evolution of computer-aided
techniques over time (Cheng and Bateman 2008; Bi and Cochran 2014). When the
computer-aided tools such as quality control (QC), total quality management (TQM),
supply chain management (SCM), enterprise requirements planning (ERP-I), enter-
prise resources planning (ERP-II), product lifecycle management (PLM), software
as a service (SaaS), and platform as a service (PaaS), and Infrastructure as a Service
(IaaS) became available, manufacturing processes were able to be automated by using
CNCs, flexible manufacturing systems (FMSs), computer-integrated manufacturing
(CIM), distributed manufacturing (DM), and predictive manufacturing (PM). On
the other hand, computer-aided technologies were driven by the emerging needs of
manufacturing systems to deal with the increasing volume, variety, velocity of the
data in the globalized business environment.
20 1 Human Civilization, Products, and Manufacturing

Identification of
business opportunity

Functional Functional modules

Design modules in cyber in physical
phase manufacturing manufacturing
environment System environment
configuration

Human machine
interface (HMI), and
Construction human machine
phase cooperation (HMC)

Human-robot
collaboration (HRI)

Cyber-physical
systems (CPS)
Operation
phase
Human cyber-physical
system (HCPS)

Collaborative systems

Sensor-guided
Dissolving operations
phase
Human-guided system

Maximize
Cyber system automation Physical system

Selected high-level functional module Available high-level functional module

Selected middle-level functional modules Available middle-level functional modules
Selected low-level functional module Available low-level functional module

Relation of functional modules Human factors

Fig. 1.20 Computers for the automation in modern manufacturing system

1.6 Emerging Digital Technologies

Digital manufacturing technologies refer to some computer-aided tools that are used
in product designs, manufacturing processes, services, supply chain management,
and system designs and operations. With the digital manufacturing, system elements
1.6 Emerging Digital Technologies 21

Enterprise System for Manufacturing

Quality Control TQM ERP-I ERP-II SCM PLM SooS/PaaS/IaaS

Predictive
Distributed Manufacturing
Manufacturing
Computer
Integrated Manufacturing
Big data
Flexible
Manufacturing Systems Analytics
NC/CNC
Workstation
Enterprise
Project Management

Business
Stream data Intelligence
Information Technology (IT)
PunchCard Mainframe Servers Internet Cloud

Before 1950’s 1960s 1970s 1980s 1990s 2000s 2010s (Timeline)

Fig. 1.21 Evolution of computer-aided technologies in manufacturing (Cheng and Bateman 2008;
Bi and Cochran 2014)

and processes are networked and connected to fully utilize real-time data in opti-
mizing a holistic manufacturing system. Digital technologies are widely adopted
to enhance the competitiveness and in expanding the landscapes of manufacturing
enterprises. In recent years, digital technologies have been progressing exponentially,
especially by adopting new technologies at the following aspects.

1.6.1 Internet of Things (IoT)

Internet of things (IoT) is a network of smart things. A smart thing can be anything
with sensing, data processing, or decision-making capabilities. IoT is the extension
of the Internet with the connections to physical devices via transmission control
protocol/Internet protocol (TCP/IP). As shown in Fig. 1.22, IoT connects the phys-
ical twin and digital twin seamlessly in digital manufacturing and IoT makes any
interactions among the things possible. As shown in Fig. 1.23, an interaction possibly
occurs to a cyber and physical system (CPS), machine-to-machine, human to machine,
human to human, machine to human, machine to infrastructure, and machine to
environment.
22 1 Human Civilization, Products, and Manufacturing

Velocity
Data Management

Variety
Storage
Infrastructure
Data CollecƟon Query
SQL Data Utilization
Internet NoSQL Database
SaaS, PaaS, IaaS
Sensors Hadroop
Sales Prediction
Log Files MapReduce
Virtual Manufacturing
Manual Recording Data Analytics
TQM, SCM, PLM
ConvenƟonal RDBMS Security Assurances
Predictive Manufacturing
Industrial Networks ….
Lean Production
Tracking Systems
Project Management
RFIDs, WSNs
ERP-I and ERP-II
Social MIedias
Enterprise Alliance
Surveys
….……
……

(IoT)

User User User User

User User

Fig. 1.22 Internet of Things (IoT) in enterprise systems (Bi and Cochran 2014)

Cyber
Machine to Human to Human to Machine to Machine to Machine to
physical
Machine Machine Human Human Infrastructure Environment
systems

Internet of
Things

Fig. 1.23 IoT makes any interaction possible

1.6.2 Big Data Analytics (BDA)

Enterprises are forced to make more product variants in mass customization with an
enhanced level of customers’ satisfactions (Pasche 2008); their enterprise systems
should be capable of dealing with the increasing volume, velocity, and variety (3 V)
of data due to system uncertainties and complexity shown in Fig. 1.24.
Nowadays, enterprises are deluged in the Internet with the data scale from
terabytes (240 bytes), petabytes (250 bytes) to exabytes (260 bytes), and traditional big
analysis methods became struggling in processing heterogeneous, unstructured, and
1.6 Emerging Digital Technologies 23

Variety of Varity of Variety of

market design available
needs solutions materials

Variety of Variety of
manufacturing Resources Variety of
and assembly and tools supplies
processes

Fig. 1.24 Uncertainties and complexity in manufacturing

dynamic data. Big data analytics (BDA) aims to analyze a big data and explore infor-
mation, knowledge, and wisdom at the unprecedented scale (Cloud Security Alliance
2013). As shown in Fig. 1.25, BDA is used in digital manufacturing to support real-
time, short-term, and long-term decision-making activities over the product lifecy-
cles. BDA is expected to achieve the goals of computerization, connectivity, visibility,
transparency, predictivity, and adaptability based on the capabilities of BDA tools
(KPMG 2018; Bi and Wang 2020).

Fig. 1.25 The goals of BDA in digital manufacturing (KPMG 2018; Bi and Wang 2020)
24 1 Human Civilization, Products, and Manufacturing

Virtual World

Cyber System

Pre-
Prediction Optimization
processing
B-Interaction F-Interaction
Data
Actuation
acquisition
Physical System

Process 1 Process 2 … Process n

Physical World

Fig. 1.26 Main components in CPS

1.6.3 Cyber-Physical Systems (CPS)

A cyber-physical system differs from stand-alone embedded systems, sensors or other

conventional mechatronic systems in sense that the cyber electric and electronic
system is fully coupled with the physical system to be controlled. The studies of
CPS were prioritized to advance the manufacturing technologies for Industry 4.0.
Existing works on CPS were at their infant stages and limited to the exploration of
basic concepts, system architecture, technologies, and challenges. Figure 1.26 shows
the main components of a cyber-physical system that include a number of physical
processes in the physical system, a set of functional modules for data acquisition,
pre-processing, prediction, and optimization in the cyber-system, and the functional
modules for data acquisition and actuation as the interactions (Schmidt and Ahlund
2018). CPS adds the capabilities of computation, communication, and controls (3C)
to a physical system, which allows a cyber-system to interact with physical processes
directly. A CPS is featured by real-time sensing and dynamic controls to respond to
the changes and uncertainties in applications.

1.6.4 Cloud Computing (CC)

An enterprise often faces some dilemmas in determining computing resources for

its information system. On the one hand, the enterprise must optimize its system
resources for core manufacturing businesses in the physical manufacturing system.
On the other hand, products, manufacturing processes, and business environment
become more and more complicated and dynamic, and high computing capabilities
are expected to process big data for decision-making units to respond to changes
1.6 Emerging Digital Technologies 25

Enterprises Service Models Deployment Models

Cloud Core Public Cloud

SaaS S P M

FaaS
Businesses in Community Cloud
information PaaS P
flow S M
CaaS
Hybrid Cloud
IaaS
S P M

Businesses in
MaaS Private Cloud
materials flow HaaS I M

S - Service P -Platform I - Infrastructure M - Manufacturing F - Function C - Container H - Hardware

Fig. 1.27 Main components in CPS (Pedone and Mezgar 2018)

and uncertainties promptly. As shown in Fig. 1.27, cloud computing (CC) offers a
technical solution for enterprises to utilize outsourcing computing resources over
the Internet. The manufacturing businesses in both of the information and materials
flows can be defined as services; and services will be provided in different service
models such as infrastructure as a service (IaaS), function as service (FaaS), and
manufacturing as service (MaaS). Services are available in a public cloud, community
cloud, hybrid cloud, or private cloud over the Internet (Pedone and Mezgar 2018). For
small and medium-sized enterprises (SMEs), CC allows to minimize the information
infrastructure and reduce the cost of the enterprise system since CC is dynamic and
flexible.

1.6.5 Blockchain Technology (BCT)

Manufacturing systems used to be closed, and system controls are centralized. With
the emerging needs of (1) collecting and sharing data from smart things in distributed
environment, (2) collaborating with business partners across system boundaries,
and (3) utilizing external resources to enhance business capabilities, the controls of
modern manufacturing systems are distributed and decentralized. Service-oriented
architecture (SOA) allows enterprises to use distributed and external resources to
expand manufacturing capabilities. However, some challenges of SOA have not been
addressed satisfactorily (Viriyasitavar et al 2019a, b).
As shown in Fig. 1.28, blockchain technology (BCT) was proposed to assure the
trustiness, security, and privacy of the services over the Internet. The performance of
a decentralized system can be measured by some performance key indicators (PKIs)
26 1 Human Civilization, Products, and Manufacturing

Performance Key Indicators (KPI)

4 3
Certificate
Authority (CAs)

Other systems Users Verifying certificate

of validators

Service Cloud
Repository services

Services Oriented Applications (SOAs)

Application Program Interfaces (APIs)

BT-enabled
smart things
IoT Services Cyber-Physical Quality of
1 Systems Services (QoS)

Fig. 1.28 Blockchain technologies in service-oriented architecture (Viriyasitavar et al. 2019a, b)

including scalability, security, openness, and flexibility. Three-layer system architec-

ture shows an integration of blockchain technology (BCT) with IoT, SoA, PKIs, and
other available services over the Internet. BCT-based applications at the first level
ensure the validity of data from smart things such as sensors, CPSs, and quality of
services (QoS). SoA at the second layer ensures that the manufacturing processes are
executed and delivered in the form of services, and SoA deals with the consistency of
data and interactions. PKIs at the third layer establishes the trustiness of validators.
The blockchain validators should be clarified and endorsed by Certificate Authority
(CA) for each system.

1.6.6 Rapid Prototyping (RP)

An engineering design process is generally iterative in which errors, defects, and

omissions are gradually eliminated to find a right solution. A traditional engi-
neering process involves in the validation where a physical product or system is
built and tested against expected functionalities. This becomes cost-forbidden when
1.6 Emerging Digital Technologies 27

(a) Traditional material removal manufacturing

(b) Rapid prototyping (additive manufacturing) such as 3D printing

Fig. 1.29 Examples of material removal processes and rapid prototyping

the product or system is extremely complicated or the number of iterations is large.

Rapid prototyping (RP) is a cost-effective technique for the fast fabrications of phys-
ical products, models, assemblies, or systems. RP is mostly implemented by an addi-
tive manufacturing process such as 3D printing. Rapid prototyping is referred as solid
freeform fabrication (SFF), additive manufacturing (AM), or direct manufacturing
(DM). In contrast to traditional manufacturing processes.
Figure 1.29 shows some example products from material removal processes and
rapid prototyping. RP has exposed the following advantages: (1) there is no needs
for tooling, forming, or fixturing, and RP can be operated more safely since no
machining force is involved. (2) It uses layered fabrication, which is applicable to
make products with any complex geometries. (3) RP is a near net shape process
which has little waste. 4) RP can shorten the development time significantly, since
no custom tools or programming is needed to make special products.

1.7 Digital Skills in Modern Manufacturing

With ever-increasing complexity of products and manufacturing processes, more

and more manufacturing operations and decision-making support becomes beyond
the reach of humans. Machines and computers are replacing humans not only for
repetitive and boring tasks, but also for complicated decision-making supports at
different levels and domains of manufacturing systems. On the other hands, humans
and machines have to collaborate with each other closely to tackle with complex tasks
(Bi et al. 2021), where operators, technicians and engineers are required to enhance
their digital skills to work on machines and make decisions in various manufacturing
businesses.
28 1 Human Civilization, Products, and Manufacturing

Muro (2017) discussed the trends of the requirements of digital skills in advanced
manufacturing at the occupations from laborers and machinists to aerospace engi-
neers. As shown in Fig. 1.30, the requirements for digital skills were increased from
2002 to 2016 with no exception. Taking the examples of production workers and
tool and die makers, they used to heavily rely on their hand-on skills to fulfill their
job responsibilities; however, more and more advanced machines demand higher and
higher digital skills from human workers to operate, interact, and program machines.
The average digital skill score of all major occupations was increased to 39 in 2016
from 24 in 2002.
In recent years, information technologies (IT) have been advanced rapidly;
however, the adoption of new ITs has been lagged more or less. Figure 1.31 shows the
relation of annual productivity changes and mean digital scores in different sectors
(Muro 2017). The digital score of the occupations in advanced manufacturing was
39 in 2016, and the digital technology adoption in advanced manufacturing showed

Fig. 1.30 Digital skill ratings for occupations in advanced manufacturing (Muro 2017)

Fig. 1.31 Annualized productivity growth by sector versus digital score (2010–2016) (Muro 2017)
1.7 Digital Skills in Modern Manufacturing 29

limited impact on the improvement of annual productivity change. The digital tech-
nologies have revolutionized some industry sectors especially for oil and gas extrac-
tion, since the new technologies such as big data analytics enhanced the exploitation
capabilities to find new natural resources. It was surprising to note that the educa-
tional services were less benefited from digital technologies. Conventional designs
of engineering curricula must be radically improved to fill the gap of the engineering
education and emerging demands of digital skills in manufacturing sector (Bi and
Wang 2020).

1.8 Organization of Book

Solving an engineering problem is generally an iterative process in the sense that

the information of the final solution is gradually accumulated; it implies that design
constraints in the formulated problem cannot be fully satisfied in the solving process
until all the details of implementation are provided in the solution. This poses the chal-
lenge of using computer-aided technologies to practice the first-time right from the
digital twin to physical twin due to (1) the uncertainties and complexity of products
and systems and (2) the needs of verification and validation in the virtual environment.
This book emphasizes on engineers’ knowledge and skills of using digital manu-
facturing tools to solve various engineering problems at different domains and levels
of manufacturing systems. As shown in Fig. 1.32, both products and system resources
have lifecycles that affect manufacturing systems. A manufacturing system involves a
set of value-added and non-value-added manufacturing operations along the product
lifecycle, and decision-making supports are needed to use various system resources
in fulfilling these operations. In digital manufacturing, the decisions on system opera-
tions are made by functional units in the digital twin, and engineers should compete to
utilize computer-aided technologies to make right decisions at the first time. Consid-
ering the time limit of a typical engineering course, this book selects the following
four main subjects that are shown in Fig. 1.32.
Chapter 2 introduces computer-aided design (CAD). In CAD, engineers use
computers to design, analyze, model, and evaluate products, processes, and systems.
Computer-aided theory and methodologies are introduced to represent products,
processes, and systems, and parametric modeling and knowledge-based engineering
(KBE) are emphasized for knowledge reuse and exploration. Solid modeling tech-
niques are introduced to create virtual models as the representations of geometric
shapes, features, design intents, and topology and structures of products. Computer-
aided kinematic and motion analyses are discussed for machine designs.
Chapter 3 introduces computed-aided engineering (CAE). In CAE, engineers use
computers to model and analyze the responses of the product, process, or system that
is subjected to the specified loads and constraints in applications. CAE is essential
to eliminate the hidden mistakes and omissions in the virtual world before the digital
twin is released to the physical world for implementation. The importance of CAE in
manufacturing is discussed, different CAE tools are introduced, and the focus is put
30 1 Human Civilization, Products, and Manufacturing

Fig. 1.32 Digital manufacturing makes the first time right from digital twin to physical twin

on finite element analysis (FEA). The theoretical fundamental of FEA is introduced,

the general procedure of FEA modeling and simulation is presented. Some typical
FEA applications in manufacturing are discussed.
Chapter 4 introduces computer-aided manufacturing (CAM). In CAM, engineers
use a computer to design, model, analyze, optimize, and control manufacturing
processes. CAM is complex since it involves machines, tools, materials, numerous
operating factors, and multiple but conflicting performance criteria such as operating
time, cost, accuracy, flexibility, and adaptability. CAM and relevant concepts are
discussed, and computer-aided technologies are introduced for modeling and simu-
lation of fixtures, mold and dies, and manufacturing processes. The programming
tools are introduced for the controls of machine operations for specified products.
Chapter 5 introduces computer-integrated manufacturing (CIM). In CIM, engi-
neers use computers to organize functional units and manufacturing operations
at system level. A manufacturing system consists of various functional units for
designing, manufacturing, assembling, transporting, managing, marketing, main-
taining, and services at different levels and domains. The enabling technologies,
such as group technology (GT), cellular manufacturing (CM) and discrete event
1.8 Organization of Book 31

dynamic simulation (DEDS) are discussed in detail. Computer-aided evaluation of

manufacturing systems are introduced with the focus on cost and sustainability.
Chapter 6 introduces digital manufacturing (DM). In DM, engineers use
computers to practice the first time right from a digital twin in the virtual world
to a physical twin in the physical world. The functional requirements of DM are
discussed, and new enterprise architecture is presented to deal with the complexity,
uncertainties, and dynamics of system operations. Two digital technologies, reverse
engineering (RE) and additive manufacturing (AM), are introduced as the exam-
ples of emerging digital technologies in digital manufacturing. There is no universal
digital solution to any manufacturing system; therefore, applied studies are more or
less required in developing a holistic digital solution to individual enterprises. Some
research case studies are provided to explore the applications of digital technologies
in various applications.
Design Problems

Problem 1.1. List some quantified measures to reflect the contributions of different
industry sectors to human society, and discuss the importance of
manufacturing to human civilization.
Problem 1.2. What are the main driving forces to develop new manufacturing
technologies? Explain why.
Problem 1.3. What are the main pulling forces to develop new manufacturing
technologies? Explain why.
Problem 1.4. What are the main reciprocating driving forces to advance manu-
facturing technologies? Explain why.
Problem 1.5. Formulate an engineering program for the following engineering
projects, respectively.
(a) A local company produces various end-effectors for robots
and other automated systems. A recent adventure is to develop
new end-effectors for collaborative robots that work in an open
environment with humans. The company meets the challenge
to mount their end-effectors on collaborative robots due to
different design standards at the interfaces. A mount of an
end-effector on a collaborative robot must (1) have the flex-
ibility of orientation shift between 0° and 90°, (2) meet the
safety standards of coving all sharp edges and corners, and (3)
minimize the impact on weight, robotic control, and cost.
(b) A truck assembly plant builds 1/2–1 t trucks. At the entrance
of the assembly line, chassis frames are separated by spacers
in stack. After a chassis frame is transported to the assembly
line, human operators remove spacers from the frames and
transport them to storage containers. The company has the
difficulty in recruiting human operators at such workstations
and the productivity of labors for such operations is low. The
company expects to replace human operators by automation to
32 1 Human Civilization, Products, and Manufacturing

locate four spacers on a chassis frame, to remove and transport

spacers to the designated areas within the production cycle of
one truck per minute.
(c) The precision cooling business unit of a client company has
a production line with brazing furnace for heat treatment
processes of heat sinkers. One human worker uploads parts at
one end and unload parts at the other end of the heat treatment
line. The company has some used robots that have the full
capabilities of loading and unloading heat sinkers. Develop
an automated solution of using used robots to replace human
operators.

Problem 1.6. Have an example of emerging digital technologies of your interest,

debrief the concept and relevant methodologies and tools, and
discuss its role in addressing the challenges in modern manufac-
turing.
Problem 1.7. Have a discussion of the importance of digital skills in engineering
practice, and have a wish list of the digital skills you would like to
master

References

Amato P (2012) Swarm-intelligence strategy for diagnosis of endogenous diseases by nanobots.

https://pdfs.semanticscholar.org/58cf/550b97fe50ca8718c31c98bcdecbb9e06aef.pdf
Babbitt CW, Kahhat R, Williams E, Babbitt G (2009) Evolution of product lifespan and implications
for environmental assessment and management: a case study of personal computers in high
education. Environ Sci Technol 43(13):5106–5112
Bi ZM (2018) Finite element analysis applications: a systematic and practical approach, 1st edn.
Academic Press, ISBN 10 018099526
Bi ZM, Cochran D (2014) Big data analytics with applications. J Manag Anal 1(4):249–265. https://
doi.org/10.1080/23270012.2014.992985
Bi ZM, Wang XQ (2020) Computer aided design and manufacturing, 1st edn. Wiley-ASME Press
Serious. ISBN 10 1119534216
Bi ZM, Lin Y, Zhang WJ (2010) The general architecture of adaptive robotic systems for
manufacturing applications. Robot Comput-Integr Manuf 26(5):461–470
Bi ZM, Gruver WA, Zhang WJ, Lang SYT (2006) Automated modeling of modular robotic
configurations. Robot Auton Syst 54 (12):1015–1025
Bi ZM, Miao ZH, Zhang B, Zhang WJ (2020) The state of the art of testing standards for integrated
robotic systems. Robot Comput Integr Manuf 63(June 2020):101893
Bi ZM, Luo M, Miao Z, Zhang B, Zhang WJ, Wang L (2021) Safety assurance mechanisms
of collaborative robotic systems in manufacturing. Robot Comput Integr Manuf 67 (February
2021):102022
Cheng K, Bateman RJ (2008) e-Manufacturing: characteristics, applications and potentials. Prog
Nat Sci 18(2008):1323–1328
Cheng K, Pan PY, Harrison (2000) The Internet as a tool with application to agile manufacturing: a
web-based engineering approach and its implementation issues, Int J Prod Res 38(12):2743–2759
References 33

Cloud Security Alliance (2013) Big data analytics for security intelligence. https://downloads.clo
udsecurityalliance.org/initiatives/bdwg/Big_Data_Analytics_for_Security_Intelligence.pdf
Digital Engineering (2016) By the numbers: product complexity. http://old.digitaleng.news/de/by-
the-numbers-product-complexity/
Jiang H, Zhao S, Yin K, Yuan Y, Bi ZM (2014) An analogical induction approach to technology
standardization and technology development. Syst Res Behav Sci 31(3):366–382
KPMG (2018) Supply chain big data series part 1. https://advisory.kpmg.us/content/dam/advisory/
en/insights/pdfs/2018/supply-chain-big-data-part-1-shaping-tomorrow.pdf
López-Gómez C, Leal-Ayala D, Palladino M, O’Sullivan E (2017) Emerging trends in global
advanced manufacturing: challenges, opportunities and policy responses. http://capacitydevelop
ment.unido.org/wp-content/uploads/2017/06/emerging_trends_global_manufacturing.pdf
McCloy D (1984) Industry: structures and operation. Technology 1984:286–304
Muro M (2017) Get with the program digitalizing America’s advanced manufacturing sector. www.
nacfam.org/wp-content/uploads/2017/03/Digitalizingusmfg_090717_v3.pptx
National Science and Technology Council (NSTC) (2008) Manufacturing the future. https://nifa.
usda.gov/sites/default/files/resource/nanotech__manufacturing_rd.pdf
Pedone G, Mezgar I (2018) Model similarity evidence and interoperability affinity in cloud ready
industry 4.0 technologies. Comput Indus 100:278–286
Prasad B (1997) Analysis of pricing strategies for new product introduction. Pricing Strat Pract
5(4):132–141
Pasche M (2008) Product complexity reduction—not only a strategy issue. In: 11th QMOD confer-
ence. quality management and organizational development attaining sustainability from organi-
zational excellence to sustainable excellence, 20–22 August 2008, Helsingborg, Sweden. http://
www.ep.liu.se/ecp/033/080/ecp0803380.pdf
Ruby DA (2003) The creation of wealth and economic growth. http://www.digitaleconomist.org/
wth_4020.html
Schmidt M, Ahlund C (2018) Smart buildings as cyber-physical systems: data driven predictive
control strategies for energy efficiency. Renew Sustain Energy Rev 90(2018):742–756
Sullivan N (2020) What is a civilization?—definition & common elements. https://study.com/aca
demy/lesson/what-is-a-civilization-definition-common-elements.html
UNPD (2020a) United nations development programme—human development reports. http://hdr.
undp.org/en/content/human-development-index-hdi
UNDP (2020b) Inequalities in human development in the 21st century briefing note for countries
on the 2019 human development report, USA. http://hdr.undp.org/sites/all/themes/hdr_theme/
country-notes/USA.pdf
Viriyasitavar W, Xu LD, Bi ZM, Sapsomboon A (2019a) New Blockchain-Based Architecture for
Service Interoperations in Internet of Things. IEEE Internet Things J 6(7):739–748
Viriyasitavar W, Xu LD, Bi ZM, Pungpapong V (2019b) Blockchain and Internet of things for
modern business process in digital economy—the state of the art. IEEE Trans Comput Soc Syst
6(6):1420–1733
Wikipedia (2021a) Human mission to Mars. https://en.wikipedia.org/wiki/Human_mission_to_
Mars
Wikipedia (2021b) Product lifecycle. https://en.wikipedia.org/wiki/Product_lifecycle
Williams T (2000) The Purdue enterprise reference architecture and methodology (PERA) institute
for interdisciplinary engineering studies. Purdue University. http://citeseerx.ist.psu.edu/viewdoc/
download?doi=10.1.1.194.6112&rep=rep1&type=pdf
Wang G (2006) Introduction to CAD/CAE. http://www2.ensc.sfu.ca/~gwa5/index_files/25.353/ind
exf_files/1Introduction-06-1.pdf
Woolf SH, Aron L, Dubay L (2015) How are income and wealth linked to health and longevity?
Available online, https://community-wealth.org/sites/clone.community-wealth.org/files/downlo
ads/paper-woolf-et-al.pdf (accessed on April 18, 2021)
Chapter 2
Computer-Aided Design

Abstract In computer-aided design (CAD), engineers use computers to design,

analyze, model, simulate, and evaluate a product, process, or system. In this chapter,
the application of CAD in the creations of virtual models of products or systems
is focused on. CAD was originally developed to represent geometric information
of objects by using computer-drawing software. With the continuous evolution in
several decades, the capabilities of CAD tools have been expanded significantly
beyond computer graphics. In this chapter, various CAD techniques are introduced
to model products, processes, and systems at different levels and aspects, parametric
modeling, and knowledge-based engineering (KBE) are specially introduced, and
basic solid modeling techniques are discussed to model geometrics, features, design
intents, and assembling relations; finally, kinematic and motion analyses are explored
to machine designs.

Keywords Computer-aided designs (CAD) · Computer-aided geometric modeling

(CAGM) · Parametric modeling · Motion analysis

2.1 Introduction

Product or system development involves an iterative design process, from (1) the
identifications of customers’ requirements (CRs), (2) interpreting CRs into design
specifications, defining design space for different design concepts, (3) evaluating
and comparing design concepts, (4) implementing the selected design concepts, and
finally to (5) verifying and validating design solutions (Bi 2018). Modern products or
systems are mostly complex sufficient which are far beyond designers’ capabilities
to fulfill all design tasks manually. Computer-aided technologies, such as Computer-
Aided Design (CAD), Computer-Aided Manufacturing (CAM), Computer-Aided
Process Planning (CAPP), and Computer-Integrated Manufacturing (CIM), can relief
designers from manual tasks for repetitive, routine, error-prone calculations, anal-
yses, and drawings. Advanced computer-aided tools, products, or systems with better
performances can be designed in shortened lead times and at a reduced cost (Lyu
et al. 2017).

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 35

Z. Bi, Practical Guide to Digital Manufacturing,
https://doi.org/10.1007/978-3-030-70304-2_2
36 2 Computer-Aided Design

Computer-aided design (CAD) is the technology of using computers to design,

model, analyze, and evaluate a product or system and document the design process.
CAD is a powerful and effective tool for engineers to conduct research, innova-
tion, new product design, and development. CAD software has been widely used
in the manufacturing sector for the purposes of (1) accelerating and optimizing
product or system designs, (2) facilitating collaboration and cooperation via graphic
communication, (3) developing and maintaining the databases of products, processes,
and systems, (4) reducing the cost and lead time of product development, and (5)
improving the productivity of designers (Wikipedia 2020a). A CAD tool offers the
flexibility for engineers to lay out and create virtual models for products and systems
and ensure them to fulfill the anticipated functions even before their physical replica
is made. CAD has become an essential technology that is used almost in every appli-
cation such as aerospace, automotive, shipbuilding, military, healthcare, industrial
and architectural design, and textile and the fashion world (Jhanji 2018).
CAD is used to automate engineering design processes by which human intel-
ligence, knowledge, innovation, and creativity are utilized in developing products
or systems to meet engineering specifications and customers’ requirements (CRs).
Generally speaking, an engineer design process corresponds to a series of loosely
structured, open-ended activities to (1) define a problem with design needs, criteria,
and constraints, (2) generate, model, and evaluate design options, and (3) deter-
mine, implement, and validate the design solution (Chang 2014). Computer-aided
geometric modeling (CAGM) is a branch of CAD that mainly concerns the repre-
sentation of geometries and shapes of objects. Figure 2.1 shows that CAGM is the
foundation of CAD since the design of a product or system generally starts from

Computer
CAD Applications on Virtual Models from CAGM

Computer Aided Visualization

Geometric Modelling
(CAGM) Technique
Engineering
and Algorithms
Drawing,
Dimensioning and
Tolerances (GT&T)
Mathematic and
Graphic User Graphic
Interfaces (GUI) Representations of
Motion Simulation
Objects

Engineering Analysis
and Parametric Study

Many Other
Applications of
Information of Solids

Fig. 2.1 Role of CAGM in computer-aided systems (CAD) (Bi and Wang 2020)
2.1 Introduction 37

Table 2.1 Seven components of a CAGM system

Component Description
Input (I): Gives the information of design needs such as specifications, topologies, features,
dimensions, and design intents.
Variables (V): Specifies the information of the design space such as (1) the representations of design
intents, parameters, features, and assembly mates, (2) the modelling strategies,
techniques, and topologies and sequences of graphic entities.
Outputs (O): Are the outcomes be generated in the design process such as solid models,
engineering drawings, rendered scenes, animations, and the physical data for
sequential engineering analyses.
Goal (G): Is the set of soft objectives to be optimized in the design process such as customer
satisfactions, productivity, time, cost, flexibility to adopt the changes, and
manufacturability.
Computer The physical entities used to execute the design process such as computing hardware,
Environment (E): software, storages, and the devices for inputs, outputs and interfaces.

Constraints (C): Is the set of hard objectives that a design solution must satisfy such as the
functionalities, compliances, standardization, accessibility, ergonomic needs, and
free of illegality, redundancy, conflicts or interferences.
Resources (R): Include other relevant resources to support and facilitate the design process such as
designers’ expertise, design templates, libraries and utilities, toolboxes, add-ins tools,
parametric modelling tools, macro supports, reverse engineering (RE), and machine
learning.

creating a virtual geometric model of objects. CAGM aims to represent the geome-
tries, shapes, dimensional and spatial relations of objects. Once an object is virtually
modeled, any associated information can be obtained and utilized to support rele-
vant decision-making activities such as design optimization. For examples, (1) a
solid model is often used to visualize and evaluate the conceptual design before
the physical product is actually made; (2) engineering drawings are created from
a virtual product model to facilitate its manufacturing processes, and engineering
drawings are generated directly from solid models with all relevant information
such as geometric dimensioning and tolerances (GD&T); (3) other engineering anal-
yses, such as computer-aided engineering (CAE) and computer-aided manufacturing
(CAM), can be conducted at any design stage when the virtual model is ready.
In CAGM, applied mathematics and computational geometry are integrated
to represent the shapes of objects, associated behaviors, and properties. From a
system perspective, a CAGM consists of the following seven components, which
are described in Table 2.1 and Fig. 2.2.

2.2 Computer-Aided Design (CAD) System

A CAD system deals with data that is intangible, and it is helpful to understand a CAD
system by comparing it with an equivalent system that deals with tangible materials.
From this perspective, a CAD system is equivalent to a manufacturing system in terms
of having a series of transferring processes from inputs to outputs (see Fig. 2.3). In
38 2 Computer-Aided Design

C: the functionalities, compliances,

E: computing hardware,
standardization, accessibility,
software, storages,
ergonomics, and free of illegality,
devices for inputs,
redundancy, conflicts or
outputs and interfaces
interferences, ….

O: solid models, engineering

I: specifications, design drawings, rendered scenes,
intents, topology, animations, physical data
features, dimensions… for sequential engineering
Bottom-up analyses …
modelling
V: representations of design
intents, parameters, Top-bottom G: functionalities, satisfaction,
features, and mates, modelling productivity, time, cost,
modelling strategies, flexibility to adapt
techniques, and CAGM System changes,
sequences… manufacturability, …

Add-ins

Collaborative platforms
Design templates

Macros supports
Libraries and utilities

Parametric modelling
Design toolboxes
Designers

R: resources

Abbreviation Meanings Abbreviation Meanings

I Inputs C Constraints
V Variables O Output
E Computer environment G Goals
R Resources

Fig. 2.2 Description of computer-aided geometric modeling (CAGM) system

Product needs, Metals, plastics,

design intents, ceramics, alloys, Processes
knowledge, Processes (data): composites, (materials):
expertise.. Sketching, utilities… Casting, powder
extruding, metallurgy,
Data patterning, Materials injecting
Databases, parametric modelling,
modelling, Industrial materials metal forming,
component models, in sheets or rolls,
drawings, standards, assembling, machining,
lofting, sweeping, parts, components, additive
library data, working and functional
data, geometric animating, and manufacturing,
converting… products such as assembling,
models, views, and appliance, machines,
videos… transporting….
instruments, …

(a) CAD System (b) Manufacturing System

Fig. 2.3 Transformation similarity of a CAD system and manufacturing system

2.2 Computer-Aided Design (CAD) System 39

a CAD system, certain types of data inputs are processed and transferred into other
types of data outputs through a series of computing processes (Bi and Cochran 2015).
Types of data can be product needs, design intents, expertise, databases, component
models, drawings, standards, library data, working data, geometric models, and
views, animation, and videos. Types of computing processes can be performed by
sketching, extruding, parametric modeling, assembling, lofting, sweeping, animating,
and converting, and more add-ins functions for different applications.
System architecture describes constitutive components, their relations, and the
rules for the changes of these system factors over time (Niu et al. 2013). Figure 2.3
shows a comparison of the architecture of a CAD system and a manufacturing
system. Similar to a manufacturing system architecture that consists of machine
tools, controllers, material handling facilities, and other manufacturing resources
that are required to transfer raw materials to finished products, a CAD system archi-
tecture consists of computing hardware, software, and input and output devices that
are required to transfer, share, process, and produce output data.
As shown in Fig. 2.4, the hardware of a CAD system is one or a number of
networked computing resources. A computer resource typically consists of (1) a
central processing unit (CPU) to process data, (2) memory and storage to keep
data, and (3) an operating system to manage the resource. The hardware determines
the capacities of the CAD system in terms of processing speeds, and volume and
types of data. The computing capacity used to be measured by the number of tran-
sistors on an integrated circuit and it was doubled every 18 months according to
Moore’s law. Since the advancement of classic digital computing (CDC) in the
forms of binary digital electronics seemed to approximate its limits, new computa-
tional models such as analog computing (AC), neuro-inspired computing (NC), and
quantum computing (QC) were expected to continuously improve the computing
capabilities of hardware systems exponentially (Intel 2002; Shalf and Leland 2015).
Moreover, the computing capacity of a CAD system can be increased by parallel

Computer and Machine i

storage i

Computer
Input: Machine i+1
and storage i- Input and Machine i-1
Keyboards, 1 Output:
Input mice, space Input
Computer and End-
balls, scanners, effectors,
haptic devices… storage i+1
Material robots,
Data Output: conveyers,
s
Screens, labors,
printers, So ware i cranes, Controller i
Output plotters, VR Output
AGVs,
devices, videos, material-
digital files… handling,…

So ware i+1 Controller i-1 Controller i+1

So ware i-1

(a) CAD System (b) Manufacturing System

Fig. 2.4 Architecture similarity of a CAD system and manufacturing system

40 2 Computer-Aided Design

computing by a number of CPUs or computers in the distributed environment or

cloud computing by the cloud services over the Internet (Zissis et al. 2017; Wang
et al. 2017). An input device is to input raw data and the designer’s intents in a CAD
system; common types of input devices keyboards, mice, space balls, haptic devices,
barcode readers, and scanners. An output device is to output design outcomes for
their usages; common types of output devices are screens, printers, plotters, Virtual
Reality (VR) devices, videos, and digital files. Finally, the software determines the
functionalities of a CAD system. CAD software can be either open-source, non-
commercial free or commercial. Popular open-source or free CAD software packages
are Autodesk 123D, freeCAD, SketchUp, Onshape, TinkerCAD, 3D System Cubify,
OpensCAD, RepoCAD, and Blender (Junk and Kuen 2016). Examples of popular
commercial CAD systems are Solidworks, AutoCAD, Creo, Inventor, Solid Edge,
Catia 3D Experience, and NX for Product Design (Hooper 2020). Without losing the
generality, Solidworks is selected in this chapter to illustrate the functionalities of
CAD system.

2.3 Properties of Solid Models

The geometric representation of three-dimensional (3D) solids plays important role

in making discrete products. A system for the geometric representation of objects
consists of four components (1) symbol structures of solid objects, (2) processes in
which geometric information such as volume and dimensions are defined, (3) input
mechanisms for creating and editing representations in an iterative process, and
(4) the devices to output the graphic results. As shown in Fig. 2.5, the sub-system
for entering, processing, modifying the graphic representations of objects is called
computer-aided graphic modeling system (Requicha 1980).
In CAGM, solids are postulated as the abstract geometric entities that are bounded,
closed, regular, and semianalytic in the subset of 3D Euclidean space (E 3 ). A very

Computer aided graphic modelling system

Process 1 Representation 1
Definition of Definition of Representations
Object Element of Object
Outputs

Geometry Types Geometry …

Inputs

Process i

Representation
Commends to j
manipulating Definition of
geometric Commend
features Types

Fig. 2.5 Four components of a computer graphic modeling system

2.3 Properties of Solid Models 41

small subset of E 3 can be used to represent physical solids since an abstract solid must
possess the properties of rigidity, homogeneity, finiteness, closure, finite describa-
bility, and determined boundaries that are described in Table 2.2 (Requicha 1980).
The properties in Table 2.2 can be used to justify if a computer model is a legal solid
model or an operation over the solids is legal or illegal. The example in Fig. 2.6 is not
a legal solid model; it does not possess the properties of homogeneity and determined
boundaries due to the existence of dangling line and face. The example in Fig. 2.7
shows that the different algorithms for the same Boolean operation may produce a
legal or illegal solid model.

Table 2.2 The properties of solid models

Properties Description
Rigidity: A solid model has a given shape or configuration that is
independent of the orientation and location of solid.
Homogeneity: Any position in E3 can be defined as one of ‘interior’,
‘exterior’, and ‘boundary’ unambiguously; the solid cannot
have dangling or isolated elements.
Finiteness: A solid must have a finite volume of space.
Closure: Rigid motions and the Boolean operations of solids must
sustain all properties of solids.
Finite describability: A solid has a finite set of elements such as ‘faces’ to ensure
that the solid is representable as a computer model.
Determined boundaries: A solid has a set of clearly defined boundary faces to
determine the interior volume.

Dangling line

Dangling face

Solid cube

Fig. 2.6 Example of illegal solid model

42 2 Computer-Aided Design

Fig. 2.7 Example of illegal solid operation

2.3.1 Legality of Solid Models

A physical body (B) can be represented by (1) a set of graphic elements such as
vertices (V ), edges (E), loops (L), genus (G), and faces (F) and (2) their dependent
relations. The meanings of a loop and genus are illustrated in Fig. 2.8.
As a pyramid example shown in Fig. 2.9, the finite volume is bounded by five
faces (F 1 , F 2 , F 3 , F 4 , and F 5 ), each face is formed by a number of edges (e.g., F 1 is
formed by the edges of E 1 , E 2 , E 3 , and E 4 ), and each edge is formed by two vertices
(e.g., E 1 is formed by the vertices of V 1 and V 2 ). The graphic elements of vertices,
edges, faces, and bodies are at different levels but with some dependences. To have
the complete in a solid model, a data structure is required to represent the topological
relations of different graphic elements. Figures 2.10 and 2.12 show a hierarchical
and network data structure for a pyramid model, respectively.
Note that the hierarchical data structure in Fig. 2.10 includes some redundant
information since a graphic element at a low level has its relations to multiple elements
at a high level. For example, a vertex can be an end point of two edges, and an edge can
be a boundary line of two faces. A network data structure in Fig. 2.11 can eliminate
such redundancies since data pointers are used to represent the topological relations

Fig. 2.8 The meanings of loops and genus

2.3 Properties of Solid Models 43

Fig. 2.9 Vertices, edges, and faces of a pyramid

Fig. 2.10 Hierarchical structure of pyramid object

Fig. 2.11 Network structure of pyramid object

44 2 Computer-Aided Design

Fig. 2.12 Example of solid

object in Example 2.1

of graphic elements. Moreover, the number of data pointers for a certain graphic
element can be determined by the number of relations of this element with others.
The Euler-Poincare equation can be applied to justify if a solid model is legal
(Sanchez-Cruz et al. 2013). Euler and Poincare proved that a polyhedral is homo-
morphic to a sphere; thus, for a model with the given numbers of faces (F), edges
(W ), vertices (V ), bodies (B), loops (L), and genus (G), it is topologically valid if
one of the first two of the following three conditions is satisfied:
⎫
F − E − V − L = 2(B − G) → General object ⎬
F−E+V =2 → Simple solid (2.1)
⎭
F −E+V −L = B−G → O pen object

Based on which condition has been satisfied, the corresponding object type is
general object, simple solid, and open object, respectively.
Example 2.1 Determine the numbers of different graphic elements, and use the
Euler-Poincare equation to classify the object type.

Solution For the object in Fig. 2.12, V = 24, E = 36, F = 1 4, G = 1, L = 2, and B

= 1. Accordingly, the first conditions in the Euler-Poincare equation is satisfied, i.e.,

24 − 36 + 14 − 2 = F − E − V − L ⇔ 2(B − G) = 2(1 − 1)

Therefore, the represented model is a general object.

2.3.2 Geometric Constraints

Depending on the physical process in generating the shape of solid, the corresponding
solid model is valid when it meets the geometric constraints such as the following.
Legality. A solid model must be legal that has all of the properties in Table 2.2.
It should meet one of the first two conditions in the Euler-Poincare equation. Two
2.3 Properties of Solid Models 45

Fig. 2.13 Illegal model examples (open objects)

well-cited examples in Fig. 2.13 are illegal solid models since the third condition in
the Euler-Poincare equation is satisfied.
Free of Interference. In representing a physical solid, no overlapped material is
allowed at any position within the solid volume. Therefore, no interference is allowed
in the corresponding solid model. Figure 2.14 shows the representation of a tapered
spring. When the wire diameter is smaller than the pitch of the spring, the solid
model is valid (see Fig. 2.14a). When the wire diameter is large than the pitch of
the spring, the interference occurs between two neighboring rings, the solid model
becomes invalid; unless that the overlapped materials are merged at the interfered
volume (see Fig. 2.14b).
Accessibility. When the shaping process of a solid object is concerned, the finite
volume for a solid must be accessible in the manufacturing process. While different

Fig. 2.14 Interference example

46 2 Computer-Aided Design

manufacturing processes shape the geometric shape of solid in different ways, the
accessibility to the finite volume can be different. In other words, an accessible solid
volume for one manufacturing process does not mean that it will be accessible to
another manufacturing process. Accessibility and accessibility have been specially
considered in creating an assembly model of products. Figure 2.15 shows an example
of a combined solid model with a total of four objects, and three objects are within the
enclosed cavity of the fourth object. The virtual model is valid if the combined solid
is made by additive manufacturing; otherwise, the solid volumes of three objects are
inaccessible.
Manufacturability. A geometric constraint might be relevant to the required
tooling in a shaping process of solid object. Taking an example of a product from
the casting process in Fig. 2.16, the geometric shape of the product is determined by

Enclosed 4
body with an
enclosed
cavity 1
Two solids in the
enclosed cavity of
A solid in the other solid
enclosed cavity of 2 3
other solid

Fig. 2.15 Example of accessible/inaccessible solid volume

Parting
plane

Core Core
placement is placement is
feasible for infeasible for
internal internal
surface surface

(a). Solid with feasible internal surface (b). Solid with infeasible internal surface

Fig. 2.16 Consideration of manufacturability of solids

2.3 Properties of Solid Models 47

the cavity formed in the mold assembly. An internal surface of casting is defined by
a core, and one has to ensure the core is feasibly removed from the mold assembly
without causing damage to internal surface. From this perspective, one of the solid
models of casting in Fig. 2.16a is valid since it is feasible to place a removable core
to generate an internal surface.

2.4 Graphic Modeling Techniques

As shown in Fig. 2.17, the representation of a physical object consists of the graph-
ical elements at different levels from points, lines, surfaces, volumes, features to
design intents. Accordingly, the modeling techniques can be classified based on their
capabilities in creating and manipulating these graphical elements into wireframe
modeling, surface modeling, solid modeling, space decomposition, feature-based
modeling, and knowledge-based modeling. The higher level of graphic elements is,
the more advanced the computer-aided graphic modeling (CAGM) tool is required.

Computer Aided Applications in Product Lifecycle

Design Intent

Feature

Knowledge-
based
Volume modelling
Solid
modelling
Surface
Surface Feature-based
modelling modelling

Line Space
Wireframe decomposition
modelling Boundary
representation
Point modelling

Fig. 2.17 Classification of computer-aided geometric modeling (CAGM) tools

48 2 Computer-Aided Design

Fig. 2.18 Modeling techniques versus types of graphic elements

The example in Fig. 2.18 shows the differences of using these modeling techniques
to model a physical object formed by two or more rectangular blocks in a golden
ratio, i.e., Width: Length = 1: 1.618.

2.4.1 Wireframe Modeling

In wireframe modeling, graphic elements are points for vertices and lines for edges,
and an object is represented by its boundary edges. Edges can be circles, arcs, straight
or curvy lines. Since no upper level graphic information such as faces or volumes is
available, the visibility of a graphic element is not detectable, and the information of
surface areas or masses is not available. Since creating a wireframe model requires to
input the coordinates of vertices, wireframe modeling is ineffective to model complex
shapes due to a large number of data points. Moreover, a wireframe model may cause
the ambiguity of the represented geometry. However, since points and lines are two
primary types of graphic elements, wireframe modeling can be used as the supporting
technique of other modeling methods.
2.4 Graphic Modeling Techniques 49

2.4.2 Surface Modeling and Boundary Representation

Modeling (B-Rep)

In surface modeling, a physical object is modeled as a set of finite surface patches

in free forms. A surface model has sufficient information to determine the visibility
of graphic elements. However, a surface model does not include the information of
volumes or masses; since no thickness of volume information is available.
Surface modeling can be advanced to be boundary representation modeling (B-
Rep). B-Rep further defines a finite and closed cover of a solid over a surface model.
To define the volume of a solid, each constitutive surface patch is treated as a half
space to separate inside space from outside space in a Cartesian space:

H = { P : P ∈ E 2 and f ( P) < 0} (2.2)

where H is the set of all points in the inside half space; P is an arbitrary point in
the Cartesian space E 3 ; f (P) = 0 is the function of surface patch; f (P) < 0 is the
constraint for any point at the negative direction of normal vector of the surface patch.
As shown in Fig. 2.17(b), the normal vector of a surface patch defines the inside
half space of the patch; each surface patch divides the Cartesian space into two
regions of infinite extension. Furthermore, if a point in the Cartesian space is at the
inside half space of all boundary faces, it is in the volume of the solid. Accordingly,
the volume of the solid S is the intersection of the half spaces H i where (i = 1, 2,…N)
of all surface patches as,
 N 
 
S= Hi (2.3)
i=0

where S is the solid pace, N is the number of boundary surface patches, and H i is
the half space of the i-th surface patch (i = 1, 2,…N).

2.4.3 Solid Modeling and Space Decomposition

In solid modeling or space decomposition modeling, the volumetric information is

defined by specifying all geometric dimensions in a 3D space.
Figure 2.19 shows three ways to specify 3D volumetric dimensions in solid
modeling. The first way (see Fig. 2.19a) gives the dimensions of a solid in a 3D
space by a 2D sketch and 1D depth along the direction of extrusion. The second
way (see Figs. 2.19b and 2.18c) gives the dimensions of a solid in a 3D space by
one 1D path, 1D thickness and 1D width over the cross-section. The third way (see
Fig. 2.19c) gives the dimensions of solid in a 3D space by one 1D path and 2D profile.
50 2 Computer-Aided Design

Fig. 2.19 Three ways of specifying 3D dimensions in solid modeling

In space decomposition modeling, a solid is represented by a set of isomorphic

3D cells. The cell size determines the precision of the geometric representation, and
it should usually be several-order smaller than the dimensions of the solid. As shown
in Fig. 2.18d, a continuous space where the solid locates is discretized into a 3D array
of isomorphic cells, the state of each cell is checked to see if it locates in the finite
volume, and the solid model is the collection of all cells in the volume of solid. Space
decomposition modeling is widely used to represent a continuous domain by a set
of discretized elements and vertices in numerical simulations such as finite element
analysis (FEA).

2.4.4 Feature-Based Modeling

In feature-based modeling, basic graphic elements such as drawing references and

solid primitives are treated as features, and a solid object is composed of a set of
constitutive solid features by the composition operations so-called Constructive Solid
Geometry (CSG). Note that a feature is unnecessary a solid element; a feature such
as a reference coordinate system is defined to create, position, and dimension solid
features efficiently.
Figure 2.18e shows that the solid is formed by uniting feature-A and feature-
B at the locations these two features are placed. Feature-based modeling is very
flexible in the sense that any pre-built graphic element can be defined as a feature for
the purpose of reusing (Fig. 2.20). shows 8 most commonly solid primitives which
include cube, rectangular block, prism, sphere, cylinder, tapered cylinder, cone, and
torus. In creating a complex geometry, solid primitives can be tailored by some solid
modeling tools such as extruding, revolving, sweeping, and lofting. These tools will
be discussed in the coming sections Table 2.3.
As shown in Fig. 2.21, three basic Boolean operations of CSG for the composition
are union (∪), intersection (∩), and difference (\). Note that (1) the operation of
difference depends on the order of features, and (2) the operations take place when
the positions and orientations of the features are given.
2.4 Graphic Modeling Techniques 51

Table 2.3 Common features and illustrations

Type Operation Illustration
Sketched Extrude (2D-sketch + 1D depth) 1-D depth 2

1
2D-sketch

Revolve (2D-sketch + 1D axis) 1-D Axis

2

1
2D-sketch

Sweep (2D enclosed profile + 1D path in 2-D profile

2D or 3D) 1

1
2
1-D depth in
2
2-D or 3-D

Boundary boss/base (Two 2D profiles +

path normal controls) 2

1 2-D profiles

1
2

2
Path normal control

Loft (a set of 2D profiles + the ordered

connections of vertices on the profiles)

1
2

2
Connections

(continued)
52 2 Computer-Aided Design

Table 2.3 (continued)

Type Operation Illustration
Build-in Hole wizard (the set of hole positions + Hole positions
hole type + dimensions) 1

2 Preview of types
3 and dimensions

Thread (Thread edge + end-conditions + Thread edge

thread dimensions)
1

2 Preview of
3 thread

Fillet/Chamfer (a set of selected edges or

faces + size of fillet or chamfer)
Fillet
1 edges

2
Preview of fillets

Pattern (linear or circular) (the direction Objected to be

of pattern + the instance (features, solids, patterned 2

or other entities) + other parameters such 1

as spacing, and number of instances) 3

3
1Direction
Preview of patterns

Rib (an open 2-D sketch + rib thickness

+ the direction of material)
Thickness
2 2

3 3
Material
direction
1

1
Open sketch

(continued)
2.4 Graphic Modeling Techniques 53

Table 2.3 (continued)

Type Operation Illustration
Draft (a set of surfaces to be drafted + the
Direction of
draft angle + the direction of drafting) drafting
2

Surfaces to
1 be drafted
3

1 3 Draft angle

Shell (the face to create the shell + the

wall thickness)
2
1

1 2
Wall The face to
thickness create shell

Wrap (Wrap type + warp method + wrap Wrap sketch

sketch + wrap face + warp thickness) 3

3
4
5
4 1 Preview
Wrap face 2 of
5 warping

Reference Point (the relation of the selected entity

with the point + the selected entity)
Selected face
1 1 2-D profiles

2
Center of face

Axis (the relation of the selected entity

with the axis + the selected entity) Selected line
1 1

2
Align axis with line

(continued)
54 2 Computer-Aided Design

Table 2.3 (continued)

Type Operation Illustration
Plane (the relation of the selected entity
with the axis + the selected entity)
Selected face
1 1

2
Align plane with the selected face

Coordinate system (the origin + two of

three axes)
1 Z-axis Y-axis
4 3
2
3

Origin 1

X-axis 1

(a) cube (b) rectangular block (c) prism (d) sphere

(e) cylinder (f) tapered cylinder (g) cone (h) torus

Fig. 2.20 Eight types of commonly used solid primitives

To place a 3D feature at the right position and orientation for a CSG operation in
a given coordinate system, one or more coordination transformations can be applied.
A coordinate transformation (CT) is a mathematical process of transforming a set
of graphic features into a new position and orientation coordinate system by trans-
lating, copying, rotating, mirroring, and scaling these features. Figure 2.22 shows
the example of the transformed object using these four operations, respectively.
2.4 Graphic Modeling Techniques 55

Fig. 2.21 Three basic Boolean operations of 3D features

Fig. 2.22 Four operations of

coordinate transformation
(CT)

Assume that (1) a solid is constructed from a set of solid primitives S i,pi at the
respective positions pi (i = 1, 2, 3,…, N), where N is the number of solid primitives;
(2) a Boolean operation (‘∪’, ‘∩’ or ‘\’) over two primitives is denoted as ⊗, then
the solid is represented as a series of composition operations in CSG as

Sc = ((S1, p1 ⊗ S2, p2 ) ⊗ S3, p3 . . .) (2.3)

56 2 Computer-Aided Design

S2,p2
Sc

(a) S1,p1

S1,p1
(b) S2,p2
S3,p3 S4,p4

4 R0.095
(c) S3,p2

(d) S4,p4 (e) Sc=((( S1,p1 S2,p2)\ S3,p3) \ S4,p4)

Fig. 2.23 Using CSG to build a model for a Lego piece

where S c stands for the composed solid, S i,pi (i = 1, 2, 3, …N) is solid primitive i at
its respective position pi , and ⊗ is one of the Boolean operations (‘∪’, ‘∩’ and ‘\’).
Figure 2.23 shows the series of logical operations to model a Lego piece S c from S i,pi
(i = 1, 2, 3, 4); note that a solid primitive, such as S 2,p2 and S 4,p4 , can be a composed
solid itself.
In feature-based modeling, a solid is modeled by a set of solid primitives and the
corresponding logical operations over these primitives. Accordingly, the data struc-
ture of the solid is a graph for the order and dependences of the Boolean operations
of solid primitives. Therefore, a CSG model consists of
(1) A binary tree for the Boolean operations,
(2) A set of the links to solid primitives as the outer leaf nodes of tree, and
(3) A set of intermediate components as the interior nodes of tree as the results of
the Boolean operations over solid primitives or components.
Figure 2.24 shows a data structure example. The model has two data types for (1)
solid primitives and (2) composition operations, respectively. For the types of solid
primitives, the model includes three features, i.e., two blocks and one cylinder. For
the types of logical operations, the model includes one union (∪) operation of two
blocks and one difference (\) operation of the united component and the cylinder.
Since feature-based modeling creates a model for an object from solid primitives.
Since solid primitives include volumetric information as well as graphic elements at
low-levels such as vertices, edges, and boundary faces, a CSG model has the complete
solids. Therefore, featuring-based modeling requires less steps than wireframe
modeling and surface modeling to create a virtual model. However, feature-based
modeling specifically emphasizes the representation of graphic contents of phys-
ical objects, it can be further improved as knowledge-based modeling to incorporate
design intents in modeling practices.
2.4 Graphic Modeling Techniques 57

Final
model

Intermediate
component

Tree of
composition
operations

A set of solid
primitives

Fig. 2.24 Example pf CSG data structure

2.4.5 Knowledge-Based Modeling

Products are designed and manufactured with purposes; so are the geometries and
shapes of products. Even though there are numerous ways to create a virtual model for
a given object, it is very helpful for a modeler to take into account design knowledge
in modeling graphic objects. Incorporating design knowledge in a solid helps in (1)
highlighting main attributes of products, (2) simplifying the modeling processes,
(3) facilitating product evolution and the reuse of product designs, (4) reducing
product design times and improving the productivity of modelers, and (5) sharing
and exchanging design concepts in communication and design collaboration.
The mechanism of incorporating design knowledge in a solid model is the use of
design intents. Design intent is to describe how a design feature is represented and
how it is guided when it needs changes. Design intents are embedded in a graphic
model and corresponded to certain dimensions and relations; a change made on
a design intent updates these graphic contents (i.e., dimensions and relations) auto-
matically. Since design knowledge is far beyond graphic contents themselves, design
intents are not just sizes and shapes of features, the scope of design intents covers
tolerances, requirements or constraints of manufacturing and assemblies, relation-
ships of features and dimensions, and other design factors in product lifecycles (Bi
and Wang 2020).
From the perspective of solid modeling, design intents are represented by various
types of design parameters in solid models. Using design parameters ensures that
the model can always meet the requirements specified by design intents. Taking an
example of the sketch in Fig. 2.18c, the design intent of the golden ratio of length and
width of the rectangle was represented as a design equation in a parametric model;
58 2 Computer-Aided Design

no matter what the length or width is, they are in a golden ratio. From the perspective
of product use, if the design intents of a product cannot be adequately represented,
the product model might be misleading and sometimes even useless.

2.5 Design Parameters, Features, and Intents

Either feature-based modeling or knowledge-based modeling is a type of parametric

modeling method. In parametric modeling, various parameters are defined to repre-
sent positions, dimensions, geometries, and relations of features. Updating a param-
eter changes all of the associated features of a solid model automatically. From this
perspective, a parameterized model is embedded with design knowledge and intelli-
gence so that design features can be defined and modified with the minimized manual
efforts.

2.5.1 Design Features

Design features are high-level representations of building blocks of a solid by which

graphic elements, engineering knowledge, and design intents are structured. To repre-
sent a solid model, the modeler should firstly understand the object to be modeled
in terms of its constitutive building blocks and the corresponding composition oper-
ations. The modeling procedure follows after these building blocks are identified,
these building blocks are models as design features, respectively, and then structured
as a solid model based on the specified Boolean operations. Figure 2.25 illustrates
the directions of understanding and modeling procedures of an exemplified solid.
As shown in Fig. 2.25, graphic features can be classified into sketched features
and build-in features; in addition, the modeling references or any reusable entities
can be treated as features. Table 2.3 shows some common features and examples
under each catalogue.
As shown in Table 2.3, (1) the examples of sketched features are extrude, revolve,
sweep, boundary boss/base, and loft; in a sketched feature, the main attributes are
defined in two-dimensional (2D) or 3D sketch(s). (2) The examples of the build-in
features are hole wizard, thread, fillet, chamfer, pattern, draft, rib, and shell. In a
build-in feature, the main attributes are determined by computer algorithms based on
users’ input(s). From a manufacturing perspective, most of the features are associated
with certain manufacturing processes, for example, a fillet on a shaft shoulder relates
to a turning operation, a counter-bore hole relates to a drilling operation, and a
draft angle relates to a variety of the manufacturing processes; the part geometry is
determined by that of molds, dies or tools. (3) Examples of reference features are
points, axes, planes, and coordinate systems.
2.5 Design Parameters, Features, and Intents 59

Fig. 2.25 Understanding and modeling procedures in feature-based modeling

2.5.2 Design Parameters

Design parameters are at the lowest level of a solid model, and design parameters
include everything that are required to represent design features and their relations.
Table 2.4 shows some common types of design parameters and examples in feature-
based modeling.

Table 2.4 Common types of design parameters and their applications

Parameter Data Type Example
Dimensional Numerical (1) the height and diameter of cylindrical feature,
variables (2) the length, the height, and the width of a rectangle
block, and
(3) the radius of a fillet feature.
Constraints Boolean (1) parallel, perpendicular, angular relations of two lines,
(2) equal radii of two arcs,
(3) convert entities in sketching, and
(4) the symmetric relations of two objects about a plane.
Operations Mixed data and logic (1) the direction and instances of a pattern feature,
(2) drafting over existing face, and
(3) the direction and the thickness of a rib feature.
Equations Mathematic (1) the link value of multiple dimensions, and
expression (2) the dependent relation of two or more dimensional
variables.
60 2 Computer-Aided Design

Design parameters for

position and orientation:
ALF,
DX, and
DY

Design parameters for

sizes:
L,
W, and
H

Fig. 2.26 Example of defining design parameters for sizes and positions of object

2.5.2.1 Dimensional Variables

Generally, a graphic feature has a number of dimensions to describe (1) the geometric
size and (2) the relative position and orientation with respect to a reference coordi-
nate system. By defining a design parameter to a dimension, the design parameter
applies a hard constraint to the dimension; since the dimensional value will be
fully controlled by the design parameter. Figure 2.26 shows an example of the fully
constrained object where the design parameters have been defined for its position,
and orientation, and sizes in three dimensions.
A constraint is hard implies that the change of the corresponding dimension has to
be made by updating the value of the design parameter. Note that the default setting of
a dimension in a modeling process is a soft constraint; in other words, the dimensional
value can be changed by an undergoing algorithm in computer-aided modeling tool.
Figure 2.27 shows the difference of the hard and soft constraints; a fillet radius in
the case of (a) hard constraint is fully controlled by its design parameter.
It is a good practice to define a design parameter to a critical dimension as a
hard constraint in modeling, especially at the stage of the sketching process of 3D
features. In Solidworks, the dimensions in a sketch are generally created by using
the smart-dimension tool in sketching, and additional dimension(s) of a 3D feature
are automatically generated by the software based on the modeler’s inputs when the
feature is created.

2.5.2.2 Constraints

A constraint refers to a dependent relationship of two or more geometric enti-

ties. Constraints are mostly used to represent the design intents for the relations
of geometric entities in sketching processes. Figure 2.28 shows a list of common
relations of graphic entities such as axes, line segments, and planes.
2.5 Design Parameters, Features, and Intents 61

(a) Defining a design

parameter for the size of
fillets create a hard
constraint of fillet radii in
the model, i.e., fixed size.

(b) Default non design

parameter of the size of
fillets corresponds to a
soft constraint to the
fillet radii in the model
i.e., changeable size.

Fig. 2.27 Difference of hard and soft constraints on a dimension

(a). Perpendicular (b). Parallel (c). Tangent (d). Coincident (e). Coincident

(f). Collinear (g). Horizontal (h). Vertical (i). Equal (j). Fix (k). Symmetric

Fig. 2.28 Common constraints for geometric relations of graphic entities in sketching

In the sketching process, a modern CAD tool can automatically detect applicable
constraints of a graphic entity with others. Once a constraint is detected and becomes
active, the system gives the feedback, and the modeler has the option to apply this
constraint to the selected entities. A constraint will be automatically accepted when
no action is taken when the constraint is active; otherwise, the modeler needs further
actions to remove or change a constraint.
A well-structured sketch usually involves in many constraints to represent
different design intents. Constraints in a sketch can be created, modified, or removed
by the Display/Delete Relations tool in Solidworks. Figure 2.29a is the icon for
the Display/Delete Relations for activation. Figure 2.29b shows a scenario when
the selected entities are a point and a line. The system identifies three applicable
constraints: (1) the point is a midpoint of the line segment, (2) the point is coinci-
dent with the line segment, or (3) both point and line entities are fix in the sketch.
Figure 2.29c shows the case when the selected entities are two line segments. The
applicable constraints identified by the system include horizontal, vertical, collinear,
perpendicular, parallel relations of two lines, an equal length of two lines, and both
line segments are fixed in the sketch. Figure 2.29d illustrates the case when two
points are selected, the applicable constraints identified by the system include the
62 2 Computer-Aided Design

(a). Activating ‘Display/Delete Relations’ tool

(b). Selected line and point

(c). Two selected lines (d). Two selected points

Fig. 2.29 Examples of applicable constraints of selected entities

relations of horizontal, vertical, merge of two points, or both of the points are fix
in the sketch. Therefore, the system determines the types of applicable constraints
based on the selected number and types of graphic entities.

2.5.2.3 Operations

An operation is to create a graphic feature from a sketch and add one or two dimen-
sions to the features. Certainly, design parameters should be defined to represent
these dimensions and other associated options related to the operation. Figure 2.30
shows the setup of a linear pattern operation. The design parameters involved in the
operation include the direction, spacing, and the number of instances of the first and
the second directions of patterning, and the feature(s) or bodies to be patterned.
Figure 2.31a shows the setup of a surface thickening operation. The design param-
eters involved in the operation include the selected surface, the direction, and the
dimension of thickening. Figure 2.31b shows the setup of a revolving operation.
The design parameters involved in the operation include the axis, the angle, and the
direction of revolving, and the selected 2D sketch.

2.5.2.4 Design Equations

A design equation is used to express the dependence of one design parameter on

other design parameters. Using design equations is effective when a set of design
parameters are associated with each other by certain rules or standards; therefore,
design equations are widely used to model machine elements and other standardized
2.5 Design Parameters, Features, and Intents 63

Feature(s) or
bodies to be
patterned
1
7

2
3

5 The 2nd 1
6 direction of 4 The 1st direction
pattern of pattern

Spacing in Spacing
st
in
2 the 2nd 5 the 1
7
direction direction

3 Instances in Instances
st in
6 the 1
the 2nd
direction direction

Fig. 2.30 Design parameters in the operation of linear pattern

Revolving
2-D surface axis 1
1

1
1
2
3
2 2-D
sketch
3
2 3
2 3
2-side Dimension Revolving
thickening of thickening angle

(a) Design parameters in surface (b) Design parameters in revolving

thickening operation operation

Fig. 2.31 Design parameters in the operations of surface thickening and revolving

products. Note that standardization is critical to reduce product cost, and standard-
ization makes independent parameters dependent. Therefore, design equations can
be used to represent the design intents where geometric dimensions are adhered to
a specified set of rules. Figure 2.32 shows the models of a gear or a nut where a
number of design equations have been defined to represent the dependences of these
parameters in two models, respectively.
A design equation is a mathematic expression that follows the syntax of the
programming languages of a CAD system. In such an expression, a dependent vari-
able on the left side of the assigning symbol ‘=’, and its value is determined by the
expression on the right side of ‘=’. Variables in the expression of a design equation
can be either of a global or independent variable. The design variables are mostly
numeric for dimensioning; however, design variables can also be defined in other
64 2 Computer-Aided Design

(a) Design equations in a

helical gear model

(b) Design equations in a nut model

Fig. 2.32 Machine element models with design equations

formats such as the states of features, material, and document properties. In addi-
tion, a CAD system supports most common operations over design variables. For
example, Solidworks supports a list of the legal operations as shown in Table 2.5.
Figure 2.33a shows how to access the equations tool in Solidworks, which is listed
in the insert menu. Once the equations tool is activated, the window of managing
equations is opened as shown in Fig. 2.33b. It consists of three fields for global
variables, feature variables, and dimensional variables, respectively. A variable on
the left side is determined by the expression on the right side of the equation, and its
evaluated value is seen in the last column.

Table 2.5 Legal mathematic operations in design equation tool of Solidworks

Operator or Operator or
Meanings Meanings
Function Function
+ Addition arccos(x) find the angle of the cosine ratio x
- Subtraction atn(x) find the angle of the tangent ratio x
* Multiplication arcsec(x) find the angle of the secant ratio x
/ Division arccosec(x) find the angle of the cosecant ratio x
^ Exponentiation arccotan(x) find the angle of the cotangent ratio x
sin(x) fine the sine ratio of angle x abs(x) find the absolute value of x
cos(x) fine the cosine ratio of angle x exp(x) find e raised to the power of x
tan(x) fine the tangent ratio of angle x log(x) find the natural log of x to the base of e
sec(x) fine the secant ratio of angle x sqr(x) find the square root of x
cosec(x) fine the cosecant ratio of angle x int(x) find the integer of x
cotan(x) find the cotangent ratio of angle x sng(x) find the sign of x as ‘-1’ or ‘1’.
arcsin(x) find the angle of the sine ratio x pi refers to the ratio of the circumference
2.5 Design Parameters, Features, and Intents 65

Global
variables
Features

Dimensions

(b). Managing equations window with the sections of ‘global

(a). Access ‘equations’ tool variables’, ’features’, and ‘dimensions’

Fig. 2.33 Using the design equations tool in Solidworks

2.5.2.5 Design Tables

A design table is used to represent a set of configurations in one part or assembly

model. Using a design table is able to represent a part or product family by a single
model. One part or assembly is referred to as a configuration in the family. Typi-
cally, a model of a part family consists of (1) a set of design parameters for major
dimensions and the states of optional features (2) a design table that includes a set
of configurations with the assigned dimensions or states of these design parameters.
A design parameter in a design table differs from that in a design equation; in
sense that the former is discrete or logical variable, while the latter is continuous. In
addition, a number of design parameters can be dependent variables; while a design
equation can only deal with one dependent variable.
A design parameter can be any graphic entity that has an associated parameter
identity (ID) in a solid model. Common types of design parameters are constraints,
dimensions, and states of features, parts, and mates. Therefore, the design table tool
is capable of (1) including a variety of sketches, features, parts, or assemblies in
a single model, (2) creating new components from existing components, and (3)
incorporating the modeling intelligence such as parametrical constraints and design
equations in a design table. Figure 2.34 shows some part and product families, which
can be modeled by a design table with a set of the design parameters for features,
parts, components, and assembling mates. To sustain the consistency of data sources,
only one design table is allowed in one part or assembly model.
In a design table, the variants of part or product are represented by the config-
urations, and the configurations can be modeled manually or directly generated by
computers. As shown in Fig. 2.35a, a design parameter can be assigned with different
values manually to create different configurations. Alternatively, a design table can
be defined to assign different values to design parameters in order to create new
configurations automatically (see Fig. 2.35b).
When a design table is defined for a model with some existing configurations, these
configurations will be automatically included in the table. Available configurations
66 2 Computer-Aided Design

(a). Feature family (b). Part family

(c). Component family (d). Product family

Fig. 2.34 Examples of using a design table for part families

(a). Manually created configurations (b). Configurations generated

before a design table is defined automatically from a design table

Fig. 2.35 Example of configurations in a design table

can be viewed and edited under the configurations tool of the model tree as shown
in Fig. 2.35. Finally, if a model has a design table already, no new design table is
allowed, and the changes on the configurations can only be made by editing the
existing design table.
The design table tool in Solidworks is implemented by the external Microsoft
Excel program. Therefore, Microsoft Excel must be available to apply the design
table tool. A design table is directly associated to its model in sense that all of the
controlled attributes in the table must have internal links to the corresponding design
parameters such as dimensions or feature states in the model.
Figure 2.36 illustrates the procedure to create a design table in a part or assembly
model. Firstly, all controllable dimensions and features must be defined in the model.
Secondly, the modeler has the options of (1) manually define a few configurations or
(2) proceed to create a design table directly. Thirdly, the design table tool under the
insert menu is activated, and the modeler then selects and adds the dimensions and
features as controllable attributes into the design table. Table 2.6 shows the types of
design parameters which can be controlled by the design table. Fourthly, a number of
2.5 Design Parameters, Features, and Intents 67

Create a solid model

Insert a Add controllable dimensions and
with well-defined
design table features in the table from a solid model
design parameters 1 2 3

Manually define Create and modify configurations and

several specify the values of dimensions and
configurations 2’ the states of features
4

Check all configurations and make

Revise when corrections when needed by using
it is needed
configurationstool in the model tree
6 5

Fig. 2.36 The procedure of defining a design table in a model

Table 2.6 Types of legal design parameters in a design table (Solidworks 2018)

configurations are defined by assigning different values to design parameters. Fifthly,

the newly created configurations are previewed and checked under the configurations
tool of the model tree. If something is wrong with certain configuration, it has to be
fixed by editing the design table.
Figure 2.37 shows the procedure of defining a design table in Solidworks.
As shown in Fig. 2.37a, the design table tool can be accessed by clicking ‘Insert’
in the menu bar → ‘Tables’ in the expended list → ‘Design Table’ in the expended
list. A popup window in Fig. 2.37b shows that there are three options in creating
a design table: (1) the blank option is for an empty template, (2) the auto-create
option is to provide guides in adding dimensional variables into the table template,
and (3) the from file option provides the flexibility to import a pre-defined design
table. Figure 2.37c shows the result of selecting the auto-create option, a pop-up
window will list the majority of legal design parameters such as dimensional variables
and other model attributes. The modeler can choose the design parameters with
variants and insert them in the design table. The modeler can also insert other design
parameters that are not listed at the phase of editing. As shown in Fig. 2.37d, once
all the controllable attributes are added, the modeler can create a new configuration
one by one by adding a new row with specified values for controllable attributes in
the design table. After all, configurations are inserted, they can be previewed and
checked in the configurations of the model tree as shown in Fig. 2.37e.
68 2 Computer-Aided Design

(a). Insert a design table

(b). Three options in

creating a table

(c). Add
(e). Review configurations (d). Insert configurations parameters in table

Fig. 2.37 Interface of defining a design table in Solidworks

Example 2.2 Create a model for an optical encoder family with six levels of
resolutions as shown in Fig. 2.38.

Solution To plan the modeling process, two identified design intents are to (1) use
design equations to represent the dependences of dimensional variables and (2) use
a design table to represent all variants (e.g., configurations) of the encoder model.
Accordingly, Fig. 2.39 shows the three major steps of the modeling process. Firstly,
the basic model of an optical encoder is created. The base feature is a solid circle with
a central hole, and it is then added six pairs of features (see Fig. 2.39a). Each pair
involves a cut feature and corresponding circular pattern, and its radical dimension
and the number of the pattern instances will be renamed and treated as design param-
eters later. Secondly, design equations are defined to represent the dependences of
the dimensions of the cut features in these pairs as shown in Fig. 2.39b. Thirdly, a
design table is created that includes six configurations as the variants of the encoder
model.
Figure 2.39c shows the layout of a design table: (1) the first row is the table title
that indicates the associated ID of the model. (2) The first field of the second row (e.g.,
the highlighted black field) must not be edited since it tells the contents in the second
row are design parameters, which include the internal links to the corresponding
dimensions, features, or other properties in the model. In this example, the second row
includes (a) the dimensional parameters for inside and outside diameters and (b) the
states of cut and pattern features. (3) The fields in the first column (e.g., the highlighted
blue fields) lists the configurations. In this example, it lists six configurations for
different levels of resolutions. (4) The rest fields of the table (e.g., the highlighted
2.5 Design Parameters, Features, and Intents 69

Fig. 2.38 Design table for optical encoder family

(a). Create the model with all

controllable dimensions and features

(b). Use design equations to define the relations of

dimensions and features

Title of design table

List of controllable
variables and properties
Template of a table
design table

Variables, equations, and

states for controllable
List of configurations variables and properties
(c). create a design table for part family

Fig. 2.39 Example of creating a part model with a design table

70 2 Computer-Aided Design

yellow fields) assign a value to a dimensional parameter or a state (e.g., ‘U’ for
unsuppressed or ‘S’ for suppressed) of a feature. In addition, the design table can be
formatted by the Microsoft Excel tool.

2.5.3 Design Intents

Design intent represents (1) the purpose of a design feature, (2) the constraints that
a feature has to satisfy, (3) the way of modeling a feature, or (4) the guidance of
modifying such a feature.
Any feature on an object should be designed with purposes, so is a graphic feature
of a solid object. In feature-based modeling, it is the modeler’s responsibility to
identify design intents and the appropriate ways of representing these design intents.
Taking an example of the model in Fig. 2.18f, one design intent is the golden
ratio of the length (L) and width (W ) of rectangle, which is widely used to represent
the beauty and ergonomic appearance of real-world objects (see some examples
in Fig. 2.40). Therefore, a design intent of L/D = 1.618 has been identified and
represented as a design equation for the relation of L and D in Fig. 2.40f. Such
a design intent should be considered when the object examples in Fig. 2.40 are
modeled.
Design intents are the modeler’s understanding of the object to be modeled. From
the perspective of modeling, there are numerous methods to create the same graphic
feature, and design intents are (1) the methods selected to create certain features and

Fig. 2.40 Examples of objects with a golden ratio of their dimensions in the real world
2.5 Design Parameters, Features, and Intents 71

Fig. 2.41 Example of considering design intents in modeling process

(2) the rationales why such selections are made. Therefore, design intents should be
considered at every step of modeling procedure.
Example 2.3 Create a model for a flange in Fig. 2.41 and discuss some exemplifying
design intents in the modeling process.

Solution Other than a solid primitive, a solid consists of multiple features that will be
created, respectively. Therefore, the modeling process depends on the decomposition
of the solid into the features. There are many ways to decompose a solid into features,
and it will be very helpful to incorporate a design intent in identifying features and
planning the steps of creating these features.
Figure 2.42a shows the design intent to minimize the number of features; accord-
ingly, the modeling steps are minimized. It is appropriate when the information of
the solid is detailed and finalized, and the future change occurring to the solid will
be rare. The main disadvantages of such a design intent are (1) the complexity of
features, (2) the coupling of dimensional variables of multiple graphic attributes,
and (3) the inefficiency in making the changes on an existing model. Figure 2.43a
shows the implementation of the modeling process using this design intent. The solid
model consists of two features in the model tree; e.g., one extrude bass (F1) and one
extrude cut (F2). The right side of Fig. 2.43a shows the logical operations of these
two features.
Figure 2.42b shows the design intent to maximize the flexibility of the modeling
process; accordingly, the features are decomposed to the lowest layer. The number
of features, or the number of the steps in the modeling process, will be maximized;
in return, the advantages of such a design intent are (1) the intuitive way to plan
72 2 Computer-Aided Design

F1 \ F2

(a). minimizing the number of features:

(Solid = F1 \ F2)

F3 F8

\ F9

\ F1 F2
F6

F7
F5
F4

(b). maximizing the flexibility of modelling process:

(Solid = ((((F1 F2) F3) \ F4)\ (F5 F6)) \ ((F7 F8) F9))

Fig. 2.42 Basic design intents in planning modeling process

the modeling process, which is extremely productive in the conceptual design of

products, (2) the simplicity of creating features, (3) the highest level of reusability
of created features, (4) the independence of design features which allows to make a
change on an individual feature with the minimized influence on others, and (5) the
flexibility and expandability of making changes on an existing model. Decomposing
a solid into the features at the lowest level provides the best resolution of modeling;
therefore, we can call it the strategy for precise modeling, and we strongly recommend
the readers to consider the flexibility and reusability in planning the modeling process.
Figure 2.43b shows the implementation of the modeling process using this design
intent. The solid model consists of nine features in the model tree; e.g., two extrude
basses (F1 and F2), two extrude cuts (F5, F7), three fillets (F3, F4, F9), one chamfer
(F8), and one circular pattern (F6). The right side of Fig. 2.43b shows the logical
operations of these nine features to represent the solid with all these features.
Whenever there are some options in the modeling process, design intents should
be considered to select an optimal option. Taking an example of creating the first
solid feature from a 2D sketch in Fig. 2.43a or Fig. 2.43b, the modeler has to select
the sketch plane, it will be helpful to consider that the first sketch-plane determines
the orientation of the solid with respect to the default world coordinate system.
2.5 Design Parameters, Features, and Intents 73

Fig. 2.43 Implementations of modeling process

Accordingly, it determines the contents of the standardized front view, right view,
and top view. Note that the front view is usually the main view to include the main
dimensions of the solid. Therefore, the design intent at the first step of modeling
should ensure that the first sketch plane includes the most significant dimensions of
the solid.
In a graphic model of solid, the positions of orientations of graphic features are
determined with respect to a specified coordinate system. A default coordinate system
in Solidworks is shown in Fig. 2.44. It consists of an origin (O), and three planes for

Default planes Design Intent: align

and origin in a the solid origin and
part template orientation with the
default coordinate
system

Fig. 2.44 Default coordinate system to locate and orientate a graphic feature
74 2 Computer-Aided Design

(a). use the front or right plane as the

sketch plane for the first revolve

(b). use the top plane as the sketch (c). ensure the front plane includes
plane for the first extrude main dimensions

Fig. 2.45 Design intent in selecting the first sketch plane

front (X-Z), top (X-Y ), and right (Y-Z) views. Three axes (X, Y, Z) of the coordinate
system are dependent, which can be derived from the aforementioned information.
In general, it is beneficial to align the original and orientation of the solid with the
default coordinate system; so that the position and orientation of the solid can be
known and easily determined at the beginning of modeling process. To implement
this design intent, Fig. 2.45a shows that a right or front plane should be selected as
the sketch plane when the first feature is revolved. Figure 2.45b shows that a top
plane should be selected as the sketch plane when the first feature is extruded. In
addition, the sketch should be centralized at the origin.

2.6 Modeling Procedure

In feature-based or knowledge-based modeling, a solid is decomposed as a set of

features, which can be modeled, respectively. The attributes of a feature to be defined
are geometry and shape, dimensions, constraints, and operator. The solid model
is then created by applying logical operations (e.g., union ∪, intersection ∩, and
difference \) on the graphic features. As shown in Fig. 2.46, the constitutive features
of the solid are identified, created, and then added or removed from the solid volume
using logical operations. For the geometry and shape, a graphic feature can be either
of (1) a built-in feature fillet, chamfer, rib, or hole or thread wizard or (2) a sketched
feature such as extrude, revolve, loft or sweep. The modeling tool for a built-in
feature provides a wizard for the modeler to input geometric data; while a sketched
feature requires the modeler to define 1D or 2D dimensions in sketches. Note that
dimensioning a graphic feature requires to select or create certain references such
as points, axes, planes, and coordinate systems. The corresponding tools for feature-
based modeling are grouped in Fig. 2.46 and illustrated in Fig. 2.47 correspondingly.
2.6 Modeling Procedure 75

Fig. 2.46 The classification of feature-based modeling tools based on their roles on feature attributes

Fig. 2.47 Feature-based modeling tools in Solidworks

Figure 2.47 shows some major feature-based modeling tools in Solidworks, and
these tools are associated with the need of creating and modifying different features
and attributes in Fig. 2.46. Modelers should be aware of these tools and know-how
to access when they are needed. As shown in Fig. 2.47, (1) the model tree is used
to represent the features and their relations in a solid; the information of any entity
can be accessed using the interface in the model tree; (2) the feature manager in the
second row shows available tools to create different features; (3) logical operators
76 2 Computer-Aided Design

in Fig. 2.46 are functioned by selecting feature types and specifying appropriate
references; (4) different graphic features (e.g., built-in or sketched) in Fig. 2.46
correspond to different modeling tools in the feature manager; (5) the sketches, the
dimensions, and the constraints in Fig. 2.46 are defined mainly by the tools in the
sketch manager and other tools for parametric modeling such as linked values, design
equations, and design tables; (6) the references in Fig. 2.46 are defined and managed
using the reference geometry manager; in editing a model, any existing graphic
elements (vertices, edges, and planes) can be used as the references to modify or
create new features; (7) to create high-level built-in features, the Solidworks includes
many adds-in tools such as weldment, fastening, mold, and surface tools as the tool
wizards.
Feature-based or knowledge-based modeling is very effective to deal with the
complexity and changes in product designs. The principal strategy is ‘divide and
conquer’; a solid is modeled as a set of graphic features, and each feature is modeled
or modified at a time to minimize its impact on other features. Figure 2.48 shows a
routine procedure to define a graphic feature in the modeling process. It begins with
the determination of design intents; note that a design intent represents the selection
of modeling strategies for the given feature. For example, a cylinder can be created

Decide design intents

for a given feature

Built-in feature Sketched feature

Is it a built-in
feature? No
Yes
Select or create a
sketch plane

Adjust design Activate the Create sketches with defined

Intents feature tool Create next geometry, dimensions, and
sketch relations

No
Select entities on existing All required sketches
features in feature are made?
property window
Yes
Activate the
feature tool

Specify dimension(s) and settings

in feature property window

Confirm the
creation of feature

No Yes
Does the created feature End
meet needs?

Fig. 2.48 Divide and conquer: one feature at a time in knowledge-based modeling
2.6 Modeling Procedure 77

by extruding, revolving, even lofting or sweeping, the modeler should choose the
simplest tool in modeling. Note that a design intent is the modeler’s choice when
there are a number of modeling options. As discussed early, a graphic feature can
either be of built-in feature or sketched feature. A built-in feature such as a fillet
and surface thickness can be created upon an existing feature, and the inputs can
be specified directly in the feature property window. A sketched feature such as
an extrusion requires to create one or more sketches to define the geometry and
dimensions. For complex solid, specifying an appropriate reference for sketching
and dimensioning is not always a trivial task; in many cases, new references should
be specially defined to define the position and orientation of the feature subjected to
certain constraints.

2.7 Assembly Modeling

With few exceptions of simple products, discrete products are assembled from parts
and components due to many different reasons: (1) parts or components are from
different materials, through different processes, or made at different locations or
facilities; (2) sizes of parts are too large to be processed as a whole on machine tools;
(3) finish products are large-scale that are inappropriate for transportations; (4) prod-
ucts are machines that have relative motions of their components; (5) product variants
are made by selecting different modules and assembling them in different ways (see
Fig. 2.49 for examples). Therefore, engineering products are mostly assembled units.
On average, 50% of the total manufacturing cost is tied to the assembly processes
of products; manufacturing enterprises must know how to reduce the cost of assembly

(a). parts from different (b). products are too large to be

materials and processes processed as a whole

(e). product variants from a

modularized platform

(c). parts have to be transported (d). machines that need relative

and assembled on sites motions among components

Fig. 2.49 Products are assembled for many reasons

78 2 Computer-Aided Design

processes to gain their competitiveness. In early time, the designs of assembly

processes were less emphasized than product designs. This might lead to the situation
that some problems relevant to assembly processes were discovered too late and too
expensive to be fixed. If an omission or error relating to assembly processes can be
identified and fixed at the product design stage, the enterprise can save significant
costs.
Design for assembly (DFA) is a design methodology by which the constraints
and implementation of assembly processes are taken into consideration at the design
stages of products. The primary goal of DFA is to reduce the cost, the time, and
the complexity of assembly processes, and two basic approaches of DFA are (1)
to simplify product structures by reducing the total number of parts in a product
and (2) optimize the features on the product for easy grasping, moving, orienting,
inserting, and fitting (Wikipedia 2020b). As shown in Fig. 2.50, Boothroyd (1994)
gave the examples of using DFA for the motor-drive assembly and reticle assembly;
the numbers of parts and assembly operations in these products were significantly
reduced. Therefore, the costs of these products were greatly reduced due to the
simplification of assembly processes via DFA.

DFA practice

Origin DFA
design
Number of parts 19 7
Assembly time 160.0 46.0
(in seconds)
Total cost (US $) 35.44 21.73

(a) Motor assembly

DFA practice

Origin design DFA

Number of parts 47 12
Assembly time
(in hours) 2.15 0.33

Number of assemblies 58 13

(b) Reticle assembly for thermal gunsight

Fig. 2.50 Examples of using DFA in product designs (Boothroyd 1994)

2.7 Assembly Modeling 79

2.7.1 Terminologies in Assembly Modeling

Other than geometric and feature modeling of products, the functional under-
standing of assemblies is crucial to clarify product designs (Gui and Mantyla 1994).
Assembly modeling aims to represent and visualize the assembly relations of parts
and components in a product model; the parts and components are presented as solid
models.
An assembly model consists of a set of components for the desired functions
and the connections that apply constraints on components. Generally, an assembly
model uses a hierarchical structure for the connections of components. To model the
assembly of a product, one has to understand the following terminologies.
(1) Degrees of Freedom (DOF)

Degrees of freedom are a number of independent variables to describe the position

and orientation of a physical body in space. As shown in Fig. 2.51a, a free body in a
3D space has six DOF since three independent variables (T x , T y , and T z ) are used to
describe the translations along three axes (X, Y, and Z), and other three independent
variables (Rx , Ry , and Rz ) are used to describe the rotations around three axes (X,
Y, and Z). As shown in Fig. 2.51b, a constraint will reduce DOF of a body since it
confines one or a few DOF of the body. For example, if the body has to be on a plane
(e.g., the XY-plane), the constraint will apply to the free motions of Rx , Ry and the
remained DOF of the body on the plane will be three DOF (e.g., T x , T y , and Rz ).
(2) Assembly Mates and Constraints

Assembly mates create geometric relations of components. When an assembly mate

is added to two components, their relative motions along one or a few directions are
confined, and the rest of the unconfined motions are referred to as allowable motions.
Therefore, the mates in an assembly model determine the DOFs of the components.
A product without any allowable relative motion is called a structure; a product with
some allowable relative motion is called a machine, and the number of allowable
DOF is referred to as the DOF of the machine.

Tz Ty
Ty
Tx

Ry RZ Rx RZ Tx

(a) 6 DOF in three-dimensional space (b) 3 DOF in two-dimensional space

Fig. 2.51 DOF of a free body in 3D or 2D space

80 2 Computer-Aided Design

Table 2.7 Possible mating relations of two components with points, edges, and faces
Component I
Point Edge Face
Component II

Point A D E
Edge D B F
Face E F C

An assembly mate involves two or more graphic elements from two components.
Note that three basic types of graphic elements are points, edges, and faces; Table 2.7
shows six possible mating types (A, B, C, D, E, and F) of the graphic elements from
two components (components I and II).
Assume that component I is fixed, Table 2.8 shows the numbers of confined and
allowable DOF of component II when the aforementioned mating types are applied,
respectively.
In a computer model, a mating relation is unnecessary to have a physical contact.
Therefore, a broad scope of mating relations has been introduced to confine the rela-
tive motions of two or more components. The types of mates are classified as standard
mates, advanced mates, and mechanical mates. Tables 2.9, 2.10 and 2.11 show the
types of standard mates, advanced mates, and mechanical mates, respectively.
(3) Root Components

In an assembly model, a root component determines the position and orientation of

the assembly in a reference coordinate system. For simplicity, a root component is
‘fixed’ in the reference coordinate system by defaults. A root component usually has
a relatively large size, and importantly, has more mates with others in comparison to
other objects in an assembly. Therefore, a root component should be selected based
on the significance of a component in defining assembling relations with others.
Figure 2.52 shows some examples of root components in respective products. In
these examples, root components are highlighted by ‘blue’ color.

2.7.2 Modeling Methods

Products can be modeled in two methods, i.e., (1) bottom-up methods and (2) top-
down methods.
Bottom-up methods are more traditional. Firstly, parts are designed and modeled;
secondly, parts are inserted into an assembly model, and mates are defined among
the parts to determine the spatial relations of parts in the assembly model. Individual
parts should be edited at the part level, and the changes occurring to parts to be
updated in the assembly model automatically.
Table 2.8 Confined and allowable DOF of components subjected to different mating relations
Mating type Degrees of Freedom (DOF) Illustration
Confined Allowable
A (point-point) T x , T y , T z Rx , Ry , Rz
Z Z

A
2.7 Assembly Modeling

O
O
X
X

B (edge-edge) T x , T z , Rx , Rz T y , Ry Z Z

O
O
X
X

C (face-face) T y , T z , Rx Ry , Rz , T x Z
Z
C

O
O
X
X

(continued)
81
Table 2.8 (continued)
82

Mating type Degrees of Freedom (DOF) Illustration

Confined Allowable
D (point-edge) The point must be one edge, which confines two of (T x , T y , T z ) T x , Rx , Ry , Rz Z
Z
D

O
O
X
X

E (point-face) The point must be one face, which confines one of (T x , T y , T z ) T x , T y , R x , R y , Rz Z

Z
E

O
O
X X

F (edge-face) Z Z
F

O O

X
X
2 Computer-Aided Design
2.7 Assembly Modeling 83

Table 2.9 The types of standard mates

Type Symbol Applicable graphic elements
Coincident Point-point, point-line, point-plane, line-plane, plane-plane

Parallel Line-line, line-plane, plane-plane

Perpendicular Line-line, line-plane, plane-plane

Tangent Arc-line, arc-arc, arc-cylinder, cylinder-cylinder, plane-cylinder

Concentric Arc-arc, arc-cylinder, cylinder-cylinder

Lock Any graphic element

Table 2.10 The types of advanced mates

Type Symbol Applicable graphic elements
Symmetric Two points/lines about a line, two points/lines/planes/any objects
about a plane

Width One or two planes in the middle of two other planes

Path mate Point-path, point-trajectory

Line/line coupler Two parallel planes

Table 2.11 The types of mechanical mates

Type Symbol Applicable graphic elements
Cam Cam: a set of faces cam follower: a face or vertex

Hinge Arc-arc, arc-cylinder, cylinder-cylinder

Gear Axis-axis, axis-cylinder, cylinder-cylinder

Rack pinion Rack: edge, sketch line, centerline, axis, cylinder

pinion: cylinder, arc, axis, revolved surface

Screw Axis-axis, axis-cylinder, cylinder-cylinder

Universal joint Axis-axis, axis-cylinder, cylinder-cylinder

84 2 Computer-Aided Design

Root component Root component

(a). board for circuit (b). base for machine tool

Root component
(e). base for manipulator

Root component Root component

(c). chassis for vehicle (d). shaft for engine

Fig. 2.52 Examples of selecting a root component

A bottom-up design method is suitable for the scenario that a product is made
of previously constructed, off-the-shelf parts, and standard components like hard-
ware, pulleys, motors, etc. These parts are standardized; they should not be changed
in engineering designs; instead, engineers select different components when the
requirements are changed. In using a bottom-up method, firstly, the parts at the
lowest level are modeled or obtained. Secondly, an assembly model is created and the
root component is selected, inserted, and placed in the reference coordinate system.
Thirdly, other part models are inserted in sequence; when a part is inserted, its spatial
relations and constraints with existing parts are defined by mates. The assembly
model is completed when all parts are inserted and the mates among the parts are
fully defined. Using a bottom-up method allows engineers to focus on the details of
the assembly relations since parts and components are modeled individually.
Example 2.4 Create a 6-DOF modular robot configuration from a modular robotic
system. As shown in Fig. 2.53, the modular robotic system consists of standardized
modules including rotary joints, linear joints, links, wrists, and grippers (Bi et al.
2008).

Solution As shown in Fig. 2.54, three critical design issues are involved in the
application of a reconfigurable system. These design issues are architecture design,
configuration design, and control design.
Architecture design determines the types of system modules and the connections
of system modules. Modules in a reconfigurable system are encapsulated; in other
words, the internal implementation of a module does not affect the interfaces of the
module with others. The connections of a module are the options by which the module
can be interacted with others. The design of system architecture aims to produce
2.7 Assembly Modeling 85

(a). gripper types

(b). link types

(c). wrist types

(d). rotary joint types

(e). linear joint types

Fig. 2.53 An example of a modular robotic system

Fig. 2.54 Task-oriented design of modular robotic systems (Bi et al. 2008)

as many system configurations as possible with a given set of modules. Note that
different configurations can be used to fulfill different tasks; the more configurations a
system can generate, the better capabilities of the reconfigurable system can deal with
changes and uncertainties in a dynamic environment. Architecture design is involved
at the phase of reconfigurable system design. In configuration design, it is assumed
86 2 Computer-Aided Design

that the reconfigurable system is given; for example, the types and the numbers of
robotic modules in Fig. 2.52 are given. Configuration design is to select a set of
modules and configure modules into a robot to fulfill the functional requirements
of a given task optimally. Configuration design is involved at the phase of system
application. Active modules in a reconfigurable system have their local controls;
however, there are system-level goals when modules are assembled into a robot.
Therefore, control design is to coordinate system modules, so that these modules can
collaborate with each other to fulfill given tasks satisfactorily. The control design is
involved at the phase of system operation (Bi et al. 2008).
Since the problem does not specify the functional requirements of the 6-DOF
robot in terms of the workload, trajectory, velocity, and acceleration of toolpath,
arbitrary three 1-DO joints are selected to build a 6-DOF model. As shown Fig. 2.53,
the demonstrated robot has 6 DOF; it consists of 1 R-90 rotary joint, 1 L-70 linear
joint, 2 R-70 rotary joints, 1 W-70 wrist, 1 link module for the connection of R-90
and L-70 modules, and 1 angled link module for the connection of L-70 and R-70.
Note that a W-70 has 2 DOF. In the robot assembly, each active module has one or
2 DOF which are driven by respective motors (Fig. 2.55).
In a top-down assembly method, parts and components are created during the
course of assembly modeling. The details of part models are not available when the
assembly relations are defined, and constitutive parts or components are created one
by one based on the conceptualized structure and assembly relations of products.
Note that even though parts are created during the course of assembly modeling,
they can be saved, either internally or externally, as individual models. Top-down
modeling allows engineers to utilize geometric relations and constraints in the

Fig. 2.55 A 6-DOF configuration of the modular robot system

2.7 Assembly Modeling 87

high-level structure for newly created parts. In such a way, geometric modeling
and assembly modeling can be proceeded simultaneously. Engineers can view the
assembly relations of the part when the part is modeled.
A top-down method reduces the rework when the assembly relations are changed
in a product design since the derived parts are associated with the assembly constraints
when they are modeled. Therefore, top-down modeling is very useful at the concep-
tual design stage; it is widely used in tooling design since the geometries and spatial
arrangement of tooling depend on the parts to be manufactured. In practice, a top-
down method can be used to create an assembly model partially, i.e., a few critical
parts in assembly or some key features of parts. Engineers can use the top-down
modeling method to layout an assembly including key parts customized to assembly
relations (Dassault Systems 2020).
Example 2.5 A customer wants to have a dining table with the features shown in
Fig. 2.56: (1) the table takes a footprint of a circular area of ∅ 2000 mm; (2) the table
is supported by a hexagonal frame that serves for six peoples; (3) the table has an
overall height of 1200 mm and stands on four legs; (4) each side of the table has a
drawer with the height of 200 mm; (5) steel materials are used for all support parts
and wood materials are used for the parts with a large surface and the parts for six
drawers. Use top-down modeling to design a dining table to meet the aforementioned
requirements.

Solution The top-down method is used to create six parts for a dining table. The main
dimensions of the parts are determined based on the drawing of the conceptual design

Metal frame for

the supports

Wood or glass
materials for two
large plates

An overall diameter
with a height of ~
1200 mm

Leg’s sizes with a

height of 1000
Drawer’s sizes with mm
a height of 200 mm

Fig. 2.56 Requirements of new dining table in top-down design example

88 2 Computer-Aided Design

in Fig. 2.56. As the result, Figs. 2.57 and 2.58 show the exploded and collapsed views
of the assembly model, respectively. The bills of materials (BOM) table in Fig. 2.57
shows the dining table is assembled from 17 parts.

Fig. 2.57 Top-down modeling for the design of a dining table

Fig. 2.58 Collapsed assembly model of dining table in Example 2.6

2.7 Assembly Modeling 89

2.7.3 New Assembly-Level Features

In assembly modeling, mates are defined to represent the spatial relations of parts in
upper level components. To meet assembly constraints, new geometric or reference
features can be created for parts or components at the assembly level; such features
are called assembly-level features. An assembly-level feature differs from a feature in
a part model in sense that the references, sketches, and dimensions of an assembly-
level feature are associated with the assembly model. Assembly modeling supports
the creation of assembly-level features. Figure 2.59 shows the modeling tools of
SolidWorks in the assembly commend group, and the following tools relate to the
assembly-level features: (1) Insert Component. A part or component is added into
the assembly model; the part can be newly created in the assembly model using the
top-down method; in other words, the new part at least takes a reference plane from
the assembly model as its sketch plane for the first solid feature as the correspondence.
(2) Component Pattern. A linear or circular pattern is defined by selecting one or
a group of parts and specifying the number and spacing of the pattern. (3) Smart
Fasteners. A standard faster is defined for a pair of selected parts at their contact
surfaces. (4) Assembly Features. The tools are used to (i) create geometric features
relevant to assemblies such as hole series, hole wizard, simple hole, extruded cut,
revolved cut, swept cut, fillet, and chamfer, and (bi) define an assembly relation such
as weld bead and belt/chain.

Example 2.6 Create a new part in Fig. 2.60a to connect the handle to the body of
the light.

Add a new or Create a linear or Create a standard Add different

existing part or circular pattern of fastener at types of assembly
component part or component interface features

Assembly
commend
group

Fig. 2.59 Interface for adding assembly-level features

90 2 Computer-Aided Design

Need a new part to

connect the handle
to the light body 1
2

(b) Insert a new part in assembly

5 3

(e) Use two existing faces as the (d) Use an existing edge to convert
ends of the extruded pin it as the diameter of the pin

Fig. 2.60 The procedure of adding a new part in the assembly process

Solution Figure 2.60b–f illustrates the steps to insert a new part using the information
of other parts in the assembly model (Fig. 2.60a). Firstly, the new part option under
the insert tool is activated to begin the process of creating a new part in the assembly
model. Secondly, the symmetric plane of the light body is selected as the associated
sketch plane of the first feature of new part. Thirdly, the hole profile of the handle
is projected on the sketch plane and converted as a circle in the sketch. Fourthly, an
extruded feature is created by specifying the newly created sketch and two sides on
the handle as the bounds on two extrude directions, respectively. Finally, a new part
is fully defined, and the inserting process is terminated and returned to the assembly
model with a newly created pin.

Example 2.7 A characterized human chest model is modified to investigate the

impact of sternotomy wires on the transmission of ultra-wideband (UWB) in a wire-
less body area network (WBAN). Figure 2.61a shows the original model that consists
of the layered models of cloth, skin, fat, muscle, sternal, bone, and heart in chest; note
that different layers correspond to different dielectric properties in the simulation.
Figure 2.61b and c show the models of wires and heart valve that will be mounted on
sternal and heart, respectively. Modify the simulation model by inserting wires in the
chest model so that it can be used for the numerical simulation of signal transmission
(Särestöniemi et al. 2019).
2.7 Assembly Modeling 91

(a) chest assembly model

(original model)

(b) wire model (d) Chest assembly model

(c) heart value model (modified model)

Fig. 2.61 Example of eliminating interferences in assembling (Särestöniemi et al. 2019)

Solution In an assembly model, the interference of solid objects must be elimi-

nated. When a new part is inserted, possible interferences should be detected. When
an interference is identified, it can be eliminated by adding a cavity feature at the
assembly level. Figure 2.61d shows the assembly model is modified in such a way
that the cavity features are created in the part models of heart, sternal, fat, bone, and
muscle; these cavities are used to accommodates the heart valve and wires without
interference.

2.7.4 Exploded Views and Bill of Materials

An assembly model represents constitutive parts and their spatial relations in a

product or system. When the number of constitutive parts increases, it becomes diffi-
cult to view the assembly relations in one configuration. Defining exploded views
provides a useful way for engineers to look into the assembly relations in multiple
configurations. In each configuration, parts can be moved and placed at exploded
positions, and the movement traces can be modeled as exploded lines to represent
assembling paths. Moreover, the exploding steps can be animated and recorded so
engineers can review the assembling processes of a product vividly. As shown in
Fig. 2.62, the tools in assembly modeling also provide the function of creating the
bills of materials (BOM) automatically, The BOM list all types of parts as well as the
numbers of each part type. Exploded views help engineers to visualize the assembly
plan for verification and validation purpose effectively.
92 2 Computer-Aided Design

Tools to create a table of Tools to create an

bill of materials (BoM) exploded view

Exploded views as
configuration

Automatically generated
exploding lines

Bill of materials (BoM)

Exploding
steps

Fig. 2.62 Interface to create an exploded view

Figure 2.63 shows an example of an engine assembly model. Figure 2.63a is a

collapsed configuration with all mates among the parts, and Fig. 2.64b is an exploded
view to show all parts as well as the parts for assembling processes. The exploded
view helped to make an automated assembly plan (Yu et al. 2014).

(a) Collapsed model

(b) Exploded model

Fig. 2.63 Example of exploded view—cylinder gear reducer (Yu et al. 2014)
2.8 Kinematic and Dynamic Modeling 93

(a). Binary link (c). Ternary link

(b). Quaternary link

(d). Hexagon link (e). Pentagon link

Fig. 2.64 Classification of links

2.8 Kinematic and Dynamic Modeling

Modern machines or products are mostly mechatronic systems; however, they are
built upon mechanical systems. In designing a mechanical system, mechanics investi-
gates the kinematic and dynamic behavior of a mechanical system, which is subjected
to mechanical loads such as displacement constraints and driving forces.
A mechanical system consists of a set of links that are connected by joints.
A mechanical system has a special link called an end-effector to perform tasks.
Assembly of links and joints ensures that the end-effector can move with specified
degrees of freedom (DOF). To model a mechanical system, links and joints are firstly
represented adequately.

2.8.1 Link Types

In a mechanical system, any rigid body in the structure is modeled as a conceptual

link, and the mechanical structure is the assembly of links and joints. In the machine
design theory, links are classified based on the number of connections to other objects.
As shown in Fig. 2.64, a conceptual link is called a (a) binary link, (b) ternary link,
(c) quaternary link, (d) pentagon link, and (e) hexagon link if the link has the number
of connections as 2, 3, 4, 5, and 6, respectively. Note that a connection of a link to
other objects implies that additional motion constraints are added to constrain the
motion of the link relative to others.
94 2 Computer-Aided Design

2.8.2 Joint Types and Degrees of Freedom (DOF)

The motion of an object or a system is represented by degrees of freedom (DOF) of

motion. DOF are the number of minimal and independent variables that are required
to describe the position and orientation of an object or system in space at any instant
of time.
As shown in Fig. 2.65a, a free body in the 3D space possesses six DOF; since
the body can be translated along X-, Y-, Z-axes and rotated along X-, Y-, Z-axes,
respectively. As shown in Fig. 2.65b, a free body in a 2D space possesses three DOF
since it can be translated along X- and Y-axes and be rotated along Z-axis; the rotating
axis is always perpendicular to the plane O-XY for the X and Y translation.
When two links are joined, some constraints apply to the jointed links. The type
of joint determines the degrees of motion to be confined. Figure 2.66 shows six joint

Y Y

X X

Z
(a). 6-DOF for an object in 3D space (b). 3-DOF for an object in 2D space

Fig. 2.65 Free object and motion degrees of freedom (DOF) in 3D and 2D spaces

(a). Prismatic joint (b). Revolute joint (c). Screw joint

(d). Cylindrical joint (e). Universal joint (f). spherical joint

Fig. 2.66 Classification of joints and the degrees of freedom of motion

2.8 Kinematic and Dynamic Modeling 95

types where the direction(s) of unconstrained motion(s) are illustrated where the rest
of the directions are constrained motions.
A prismatic joint in Fig. 2.66a has one translational motion; a translation or
rotation along any one of five other directions is fully constrained. A revolute joint in
Fig. 2.66b has one rotational motion. A screw joint in Fig. 2.66c allows the translation
and rotation along the same axis simultaneously, but these two motions are coupled
and the joint only has one DOF. Differing from a screw joint, a cylindrical joint in
Fig. 2.66d has one transition and one rotation along the same axis, while these two
motions are independent; the joint has two DOF. A universal joint in Fig. 2.66e and
a spherical joint in Fig. 2.66f have two and three rotations without any translations,
respectively.

2.8.3 Kinematic Chains

A mechanical system without the consideration of its energy source and ground
component is called a kinematic chain. A kinematic chain specifically refers to the
topology of the assembly of rigid bodies (or links) by joints.
Figure 2.67a shows an open-loop kinematic chain where links carry one upon
another in a series. Due to a few constrained motions occurring to links, an open-
loop kinematic chain usually has a large motion range but with the limited rigidity
to carry external loads. In addition, any link or joint brings new sources of error; the
errors from these links and joints are stacked up linearly in an open-loop kinematic
chain. Theoretically, an open-loop chain has a relatively low accuracy.
Figure 2.67b shows a closed-loop kinematic chain that includes two or several
special links with more connections to others. These links are connected to a group of
others in parallel. The load on a multi-connected link is shared by a set of connected
links. Therefore, a closed-loop kinematic chain has the better capability to carry the
external load. In addition, since the same link is connected with others in parallel, the
errors at joints are averaged instead of stacked. Therefore, a closed-loop kinematic

(a). open-loop (b). closed loop (c). Combined chain

Fig. 2.67 Three types of kinematic chains

96 2 Computer-Aided Design

chain is expected to have a better motion accuracy; however, it has a relatively small
range of motion due to the constraints by multiple links.
Figure 2.67c shows a hybrid kinematic chain that includes both open-loop or
closed-loop chains. A hybrid chain makes the tradeoff between the loading capability
and the range of motion. The closed-loop sub-chains are used where the loads are
large to the machine, and the open-loop sub-chains are used where the machine needs
a large accessible space for given tasks.

2.8.4 Mobility of Mechanical Systems

When the links, the joints, and the assembly topology are given, the mobility of
a mechanical system can be determined. The mobility of a mechanical system’s
mobility (M) is quantified as the number of DOF of system.
DOF of a mechanical system is defined with respect to a selected reference frame
called a ground reference frame as,


j
M = λ(l − j − 1) + fi (6.1)
i=1

where
M be degrees of freedom (DOF) of system,
l is the total number of links, including fixed link,
n is the total number of joints,
f I is the degree of freedom of relative motion between element pairs of i-th joint,
and
λ is an integer λ = 3 for a plane mechanism and λ = 6 for a spatial mechanism.
If a mechanism is planar, and all of the joints are low-pair joints (prismatic joint
or revolute joint), Eq. (6.1) can be simplified as the Gruebler’s equation as

M = 3(l − 1) − 2 j (6.2)

where
M be degrees of freedom (DOF) of a planar mechanism,
l is the total number of links, including the fixed link, and
j is the total number of low-pair joints.
In a machine, independent motions occur to active joints, and active joints are
driven by motors. The motions of active joints are transferred to the end-effector
link where the task is performed. Figure 2.68 shows some simple machines with
the driving motion (in red) and the driven motion at the end (in black). All of the
example machines except Fig. 2.68g have 1-DOF input and output. It implies that
2.8 Kinematic and Dynamic Modeling 97

(a). Package mover on assembly bench (b). Can crusher (c). Simple press

(d). Car window wiper (e). Microwave carrier to assist (f). Lift platform in truck
people on wheelchair

(h). Box closer (i). Mechanism on exercise

(g). Front loader machine

Fig. 2.68 Examples of simple machines

many feasible solutions are available to satisfy the same motion requirement in the
application.
At the conceptual design stage, the designer should be able to analyze degrees of
freedom of motion when the assembly model of a machine is given.

Example 2.8 Evaluate the DOF of the mechanism shown in Fig. 2.68a.

Solution As shown in Fig. 2.68a, the mechanism is a planar mechanism. Let the DOF
of a link in a plane be = 3. The mechanism includes six links, and seven joints, i.e., l
= 6 and j = 7. All joints are either of 1-DOF rotational or translational. Accordingly,
f i = 1 for (i = 1, 2,… 7). Using Eq. (6.1) finds that,


j
M = λ(l − j − 1) + f i = 3(6 − 7 − 1) + 7(1) = 1 (6.3)
i=1

Machine design is a complicated process. With the computer implementation

of the aforementioned kinematic and dynamic modeling methods, computer-aided
design tools are able to assist engineers to design and optimize a machine for the
expected functional requirements. Figure 2.69 shows that engineers could interact
and fully utilize the capabilities of computer-aided design tools in virtual machine
98 2 Computer-Aided Design

Fig. 2.69 Main functional modules for virtual machine design in SolidWorks

design (Markkonen 1999). Firstly, a design process begins with the preparation of
a CAD model of a machine including geometries, dimensions, material properties
of parts and components, mating relations in the assembly model, and boundary
conditions in the application environment. Secondly, a simulation model is defined
to establish a mathematical model for kinematic and dynamic behaviors of system.
Thirdly, model parameters are specified; main modal parameters include the proper-
ties of motors, the profiles of expected motions, the duration of simulation, and design
variables to be investigated. Fourthly, the simulation is performed to find the solu-
tion to the formulated mathematic model numerically. Fifthly, the simulation result is
analyzed and verified, and an iterative process is repeated to precedent steps until the
design and analysis goal is achieved. It should be noted that even CAD tools are avail-
able to take over many critical tasks, engineers’ involvement in the computer-aided
design process is essential to a successful machine design. For a machine design at
different stages, design scopes and goals are different, and different design tools are
needed at different stages. SolidWorks provides a comprehensive toolset to support
virtual machine design.

2.8.5 Motion Simulation

SolidWorks Motion Analysis is to create a simulation model to study the position,

velocity, acceleration, and torque of a mechanism, which is subjected to external
loads. Virtual motion analysis brings significant benefits to machine design. (1) The
number of physical prototypes can be minimized since the simulation will be able
to identify most of the potential design errors and omissions. (2) Virtual analysis
takes much less time than the experiments. It supports parametric studies to opti-
mize the design before it is prototyped; this allows exploring more design options
2.8 Kinematic and Dynamic Modeling 99

at a very early stage. (3) Simulation-based optimization can be utilized to improve

machine design. (4) The simulation is integrated with all the other engineering anal-
yses; it becomes possible to gain the quantified insights for additional engineering
analysis and investigate the design feasibility over the entire product lifecycle. In the
following, the main steps of a motion simulation are discussed.

2.8.5.1 Model Preparation

The first step for motion simulation is to prepare a CAD model of machine. The
machine model includes all constitutive parts, and the connections of the parts are
presented in a simplified form. In addition, the materials are specified for all solid
bodies, so that the strengths of parts are determined in terms of allowable stresses.
The forces acting on parts in the process of machine operation are then determined
based on the required motion of machine.
When all parts are modeled, the next step is to model the assemblies of parts.
If two parts have a static spatial relation in the machine, these two parts should be
grouped since no relative motion is allowed between them. If two parts involve in a
relative motion with each other, a correct type of joint must be selected. The assembly
relations of parts are modeled as mates in assembly modeling. As shown in Fig. 2.70,
the mates in the SolidWorks are catalogued into (a) standard mates, (b) advanced
mates, and (c) mechanical mates. If two parts are bonded, the mates of these two
parts are likely a combination of a few of standard mates in (a). Advanced mates
are applied in a scenario where more than two entities are involved (e.g., symmetric
and width) or a coupling of two motions occurs (e.g., path mate and linear coupler).
Mechanical mates are special mates to represent the motions of typical machine
elements.
Figure 2.71 shows an assembly model example of a Yumi robot for motion simu-
lation. The robot has two mechanical arms, and each arm has a 7-DOF motion. Each
DOF is enabled by an active rotary joint. In the assembly model, each DOF is modeled
as a combination of one ‘coincident’ mate of two planes and one ‘concentric’ mate
of two cylindrical surfaces from two joined parts.

Distance
mate
Angle
mate
(a). standard mates (b). Advanced mates (c). Mechanical mates

Fig. 2.70 Mates in assembly modeling of SolidWorks

100 2 Computer-Aided Design

(a). 14-joints (ABB 2018) (b). Assembling modeling

Fig. 2.71 Assembly modeling of ABB yumi robot

2.8.5.2 Creation of Simulation Model

SolidWorks has three functional modules for a motion study: Animation, Basic
Motion, and Motion Analysis. The animation tool simulates the kinematic behav-
iors of models without the consideration of dynamics. Users can use the animation
to visualize possible motions using an assembly model. Both of the basic motion and
motion analysis tools are used to simulate kinematic and dynamic behaviors with
the consideration of dynamic properties and driving forces but with different levels
of calculation accuracies. Galliera (2010) gave the comparison of three simulation
tools as shown in Table 2.12.
As shown in Fig. 2.72a, the Motion Analysis tool is included in ‘Add-Ins’ in
‘Options’ of SolidWorks. It is not loaded as default; therefore, a user has to activate
the tool before it can be accessed. A new motion study can be created by right-
clicking the motion study tab as shown in Fig. 2.72b. After the motion analysis has
been activated, the list of the options under ‘animation’ includes Animation, Basic
Motion, and Motion Analysis.

2.8.5.3 Define Motion Variables

As shown in Fig. 2.73, a motion study includes a set of motion variables for motor,
spring, damper, force, contact, and gravity. These motion variables should be defined
for the machine to be simulated. For example, a set of motors must be defined for all
the active joints in a machine. In addition, quite a few properties have to be specified
when a motor is defined in the motion analysis.
Figure 2.74 shows the interface in defining a motor in the motion study. Firstly,
the motion of a motor can be translational or rotational. Secondly, the motion is
2.8 Kinematic and Dynamic Modeling 101

Table 2.12 Comparison of SolidWorks Animation, Basic Motion, and Motion Analysis
Types Solvers Description

Animation 3D Dimensional The 3DDCM solver is capable of positioning parts in an assembly model or in a
Constraint mechanism. The animation can be used to build, modify and animate the
Manager assembly model mainly for the visualization of the changes occurring to
(3DDCM) by geometries, appearances, dimensions, and constraints such as mates. An
D-Cube animation can be defined as a smooth interpolation of multiple static views, or
referred as keys, in a given animation time.

Basic Ageia PhysX The Ageia PhysX is a physics solver primarily for the animations in games. The
Motion basic motion tool simulates how objects behave, move and react for life-like
motion and interaction. Adopting the Ageia PhysX in the basic motion makes the
simulation look realistic but the motion is not precise.
The basic motion tool is capable of approximating the functions of motors,
springs, collisions, and gravity effect. It is physics-based, which allows updating
the simulation quickly with less computation. It suits the best for the presentation-
worthy animation.

Motion ADAMS solver The ADAMS solver is a sophisticated tool to analyze the kinematic and dynamic
Analysis behaviors of mechanical systems. The motion analysis tool aims to analyze the
forces, torques, contact forces, and power consumption accurately. The
simulation result over time can be exported for other engineering analysis after
the simulation is completed.
The motion analysis tool is used to simulate and analyze a machine with the
consideration of driving forces, springs, dampers, and frictions. The kinematic
solver takes into account of motion constraints, material properties, mass, and
component contacts.

(a). Activate Motion in Adds-Ins

(b). Create New Motion Study by right-clicking (c). Select Motion Analysis in the drop-
the blank area above the status bar down menu of Animation

Fig. 2.72 Create a Motion Study for a machine model

Fig. 2.73 Types of motion variables in a Motion Study

102 2 Computer-Aided Design

Moving body Reference body

Motion direction

Selection of
motion profiles

Fig. 2.74 Defining a motor in a SolidWorks motion study

associated with a moving body, and the motion is relative to a reference body along
a specified direction; therefore, the moving body, the reference body, and the motion
direction must be specified. Thirdly, the profile of a motion can be one of Constant
Speed, Distance, Oscillating, Segments, Data Points, Expression, or Servo Motor
from the drop-down list of Motion. Fourthly, the direction of the motion must be
specified.

2.8.5.4 Setting Simulation Parameters

Other than motion variables, the motion study allows to customize the properties
of a simulation model shown in Fig. 2.75a. The user can (1) specify the number
of frames per seconds in calculation; (2) decide whether or not the simulation can
be visualized in the course of calculation; (3) refine the accuracy of 3D contact or
of the representation of solid geometry; (4) specify the cycle settings; (5) specify
the solving algorithm and the tolerance as the criterion of termination as shown in
Fig. 2.75b.

2.8.5.5 Motion Simulation

The motion analysis does not initialize the calculation automatically when some
changes are made in the simulation model; therefore, the user has to accept the
changes for a new simulation by clicking the ‘calculate’ icon to run the simulation
shown in Fig. 2.76. After the calculation is completed, the user can use the animation
tool to review the motion of the machine over time. In addition, the simulation result
can be saved and exported to external sources in an .AVI or other formats.
2.8 Kinematic and Dynamic Modeling 103

(a). activate Motion Study Properties

(c). Advanced setting for motion analysis (b). define simulation properties

Fig. 2.75 Interface to set simulation properties

Fig. 2.76 Run a motion study simulation

2.8.5.6 Analyze Simulation Data

A Motion Analysis model involves a large number of motion variables and simulation
parameters. It should be a very rare case that a user defines all of the simulation param-
eters appropriately at his or her first iteration. The user should know what kinematic
and dynamic properties are expected from the simulation to be capable of making
engineering judgments to see if the simulation result is reasonable. Figure 2.77 shows
the selection of kinematic and dynamic variables relevant to a motion study. Any of
these variables can be selected to investigate how it is changed over time in simu-
lation. Figure 2.78 shows an example plot of the change of torque over time for a
specified motor in the machine.
104 2 Computer-Aided Design

(a). activate Results

and Plots

(b). Quantity types

(c). Quantity sub-types

(d). Component types (e). Select object and the

coordinate system

Fig. 2.77 Define a plot for change of kinematic or dynamic variable over time

Fig. 2.78 The change of driving torque over simulation time

2.8.5.7 Mechanical Event Simulation

For a machine with motion, the forces exerted on solid bodies are varied over time;
accordingly, stress distributions over the bodies are changed over time. For a safe
design, it is helpful to determine when and at what amplitude of the maximum stress
occurs to the bodies in a machine. The time-dependent loads from a motion analysis
simulation can be utilized for the structural analysis of a solid body in a machine.
By incorporating a structural analysis in a motion study, the distributions of stress,
a factor of safety, or deformation over solid bodies can be analyzed directly without
manually setting up boundary conditions and loads; since the loads are imported
automatically from the motion study results. The user can investigate the effect of
dynamic motion load on stress or deformation distribution over one part or compo-
nent. Figure 2.79 shows an example where the stress and deformation of a part in a
robot have been analyzed at the specified timeframe. In this simulation model, the
loads and boundary conditions are automatically defined in the motion study model.
A motion study can be performed on a specified time and time range. In addition,
the data flow from a motion study to structural simulation is one-directional. In other
words, the results of the stress analysis do not affect the motion study model. Note
that detailed stress analysis can be performed in SolidWorks Simulation which will
be discussed in detail in Chap. 3.
2.8 Kinematic and Dynamic Modeling 105

Fig. 2.79 Incorporating mechanical event simulation in motion analysis

Example 2.9 The modular robotic system in Fig. 2.53 is used to create an assembly
model of 3-DOF parallel kinematic machine (PKM). The diameters of base and end-
effectors are set as 600 mm and 200 mm, respectively. Each branch of the 3-DOF
parallel kinematic machine consists of one 70-mm active rotary joint, two 70-70-
110 type-A links, and two 70-mm passive rotary joints. Taking the reference home
position of the assembly model, the motions of three active joints are given in Table
2.13.
Assume that the materials are set as 1060 alloy for all components in the assembly
model, and the direction of gravity force is along negative Z b . (1) Create the assembly
model for motion study. (2) Define and analyze the motion of the end-effector based
on the given joint motion. (3) Visualize the trace of the reference (Oe ) of the end-
effector platform. (4) Export the results of displacements and driving torques of active
joints. (5) Export the results of the displacements of Oe (x e , ye , θ e ) with respective
time.

Solution Hybrid modeling is adopted to (1) build three kinematic branches from the
modules in Fig. 2.53 by the bottom-up method and (2) create the ground and top
plates based on the specified diameters by the top-down method. Figure 2.80 shows
the assembly model of 3-DOF parallel kinematic machine. It consists of a ground
platform, three 3-DOF branches, and an end-effector platform.
To create a motion study model, all active joints are defined as ‘motors’. In the
given 3-DOF parallel kinematic mechanism, each branch is built from two passive
rotary joints and one active rotary joint. The first rotary joint in each branch is selected
to define the motor for this branch as shown in Fig. 2.81. The motion properties
of the three motors are defined based on Table 2.13 accordingly. The completed
model includes three motors; in addition, the direction of gravitational acceleration
is specified to take into consideration the effect of the weights in motion.
106 2 Computer-Aided Design

Fig. 2.80 Assembly model of 3-DOF parallel kinematic machine

Trace of the reference (Oe) of end-effector platform

Joint 3

Joint 2

Joint 1

Motion definitions
of active joints

Fig. 2.81 The trace of the reference Oe subjected to the given motions of three active joints

Table 2.13 Specified joint motions

Joint No. Motion profile Motion range (°) Frequency (hertz)

1 40 1

2 Oscillation 35 0.5

3 20 1

The simulation is calculated, and the simulation results can be visualized,

analyzed, and exported. Figures 2.82, 2.83, and 2.84 show the changes of joint
displacements, driving torques, and end-effector motions over time, respectively.
2.8 Kinematic and Dynamic Modeling 107

Angular Displacement1 (deg) 40

30

20

10

0
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
Time (sec)

(a) joint angle (θ 1) over time (second)

35
Angular Displacement2 (deg)

26

17

-0
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
Time (sec)

(b) joint angle (θ 2) over time (second)

20
Angular Displacement3 (deg)

15

10

0
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
Time (sec)

(c) joint angle (θ 3) over time (second)

Fig. 2.82 Joint displacements over time
108 2 Computer-Aided Design

Motor Torque1 (newton-mm) 1599

1226

852

478

104
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
Time (sec)

(a) joint torque (τ1) over time (second)

969
Motor Torque2 (newton-mm)

732

495

258

22
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
Time (sec)

(b) joint torque (τ 2) over time (second)

1286
Motor Torque3 (newton-mm)

966

647

328

8
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
Time (sec)

(c) joint torque (τ 3) over time (second)

Fig. 2.83 Joint torques over time

2.8 Kinematic and Dynamic Modeling 109

Center of Mass Position1 (mm) 2

-8

-18

-29

-39
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
Time (sec)

(a) end-effector platform xe (mm) over time (second)

69
Center of Mass Position2 (mm)

40

11

-18

-47
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
Time (sec)

(b) end-effector platform ye (mm) over time (second)

38.0
Angular Displacement4 (deg)

37.4

36.8

36.2

35.5
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
Time (sec)

(c) end-effector platform θ e (mm) over time (second)

Fig. 2.84 The displacements of end-effector over time

110 2 Computer-Aided Design

2.9 Summary

Engineers should master the skills of using CAD tools to design, analyze, model,
simulate, and evaluate a product, manufacturing process, or system. In this chapter,
various CAD techniques are introduced to represent products, processes, and systems
at different levels and aspects. To improve the productivity of engineering designs,
engineers should understand the theory, methods, and tools of parametric modeling
and knowledge-based engineering (KBE), and use computer-aided tools efficiently
to model the geometrics, features, design intents, and assembling relations. For
machine design, engineers can benefit greatly by using the motion simulation tools
for kinematic and dynamic analysis of machines in an integrated computer-aided
environment.
Design Problems

Problem 2.1 Identify the number of vertices, edges, faces, loops, and genus of the
following objects, and use the Euler-Poincare law to justify if they are legal solid
models.

(a) (b) (c)

(d) (e) (f)

Problem 2.2 Identify the number of vertices, edges, faces, loops, and genus of the
following objects, and use the Euler-Poincare law to determine object types.
2.9 Summary 111

(a) (b) (c)

(d) (e) (f)

Problem 2.3 Using the Euler-Poincare Law to justify the validity of open objects in
the first column of Table 2.14.

Table 2.14 Examples of objects for problem 2–3

Example F E V L B G F-E+V-L=B-G
112 2 Computer-Aided Design

(a) 2 2 2 (b) Sphere with 2.5 (c) A model with 6 sphere patterns (
cube 0.25)

(d) Finish dice model

Fig. 2.85 A dice model using CSG composition

Problem 2.4 Create a dice model as shown in Fig. 2.85d from the given three features
using the CSG composition.
Problem 2.5 Review the drawing in Fig. 2.86, identify the set of solid primitives,
and determine the composition operations to create the represented solid model.
Problem 2.6 Review the drawing in Fig. 2.87, identify the set of solid primitives,
and determine the composition operations to create the represented solid model.
Problem 2.7 Create a part model for a set of six sockets as shown in Fig. 2.88d.
Problem 2.8 Bottom-up assembly modeling: manufacturers of modular robotic
systems often provide the design library of part models for users to model robots.
Download the part library for Vex robots from https://www.vexrobotics.com/iq/dow
nloads/cad-snapcad, and build a robot similar to the one shown in Fig. 2.89, and have
a motion simulation to illustrate its basic movement.
Problem 2.9 Top-bottom assembly modeling: there are thousands of rivets on a
wing of commercial aircraft (see Fig. 2.90), and riveting operations heavily rely on
human operators. Create a design concept of a multifunctional riveting tool which
can move over the surface of wing and automate riveting processes, and create the
motion simulations to demonstrate all of the required operations.
Problem 2.10 use the modular robotic system in Fig. 2.53 to build a 2D robot (open
chain) and create a motion simulation model to analyze its kinematic and dynamic
behaviors.
2.9 Summary 113

Fig. 2.86 Determine constitutive solid primitives and composition operations for solid for Problem
2.5

Fig. 2.87 Determine constitutive solid primitives and composition operations for solid for Problem
2.6
114 2 Computer-Aided Design

(b). create a ‘design table’ in the ‘tools’ menu

Title of design table

(d). six configurations in
model

(a). define controllable

variables and feature in a List of controllable variables
model and properties table
Template of a
design table
Variables, equations, and
List of configurations states for controllable
variables and properties

(c). create a design table for part family

Fig. 2.88 Example of creating a part model with design table

Fig. 2.89 Example of Vex

robot configuration

Fig. 2.90 Thousands of

rivets on aircraft wing Ty
Tx

Rz
2.9 Summary 115

Problem 2.11 Use the modular robotic system in Fig. 2.53 to build a 3D robot (open
chain) and create a motion simulation model to analyze its kinematic and dynamic
behaviors.

Problem 2.12 For a four-bar mechanism, the Grashof condition is used to justify is
the driving link has a full rotation as a crank. The Grashof condition is L + S≥ P + Q;
where L and S represent the lengths of the longest and the shortest links, and P and Q
represent the lengths of other two links. (1) Create a parametric 4-bar mechanism in
which the lengths of L, S, P, and Q can be adjusted to verify the Grashof condition.
(2) For a 4-bar mechanism with a crank, create a motion simulation to analyze the
relation of input and output angles.

References

Bi ZM (2018) Finite element analysis applications: a systematic and practical approach, 1st edn,
Academic Press. ISBN-13: 978-0128099520
Bi ZM, Wang XQ (2020) Computer aided design and manufacturing, Wiley-ASME Press Serious,
1st edn. ISBN-10 1119534216
Bi ZM, Cochran D (2015) Big data analytics with applications. J Manag Anal 1(4):249–265
Chang KH (2014) Chapter 3 solid modeling, product design modelling using CAD/CAE, the
Computer Aided Engineering Design Series, Academic Process, pp 125–167
Bi ZM, Yang SYT, Shen W, Wang L (2008) Reconfigurable manufacturing systems: the state of the
art. Int J Prod Res 46(4):967–992
Boothroyd G (1994) Product design for manufacture and assembly. Comput Aided Des 26(7):505–
520
Dassault Systems (2020) Design methods (bottom-up and top-down design). https://help.solidw
orks.com/2018/English/SolidWorks/sldworks/c_Design_Methods.htm?id=4cf2b835663b4c0
ca59e0b83aa6f7dda#Pg0
Galliera J (2010) Solvers used for animation, basic motion, and motion analysis. https://forum.sol
idworks.com/community/simulation/motion_studies/blog
Gui J-K, Mantyla M (1994) Functional understanding of assembly modelling. Comput Aided Des
26(6):435–451
Hooper T (2020) Top 3D CAD modeling software: the 50 best CAD Tools for ideation, rendering,
prototyping and more for product engineers. https://www.pannam.com/blog/best-3d-cad-mod
eling-software/
Intel (2002) Expanding Moore’s law the exponential opportunity. https://www.cc.gatech.edu/com
puting/nano/documents/Intel%20-%20Expanding%20Moore’s%20Law.pdf
Jhanji Y (2018) Chapter 11: computer-aided design—garment designing and patternmaking. In:
Automation in garment manufacturing, pp 253–290
Junk S, Kuen C (2016) Procedia CIRP 50:430–435
Lyu G, Chu X, Xue D (2017) Product modeling from knowledge, distributed computing and lifecycle
perspectives: a literature review. Computer in Industry 84:1–13
Markkonen P (1999) On multi body systems simulation in product design. Doctoral thesis, Royal
Institute of Technology, KTH, Stockholm
Niu N, Xu L, Bi ZM (2013) Enterprise information system architecture—analysis and evaluation.
IEEE Trans Ind Inform 9(4):2147–2154
Requicha AAG (1980) Representations for rigid solids: theory, methods, and systems. ACM Comput
Surv 12(4):437–464
116 2 Computer-Aided Design

Sanchez-Cruz H, Sossa-Azuela H, Braumann UD, Bribiesca E (2013) The Euler-Point Formula

through contact surfaces of voxelized objects. J Appl Res Technol 11(1):65–78
Särestöniemi M, Pomalaza-Ráez C, Bi ZM, Kumpuniemi T, Kissi C, Sonkki M, Hämäläinen M,
Iinatti J (2019) Comprehensive study on the impact of sternotomy wires on UWB WBAN channel
characteristics on the human chest area. IEEE Access 7(1):74670–74682
Shalf J, Leland R (2015) Computing beyond Moor’s Law. Computer 48(12):14–23
Solidworks (2018) Summary of design table parameters. Available online: https://help.solidw
orks.com/2018/english/SolidWorks/sldworks/r_Summary_of_Design_Table_Parameters.htm
(accessed on April 18, 2021).
Wang XV, Givehchi M, Wang L (2017) Manufacturing system on the cloud: a case study on cloud-
based process planning. Procedia CIRP 63:39–45
Wikipedia (2020a) Computer aided design. https://en.wikipedia.org/wiki/Computer-aided_design
Wikipedia (2020b) Design for assembly. https://en.wikipedia.org/wiki/Design_for_assembly
Yu J, Xu L, Bi ZM, Wang C (2014) Extended interference matrices for exploded views of assembly
planning. IEEE Trans Autom Sci Eng 11(1): 279–286
Zissis D, Lekkas D, Azariadis P, Papanikos P, Xidias E (2017) Collaborative CAD/CAE as a cloud
service. Int J Syst Sci Oper Logist 4(4): 339–355
Chapter 3
Computer-Aided Engineering (CAE)

Abstract Product design involves in modelling, simulation, and evaluation of the

behaviors of the product at the different stages of its product lifecycle (PLC).
Computer aided engineering (CAE) uses computers to model and analyze the
responses of product that is subjected to external loads and constraints. CAE has
become essential to eliminate hidden mistakes and omissions in virtual design before
the physical product is made. In this chapter, the importance of CAE in manufac-
turing is discussed, different CAE tools are introduced, and the focus is put on
finite element analysis (FEA). The theoretical fundamental of FEA is introduced, the
general procedure of FEA modelling and simulation is presented, and FEA is used to
analyze six common types of engineering problems. The SolidWorks Simulation is
used to illustrate how CAE is applied in analyzing and solving various engineering
problems.

Keywords Computer aided engineering (CAE) · Axiomatic design theory (ADT) ·

Finite element analysis (FEA) · Numerical simulation · Static analysis · Modal
analysis · And multidisciplinary system · Verification and validation

3.1 Introduction

Engineering concerns the design, construction, and use of structures, machines, and
systems. Engineering is often multidisciplinary which relates to a broad range of
particular areas such as mathematics, applied science, and types of applications
such as solid mechanics, flow mechanics, aerodynamics, and electromagnetics. In
computer-aided engineering (CAE), computers are used to model, analyze, and
simulate the behaviors of products, systems, or processes for design optimization
or solving various engineering problems. CAE tools are used in modeling, simula-
tion, and design optimization of products, processes, and systems (Bahman 2018).
The success of a modern enterprise depends greatly on the digitization of all of the
business processes including engineering processes (Krahe et al. 2019).
Enterprises benefit from using CAE at a number of aspects. (1) CAE reduces
cost and time of product development, and improves the product in its lifecycle
continuously. (2) CAE generates, evaluates, and implements product design in the

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 117
Z. Bi, Practical Guide to Digital Manufacturing,
https://doi.org/10.1007/978-3-030-70304-2_3
118 3 Computer-Aided Engineering (CAE)

minimized number of the design iterations. (3) CAE aims at the first time right
practice by reducing the needs of physical prototypes and tests. (4) CAE uses a virtual
model to evaluate the performance, reliability, and safety of products. (5) CAE can
incorporate CAD, CAM, and other computer-aided tools as the holistic integration of
data and process management over the product lifecycle. (6) CAE analyzes product
fatigue life that can reduce the cost associated with the unanticipated failure of
products or systems.
An engineering process is to convert customers’ needs into a physical product or
system to meet desired functional requirements (FRs). Following a generic procedure
is helpful to ensure that an engineering process generates an ideal product or system
in satisfying customers’ needs. No matter that a product or system is simple or
complex; the engineering process follows a generic procedure as shown in Fig. 3.1.
The engineering process involves in a number of critical steps and corresponding
activities to obtain an optimal solution. It begins with the identification of customers’
requirements (CRs), and it is followed by (1) the formulation of design problem, (2)
the definition of solution space, (3) design analysis over solution alternatives, (4)
design synthesis for the comparison and selection of the solution alternatives, and
(5) the verification, validation (V&V), and implementation of the finalized solution.
In Fig. 3.2, the case of developing an automatic fuel-recharging system (Bi et al.
2020) is used to illustrate the critical steps in applying the above engineering process.
Formulation of design problem. An engineering process aims at an optimal
solution to a given engineering problem. Therefore, the first step is to formulate the
design problem with clear CRs, design objectives, and design constraints. Engineers
need to collect as much information as possible; in particular, the information relevant
to operation environment and available resources. The engineering process for the
design challenge in Fig. 3.2 can be formulated as a design problem as developing
a system that is fully automated to (1) refuel gasolines for various cars; (2) make
payments through the network while drivers stay in their vehicles; (3) offer the drivers
with safer, cleaner, and more comfortable services than traditional self-refueling
services without an extra-cost to drivers.
Definition of solution space. A solution space consists of all possible design
options that may meet the given FRs and design constraints. After the design problem
is formulated, engineers will determine a conceptual design of product or system that
covers all of the feasible solutions. The conceptual design defines the boundaries of a
solution space in terms of the number, characteristics, and types of design variables,
the ranges and resolutions for the changes of design variables, and the constraints for
the dependences of design variables. To make the complexity of a design manageable,
some system design methodologies, such as axiomatic design theory (ADT), can be
used to identify a set of design variables (DVs) for FRs (Suh 2005). To define a
solution space by ADT, the system-level FRs are decomposed, so that each sub-FR
can be fulfilled by a given set of DVs. Accordingly, a solution space uses a hierarchical
structure to represent a set of sub-solutions and their relations.
Figure 3.3 shows the decomposition of FRs in terms of the major tasks an auto-
matic refueling system needs to perform. The highest level FR-0 is decomposed into
the second level sub-FRs as FR-11 for ‘data collection and processing’, FR-12 for
3.1 Introduction 119

Inputs of Engineering Process

• Customers’ requirements (CRs)
• Business environments
• Performance metrics
• Standards and regulaons
• Current statuses

Define Design Problem

• Design goals and constraints
• Funconal requirements (FRs)
• Available resources

Determine Soluon Space

• Design variables (DVs)
• Resoluons of variables
• Design boundaries

Design Synthesis
Design Analysis
• Pool of candidates
• Dependence of variables
• Search strategies
• Variables & constraints
• Evolving strategies
• Variables & goals
• Terminaon criteria

V&V and Implement

• Validaon and verificaon (V&V)
• Refinement
• Prototyping
• Tesng
• Connuous improvement

Outputs of Engineering Process

Activities that are heavily • The soluons to meet FRs
supported by computer aided • Evoluon of product/system
engineering (CAE) techniques • Opmized design performance
such as Finite Element Analysis • Updated status
(FEA)

Fig. 3.1 Description of an engineering process for product or system designs (Bi 2018)

Fig. 3.2 Automatic

refueling system (Fuelmatics
2020)
120 3 Computer-Aided Engineering (CAE)

Fig. 3.3 Decomposition of system FRs of automated refueling system (Bi et al. 2020)

‘controls for normal events’, FR-13 for ‘refueling operation’, and FR-14 for ‘controls
for abnormal events’. FR-11 is decomposed into the third level sub-FRs as FR-111
for ‘detecting an incoming vehicle’, FR-112 for ‘obtaining customer’s intent and
payment information’, FR-113 for ‘collecting information of vehicle and fuel’, and
FR-114 for ‘determining position and orientation of vehicle relative to the refueling
tool’. Similarly, FR-13 is further decomposed into the third level sub-eFRs as FR-131
for ‘accessing fuel port to open cover’, FR-132 for ‘selecting and gripping nozzle
for refueling’, FR-133 for ‘retrieving refueling tool’, and FR-134 for ‘closing fuel
cover and resetting refueling tool’.
Figure 3.4 shows the correspondence of DVs to the FRs in Fig. 3.3. The sub-FRs
at the third level under FR-11 are mapped to four DVs as DV-111 with the options
of visions, barcodes, and laser scanners, DV-112 and DV-113 with the options of
electronic chips, apps, and decentralized databases over the Internet, and DV-114
with the options of visions, barcodes, laser scanners, and controllable platforms. The
sub-FRs at the third level under FR-13 are mapped to four DVs as DV-131, DV-132,
and DV-114 with the options of robots, gantry systems, and specialized mechanisms
equipped with multiple functional tools, and DV-133 with the options of manual
inputs, programmable logic controllers (PLC), and IoT-based apps. The FR-12 and
FR-14 at the second level are mapped to DV-12 and DV-14 with the options of
stand-alone systems and IoT-enabled apps.
Design synthesis. A design space consists of infinite or finite number of possible
solutions depending on the number and types of design variables. To obtain an optimal
3.1 Introduction 121

DV-12: Stand-along system, IoT-enabled apps

DV-0

Fig. 3.4 The mapping from FRs to DVs of automated refueling system (Bi et al. 2020)

solution to the formulated problem, the performances of potential solutions in the

design space must be analyzed, evaluated, and compared based on the specified opti-
mization criteria. Therefore, design synthesis plays its critical role in the engineering
process in (a) selecting an initial possible solution as a reference for evaluation and
comparison; (b) executing search algorithms to explore better solutions, (c) defining
terminating conditions to control the optimization process (see Fig. 3.5).
Design analysis. To implement the design synthesis in Fig. 3.5, design analysis is
needed to evaluate and compare different solutions. Design analysis is used to model
the behaviors of a product or system, and analyze a potential solution to determine
(a) whether or not it meets all design constraints and (b) how well its performance
is against the specified optimization criteria. The evaluation results from the design
analysis are used by design synthesis to select better solutions. Generally, potential
solutions have to be evaluated in a quantitative way. Therefore, the behaviors of a
product or system have to be mathematically modeled to correspond design variables
to design constraints, FRs, and evaluation criteria. Take the mechanism (DV-131 or
DV-134 in Fig. 3.4) for opening or closing a fuel inlet as an example, performing
design analysis needs the kinematic and dynamic models of the mechanism to eval-
uate whether or not the fuel inlet is reachable, and the inlet cover can be opened or
closed without any interference within the specified period of time.
122 3 Computer-Aided Engineering (CAE)

Fig. 3.5 Design analysis and design synthesis in an engineering process (Bi 2018)

Implementation of optimal solution. Once an optimal solution is determined

at the conceptual design stage, the detail design is performed to implement optimal
solution. Since adding the details of the design solution may affect the performance
of the product or system, engineers have to refine the solution for its practical imple-
mentation. Taking the functional modules for DV-131 or DV-134 as the examples,
joint actuators specify the ranges of motions, accelerations, and power; engineers
need to re-evaluate the solution when the actuators are selected for the robot. Even
the home positions of actuators will affect the workspace of the robot. The main tasks
at the implementation stage include (a) convert the virtual model into the physical
product or system, (b) verify and validate the performance of physical system, and
refine the system whenever it is needed, and (c) perform the practice of continuous
improvement (CI) in the application.

3.2 Design Analysis Methodologies

The behaviors of a product or system are governed by certain scientific principles,

which can be mathematically represented by governing equations. Taking an example
of a solid object subjected to external loads, the responses of the solid object to
external loads can be represented by the compatibility relations, strain–displacement
relations, and the equations of motion. From this perspective, design analysis begins
with the understanding of the scientific and mathematical principles that govern the
system behaviors.
As shown in Fig. 3.6, once a mathematic model is formulated, different engi-
neering methods can be used to analyze the product or system. For example, design
analysis methods can be the types of graphic methods, experimental methods, and
computational methods. Table 3.1 gives a comparison of these analysis methods.
3.2 Design Analysis Methodologies 123

Formulated
Engineering Problems

Theory Governing Data

mathema c models

Experimental Computa onal Graphic

Methods Methods Methods

Numerical Analy cal

Methods methods

Finite element Boundary

methods element methods
Finite difference
methods

Predicted system behaviors based on given design variables

Fig. 3.6 Classification of design analysis methods

A graphic method uses the drawing tools to define the relations of inputs and
outputs of system graphically. It helps to understand system behaviors visually.
However, a graphic method is usually very preliminary, and it can only deal with
some simple problems.
An experimental method studies the relations of inputs and outputs of system
experimentally. Experimental methods are often applied in the implementation stage;
appropriate design of experiments (DoE) will be needed to ensure that experimental
results are reliable and trustable. However, an experimental method has a few of
the limitations as below: (1) experiments are conducted on a physical system while
it is only available at the implementation stage; fixing a design error discovered at
the implementation phase involves a high cost; (2) using an experimental method
means the need of additional cost on physical systems and instrumentations for
measurements; (3) an experimental method uses the enumeration to investigate the
impact of the changes of inputs on outputs of system. Since a system usually relates
to many design variables, it is infeasible or impractical to understand the system by
a large number of experiments.
 124

Table 3.1 Comparison of different design analysis methods

Method Description Features
Graphic Uses drawing tools to define the relations of inputs and • It is easy to use, but it is only applicable to simple design
outputs of system graphically problems
• It requires the users with a high-level visibility and drawing
capability; fixing an error takes a large amount of time for
repetitive works
Experiment Models the relations of inputs and outputs of system • It is very reliable and can make the experiments very specific
experimentally to given applications. It is applicable to the systems with any
level of the complexity and uncertainty
• It involves in an additional cost for setup and
instrumentation; it is used only when the physical system is
available, and it has its limitation in the number of
experiments for practical purpose
Computational Analytical Obtains system outputs based on given system inputs • It leads to an explicit mathematical model to predict the trend
analytically of system in a continuous domain
• It is only applicable to small-scale, simple design problems
with a few of design variables and well-structured governing
equations
Numerical FEA—integrating derivatives in the elements of the entire • It is used in a virtual model where no physical system is
discretized domain needed. It is generic to a wide scope of interdisciplinary
FDM—approximating derivatives by finite differences in the problems in a continuous, discrete, or mixed domain. It is
elements of the entire discretized domain applicable to the engineering systems with any level of
complexity and uncertainty
FBM—approximating the governing equations in elements • It leads to an approximated solution, and it needs multiple
of discretized boundaries of the domain steps for the verification and validation of design solutions.
The impact of system inputs on system outputs can only be
studied numerically by simulation
3 Computer-Aided Engineering (CAE)
3.2 Design Analysis Methodologies 125

A computational method uses computer models to simulate system behaviors.

Computational methods are further classified into analytical methods and numerical
methods. An analytical method uses analytical models to represent the relations
of system inputs and outputs; it applies to some simple problems where system
outputs can be obtained explicitly using analytical models. A numerical method uses
numerical models to represent the relations of system inputs and outputs, and system
outputs are obtained numerically by computer simulation. Numerical methods are
applicable to both of explicit and inexplicit mathematical models. Numerical methods
are widely used to analyze the systems with high complexity and uncertainties,
which cannot be tackled with graphic methods, experimental methods, or analytical
methods.
Numerical methods have become the default CAE tools for a number of reasons:
(1) modern products or systems become so complex for other design analysis methods
to evaluate their performances thoroughly; (2) in contrast to experimental methods,
numerical methods are able to predict system behaviors without physical prototyping;
this reduces the development cost, shortens design time, and allows to compare a
large number of design options for optimization; (3) cutting-edge CAD and CAE
tools are available for engineers to solve a wide scope of engineering problems
without sophisticated trainings.
Many numerical methods are available, and they generally use the divide and
conquer strategy to deal with the generality and complexity of different engineering
problems. As shown in Fig. 3.6, a numerical method can be one of finite element
analysis (FEA), finite difference methods (FDM), and boundary element methods
(BEM). FEA differs from FDM in approximating the derivatives in a mathematical
model. Derivatives are evaluated by the integration in FEA; while these derivatives
are evaluated by finite differences in FDM. FEA differs from BEM in tackling with a
continuous domain. FEA model has elements and nodes in the entire domain; while
a BEM model has the elements and nodes only on the boundaries of the domain.

3.3 Numerical Simulation

As popular computational tools, numerical simulation is widely used to find approx-

imate solutions to various engineering problems. In analyzing a system by numer-
ical simulation, the system behaviors should firstly be formulated as a mathematic
model; which is typically represented by some partial differential equations (PDEs)
subjected to the specified boundary conditions. Figure 3.7 shows the dependence of
a simulation model with its original engineering problem. An engineering problem
is formulated as a set of design variables for inputs (I), system characteristic (S), and
outputs (O); the behaviors of system can be represented by the relations of I, O, and
S inexplicitly as f (I, S, O) = 0.
126 3 Computer-Aided Engineering (CAE)

Problem domain
(Infinite DoFs) Problem’s outputs (e.g.
Problem’s inputs quantities to represent
(e.g. fixtures, initial State variables O for outputs, system system behaviors such as
conditions, loads) characteristics/parameters S and displacement, stress,
governing equations f (I, O, S)=0 factor of safety in solid
(Inexplicitly) mechanics

Solution domain
Model’s inputs (Finite DoFs) Model’s outputs
(Boundary conditions (Independent variables
and loads on Conversion from inputs to outputs
and other dependent
discretized elements (Explicitly) scalars or vectors) of
-1
{O}={R(S)} {I} the discretized system

Fig. 3.7 Numerical simulation as design analysis method

The simulation model differs from its original engineering problem at the
following aspects:
(1) A simulation model discretizes the continuous domain of the original problem;
in a discretized node, the behaviors on discretized nodes are used to represent
the behaviors of any point in the continuous domain.
(2) A simulation model has a limited number of degrees of freedom (DoF),
each DoF corresponds to a state variable to be solved in simulation. The
system behaviors are collectively determined by state variables on all nodes.
In contrast, the original engineering problem involves infinite number of DoF
since the points in a continuous domain are uncountable and infinite.
(3) A simulation model must be solvable to find state variables explicitly subjected
to the specified boundary conditions and loads. Solving a simulation model can
be viewed as a conversion from the given boundary conditions to state variables
to represent system behaviors approximately.
(4) A simulation model is developed in such a way that all of the given parameters
in the original engineering problem are defined as the boundary conditions and
loads on discretized nodes and elements; all independent parameters related to
system behaviors are defined as state variables or other dependent quantities.

3.4 Finite Element Analysis (FEA) and Modeling Procedure

FEA adopts the piecewise approximation where a continuous domain is discretized

into finite elements. Each element is represented by a set of nodes; the behaviors
of elements are represented by those of the nodes; a system model can then be
assembled from element models. By incorporating boundary and load conditions,
3.4 Finite Element Analysis (FEA) and Modeling Procedure 127

Validation Application of
Holistic
computer computer
solution to the solutions in
Engineering Formulated engineering engineering
applications engineering problems problem applications

Validation Solutions to
single-physics
system; solutions
to multi-physics
System components systems
Validation
Element modelling,
boundary
conditions,
assemblies of sub-
Functional modules models

Constitutive
models, governing
equations for
Materials with mass, momentum,
physical behaviors and energy
Verification

Fig. 3.8 Top-down and bottom-up procedure in FEA

the system model can be solved to obtain a numerical solution to original engineering
problem. To analyze an engineering problem, FEA consists of two procedures, i.e.,
the top-to-bottom procedure and the bottom-up procedure as shown in Fig. 3.8.
The top-down procedure is to break down system complexity. Firstly, the bound-
aries of system are defined to clarify inputs (I), parameters (S), outputs (O), and
design constraints and objectives. Secondly, the continuous domain is decomposed
into discretized elements and nodes; multiple layers of decompositions may be
performed, so that the behavior of one element and its relations with others can be
defined appropriately. Lastly, the elemental behaviors are modeled as mathematical
equations.
The bottom-up procedure is to obtain the system solution based on the sub-
solutions at the element level. Firstly, element types and analysis types are selected to
determine governing equations of elements adequately. Secondly, the solutions to the
mathematical models are defined at the element level. Thirdly, the elements models
are assembled into a system model based on the topological relations defined in
the top-down procedure. Finally, the boundary conditions are applied in the system
model; it is then solved to obtain the system solution. Since FEA is a technique
of numerical simulation, verification and validation (V&V) that must be applied to
endure the outcomes from FEA are acceptable.
To illustrate the steps in using FEA to solve an engineering problem, the idea of
FEA is adopted to calculate the workspace of a robot numerically as in Example 3.1.

Example 3.1 Calculate the workspace of an Espon robot shown in Fig. 3.9a (Espon
2020).
Solution. Firstly, the problem of workspace calculation is formulated as a design
problem with the inputs (I), outputs (O), and system parameters (S) as shown in
128 3 Computer-Aided Engineering (CAE)

Worpspace (x, y, z, θ z)

(a) 4-DOF workspace of end-effector tool

X Z
(x, y) (z) (θ z)

(b) Reachable space (c) Travel distance (d) Tool rotation

in x-y plane along z-axis along z-axis

Fig. 3.9 Decomposition of continuous domain

Table 3.2. Note that the workspace is a collection of all reachable points (x, y, z, θ z )
by the end-effector.

Secondly, the continuous domain of the workspace is discretized into three sub-
domains as shown in Fig. 3.9b–d, respectively. Since the z-axis translation, z-axis

Table 3.2 Description of the formulated problem for workspace analysis of Espon robot
3.4 Finite Element Analysis (FEA) and Modeling Procedure 129

rotation, and the motion in the x–y plane are decoupled and implemented by respective
joints, the motions in these three sub-domains can be analyzed individually. The
behavior of each element is represented by its central point, e.g., if the central point
is reachable, the corresponding cell is reachable.
Thirdly, the kinematic model is developed to describe the relations of joint motions
(θ 1 , θ 2 , z3 , θ 4 ) and the end-effector motion (x, y, z, θ z ). Since z3 and θ 4 are one-to-one
mapped to z and θ z , respectively, only the mapping of (θ 1 , θ 2 ) to (x, y) are modeled
here.
In the first sub-domain, a working point (x, y) in the workspace can be calculated
from given joint displacement (θ 1 , θ 2 ) as

x = L 1 cosθ1 + L 2 cosθ2
(3.1)
y = L 1 sinθ1 + L 2 sinθ2

A1 cosθ1 + B1 sinθ1 + C1 = 0 (3.2)

where
⎫
A1 = 2x L 1 ⎬
B1 = 2y L 1 (3.3)
⎭
C1 = L 21 − L 22 − x 2 − y 2

1−t12
Let t1 = tan θ21 , sinθ1 = 1+t
2t1
2 , and cosθ1 =
1+t12
,
1
Equation (3.2) becomes

(C1 − A1 )t12 + 2B1 t1 + (C1 + A1 ) = 0 (3.4)

The solution of from Eq. (3.4) is found as


−2B1 ± B12 − C12 + A21
θ1 = 2tan−1 (3.5)
(C1 − A1 )

Following the similar procedure to eliminate θ1 in Eq. (3.1) gets the solution of
θ2 as

−2B2 ± B22 − C22 + A22
−1
θ2 = 2tan (3.6)
(C2 − A2 )

where
130 3 Computer-Aided Engineering (CAE)

Fig. 3.10 Determine if a point in 3D space is a reachable point

⎫
A2 = 2x L 2 ⎬
B2 = 2y L 2 (3.7)
⎭
C2 = L 22 − L 21 − x 2 − y 2

Fourthly, elements are analyzed one by one to determine if their central points
have corresponding inverse kinematic solutions. If an inverse kinematic solution
exists in the joint space, the element is reachable and within the workspace in the
task space (see Fig. 3.10).
Fifthly, the reachability of all of discretized elements is checked, and all of the
reachable elements are assembled to form the workspace of the robot. Note that the
workspace consists of the sub-space on the x–y plane, z-axis translation, and z-axis
rotation.
Finally, the result from the workspace analysis is processed to visualize and
retrieve the result. The post-processing also includes deriving some relevant quan-
tities that help to understand the solution to the original engineering problem.
Figure 3.11 visualizes the workspace of the robot with the limits of the end-effector
motion along x-, y-, and z- axes.
The FEA modeling procedure can be generalized from the aforementioned
example, and it consists of the following steps.
Step 1—Decomposition. The engineering problem is formulated to identify
inputs (I), outputs (O), and characteristic parameters (S) of system. The continuous
domain about S is discretized into a set of elements, and each element is represented
by its nodes. The schematic of the discretization also defines element-to-element
and node-to-element relations. The system behaviors are represented by the field
variables on nodes, and the interpolation functions are used to describe the relation
of any point and its corresponding nodes in the element.
Step 2—Develop element models. The analysis type of the FEA model is defined
to reflect the physical behaviors of system. The analysis type determines governing
differential equations of elements in mathematical modeling. The numerical solution
3.4 Finite Element Analysis (FEA) and Modeling Procedure 131

θz ∈[-360°, 360°]

y
z

Rz=200 mm

Fig. 3.11 Post-processing of workspace analysis

to the mathematic model is obtained by the approximation such as direct methods,

minimum potential energy methods, or weighted residual methods.
Step 3—Assembly. The element models in local coordinate systems (LCSs) are
transformed into the element models in a global coordinate system (GCS), and the
transformed element models in GCS are then assembled into a system model.
Step 4—Apply boundary conditions and loads. The interactions of the system
with its application environment are analyzed to define boundary conditions (BCs)
and loads for the FEA model.
Step 5—Solve for primary unknowns. Sufficient BCs ensure that the assembled
system model is solvable. For the majority of engineering applications, a system
model refers to a set of the linear equations for the field variables on nodes. Numerous
algorithms have been developed and used to solve unknown field variables in the
system model.
Step 6—Calculate dependent variables. The variables used to represent the
system behaviors can be classified into independent variables and dependent vari-
ables. For example, stress and strain are dependent with each other. Strains can be
selected as independent variables, sequentially, the stresses are dependent, which
can be determined based on the constitutive models of materials. After independent
variables are solved from the system model, the post-processing can be performed
to evaluate dependent variables.

3.5 FEA Theory

In FEA modeling, a virtual model of a part, process, product, or a system has to

be established; the virtual model includes one or more finite volumes with specified
material properties. The divide and conquer strategy is applied to deal with the
132 3 Computer-Aided Engineering (CAE)

Distributed Concentrated
loads loads
Field variables u(x, y) for
system behaviors

Discrete
nodes Elements

Constrained boundaries

Fig. 3.12 Discretization of a continuous domain

0. As shown in Fig. 3.12, the continuous domain is decomposed into small parts
called ‘elements’. Elements are connected at the interconnections, which are called
as ‘nodes’. In return, each element is associated with a set of nodes, and the behavior
of any material point in the element is determined by the behaviors of the nodes
associated with this element. An elemental model is used to describe the behaviors
of the nodes of the element; note that the behaviors of the nodes are represented by
the field variables. A system model is an assembly of the elemental models of all
of discretized elements in the continuous domain. By applying boundary conditions
and loads to a system model, the behaviors of all nodes in the continuous domain
can be determined as the solution to the system model.

3.5.1 Governing Equations of Engineering Problems

Ideally, an analytical model to an engineering problem is desired; since it leads to

an exact solution. Unfortunately, an analytical solution does not always exist; even
if it exists, it is mostly impractical to obtain it with a reasonable amount of time and
level of the required expertise. In contrast, FEA has been a powerful CAE tool due
to its capability in dealing with the complexity of various engineering problems at
different aspects: (1) a continuous domain of a real-world design problem is often
irregular and very complex; moreover, it is difficult to represent the boundaries of a
domain accurately. (2) The solid domain may consist of a set of sub-domains with
different properties, or one material with anisotropic nature. (3) The behaviors of the
engineering system may involve in multiple disciplines. (4) The engineering system
has the complex interactions with its application environments, which brings the
challenges in defining boundary or loading conditions.
3.5 FEA Theory 133

Table 3.3 shows a few examples of engineering problems with the corresponding
laws of mechanics and differential equations. As long as an engineering problem is
formulated as a mathematic model with specified boundary conditions, numerical
simulation such as FEA can be used to find the system solution.
FEA is a generic technique to solve a mathematic model subjected to specified
boundary conditions. If different engineering systems can be formulated as the same
mathematic model, the same solver in the FEA implementation can be used to solve

Table 3.3 Mathematic models of common physical phenomena in engineering

Engineering applications Laws of mechanics and differential equations
Solid mechanics Hooke’s law


εx x = E σx x − v σ yy + σz
1

ε yy = E σ yy − v(σx x + σzz )
1


εzz = E σzz − v σx x + σ yy
1

τx y τ
γx y = G γ yz = Gyz γx z = τGx z
where [σ ]T and [ε]T are stress and strain vectors as where
 
[σ]γ = σx x σ yy σzz τx y τ yz τx z εx x = ∂u
ε yy = ∂v
εzz = ∂w
  ∂x ∂y ∂z
∂u ∂v ∂v
[ε]T = εx x ε yy εzz γx y γ yz γx z γx y = ∂y + ∂x γ yz = ∂z + ∂w
∂y γx z = ∂u
∂z + ∂w
∂x
E and G are young’s modulus and shear modulus,
respectively
Dynamics Lagrange’s equation

d ∂T ∂T ∂
where dt ∂ q̇i − ∂qi + ∂qi = Q i (i = 1, 2, 3 · · · n)
t = time
T = kinetic energy of the system
qi = coordinate system
q̇ = time derivate of coordinate system
 = potential energy of the system
Qi = non-conservative forces or moments

Heat transfer problem Fourier law of conduction q = −k · ∇T

Radiation over Convection on

surface surface Newton’s law of convection (BC)q = h A Ts − T f
Stefan-Boltzmann law of radiation (BC)
Tc

P = eσ A T 4 − TC4
where
q = heat flux
Conduction in
k = conductivity coefficient
T
body T = temperature
h = convection coefficient
T s = temperature on surface
T f = temperature in surrounding fluid
P = net radiated power
A = radiating area
σ = Stefan’s constant
E = emissivity
T c = temperature of surrounding
(continued)
134 3 Computer-Aided Engineering (CAE)

Table 3.3 (continued)

Engineering applications Laws of mechanics and differential equations
Electromagnetics Maxwell’s equations
Gauss’ law for electricity
∇ · (ε0 E + P) = ρ
Gauss’s law for magnetism
∇·B=0
E = electric field Faraday’s law of induction
B = magnetic field
∇ × E = ∂∂tB
D = electric displacement
Ampere’s law
H = magnetic field strength ∂D
ρ = charge density ∇×H =J+ ∂t
ε0 = permittivity
D = (ε0 E + P)
μ0 = permeability where
J = current density B = μ0 (H + M)
C = speed of light
∇ · (•) and ∇ × (•) are the divergence and curl
P = polarization
operations of vectors, respectively
I = electric current
M = magnetization
Fluid mechanics Law of conservation of mass
∂ρ ∂ρu ∂ρv ∂ρw
∂t + ∂ x + ∂ y + ∂z = 0
where
t = time,
dz ρ = density,
u, v, and w are the velocities along x, y, and z axis,
respectively
For a steady flow: ∂ρ
∂t = 0
dy From impressive flow where the density is constant:
∂u ∂v ∂w
∂x + ∂y + ∂z =0
dx

these problems. Many engineering problems have similar governing equations, even
though these problems are formulated in different disciplines. Figure 3.13 has shown
a variety of engineering problems, which are governed by the well-known Poisson
equations (Chandrupatla and Belegundu 2012). If an FEA tool includes a solver to
the Poisson equations, it can be applied to solve all of these problems.

3.5.2 Solution by Direct Formulation

In direct formulation, an element model is derived from the governing equations

directly. Figure 3.14 shows an example of modeling a spring element by direct
formulation.
The spring element in Fig. 3.14 is composed of two nodes, which are denoted as
node i and node j, respectively. The behavior of the element is represented by the
displacements at two nodes U i and U j . U i and U j can be determined by the conditions
3.5 FEA Theory 135

Fig. 3.13 Poisson Equations in different engineering problems

Xj
Xi
Ui Uj

Fi Fj
k

Node_i Node_j
(a) An element in a spring system

Fi F=k ΔU F=k Δ U Fj
=k (Ui-Uj) =k (Uj -Ui )
(b) Force on node i
(c) Force on node j

Fig. 3.14 Modeling of a spring element by direct formulation

of the force balance in the free body diagrams (FBDs) of node i and j in Fig. 3.14b,
c. Note that the behavior of a spring element is governed by the Hook’s law as

F = k ·
U (3.8)
136 3 Computer-Aided Engineering (CAE)

where.
F: the force passes through the spring,
K: the spring coefficient, and
ΔU: the length change of the spring
In an FEA model, the behavior of an element is represented by the field variables
of the associated nodes; in Fig. 3.14, the behavior of the spring element is represented
by U i and U j . By applying the forces F i and F j on the nodes, the forces in the FBDs
of two nodes must be balanced, i.e.,

 
k
Ui − U j  = Fi
(3.9)
k U j − Ui = F j

Equation (3.9) involves in two equations for two field variables U i and U j ; it is
not solvable until the forces F i and F j are known. In turn, F i and F j are determined
by boundary conditions. For example, if the left node of the spring element is fixed
and the right node is applied a force of F, the boundary conditions generate two
additional equations for Eq. (3.9), i.e., U i = 0, and F j = F; then, the element model
becomes solvable to obtain the element solution as F i = − F and U j = F/k.
The above spring element model is generic and can be extended to solve a variety of
simple engineering problems. When system elements respond external loads linearly,
they can be modeled as equivalently as spring elements. In an equivalent spring
system, the solution can be derived by mapping their physical parameters, external
loads, and field variables into spring rate (k), forces (F i and F j ), and the displace-
ment in a spring element (U i and U j ). Table 3.4 gives the mappings of some basic
engineering problems to equivalent spring systems.
Taking an example of a resistor element in Table 3.4, its element model can be
defined by mapping all of physical properties and quantities to those of a spring
element, i.e.
⎫
1
R
→k ⎬
V →U (3.10)
⎭
I →F

Replacing the corresponding parameters and variables in Eq. (3.9) gets the model
of a resistor element (R) as
    
1/R −1/R Vi I
− = i (3.11)
1/R 1/R Vj Ij

Example 3.2. Find the voltage drop in each resistor of the circuit in Fig. 3.15.
Solution. Each resistor is treated as an equivalent spring element; accordingly, the
circuit in Fig. 3.15a is decomposed in the equivalent spring system in Fig. 3.15b. It
3.5 FEA Theory 137

Table 3.4 Similarity of a spring system with other engineering systems

Engineering problems Equivalent terms in a spring element
Stiffness Load Field variable U
k F
EA
Pi L P u

ui E, A, L Axial Member
E: Young’s modulus
A: Cross-section area
L: length
P: Axial load
uj Pj Y: Displacement

GJ
L T θ

π D4
128μ Q p

1
R I V

kA
L q T

consists of 4 elements and 4 nodes. Table 3.5 shows the result of decompositions and
corresponding element models.

Assembling the element models in Table 3.5 gets the system model as
⎡ ⎤⎡ ⎤ ⎡ ⎤ ⎡ ⎤
0.2 −0.2 0 0 V1 I1,1 1
⎢ ⎥⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎢ −0.2 0.2 + 0.1 + 0.05 −0.1 − 0.05 0 ⎥⎢ V2 ⎥ ⎢ I2,1 + I2,2 + I2,3 ⎥ ⎢ 0 ⎥
⎢ ⎥⎢ ⎥ = ⎢ ⎥←⎢ ⎥ (3.12)
⎣ 0 −0.1 − 0.05 0.1 + 0.05 + 0.1 −0.1 ⎦⎣ V3 ⎦ ⎣ I3,2 + I3,3 + I3,4 ⎦ ⎣ 0 ⎦
0 0 −0.1 0.1 V4 +I4,4 −1
138 3 Computer-Aided Engineering (CAE)

R1=5 k1=0.2
1
I1 =1
2

I=1A R2=10 R3=20 k2=0.1 k3=0.05

4
I4= 1
R4=10 V4= 0 k4=0.1 3
(a) Circuit system (b) Equivalent spring system

Fig. 3.15 FEA modeling of a circuit example

Fig. 3.16 Types of potential energy in a solid domain

Table 3.5 The decomposition of equivalent spring system

Element Node i Node j ke (1/R) Element model
    
1 ➊ ➋ 0.2 0.2 −0.2 V1 I1,1
=
−0.2 0.2 V2 I2,1
    
2 ➋ ➌ 0.1 0.1 −0.1 V2 I2,2
=
−0.1 0.1 V3 I3,2
    
3 ➋ ➌ 0.05 0.05 −0.05 V2 I2,3
=
−0.05 0.05 V3 I3,3
    
4 ➌ ➍ 0.1 0.1 −0.1 V3 I3,4
=
−0.1 0.1 V4 I4,4
3.5 FEA Theory 139

The circuit system has two boundary conditions: (1) the current at nodes 1 and 4
are given as 1 A and − 1 A, respectively, and no additional current is input at node 2
and node 3; (2) node 4 is grounded as a reference (V 4 = 0 v). Therefore, the system
model Eq. (3.12) can be simplified as

(3.13)

Solving Eq. (3.13) obtains the solution at [V] = [21.6667, 16.6667, 10, 0]T , and
the voltage drops over the resistors R1 , R2 , R3 , and R4 are 5, 6.6667, 6.6667, and 10
voltage, respectively.

3.5.3 Principle of Minimum Potential Energy

The behavior of an engineering system can be described from the perspective of

energy and work. In a computational domain, energy can be transferred from one
form to another, but it never vanishes. For a system in the steady condition, the
principle of the minimum potential energy can be utilized to model system behaviors.
Figure 3.16 shows a solid domain subjected to a set of external loads F i (i = 1,
2,….n). The response of the solid domain is represented by (1) the displacements
ui (i = 1, 2, ….n) at the locations where the external loads are applied and (2) the
distributions of stress σ (x, y, z) and strain ε(x, y, z) in the solid domain. The total
potential energy in the solid domain is expressed as

1  n
=  − W = ∫ σ (x, y, z) · ε(x, y, z)dv − Fi · u i (3.14)
V 2
i=1

where is the total potential energy,  is the strain energy in solid domain, and W
is total work done by external forces F i (i = 1, 2,….n).
The principle of minimum potential energy states that an equilibrium system meets
the conditions of minimizing the total potential energy of the system as

∂ ∂( − W )
= = 0, (i = 1, 2, . . . n) (3.15)
∂u i ∂u i

Example 3.3 Use the principle of minimum potential energy to find the displacement
at the free end in Fig. 3.17.
Solution. The solid domain in Fig. 3.17a is decomposed into an FEA model in
Fig. 3.17b, which consists of 4 elements and 4 nodes. The details of FEA model are
summarized in Table 3.6.
140 3 Computer-Aided Engineering (CAE)

Fig. 3.17 A two-dimensional plate subjected to axial load

Table 3.6 The decomposition of equivalent spring system

Element Node i Node j ε A(in2 ) Strain energy (lbf.in)
u 2 −u 1  10 E A1
u 2 −u 1 2
1 ➊ ➋ 10 10 0 2 10 dx
u 3 −u 2  10 E A2
u 3 −u 2 2
2 ➋ ➌ 10 2 0 2 10 dx
u 3 −u 2  10 E A3
u 3 −u 2 2
3 ➋ ➌ 10 2 0 2 10 dx
u 4 −u 3  10 E A4
u 4 −u 3 2
4 ➌ ➍ 5 10 0 2 5 dx

Using Eq. (3.14) gets the total potential energy as

= 1.45 × 107 (u 2 − u 1 )2 + 5.8 × 106 (u 3 − u 2 )2 + 5.8 × 107 (u 4 − u 3 )2

− (R)u 1 − (5000)u 4 (3.16)

Using Eqs. (3.15) and (3.16) determine the conditions for minimum potential
energy as

∂
⎫
∂u 1
=0→ −2.9 × 107 (u 2 − u 1 ) = R ⎪
⎪
⎪
∂
∂u 2
= 0 → 2.9 × 107 (u 2 − u 1 ) − 11.6 × 106 (u 3 − u 2 ) = 0 ⎬
∂ (3.17)
= 0 → 11.6 × 106 (u 3 − u 2 ) − 11.6 × 107 (u 4 − u 3 ) = 0 ⎪
⎪
∂u 3
∂
⎪
⎭
∂u 4
=0→ 11.6 × 107 (u 4 − u 3 ) = 5000

Equation (3.17) can further be simplified as

⎡ ⎤⎡ ⎤ ⎡ ⎤
10 −10 0 0 u1 R
⎢ −10 14 −4 0 ⎥ ⎢ ⎥ ⎢ ⎥
2.9 × 106 ⎢ ⎥⎢ u 2 ⎥ = ⎢ 0 ⎥ (3.18)
⎣ 0 −4 44 −40 ⎦⎣ u3 ⎦ ⎣ 0 ⎦
0 0 −40 40 u4 5000
3.5 FEA Theory 141

Applying the boundary condition that the left side is fixed, i.e., u1 = 0, Eq. (3.18)
leads to the solution of [u] as
T
[u] = 0.0 1.72 × 10−4 6.03 × 10−4 6.47 × 10−4

3.5.4 Weighted Residual Methods

The direct method and the principle of the minimum potential energy are applicable
to a limited number of engineering problems; they lack the flexibility to solve a
broad scope of partial differential equations (PDEs). There is a demand to develop
the methods that are more generic for PDEs. In this section, the object under an axial
load in Fig. 3.18 is used as an example to illustrate the concept of approximated
solution, and the weighted residual method is discussed to approximate the solution
to a PDE. The object has a varying cross-sectional area along its length to carry the
load.
Drawing the FBD of the finite segment of the solid domain identifies all of three
forces and the condition of force equivalent as
⎫
p(x) = E · b(x)(1) u(x)−u(x+
x)
x ⎬
p(x +
x) = E · b(x +
x)(1) u(x+
x)−u(x) (3.19)

x ⎭
p(x +
x) + b(x) · (1) · ρ · g ·
x = p(x)

Simplifying Eq. (3.19) gets

Fig. 3.18 A two-dimensional plate subjected to an axial load

142 3 Computer-Aided Engineering (CAE)

u(x +
x) − u(x)
E · {b(x +
x) − b(x)} + b(x) · ρ · g ·
x = 0 (3.20)

x
Taking the limit of
x → 0 in Eq. (3.20) gets

d(b(x)) d(u(x))
E + b(x) · ρ · g = 0 (3.21)
dx dx
Adding the boundary conditions in Eq. (3.21) gets the mathematic model of the
boundary value problem (BVP) as
⎫
d(u(x))
dx 
+ f (x) = 0 ⎪
⎬
d(u(x)) 
d x x=L
= E·WP
(3.22)
L ⎪
⎭
u(x)|x=L = 0

 
where f (x) = b(x) · ρ · g/ E · d(b(x))
dx
The exact solution u(x) of the displacement to the boundary value problem
Eq. (3.22) is defined as
⎫
(x) = d(u(x))
dx
+ f (x) ≡ 0 x ∈ [0, L] ⎪
⎬
d(u(x)) 
d x x=L
= E·W
P
(3.23)
L ⎪
⎭
u(x)|x=L = 0

where R(x) is the residual function over the solid domain of x ∈ [0, L].
When an approximate solution u(x) is used in Eq. (3.23) instead of the exact
solution u(x), the first equation cannot be satisfied strictly. In the approximation, the
conditions of a weak solution are relaxed as
L L   ⎫
0 v(x) · (x)d x = 0 v(x) · d(u(x))
dx
+ f (x) dx = 0 ⎪
⎪
⎬
d(u(x)) 
d x x=L
= E·W L
P
⎪
(3.24)
⎪
⎭
u(x)|x=L = 0

where (x) is the residual function for u(x) over the solid domain of x ∈ [0, L]; v(x)
is a test function.
For any function u(x), the solution to BVP can be approximated by a part of the
Taylor expansion of the function u(x) as

∂u(x0 ) 1 ∂ 2 u(x0 ) n
2 + · · · + 1 ∂ u(x 0 ) (x − x )n
ũ(x) = u(x0 ) + (x − x0 ) + (x − x 0 ) 0
2x 2! ∂ x 2 n! ∂ x n
(3.25)
3.5 FEA Theory 143

Let x 0 = 0, Eq. (3.25) can be simplified as a polynomial equation as

ũ(x) = c0 + c1 x + c2 x 2 + . . . + c N −1 x N −1 (3.26)

where ci (i = 0, 1, 2,…, N−1) are N constants which are determined by the weighted
residual method.
Note that a total of N equations are needed to solve N unknown constants in
Eq. (3.26), a few of the equations correspond to the boundary conditions of PDE,
and the rest of the equations are defined by using different test functions in the
expression of the residual over the domain. Table 3.7 shows some commonly used
test functions.
A weighted residual method is reduced to find a set of coefficients ci (i = 0, 1, 2, …,
N−1) in u(x); it will be based on the conditions for a weak solution to the BVP as
below.
(1) Determine the number of the polynomial terms (N). This number should be
larger than the sum of the numbers of differential equation(s) and BCs in BVP,
i.e.,

N ≥ N D + N BC (3.27)

Taking an example of BVP in Eq. (3.23), it has one differential equation (N D =

1) and two BCs (N BC = 2). Therefore, the included polynomial terms of the Taylor
expansion in the approximation should be equal to or larger than N D + N BC = 3.
The more polynomial terms are used, the better approximation can be achieved but
with much more computation.

Table 3.7 Common test functions in a weighted residual method

Test function Expression Description

Collection method 1 x = xi The test function has the unit value
vi (x) = only at the specified point (x i ) in the
0 x ∈ (a, b), x
= xi domain (a, b) of x
Subdomain method vi (x) = The test function has the unit value
only in the specified subdomain
1 x ∈ (xi , xi+1 )
(xi , xi+1 ) in the domain (a, b) of x
/ (xi , xi+1 )
0 x ∈ (a, b), x ∈
∂(c0 ,c1 ,...c N −1 ,x)
Least-square method vi (x) = αci The test function aims to minimize
the least square residuals over the
domain; the number of the test
functions determines that of the
constants to be determined
Galerkin method vi (x) = φi (x) One of the sub-functions in used the
approximated solution ũ(x) as the
test function
144 3 Computer-Aided Engineering (CAE)

(2) BCs in BVP are applied to define N BC number of the equations, which should
be satisfied by ũ(x).
(3) Select (N-N BC ) number of test functions, and apply the condition Eq. (3.24)
for a weak-solution. This generates (N-N BC ) number of new equations, which
should be satisfied by ũ(x) . Combining with N BC equations for boundary
conditions, a system of N number of the equations are developed for N number
of the constants ci (i = 0, 1, 2, …, N−1) in u(x).
(4) Solve N equations simultaneously to obtain c0 , c1 , …, cN−1 for the complete
solution of u(x) to VBP.
In above procedure, the number of the polynomial terms (N) and the types of test
functions can be chosen arbitrarily as long as N satisfies Eq. (3.27). Different test
functions lead to the different approximated solutions to BVP; but all of them have
to meet the conditions for a weak solution in Eq. (3.24).
Example 3.4 Find a weak solution to the following boundary value problem:

d 2u
= 6x − sin(x), 0 ≤ x ≤ 1
dx2
subjected to :
⎫
u(x)|x=1 = sin(1)⎬
du (3.28)
|x=0 = 0 ⎭
dx

Solution. The weighted residual method is used in the steps as follows.

Step 1. Equation (3.28) has one differential equation (N D = 1) and two boundary
conditions (N BC = 2). Therefore, the number of the polynomial terms (N) must be
larger than N D + N BC = 3. Let N be 4, and the approximate solution is assumed as

ũ(x) = c0 + c1 x + c2 x 2 + c3 x 3 (3.29)

Equation (3.29) has unknown constants (c0 , c1 , c2 , c3 ); thus, four equations about
(c0 , c1 , c2 , c3 ) are derived from the conditions of the weak solution.
Step 2. Equation (3.28) involves in two boundary conditions, substituting
Eq. (3.29) into Eq. (3.28) gives

ũ(x)|x=0 = c0 + c1 (1) + c2 (1)2 + c3 (1)3 = sin(1)

d ũ(x)  (3.30)
dx 
= c1 + 2c2 (0) + 3c3 (0)2 = 0
x=0

Using Eq. (3.30) in Eq. (3.29) gets

ũ(x) = (sin(1) − c2 − c3 ) + c2 x 2 + c3 x 3 (3.31)

Equation (3.31) is used to obtain the first and second derivatives of ũ(x) as
3.5 FEA Theory 145

d ũ(x)
dx
= 2c2 x + 3c3 x 2
d 2 ũ(x) (3.32)
dx2
= 2c2 + 6c3 x

Step 3. Equation (3.31) include two unknown constants (c2 , c3 ). Therefore, two
different test functions have to be selected to define another two equations by using
the condition of a weak-solution (Eq. 3.23). Using Eq. (3.32) into Eq. (3.28) yields
the expression of the residual as
! 1
(x) = v(x) · ((2c2 + 6c3 x) − 6x + sin(x))d x = 0 (3.33)
0

Next, different methods are used in selecting test functions v(x) to evaluate
residuals in Eq. (3.33).
(a). In a collocation method, the test function is selected as
"
1 x = xi
vi (x) = (3.34)
0 x ∈ (a, b), x
= xi

where x i is an arbitrary position in the domain of interest.

Since only two more equations are needed to solve c2 and c3 ; the condition of
the weak-solution is applied twice. Let x 1 = 1/3 and x 2 = 2/3 in Eq. (3.24) and
substituting it in Eq. (3.33), respectively,



 
1 =
2c2 + 6c3
13  − 6
13  + sin(x) = 0
(3.35)
1 = 2c2 + 6c3 23 − 6 23 + sin(x) = 0

Solving Eq. (3.35) yields c2 = −0.01801 and c3 = 0.8544. The approximated

solution becomes

ũ(x) = 0.00508 − 0.01801x 2 + 0.8544x 3 (3.36)

(x) by using the collocation method is plotted in Fig. 3.19. ũ(x) in Eq. (3.36)
satisfies the condition of a weak solution in sense that there is no error at a position
specified by the test function in Eq. (3.34).
(b). In the sub-domain method, the test function is defined as
"
1 x ∈ (xi , xi+1 )
vi (x) = (3.37)
0 x ∈ (a, b), x ∈ / (xi , xi+1 )

where (ai , bi ) is an arbitrary sub-domain of interest.

146 3 Computer-Aided Engineering (CAE)

0.07

0.06

0.05 R(x)
0.04

0.03

0.02
x1=1/3 x2=2/3
0.01

0
0 0.2 0.4 0.6 0.8 1
-0.01
x

Fig. 3.19 Approximated solution using collocation methods

Two additional equations are needed to solve c2 and c3 . Therefore, set two subdo-
mains as (0, 0.5) and (0.5, 1.0), and use Eq. (3.27) in Eq. (3.3) to test the condition
of a weak solution as

  ⎫
1 = 00.5 ((2c2 + 6c3 x) − 6x + sin(x))d x = c2 + 43 c3 + 41 − cos 21 = 0 ⎬
1 
2 = 0.5 ((2c2 + 6c3 x) − 6x + sin(x))d x = c2 + 49 c3 − 49 + cos 21 − cos(1) = 0 ⎭
(3.38)

Solving Eq. (3.38) yields c2 = − 0.01499 and c3 = 0.85676, and the approximated
solution becomes

ũ(x) = 0.000299 − 0.01499x 2 + 0.85676x 3 (3.39)

(x) by using the subdomain method is plotted in Fig. 3.20. ũ(x) in Eq. (3.39)
satisfies the condition of a weak solution in the sense that there is no error in the
sub-domain specified by the test function in Eq. (3.37).
(c). In a least-square method, the test function is selected as

∂(c0 , c1 , . . . c N −1 , x)
vi (x) = (3.40)
∂ci

where ci is an elected constant to be determined. The motivation of a least-square

method is to minimize the absolute value of the residual integral of the whole domain.
To achieve this, c2 and c3 must be chosen to minimize the residual integral,
3.5 FEA Theory 147

0.02
R(x)
0.01

0
0 0.25 0.5 0.75 1
-0.01

-0.02 Subdomain 2

-0.03

-0.04

-0.05

Fig. 3.20 Approximated solution using sub-domain methods

! 1
2
R̃ = R (x)d x = 0 (3.41)
0

Minimizing R̃ leads to the following conditions of c2 and c3 ,

∂ R̃
1 ∂R
∂c2
=2 ∂c2
R(x)d x = 0
∂ R̃
01∂R
(3.42)
∂c3
=2 0 ∂c3 R(x)d x = 0

For the BVP in Eq. (3.28), the test functions are defined as

∂2 (c0 ,c1 ,c2 ,c3 ,x)

v1 (x) = ∂c2
=2
∂n(c0 ,c1 ,c2 ,c3 ,x) (3.43)
v2 (x) = ∂c3
= 6x

Using Eq. (3.43) for the condition of a weak solution yields

1
1 = 0 (2)((2c2 + 6c3 x) − 6x + sin(x))d x = (2)(2c
2 + 3c3 − 3 − cos(1)) = 
0
1 1
2 = 1 (6x)((2c2 + 6c3 x) − 6x + sin(x))d x = (6) c2 + 2c3 − 2 + 0 x sin(x)d x = 0
(3.44)

Solving Eq. (3.44) yields c2 = −0.012462 and c3 = 0.855076, and the

approximated solution becomes

ũ(x) = −0.00114 − 0.012462x 2 + 0.855076x 3 (3.45)

148 3 Computer-Aided Engineering (CAE)

0.03

0.02

0.01

0
0 0.2 0.4 0.6 0.8 1
-0.01

-0.02

-0.03

-0.04

-0.05

-0.06

Fig. 3.21 Approximated solution using least-square methods

(x) by using the Least square method is plotted in Fig. 3.21. ũ(x) in Eq. (3.45)
satisfies the condition of a weak solution in the sense that the integral of absolute
residual over the whole domain is minimized.
(d). In a Galerkin method, the test functions are selected from the constitutive
functions in the expression of the approximated solution ũ(x) in Eq. (3.31), which
includes three constitutive functions of

φ1 (x) = 1, φ2 (x) = x 2 , φ3 (x) = x 3 (3.46)

φ 1 (x) and φ 2 (x) are selected as test functions to determine c2 and c3 ,

1
1 = 0 φ1 (x)((2c2 + 6c3 x) − 6x + sin(x))d x = (2c2 + 3c3 − 2 − cos(1)) = 0
1 1
R 2 = 1 φ2 (x)((2c2 + 6c3 x) − 6x + sin(x))d x = 23 c2 + 23 c3 − 23 + 0 x 2 sin(x)d x = 0
(3.47)

Solving Eq. (3.47) yields c2 = − 0.019815 and c3 = 0.8599773, and the

approximated solution becomes

ũ(x) = 0.00131 − 0.019815x 2 + 0.85599773x 3 (3.48)

(x) by using the Galerkin method is plotted in Fig. 3.22. ũ(x)in Eq. (3.36) satisfies
the condition of a weak solution in the sense that the integral of a weighted residual
over the whole domain is minimized.
In using a weighted residual method, different test functions are defined to satisfy
the condition of a weak solution. By specifying certain test function, the accuracy
of the approximated solution ũ(x) varies from one BVP to another. Therefore, the
3.5 FEA Theory 149

0.02

0.01

0
0 0.2 0.4 0.6 0.8 1
-0.01

-0.02

-0.03

-0.04

-0.05

-0.06

-0.07

Fig. 3.22 Approximated solution using the Galerkin methods

performances of different weighted residual methods are not comparable. The accu-
racy of an approximation method can be verified when an exact solution is available
to original BVP.
Note that the exact solution is validated by comparing the approximated solution
with the exact one if it is available. The exact solution to the BVP in Eq. (3.28) can
be found by the integration with the given boundary conditions as

u(x) = x 3 + sin(x) − x (3.49)

Table 3.8 and Fig. 3.23 show the comparison of the exact solution (Eq. (3.49)) with
the approximated solutions (Eqs. (3.36), (3.39), (3.45), and (3.48)) over the whole
domain of x ∈ (0, 1). The comparison has shown that all of the four weight-residual
methods have resulted in a weak solution ũ(x) with an acceptable approximation
over the whole domain.
In the practice of FEA modeling, the Galerkin method is the most commonly used
one among these four weighted residual methods. In this book, the Galerkin method
will be used as an example to develop element models in FEA.

3.5.5 Types of Differential Equations

The system behavior of an engineering system can be represented by a mathematical

model, especially, partial differential equations (PDEs). Table 3.3 has shown some
engineering systems with corresponding PDEs. The solution to a PDE is that to
150 3 Computer-Aided Engineering (CAE)

Table 3.8 The approximations by using four different test functions

0.9

0.8
u(x)
0.7
collocaƟon
0.6 sub-domain
Least-square
0.5
Galerkin
0.4
u(x)

0.3

0.2

0.1

0
0 0.2 0.4 0.6 0.8 1
-0.1
x

Fig. 3.23 Comparison of exact and approximated solutions

3.5 FEA Theory 151

Fig. 3.24 Domain of

Dependence and Influence of
an Elliptic PDE

an engineering system subjected to the boundary conditions, PDEs can be elliptic,

parabolic, or hyperbolic.
In a 2D domain, an elliptic PDE only contains the derivatives with respect to
spatial variables (x, y), and no preferred path exists to seeking the solution to PDE.
Consequently, the solution at a certain point depends on that of all other points in
the physical domain. An elliptic PDE is used to represent an equilibrium problem
in a closed domain D(x, y), which exists in all of the engineering fields. Figure 3.24
shows the closed solution domain D(x, y) and its boundary B. An elliptic PDE is
usually an ordinary differential equation, and a typical example of an elliptic PDE is
the Laplace equation as

∂ 2φ ∂ 2φ
cx + c y 2 + cφ = 0 (3.50)
∂ x
2 ∂ y

where φ (x, y) is a state variable such as the temperature in a heat transfer problem.
It is subjected to the boundary conditions of

 1 : φ = φ0
(3.51)
2 : aφ + b ∂∂ 2φn = 0
2

where Γ 1 and Γ 2 correspond to essential and natural boundaries and ∂φ ∂n

denotes the
derivative normal to the boundary.
A parabolic PDE contains only second-order spatial derivatives; the preferred
path in seeking the solution to PDE is along the line or surface of the constant
time. At each time step, the solution at one point are dependent on that of all other
points. A propagation problem is an initial-value problem in an open domain. One
of independent variables, for example, of time, does not have a closed range. The
solution f (x, y, t) in the domain D(x, y, t) is marched forward from the initial state,
guided and modified by the given boundary conditions. A propagation problem is
152 3 Computer-Aided Engineering (CAE)

governed by a parabolic or hyperbolic PDE. The majority of the propagation problems

are unsteady problems. Taking an example, the following equation stands for an
unsteady propagation problem,
# 2 $
∂φ ∂ φ ∂ 2φ
=α 2 + 2 (3.52)
∂t ∂ x ∂ y

where φ (x, y) is a state variable and its initial condition is given as


φ x, y, t0 = φ0 (x, y) at t = t0

The solution to a parabolic or hyperbolic PDE is propagating from the initial

property distribution at time t o . The solution is marched forward in the given time
step in the time domain.
A hyperbolic PDE corresponds to an eigenvalue problem. The solution to a hyprt-
bolic PDE exists only for special values (i.e., eigenvalues) of system parameters. An
eigenvalue problem is solved to obtain eigenvalues and the corresponding system
configurations A typical hyperbolic PDE is shown as

∂ 2φ ∂ 2φ
+ + λφ = 0 (3.53)
∂2x ∂2 y

where φ (x, y) is a state variable and the boundary conditions are given as

φ = φ0 on 1
∂φ
+ αφ = 0 on 2
∂n
The numerical solution treats a continuous domain as a set of discrete nodes
and elements. It calculates the values of the state variables on these discrete nodes.
The calculation is repeated whenever a system parameter or boundary condition is
changed in the model.

3.5.6 Solutions to Different PDEs

The solutions to three types of PDEs for 2D rectangle and triangle elements are
discussed in the section.
3.5 FEA Theory 153

3.5.6.1 Solution to Elliptic PDEs

An equilibrium problem represented by elliptic PDEs is a steady problem, and the

solution at any position is influenced by these of all other points in the domain, and the
solutions to all nodes in the domain are modeled and obtained simultaneously. In this
section, the Galerkin method is applied to develop an element model for an elliptic
PDE in Eq. (3.50), which is subjected to the boundary conditions of Eq. (3.51).
The weak-form solution to the PDE meets the condition of
! # $
∂ 2φ ∂ 2φ
R = w(x, y) cx 2 + c y 2 + cϕ − 0 d xd y = 0 (3.54)
∂ x ∂ y

where w(x, y) is the selected test function and cx , cy and cφ are constants.
Equation (3.54) can be expressed by the reduced derivatives as below
⎧     ⎫
! ⎨ ∂ w(x, y) ∂φ ∂ w(x, y) ∂φ ⎬
∂x ∂w(x, y) ∂φ ∂y ∂w(x, y) ∂φ
R= cx { − } + cy { − } + cϕ · w(x, y) d xd yφ
⎩ ∂x ∂x ∂x ∂y ∂y ∂y ⎭
(   ! (  
∂φ ∂w(x, y) ∂φ ∂φ
= cx ∂ w(x, y) dy − cx d xd y + c y ∂ w(x, y) dx
∂x ∂x ∂x ∂y
) *+ , ) *+ , ) *+ ,
I II III
! !
∂w(x, y) ∂φ
− cy
∂y ∂y
d xd y + w(x, y) · cφ d xd y (3.55)
) *+ , ) *+ ,
IV V

where the residual caused by approximation consists of five parts (i.e., I, II, III, IV,
and V ). The parts of I and III are the integrals along the boundary of the 2D domain.
The parts of II, IV and V are the integrals in the domain.
Figure 3.25 shows a heat transfer domain, which is governed by an elliptic PDE.
The continuous domain has its two types of boundaries, i.e., essential boundary Γ 1

Fig. 3.25 Rectangle and triangle elements in a heat transfer domain

154 3 Computer-Aided Engineering (CAE)

and natural boundary Γ 2 . In addition, rectangle and triangular elements are shown
as the examples of discretized elements and nodes.
To solve Eq. (3.55) for an elemental model, this section limits the discussions
on the parts of II, IV and V, since the parts of I and III are the integrals over the
boundary surfaces rather than the others in the solid domain.
For a rectangle element in Fig. 3.25, a state variable at a given position (x, y) can
be determined by the values at four nodes (φ i , φ j , φ m , φ n ) using the shape functions
as

φ = [S]{ϕ} = Si (x, y) · φi + S j (x, y) · φ j + Sm (x, y) · φm + Sn (x, y) · φn

(3.56)

where φ(x, y) is the state variable at an arbitrary position (x, y) in the element, φ i ,
φ j , φ m , and φ n are the values of the state variable at nodes, and the shape functions
in a rectangle elements are


w−y  ⎫
Si (x, y) = l−x

l w ⎪
⎪
⎬
S j (x, y) = xl w−yw (3.57)
Sm (x, y) = xl wy ⎪
⎪

y ⎭
Sn (x, y) = l−x l w

where l and w are the length and width of the rectangle element.
Since the Galerkin method is used, the shape functions S i (x, y), S j (x, y), S m (x, y),
and S n (x, y) in Eq. (3.56) are used as the test functions, respectively. Substituting the
shape functions into Eq. (3.55) obtain the parts of II, IV, and V as below.
Part II for the rectangle element is found as
! !
∂[S]T ∂φ ∂ S]T d[S][φ])
cx d xd y = cx d xd y = [K x ][φ] (3.58)
∂x ∂x ∂x ∂x

where
⎡ ∂ Si
 ∂S   ⎤
· ∂∂Sxi d xd y ∂∂Sxi · ∂ xj d xd y ∂ Si
· ∂∂Sxm d xd y ∂ Si
· ∂∂Sxn d xd y
⎢ ∂x
∂S
∂ Si
 ∂ Sj ∂S  ∂x
∂ Sj  ∂x
∂ Sj ⎥
⎢ · ∂ xj d xd y · ∂ xj d xd y · ∂∂Sxm d xd y · ∂∂Sxn d xd y ⎥
[K x ] = cx ⎢  ∂x
∂ Sm
 ∂x
∂S  ∂x  ∂x ⎥
⎣ · ∂∂Sxi d xd y ∂∂Sxm · ∂ xj d xd y ∂ Sm
· ∂∂Sxm d xd y ∂ Sm
· ∂∂Sxn d xd y ⎦
 ∂x
∂ Sn
 ∂ Sj  ∂x  ∂x
· ∂∂Sxi d xd y ∂∂Sxn
∂x
· ∂ x d xd y ∂ Sn
∂x
· ∂∂Sxm d xd y ∂ Sn
∂x
· ∂∂Sxn d xd y
⎡ ⎤
2 −2 −1 1
cx w ⎢
⎢ −2 2 1 −1 ⎥
⎥
=
6l ⎣ −1 1 2 −2 ⎦
1 −1 −2 2
3.5 FEA Theory 155

and [φ]T = φi φ j φm φn
Part IV for the rectangle element is found as
! !
∂[S]T ∂φ ∂[S]T ∂[S][φ]) 
cy d xd y = c y d xd y = K y [φ] (3.59)
∂y ∂y ∂y ∂y

where
⎡  ∂S   ⎤
∂ Si
· ∂∂Syi d xd y ∂∂Syi
∂y
· ∂ yj d xd y ∂ Si
∂y
· ∂∂Sym d xd y ∂ Si
∂y
· ∂∂Syn d xd y
⎢ ∂S∂ Si
 ∂ Sj ∂S  ∂ Sj  ∂ Sj ⎥
 ⎢
⎢ · ∂ yj d xd y
∂y
· ∂ yj d xd y · ∂∂Sym d xd y · ∂∂Syn d xd y ⎥
⎥
K y = cy ⎢   ∂y  ∂y  ∂y
· ∂∂Syn d xd y ⎥
∂ Sm ∂S
⎣ · ∂∂Syi d xd y ∂∂Sym
∂y
· ∂ yj d xd y ∂ Sm
∂y
· ∂∂Sym d xd y ∂ Sm
∂y ⎦
∂ Sn
 ∂S  
· ∂∂Syi d xd y ∂∂Syn
∂y
· ∂ yj d xd y ∂ Sn
∂y
· ∂∂Sym d xd y ∂ Sn
∂y
· ∂∂Syn d xd y
⎡ ⎤
2 1 −1 −2
cy l ⎢
⎢ 1 2 −2 −1 ⎥
⎥
=
6w ⎣ −1 −2 2 1 ⎦
−2 −1 1 2

Part V for the rectangle element is found as

⎡ ⎤ ⎡ ⎤
!  Si d xd y 1
⎢ Sj d xd y ⎥ cφ (lw) ⎢ 1 ⎥
[Q] = [S] cφ d xd y = cφ ⎢
T  ⎥ ⎢ ⎥
⎣ Sm d xd y ⎦ = 4 ⎣ 1 ⎦ (3.60)

Sn d xd y 1

Substituting Eqs. (3.58)–(3.60) into Eq. (3.54) gets the rectangle element model
with no boundary edge (no Part I or III) as


[K x ] + K y {φ} = [ Q] (3.61)

where [K x ] and [K y ] are obtained in Eqs. (3.58) and (3.59), and the load vector [Q]
is given in eq. (3.60).
Similarly, for a triangle element in Fig. 3.25, a state variable at a given position
(x, y) can be determined by the values at four nodes (φ i , φ j , φ k ) using the shape
functions as

φ = [S]{φ} = Si (x, y)φi + Si j (x, y)φ j + Sm (x, y)φm + Sn (x, y)φn (3.62)

where φ(x, y) is the state variable at an arbitrary position (x, y) in the triangle element,
and φ i , φ j and φ k are the values of state variables at nodes. The shape functions in a
triangular element are given as (Bi 2018)
156 3 Computer-Aided Engineering (CAE)

|
i | αi +βi ·x+δi ·y ⎫
Si (x, y) = |
| = |
| ⎪
⎬
|
j | α j +β j ·x+δ j ·y
S j (x, y) = = (3.63)
|
|
|
k |
|
|
αk +βk ·x+δk ·y
⎪
⎭
Sk (x, y) = |
| = |
|

where the constants α, β, δ, |

| are determined by the coordinates of three nodes (x i ,
yi ), (x j , yj ), and (x k , yk ) as
  ⎫
 1 xi yi  ⎪
⎪
  ⎪
⎪
|
| =  1 x j y j  ⎪
⎪
1 x y  ⎪
⎬
k k
(3.64)
αi = x j yk − xk y j , α j = xk yi − xi yk , αk = xi y j − x j yi ⎪
⎪
⎪
⎪
βi = y j − yk , β j = yk − yi , βk = yi − y j ⎪
⎪
⎪
⎭
δi = xk − x j , δ j = xi − xk , δk = x j − xi

In using the Galerkin method, the shape functions in Eq. (3.62) are used as the
test functions. Using these functions in Eq. (3.55) obtain the parts of II, IV, and V as
below.
Part II for the triangle element is found as
! !
∂[S]T ∂φ ∂[S]T ∂[S][φ])
cx d xd y = cx d xd y = [K x ][φ] (3.65)
∂x ∂x ∂x ∂x

where
⎡  ∂ Sj  ⎤
∂ Si
· ∂∂Sxi d xd y ∂ Si
· d xd y ∂ Si
· ∂∂Sxk d xd y
⎢ ∂x
∂ Si ∂S  ∂x
∂ Sj
∂x
∂ Sj  ∂x
∂ Sj ⎥
[Kx ] = cx ⎣ · ∂ xj d xd y · d xd y · ∂∂Sxk d xd y ⎦
 ∂x
∂ Sk
 ∂x ∂x
∂ Sj  ∂x
∂x
· ∂∂Sxi d xd y ∂ Sk
∂x
· ∂x
d xd y ∂ Sk
∂x
· ∂∂Sxk d xd y
⎡ ⎤
βi2 βi β j βi βk
cx ⎢ ⎥
= ⎣ β j βi β 2j β j βk ⎦
4A
βk βi βk β j βk2


and [φ]T = φi φ j φk
Part IV for the triangle element is found as
! !
∂[S]T ∂φ ∂ S]T ∂[S][φ]) 
cy d xd y = c y d xd y = K y [φ] (3.66)
∂y ∂y ∂y ∂y

where
3.5 FEA Theory 157
⎡ ∂ Si
 ∂ Sj  ⎤
· ∂∂Syi d xd y
∂y
∂ Si
∂y
· ∂y
d xd y ∂ Si
∂y
· ∂∂Syk d xd y
 ⎢ ∂S
∂ Si
 ∂ Sj ∂ Sj  ∂ Sj ⎥
K y = cy ⎣ · ∂ yj d xd y
∂y ∂y
· ∂y
d xd y ∂y
· ∂∂Syk d xd y ⎦
 ∂ Sk
 ∂ Sj 
· ∂∂Syi d xd y
∂y
∂ Sk
∂y
· ∂y
d xd y ∂ Sk
∂x
· ∂∂Syk d xd y
⎡ ⎤
δi2 δi δ j δi δk
cy ⎢ ⎥
= ⎣ δ j δi δ 2j δ j δk ⎦
4A
δk δi δk δ j δk2

Part V for the triangle element becomes

⎡ ⎤ ⎡ ⎤
! S d xd y 1
 i
cφ A ⎣ ⎦
[Q] = [S]T cφ d xd y = cφ ⎣ Sj d xd y ⎦ = 1 (3.67)
 3
Sk d xd y 1

Substituting Eqs. (3.65)–(3.67) into Eq. (3.54) gets the triangle element model
with no boundary edge (no Part I or III) as


[K x ] + K y {φ} = [ Q] (3.68)

where [K x ] and [K y ] are defined in Eqs. (3.65) and (3.66) and the load vector [Q] is
defined in Eq. (3.67).

3.5.6.2 Solution to Parabolic PDEs

A parabolic PDE is used to model a unsteady propagate problem. The solution in the
continuous domain is evolved with respect to time, and the initial condition affects
the time-dependent solution. In the following, the procedure of element modeling is
illustrated by using the parabolic PDE as

∂ 2φ ∂ 2φ ∂φ
cx + c y + cφ = α (3.69)
∂2x ∂2x ∂t

where φ(x, y) is the state variable in the element,

cx , c y , cφ , α are the constants, and the initial conditions are


φ x, y, t0 = φ0 (x, y) at t = t0

The solution to Eq. (3.69) is approximated by satisfying the condition of R = 0

where the residual R is expressed as
⎧     ⎫
! ⎨ ∂ w(x, y) ∂φ ∂ w(x, y) ∂φ
∂x ∂w(x, y) ∂φ ∂y ∂w(x, y) ∂φ ∂φ ⎬
R= cx { − } + cy { − } + cφ · w(x, y) − α · w(x, y) d xd y
⎩ ∂x ∂x ∂x ∂y ∂y ∂y ∂t ⎭
158 3 Computer-Aided Engineering (CAE)
(   ! (  
∂φ ∂w(x, y) ∂φ ∂φ
= cx ∂ w(x, y) dy − cx d xd y + c y ∂ w(x, y) dx
∂x ∂x ∂x ∂y
) *+ , ) *+ , ) *+ ,
I II III
! ! !
∂w(x, y) ∂φ ∂φ
− cy
∂y ∂y
d xd y + w(x, y) · cφ d xd y − αw(x, y) ·
∂t
d xd y (3.70)
) *+ , ) *+ , ) *+ ,
IV V VI

Equation (3.70) includes the parts of I, II, III, IV, and V which are the same to
Eq. (3.55) and discussed in the previous section. Only part VI is new.
To evaluate part VI, let ∂φ(x,y)
∂t
be a function of the state variable φ(x, y) which
can be evaluated by taking the derivative of the interpolation in the element as

∂φ(x, y)
= [S]{φ̇} (3.71)
∂t
Therefore, part VI for a rectangle element can be found as
- . - .
∫[S]T [S] φ̇ d xd y = [C] φ̇ (3.72)

where
⎡  2    ⎤
 Si d xd y Si S2 j d xd y  Si Sm d xd y  Si Sn d xd y
⎢ S j Si d xd y S j d xd y S S d xd y S S d xd y ⎥
[C] = α ⎢
⎣  S Si d xd y  S S j d xd y  i 2m  j n ⎥
⎦
 m  m  Sm d xd y S
 2
m S n d xd y
Sn Si d xd y Sn S j d xd y Sn Sm d xd y Sn d xd y
⎡ ⎤
2112
(lw)α ⎢⎢1 2 2 1⎥
⎥
= ⎣
18 1 2 2 1⎦
2112

[ϕ̇] is the vector for the derivative of state variable with respect to time.
Merging Eqs. (3.61) and (3.72) yields the rectangle element model to Eq. (3.69)
as


[K x ] + K y {φ} + [C]{φ} = [ Q] (3.73)

where [K x ], [K y ] and [Q] are given in Eq. (3.61) and [C] is given in Eq. (3.72) for
the rectangle element.
Similarly, [C] in part VI for a triangle element can be found as
⎡  2   ⎤ ⎡ ⎤
 Si d xd y Si S2 j d xd y  Si Sk d xd y Aα
211
[C] = α ⎣ S j Si d xd y S d xd y S S d xd y ⎦ = ⎣1 2 1⎦ (3.74)
  j  k2j 12
Sk Si d xd y Sk S j d xd y Sk d xd y 112
3.5 FEA Theory 159

Merging Eqs. (3.68) and (3.74) yields the triangle element model to Eq. (3.69) as

 - . 
[K x ] + K y {φ} + [C] φ̇ = Q (3.75)

where [K x ], [K y ] and [Q] are given in Eq. (3.61) and [C] is given in Eq. (3.74) for
the triangle element.
- .
Note that φ̇ depends on {φ} and changes over time, the time-dependent solutions
are obtained iteratively based on the given initial condition. For example,
- . the forward
difference method can be used in Eq. (3.73) or Eq. (3.75) to evaluate φ̇ at time step
- .
t based on{φ}. Once φ̇ at time t is obtained, {φ} at the next time step t + Δt can
be found as
- .
{φ}t+
t = {φ}t + φ̇ t
t (3.76)
- .
Once {φ}t+
t is known, φ̇ t can be obtained from two time steps as

- . 1 - .
φ̇ t = {φ}t+
t − {φ}t (3.77)

t
Substituting Eq. (3.77) back to Eq. (3.73) or Eq. (3.75) finds

[C]
 

{φ}t+
t − {φ}t = Q − [K x ] + K y {φ}t (3.78)

t

- .Rearranging Eq. (3.78) leads to an explicit element model without the term of
φ̇ t as
# $
[C]  [C]

{φ}t+
t = Q + − [K x ] + K y {φ}t (3.79)

t
t

In Eq. (3.79), the time step Δt must be set appropriately to the convergence and
stability.

3.5.6.3 Solution to Hyperbolic PDEs

A hyperbolic PDE can be used to model an eigenvalue problem, which has the
solution only for specific values (i.e., eigenvalues). In a 2D domain, the Helmholtz
equation in Fig. 3.26 is used as an example to model an eigenvalue problem.
The hyperbolic PDE for such an eigenvalue problem is

∂ 2φ ∂ 2φ
+ 2 + λφ = 0 (3.80)
∂ x
2 ∂ y
160 3 Computer-Aided Engineering (CAE)

Fig. 3.26 Helmholtz equation for an eigenvalue problem

where φ(x, y) is the state variable and the boundary conditions are given as

 1 : φ = φ0
(3.81)
2 : aφ + b ∂φ
∂n
=0

The solution to Eq. (3.80) is approximated by satisfying the condition of R = 0

where the residual R is expressed as
⎧     ⎫
! ⎨ ∂ w(x, y) ∂φ ∂ w(x, y) ∂φ ⎬
∂x ∂w(x, y) ∂φ ∂y ∂w(x, y) ∂φ
R= { − }+{ − } + w(x, y)λφ dxdy
⎩ ∂x ∂x ∂x ∂y ∂y ∂y ⎭
(   ! (  
∂φ ∂w(x, y) ∂φ ∂φ
= ∂ w(x, y) dy − dxdy + ∂ w(x, y) dx
∂x ∂x ∂x ∂y
) *+ , ) *+ , ) *+ ,
I II III
! !
∂w(x, y) ∂φ
− dxdy + w(x, y) · λφdxdy (3.82)
∂y ∂y
) *+ , ) *+ ,
IV V

where w(x, y) is a selected test function; part I and III are the integrals along the
boundary, part II, IV, and V are three integrals over the continuous domain. The
discussion on the integrals along the boundary (part I and III) will be left out here;
interested readers can find the details in other literatures (Bi 2018). Here, the integrals
over the domain are discussed in modeling rectangle and triangle elements.
In using the Galekin method, the shape functions are used as the test functions in
evaluating the residual in Eq. (3.82), and part II, IV, and V are obtained accordingly.
For part II in the rectangle element,
3.5 FEA Theory 161

∂[S]T ∂φ ∂[S]T ∂([S][φ])

∫ d xd y = ∫ d xd y = [K x ][φ] (3.83)
∂x ∂x ∂x ∂x

where
⎡ ∂ Si
 ∂S   ⎤
· ∂∂Sxi d xd y ∂∂Sxi · ∂ xj d xd y ∂ Si
· ∂∂Sxm d xd y ∂ Si
· ∂∂Sxn d xd y
⎢ ∂x
∂S
∂ Si
 ∂ Sj ∂S  ∂x
∂ Sj  ∂x
∂ Sj ⎥
⎢ · ∂ xj d xd y · ∂ xj d xd y · ∂∂Sxm d xd y · ∂∂Sxn d xd y ⎥
[Kx ] = ⎢  ∂x
∂ Sm
∂ Si
 ∂∂Sxm ∂S  ∂x
∂ Sm
 ∂x ⎥
⎣ · ∂ x d xd y · ∂ xj d xd y · ∂∂Sxm d xd y ∂ Sm ∂ Sn
· ∂ x d xd y ⎦
 ∂x
∂ Sn
 ∂x ∂S  ∂x  ∂x
· ∂∂Sxi d xd y ∂∂Sxn
∂x
· ∂ xj d xd y ∂ Sn
∂x
· ∂∂Sxm d xd y ∂ Sn
∂x
· ∂∂Sxn d xd y
⎡ ⎤
2 −2 −1 1
w ⎢ −2 2 1 −1 ⎥
= ⎢ ⎥
6l ⎣ −1 1 2 −2 ⎦
1 −1 −2 2
T 
and ϕ = φi φ j φm φn .
For part IV in the rectangle element,
! !
∂[S]T ∂φ ∂[S]T ∂[[S][φ]) 
d xd y = d xd y = K y [φ] (3.84)
∂y ∂y ∂y ∂y

where
⎡  ∂S   ⎤
∂ Si
∂y
· ∂∂Syi d xd y ∂ Si
∂y
· ∂ yj d xd y ∂ Si
∂y
· ∂∂Sym d xd y ∂ Si
∂y
· ∂∂Syn d xd y
⎢ ∂ Si ∂S  ∂ Sj ∂S  ∂ Sj  ∂ Sj ⎥
 ⎢ · ∂ yj d xd y · ∂ yj d xd y · ∂∂Sym d xd y · ∂∂Syn d xd y ⎥
Ky =⎢
⎢
∂y  ∂y  ∂y  ∂y ⎥
· ∂∂Syn d xd y ⎥
∂ Sm ∂S
⎣ ∂y
· ∂∂Syi d xd y ∂ Sm
∂y
· ∂ yj d xd y ∂ Sm
∂y
· ∂∂Sym d xd y ∂ Sm
∂y ⎦
∂ Sn
 ∂S  
∂y
· ∂∂Syi d xd y ∂ Sn
∂y
· ∂ yj d xd y ∂ Sn
∂y
· ∂∂Sym d xd y ∂ Sn
∂y
∂ Sn
· ∂ y d xd y
⎡ ⎤
2 1 −1 −2
l ⎢
⎢ 1 2 −2 −1 ⎥
⎥
=
6w ⎣ −1 −2 2 1 ⎦
−2 −1 1 2

For Part V in the rectangle element,

∫[S]T λ[S][φ]d xd y = [M][φ] (3.85)

where
162 3 Computer-Aided Engineering (CAE)
⎡  2    ⎤ ⎡ ⎤
Si d xd y S S d xd y S S d xd y S S d xd y 21 1 2
⎢  S S d xd y  i 2j  i m  i n
⎢ i S j d xd y Si Sm d xd y S j Sn d xd y ⎥ ⎢
⎥ (lw) ⎢ 1 2 2 1⎥
⎥
[M] = ⎢  j   2  ⎥= ⎢ ⎥
⎣ Sm Si d xd y Sm S j d xd y S d xd y Sm Sn d xd y ⎦ 18 ⎣ 1 2 2 1⎦
   m  2
Sn Si d xd y Sn S j d xd y Sn Sm d xd y Sn d xd y 21 1 2

Substituting Eqs. (3.84–3.86) into Eq. (3.82) yields the element model as
-
 .
[K x ] + K y − λ[M] {φ} = 0 (3.86)

where [K x ], [K y ] and [M] are given in Eqs. (3.84–3.86).

A specific solution
-
λi , (or an eigenvalue)
. to Eq. (3.86) must ensure that the
determinate of [K x ] + K y − [M] is 0, and the corresponding φ i is an
eigenvector.
The similar procedure is applied to a triangle element to obtain the triangle element
model in the same format of Eq. (3.86) but with different [K x ], [K y ] and [M] as below
⎡  ∂ Sj  ⎤
∂ Si
· ∂∂Sxi d xd y ∂ Si
· d xd y ∂ Si
· ∂∂Sxk d xd y
⎢ ∂x
∂ Si ∂S  ∂x
∂ Sj
∂x
∂ Sj  ∂x
∂ Sj ⎥
[K x ] = ⎣ · ∂ xj d xd y · d xd y · ∂∂Sxk d xd y ⎦
 ∂x
∂ Sk
 ∂x ∂x
∂ Sj  ∂x
∂x
· ∂∂Sxi d xd y ∂ Sk
∂x
· ∂x
d xd y ∂ Sk
∂x
· ∂∂Sxk d xd y
⎡ ⎤
βi2 βi β j βi βk
1 ⎢ ⎥
= ⎣ β j βi β 2j β j βk ⎦ (3.87)
4A
βk βi βk β j βk2
⎡  ∂ Si ∂ Si  ∂S  ⎤
· d xd y ∂∂Syi · ∂ yj d xd y ∂∂Syi · ∂∂Syk d xd y
 ⎢  ∂∂Syi ∂∂Syj  ∂ Sj ∂ Sj  ∂ S j ∂ Sk ⎥
K y = ⎣ ∂ y · ∂ y d xd y ∂ y · ∂ y d xd y ∂ y · ∂ y d xd y ⎦
 ∂ Sk ∂ Si  ∂ Sk ∂ S j  ∂ Sk ∂ Sk
∂y
· ∂ y d xd y ∂ y · ∂ y d xd y ∂ x · ∂ y d xd y
⎡ ⎤
δi2 δi δ j δi δk
1 ⎢ ⎥
= ⎣ δ j δi δ 2j δ j δk ⎦ (3.88)
4A
δk δi δk δ j δk2
⎡  2   ⎤ ⎡ ⎤
 Si d xd y Si S2 j d xd y  Si Sk d xd y A⎣
211
[M] = ⎣ S j Si d xd y S d xd y S S d xd y ⎦ = 1 2 1⎦ (3.89)
  j  k2j 12
Sk Si d xd y Sk S j d xd y Sk d xd y 112
3.5 FEA Theory 163

3.5.7 System Model and Solution

3.5.7.1 Build System Model

In an FEA model, the number of independent variables to be determined is called

degrees of freedom (DoF). The size of a system model is determined by the DoF of
the model as

[K ] N ×N {U} N ×1 = {F} N ×1 (3.90)

where N is the degrees of freedom; {U}is the vector of N independent variables;

{F}is the vector of N loads over the corresponding independent variables; [K] is a
N × N relational matrix of {U} and {F}.
The system model is assembled from element models. Assume that the FEA model
has N e elements, and each element involves in NDi independent variables (nd i,1 nd i,2 ,
…, nd i, NDi ). The system model is assembled as

e
i=N
[K ] N ×N = [ Ai ]T [K i ][ Ai ] (3.91)
i=1

where [K i ] is the NDi × NDi relational matrix of the NDi –th element, [ Ai ]T is the
transpose of [ Ai ], and [ Ai ] is a NDi × N assistive matrix defined as

⎡ nd i,1 nd i,2 nd i,nd i ⎤

. . . . . . . . . . . . ... . . . . . .
⎢... 1 ... 0 ... 0 ...⎥
⎢ ⎥
[ Ai ] = ⎢
⎢ . . . . . . .. ... . . . . . . . . . ⎥
⎥ (3.92)
⎢ ⎥
⎢... 0 ... 1 ... 0 ...⎥
⎢ ⎥
⎣... ... ... ... ... ... ...⎦
... 0 ... 0 ... 1 ...

where all other entities in [ Ai ] are zeros.

Example 3.5 Figure 3.27a shows a FEA model for a 2D heat transfer problem.
The model is to find the temperature distribution, and the FEA model consists of
10 nodes for three triangular elements and three rectangle elements. Build a system
model based on the given element models as follows:
The stiffness matrices [K i ] for elements are found as

Triangle elements (1), (4) and (6) : Rectangle elements (2), (3) and (4) :
164 3 Computer-Aided Engineering (CAE)

Fig. 3.27 a Elements and nodes b Relations of nodes and elements

⎡ ⎤
⎡ ⎤ 4 −1 −2 −1
0.5 −0.5 0 ⎢ −1 4 −1 −2 ⎥
[K](e) = ⎣ −0.5 1 −0.5 ⎦ [K](e) =⎢
⎣ −2
⎥
−1 4 −1 ⎦
0 −0.5 0.5
−1 −2 −1 4

Solution. Each node in the FEA element has one DoF, and the model has 10 nodes.
Therefore, the system model has N = 10 DoF, and [K] for the system has a size of
10 × 10.
The system model is assembled from element models. The model as N e = 6
elements, the involved DoF in these elements are described in Fig. 3.27b.
To use Eq. (3.91) to assemble system model, [Ai ] (i = 1, 2, …, 6) are determined
based on Fig. 3.27b as
⎡ ⎤
1 000000000
⎣
[ A1 ] = 0 1 0 0 0 0 0 0 0 0⎦
0 000100000
⎡ ⎤
0 100000000
⎢0 0 1 0 0 0 0 0 0 0⎥
[ A2 ] = ⎢
⎣0
⎥
0 0 0 0 1 0 0 0 0⎦
0 000100000
⎡ ⎤
0 0 1 0 0 0
0 0 0 0
⎢0 0 0 1 0 0⎥
0 0 0 0
[ A3 ] = ⎢
⎣0
⎥
0 0 0 0 0⎦
0 1 0 0
0 0 0 0 0 1
0 0 0 0
⎡ ⎤
0 000100000
[ A4 ] = ⎣ 0 0 0 0 0 1 0 0 0 0⎦
0 000000100
3.5 FEA Theory 165

⎡ ⎤
0 0 0 0 0 1
0 0 0 0
⎢0 0 0 0 0 0⎥
0 1 0 0
[ A5 ] = ⎢
⎣0
⎥
0 0 0 0 0⎦
0 0 0 1
0 0 0 0 0 1
0 0 1 0
⎡ ⎤
0 000000100
[ A6 ] = ⎣ 0 0 0 0 0 0 0 0 1 0⎦
0 000000001

Substituting the above assistive matrices and given stiffness matrices in Eq. (3.91)
yield the stiffness matrix of the system as
⎡ ⎤
0.5 −0.5 0 0 0 0 0 0 0 0
⎢ −0.5 −1 0 −1.5 −2 0 0 ⎥
⎢ 0.5 0 0 ⎥
⎢ ⎥
⎢ 0 −1 8 −1 −2 −2 −2 0 0 0 ⎥
⎢ ⎥
⎢ 0 0 −1 4 0 −2 −1 0 0 0 ⎥
⎢ ⎥
 T   ⎢ 0
i=6
−1.5 −2 0 5 −1.5 0 0 0 0 ⎥
[K ]10×10 = Ai K i Ai = ⎢
⎢ 0
⎥
⎢ −2 −2 −2 −1.5 13 −2 −1.5 −2 0 ⎥
⎥
i=1 ⎢ 0 −2 −1 0 −2 8 −2 −1 0 ⎥
⎢ 0 ⎥
⎢ ⎥
⎢ 0 0 0 0 0 −1.5 −2 5 −2 0 ⎥
⎢ ⎥
⎣ 0 0 0 0 0 −2 −1 −1.5 5 −0.5 ⎦
0 0 0 0 0 0 0 0 −1 0.5
(3.93)

3.5.7.2 Solution to Steady Problems

When FEA is used to model a steady engineering problem, the system model becomes
a set of linear equations in Eq. (3.90), and the solution to the steady problem is found
as

{U} = [K ]−1 {F} (3.94)

where {F} is the vector of loads, {U} is the vector of state variables, and [K] is the
matrix for the relation of loads and state variables, such as the stiffness matrix in
static analysis. [K]−1 is an inverse matrix [K].
Note that Eq. (3.94) can be linear or nonlinear. A linear system model have all
constant entities in [K] and {F}; otherwise, the system model is nonlinear. To solve
a nonlinear system model, an iterative procedure is applied to solve {U}. At each
iteration, the system model is simplified as a linear system model for {U}. The
iterative procedure will be terminated until all of the constraints are satisfied.
Since an FEA model for a real-world engineering problem involves in thousands of
unknown variables, an effective algorithm to find the solution to Eq. (3.94) is critical.
166 3 Computer-Aided Engineering (CAE)

Solvers to a linear system model can be classified into direct or iterative methods.
A direct method, such as Gaussian elimination or LU decomposition, obtains the
solution in a definite number of steps, and the computation can be predicted based
on the size of [K]. The computation of an iterative method is hard to be estimated since
the number of iterations to find a converged solution depends on the characteristic
of [K]. Theoretically, a direct method can find an exact solution to a system of
linear equations. However, it suffers a number of significant disadvantages in its
implementation. For example,
(1) A direct method provides an open-loop solution. It implies that the solution
does not have the mechanism to compensate round-off errors in numerical
calculation; however, computation errors are accumulated rapidly in the solving
process.
(2) The implementation of a direct method is very intrigue if all of the special cases
are considered. For example, a special treatment is needed when the coefficient
at the pivoting position becomes zero in the Gaussian elimination.
(3) A direct method requires a large amount of computation when the number of
unknown variables increase. In addition, it lacks the flexibility to reduce the
computation for the system model with a sparse matrix.
An iterative method is an ideal alternative since it provides a closed-loop solu-
tion by iterations. The accuracy of the solution is evaluated at every iteration, and
the iterative process will not be terminated until the solution reaches the required
accuracy.
To solve the system model iteratively, rewrite [K]{U} = {F} as ([K L ] + [ K D ] +
[K U ]){U} = {F}where the original full [K] is decomposed into three sub-matrices,
i.e., lower, diagonal, and upper sub-matrices as
⎡ ⎤ ⎡ ⎤
0 0 ... 0 k1,1 0 ... 0
⎢ .. ⎥ ⎢ .. ⎥
⎢ k2,1 . ... ...⎥ ⎢ . ... ... ⎥
[K L ] = ⎢ ⎥, [K D ] = ⎢ 0 ⎥,
⎢ ⎥ ⎢ ⎥
⎣ . . . . ...
. . ...⎦ ⎣ . . . ... ..
. ... ⎦
k N ,1 . . . k N ,N −10 0 ... 0 k N ,N
⎡ ⎤ (3.95)
0 k1,2 . . . k1,N
⎢ ⎥
⎢ 0 ... . . . . . . ⎥
⎢
[K U ] = ⎢ ⎥
⎥
⎣ . . . . . . . . . k2,1 ⎦
0 ... 0 0

Accordingly, the system solution Eq. (3.94) can be written as

{U} = [K D ]−1 {{F} − ([K L ] + [K U ]){U}} (3.96)

where [K L ], [K D ], and [K U ] are lower, diagonal, and upper matrices shown in Eq.
(3.95).
3.5 FEA Theory 167

Equation (3.97) can further be written in an iterative form from the k-th step to
the k + 1-th step as
- .
{U}(k+1) = [K D ]−1 {F} − ([K L ] + [K U ]){U}(k) (3.97)

where {U}(k) and {U}(k+1) are the iterated solutions at the k-th step to the k + 1-th
steps, respectively.
The interactive process (i.e., Eq. (3.97)) will continue until the termination
condition is satisfied,

{U}(k+1) − {U}(k)  < ε (3.98)

where ε is the specified accuracy of system solution.

The above method is referred as the Jacobi’ method. Assumed that the diagonal
entries in [K] are nonzero, the scalar version of Eq. (3.99) is
 
/
N
Ui(k+1) = 1
kii
Fi − ki j U (k)
j , (i = 1, 2, . . . N ) (3.99)
j=1; j
=i

Note that an iterative method can guarantee the convergence of the iterative process
under the conditions of

/
N  
|kii | > ki j  f ori = 1, 2, . . . N (3.100)
j=1; j
=i

And
 
|K D |−1 (K L + K U ) = i=1,2,...N max ki j /kii  < 1 (3.101)

The conditions in Eqs. (3.100) and (3.101) ensure that the converged solution will
be guaranteed. Not all of FEA models can strictly meet these conditions, but they
can still be solved by an iterative method effectively.

Example 3.6 Use the iterative method to solve [K]{U} = {F} where.
⎡ ⎤ ⎧ ⎫
2 −1 0 0 0 ⎪
⎪ 5⎪
⎪
⎢ −1 2 −1 0 0 ⎥ ⎪
⎪ ⎪
⎪
⎢ ⎥ ⎪
⎨2⎪⎬
⎢ ⎥
[K ] = ⎢ 0 −1 2 −1 0 ⎥ and {F} = 2
⎢ ⎥ ⎪
⎪ ⎪
⎪
⎣ 0 0 −1 2 −1 ⎦ ⎪
⎪2⎪⎪
⎪
⎩ ⎪
0 0 0 −1 2 5⎭
Solution. Let {U}(0) = [1 1 1 1 1] and the terminate tolerance ε = 0.0001.

Using Eq. (3.99) to execute the iterative process, the process to obtain the final
solutions is.
168 3 Computer-Aided Engineering (CAE)

Step {xi}(k) Convergence

0 [1 1 1 1 1]
1 [3.0000 2.0000 2.0000 2.0000 3.0000] 3.5000
10 [6.5762 8.6270 9.1523 8.6270 6.5762] 0.7750
20 [7.6621 10.4369 11.3242 10.4369 7.6621] 0.1839
30 [7.9198 10.8664 11.8396 10.8664 7.9198] 0.0436
40 [7.9810 10.9683 11.9619 10.9683 7.9810] 0.0104
50 [7.9955 10.9925 11.9910 10.9925 7.9955] 0.0025
60 [7.9989 10.9982 11.9979 10.9982 7.9989] 5.8325e-04
70 [7.9997 10.9996 11.9995 10.9996 7.9997] 1.3841e-04
73 [7.9998 10.9997 11.9997 10.9997 7.9998] 8.9898e-05

3.5.7.3 Solution to Eigenvalue Problems

To understand the response of an engineering system to dynamic loads, eigenvalues

must be found so that the stead ability of system under the stimulations of different
frequencies can be evaluated. An eigenvalue problem aims to find the eigenvalues
and eigenvectors for the given system model.
Let a system model of an eigenvalue problem be

[K ]{u} − λ[M]{u} = 0 (3.102)

An eigenvalue λi should satisfy the condition for a valid solution of the system
model,

det([K ]{u} − λ[M]) = 0 (3.103)

where det(•) is the determinant of the matrix (•).

Equation (3.103) is commonly referred as the characteristic equation in an eigen-
value problem of Eq. (3.102). λi that satisfies Eq. (3.103) is called as an eigenvalue
(i = 1, 2, …N), and an eigenvector ui corresponding to λi satisfies the condition of

([K ] − λi [M]){u} = 0 (3.104)

If a system model with a few DoF, eigenvalues to Eq. (3.102) can be obtained by
solving the polynomial equation about λ, which is defined based on Eq. (3.103). For
a system model with a large number of DoF, the following numerical method is used
to find a set of the most significant eigenvalues.
Assume that the eigenvalue problem has N DoF, i.e., the matrix [K] or [M] has the
size of N × N, a numerical method aims to find M number of the lowest eigenvalues
and the corresponding eigenvectors, and these eigenvalues are found iteratively in
the following procedure:
(1) select M non-zero vectors for a N × M modal matrix [X i ], where each of its
columns corresponds to one vector; set the iteration index i = 1,
3.5 FEA Theory 169

(2) solve the system of linear equations [K]{Y i } = [M]{X i } to get N × M modal
matrix {Y i },
(3) build a new eigenvalue problem with a reduced size of M × M as

[K ]i∗ {u}∗ − λ[M]i∗ {u}∗ = 0

where [K ]i∗ = {Y i }T [K ]{Y i } and [M]i∗ = {Y i }T [M]{Y i }

(4) solve the new eigenvalue problem for M eigenvalues and eigenvalues as
 
(i) (i)
λ(i)
m = diag λ 1 , λ 2 , · · · λ (i)
m
∗   (i) (i) 
u = u1 , u2 , · · · um(i)

(5) update the iteration as i ← i + 1, [X i ] ← [X i+1 ] = normc ([Y i ][u* ], where

normc(•) is a function to make every vector (column) in a matrix as a unit
vector;
(6) check the terminating condition as
0     (i) 1
 λ(i) − λ(i−1)   λ(i) − λ(i−1)   λm − λ(i−1) 
 1   2 
max  1
,  2
, · · ·  m  ≤ε

 λ (i)   λ (i)  (i)
λm
1 2

If it is satisfied, the iteration is terminated. Otherwise, return to step 2) for the

continuation.

Example 3.7 Find the first two natural frequencies of the following system model
with the given accuracy ε of 0.001:

[M]{ü} + [K ]{u} = 0 (3.105)

where
⎡ ⎤ ⎡ ⎤
20 5 0 2 −1 0
[M] = ⎣ 5 10 5 ⎦, [K] = ⎣ −1 2 −1 ⎦
0 5 10 0 −1 1

Solution. Equation (3.105) can be turned into an eigenvalue problem with three DoF
(N = 3) as.

[K ]{u} − λi [M]{u} = 0 where λi = ωi2 (3.106)

Since the first two natural frequencies (M = 2) of the system model is interested,
let assume that an initial vector as {u}2×3 , i.e.,
170 3 Computer-Aided Engineering (CAE)
⎡ ⎤
 10
[u]23 = u1 u2 = ⎣0 1⎦
00

Let i = 1 and [X 1 ] = [u]23 to initialize the iterative process.

Iteration 1. Solving the equation set [K][Y i ] = [M][X i ] yields
⎡ ⎤
25 20
[Y i ] = ⎣ 30 35 ⎦
30 40

The reduced eigenvalue problem at this iteration becomes

[K ]∗ [u]∗ − λ[M]∗ [u]∗ = 0

where
 
∗ 650 575
[K ] = [Y i ] [K ][Y i ] =
T
575 650
 
4.7 5.1125
[M]∗ = [Y i ]T [M][Y i ] = 1.0 × 104
5.1125 5.725

The eigenvalues and eigenvectors ofthe above reducedproblem can be found as

0.7455 0.0450
λ1 = 0.1052, λ2 = 0.0113, and [u]∗ = .
−0.6665 0.9990
Next, [X i ] can be updated as
⎛⎡ ⎤⎞ ⎡ ⎤

 5.3075 21.1040 0.7697 0.3582
⎜⎢ ⎥⎟ ⎢ ⎥
[X i ] = nor mc [Y i ][u]∗ = nor mc⎝⎣ −0.9625 36.3137 ⎦⎠ = ⎣ −0.1396 0.6164 ⎦
−4.295 41.3087 −0.6229 0.7012

Iteration 2. Solving the equation set [K][Y i ] = [M][X i ] with the updated [X i ]
yields
⎡ ⎤
7.1085 31.8025
[Yi ] = ⎣ −0.4800 53.3581 ⎦
−7.4069 63.4522

The reduced eigenvalue problem at this iteration becomes

[K]∗ [u]∗ − λ[M]∗ [u]∗ = 0

3.5 FEA Theory 171

where
 
650 575
[K ]∗ = [Y i ]T [K ][Y i ] =
575 650
 
∗ 4.7 5.1125
[M] = [Y i ] [M][Y i ] = 1.0 × 10
T 4
5.1125 5.725

The eigenvalues and eigenvectors ofthe above reducedproblem can be found as

1.000 −0.0069
λ1 = 0.1052, λ2 = 0.0113, and [u]∗ = .
0.0054 1.0000
Next, [X i ] can be updated as
⎛⎡ ⎤⎞ ⎡ ⎤

 7.2799 31.7526 0.7175 0.3575
⎜⎢ ⎥⎟ ⎢ ⎥
[X i ] = nor mc [Y i ][u]∗ = nor mc⎝⎣ −0.1923 53.3601 ⎦⎠ = ⎣ −0.0190 0.6008 ⎦
−7.0647 63.5019 −0.6963 0.7150
 i+1 i+1   i+1 i+1 
 λ −λ   λ −λ 
Check the convergence: max  1 λi+1 1 ,  1 λi+1 1  = 0.05 > ε; therefore, the
1 1
iteration has to be continued.
Iteration 3. Solving the equation set [K][Y i ] = [M][X i ] with the updated [X i ]
yields
⎡ ⎤
7.1139 31.6789
[Y i ] = ⎣ −0.0274 53.2034 ⎦
−7.0852 63.3574

The reduced eigenvalue problem at this iteration becomes

[K ]∗ [u]∗ − λ[M]∗ [u]∗ = 0

where
 
∗ 0.1514 −0.0000
[K ] = [Y i ] [K ][Y i ] = 1.0 × 10
T 3
−0.0000 1.5700
 
0.0151 −0.0000
[M]∗ = [Y i ]T [M][Y i ] = 1.0 × 105
−0.0000 1.3908

The eigenvalues and eigenvectors ofthe above reduced

 problem can be found as
1.000 −0.000
λ1 = 0.1001, λ2 = 0.0113, and [u]∗ = .
0.000 1.0000
Next, [X i ] can be updated as
172 3 Computer-Aided Engineering (CAE)
⎛⎡ ⎤⎞ ⎡ ⎤

 7.1143 31.6788 0.7086 0.3576
⎜⎢ ⎥⎟ ⎢ ⎥
[X i ] = nor mc [Y i ][u]∗ = nor mc⎝⎣ −0.0267 53.2034 ⎦⎠ = ⎣ −0.0027 0.6006 ⎦
−7.0844 63.3575 −0.7056 0.7152
 i+1 i   i+1 i 
 λ −λ   λ −λ 
Check the convergence: max  1λi+1 1 ,  2λi+1 2  < ε. The convergence condi-
1 2
tion is satisfied.
Finally, the first two natural frequencies are found as ω1 = 0.3164 and ω2 =
0.1063.

3.5.7.4 Solution to Transient Problems

In a transient problem, the dynamic response of a system in a transient period is of

interest. Other than the spatial domain, the time domain has to be discretized. The
solution to a transient problem can be treated as a set of sub-solutions in a series
of time steps from the initial time to the terminate time. The solving procedure is
step by step where the sub-solution in the next step is updated based on those in
precedent steps. While some special consideration have to be taken in determining
time steps and updating system models, the sub-solution at each step is similar to an
equilibrium problem. In this section, a transient heat transfer problem is used as an
example to illustrate the solving procedure for a transient problem.
In a transient heat transfer problem, the temperature distribution in a body varies
over time; the transient response is critical to some manufacturing processes such
as heat treatments. To find the solution to a transient heat transfer problem, the time
domain is discretized into steps, while the time step is critical to both of the stability
of the iterative process and accuracy of sub-solutions: if the time step is too small, a
spurious oscillation may occur which leads to meaningless results. If the time step
is too large, the solutions may not be converged since the temperature gradients are
evaluated appropriately.
To discretize the time domain, Biot number (Bi = h
x/k) and Fourier number
(F0 =
t/(
x)2 ) are used to determine a reasonable time step, where h and k are
convective and conductive coefficients, respectively; α is the thermal diffusivity;
x
denotes the mean length of an element in one spatial direction and
t is the time
step.
When the Biot number Bi is less than 1 (Bi < 1), the time step should be determined
to meet the condition of convergence of 0.1 ≤ F0 ≤ 0.5; therefore,

(
x)2 F0

t = (3.107)
α
When the Biot number Bi is larger than 1 (Bi > 1), the time step should be
determined by considering both of the Biot and Fourier numbers as (Fo)(Bi) = b,

(
x)ksolid (
x)ρc

t = b= b (3.108)
αh h
3.5 FEA Theory 173

where b = F 0 ·Bi ; ρ and c are the density and specific heat of materials.
Assume that a transient heat transfer problem is formulated as the following
differential equations (Bi 2018),
" 
∂T
[C] · + [K] · {T} = {Q} (3.109)
∂t

where [C] and [K] are the heat storage and heat transfer matrices, respectively; {Q}
is the vector of the heat loads; T} and { ∂T
∂t
} are the vectors of temperatures and
temperature gradients, respectively.
Discretizing the time domain by setting the time step of
t gets the iteration
equation at step i as

- .(i)
t 2 - .(i)
{T }(i+1) {T }(i) +
t Ṫ + T̈ (3.110)
2
The flexibility of Eq. (3.110) can be improved by adjusting the weight in the front
- .(i)
of T̈ by the Euler parameter θ ∈ (0, 1) as
- .(i) - .(i)
{T }(i+1) {T }(i) +
t Ṫ + θ
t 2 T̈ (3.111)

A different value of the Euler parameter corresponds to a different scheme of the

iteration, i.e.,
⎧
⎪
⎨ 0 f or war d di f f er ence(Euler )
θ = 1/2 Crank − N icolson di f f er ence (3.112)
⎪
⎩ 1 Backwar d di f f er ence

In Eq. (3.111), the second-order of temperature can be approximated by the first-

order derivative at two time steps as
- .(i+1) - .(i)
- .(i) Ṫ − Ṫ
T̈ = (3.113)
t
Substituting Eq. (3.113) into (3.111) yields
- .(i) - .(i+1)
{T }(i+1) {T }(i) + (1 − θ )
t Ṫ + θ
t Ṫ (3.114)
- .(i+1)
Equation (3.114) shows that Ṫ depends on {T }(i+1) as

(i+1)  # $
- .(i+1) {T } − {T }(i) 1 - .(i)
Ṫ = − − 1 Ṫ (3.115)
θ
t θ
174 3 Computer-Aided Engineering (CAE)

- .(i)
Since both of {T }(i) and Ṫ are known at the previous time step i; using
Eq. (3.115) in Eq. (3.114) gets the explicit equation of iteration as
# $ # $
[C] (i+1) [C] (i) 1 - .(i)
+ [K ] {T } = { Q} + {T } + − 1 [C] Ṫ (3.116)
θ
t θ
t θ

[C]
where [K m ] = θ
t + [K ] is the modified heat transfer matrix, and.
- . [C]

 - .(i)
Q m = { Q} + θ
t {T }(i) + θ1 − 1 [C] Ṫ is the modified vector of heat
loads.

Example 3.8 A 2-in thick plate of 1060 alloy is uniformly heated to 300 °C. It is
air-quenched in the forced air. Assume that the temperature of the plate surface is
immediately set to 25 °C. The materials density ρ = 0.0975437 lb/in3 , the conduc-
tivity k = 0.002675 Btu/(s·in·°C), and the specific heat c = 0.214961 Btu/(lb·°C).
Determine the cooling curve of temperature at the neutral plane of the plate with
respect to time in 5 s.
Solution. Figure 3.28 shows a 1D transient heat transfer element described in the
local coordinate system (LCS). It consists of node i and node j. The state variables
at two nodes are the temperatures (Ti , T j ) and temperature gradient (Ṫi , Ṫ j ). The
elemental behavior is governed by

∂2T ∂T
k −c + q̇ = 0 (3.117)
∂x 2 ∂t

where k is conductivity, ρ is the density, and c is the specific heat.

The element model is developed based on Eq. (3.117) as follows. Let the shape
functions of an element be S i (x) = 1−x/L and S j (x) = x/L, using the shape functions as
the test function in the Galerkin method leads to the weak-form solution Eq. (3.117)
as (Bi 2018),
" ( p+1)
T - .
[K m ] i = Qm (3.118)
Tj

Fig. 3.28 1D transient heat

transfer element
3.5 FEA Theory 175

where p is the step index and [K M ] is the modified conductive matrix as

#    $
ρc AL 2 1 k A 1 −1
[KM ] = + (3.119)
6θ
t 1 2 L −1 1

[QM ] is the modified load vector determined by the state of the precedent step
(i.e., step i) as
"   " ( p) # $  " ( p)
- . Qi ρc AL 2 1 Ti 1 ρc AL 2 1 Ṫi
Qm = + + −1
Qj 6θ
t 1 2 T j θ 6 1 2 Ṫ j
(3.120)

After the temperature at (p + 1) have been found, the temperature gradients can
be derived by .
" ( p+1)  # $" ( p)
( p+1) ( p)
Ṫi 1 Ti − Ti 1 Ṫi
= ( p+1) − −1 (3.121)
Ṫ j θ
t Tj − T j(i) θ Ṫ j

Since the plate is symmetric about its neutral axis, a 1D FEA model is developed
for half of the plate thickness. Accordingly, Fig. 3.29 shows a decomposition which
leads to an FEA model with 5 elements and 6 nodes. Every element has the length
of 0.2-in.
Let the Fourier number be F 0 = 0.5, and the step time Δt is found from Eq. (3.107)
as

(
x)2 F0 (0.2)2 (0.5)

t = = = 0.1568(s)
α (0.002675/(0.0975437*0.214961))

If the history of the temperature change in 5 s is concerned, the total number N

of the time steps is

(5)
N= = 32(steps)

t

Fig. 3.29 Nodes and (1) (2) (3) (4) (5)

elements in 1D transient heat
transfer problem 1 2 3 4 5 6
176 3 Computer-Aided Engineering (CAE)

Let the Euler parameter θ = 0.5 in Eq. (3.121) where the Crank-Nicolson differ-
ence is used in the approximation. At each time step p from 1 to N, the system model
can be assembled from 5 element models expressed in Eq. (3.118) as
⎧ ⎫
⎧ ⎫( p+1) ⎪ ( p) ( p) ( p) ( p) ⎪
⎪ T ⎪ ⎪
⎪ 2c1 T + c1 T + 2c2 Ṫ + c2 Ṫ ⎪
⎪
⎪
⎪ ⎪
1
⎪ ⎪
⎪
1 2 1 2
( p) ⎪
c1 T1 + 4c1 T2 + c2 Ṫ1 + 4c2 Ṫ2 + c1 T3 + c2 Ṫ3 ⎪
⎪ ⎪ ⎪ ( p) ( p) ( p) ( p) ( p) ⎪
⎪
⎪ T2 ⎪
⎪ ⎪
⎪ ⎪
⎪
⎪
⎨ ⎪ ⎬ ⎨ c T ( p) + 4c T ( p) + c Ṫ ( p) + 4c Ṫ ( p) + c T ( p) + c Ṫ ( p) ⎪
⎪ ⎬
(G) T3 ( p) 1 2 1 3 2 2 2 3 1 4 2 4
[K ] = { Q} = ( p) ( p) ( p) ( p) ( p) ( p) ⎪
⎪
⎪ 4⎪
T ⎪
⎪ ⎪
⎪ c1 T3 + 4c1 T4 + c2 Ṫ3 + 4c2 Ṫ4 + c1 T5 + c2 Ṫ5 ⎪
⎪
⎪ ⎪ ⎪
⎪ ⎪
⎪
⎪ ⎪ ⎪ ( p) ( p) ( p) ( p) ( p) ( p) ⎪
⎪
⎪ T5 ⎪
⎩ ⎪
⎭
⎪
⎪
⎪ c1 T4 + 4c1 T5 + c2 Ṫ4 + 4c2 Ṫ5 + c1 T6 + c2 Ṫ6 ⎪ ⎪
⎪
T6 ⎪
⎩ ( p) ( p) ( p) ( p) ⎪
⎭
c T 1 5 + 2c T 1 6 + c Ṫ
2 5 + 2c Ṫ
2 6
(3.122)

where
⎡ ⎤
0.0312 −0.0045 0 0 0 0
⎢ −0.0045 0.0624 −0.0045 ⎥
⎢ 0 0 0 ⎥
⎢ ⎥
⎢ 0 −0.0045 0.0624 −0.0045 0 0 ⎥
[K ](G) =⎢ ⎥ (3.123)
⎢ 0 0 −0.0045 0.0624 −0.0045 0 ⎥
⎢ ⎥
⎣ 0 0 0 −0.0045 0.0624 −0.0045 ⎦
0 0 0 0 −0.0045 0.0312

c1 and c2 are two constants as below


c1 = ρc AL
= 0.0089

6θ
t 1  ρc AL (3.124)
c2 = 1 − θ 6 = −0.00069893

At each step p, after {T}(p+1) is found from Eq. (3.118), the temperature gradients
on nodes can be updated based on Eq. (3.121) as
⎧ 
 ( p) ⎫
⎪
⎪ θ
t
1
T1
( p+1)
− T1
( p)
− θ1 − 1 Ṫ1 ⎪ ⎪
⎧ ⎫( p+1) ⎪⎪ ⎪
⎪ Ṫ1 ⎪ ⎪
⎪

 ( p) ⎪ ⎪
⎪
⎪
⎪ ⎪
⎪ ⎪
⎪ 1
T2
( p+1)
− T2
( p)
− θ − 1 Ṫ2 ⎪
1
⎪
⎪
⎪ ⎪
⎪ ⎪
⎪ θ
t 
⎪
⎪
⎪
⎪ Ṫ2 ⎪
⎪ ⎪
⎪  ⎪
⎪
⎨ ⎬ ⎨ 1 T ( p+1)
− T3
( p)
− θ − 1 Ṫ3 ⎬
1 ( p)
Ṫ3 θ
t 3 
=
 ( p) (3.125)
⎪
⎪ Ṫ4 ⎪
⎪ ⎪
⎪ 1
T4
( p+1)
− T4
( p)
− θ1 − 1 Ṫ4 ⎪ ⎪
⎪
⎪ ⎪
⎪ ⎪
⎪ θ
t 
⎪
⎪
⎪
⎪ Ṫ ⎪
⎪ ⎪
⎪  ⎪
⎪
⎪
⎩ ⎭
5 ⎪ ⎪
⎪ 1
T5
( p+1)
− T5
( p)
− θ − 1 Ṫ5 ⎪
1 ( p)
⎪
Ṫ6 ⎪
⎪ θ
t  ⎪
⎪
⎪
⎪
 ⎪
⎩ 1 25 − T ( p+1) − 1 − 1 Ṫ ( p) ⎪ ⎭
θ
t 6 θ 6

Note that the calculation for the temperature gradient of the boundary node 6 is
different since the temperature at this node is fixed as 25 °C. Accordingly, when the
calculation result at step p is moved to next step p + 1, the temperature at node 6
3.5 FEA Theory 177

must be set back to 25 °C as the boundary constraint. The temperature changes on

nodes with respect to time are obtained in Table. 3.9.
The plots of temperature on node 1 (i.e., the center of plate) are illustrated in
Figs. 3.30 and 3.31 respectively.

Table 3.9 The History of nodal temperatures in 5 s

Fig. 3.30 Temperature at

Center of Plate with respect
to Time
178 3 Computer-Aided Engineering (CAE)

Fig. 3.31 Temperature

Gradient at Center of Plate
with respect to Time

3.6 CAE Implementation—SolidWorks Simulation

Numerous commercial and open-source computer-aided engineering packages have

been developed based on the aforementioned FEA theory. In the following section,
the Solidworks Simulation is used as an example to illustrate the capabilities of
commercial CAE tools for pre-processing, processing, and post-processing of FEA,
which are discussed in Sect. 3.4. Figure 3.32 shows the Graphic User Interfaces and
main functional modules for FEA modeling.

3.6.1 Computational Domain

FEA aims to investigate the system response in a continuous domain subjected to

boundary conditions. A continuous domain can be in one-dimensional (1-D), two-
dimensional (2-D), or three-dimensional (3-D). FEA begins with the construction of
solid models as a computational domain. A solid model with the geometric infor-
mation of vertices, edges, and boundary surfaces is sufficient to FEA; however, it
is desirable to use an advanced solid model that includes parametrized sketches,
dimensions, and features. A parametric model simplifies the changes and supports
the system integration. Modern CAE systems are integrated with their solid modeling
tools to create native solid models; in addition, CAE tools can use solid models from
other computer-aided platforms directly. For example, SolidWorks Simulation is
integrated in the solid modeling environment, and allow to import and export solid
models in over 30 formats of solid models as shown in Fig. 3.33.

3.6.2 Materials Library

In an FEA model, material properties are defined in a computational domain. Material

properties are classified into five groups as (1) physical properties such as density
3.6 CAE Implementation—SolidWorks Simulation 179

Mesh Tools CAD Modeling Tools Materials Tools

Isolate object and

applicaon from
environment

PROPROCESSING
Create CAD Assemble parts Assign material Parametric Study
Analysis Models model into as assembly properties

PROCESSING
Discretize solid
object(s) into nodes and
elements

Simulation based design optimization

Validation of Design and Analysis

Develop element models

and assemble them into a Solvers
system model

Define boundary conditions and loads,

and modify system models and load
vectors based on boundary conditions

Solve system Results Tools

Loading Tools
model

POST-PROCESSING
Evaluate dependent variables
and scales, visualize or
analyze results

Verification of
solutions

Restraints Tools Contact Tools

Scope for Scope for
validaon verificaon

Fig. 3.32 Solidworks Simulation for computer-aided engineering (CAE)

and melting temperature, (2) mechanical properties such as the modulus of elasticity,
Poisson’s ratio, yield strength, and hardness, (3) thermal properties such as thermal
conductivity, specific heat, and the coefficient of thermal expansion, (4) electric
properties such as resistivity and di-electricity, and (5) acoustic properties such as
compression wave velocity, and shear wave velocity, and bar velocity. Different
analysis types require the information of different material properties. For example,
the modulus of elasticity, the Poison’s ratio, yield strength, and the density must be
180 3 Computer-Aided Engineering (CAE)

Fig. 3.33 Compatible file

formats in Solidworks

defined in a static and modal analysis. The specific heat and the thermal conductivity
have to be defined in heat transferring analysis.
Engineering systems mostly use conventional materials such as metals and alloys,
plastics, and construction materials. The properties of those materials are well studied
and documented. A CAE tool usually provides a material library for conventional
industrial materials; it also provides the interface and template for users to define
custom materials for solids.
In the SolidWorks Simulation, the material library has the levels of Library, Cate-
gory, and Materials, and custom materials are defined in a Custom Material Library
as shown in Fig. 3.34. To create a new material, the designer begins with the construc-
tion of a custom material library, follows with the creation of a new material category
in this custom library, and finally, defines a new material in the new custom category.

3.6.3 Meshing Tools

The continuous domain in an original engineering problem has infinite degrees of

freedom (DoF). The continuous domain is discretized as a discrete FEA model
3.6 CAE Implementation—SolidWorks Simulation 181

Material Library in
SolidWorks

Library Build-in Solidworks Imported Custom …….

Material Library Material Library Material Library
Level

Category Steel Iron Aluminum Copper Plastics …….

Level Alloys Alloys

Material 1023 Carbon AISI AISO 4340 AISI Type

1020 …….
Level Steel Sheet Steel, Annealed A2 Tool Steel

Fig. 3.34 Structure of material library in Solidworks Simulation

with a finite number of nodes and elements by meshing. The SolidWorks Simu-
lation supports automatic meshing that is capable of (1) estimating the element sizes
based on the volume, surface area, and other geometric attributes of a computational
domain, and (2) generating elements and nodes based on element sizes and mesh
controls. The scale of a system model increases with an increase of elements and
nodes, and the system solution can be generally improved when the numbers of
elements and nodes are increased. To reduce the approximation errors at some areas
of interest, local mesh controls can be applied to refine element sizes on the selected
features such as edges, faces, and solids.
An CAE tool supports many element types for different analyses. Figure 3.35
shows three element types with different degrees of freedom. For a bulky object,
3D solid elements are suitable. For thin objects, 2D shell elements can be used to
simplify the FEA model. For sparse structures built from trusses and beams, 1D truss
or beam elements are appropriate.
When an assembly model is analyzed, manual intervenes are often required to
(1) ensure there is no interference of solids, and (2) define mesh controls to achieve

(a). 1-D elements (b). 2-D elements (c). 3-D elements

Fig. 3.35 1D, 2D and 3D elements

182 3 Computer-Aided Engineering (CAE)

Identified
interferences

Fig. 3.36 Detecting and removing interferences in an assembly model

compatible meshes when they are needed. Interface detection in the SolidWorks can
be used to identify and eliminate an interference of two solids. Figure 3.36 shows the
interface of the interface detection tool to detect the interferences in an object group.
When an interference is detected, the assembly model must be revised to eliminate
the mutual intrusions of solids.
As assembly model involves in some boundary surfaces where the parts make
contacts with each other. If no relative motion is allowed on a boundary surface, a
bonded contact can be defined for this boundary surface. A bonded contact means that
the nodes on the contact surface of the respective parts have the same displacements
in system modeling. The mesh at a contact surface can be either a compatible mesh
or an incompatible mesh as shown in Fig. 3.37. In a compatible mesh, the nodes at
the contacts are shared by two solids; in an incompatible mesh, the nodes on two
solids do not have one-to-one correspondences. Since nodes in an incompatible mesh

Object A Object A

Interface Interface

Object B Object B

(a) compatible mesh (b) incompatible mesh

(node-node at interface) (non node-node at interface)

Fig. 3.37 Nodes in compatible and incompatible meshes

3.6 CAE Implementation—SolidWorks Simulation 183

are restrained by displacement equations, errors in numerical calculation may lead

to an artificial issue of the stress concentration at contacts. Therefore, a compatible
mesh usually achieves better accuracy than an incompatible mesh. For an assembly
model, designers should refine the meshing parameters to obtain a compatible mesh
whenever it is feasible.
For a complex model, it is common that an automatic meshing process is not
successful on the first try. When the meshing process fails, the Failure Diagnostics
can be utilized to diagnose the problems, and an iterative meshing process can be
used to modify element sizes, apply mesh controls on the features of interest, and
change the settings at boundary surfaces until the mesh in the whole computational
domain is successfully generated.
To improve the solution, the automatic mesh refinement tool can be used. It
supports the mesh refinement in two ways, i.e., h-adaptive meshing and p-adaptive
meshing. H-adaptive meshing is to refine the mesh by increasing the number of
elements and reducing element sizes, and p-adaptive meshing is to use the high-
order elements without increasing the number of elements. Both of h-adaptive
and p-adaptive methods can be performed to meet the expected meshing accuracy
automatically.

3.6.4 Analysis Types

The solution to an FEA model is developed for a given mathematic model; therefore,
two FEA models would be different if their governing equations are different. By
defining the analysis type of an FEA model, the SolidWorks Simulation provides the
solutions to a variety of engineering problems. The analysis types supported by the
Solidworks Simulation are listed in Table 3.10.

3.6.5 Boundary Conditions

To find a particular solution to differential equations, boundary conditions (BCs) must

be given in an FEA model. Taking an example of static analysis, BCs are classified
into fixtures and loads. The PropertyManager in the SolidWorks Simulation provides
an interface to define BCs on vertices, edges, faces, features, and bodies. Restrained
DoF can be zero or a non-zero fixed displacement. Note that BCs are essential to
some analysis types such as static analysis, modal analysis, and dynamic analysis.
Table 3.11 provides a list of common restraints in the SolidWorks Simulation.
Table 3.12 gives the options to define a load for a static analysis in the SolidWorks
Simulation.
184 3 Computer-Aided Engineering (CAE)

Table 3.10 Analysis types supported by SolidWorks Simulation

3.6.6 Solvers to FEA Models

In numerical simulation, an engineering problem is formulated as an FEA model,

and the solution to an FEA model is the state variables that satisfy the mathematic
equations and constraints in the model. As shown in Fig. 3.38, mathematic equations
can be classified from the perspectives of probability, linearity, time-dependence,
explicitness, and continuity.
• Deterministic versus probabilistic: the design variables in a deterministic model
are uniquely determined without uncertainty. The deterministic model leads
to certain results of design variables when the initial conditions are given. A
3.6 CAE Implementation—SolidWorks Simulation 185

Table 3.11 BCs for static analysis

Table 3.12 Load types in static analysis

probabilistic model involves in the uncertainties where the design variables are
distributed around their mean values with certain probability.
• Linear versus nonlinear: in a linear model, all of the relations among design
variables are linear; otherwise, it is a nonlinear model. The common approach
to solve a nonlinear model is the linearization, in which a nonlinear model is
linearized into a set of linear models.
186 3 Computer-Aided Engineering (CAE)

Stochastic Models
Probability
Deterministic Models

Linearity Linear Models

Nonlinear Models

Mathematic Static Models Applicable Scope

Time
Models dependency Dynamic Models of Finite Element
Analysis Modeling
Explicitness Explicit Models
Implicit Models

Continuity Discrete Models

Continuous Models

Fig. 3.38 Types of mathematic models

• Static versus dynamic: a static model is steady in the sense that the corresponding
engineering system is in equilibrium state, and the design variables are time-
invariants, and a dynamic model treats the time as another dimension of design
problems; it takes into account of time-dependent changes of design variables.
• Explicit versus implicit: an explicit model can be solved in a series of the steps
explicitly while an implicit model have to be solved iteratively.
• Discrete versus continuous: the computational domain in a discrete model
consists of discrete elements (for example, truss elements in a truss structure)
and that of a continuous model corresponding to a continuous domain.
As shown in Fig. 3.38, FEA modelling can be applied to solve all types of the
above mathematical models except a probabilistic model.
The solutions to three types of differential equations have been discussed in
Sect. 3.5.6 in detail, and the corresponding engineering problems are called as equi-
librium problems, eigenvalue problems, and propagation problems, respectively. An
equilibrium problem concerns the system states subjected to static, quasi-static, or
repetitive loads. An eigenvalue problem is an extension of an equilibrium problem
whose solutions are characterized by identifying a set of unique system configura-
tions such as resonances and bulking. A propagation problem concerns the changes
of state variables over time. The Solidworks Simulation provides four generic solvers
to the above three types of engineering problems: Auto, FFEPlus, Direct Sparse, and
Large Problem Direct Sparse.
Auto solver authorizes the software to select an appropriate solver based on the
statistics of an FEA model automatically.
3.6 CAE Implementation—SolidWorks Simulation 187

Fast Finite Elements (FFEPlus) uses an implicit method to find the solution iter-
atively. It is very effective when an FEA model involves in large DoF (typically,
a model over 100, 000 DoF). However, FFEPLus may be ineffective when (1) the
mesh is incompatible and a local bonded contact is not covered by a global bonded
contact; (2) a modal analysis which involves in external forces or gravity; (3) a linear
dynamic study which involves in base excitation; (4) a computational domain which
involves in varying moduli of elasticity; (5) a nonlinear analysis.
Direct Sparse uses an analytic method to solve a system model. ‘Sparse’ refers
to the sparsity (zeroes) of a matrix; it shows the relations of design variables and
loads. Direct Sparse is applicable to an FEA model with a small number of DoF and
nonlinear analysis with better accuracy. However, it requires large memory; therefore,
the limits of DoFs for a linear analysis, nonlinear analysis, and heat transfer problem
are 100,000, 50,000, and 500,000 DoF, respectively.
Large Problem Direct Sparse (LPDS) is an enhanced version of the Direct Sparse
solver fora large FEA models. Intensive computation should be handled by parallel
computation. A LPDS solver is used when the Direct Sparse solver is disabled due
to the limit of random access memory (RAM). LPDS should be the last resolution
to an FEA model.

3.6.7 Post-processing

Numerical simulation to an FEA model usually generates a large amount of data; the
post-processing tool helps users to classify, visualize, and export simulation data.
The post-processing in the Solidwork Simulation allows users to (1) visualize the
distributions and contours of state variables, (2) animate system responses from the
perspectives of deformations, vibrational modes, and contact behaviours and others,
(3) view trajectories, traces, paths in transit or dynamic model, (4) define sectional
views to look into state variables internally, and (5) use a probe tool to examine a state
variable or any dependent variable at selected points, edges, faces, or components.

3.6.8 Design Optimization

FEA modeling supports the simulation-based optimization. Any system optimization

requires analyzing and comparing different system solutions, and FEA modeling is
an ideal and generic tool to analyze system solutions in system optimization (Bi and
Zhang 2001). Many CAE tools, such as the SolidWorks Simulation, provide Design
Study for simulation-based optimization. A Design Study is defined to (1) discretize
the design variables of interest, (2) define a set of design options by combining a set
of design variables, (3) specify the optimization criteria based on the simulation data,
(4) run the simulations on all of design options, and (5) compare the performances
of design options and finalize the design option with the best performance.
188 3 Computer-Aided Engineering (CAE)

3.7 CAE Applications

This section discusses the applications of CAE in different types of engineering

problems.

3.7.1 Structural Analysis

A mechanical system responds to different types of loads such as forces, pressures,

and heats. In engineering design, a structural analysis serves for multiple purposes
such as (1) modeling and analyzing the system response to given loads, (2) evaluating
stress distributions to identify critical areas subjected to the given loading conditions,
and (3) determining if an engineering design is safe by comparing the maximum stress
in the computational domain and the strength of material.
Figure 3.39 describes an engineering problem in structural analysis. The compu-
tational domain is the solid with a finite volume. The loads applied in the solid are
classified as (1) volume force such as the weight by gravity, (2) surface force such as
a drag force by surrounding airflow, and (3) concentrated load such as a point load
on body. A structural analysis is for an equilibrium problem where the solution at
any position is affected by those of all other positions in solids. At a position with
an infinitesimal volume, the state variable is the stress, which is described as
T
σ = σx σ y σz τx y τx z τ yz (3.126)

Surface
loads
w

Node i fz
Concentrated Volume
loads loads
Node j fy
fx
v

u dV=dx·dy·dz
Z Fixed Free
Boundary Boundary
O
Y
X

Fig. 3.39 Engineering problem in structural analysis

3.7 CAE Applications 189

Fig. 3.40 Free body diagram (FBD) of an infinitesimal volume

where σ x , σ y , and σ z are normal stresses along X, Y, and Z; τ xy , τ xz , τ yz are the shear
stresses over YZ, XZ, and XY planes, respectively.
Figure 3.40 shows the free-body diagram (FBD) of an infinitesimal volume (dx
× dy × dz). The force-balancing equation for the solid can be defined in any of X-,
Y-, and Z-axis.
Taking an example of the force balancing equation along Y-axis, we have
/   ⎫
Fy = σ y + ∂ yy dy d xdz − σ y d xdz + τx y + ∂ xx y d x dydz − τx y d xdz ⎬
∂σ ∂τ

∂τ
+ τzy + ∂zzy dz d xd y − τzy d xd y − f y d xd ydz = 0 ⎭
(3.127)

Equation (3.127) can be simplified as

∂τx y ∂σ y ∂τ yz
+ + + fy = 0 (3.128)
∂x ∂y ∂z

Repeating the process for the force balancing equations along X- and Z-axes gets
∂τ ⎫
∂σx
∂x
+ ∂ yx y + ∂τ∂zx z + f x = 0 ⎪
⎬
∂τx y ∂σ ∂τ
∂x
+ ∂ yy + ∂zyz + f y = 0 (3.129)
∂τx z ∂τ ⎪
⎭
∂z
+ ∂ yyz + ∂σ∂z
z
+ fz = 0
190 3 Computer-Aided Engineering (CAE)

To model the system response in terms of the deformation of the solid, the strain
T
at the infinitesimal volume is described as ε = εx ε y εz γx y γzy γx z and the
constitutive relations to the stress state is defined as


 ⎫
εx = 21
σx − v
σ y + σz  ⎪ ⎪
⎬
ε y = 21
σ y − v
σ y + σz 
(3.130)
εz = 21 σz − v σx + σ y ⎪
⎪
τ τ ⎭
γx y = Gx y , γ yz = Gyz , γx z = τGx z

where E is the Young’s modulus, ν is the Poisson’s ratio, and G is the shear modulus,
which relates to the Young’s modulus as

E
G= (3.131)
2(1 + v)

The stress and strain relation in Eq. (3.130) can be converted to calculate stress
state based on given strain state as

{σ } = [ D]{ε} (3.132)

where [D] is the matrix for the stress and strain relation in the 3D space,
⎡ ⎤
1−ν ν ν 0 0 0
⎢ ν 1−ν ν 0 ⎥
⎢ 0 0 ⎥
⎢ ⎥
E ⎢ ν 0 1−ν 0 0 0 ⎥
[ D] = ⎢ ⎥ (3.133)
(1 + ν)(1 − 2ν) ⎢ 0 0 0 1
−ν 0 0 ⎥
⎢ 2 ⎥
⎣ 0 0 0 0 1
2
−ν 0 ⎦
0 0 0 0 0 1
2
−ν

Note the strain state {ε} can be determined based on the displacement
- .T
{u} = u x , u y , u z at position (x, y, z) as
⎧ ⎫
∂u x
⎧ ⎫ ⎪⎪ ⎪
⎪
⎪ ε ⎪ ⎪
⎪ ∂x
∂u ⎪
⎪
⎪
⎪ x ⎪
⎪ ⎪
⎪ y ⎪
⎪
⎪
⎪ ε ⎪
⎪ ⎪
⎪ ∂y ⎪
⎪
⎪
⎪ ⎪
⎪ ⎪ ∂u ⎪
⎬ ⎪ ⎪
y
⎨ ⎨ ∂zz ⎬
εz
{ε} = = 1 ∂u x + ∂u y (3.134)
⎪
⎪ γx y ⎪
⎪ ⎪
⎪ 2 ∂y ∂x  ⎪
⎪
⎪
⎪ ⎪
⎪ ⎪
⎪ 1 ∂u y
⎪
∂u z ⎪
⎪ γ ⎪ ⎪ ⎪
⎪
⎪ yz ⎪
⎩ ⎪ ⎪
⎭ ⎪ 2 ∂z +
⎪ ∂y  ⎪
⎪
⎪
γx z ⎪
⎪ ⎪
∂u z ⎪
⎩ 1 ∂u x + ⎭
2 ∂z ∂x

Let {U} be the vector of state variables, {S(x, y, z)} be the vector of the shape func-
tions in an element, the displacement {u} in the position (x, y, z) can be interpolated
as
3.7 CAE Applications 191

{u} = [S(x, y, z)]{U} (3.135)

where the size of [S(x, y, z)] depends on (1) the number of nodes and (2) the degrees
on freedom on each node.
Substituting Eq. (3.135) into Eq. (3.14) obtains the relation of the strain state {ε}
and the vector of state variables {U} as

{ε} = [B]{U} (3.136)

where [B] is the matrix for the relation of strain state and state variables in an element
of the solid.
In structural analysis, the principle of the minimized potential energy can be used
to obtain element models about the displacements on nodes. Assume that external
force {F} is applied on nodes, and the potential energy of an element can be evaluated
as


n
=− Fi Ui (3.137)
i=1

where is the potential energy,  is the strain energy, n is DoF of the element, F i is
an external load applied on the i-th DoF, and U i is the displacement on the i-th DoF.

Example 3.9 A wall hook has a base size of 30 × 30 × 5 mm. The hook as a
diameter of φ5 mm, and its center is offset from the top surface of base by 7.5 mm.
The wall hook is printed with ABS plastics, and the yield strength is given as S y =
35 MPa. Assume the safety factor is n d = 2.0, determine the maximum load the hook
can carry without a failure.
Solution. Using the SolidWorks Simulation, the wall hook is modeled as shown in
Fig. 3.41a. The corresponding FEA model is defined in Fig. 3.41b; it is assumed
that the hook is fixed on wall (e.g., the base plate), and the nominal force Fnominal =
10 N is applied on hook. The simulation of the FEA model results in the stress and
displacement distribution, which are shown in Fig. 3.42a, b, respectively.

The maximum von Mises stress σ max is 6.585 MPa subjected to a nominal force
Fnominal = 10 N. Since the FEA model is a linear model, the stress is proportional
to the load, and the maximum allowable load for a safety factor of n d = 2.0 can be
determined as
# $# $
Fnominal S y 10 35106
Fallowable = = = 26.58(N )
n d σmax 2 6.585106
192 3 Computer-Aided Engineering (CAE)

Hook diameter φ 5 mm with

offset from surface 7.5 mm

Fixed geometry
Base of 30 × 30 × 5 mm Nominal force 10 N
(a) Hook model (b) FEA model

Fig. 3.41 FEA model of wall hook in Example 3.9

(a) Stress distribution (b) Displacement distribution

Fig. 3.42 Simulation result of the FEA model in Example 3.9

3.7.2 Modal Analysis

Modal analysis is to address an eigenvalue problem in which a set of the natural

frequencies are determined in an engineering system. As shown in Fig. 3.43, modal
analysis is critical to many engineering systems subjected to dynamic loads; since the
frequencies of external excitations should be away from natural frequencies except
that the systems such as vibrators utilize self-vibration to implement the system
functions. The mathematic models and solutions of eigenvalue problems have been
discussed in Sect. 3.5.6. Here, the CAE application in modal analysis is discussed.

Example 3.10 A manufacturer produces an integrated ventilating system for

customers, and Fig. 3.44 shows one of its products with two blowers. In mounting
3.7 CAE Applications 193

(a). Wind turbine (b). Construction (c). machine tool

Fig. 3.43 Modal analysis for example engineering systems

Blowers

Frame
Panels

Control system

Support Piping system

Fig. 3.44 Example of integrated ventilating system with two blowers

the blowers, a misalignment of the transmission plane of a blower will cause an exci-
tation to the system. Conduct a modal analysis to confirm that two blowers operate
at the frequencies that are below to the natural frequencies of the integrated system.
Solution. The SolidWorks Simulation is used to create an FEA model for frequencies
analysis of the ventilating system. Firstly, the assembly model is imported as shown
in Fig. 3.45a, and it is simplified to (1) suppress insignificant components such as the
control system, the piping system, and panels, (2) the interfaces at the contacts of two
components are checked to eliminate possible interferences for appropriate contact
conditions, (3) the material properties are defined for all of the inclusive components.
Secondly, a modal analysis is defined, the conditions of contact surfaces are specified,
the support components are defined as fixed geometries, the applied controls are
defined for some critical parts or areas, and the mesh is created for the assembly
model as shown in Fig. 3.45b. Finally, the number of the natural frequencies of
194 3 Computer-Aided Engineering (CAE)

(a) Simplified assembly model (b) Meshed model

Fig. 3.45 Example of integrated ventilating system with two blowers

Table 3.13 The first six natural frequencies of the integrated ventilating system
Mode No Frequency (Rad/s) Frequency (Hertz) Period (s)
1 73.049 11.626 0.086013
2 91.969 14.637 0.068318
3 136.22 21.679 0.046127
4 144.25 22.958 0.043557
5 207.12 32.964 0.030336
6 256.1 40.76 0.024534

interest is specified, the FEA model is solved, the list of calculated natural frequencies
is exported in Table 3.13, and the corresponding modal shapes are shown in Fig. 3.46.

Since the excitations are from the blower motors that are operated at the speed of
1750 resolutions per minute (RPM). Therefore, the excitations have the frequency
of 29.167 (Hz) or 183.25 (rad/s). This operating frequency lies between the fourth
(22.958 Hz) and the fifth (32.964 Hz) natural frequencies of the system in Table
3.13. It implies that the mounting misalignment may lead to an amplified vibration
in operation and a premature failure of some fastening parts such as brackets and
screws.

3.7.3 Heat Transfer Systems

Three primary physical quantities in thermodynamics are temperature, energy, and

entropy, and there four thermodynamic laws that describe their relations:
• The zeroth law: when the systems A and B are in an equilibrium with the third
system C. System A is in an equilibrium with system B; the concept of temperature
is defined to describe a thermal equilibrium.
3.7 CAE Applications 195

(a) (b)

(c) (d)

(e) (f)

Fig. 3.46 The first six mode shapes from simulation

• The first law: when the energy is transferred into or out of a system in the form of
work, heat, or matter, the energy in the system changes based on the conservation
of energy. Energy can be neither created nor destroyed, but it can be transferred
from one matter to another. Making a perpetual motion machine is impossible.
• The second law: in a natural thermodynamic process, the interaction of
thermodynamic systems increase the summed entropies of systems.
• The third law: the system entropy approaches a constant value when the temper-
ature in the system approaches absolute zero. The system entropy at absolute zero
is close to zero except for non-crystalline solids; the system entropy equals to the
logarithm of the product of the quantum ground states.
196 3 Computer-Aided Engineering (CAE)

Heat transfer refers to an exchange of heat energy, which is governed by the

first and second thermodynamic laws. According to the first thermodynamic law, the
internal energy of system changes when heat transfer occurs. According to the second
thermodynamic law, heat is transferred from the matter with a high temperature to the
matter with a low temperature. Analyzing heat transfer behaviors is very important
to many engineering systems due to a number of reasons: (1) an engineering system
is usually to transfer the motion and/or energy in one format to another or from
one component to another. (2) The properties of any materials are closely related
to temperature and heat. (3) Many products and manufacturing processes are the
implementation of heat transfer processes. Figure 3.47 shows three examples of the
engineering systems whose heat transfer behaviors must be analyzed in designing
and operating the systems.
Heat can be transferred in one of three basic modes: conduction, convection,
and radiation. In heat conduction, the thermal energy is transferred in the materials
by the electronic movements and the collisions of microscopic particles such as atoms
and molecules. As shown in Fig. 3.48, the Fourier’s law can be used to represent a
heat transfer behavior as

(a). Heat treatment (b). Renewable energy (c). Air conditioning

Fig. 3.47 Example of heat transfer systems

Fig. 3.48 Heat transfer and

generation in a continuous
domain

dz

dy

dx
3.7 CAE Applications 197
⎫
qx = q̇x A x = −k x A x ∂∂Tx ⎪
⎬
q y = q̇ y A y = −k y A y ∂∂Ty (3.138)
⎪
⎭
qz = q̇z A z = −k z A z ∂∂zT

where qx , q y , andqz are heat transfer rates along x-, y-, and z-axes;
q̇x , q̇ y , andq̇z are heat fluxes along x-, y-, and z-axes;
k x , k y , andk z are the conductivity coefficients along x-, y-, and z-axes;
∂T ∂T
, , and ∂∂zT are the temperature gradients along x-, y-, and z-axes;
∂x ∂y
A x , A y , and A z are the effective areas of conductions perpendicular to x-, -y, and
z-axes, respectively.
In heat convection, the heat is transferred by the fluid motion over the surface of
an object. Heat convection occurs when the surface temperature is different from the
surrounding environment, which is governed by the Newton’s law as,


qh = h A Ts − T f (3.139)

where qh is the rate of heat convection over the surface;

A is the effective area of heat convection;
T s is the temperature on surface;
T f is the fluid temperature.
In heat radiation, the heat is transferred in the form of electromagnetic waves. The
thermal energy is converted into the electromagnetic energy and emitted as the radi-
ation over surfaces. Heat radiation is governed by the Stefan-Boltzmann law
as,


qr = σ ε A Ts4 − TO4 (3.140)

where σ is the Stefan-Boltzmann constant;

ε is the coefficient of surface emission;
A is the effective area of heat radiation;
T 0 is the absolute temperature of the environment;
T s is the absolute temperature over the surface;
qr is the radiant heat flow over the surface.
In an infinitesimal volume (dxdydz), the flows of heat transfers are shown in
Fig. 3.48. The heat may be conducted along x, y, and z axes. The amount of the
thermal energy in the volume is determined by (1) heat conduction along three axes,
(2) the temperature change in the volume, and (3) the generated energy. heat for
temperature changes of the mass in unit, and (4) the heat generated from any source
in unit. The corresponding amount of thermal energy for each heat transfer mode
can be calculated as below.
The energy E in that is flowed into the volume by conduction is,
198 3 Computer-Aided Engineering (CAE)

∂T ∂T ∂T
E in = qx + q y + qz = −k x dydz − k y d xdz − k z d xd y (3.141)
∂x ∂y ∂z

where qx , qy , and qz are given in Eq. (3.138) and ∂∂Tx , ∂∂Ty , and ∂∂zT are the temperature
gradients along x, y, and z axes.
The energy E out that is flowed out of the volume by conduction is,

E out = qx+d x + q y+dy + qz+dz



∂T  ∂T
∂T 
∂ ∂ ∂y ∂
∂x ∂z
= qx + q y + qz − k x d xd ydz − k y d xd ydz − k z d xd ydz
∂x ∂y ∂z
(3.142)


∂T
∂ ( ∂∂Tx ) ∂ ∂ y∂( ∂T )
where ∂x
, and ∂z∂z are the second derivatives of temperature along x, y, z
∂y
,
axes, respectively.
The amount of energy E g generated by the internal source is,

E g = q̇d xdydz (3.143)

where q̇ is the rate of heat generated in unit.

The amount of the energy E s due to the temperature change is,

∂T
E s = ρc d xd ydz (3.144)
∂t

where ρ and c are the density and specific heat of materials, and ∂∂tT is the temperature
gradient over time.
According to the first thermodynamic law, the thermal energy in the volume must
be conserved; therefore, we have,

E in − E out + E g = E s (3.145)

Substituting Eqs. (3.141)–(3.144) into Eq. (3.145) yields the governing equations
of heat transfer behaviors as,
# $ # $ # $
∂T ∂T ∂T ∂T ∂T ∂T ∂T
kx + ky + kz −c + q̇ = 0 (3.146)
∂x ∂x ∂y ∂y ∂z ∂z ∂t

When q̇ = 0 and ∂∂tT = 0, Eq. (3.146) becomes a differential equation for an

equilibrium problem; when
= Tt 0, Eq. (3.147) becomes a differential equation for a
propagating problem. The FEA solutions to both types of differential equations have
been discussed in Sect. 3.5.6.
3.7 CAE Applications 199

Note that Eq. (3.146) describes the heat transfers within a given volume; however,
the heat convection and radiation may occur on the boundary surfaces. Therefore,
the heat convection and radiation should be modelled as the boundary conditions.
Figure 3.49 shows five types of the boundary conditions in a heat transfer problem:
(a) a fixed temperature, (b) a given heat flux, (c) a coefficient for convection, (d) the
parameters for radiation, and (e) an insulated condition when no heat transfer occurs.

Example 3.11 A heatsink model of a central processing unit (CPU) is shown in

Fig. 3.50. Assume that the heatsink uses copper materials, the heat transfer coefficient
under the forced air is h = 15 W/(m2 ·K), and the ambient temperature of the operating
environment is given as T 0 = 25 °C. Find (1) the maximum temperature in CPU

qs
Ts qs
Ts
Ts(x,y, t) Ts(x,y, t)

qs=dT/ds

(a) Fixed temperature on surface (b) Given heat flux on surface

Tf

Ts
Ts

T(x,y, t) T(x,y, t)

(c) Convection through surface (d) Radiation through surface

qs=0

Ts(x,y, t)

(e) Insulated surface

Fig. 3.49 Five types of boundary conditions in a heat transfer problem

200 3 Computer-Aided Engineering (CAE)

Heatsink: cooper, h = 15 W/(m2 •K),T0 =25ºC All exposed surfaces over

heatsink as BCs for
convections

CPU: size 22 × 26 × 2 mm, 0.5 watts CPU has heat power of 0.5 W

(a) Heatsink model (b) Thermal analysis model

Fig. 3.50 Heat transfer model in Example 3.11

and (2) the maximum heat power the heatsink can dissipate when the maximum
temperature can be below T max = 60 °C.
Solution. The heat transfer model for the heatsink has been defined in Fig. 3.50b. Two
boundary conditions are (1) the convection condition on all of the exposed surfaces
of heatsink, and (2) the heat power defined in the body of CPU. The materials of
CPU is set as silicon. The mesh is created using the default settings. Running the
simulation yields the results of Fig. 3.51a, b for the distributions of the temperature
and temperature gradients, respectively. It has been found the maximum temperature
at the center of CPU is T max = 30.95 °C.

The above thermal analysis model is used to create a design study as shown in
Fig. 3.52. The heat power (P) from CPU is set as the simulation variable, and the
analysis range is set as P ∈ [0.5, 3] with a step of 0.25 (W ). The result of the design

(a) Temperature (b) Temperature gradient (K/m)

Fig. 3.51 Simulation result of the heat transfer model in Example 3.11
3.7 CAE Applications 201

Fig. 3.52 Design study for different level of heat power from CPU

65

60
Max temperature ( ºC)

55

50 (2.75 W, 57.71ºC )
45

40

35

30

25
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 2.75 3
CPU heat power (W)

Fig. 3.53 Max temperature (°C) versus heat power (W)

study in Fig. 3.53 shows that the heatsink reach its maximum temperature T max =
57.71 °C < 60 °C when the heat power of CPU P = 2.75 (W ).

3.7.4 Transient Heat Transfer Problems

The above thermal analysis is for an equilibrium-engineering problem; the steady

solution is found when the input and output heat powers in solids are balanced.
However, the changes of state variables at the transient stages are more critical
in some manufacturing processes such as heat treatments and injection molding
processes. An FEA model for a transient problem requires discretizing the time
domain; the solution at each step is approximated as an equilibrium problem whose
boundary conditions are updated based on the simulation results at previous steps.
In this section, the heat treatment of an aluminum part is used as an example to show
how a CAE tool such as SolidWorks Simulation can be used to solve a transient
problem.
202 3 Computer-Aided Engineering (CAE)

Example 3.12 The par geometry is shown in Fig. 3.54a; it is made of 6061-T6
aluminum alloy, the initial temperature of the heat treatment process is T 0 = 532 °C.
The ambient temperature is T a = 30 °C and the coefficient of heat convection is h
= 60 W/(m2 ·K), find out the cycle time of the heat treatment process for the parts to
reach a temperature of T f = 35 °C.
Solution. The SolidWorks Simulation is used to define a transient thermal analysis
model shown in Fig. 3.55. The duration of simulation is set as 300 s, and the step for
the discretization over the time domain is 10 s. Figure 3.56 shows that the FEA model

Fig. 3.54 Geometry and properties of part in Example 3.12

Fig. 3.55 Define a transient thermal analysis in SolidWorks Simulation

3.7 CAE Applications 203

Convection over all

exposed surface

Temperature of body at
543 °C when t = 0 s

Fig. 3.56 Boundary conditions (BCs) of transient thermal analysis model

(a) Temperature (°C) (b) Temperature gradient (°C/m)

Fig. 3.57 The distributions of temperature and temperature gradients at 300 s

includes two boundary conditions: (1) the initial temperature in the solid is T 0 = 532°
C at t = 0 (s) and (2) the convections over all exposed surfaces have h = 60 W/(m2 ·K).
Figure 3.57 shows the distribution of temperature and temperature gradient when t
= 300 s. Figure 3.57 shows the cooling curves of the part in heat treatment. It has
been found that the maximum temperature of the part T max = 34.77 °C < 35 °C at
the time step of t = 240 s. Therefore, the cycle time of the heat treatment process is
240 s (Fig. 3.58).

3.7.5 Fluid Mechanics

Fluid mechanics is a branch of continuous mechanics. Fluid mechanics studies the

behaviors of fluid systems subjected to static and dynamics loads. The matter in a
204 3 Computer-Aided Engineering (CAE)

600

MaxTemp MinTemp TargetTemp

500
Max & Min Temperature (°C )

400

300

200

100
(240 s, 34.77 C )

0
0 50 100 150 200 250 300
Time (seconds)

Fig. 3.58 The max, min, and target temperatures over time in heat treatment

fluid system should be modelled as continuous masses rather than discrete particles;
since the matter such as fluid or gas does not hold a shape like a solid (Bar-Meir 2008).
In addition, the density of the matter may vary with time and position. Figure 3.59
shows continuous mechanics includes the branches of solid mechanics, fluid structure
interaction, and fluid mechanics. Fluid mechanics can be further classified into fluid
statics, and fluid dynamics. Fluid systems can also be classified from the perspec-
tive of applications such as internal/external flows, single/multi-phase flows, and
laminar/turbulent flows, pipe flows, and groundwater flows. In this section, FEA
modelling of fluid systems is discussed.

Solid Pipe flow

Mechanics
problems
Stability

Internal or
External flow
Fluid
Continuous Structure Single/multi-
Mechanics Interaction Fluid
Mechanics Statics phase flow
Boundaries

Laminar or
problems

Fluid Fluid turbulent flow

Mechanics Dynamics
Groundwater
flow

Fig. 3.59 Examples of engineering problems in fluid mechanics

3.7 CAE Applications 205

Solid and fluid respond shear stresses differently. A fluid deforms continuously
subjected to shear stresses, and it does not return to its original state after the defor-
mation. In contrast, a solid exhibits a fine deformation that does not change with
time. A fluid moves or flows subjected to shear stresses. A fluid can be a liquid or
gas.
The earliest study on computer simulation of fluid systems was for potential
flows in the 60 s. A potential flow was modelled as a velocity field of incompressible
fluids; it was later expanded to compressible fluids (Ritchmeyer and Morton 1967).
A fluid flow transfers mass, momentum and energy, and its motion is governed by
the Navier–Stokes equations for mass, momentum and energy conservations. The
numerical simulation to the Navier–Stokes equations was implemented in the 70’s,
and nowadays, CAE has been widely used to analyze fluid systems in heat, ventilation
and air-conditioning (HVAC) and manufacturing industry.

3.7.5.1 Constitutive Models

A fluid flow is different from solids in sense that a fluid flow can sustain a hydrostatic
stress but not shear stress. Therefore, the behavior of a fluid flow is not represented
by the displacement like a solid; it is represented by the flow velocity (v) in a three
dimensional space as,

v(x, y, z) = u(x, y, z) · i + u(x, y, z) · j + w(x, y, z) · k (3.147)

where (x, y, z) is an arbitrary position in the 3-D space; v(x, y, z) is the vector of the
fluid flow velocity; and u(x, y, z), v(x, y, z), and w(x, y, z) are the components of the
velocity vector along x-, y- and z- axes, respectively.
With the given velocity, the strain rate εij (i, j = 1, 2, 3) in the fluid flow can be
derived as,
⎫ ⎫
1 ∂u ∂v ⎪
ε̇11 = ∂∂ux ⎪
⎬ ε̇12 = 2 ∂y
+ ∂x  ⎪
⎬
∂v
ε22 = ∂ y and ε̇23 = 1 ∂v
+ ∂w (3.148)
⎪
⎭ 2 ∂z

∂u ∂y ⎪
⎪
ε̇33 = ∂w ∂w ⎭
∂z ε̇13 = 2 ∂z + ∂ x
1

where u, v, and w are the components of the velocity vector given in Eq. (3.147).
Equation (3.148) can be further written in the matrix form as,
T
ε̇ T = ε̇11 ε̇22 ε̇33 2ε̇12 2ε̇23 2ε̇13 = [S][v] (3.149)

T
where [v] = u(x, y, z) v(x, y, z) z(x, y, z) is given in Eq. (3.147).
the subscripts 1, 2, and 3 represent x, y, and z axes, respectively, and
206 3 Computer-Aided Engineering (CAE)

[S] is the operator matrix for the strain rate given by Eq. (3.148),
⎡ ⎤
∂
x
00
⎢ 0 ∂ 0 ⎥
⎢ ∂y ⎥
⎢ ∂ ⎥
⎢ 0 0 ∂z ⎥
[S] = ⎢
⎢1 ∂ 1 ∂ 0 ⎥
⎥ (3.150)
⎢ 2 ∂y 2 ∂x ⎥
⎢ 1 ∂ 1 ∂ ⎥
⎣ 0 2 ∂z 2 ∂ y ⎦
1 ∂
2 ∂z
0 21 ∂∂x

Once the strain rate is obtained, the stress σ ij can be derived accordingly. However,
two constants should be defined to calculate the stress from the strain rate. The first
one is called the deviatoric stress τ ij , which relates to the deviatoric strain rate as,
# $
ε̇kk
τi j = 2μ ε̇i j − δi j (3.151)
3


where the term ε̇i j − δi j ε̇3kk is called the deviatoric strain rate,
δi j is the Kronecker delta constant,
the subscripts i, j, k = 1, 2, 3, and 1, 2, and 3 represent x, y, and z axes, respectively,
and
μ is the coefficient for the viscosity.
The second constant is the pressure in a fluid flow, which depends on the strain
rate. The pressure is calculated from the volumetric strain rate as,

p = −κ ε̇ii + p0 (3.152)

where κ is the coefficient of volumetric viscosity similar to the bulk modulus for an
elastic solid; p0 is the initial pressure.
he constitutive model for the stress–strain relation is derived from Eqs. (3.149)–
(3.152) as,
# $ 
2
σi j = τi j − δi j p = 2με̇i j + δi j κ − μ ε̇ii + p0 (3.153)
3

Equation (3.153) can be further simplified since the volumetric viscosity can be
ignorable. Therefore, the constitutive model of the fluid flow is written as,

   
∂v j
τi j = 2μ ε̇i j − δi j ε̇3kk = μ ∂∂vx ij + ∂ xi
− δi j ∂∂vxkk
(3.154)
σi j = τi j − δi j p

Equation (3.154) of the fluid flow is similar to the constitutive model of a solid.
It is worth to note that the characteristic of a fluid flow depends on viscosity (μ)
3.7 CAE Applications 207

in Eq. (3.154). If the viscosity is constant, the fluid flow is a Newtonian fluid flow;
otherwise, the fluid flow is a non-Newtonian fluid flow whose viscosity varies with
the strain rate.

3.7.5.2 Mass Conservation

The mass of a matter can be not created or vanished. Since a fluid does not hold a
shape and it flows under an external force, it is convenient to evaluate the mass of
the fluid in terms of the flow rate passing a given section of the container. As shown
in Fig. 3.60, the mass rates of the fluid flow at two sections of an enclosed volume
should be the same,

ṁ 1 = ṁ 2 = ρ1 v1 A1 = ρ2 v2 A2 (3.155)

where ṁ 1 and ṁ 2 are the scalar mass rates; ρ 1 and ρ 2 are the densities, v1 and v2 are
the scalar velocities, and A1 and A2 are the sectional areas.
In a 3-D space, the mass flow is defined as a vector shown in Fig. 3.61. An
infinitesimal volume with the size of dx × dy × dz is considered, the velocity of
flow is represented by (u, v, w), the density of fluid is ρ, the rate of density change is
dρ/dt, and a source of mass is Ṡm . The mass conservation is expressed as,

Fig. 3.60 Mass conservation

in an enclosed volume

Fig. 3.61 Representation of

the fluid flow in an
infinitesimal volume
208 3 Computer-Aided Engineering (CAE)

dm
ṁ in − ṁ out = (3.156)
dt

where ṁ in and ṁ out are the masses flow in and out of the infinitesimal volume, and
dm
dt
is the rate of mass change over time.
Expanding Eq. (3.156) with the variables in Fig. 3.61 yields,

ρ(udydz + vd xdz + wd xdy)

⎛# $ # $ ⎞
(ρu) (ρv)
ρu + d x dydz + ρv + dy d xdz+
⎜ x y ⎟
−⎜⎝# $ ⎟
⎠
(ρw)
ρw + dz d xd y
z
dρ
+ d xd ydz = Ṡm d xd ydz (3.157)
dt
Equation (3.156) can be further rewritten as,

dρ
∇(ρv) + = Ṡm (3.158)
dt

where ∇(ρv) = ∂(ρu) ∂x

+ ∂(ρv)
∂y
+ ∂(ρw)
∂t
is the divergence of fluid flow and Ṡm is the
rate of mass generation.
In fluid mechanics, Eq. (3.158) is referred as the continuity equations described
in a variety of forms.

3.7.5.3 Momentum Conservation

Energy may be transferred from one form to another. In a fluid flow, energy exists in
three forms as below.
The first form is potential energy (PE) that is relevant to the altitude of the fluid
as,

PE = W · H (3.159)

where PE is the amount of potential energy, W and H are the weight and altitude of
the fluid weight, respectively.
The second form is kinetic energy (KE) that is associated with the fluid movement
as,

v2
KE = W · (3.160)
2g
3.7 CAE Applications 209

where KE is the amount of kinetic energy, v the velocity of the fluid flow.
The third form of energy is determined by the flow pressure (FP) as,

W·p
FP = (3.161)
g

where p is pressure of the fluid, and g is the gravity acceleration.

Adding the energies in three forms in Eqs. (3.159–3.161) finds the total energy
of the fluid as,
# $
v2 p
E = PE + K E + FP = W H + + (3.162)
2g g

When the fluid flow is steady, some energy is consumed to overcome the friction
along the streamline. In such a case, the momentum of the flow should be conserved
and this leads to the Bernoulli’s equation as,

v12 p1 v2 p2
H1 + + = H2 + 2 + + hL (3.163)
2g g 2g g

v2 v2
where H1 and H2 are the heads, p1 and p2 are the pressures, and 2g1 and 2g2 are
velocity heads at two points over the streamline, respectively, and hL is the head loss
by friction, fittings, bends, and valves as (Acharya 2016),
2
h L = k 2g
v
(minor loss f r om valves, f itting and bents)
L v2 (3.164)
h L = f D 2g ma jor loss f rom f riction and pumping

where k and f is the friction factors, L is the distance of two sections, v is the average
velocity, and g is the gravity acceleration.
In an open space, the condition of momentum conservation can be expressed as,
⎫
+ (∂ x ) + ∂(ρuv)
∂ ρu 2 ∂τ
∂(ρu)
∂t ∂y
+ ∂(ρuw)
∂z
− ∂τ∂ xx x − ∂ yx y − ∂τ∂zx z + ∂∂ px − ρ fx = 0 ⎪
⎪
⎬
+ (∂ y ) + ∂(ρuv)
∂ ρv 2
∂τ ∂τ ∂τ
∂(ρv)
+ ∂(ρvw) − ∂ yyy − ∂ xx y − ∂zyz + ∂∂xpy − ρ fy = 0 (3.165)
∂ ∂x ∂z ⎪
⎪
⎭
+ ( ∂z ) + ∂(ρuw)
∂ ρw2 ∂τ
∂(ρw)
∂ ∂x
+ ∂(ρvw)
∂y
− ∂τ∂zzz − ∂τ∂ xx z − ∂ yyz + ∂∂zp − ρ fz = 0

3.7.5.4 Energy Conservation

Equation (3.165) is applicable to an isothermal state where no heat transfer occurs. If

a fluid flow involves in the temperature change, the thermal energy must be taken into
210 3 Computer-Aided Engineering (CAE)

the consideration in the energy conservation. When temperature changes, density (ρ)
depends on pressure (p) and temperature (T ) as,

ρ = ρ( p, T ) (3.166)

The density function Eq. (3.166) depends on the fluid type. For example, the
density of ideal gas can be expressed explicitly as,
p
ρ= (3.167)
R·T

where R is the universal gas constant.

There are other three types of energy (1) intrinsic energy e, (2) the kinetic energy
relevant to the fluid motion, and (3) the energy relevant to pressure of fluid; therefore,
the equivalent heat of the system is found as,
p vi v j p
H=E+ = e( p, T ) + + (3.168)
ρ 2 ρ

where E is the sum of intrinsic energy and kinematic energy; H called as is the
enthalpy of the system, e(p, T ) is the intrinsic energy, and vi (i = 1, 2, 3) corresponds
to u, v, and w in Fig. 3.61.
The substantial derivative of the total energy is found as,
# $ # $
d(ρ E) ∂ρ ∂ρvi ∂ρ E ∂ρ Evi
=E + + + (3.169)
dt ∂ ∂ xi ∂t ∂ xi

Heat energy can be also transferred by convection, conduction, and radiation. For
example, the heat flux qi relates to the temperature gradient as,

∂T
qi = −k (3.170)
∂ xi

where k is the conductivity coefficient, and x i (i = 1, 2, 3) correspond to x, y, z axis,

respectively.
In addition, energy can be dissipated due to internal stresses that can be calculated
as,

∂
 ∂
 ∂ pi

σi j v j = τi j v j − pv j (3.171)
∂ xi ∂ xi ∂x j

As the result, the energy conservation with the consideration of thermal energy
becomes,
3.7 CAE Applications 211
# $
∂(ρ E) ∂(ρvi H ) ∂ ∂T ∂

+ − k − τi j v j − ρgi vi − q H = 0 (3.172)
∂ ∂ xi ∂ xi ∂ xi ∂ xi

Note that in many cases, not all of the equations for mass, momentum, and energy
conservations are needed to find the solutions to fluid mechanics problems.

3.7.5.5 SolidWorks Flow Simulation

Fluid mechanic problems can be solved in SolidWorks Flow Simulation, which is

an intuitive computational fluid dynamics (CFD) solution to fluid flows. It adopts
the Cartesian-based meshes for solid–fluid and soli-solid interfaces to increase the
flexibility of meshing process, and the meshes are automatically based on specified
control parameters. To illustrate the application of the CAE tools in solving various
fluid mechanics problems, the SolidWorks Flow Simulation is used to evaluate the
fuel efficiency of car body.

Example 3.13 Analyze the fluid flow around an example racecar model shown in
Fig. 3.56a and estimate the drag force when the racecar run at the speed of 90 m/s
(~200 miles per hour).
Solution. The SolidWorks Flow Simulation is used to simulate the airflow around the
rackcar body. The flow Simulation tool is included in the add-ins of the SolidWorks.
It can be activated by selecting the tool in the list of add-ins under Options → Add-ins
as shown in Fig. 3.62. The SolidWorks Flow Simulation provides the project wizard
to guide users in defining a flow simulation model by following the main steps
shown in Fig. 3.63. Firstly, the CAD model is imported, the configuration of interest
is specified, and the name of flow simulation is created. Secondly, the preferable
units are specified for all of parameters and variables used in the simulation model.

(a) Example race car model (b) Flow simulation in add-ins

Fig. 3.62 Create a flow simulation model for an example racecar body
212 3 Computer-Aided Engineering (CAE)

(a) Step 1: specify a configuration in CAD model (b) Step 2: choose units for parameters of interest

(c) Step 3: define analysis type (d) Step 4: define properties of fluid flow

(e) Step 5: define wall conditions (f) Step 6: define initial and ambient conditions

Fig. 3.63 Main steps in defining a flow simulation model

Thirdly, the analysis type is given; the flow simulation can be for an internal or
external flow; moreover, it is optional to include (1) cavities or internal surfaces, (2)
heat transfer and radiation, and (3) historical data over time. An external simulation
with excluded cavities is defined in this example. Fourthly, the types and properties
of fluid are defined; the software include a design library that have many commonly
used fluids and gases. Fifthly, the wall conditions are defined for fluid–solid contact
surfaces. Sixthly, initial conditions of fluid are defined; the relative speed (i.e., 90 m/s)
of the airflow and racecar is defined along the x-axis for this example.

A flow simulation model involves in a large number of state variables, and it is

unnecessary to include all of derived variables in the solving process. Figure 3.64a
shows the interface where the variables of interest can be specified as the analysis
goals. In addition, the computational volume can be should be adjusted with the
consideration of computation time and accuracy (see Fig. 3.64b); users can also
3.7 CAE Applications 213

(a) Define analysis goals (b) Refine the computational volume

Fig. 3.64 Define analysis goals and refine computational volume

Fig. 3.65 The statistics of meshes and the solving process

specify the levels or sizes of meshes or apply control meshes in the areas of interest
when they are needed.
Figure 3.65 gives the statistic data of solid and fluid meshes and the solving
process. Figure 3.66 shows the example plots of velocity and pressure distribution
over a selected cutting plane. The goals of the flow simulation is shown in Table
3.14. It is found that the drag force along moving direction (x-axis) is 4802 (N).

3.8 Summary

An engineering problem can be usually formulated into a design problem with the
objectives, inputs, system parameters, outputs, and constraints from the perspective
of system. To solve a design problem, the possible solutions in the design space are
214 3 Computer-Aided Engineering (CAE)

(a) Velocity distribution

(b) Pressure distribution

Fig. 3.66 Visualization of velocity and pressure distribution over a cut plane

Table 3.14 The results of analysis goals in flow simulatioon

Goal Name Unit Averaged value Minimum value Maximum value
GG Bulk Av Velocity [m/s] 87.28900777 87.28787613 87.29089342
1
GG Bulk Av Velocity [m/s] 86.92918179 86.92779739 86.93108198
(X) 1
GG Bulk Av Velocity [m/s] 2.036041808 2.034343274 2.037097925
(Y) 1
GG Bulk Av Velocity [m/s] 0.038781473 0.036952233 0.040201268
(Z) 1
GG Force 1 [N] 4923.384152 4919.574211 4928.137101
GG Force (X) 1 [N] 4797.504651 4794.285348 4802.072032
GG Force (Y) 1 [N] −1089.135225 −1092.571206 −1085.872086
GG Force (Z) 1 [N] −193.4805561 −196.6832158 −187.0958669

analyzed and compared, and design synthesis is performed to obtain the optimized
solution. When the scope and complexity of a design problem increases, engineering
design goes far beyond the capabilities of analytical or experimental approaches, the
numerical methods such as CAE tools become imperative to model, simulate, and
evaluate engineering solutions. In this chapter, the fundamental theory of numerical
simulation is introduced, and the focus lies on the implementation of finite element
analysis (FEA). The SolidWork Simulation is used as an example CAE tool to analyze
engineering problems in different disciplines. From this chapter, students should
3.8 Summary 215

master the knowledge and skills of (1) formulating an engineering problem as a

design problem and (2) utilizing the CAE tools for design analysis and synthesis of
any engineering systems.
Design Problems

Problem 3.1 Formulate the following real-world problems as engineering design

problems that include design objectives, inputs, system parameters, outputs, and
design constraints.

(a) Due to a high strength-weight ratio, composite materials have been widely
adopted in aerospace products, and industrial robots are used for light
machining of composite parts. However, one challenge is to collect the dwarf
and dust generated from machining process in open space, which is shown
in Fig. 3.67; since the dissipated dusts will pollute working environment (Bi
2010). Explore an engineering solution to collect dusts at sources.
(b) An Industrial robot with an open kinematic chain usually has its limitation
of low workload for machining process. In contrast, a parallel kinematic
machine with closed kinematic chains is capable of offering a high work-
load for light machining (see Fig. 3.68a). However, the rigidity of a parallel
kinematic machine varies from one place or orientation to another (Bi 2014).
Explore an engineering solution to predict the accuracy of a parallel kinematic
machine in real time.
(c) Many machine elements are used to transfer motions and power in machines.
The wears of machine elements determine their lifespans. The lifespan of a
machine element relates to many factors such as loads, lubrications, mate-
rial properties, surface properties, pressure, and temperature. Explore an engi-
neering solution which is capable of predict fatigue life of a type actuators
shown in Fig. 3.69a.

Problem 3.2 Determine the workspace of the 2-DOF robot shown in Fig. 3.70. Note
that the motion ranges of joints A and B are θ A = (−80◦ , 80◦ ) and θ B = (−60◦ , 60◦ )

Spindle

Host

Air chamber

Adjustment

Brush
Workpiece

(a) End-effector attachment (b) Simulation of dust collection

Fig. 3.67 Illustration of design Problem 3.1(a) (Bi 2010)

216 3 Computer-Aided Engineering (CAE)

(a) Parallel kinematic machine (b) Example of stiffness distribution

Fig. 3.68 Illustration of Design Problem 3.1(b) (Bi 2014)

(a) Lead screw actuator (b) Fatigue analysis

Fig. 3.69 Illustration of Design Problem 3.1(c) (Bi and Meruva (Bi and Meruva 2019)

Problem 3.3 Determine the voltage drop in each resistor of the circuit shown in
Fig. 3.71.

Problem 3.4 Determine the maximum displacement in the truss structure in

Fig. 3.72; note that the element model of a truss member in 2-D space is shown
below,
⎡ ⎤⎧ ⎫ ⎧ ⎫
l2 lm −l 2 −lm ⎪ ⎪ ui ⎪ ⎪ fi x ⎪
⎢ ⎨ ⎪ ⎬ ⎪
⎨ ⎪
⎬
E A ⎢ lm m2 −lm −m 2 ⎥
⎥ vi = fi y
L ⎣ −l 2 −lm l2 lm ⎦⎪⎪ u ⎪ ⎪ f jx ⎪
⎩ j⎪ ⎭ ⎪
⎩ ⎪
⎭
−lm −m 2 lm m2 vj f jy

where E, A, L are the Young module, the cross-section area, and the length of truss
element, l = cosθ, m = sinθ, and θ is the direction of truss with respect to x-axis.
3.8 Summary 217

End-effector
Y
Fixed ground
support

B X
A C

Length in the
unit of mm

Fig. 3.70 A two dimensional robot in Design Problem 3.2

R1=10 Ω

I=2A R2=10Ω R3=20Ω

R4=10Ω

R5=20Ω

Fig. 3.71 The circuit diagram in Design Problem 3.3

A = 1.44×10-4 m2,
E = 2.0×1011 N/ m2
2000 (N)
γ = 0.265

60° 60°

Fixed Roller
support 1000 mm support

(a) Truss structure subjected to 2000 (N) (b) Truss section (mm)

Fig. 3.72 The truss structure in Design Problem 3.4

218 3 Computer-Aided Engineering (CAE)

Fixed support

thickness = 1 in
7 2
E =2.9×10 lbf/ in
γ = 0.3

F = 200 (lbf)
Fig. 3.73 A two-dimensional thin plate subjected to an axial load

Problem 3.5 Figure 3.73 shows a thin plate subjected to an axial load F, create a
one-dimensional model with 4 axial members to find the displacement at free end.
Note that the element model of an axial member is,
 "  " 
E A 1 −1 ui fi
=
L −1 1 uj fj

Problem 3.6 Find a weak solution to the following boundary value problem:
⎫
d2 y
+ 2x − y = 0 0 ≤ x ≤ 1 ⎪
⎬
dx2
subjected to : y(0) = 1
dy(0) ⎪
⎭
dx
=0

Problem 3.7 Figure 3.74 shows a torsional member that consists of three segments
AB, BC, and CD with different materials. Determine the angular displacement at D

φ = 50 φ = 60 φ = 50

Fixed A B C
support D = 250 ( ⋅ )

= 500 ( ⋅ ) = −1000 ( ⋅ )

GAB= 40 GPa GBC= 30 GPa GCD= 80 GPa

Fig. 3.74 A torsional member in Problem 3.7

3.8 Summary 219

relative to position A. Note that the element model of torsional member is,
 "  " 
G J 1 −1 θi Ti π d4
= J=
L −1 1 θj Tj 32

Problem 3.8 Figure 3.75 shows a truss structure in three-dimensional space with
the specified loads F = (0, 50, −100) lbf at two nodes (G and I) and F = (0, 0, −
500) lbf at node H. The nodes at the ground (A, B, C, D, E, F) are all fixed. All truss
members use the same material—Gray Case Iron, which has the Young’s modulus
of E = 9.598 × 106 psi and the Poisson’ ratio of 0.27. The yield strength is S y =
2.20 × 104 psi. The cross-section areas of all truss members are given as A = 1 in2 .
Predict the maximal stress and deflection of the truss structure.
Problem 3.9 Figure 3.76 shows a thin plate with a thickness of 0.2-in. The plate
includes three discontinuities of geometries, a Φ 1.5-in hole, a Φ 2.0-in hole, and
a 0.5-in fillet at the shoulder. The plate uses Aluminum 1060 alloy with the elastic
modulus of E = 1.0 × 107 psi, the Poisson’ ratio ν = 0.33, yield strength S y = 3.999 ×
103 psi. Determine stress concentration factors at three discontinuity sections under
a bending load.
Problem 3.10 Figure 3.77 shows a truss structure in two-dimensional space. Nodes
A and B are fixed. All truss members use the same materials Gray Case Iron, which
has the Young’s modulus E = 9.598 × 106 psi and the Poisson’ ratio of 0.27. The
yield strength is S y = 2.20 × 104 psi. The cross-section areas of all truss members
are given as A = 1.0 in2 . Use the Solidworks Simulation to estimate the first 4 natural
frequencies.
Problem 3.11 Figure 3.78 shows a fin to transfer the heat from base by convection;
the dimensions of lengths are in inch. The temperature of fluid flow is 20 °C, and
heat transfer coefficient h = 0.1 Btu/in2 ·s·°C. The coefficient of conduction of fin is
k = 3.0 Btu/in·s·°C, determine the temperature distribution in the fin.

Unit: in
Fixed nodes:
Grid size: 10-in G A, B, C, D, E, F
Material: Gray Case Iron
D
Coordinates:
H A: (10, 10, 0) in
Z Y E B: (50, 10, 0) in
I C: (90, 10, 0) in
A D: (10, 90, 0) in
F E: (50, 90, 0) in
Loads on nodes: X
G: (0, 50, -100) lbf F: (90, 90, 0) in
H: (0, 0, -500) lbf B G: (10, 50, 40) in
I: (0, 50, -100) lbf H: (50, 50, 40) in
I: (90, 50, 40) in
C

Fig. 3.75 A truss structure in Problem 3.8

220 3 Computer-Aided Engineering (CAE)

Fig. 3.76 A plane stress plate with a thickness of 0.2-in

Fig. 3.77 A Truss Structure in Problem 3.10

Base temperature
80 °C
Surrounding convection:
h = 0.1 Btu/in2⋅sec⋅°C
Tf = 20°C

Fig. 3.78 Example of heat transfer problem (length unit: inch)

Problem 3.12 Figure 3.79 shows a wall consisting of two materials, i.e., the first
layer is 2-cm with the conductivity of k 1 = 0.2 W/cm·s·°C and the second later is
6-cm with the conductivity of k 2 = 0.05 W/cm·s·°C. Other parameters related to the
boundary conditions are shown in Fig. 3.79. Create a simplified model in SolidWorks
Simulation to estimate the distribution of temperature in the wall.

Problem 3.13 A 2-D domain with an incompressible irrotational fluid flow is shown
in Fig. 3.80. The length is 2 inches and the height is 1 inch, and there is the obstacles
3.8 Summary 221

k1=0.2 W/cm⋅sec⋅°C k2=0.05 W/cm⋅sec⋅°C

Tf= −5°C
h=0.1 W/cm2⋅sec⋅°C Troom=20°C

2-cm 6-cm

1 2 3

Fig. 3.79 1-D heat transfer through wall

D C
V = 1 BTU/s/in2 on A-D

A B
Atmosphere
pressure on B-C

Fig. 3.80 Ideal fluid flow in a 2-D domain for problem 3.13 (length unit: inch)

in the middle with 4 circles of φ 0.2-in on a circle of φ 0.5-in. Assume that the
reference atmosphere pressure at the outlet is and the fluid speed at the inlet is 1.0
BUT/s/in2 . Create an FEA model and determine the distribution of the velocity in
the fluid domain.
222 3 Computer-Aided Engineering (CAE)

References

Acharya S (2016) Analysis and FEM simulation of flow of fluids in pipes. Arcada University of
Applied Science, Finland. Available online, https://www.theseus.fi/bitstream/handle/10024/106
991/Acharya_Saroj.pdf?sequence=1 (accessed on April 18, 2021)
Bahman AS (2018) Chapter 8 Computer-aided engineering simulations, in Wide Bandgap Power
Semiconductor Packaging: Materials, Components, and Reliability. Elsevier, ISSN 0922–3444,
pp 199–223
Bi ZM (2010) Design and simulation of dust extraction for composite drilling. Int J Adv Manuf
Technol 54(5):629–638
Bi ZM (2014) Kinetostatic modeling of Exechon parallel kinematic machine for stiffness analysis.
Int J Adv Manuf Technol 71:325–335
Bi ZM (2018) Finite element analysis applications: a systematic and practical approach, 1st edn.
Academic Press. ISBN-10 018099526
Bi Z, Luo C, Miao Z, Zhang B, Zhang CWJ (2020) Automatic robotic recharging systems–devel-
opment and challenges, Industrial Robot, Vol. ahead-of-print No. ahead-of-print. https://doi.org/
10.1108/IR-05-2020-0109
Bi ZM, Meruva K (2019) Modeling and prediction of fatigue life of robotic components in intelligent
manufacturing. J Intell Manuf 30(7):2575–2585
Chandrupatla TR, and Belegundu AD (2012) Introduction to finite elements in engineering. 4th
Edn. Pearson, ISBN-10: 0132162741
Espon (2020) Synthis T6 all-in-one SCARA robots. https://files.support.epson.com/docid/cpd5/cpd
55682.pdf.
Fuelmatics (2020) Fuelmatics systems. https://ibem-management.com/onewebmedia/FUELMA
TICS%20SYSTEMS%20(IBEM).pdf
Krahe C, Iberl M, Jacob A, Lanza G (2019) AI-based computer aided engineering for automated
production design–a first approach with a multi-view based classification. Procedia CIRP 86:104–
109
Suh NP (2005) Complexity: theory and applications, 1st edn. Oxford University Press, New York,
p 2005
Chapter 4
Computer-Aided Manufacturing (CAM)

Abstract In a manufacturing system, manufacturing processes are the operations

through which raw materials are transformed into final products. Manufacturing
processes use machines, tools, and labors to implement the transformation of mate-
rials. Design of manufacturing processes is generally complex since it involves in
machines, tools, materials, and numerous operating parameters. A manufacturing
process is evaluated comprehensively based on a number of conflict criteria such
as cost, accuracy, productivity, flexibility, and adaptability. Traditional intuitive
or experimental designs of manufacturing processes have their limitations in (1)
exploring a wide scope of alternative processes and (2) ensuring the first-time-right
for virtual design to physical implementation. In this chapter, computer-aided manu-
facturing (CAM) is introduced to model and evaluate manufacturing processes in
the virtual environment, the theory and enabling techniques of CAM are discussed,
and the applications of CAM are explored in designs of materials, geometric dimen-
sioning and tolerance (GD&T), designs and simulation of fixtures, molds and dies,
machining processes, and machining programming.

Keywords Manufacturing processes · Computer-aided manufacturing (CAM) ·

Modeling and simulation · Geometric dimensioning and tolerance (GD&T) ·
DimXpert · Toansist · Computer-aided fixture design (CAFD) · Design for
manufacturing (DFM) · Computer numerical controls (CNC) · Machining
programming

4.1 Introduction

Manufacturing processes are essential activities in the secondary industry. Manu-

facturing processes transform starting materials from the primary industry into the
finished products used as capital goods in the secondary or third industry or customers
products for end-users. Artificial goods in our modern society are mostly produced
from manufacturing processes.
Design of manufacturing processes follows product design, and manufacturing
processes determine the way how a virtual product model is transformed into
a physical product. Therefore, design of manufacturing processes covers many

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 223
Z. Bi, Practical Guide to Digital Manufacturing,
https://doi.org/10.1007/978-3-030-70304-2_4
224 4 Computer-Aided Manufacturing (CAM)

aspects including (1) the selections of materials, processing types, machines, and
tools; (2) geometric dimensioning and tolerancing (GD&T); (3) work-holding and
fixturing; (4) operating conditions; (5) machining programming; and (6) quality
controls. Computer-aided manufacturing (CAM) has been widely used to support
the decision-making activities in these aspects.

4.2 Material Characteristic, Structures, and Properties

As shown in Fig. 4.1a, the characteristic of a material consists of structures, prop-

erties, processes over the material, and the exhibited performance in applications.
Figure 4.1b uses the steel material as an example to illustrate the contents of its char-
acteristic: it has its metallic bonds in structure; it supports the processing types such
as costing, forging, and particulate processes. It has the properties of high elasticity,
density, yield strength, and hardness. It exhibits the performance of high strength,
ductility, and thermal and electric conductivities in applications.
Four elements in a material tetrahedron are dependent on each other. For
example, material structure has the great impact on material properties and applicable
processes. As shown in Fig. 4.2, different materials exhibit distinct properties, and
materials should be appropriately chosen for given applications. Steel materials have
high strengths and have affordable costs that have widely been applied in large-scale
construction (see Fig. 4.2a). Products in musculoskeletal healthcare are embedded
in human bodies; this requires the materials be corrosion resistant (see Fig. 4.2b).
Machine elements are mostly involved in the motions at contact surfaces; the parts
with oil impregnation from powder metallurgy (PM) will reduce wear and prolong
fatigue lives of products (see Fig. 4.2c). It will be helpful for engineers to under-
stand material structures so that right materials can be chosen to meet the functional
requirements for products.

Structure Metallic

Elasticity, density,
Characterization Steel yield strength,
Properties hardness …
characterization

Coating, forging,
particulate
Processing processing, etc. High strength,
ductile,
Performance conductive…

(a) Material tetrahedron (Sun 2009) (b) Steel tetrahedron

Fig. 4.1 Material tetrahedron and example

4.2 Material Characteristic, Structures, and Properties 225

(a) Steel metallic structure for (b) Corrosion resistant alloy for (c) Customized tools from
large construction medical products powder metallurgy
construction

Fig. 4.2 Example of selecting right material and processing for products

Component
Molecules
Electrons

Product

System
Crystal
Atoms

Part

0.1 nm
nm µm mm m km
(Angstrom)
Nuclear
Electrical
Thermal

Chemical

Mechanical
mechanism
Quantum

dynamics
Molecular

mechanics
Statistical

Micromechanics

mechanics
Damage

Element testing

analysis
Structural

engineering
System

Fig. 4.3 Disciplines for material characterizations at different scales (Kreculj and Rasuo 2018)

In a manufacturing process, certain change occurs to the material, and some

common changes of the materials include reshaped geometries, improved material
properties, and joined structures. It is clear that the material properties (1) determine
the types of manufacturing processes that can be applied to the given material and (2)
affect operating conditions when certain manufacturing process is applied. Taking an
example of shaping a part geometry by machining, since unwanted material must be
cut away from a workpiece, the material properties affect the selections of machines,
cutting tools, and operating conditions such as cutting speed, feed rate, and depth.
As shown in Fig. 4.3, material structures and properties are characterized in a
varying scale from the sub-atom level to the system level. Nuclear properties are
226 4 Computer-Aided Manufacturing (CAM)

determined by atomic structures, while the mechanical properties will be determined

by all of the characteristics and attributes cross over the scales. The study on material
structures and properties is interdisciplinary; the materials at different scales fall in
the different disciplines relevant to the material science.
At molecule level, numerous materials are formed by combing protons, neutrons,
and electrons, and the material properties at that level are direct results of (i) consti-
tutive structures and (ii) processing of property enhancement to change material
structures. Atoms are the basic structure of matter. An atom consists of a nucleus
with positive charges and a set of electrons with negative charges. A planetary model
by Niels Bohr can be used to represent an atomic structure. As shown in Fig. 4.4a,
electrons orbit around the nucleus at certain distances called as orbits or shells, and
the maximum number of electrons in the n-th shell is 2n2 . Note that a nucleus consists
of neutrons and protons, and the positive charges of the nucleus are associated with
protons. The similarities of the elements are determined by their atomic structures,
Fig. 4.4b–d shows the examples of nonmetal, metal, and noble gas, respectively.
Atoms are distinguished by atomic numbers and elements. Up to now, 118
elements have been discovered. As shown in Fig. 4.5, the discovered elements
have been classified into element families based on their similarities and relations.
Elements are organized in columns and rows. The elements in the same column
have their similarities. For example, the elements in the very right column including
helium, neon, argon, krypton, xenon, and radon are noble gases; these elements
exhibit the great chemical stability. The elements in the left and middle sections of
the table are metallic. The elements at the right section are nonmetal and the elements

Fig. 4.4 Description of atomic structure and examples

4.2 Material Characteristic, Structures, and Properties 227

Fig. 4.5 Periodic table of the elements (Science Notes 2020)

in the diagonal transition zone of metal and nonmetal elements are metalloids or
semimetals (Science Notes 2020).
Molecules are built from atoms by bonds. Therefore, the material strengths at
the molecular level are determined by the types of bonding relations. Bonds can be
classified into (1) primary bonds for the formulation of molecules and (2) secondary
bonds for the connections in a set of molecules. Valence electrons in the outermost
shell of atoms are involved in bonding.
The atoms in a primary bond exchange the valence electrons with a strong atom-
to-atom attraction. In an ionic bond shown in Fig. 4.6a, the electron(s) at the outer
shell of the atom are moved to the outer shell of the other atom to form the mutual
attraction. Material properties with ionic bonds have low electrical conductivity and
poor ductility. In a covalent bond shown in Fig. 4.6b, the electrons at the outer shells
of two atoms are shared to bond two atoms. The material properties with covalent
bonds have high harnesses and low electrical conductivity. In a metallic bond shown
in Fig. 4.6b, the valence electrons of metallic atoms are moved freely to form an
electron cloud that generates the strong attractions among the atoms. The material
properties with metallic bonding have high electrical and thermal conductivities,
high strength, and high melting points.
Secondary bonds generate the attraction forces of molecules. In contrast to primary
bonds, there is no electron transfer or sharing; therefore, the secondary bonds are
weaker than primary bonds. Secondary bonds can be classified into dipole force,
London force, and hydrogen bonding. As shown in Fig. 4.7a, the dipole force is
formed for the molecule with positive and negative poles. Therefore, the positive pole
of one molecule is attracted by the negative pole of the other molecule. As shown
228 4 Computer-Aided Manufacturing (CAM)

Fig. 4.6 Three primary bonding types

in Fig. 4.7b, even though molecule does not have a permanent pole, it generates
an instanton pole due to the high movement of electrons. A London force reflects
the attraction of molecules with an instanton pole. As shown in Fig. 4.7c, hydrogen
bonding refers to the intermolecular force of the molecules with hydrogen atoms.
Microstructures of materials can be crystalline or amorphous. In a crystalline
structure, molecules are arranged in three-dimensional lattices, and Fig. 4.8 shows a
comparison of three basic lattice structures, i.e., body-centered cubic (BCC), face-
centered cubic (FCCn), and hexagonal closed packed (HCP). In an amorphous struc-
ture, atoms are not arranged in a lattice structure. Note that microstructure of material
can be changed by certain manufacturing processes. For example, both crystalline
4.2 Material Characteristic, Structures, and Properties 229

Fig. 4.7 Three secondary bonding types

Fig. 4.8 Microstructures of materials

230 4 Computer-Aided Manufacturing (CAM)

and amorphous structures exist in glass, the higher level the crystalline is, the better
strength the glass can achieve, and the higher density the glass has.
The imperfections are usually the most critical locations where a fracture of the
material is mostly initialized. An imperfection involves in the discontinuity of mate-
rial to pass over the stress on the neighboring areas. Micro-level defects include (1)
point defects, (2) line defects, and (3) surface deflects that are illustrated in Fig. 4.9.
Point defects include ‘vacancy’ where an atom is missed, ‘interstitial’ where a rela-
tively small atom is filled in the gap of other atoms, and ‘substitutional’ where an
atom is replaced by another atom type. When the crystal structure of an alloy is
considered, two other types of point defects include a ‘Frenkel defect’ where the
small atoms are dislocated, and ‘ion-pair-vacancy’ where a pair of an ion and atom
has been missed. Line defects include ‘edge dislocation’ where the atoms along a
line of lattice are missed and ‘screw dislocation’ where the atoms along a spiral path

Fig. 4.9 Types of micro-level imperfections (Kailas 2020)

4.2 Material Characteristic, Structures, and Properties 231

Fig. 4.10 Classification of engineering materials

in the lattice are missed. Surface defects are mainly for the boundaries of grains;
when crystalline grains along different orientations are met, irregularities occur to
the boundaries.
In applications, engineering materials can be classified in different ways. For
example, Fig. 4.10 shows that materials can be classified into metallic, ceramic,
and polymeric based on the types of atomic bonding forces. In addition, different
materials can be structured to form composite materials. Moreover, the materials in
each category can be further grouped based on chemical compositions, mechanical
properties, and physical properties.
Metals are classified into (1) ferrous metals that contain iron as one of constitutive
elements and nonferrous metals that are free of iron element; a ferrous metal is
generally magnetic in nature. Many other metals fall into the group of nonferrous
metals, the examples of nonferrous metals are copper, aluminum, and lead.
Ceramic is composed of inorganic and nonmetallic elements through the processes
of heating and consequent solidification. Ceramic is brittle and with a high melting
point, high elastic module, and high strength. Ceramic materials are used to make
cutting tools such as drills, cutting chips, and grinding wheels. However, ceramic
products are also vulnerable to be broken under shock loads. Depending on the level
of crystalline, ceramics can be classified into crystalline ceramics and glasses.
Polymer is composed of recurring molecular structures as macromolecules. Poly-
mers are classified as thermoplastic, thermoset, and elastomer. Due to the difference
of molecular structures, polymers are highly diversified in the mechanical properties
such as strength, toughness, and hardness.
A composite material is formed by combing two or more distinct materials, and
each constitutive material retains its properties. Materials are combined to create a
new composite material with the properties that cannot be achieved by any of consti-
tutive components alone. Composite materials differ from conventional materials
since the material properties depend on hown different materials are composed in
manufacturing processes. Therefore, design and manufacture of composite materials
are discussed in the following section.
232 4 Computer-Aided Manufacturing (CAM)

4.3 Composite Materials

A composite material consists of two phase types, i.e., reinforcing phase and matrix
phase. The materials for the reinforcing phase are usually light and strong; reinforcing
phases are present in the form of particles, fibers, or sheets. The materials for the
matrix phase are tough and ductile since the matrix material needs to ensure the
deflection to pass external forces cross the matrix and protect reinforcing materials.
Figure 4.11 shows that composite materials are classified based on the types
of reinforcing and matrix materials. All conventional materials including metal,
polymer, and ceramic can be used as matrix materials. Aluminum, titanium, and
magnesium are widely used as the materials of metal matrix. Polymer matrix mate-
rials can be thermoset and thermoplastic; thermoset polymers are stronger and stiffer
than thermoplastic polymers; however, thermoplastic polymers have better ductility
and are reusable. The materials for ceramic matrix include Si3 N4 and Al2 O3 . A
ceramic material has very high specific modulus and mechanical strength at high
temperature.
In Fig. 4.11, reinforcing materials are classified in terms of the shapes and material
types. The shapes of the reinforcing material particles, flakes, whiskers, and long
fibers are shown in Fig. 4.12. Common materials used as the reinforcements in the
composites include carbon, glass, aramid, graphite, silicon carbide, and boron.

Fig. 4.11 The classification of composite materials (Bi et al. 2009)

4.3 Composite Materials 233

Fig. 4.12 Shapes of reinforcing materials

The composites with the reinforcing materials of Fig. 4.12a–c can be modeled as
isotropic materials that exhibit the similar properties in any direction. For isotropic
material, the generic Hooke’s law is used to describe the relation of stress and strain as
⎡ ⎤ ⎡ ⎤⎡ ⎤
σ11 1−ν ν ν 0 0 0 ε11
⎢ σ22 ⎥ ⎢ ν 1−ν ν 0 ⎥ ⎢ ε22 ⎥
⎢ ⎥ ⎢ 0 0 ⎥⎢ ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎢ σ33 ⎥ E ⎢ ν ν 1 − ν 0 0 0 ⎥⎢ ε33 ⎥
⎢ ⎥= ⎢ ⎥⎢ ⎥
⎢ σ23 ⎥ (1 + ν)(1 − 2ν) ⎢ 0 0 0 1 − 2ν 0 0 ⎥⎢ ε23 ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎣ σ13 ⎦ ⎣ 0 0 0 0 1 − 2ν 0 ⎦⎣ ε13 ⎦
σ12 0 0 0 0 0 1 − 2ν ε12
(4.1)

where [σ ] and [ε] are the vectors for stresses and strains, E is Young’s modules, and
ν is Poisson’s ratio.
A composite with the reinforcing materials of Fig. 4.12d is usually referred as
laminated composite, and plies are basic building elements of a laminated composite.
A ply is a thin orthotropic material which is characterized by four parameters, i.e.,
(1) longitudinal stiffness E 11 , (2) transverse stiffness E 22 , (3) shear stiffness G12 , and
(4) Poisson’s ration ν12 . Accordingly, the constitutive model of a ply becomes
⎡ ⎤ ⎡ ⎤⎡ ⎤
σ11 Q 11 Q 12 0 ε11
⎣ σ22 ⎦ = ⎣ Q 21 Q 22 0 ⎦⎣ ε22 ⎦ (4.2)
σ12 0 0 Q 66 ε12

where
 ⎫
2 ⎪
Q 11 = E 11 /1 − ν12 ⎪
2 ⎪
⎪
Q 22 = E 22 / 1 − ν12 ⎪
⎬
Q 12 = ν12 E 22 (4.3)
⎪
⎪
Q 21 = Q 12 ⎪
⎪
⎪
⎭
Q 66 = G 12

As shown in Fig. 4.13a, a laminated composite is built-up from plies, the constitu-
tive model of the laminated composite is developed based on that of plies. As shown
in Fig. 4.13b, the transformation of the constitutive model of a ply is determined by
the direction (θ ) of the ply coordinate system in the laminate coordinate system. Let
234 4 Computer-Aided Manufacturing (CAM)

Fig. 4.13 Plies in laminated composite

m = cosθ and n = sinθ, we have

⎫
Q 11 = Q 11 m 4 + 2(Q 12 + 2Q 66 )m 2 n 2 +Q 22 n 4 ⎪
⎪
⎪
⎪
Q 12 = (Q 11 + Q 22 − 4Q 66 )m n +Q 12 (m + n )
2 2 4 4 ⎪
⎪
⎪
⎬
Q 22 = Q 11 n + 2(Q 12 + 2Q 66 )m n +Q 22 m
4 2 2 4
3⎪ (4.4)
Q 16 = (Q 11 − Q 22 − 2Q 66 )m n+(Q 12 − Q 22 + 2Q 66 )mn ⎪
3
⎪
Q 26 = (Q 11 − Q 22 − 2Q 66 )mn 3 +(Q 12 − Q 22 + 2Q 66 )m 3 n ⎪
⎪
⎪
⎪
⎭
Q 66 = (Q 11 + Q 22 − 2Q 12 − 4Q 66 )m n +Q 66 (m + n )
2 2 4 4

Figure 4.14 shows the free-body diagram at a point of composite material; the
constitutive relation of stress and strain of laminated material can be found as
⎡ ⎤ ⎡ ⎤⎡ ⎤
σ11 C11 C12 C13 C14 C15 C16 ε11
⎢ σ ⎥ ⎢ C C C C C C ⎥⎢ ε ⎥
  ⎢ 22 ⎥ ⎢ 21 22 23 24 25 26 ⎥⎢ 22 ⎥   
⎢ ⎥ ⎢ ⎥⎢ ⎥
N ⎢ σ ⎥ ⎢ C C C C C C ⎥⎢ ε ⎥ A B ε
← ⎢ 33 ⎥ = ⎢ 31 32 33 34 35 36 ⎥⎢ 33 ⎥ →
M ⎢ σ23 ⎥ ⎢ C41 C42 C43 C44 C45 C46 ⎥⎢ ε23 ⎥ B D K
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎣ σ13 ⎦ ⎣ C51 C52 C53 C54 C55 C56 ⎦⎣ ε13 ⎦
σ12 C61 C62 C63 C64 C65 C66 ε12
(4.5)

where [C]⎡ is stiffness

⎤ matrix;
⎡ ⎤
σ11 σ23
N = ⎣ σ22 ⎦ and M = ⎣ σ13 ⎦ are the vectors of membrane loads and bending
σ33 σ12
loads, respectively;
⎡ ⎤ ⎡ ⎤
ε11 ε23
ε = ⎣ ε22 ⎦ and K = ⎣ ε13 ⎦ are the vectors of strains and curvatures,
ε33 ε12
respectively;
4.3 Composite Materials 235

Fig. 4.14 Free-body diagram of a material point in composite

⎡ ⎤ ⎡ ⎤ ⎡ ⎤
C11 C12 C13 C14 C15 C16 C41 C42 C43
A = ⎣ C21 C22 C23 ⎦, B = ⎣ C24 C25 C26 ⎦ = ⎣ C51 C52 C53 ⎦, and D =
C31 C32 C33 C34 C35 C36 C61 C62 C63
⎡ ⎤
C44 C45 C46
⎣ C54 C55 C56 ⎦.
C64 C65 C66
In Eq. (4.5), A, B, and D are extensional stiffness matrix, stretching-bending
coupling matrix, and flexural stiffness matrix that are assembled from the constitutive
models of plies (Eq. 4.4) as
  ⎫
Ai j = nk=1 Q i j k (h k − h k−1 ) ⎪
   ⎬
Bi j = 21 nk=1 Q i j k h 2k − h 2k−1 (4.6)
   ⎪
Di j = 13 nk=1 Q i j k h 3k − h 3k−1 ⎭

To use laminated materials safely, the materials should be analyzed at (1) micro-
level and (2) macro-level. Micro-level analysis is at the scale of fiber diameter; the
reinforcing and matrix materials are treated as isotropic and modeled, respectively.
It is mainly used to investigate the stress distribution at the interface of fiber and
matrix where a crack may likely be initiated.
236 4 Computer-Aided Manufacturing (CAM)

In general, micro-level analysis involves in the intensive computation. Therefore,

unit cells are often defined and analyzed to make the amount of computation manage-
able in simulation. Figure 4.15 shows some common models of unit cells in repre-
senting composites. Since both fiber and matrix materials are treated as isotropic
materials, respectively, conventional failure criteria (see Fig. 4.16) can be used in
micro-level analysis of composites.
Macro-level analysis takes into consideration the part geometry which is made of
composite materials; the plies in laminates are treaded as homogeneous anisotropic
materials. Consequentially, the strengths of the laminated composite are character-
fT fC fT fC f
ized as σ11 , σ11 , σ22 , σ22 , and τ 12 , and Table 4.1 shows three commonly used
failure criteria for composites in a macro-level analysis.

Example 4.1 In the spider lift shown in Fig. 4.17a, a few of the highlighted compo-
nents are made of composite materials. These components are adopted to avoid a
direct circuit flow from a power line to ground. The parts are prototyped and tested
in the real-world application environment. However, a failure occurred to one proto-
typed component as shown in Fig. 4.17b, and the user wanted to find out the cause
of failure in the product design (Bi and Mueller 2016). The CAD model of the failed
part is shown in Fig. 4.7c; establish the modeling procedure of FEA for the analysis
of composite parts.

Fig. 4.15 Examples of unit cell models for micro-level composite analysis (Tuttle 2020)
4.3 Composite Materials 237

Fig. 4.16 Failure criteria for isotropic materials (Pilkey et al. 2020)

Table 4.1 Commonly used failure criteria of composite materials

Failure criterion Formula Strengths and stresses
⎫ fT fC
Maximum stress failure fC
−σ11 < σ11 <
fT
σ11 ⎪
⎪ σ11 and σ11 are tensile and
criterion ⎬
fC fT compression strengths along
−σ22 < σ22 < σ22
⎪
⎪ longitude direction
|τ12 | < τ12
f ⎭ fT fC
σ22 and σ22 are tensile and
Tsai-Hill failure criterion (σ )2
+ (σ )2
+ (τ )2
− compression strengths along
 11 2  22 2  12 2
fT
σ11
fT
σ22
f
τ12 transverse direction
σ11 σ22 f
τ12 is the shear strength in the
 
fT 2
<1
σ11 laminate plane
Tsai-Wu failure criterion σ11 , σ22 , and τ12 are the longitude
X 1 σ11 + X 2 σ22 + X 11 σ11
2 +
stress, and transverse stress, and
X 22 σ22
2 + X τ2 +
66 12 shear stress, respectively
2X 12 σ11 σ22 < 1 X 1 , X 2 , X 11 , X 12 , X 22 , and X 66
are experimental determined
material strength parameters

Solution The Solidwork Simulation is used to analyze a composite part as follows.

Firstly, the properties of the composites are defined. Figure 4.18 shows the interface
to define new material properties. ‘Linear elastic orthotropic’ is selected as the model
type to specify the attributes of composite materials.
Secondly, the laminates of the composite material are configured. Figure 4.19a
shows the interface to define a surface which is made of composites. Figure 4.19b
238 4 Computer-Aided Manufacturing (CAM)

Fig. 4.17 Tie rod in a spider lift (Bi and Mueller 2016)

Fig. 4.18 Interface for definition of composite materials

4.3 Composite Materials 239

Fig. 4.19 Configure the plies of composite materials on surface

Fig. 4.20 Failure criteria of

composites

shows its main attributes, which include the number, the thicknesses, and the materials
of plies; the attributes also include the directions of plies in each layer.
240 4 Computer-Aided Manufacturing (CAM)

Thirdly, the failure criteria are specified for each solid body. As shown in Fig. 4.20,
the Solidworks Simulation supports all three options of failure criteria in Table 4.1 to
define the factor of safety for composite materials. In addition, the software supports
the automatic selection of failure criteria based on the assigned material properties.
Fourthly, the steps in FEA modeling are followed to (1) create part geometry,
(2) assign material properties, (3) create meshes, (4) define boundary conditions and
loads, (5) run the model, and (6) post-process the simulation results. Figure 4.21a–c
shows the outcome examples for the distributions of displacement, stress, and the
factor of safety in the first ply of the part surface.

Fig. 4.21 Example of simulation results over plies

4.4 Geometric Dimensioning and Tolerancing (GD&T) 241

4.4 Geometric Dimensioning and Tolerancing (GD&T)

Geometric dimensioning and tolerancing (GD&T) specify the desired size and accu-
racy of a feature on part; GD&T serves as the specifications for dimensional and
shape controls of parts and fits. GD&T is critical since that the selection and option
of a manufacturing process must meet the GD&T specifications for quality control
of product. Figure 4.22 shows the relations of common manufacturing processes and
typical tolerance ranges, and Fig. 4.23 shows the impact of the selection of manu-
facturing processes on the production times (Vorburger and Raja 1990), which are
directly related to manufacturing costs.
Proper application of GD&T ensures that (1) parts are interchangeable in mass
production; (2) right references are used to define the attributes of part features;
(3) avoid the ambiguity in understanding the functions of machined features; (4)
eliminate a tolerance stack-up for better part designs; and (5) clarify the criteria

Fig. 4.22 Accuracy of common manufacturing processes (Vorburger and Raja 1990)
242 4 Computer-Aided Manufacturing (CAM)

Fig. 4.23 Typical relations of surface finish to production time (Vorburger and Raja 1990)

for inspection and quality controls. GD&T communicates dimensional and toler-
ance requirements for the implementation of manufacturing and assembly processes
(Steeves 2016). GD&T practices are governed by standards, especially the standards
from ASME (Y14.5 M-1994 for 2D, Y14.41–2003 for 3D) and ISO (ISO 1101 for
2D and ISO 16,792 for 3D). While GD&T provides the necessary information for
the inspections in product quality control, a tight tolerance should only be applied to
some critical part features.

4.4.1 Datum Systems

Most of the GD&T specifications are defined based on the specified reference(s),
which are commonly known as the datums in GD&T. A datum is an entity such as
4.4 Geometric Dimensioning and Tolerancing (GD&T) 243

Fig. 4.24 The 3–2-1 rule and three orthogonal datum planes

a point, axis, line, plane, or coordinate system that is used as a reference when the
part is dimensioned, measured, manufactured, or inspected. To locate and orientate a
machined feature on part, a datum system with six degrees of freedom (DOF) must be
applied. The 3–2-1 rule uses the minimum number of contact points as the references
to locate and orientate the features on a part. As shown in Fig. 4.24, a datum system
based on the 3–2-1 rule consists of three datum planes, i.e., the primary plane, the
secondary plane, and the tertiary plane. When such a datum system is used to locate
and orient a part, there will be 1, 2, and 3 points on the part that make the contacts
to the tertiary plane, secondary plane, and the primary plane, respectively.

4.4.2 Geometric Tolerancing

The degrees of accuracy of geometric features such as lines, arcs, curves, planes,
surfaces, and cylinders are specified by geometric control symbols and associated
tolerances. As shown in Fig. 4.25a, the ASME standards have a total of 14 control
symbols, which are used to control the geometric tolerances in form (F), orientation
(O), profile (P), runout (R), and location (L), respectively. The main attributes of
each control symbol have been described in Fig. 4.25b. A control symbol has its
graphic icon, specifies the inspection method, indicates the datum when it is needed,
and provides the link to the corresponding section in ASME Y14.5 M-1994 for the
detailed explanation of the control symbol. Figure 4.26 gives a full table of GD&T
control symbol developed by Allsup (2009) that helps users to be familiar with all
control symbols and their relations. Figures 4.27, 4.28, 4.29, 4.30 and 4.31 show the
examples of (1) how to use control symbols and (2) how to measure the attributes of
features to validate the required degrees of accuracy.
244 4 Computer-Aided Manufacturing (CAM)

Fig. 4.25 14 control symbols and the template for their attributes

Fig. 4.26 Table of geometric control symbols by Allsup (2009)

4.4.3 Basic Concepts of Dimensioning

Dimensions define the nominal geometries and allowable variations for the features of
parts and assemblies. A dimension should have a tolerance since every manufactured
feature is subject to variation. The limits of an allowable variation must be specified
to control the product quality. An engineering drawing should be dimensioned to
reflect the functions of the features of parts, and the interpretation of a dimension
must be unambitious.
A basic size is a nominal size which is used as a reference of ideal size from
which the limits are calculated. A basic size is common to two fitting parts at their
interface. A maximum material condition (MMC) is a dimension of a part feature
4.4 Geometric Dimensioning and Tolerancing (GD&T) 245

Fig. 4.27 Form control symbols and measurements

Fig. 4.28 Orientation control symbols and measurements

Fig. 4.29 Profile control symbols and measurements

246 4 Computer-Aided Manufacturing (CAM)

Fig. 4.30 Runout control symbols and measurements

Fig. 4.31 Location control symbols and measurements

when it is composed of the most of material, and a least material condition (LMC)
is a dimension of a part feature when it is composed of the least material. Limits are
the extreme values of a dimension; tolerance is a difference of the lower and upper
limits. A tolerance can be specified by (1) the limits from the basic size or (2) the
general notes in block or other places in engineering drawing.
When a part has multiple dimensions in one direction, either chain dimensioning
or baseline dimensioning can be applied. As shown in Fig. 4.32, chain dimensioning
defines multiple dimensions in a sequence in which the next dimension is placed
directly adjacent to the previous one, while baseline dimensioning defines multiple
dimensions concurrently based on the same reference.
Tolerance stacking refers to the combination of the tolerances of multiple dimen-
sions. As shown in Fig. 4.32a, the tolerance of an indirect dimension of two entities
4.4 Geometric Dimensioning and Tolerancing (GD&T) 247

Fig. 4.32 The schematics of chain and baseline dimensioning

(A1 A5 ) is a stacked-up from those of the tolerances of the dimensions (A1 A2 , A2 A3 ,

A3 A4 , A4 A5 ) from where the indirect dimension is formed, i.e.,
⎫
N or minal o f A1 A5 = A1 A2 + A2 A3 + A3 A4 + A4 A5 = 3.75 ⎪
⎬
U pper limits o f A1 A2 + A2 A3 + A3 A4 + A4 A5 +0.04 (4.7)
T olerance o f A1 A5 = = ⎪
⎭
Lower limits o f A1 A2 + A2 A3 + A3 A4 + A4 A5 −0.04

In the schematic of the baseline dimensioning in Fig. 4.32b, the common refer-
ence (B1 ) is used to dimension the entities (B2 , B3 , B4 , B5 ) along the Y-axis, and
no tolerance stack-up occurs since each entity is dimensioned from the common
reference.

Example 4.2 Determine the dimensioning schematic to obtain a better accuracy of

the dimension A2 A5 in Fig. 4.33.

Fig. 4.33 The schematics of chain, baseline, and direct dimensioning

248 4 Computer-Aided Manufacturing (CAM)

Solution For the baseline dimensioning in Fig. 4.33a, the dimension of A2 A5 is

indirectly derived as
⎫
N or minal o f A2 A5 = A1 A5 − A1 A2 = 3.50 ⎪
⎬
U pper limit o f A1 A5 − Lower limit o f A1 A2 +0.02 (4.8)
T olerance o f A2 A5 = = ⎪
⎭
Lower limit o f A1 A5 − U pper limit o f A1 A2 −0.02

For the chain dimensioning in Fig. 4.33b, the dimension of A2 A5 is indirectly

derived as
⎫
N or minal o f A2 A5 = A2 A3 + A3 A4 + A4 A5 = 3.50 ⎬
U pper limits o f A2 A3 + A3 A4 + A4 A5 +0.03
T olerance o f A2 A5 = = ⎭
Lower limits o f A2 A3 + A3 A4 + A4 A5 −0.03
(4.9)

For the direct dimensioning in Fig. 4.33c, the dimension of A2 A5 is directly

obtained as A2 A5 : 3.50+0.01
−0.01 . By comparing the tolerances of A2 A5 in three
schematics, direct dimensioning yields the most accuracy for the dimension A2 A5 .

4.4.4 Engineering Fits

When parts are assembled into products, engineering fits are used to represent the
relations of matting parts. The amount of an engineering fit determines whether or
not two parts can move or rotate with each other independently. An engineering fit
is generally referred as a pair of shaft and hole. As shown in Table 4.2, engineering
fits can be classified into four types, i.e., clearance fits, interference fits, transition
fits, and line fits.
Engineering fits are classified based on the maximum and minimum clearances.
The maximum clearance refers to the maximum amount of space that occurs at the
interface of two fitting parts, and the minimum clearance is the minimum amount of
space that occurs at the interface of two fitting parts. The minimum clearance is also
called as allowance. The clearances of a fit are evaluated based on the shaft and hole
dimensions under MMC and LMC conditions as

Table 4.2 Types of engineering fits and conditions

Fit type Description Condition
Clearance fit There is always space at the interface of two parts Min. clearance > 0
Interference fit There never be a space at the interface Max. clearance ≤ 0
Transition fit A shaft and hole may be either of interfered or shaped Max. clearance > 0
Min. clearance < 0
Line fit There is a space or a contact at the interface Max. clearance > 0
Min. clearance = 0
4.4 Geometric Dimensioning and Tolerancing (GD&T) 249

Max. Clearance = L MC hole − L MCsha f t
(4.10)
Min. Clearance = M MC hole − M MCsha f t

where M MC hole and L MC hole are the hole sizes and M MC sha f t and L MC sha f t are
the shaft sizes under MMC and LMC, respectively.
The third column in Table 4.2 gives the conditions of the minimum and maximum
clearances for different fit types.

Example 4.3 Determine the fit types of shafts and holes in Table 4.3.

Solution Equation (4.10) is used to evaluate the maximum and minimum clearances,
and the clearances are then used determine the fit types, and the results are shown in
Table 4.4 (Fig. 4.34).
In ASME standards, running or sliding (RC) fits are classified into 9 levels from
1 for the tightest to 9 for loosest. The locational fits are classified into (1) locational
clearance (LC) fits with 11 levels, (2) locational transition (LT) fits with 6 levels,
and (3) locational interference (LN) fits in 5 levels. The corresponding tolerances
are standardized. Table 4.5 gives the excerpted tolerances for the fits of RC5 to RC9
and a basic size range of (0.0 in, 19.69 in). In the table, the first column is the range
of basic size, the other columns are for standard limits for shaft and hole fits.

Example 4.4 Use Table 4.5 to determine the tolerances of hole and shaft fit with
RC8 for a basic size of 0.50.

Table 4.3 The shaft and hole dimensions (Example 4.3)

Case Shaft Hole
1 1.50000.0
−0.005 1.50000.0
−0.005
2 0.850+0.007
+0.002 0.8500.0
−0.005
3 0.575+0.003
−0.002 0.5750.0
−0.004
4 1.2500.0
−0.003 1.250+0.005
−0.0

Table 4.4 Determining fit types by maximum and minimum clearances

Case LMCshaft MMCshaft LMChole MMChole Max Min Fit type
clearance clearance
1 1.495 1.500 1.506 1.501 0.001 0.011 Clearance
2 0.852 0.857 0.850 0.845 −0.002 −0.012 Interference
3 0.573 0.578 0.575 0.571 0.002 −0.007 Transition
4 1.247 1.250 1.255 1.250 0.005 0 Line
Note that an engineering fit is for the relation of two mating parts; either a hole-basis system or
a shaft-basis system can be used to specify the fit. As shown in Fig. 4.44, in a hole-basis system,
the hole size remains constant to determine shaft sizes based on the specified fits. In a shaft-basis
system, the shaft size remains constant to determine hole sizes based on the specified fits
250 4 Computer-Aided Manufacturing (CAM)

Fig. 4.34 Hole-basis system and shaft-basis system

Solution Firstly, the basic size is used to locate the row where the size falls in the given
range: 0.5(0.40, 0.71). Secondly, the fit RC8 is used to locate the column. Thirdly,
the tolerances at the intersections of specified column and row are interpolated as
the dimensions of hole and shaft as
−0.0035
H ole : 0.0028
0.50000.0000 Sha f t : 0.5000−0.0051

Note the limits in Table 4.5 has a unit of thousandth of an inch.

In the ISO system, an alphanumeric code called the international tolerance grade
number (IT#) is used to specify the tolerance ranges of fits, an uppercase represents

Table 4.5 Tolerance of RC5 to RC9 fits

Fig. 4.35 Example of fit

specifications in ISO system
4.4 Geometric Dimensioning and Tolerancing (GD&T) 251

the hole tolerance, and lowercase represents the shaft. Figure 4.35 shows an example
of the fit specifications in the ISO system.

4.4.5 Computer-Aided GD&T (DimXpert)

A complete part or assembly model must include the GD&T information for all of the
features and relations, while manually adding GD&T annotations in a CAD model is
trivial and error-prone. Solidworks DimXpert is an automated GD&T tool, which is
used to add both of geometric tolerances and dimensional tolerances in CAD models
(see Fig. 4.36). The dimensions and tolerances by DimXpert are compatible to ASME
Y14.5, Y14.41, and equivalent ISO standards. DimXpert supports engineers in using
dimensions and constraints to express the design intents and define the specifications
for manufacturing processes and inspections of parts.
DimXpert supports the full annotations in 3D models which are viewable as
edrewing. It helps to avoid manual errors and maintains the consistency of GD&T
settings. The TolAnalyst tool can be used to identify over- or under-toleranced parts
graphically, and it can analyze tolerance stacks automatically. Especially, DimXpert
is an ideal tool to create GD&T annotations for the machined parts such as those
with turned, drilled, and milled features. DimXpert has been widely used to annotate
sheet metals, castings, and machined elements such as gears, threads, and cams.
DimXpert can be used to automatically identify machined features. As shown
in Fig. 4.27, the identified features will be listed in the Features Tree under the
DimXpertManager. Figure 4.28 shows that identifiable machined features include
hole, cone, countersink, counterbore, plane, slot, notch, boss, cylinder, width, pocket,
surface, chamfer, and fillet (Figs. 4.37 and 4.38).
The ShowToleranceStatus tool in DimXpert is capable of detecting over-
dimensioned or under-dimensioned feature. As shown in Fig. 4.39, running this
tool in a CAD model colors the features in such a way that the features in a red

Fig. 4.36 DimXpert for GD&T (Steeves 2016)

252 4 Computer-Aided Manufacturing (CAM)

Fig. 4.37 Manufacturing features identified by DimXpert

Fig. 4.38 Types of the supported manufacturing features in DimXpert

color are over-dimensioned, the features in a yellow color are under-dimensioned,

the features in a green color are well defined, and the features in other colors are not
detectable by DimXpert.
In creating GD&T annotations for CAD models, the settings of DimXpert can be
customized by setting the DocumentProperties in Solidworks as shown in Fig. 4.40a.
One has the options of (1) GD&T standards and (2) the methods to specify the toler-
ances. In addition, the detailed settings for the methods and styles of Size Dimension,
4.4 Geometric Dimensioning and Tolerancing (GD&T) 253

Fig. 4.39 Using the ShowToleranceStatus tool to detect under- and over-dimensioned features

Location Dimension, Chain Dimension, Geometric Tolerance, Chamfer Controls, and

Displace Options can be made through the interfaces as shown in Fig. 4.40 b–g.

Example 4.5 Figure 4.41a shows the dimensions of a machined part. Use the AutoDi-
mensionScheme tool tin DimXpert o annotate GD&T information automatically.

Solution Firstly, the part is modeled based on the dimensions specified in Fig. 4.41a.
Secondly, the AutoDimensionScheme tool is used to specify (1) the part type; (2)
tolerance type; (3) the pattern dimensioning; and most importantly, the primary,
secondary, and tertiary references (see Fig. 4.41b). Thirdly, the AutoDimension-
Scheme tool is executed to create G&D annotations automatically.

4.5 Manufacturing Processes

Manufacturing processes aim to transform raw materials into finish products. In a

manufacturing system, each manufacturing process is a value-added step in the whole
transformation from starting materials to final products. For discrete manufacturing
in which discrete parts are made, a manufacturing process is to change the geometry,
properties, and appearance of materials. Figure 4.42 shows the input and output of
a manufacturing process. The input is a workpiece with the materials at the starting
state, the manufacturing resources such as machinery, tooling, and labor are applied to
transform the workpiece into new state. In addition, a manufacturing process involves
some side effects such as wastes. A manufacturing process is also an economic
activity since it adds the value to workpiece.
As shown in Fig. 4.43, manufacturing processes can be classified in terms of the
transformation types (Groover 2012). At the highest level, manufacturing processes
254 4 Computer-Aided Manufacturing (CAM)

Fig. 4.40 Customize DimXpert settings

are classified into processing operations, assembly operations, and other non-value-
added processes. A processing operation applies to individual parts, an assembly
operation applies to two of a group of parts or components, and a non-value-added
process such as inspection or transportation may apply to parts or components.
The manufacturing processes in the category of processing operations are further
classified into shaping processes, surface treatments, and property enhancement.
A shaping process is to create geometric features on parts, and geometric features
can be created by solidification, deforming processes, particulate processes, additive
manufacturing, and material removal processes.
Depending on the amount of geometric changes over parts, shaping processes can
also be classified into primary shaping processes and secondary shaping processes.
4.5 Manufacturing Processes 255

Fig. 4.41 Example of auto dimension scheme

Fig. 4.42 Description of manufacturing process

A primary shaping process is to create a rough geometric shape of part; common

primary shaping processes are casting, powder metallurgy, injection modeling, metal
sheet forming, and forging. A secondary shaping process is to refine geometric dimen-
sions of parts; common secondary shaping processes are turning, drilling, milling,
threading, boring, shaping, planning, sawing, broaching, hobbing, and grinding.
In addition, materials can be added or removed by some unconventional ways.
Therefore, shaping processes include other types of machining processes such as
electrochemical machining (ECM) and laser beam machining (LBM).
256 4 Computer-Aided Manufacturing (CAM)

Fig. 4.43 Manufacturing processes—classification

Surface treatment aims to change the material properties over the surfaces of
parts. Surface properties can be enhanced by cleaning, coating, and deposition.
Some common surface treatments are polishing, cleaning, tumbling, sanding, elec-
troplating, deburring, buffing, painting, anodizing, coating, sand blasting, galva-
nizing, anodizing, honing, and lapping. Property enhancement is to enhance the
material properties in the bodies of parts. Material properties can be changed by the
heat treatments since temperature has the great impact on material structures and
characteristics. Some common heat treatments are hardening, annealing, tempering,
normalizing, grain refining, and shot peening.
An assembly operation is to join two or more parts or components as new compo-
nent. If an assembly is at an intermediate state of a product, it is called a subassembly
or component. Parts can be joined either permanently or non-permanently. There-
fore, assembly processes are classified into two types—the processes for nonperma-
nent and permanent joints, respectively. Nonpermanent joints are usually mechanical
joints which are formed by some machine elements such as screws and bolts, snap
4.5 Manufacturing Processes 257

fits, and shrink fits. Permanent joints can be formed by chemical bonding or mechan-
ical means such as welding, brazing soldering, riveting, press fitting, sintering, shrink
filling, stitching, and stapling.
Other manufacturing processes do not change workpieces; therefore, they do not
add values to products. The examples of non-value-added processes are inspection,
testing, prototyping, transportation, and material handling.

4.5.1 Shaping Processes

A shaping process is to generate desired geometric shape or features on parts. Part

geometry can be generated in various ways. In a shaping process, it is desirable that
the raw material has a lower strength and higher formability. Since the strength and
formability relate closely to the temperature, the operating temperature is usually
the most critical factor in a shaping process: the higher the operating temperature is,
the lower strength and higher formability the material is, and the better the material
can be shaped. Therefore, shaping processes are classified based on the temperature
range when a shaping process is performed.
Figure 4.44 shows the classification of shaping processes based on the operating
temperature in comparison with the melting temperature T m of the raw material.
Cold working, warm working, hot working, sintering, and casting are performed
at the temperature ranges of below 0.3·T m , (0.3·T m , 0.5·T m ), (0.5·T m , 0.75·T m ),
(0.75·T m , T m ), and above T m, respectively. The operating temperature is used to
distinguish one type of the shaping process from another. For example, when an
operating temperature exceeds 0.5·T m , the material can be shaped by recrystallization
or solidification.
Due to the differences of operating temperatures and amount of geometric
changes, the raw materials have different starting conditions for various shaping

Fig. 4.44 Classification of shaping processes based on operating temperature

258 4 Computer-Aided Manufacturing (CAM)

Fig. 4.45 Raw materials in different shaping processes

processes. As shown in Fig. 4.45, the starting material in a casting process is in the
liquid state and the material does not hold the desired shape until it has been solidi-
fied in a cavity. The starting material in a particulate process is in the powder state,
and the starting material is pressed in a mold and heated until that sintering occurs
to form the crystallized geometry. The starting material in an injection molding is
small particles and the particles are heated and pressed into a mold and then solid-
ified as a finished part. The starting material in a deforming process is in a bulk or
sheet form from primary shaping processes. The material is deformed mechanically
with/without thermal loads. The starting material in additive manufacturing can be in
the state of powder, strip, or fluid state. The material is heated, solidified, or sintered
layer by layer to create part geometry.

4.5.2 Design and Planning of Manufacturing Process

Design of manufacturing processes is to select manufacturing methods and design

their implementations. Manufacturing processes should be designed to make prod-
ucts economical and competitive. Planning of manufacturing processes is to trans-
form the specifications and features of products into the instructions of machine
operations. Typical tasks in planning a manufacturing process include selection
of machines and tools, programming of machines, and determination of process
parameters.
Design of manufacturing processes is critical and greatly affected by other design
activities in the product lifecycle. Product development is usually an iterative process
4.5 Manufacturing Processes 259

Fig. 4.46 Computer-aided design in reducing manufacturing cost

where continuous changes are made to satisfy design constraints at different design
stages. As shown in Fig. 4.46, the requirements of manufacturing processes should
be considered at the earliest possible stage to reduce cost in changing the solutions
relevant to manufacturing. Similar to CAD tools, computer-aided techniques help to
identify manufacturing defects at the stage of product design.

4.6 Simulation of Manufacturing Processes—Mold Filling

Analysis

In designing a manufacturing process, it is challenging to avoid defects and opti-

mize process parameters due to many sources of uncertainties. Computer simula-
tion helps designers to understand the impact of process parameters, and optimize
manufacturing processes in the digital environment. In this section, some basic anal-
yses of injection modeling processes are discussed to illustrate how computer-aided
manufacturing (CAM) can be deployed for design and simulation of manufacturing
processes.

4.6.1 Injection Molding and Machines

Injection molding is one of the commonly used methods to make net-shape parts.
In injection molding, polymer is heated to form a plastic flow and then fill in a
mold cavity under pressure. The material in the cavity is solidified as molded part(s).
Injection molding is widely used to make parts with complex shapes. To improve
productivity, one mold may contain multiple cavities, so multiple moldings can be
made in a production cycle.
260 4 Computer-Aided Manufacturing (CAM)

Figure 4.47 shows the basic components of an injection molding machine. Two
main components are an injection unit and a clamping unit. The injection unit has a
barrel to keep, heat, and feed the material and a screw to mix, pressurize, and inject
the material into the mold. The clamping unit is to hold the mold when the mold
is filled and the part is solidified; the clamping unit is released when the part(s) are
ejected from the mold.
As shown in Fig. 4.48, an injection modeling cycle involves the following steps.
Firstly, the clamping unit is set to close the mold. Secondly, the raw material is fed,
heated, and injected into the mold. Thirdly, the melt in cavity is solidified when the

Fig. 4.47 Description of injection molding machines

Fig. 4.48 Main steps of injection molding process

4.6 Simulation of Manufacturing Processes—Mold Filling Analysis 261

mold is filled. Fourthly, the clamping unit is released to open the mold, and the part
is ejected from the cavity. The ejected part may need the post-processing to remove
unwanted materials such as solidified materials in runners from the part. Note that a
runner is a path by which the melt is filled into the enclosed cavity.

4.6.2 Moldability and Design Factors

Moldability measures how well the part geometry may conform to the mold when
the molten plastic is injected to the mold. An injection molding process should
be designed to avoid any defects when the material is filled into the mold cavity.
Figure 4.49 shows the impact of operating temperature and pressure on the mold-
ability. These two variables are most critical since they determine the feasible working
range (area 1) of an injection modeling process. Too low or high pressure will cause
the defects of short shot (area 2) and flash (area 3), respectively. Too low or high
temperature will cause the defects of melting (area 4) and thermal degradation (area
5), respectively.
The cycle time increases when the mold temperature increases, and this takes a
longer time to crystallize the melt in the mold. Therefore, mold temperature affects
the cooling rate. A high cooling rate helps in reducing the crystallization time and
reducing shrinkage of the molded parts. However, if too less time is used in cooling
the part, it is prone to shrink or causes warpage after the molding process.
A high pressure helps to fill the mold quickly to avoid premature freezing of the
flow of melt. Note that the clamping mechanism is used to hold the model when a
high pressure is applied, and the clamping force is determined by the pressure and
the effective area of cross section of the melt flow.

Fig. 4.49 Temperature and

pressure for moldability of
injection molding processes
262 4 Computer-Aided Manufacturing (CAM)

4.6.3 Design of Injection Molding Systems

Figure 4.50 shows the procedure of designing an injection molding system; it consists
of five major tasks as given below.
(1). Firstly, parts are modeled and analyzed to clarify the requirements of injection
modeling; more specifically, the parameters such as one-shot weight, required
cores, number of cavities, injection locations, and clamping force and stroke
should be determined.
(2). Secondly, the mold cavities are designed to (i) select molding types such as
hot runners, cold runners, and conventional sprues; (ii) select the materials
for mold bases, inserts, plates, and cores; and (iii) tailor the mold cavities to
geometric shapes of parts.
(3). Thirdly, the mold cavities are optimally placed, common placement options are
star, symmetrical, and in-line arrangement. The mold assembly is modeled,
simulated, and verified.
(4). Fourthly, other components for injecting and cooling melts and ejecting parts
are designed. Especially, the gating system is designed; common layouts of
a gating system are conventional, pinpoint, sub-marine, flash, tab, disk, and
diaphragm. The ejection component is designed; some common options are

Fig. 4.50 Design of injection molding process

4.6 Simulation of Manufacturing Processes—Mold Filling Analysis 263

pins, stripper rings, stripper plates, slides, and side cores). The type of venting
parting lines is specified.
(5). Fifthly, the injection molding system is reviewed as a whole against the
specified functional requirements; this ensures that all design constraints are
satisfied before the injection molding system is fabricated.

4.6.4 Design Variables and Considerations

The discussion in Sect. 4.5.2 shows that two main factors to affect the moldability are
operating temperature and pressure. However, injection modeling involves numerous
other design variables, which might affect the success of the injection modeling
process.
(1) Shrinkage

The material in an injection modeling process involves in the change of temperature

in a wide range. The thermal expansion coefficient of the material varies with the
temperature change. The higher the temperature is, the higher the thermal expansion
coefficient is. Accordingly, the dimensions of part in the mold cavity will be shrunk
in crystallization, and such shrinkage has to be compensated in the design process.
The amount of shrinkage depends on (1) the type of the material and (2) the range
of temperature changes.
Table 4.6 shows the percentage of the shrinkage for four commonly used materials.
To obtain an expected dimension, the corresponding dimension of a mold cavity is
compensated as

Dc = D p 1 + S + S 2 (4.11)

where Dc and Dp are the corresponding dimensions on the mold and the part, respec-
tively, and S is the percentage of the shrinkage given in Table 4.6. To yield better
dimensional accuracy, the following approaches can be adopted:
(1) Increasing the operating pressure to accelerate the injection of the melt into
the mold cavity.
(2) Increasing the operating temperature to reduce the viscosity of the polymer
melt and increase the density of packing in the mold.

Table 4.6 Shrinkage of four commonly used plastics

264 4 Computer-Aided Manufacturing (CAM)

(3) Increasing the compaction time and supplying the melt in the cooling process.

(2) Economic factor

The cost for molds and tooling is a significant part of the cost to manufacturing
processes, since the parts with different geometries or features require different molds
and tooling, which are fabricated in a dimensional tolerance more strict than the parts
(Bi et al. 2001). In designing an injection molding process, various process types
are available, and these processes have to be analyzed and compared from the cost
aspects. A high tooling cost is only allowed when the volume of certain parts is high.
Overall, injection molding is economically sound when the volume of a product
is over 10,000. When the product volume is lower than 1000, a compression molding
process should be used; when the product volume is in the range of (1000, 10,000),
a transfer molding process should be taken into consideration.
(3) Design features of molded parts

The more complex a part is, the high cost the mold and tooling are. From this
perspective, adopting injection modeling is beneficial since multiple functional parts
with same materials can be made as an assembly when it becomes feasible.
Design for manufacturing (DfM) can be applied in designing a molded part to
minimize injection defects such as molded-in stress, flashes, sink marks, and surface
blemishes. This is achieved by satisfying the requirements of an injection molding
process for flow lengths, weld line locations, injection pressures, clamping require-
ments, scrap rate, easiness of part assembly, and the needs of secondary operations
such as degating, painting, and drilling. In designing a molded part, the following
features must be paid a special attention:
(1). Wall thickness: A thick wall consumes more materials and tends to cause
warping since the shrinkage along the depth is uneven; it takes a long time
to harden the materials. It will be desirable to have thin walls with evenly
distributed thickness.
(2). Reinforcing ribs: To enhance the stiffness and strength at certain cross sections,
adding reinforcing ribs is a preferable option. However, the thickness of a
reinforcing rib must be less than the wall to reduce sink marks on walls.
(3). Corner radii and fillets: A sharp external or internal corner causes the stress
concentration in the application, and it affects the smoothness of melt flow
in injection molding. Sharper corners must be avoided to eliminate surface
defects on mold parts.
(4). Holes: It is mostly feasible to include holes in a molded part, while the impact
of having holes on the complexity of the mold and part ejection must be taken
into consideration.
(5). Drafts: To eject parts from the mold, the part must have the drafts on
lateral surfaces. The drafts recommended for thermosetting and thermoplastic
materials are ~1/2° to 1° and ~1/8° to 1/2°, respectively.
4.6 Simulation of Manufacturing Processes—Mold Filling Analysis 265

(6). Tolerances: Shrinkage is predictable under closely controlled conditions;

however, loose dimensional tolerances are preferable since that the uncer-
tainties of process parameters and variations of part geometries affect the
shrinkage and distribution over surfaces.

4.6.5 Defects of Molded Parts

Similar to other manufacturing processes, injection molding involves many design

factors. It is very challenging to develop a realistic mathematic model for process
optimization. Therefore, DfM is an iterative process to make the trade-offs among
some mutually conflicted design criteria, and the basic requirement is to minimize the
defects of parts, which might be caused by different reasons such as the design flaws
of parts, molds, and tools, not well-prepared materials, and inappropriate operating
conditions. Injection modeling leads to the following common defects.
(1). Weldlines

Weldline is an ill-filled line when two or more melt flows meet and remain undis-
solved. The weldline is a boundary of two melt flows. Figure 4.51 shows the process in
which a weldline is formed by two flows. To alleviate weldline defects, the following
methods can be used: (1) increasing the injection flow rate and increasing the temper-
ature of the mold; (2) reducing the temperature of the melt and increasing the injec-
tion pressure; and (3) optimizing the gate positions and the melt flows with less or
insignificant merging locations.
(2). Flashes

Flashes are formed at the parting line/surface of the mold or at some locations where
the ejectors are installed. Figure 4.52 shows the case where the flashes are developed
at the parting plane. A flash is a phenomenon where the molten polymer smears out
or sticks to the seam or gap between two surfaces. The common causes of flashes are

(a). melt fronts (b). weld line (c). weld line (d). weld line
approach initialized gradually developed completely formed

Cavity Weld
line

Filled Cavity

Melt Melt Melt Melt

Fig. 4.51 Process of forming a weldline by two flows

266 4 Computer-Aided Manufacturing (CAM)

Sprue
Mold 1 Mold 2

Gate Runner

Parting plane

(a). without flash on parting plane

Sprue
Mold 1 Mold 2

Gate Runner

Flash

Parting plane

(b). with flashes on parting plane

Fig. 4.52 Flashes in injection molding

(1) poor conformance at two contact surfaces; (2) low viscosity of molten polymer;
(3) high injection pressure; and (4) low clamping force.
The following methods can be used to reduce flashes (1) avoiding excessive thick-
ness difference in part design; (2) reducing injection flow rate; (3) applying well-
balanced pressure and clamping force to the mold assembly; (4) increasing clamping
force; (5) improving the surface quality of the parting plane, ejector pins, and holes.
(3). Short Shots

Short shots are voids that are not fully filled by molten materials. Figure 4.53 shows
the difference of parts with and without short shots. A short shot may occur at a
location with a long flow distance to the gate. Common causes of short shots are
(1) insufficient one-shot weight, (2) a low injection pressure, and (3) a low injection
speed; this leads to the situation that the melt flow freezes before the mold is fully
filled. Short shots can be improved by (1) increasing the injection pressure; (2) using
air vents or degassing device for a high flow pressure drop; and (3) optimizing the
mold design to smooth the melt flow.
4.6 Simulation of Manufacturing Processes—Mold Filling Analysis 267

Sprue
Mold 1 Mold 2

Gate Runner

Parting plane

(a). without short shorts

Sprue
Mold 1 Mold 2

Gate Runner Short

shots

Parting plane

(b). with short shots in cavity

Fig. 4.53 Part with short shots

(4). Warpages

A warpage is formed when the part is removed from the mold. Figure 4.54 shows a part
with warpages, which might be caused by (1) uneven shrinkage due to the thickness
difference or temperature derivation and (2) low injection pressure and insufficient
packing. To reduce warpages, the following methods can be used: (1) extending the
cooling time before the part is ejected, (2) optimizing the positions of ejection pins,

Fig. 4.54 Part with

warpages

Part Part
designed
molded

(a). without warpage (b). with warpages

268 4 Computer-Aided Manufacturing (CAM)

Fig. 4.55 Part with sink

marks
Part
designed Part
molded

(a). without sink mark (b). with sink marks

(3) enlarging draft angles, (4) balancing cooling lines, and (5) increasing packing
pressure.
(5). Sink marks

Sink marks may be generated when the melt in some areas (such as ribs or bosses)
shrinks more than adjacent areas. Figure 4.55 shows the part with sink marks, which
are caused by insufficient melts in the area before the flow is closed due to the
solidification. The methods to reduce sink marks include to (1) increase packing
pressure for more melt filling into the cavity, (2) keep the gates open for a longer
time, and (3) reduce the temperature of the mold to strengthen the surface.

4.6.6 Mold Flow Analysis

Design of injection molding processes used to rely on designers’ experiences with

the aid of the trial-and-error methods. Nowadays, computer simulation tools provide
the scientific methods to support the designs of injection molding processes. Mold
flow analysis aims to simulate an injection modeling process and optimize the design
of molds and process variables.
Figure 4.56 shows the inputs and outputs of Solidworks Plastics for mold flow
analysis (Eastman 2019). The system inputs include part model, material properties,
thermal properties, melt and mold temperatures, runners, gating systems, and cooling
systems. The system outputs include the simulated results for flows and fills, loca-
tions of weldlines, pressure to fill, pressure patterns, clamping force, temperature
patterns, shear patterns, filling, temperature distribution, shear thinning, freezing
and reheating, and temporary stoppages of flow. Mold flow analysis can be used for
the simulation-based optimization of injection molding processes.
As shown in Fig. 4.57, SolidWorks Plastics is integrated in the 3D solid modeling
environment, and the part model can be directly utilized for mold flow analysis. A
user-friendly wizard is provided to define and run a mold flow analysis step by step.
Figure 4.57a shows the access to the SolidWorks Plastics, which is an add-ins of
the SolidWorks. Figure 4.57b shows that the wizard for mold flow analysis is run to
proceed automatic meshing process.
Following the wizard, analyzing the injection process for a mold part straight-
forward. Figure 4.58 shows six critical steps in a mold flow analysis: (1) generate
4.6 Simulation of Manufacturing Processes—Mold Filling Analysis 269

Outputs
Inputs 1. Flow and fill
1. Part and mold parameters
design 2. Weld line locations
2. Material flow 3. Pressure to fill
Characteristics 4. Pressure patterns
Computer 5. Clamping force needs
3. Heat transfer Aided Mold
properties 6. Temperature patterns
Flow Analysis 7. Shear patterns
4. Melt temperature
5. Mold temperature 8. Filling
6. Runner and gating 9. Temperatures
systems 10. Shear thinning
7. Cooling systems 11. Freezing and reheating
12. (Temporary stoppage
of flow)

Fig. 4.56 Solidworks Plastics for mold flow analysis

(b). CommendManager SolidWorks

(a). Activate SolidWorks Plastics in add-ins of Options Plastics become available

Fig. 4.57 Activate SolidWorks Plastics tool

mesh, (2) define material, (3) specify process parameters, (4) define the locations
of injection, (5) run simulation, and (6) review the results. Figure 4.59a shows that
the Solidworks Plastics has its capability of generating a mesh automatically on a
surface or solid model. Figure 4.59b shows that injection locations may be automat-
ically defined at step 4. Through the simulation, the designers can view and retrieve
any type of the simulated data related to pack, flow, and warp. Figure 4.60 shows
the list of attributes relevant to the mold flow. Figure 4.61 shows some examples
of critical data from mold flow analysis that include the distribution of shrinkage,
cooling time, pressure, mold temperature, and sink marks.
The SolidWorks Plastics also provides a set of advanced tools to control the
process variables that the mold flows. Figure 4.62 shows the interface to access these
270 4 Computer-Aided Manufacturing (CAM)

1 2 3

6 5 4

Fig. 4.58 Steps in mold analysis wizard

(a). Generate mesh (b). Injection location

Fig. 4.59 Interfaces in the steps of generate mesh and injection location
4.6 Simulation of Manufacturing Processes—Mold Filling Analysis 271

Fig. 4.60 List of attributes in mold flow analysis

(a). Shrinkage (b). Pressure distribution (c). Sink marks

(d). Cooling time (e). Mold temperature

Fig. 4.61 Critical attributes in mold flow analysis

Wizard Mold flow analysis Mold layout analysis

Fig. 4.62 Advanced tools in SolidWorks Plastics

272 4 Computer-Aided Manufacturing (CAM)

tools: (1) polymer to change material properties; (2) filling settings to setup injec-
tion molding parameters such as filling time, melt temperature, mold temperature,
injection pressure limit, and clamp force limit; (3) injection location to add or change
injection locations; (4) flow + pack + warp to specify the simulation type of interest;
and (5) flow results to visualize the results.

4.7 Designs of Tools, Dies, and Molds (TDM)

Tools, dies, and molds (TDM) are essential to most of manufacturing processes
such as casting, forging, injection modeling, and material removal processes. The
geometries and motions of TDM determine part geometries. In particular, a mold
or die should have the same exterior surfaces to the parts at their contact areas.
From this perspective, the concepts of dies and molds are similar and they are used
alternatively. However, a die is used to shape the materials in a deforming process
such as forging and stamping operations, and a mold is used to shape the materials
in a forming process such as casting or injection modeling. Moreover, a mold differs
from a die in sense that a mold assembly is taken apart horizontally, but a die is taken
apart vertically.
Due to low volumes and highly diversified TDM products, toolmakers are mostly
small- and medium-sized enterprises (SMEs) who need highly skilled employees to
design and fabricate tools. However, tooling is essential to all manufactured prod-
ucts, especially nearly half of tooling consumptions are in automotive and defense-
related manufacturing. TDM companies offer the services to design, fabricate, and
test machining tools for manufacturing enterprises. The growth of the manufacturing
industry is greatly influenced by that of the TDM industry. Canie (2010) provided the
data of import and export trades of TDM products in the Unites States from 1997 to
2010 in Table 4.7; it showed that the trade deficit was increased from $ 897 million
to $ 4, 391 million. Currently, the companies in the United States mainly rely on
foreign suppliers to make custom dies and molds. The shrinkage of TDM industry
relates to the shrinkage of whole manufacturing industry. The shrinkage of the US
manufacturing industry has been evidenced by many indicators such as employment
rates, annual payrolls, and the number of newly established companies.
To strengthen a nation’s manufacturing, expanding the TDM industry is critical,
while US manufacturing is facing the challenge in attracting the talents to TDM

Table 4.7 Import and export trades of TDM from 1997 to 2010 in USA (Canie 2012)
4.7 Designs of Tools, Dies, and Molds (TDM) 273

industry. The statistics from the National Tooling & Machining Association (NTMA)
showed that 95% of toolmakers had openings even in the economy with a high
unemployment rate. Using new technologies was expected to improve the efficiency
as much as 20% in TDM companies.

4.7.1 Design Criteria of TDM

Tooling design is a branch of manufacturing engineering that is composed of the

methods and procedures to analyze, plan, design, make, and use tools, dies, and
molds in manufacturing. TDM products are highly diversified since each product
has to be tailored to specified functions, geometries, and manufacturing processes.
While the functions of TDM products are highly diversified, the following design
criteria are commonly applicable to any TDM products.
(1). Cost

Cost is an essential indicator to any manufactured products including TDM prod-

ucts. Since TDM products usually have low volumes, the unit cost is often one of
the primary concerns. This is mostly critical to the companies who make house-
hold appliances, electronics, power hand tools, and housewares. Attractive price was
identified as the leading competitiveness by toolmakers (United States International
Trade Commission 2002).
The total cost of TDM is associated with the tasks involved in different phases of
its product lifecycle. Figure 4.63 gives an estimation of cost at different phases. It is
clear that adopting new technologies in designing, manufacturing, and assembling
of TDM is expected to reduce cost. Therefore, computer-aided technologies have
been widely used to improve productivity and competitiveness, increase machining
capacity, and replace highly skilled labor.
(2). Lead Time

TDM products are mostly customized and manufactured in pull production. From
the time when a user places an order to the time when the product is made and

TDM cost (100%)

Material Secondary
TDM design Tooling base Manufacturing elements (screws,
(11.0%) selections (34.7%) (33.5%)
15.0% ejectors) (5.8%)

Fig. 4.63 Cost breakdown of TDM products

274 4 Computer-Aided Manufacturing (CAM)

Lead time for Custom Delivery (in Weeks )

12
11.6
10
9.3
8 8.7
7.9
6

0
Best in class tool Average tool Laggard tool shops Optimized practice
shops shops (Cimatron 2017)

Fig. 4.64 Average lead times of TDM manufacturers

delivered to user, there is a period of the waiting time called a lead time. Due to the
complexity and high-quality requirements of TDM products, average lead times of
TDM products are long and usually in weeks. Figure 4.64 shows the performance
indicators of the toolmakers based on the average lead time of their TDM products
(Cimatron 2017); the toolmakers with the optimized practice can gain a time window
of 3.7 weeks in capturing emerging business opportunities in comparison with less
competitive toolmakers.
(3). Complexity

TDM products are highly diversified in terms of sizes, geometries, and applied manu-
facturing processes, but toolmakers have to make these at the same facilities to sustain
manufacturing capabilities. (1) One toolmaker usually serves diversified customers
in different industries; the requested TDM products are very different even if they are
made from same manufacturing facilities. (2) Toolmakers are forced to encompass
a wide range of tooling of a particular construction from low-end, simple fabrica-
tion, to highly intricate and technologically advanced products. The complexity of
TDM products and manufacturing processes brings the challenge in (1) defining the
commonalities of products for mass production and (2) reusing knowledge for new
product development.
(4). Precision

The dimensional accuracy of the manufactured product relies on the precision of

the toolings in the manufacturing processes. From this perspective, the precision of
selected tooling should be determined by the required accuracy of products. Tooling
with a low precision is suitable to some products such as plastic buckets in which the
dimensional variants do not affect product uses. An average precision of tooling is
applicable to the parts such as electronic components, keyboards, or clock covers. A
4.7 Designs of Tools, Dies, and Molds (TDM) 275

high precision of tooling is required for parts that be assembled with other compo-
nents under the tight tolerances such as cellphone housing and closures of food
containers. The tooling error will be passed over to products whose accuracy is
adversely affected. Typically, a tooling for high-precision products is expected to
have a repeatable dimensional tolerance of ±0.00005 inch.
(5). Quality

Quality of a TDM product is measured by durability, maintainability, and produc-

tivity. The durability is measured by the product life span in making parts without
the quality concern on parts or tools. The productivity is evaluated by the downtime
and the throughput of parts in a unit time. The maintainability is assessed by the cost
and easiness to prolong the life span, repair wears, prevent premature failure, and
meet new requirements.
(6). Materials

Many factors should be taken into consideration in selecting materials for TDM
products. Some common selection criteria are strengths, toughness, hardness, hot
hardness, machinability, and wear and corrosion resistance. These factors might
be conflicted with each other, for example, the criterion for a high strength and
toughness is conflicted to the criterion for better machinability. Some trade-offs must
be made among these criteria in selecting appropriate materials for TDM products.
For example, if a TDM must sustain very high hot strength in its application, it likely
be brittle that may be broken easily subjected to a dynamic load.

4.7.2 Computer-Aided Mold Design

The geometry of a die or mold must be designed to match the geometry of parts to
be made. In this section, molds and dies are designed based on part models.
(1) Mold Assembly

The geometry of a part is given by the cavity in the mold assembly. Figure 4.65 shows
an example of a mold assembly that consists of a core (male part), a cavity (female
parts), and a parting line or surface. Multiple cores might be needed to define complex
surfaces or some special features such as pockets, holes, and undercut features. A
parting line or surface is to match or split the cores from the cavity in mold assembly
or disassembly.
(2) Computer-Aided Mold Design

In this section, the SolidWorks Mold Tool is used as the computer-aided mold design
tool to illustrate the procedure and steps in designing a mold assembly based on
given part models. Figure 4.66 shows the interface of the Mold Tool commander
276 4 Computer-Aided Manufacturing (CAM)

Core

Other cores

Part Parting surface

Other cores

Cavity

Fig. 4.65 Main components of a mold assembly

Tools for Creation Right-click here to

Tools for Analysis Tools for of Mold
Surfaces Mold Features turn on Customize
Components

Mold Tools
Command Manager

Turning on
Mold Tools

Fig. 4.66 Mold Tool command manager

manager. If the mold tool is unavailable in the default layout, it can be activated by
right-clicking any item in the commend manager group and then checking the mold
tool in the popped list of the command groups. The user may also add the mold tool
by customizing the layout as shown in the right side of Fig. 4.66.
The mold tool commands are organized into three groups: (1) the tools used to
modify, knit, delete, and create surfaces, and fix surfaces by filling holes; (2) the tools
used to analyze the features of mold such as drafts, undercuts, and parting lines; and
(3) the tools used to create the mold components such as cores and cavities. The
procedure of computer-aided mold design is discussed in the next section.
4.7 Designs of Tools, Dies, and Molds (TDM) 277

Fig. 4.67 Procedure of 5

computer-aided mold design Load Part Model Create Parting Line

1 6
Scale Part for Create Parting
Shrinkage Surfaces
Compensation
7
2 Split Tools
Draft Analysis End
Start
8
3 Create Multi-Cores
Evaluate and Fix when Needed
Model
9
4 Create Exploded View
Create Shut-Off
Surfaces if Needed of Mold Assembly

(3). Design procedure

Figure 4.67 shows the procedure of designing a mold based on the given part model:
(1) the part is modeled and scaled to compensate the shrinkage in the manufac-
turing process;
(2) the draft angles of lateral surfaces are analyzed to verify if the part can be taken
out from the mold assembly;
(3) the faces with inappropriate drafts are fixed to ensure right drafts are placed;
(4) if a part has open space(s), add the boundary surfaces to make the core part
and the cavity part separable by using the Shut-Off Surfaces tool;
(5) define a parting line to split the core and cavity by using the Parting Line tool;
(6) use the parting line to generate parting surfaces, so that the geometries and
shapes of the core and cavity are defined, respectively;
(7) knit the parting surfaces with the other boundary surfaces as the solid for the
core and cavity;
(8) split the solid into the core and cavity parts; and
(9) check if the model has an undercut(s), and define additional core(s) for the
undercuts when it is needed.
Finally, new configurations can be defined for the exploded views of mold
assembly.

4.7.3 Shrinkage Compensation

To achieve a better dimensional accuracy of molded part, the dimensions of the

mold should be adjusted to compensate the shrinkage due to temperature changes.
The shrinkage is represented by a scale factor, which corresponds to the ratio of a
dimension between a virtual and physical part. A scale factor depends on material
type and the shape of part. Figure 4.68 a, b shows that the scale tool can be accessed
278 4 Computer-Aided Manufacturing (CAM)

Scale in Mold Tools

(b) Access Scale in Mold Tools

(a) Access Scale in Insert (c) Scale Interface

Fig. 4.68 Scale a part model to compensate shrinkage

in multiple ways such as (1) click insert > mold > scale and (2) click scale under
the command manager of the mold tool. Figure 4.68c shows that a part can be scaled
about centroid, origin, or a custom coordinate system of parts.

4.7.4 Draft Analysis

Lateral surfaces of a molded part should be tapped to ensure that the part can be taken
out from the mold correctly. The Draft Analysis in the commend manger can be used
to evaluate draft angles of lateral surfaces. Figure 4.69 shows the application of the
draft analysis tool. The direction of pulling can be defined by specifying a planar
face, linear edge, or axis, and the user can make the adjustment by using a dynamic
triad shown in the figure.
The draft angle of the tangential plane of each surface with respect to the pulling
direction must be positive and larger than a minimum value. If such a condition is
not satisfied for a surface, a draft feature should be added to the surface by using
the DraftXpert tool shown in Fig. 4.70. The DraftXpert tool requires a number of
inputs including (1) the minimum draft angle, (2) the direction of pulling, and (3) the
group of selected surfaces to be drafted. It should be noted that some features such
as undercuts do not require to have positive draft angles since they will be generated
by other cores.
4.7 Designs of Tools, Dies, and Molds (TDM) 279

Direction of Pulling
Activate
Adjustment Triad

Minimum Draft Angle

Fig. 4.69 Run Draft analysis

Minimum Draft Angle

Direction of Pulling

Selected Draft Faces

Fig. 4.70 Add positive drafts on surfaces

4.7.5 Parting Lines and Shut-Off Surfaces

A mold assembly must be separable into multiple parts. A group of the parting
lines are used to define a parting surface to separate the main core from the cavity.
Figure 4.71 shows the interface to define a group of the parting lines. The main inputs
280 4 Computer-Aided Manufacturing (CAM)

Yellow message indicates that some shut-off

surfaces are required to split the core and
cavity successfully.

A group of parting lines,

which are closed as a loop.

Fig. 4.71 Create parting lines

for the parting lines are (1) the direction of pulling, (2) the minimum draft angle, and
(3) a list of the selected edges that form a closed loop.
The parting lines are inapplicable to internal holes or openings on the part models.
The Shut-off Surface tool is used to create shut-off surfaces for these internal holes or
openings. Figure 4.72 shows the application of the Shut-Off Surface tool where all
of internal holes and openings should be selected to create virtual shut-off surfaces
on these features. This ensures that the mold is separable into core(s) and cavity
(Fig. 4.73).

4.7.6 Parting Surfaces

In using the mold tool, the solids of the core and the cavity are formed by knitting
the boundary surfaces including the parting surfaces where the core and cavity are
separated. Figure 4.74 shows that a parting surface is defined based on (1) the relation
of the surface normal to the pulling direction, (2) the parting line(s), and (3) the margin
distance of from the parting edges. Note that the margin distance should be as large
as possible, so that the parting surface is large enough to form an enclosed volume
for the core and cavity in knitting. Otherwise, the knitting process in the next step
will generate an error message about knitting failure.
4.7 Designs of Tools, Dies, and Molds (TDM) 281

(a)
Green message indicates that the core and
cavity becomes separable

A group of enclosed edges that require

the shut-off surfaces.

Fig. 4.72 Generate shut-off surfaces to make core and cavity separable

Fig. 4.73 Create a parting surface

4.7.7 Mold Components

The Tooling Split tool is used to create separated core and cavity. In using the tooling
split tool, a ketch is defined for the outside boundary of the tooling. Note that the
sketch must be within the parting surface completely to ensure the success of the
knitting operations for enclosed solids. Figure 4.74 shows the interface of using the
tooling split tool where the main inputs are (1) the extrude depths of the core and
cavity and (2) the parting surface. The software gives the default names for the solids
of core and cavity, and these solids can be renamed once they are created.
282 4 Computer-Aided Manufacturing (CAM)

Fig. 4.74 Create core and cavity by Tooling Split

4.7.8 Mold Assembly

Figure 4.75a shows that many intermediate features are generated in designing a
mold assembly. These features may or may not be useful in the following design
stages; these features include reference planes, sketches, surfaces, and solids. It is
helpful to hide the intermediate features after the mold tool is created, and it can be
set by changing the hide/show settings of features, in particular, such settings can be
saved as the configurations for the mold assembly.

Created Surface and

Solid Features.

Hide/Show
Settings

(a) Intermediate features in (b) Core and cavity with

creating a mold tool additional features

Fig. 4.75 Review mold assembly

4.7 Designs of Tools, Dies, and Molds (TDM) 283

(a) Combine multiple features as a body

(c) Save mold components as individual parts (b) Use Save Bodies

Fig. 4.76 Combine and save individual models of parts

Additional features can be added to the core and cavity parts as a fully functioned
mold assembly as shown in Fig. 4.75b. Since a mold assembly involves a number of
components, it is very helpful to save these components as individual part models.
Figure 4.76 shows the features tool by which the solids can be edited, subtracted,
combined as new graphic entities, and they can be saved as individual solids by using
the Save Body tool.

4.8 Computer-Aided Fixture Design

In discrete manufacturing, parts are shaped in a series of manufacturing processes

in which the parts must be fixed firmly to receive the operations (Bi and Zhang
2001). In a manufacturing process such as a material removal process, mechanical
forces must be large enough to remove unwanted materials from the workpieces,
and such forces also cause the deflections of machine tools and fixtures. Therefore,
GD&T of parts are greatly affected by these machine tools and fixtures. Figure 4.77
illustrates the discrepancies of nominal and actual dimensions on the machine tool and
the workpiece subjected to the machining force. Such discrepancies will eventually
cause geometric and dimensional errors of the part. Note that a fixture is a part of
the machine tool that is tailored to the workpiece.
284 4 Computer-Aided Manufacturing (CAM)

Fig. 4.77 Deflections of machine tool and workpiece in manufacturing process

GD&T of parts are affected by numerous factors such as material properties of

machine elements and workpieces, processing parameters, and most importantly,
the fixture to position and fix the workpiece. Fixture design has the direct impact
on product quality and manufacturing cost. The statistic showed that the cost on
fixtures typically took 10%–20% of total manufacturing cost, and around 40% of the
rejected products were caused by poor fixturing solutions. As a critical component of
computer-aided manufacturing (CAM), computer-aided fixture design (CAFD) has
attracted a great deal of attention, and many CAFD tools and flexible fixture systems
have been developed and reported recently (Wang et al. 2010).

4.8.1 Functional Requirements (FRs)

A fixture aims to secure the workpiece firmly at the specified position and orientation
in a manufacturing process. A fixturing solution satisfies the following functional
requirements (FRs):
(1) positioning and orienting the workpiece relative to the cutting tool;
(2) providing the support to reduce the deflection of the workpiece subjected to
external force; and
(3) clamping the workpiece to resist the external loads without any rigid motion.
Figure 4.78 shows a free body in a three-dimensional (3D) space; it has six degrees
of freedom (DoF), i.e., three DoF for translations (T x , T y , T z ) and three other DoF for
the rotations (Rx , Ry , Rz ) along X-, Y-, and Z-axes, respectively. A fixture usually fixes
all of the rigid motions of the part. As shown in Fig. 4.78, if no constraint applies
to an object in a 3D space, the object may have a total of 12 possible directions of
motion. To fix a workpiece in the machining process, any rigid motion along these
12 motion directions must be eliminated. To achieve this goal, a fixturing system
usually consists of three components, i.e., base supports, locators, and clampers as
shown in Fig. 4.79.
4.8 Computer-Aided Fixture Design 285

Fig. 4.78 Degree of freedoms (DoF) of a solid body

Fig. 4.79 Three types of

Clamps Workpiece
components in a fixturing
system

Locators

Supports

In designing these components, the constraints at the following aspects are taken
into consideration, so that the fixturing system can meet the aforementioned FRs
(Basha and Salunke 2013).
(1) Geometric control. All fixture components must be placed on datum surfaces
with adequate supports. No interference is allowed between a fixture compo-
nent and others such as a cutting tool. In addition, fixture components should
be optimized at the aspects of loading and unloading cycle times, the number of
fixture components, accessibility, and the feasibility of the supports of multiple
operations.
(2) Dimensional control. Any solid will be deformed subjected to external
loads, the deformation of a fixturing system adversely affects the accuracy of
machining operations. While the deformation is unavoidable due to clamping
forces or operating forces, it should be minimized to meet the tolerance
requirement.
286 4 Computer-Aided Manufacturing (CAM)

Table 4.8 Other quantifiable evaluation criteria in fixture design

(3) Mechanical control. A workpiece is fixed by the locators of the fixturing

system. GD&T of a finished part depends not only on the deflection of the
fixturing system, but also on the positioning accuracy of the locators; these
locators must be placed optimally for the better dimensional accuracy of parts.
To this end, the quantified criteria in Table 4.8 should be evaluated in designing
a fixturing system.

4.8.2 Axioms for Geometric Control

A free body in Fig. 4.78 has 12 possible directions of motions. To immobilize a solid
body, the 3–2-1 principle can be applied to confine these motions in which three
planes are used to fix the object. A primary plane in Fig. 4.79 confines the rotations
about x-axis or y-axis (Rx and Ry ) and the translation along z-axis (T z ). A secondary
plane in Fig. 4.80a confines the rotation about z-axis (Rz ) and the translation along
y-axis (T y ), and a tertiary plane in Fig. 4.80b confines the translation along x-axis
(T x ). In addition, the primary and secondary planes are selected to counteract external
forces including clamping and operating forces (Fig. 4.81).
The following axioms can be defined for geometric controls based on the 3–2-1
principle:
(1) Six locators are essential and sufficient to fix a prismatic rigid body; more or
less locators may cause the uncertainty in positioning.
(2) One locator eliminates one DoF motion of the body.
(3) Two directions of each DOF must be restrained to immobilize the body.
(4) The numbers of the locators associated to the primary, the secondary, and the
tertiary planes are 3, 2, and 1, respectively.
(5) Locators should be placed as widely as possible to stabilize the object, and
locators should be placed on the most accurate features to achieve the best
possible accuracy.
4.8 Computer-Aided Fixture Design 287

Fig. 4.80 The primary plane in fixturing system

Fig. 4.81 Secondary and tertiary planes in fixturing system

288 4 Computer-Aided Manufacturing (CAM)

4.8.3 Axioms for Dimensional Control

Better dimensional tolerance can be achieved by following the axioms given below.
(1) To avoid tolerance stacks, a locator should be placed on a reference surface
from which the body is dimensioned along this DoF.
(2) When a geometric tolerance is defined between two planes, the primary plane
must be selected as the reference planar surface; the primary plane makes its
contacts to three locators shown in Fig. 4.82.
(3) To locate the centerline of cylindrical surface, the locators must straddle the
centerline.
(4) Locators should be placed on machined surfaces to achieve better dimensional
accuracy when it is possible.
(5) When an axiom for geometric control in Sect. 8.5.2 has conflicted an axiom
for dimensional control, the preference is given to the axiom for dimensional
control.

4.8.4 Axioms for Mechanical Control

To minimize the deformation in the operation when a fixturing system is used, the
following axioms apply:
(1) Place a locator directly opposite to the operating force to minimize its
deformation.
(2) Place a locator directly opposite to the clamping force to minimize its
deformation.
(3) If an external force cannot be reacted directly from any locator, consider to
place a fixed support (no locator) opposite to the force, and avoid an initial
contact of a fixed support with the body.
(4) Apply a clamping force toward locators.
(5) The moment by a clamping force about any of possible axes must sufficiently
overcome that of operating force, and the operating force should be in the
direction where the body remains the contacts to locators.

Fig. 4.82 The primary plane

for GD&T
Three locators on // 0.02
reference plane
4.8 Computer-Aided Fixture Design 289

4.8.5 Form Closure and Force Closure

The conventional approaches to fixture design and planning rely heavily on the thumb
rules, axioms, and prior experiences. A valid fixture design often needs the trial-and-
error iterations. It is desirable to use more scientific methods in fixture analysis
and design. The kinematic and dynamic modeling of a fixturing system can be an
alternative approach.
In kinematic and dynamic modeling, a fixturing system is modeled as a set of fixed
rigid contacts together with an object whose motion is restrained by rigid contacts.
Figure 4.83a shows that an object is fixed by a number of contacts in a fixturing
system, and Fig. 4.83b shows the free-body diagram (FBD) of the object which is
subjected to external forces. All of the applied forces must meet the conditions for
form closure and force closure in operations.
A fixture design should be evaluated by the criteria for form closure and force
closure. For a form closure, the contacts to the object must resist all external forces
to keep the object still in operations. These contacts eliminate all DoF of the object
purely based on the geometric placement at the contacts. For a force closure, the object
maintains the expected contacts to fixture elements subjected to external forces and
moments. In practice, most of the fixtuing systems for machining processes must
be force closed, and they are achieved mainly by utilizing the frictional forces to
immobilize an object.
Kinematic and dynamic modeling aims to evaluate if a fixturing system quar-
antines the conditions for form closure and force closure so that (1) there is no
possible movement under external loads; (2) the fixturing system is fully accessible
and detachable when a part is loaded and unloaded; and (3) the axioms for geometric,
dimensional, and mechanical controls are appropriately followed to position and
secure the part with better dimensional accuracy.

Fig. 4.83 The fixed object and its free-body diagram (FBD)
290 4 Computer-Aided Manufacturing (CAM)

4.8.6 Fixture Design in Manufacturing Processes

Many manufacturing processes involve fixture designs. Figure 4.84 shows the rela-
tions of fixture design to other tasks in manufacturing process planning. The fixture
design is affected by machine tools, process parameters, tooling paths, and machining
programs which are tailored to the performed manufacturing processes. The tasks
of fixture design are logically followed one after another, while an iterative process
is usually needed since the information at earlier steps is insufficient to verify if
the design constraints in the following steps are satisfied. Figure 4.84 shows that a
fixture design involves four stages with the specified main tasks. It can be seen that
a fixture design requires the information workpiece and manufacturing processes,
and Fig. 4.85 shows the essential information for a fixture design includes products,
machine tools, manufacturing processes, and verification and quality control.
The fixturing solution has to be verified thoroughly since the fixture makes a
direct contact to various hardwares such as machine tools, cutters, and other auxiliary
systems (e.g., lubrication or coolant systems). Verification should be an integrated
part of the fixture design. Many components bring the constraints to a fixturing
system (see Fig. 4.85); in addition, not all information is available when the fixture

Fig. 4.84 Fixture design in manufacturing process planning

4.8 Computer-Aided Fixture Design 291

Fig. 4.85 Key design factors and criteria of fixture design

concept is developed. Therefore, the verification is needed since not all of fixturing
requirements can be defined in depth from the beginning.
Figure 4.86 shows a framework for the verification of fixture design (Wang 2010).
The critical tasks in the verification are analysis of geometric constraints, toler-
ance analysis, stability analysis, stiffness analysis, and accessibility analysis. The
knowledge-based engineering (KBE) approach could be applied to accelerate the
design process and eliminate design defects.

Fig. 4.86 Verification of fixture design

292 4 Computer-Aided Manufacturing (CAM)

4.8.7 Computer-Aided Fixture Design (CAFD)

Conventionally, fixture design might take days even weeks, and the performance of
the fixturing system relied heavily on the designers’ experiences. Computer-aided
techniques help to accelerate the process of the fixture design, and CAFD should be
an essential component in the cycle time of product development. Figure 4.87 shows
an integrated system for computer-aided process planning (CAPP) in which CAFD
is able to obtain all of the relevant information for fixture design. Especially, part
models can be directly uploaded for processing planning, and the results of CAFD
can be used to plan cutting paths and program machine tools. In this section, some
main aspects of CAFD are discussed below.
(1). Fixture Design Library

Fixtures are customized to the parts to be manufactured; therefore, fixtures are highly
diversified for different applications. To develop customized fixturing solutions in
a cost-effective way, modular structure should be adopted in which the fixturing
system consists of various fixture elements such as supports, locators, and clamps,
and different configurations can be assembled by selecting different elements and
assembling them in different ways. By developing a fixture design library, commonly
used fixture elements can be standardized, virtual models can be parametrized, and the
fixtures solutions to new tasks can be obtained by reusing existing fixture elements.

Fig. 4.87 CAFD in computer-aided process planning (CAPP)

4.8 Computer-Aided Fixture Design 293

(2). Interference Detection

No interference is allowed in assembling a fixture, or loading and unloading a part.

To ensure no interference occurs to the fixture, virtual analysis could be applied
to visualize assembling processes of a fixturing system, and simulate the loading
and unloading process of parts. With the virtual models of parts, fixture elements,
and machine tool, solid modeling tools are able to run interference checks to detect
possible interferences before the physical fixturing system is implemented.
(3). Accessibility Analysis

Accessibility analysis investigates possible collisions of fixture elements to

surrounding objects in (1) loading and uploading parts and (2) operating a manu-
facturing process such as moving a cutting tool along the given path. To evaluate
accessibility of an object, the motions of machine, tools, and parts in the reference
coordinate system are considered with respect to the fixture elements. The Solid-
works Motion can be utilized to detect possible collisions in the motions of multiple
objects.
(4). Analysis of Deformation and Accuracy

The primary objective of a fixturing system is to hold the part when it receives the
operation; therefore, the quality of part is affected by geometric and dimensional
deviations of the fixturing system subjected to the operating forces. The layout of a
fixturing system should be optimized to minimize the impact on the quality of parts.
To estimate the dimensional accuracy, the computer-aided engineering (CAE) tool
can be used to simulate the deformations of fixture elements and workpiece subjected
to operating forces.

Example 4.6 Figure 4.88 shows a fixture setup for the milling operation on the part
for a slot feature. The fixturing system is implemented by using the 3–2 principle. The
primary plane makes the contacts to the bottom surface of the part. The movements
along x-axis and y-axis are constrained by one and two locators, respectively. The
position of locator A can be placed at L 1 = 25, 75, and 125 mm; the position of
locator B can be placed at L 2 = 175, 225, and 275 mm; and the position of locator C
can be placed at H = 25, 75, 125, and 175 mm. The clamping forces along x-axis and
y-axis are − 2 kN and −0.5 kN, respectively, and the cutting forces are (F t_x , F t_y) =
(−4kN, −2kN). Determine the positions of the locators to optimize the dimensional
accuracy of the part.

Solution Based on the provided information, the workpiece and fixture elements are
modeled, and an FEA model for static analysis is further defined in the Solidworks
Simulation. The model for static analysis is shown in Fig. 4.89a, and an example
of the workpiece deformation from the simulation is illustrated in Fig. 4.89b. To
optimize the layout of three locators, a design study in Fig. 4.89c is defined for three
294 4 Computer-Aided Manufacturing (CAM)

Fig. 4.88 Example of deformation analysis

Fig. 4.89 The solution to fixture design problem in Example 4.6

variables (i.e., L 1 , L 2 , and H) to evaluate the displacement of workpiece. The design

study finds that the optimized positions of three locators are L 1 = 75 mm, L 2 =
175 mm, and H = 175 mm, and the minimized displacement is 2.89 microns.
4.9 Computer-Aided Machining Programming 295

4.9 Computer-Aided Machining Programming

To automate manufacturing processes, human operators are replaced by machines,

and the operation-related decisions are made by computer programs. Computer-
aided machining programming is to generate the programs to run machine tools. In
this section, machining programming is introduced, and computer numerical control
(CNC) is discussed as the programming tool for automated machining processes.
In this section, the programming of machining processes is focused, and the Solid-
Works High-Speed Machining (HSM) is used to generate, simulate, and verify control
programs.
Machining processes offer tight geometric and dimensional tolerances and high
surface finishes. Therefore, machining processes are often used as the secondary
processes to improve product quality or add machined features that are difficult to be
made by other processes such as a side hole on a molded part. However, a machining
process usually involves in a high cost due to a number of reasons: (1) reshaping a
part by machining differs from net-shape operations such as casting and injection
molding in sense that a machining process takes a long time to remove unwanted
materials gradually to shape the desired features on parts; (2) a high-performance
machine tool requires a high initial capital investment that increases the unit cost of
products to be made; (3) cutting tools are made from the materials with high strength,
durability, and toughness; however, they are consumable in machining processes; and
(4) the volumes of products that are made by machining processes, such as molds
and dies, are mostly very limited and impractical to be made in the mass production
paradigm.
To program a machine tool, it is helpful to understand controllable variables of
the machine tool in machining processes. A machining operation refers to a material
removal process in which a cutting tool cuts away unwanted materials from the

Fig. 4.90 Starting materials and finished products of machining processes

296 4 Computer-Aided Manufacturing (CAM)

part for a desired feature or dimension on parts. As shown in Fig. 4.90, the starting
material of a machining process can be a primarily processed part or raw material
with a simple shape such as block, bar, sheet, roll, beam, or tube. After machining
processes, desired features such as turning, milling, drilling, or grinding features are
made on finished parts.
Based on overall geometries, machined parts can be classified into prismatic
parts and cylindrical parts, and some common machined features such as holes,
slots, pockets, flat surfaces, and complex surface contours. To automate a machining
process, the motions of all moving machine elements have to be controlled by
programs. Figure 4.91 shows an example of machine tools in which the main moving
components are worktables, spindles, axes, cutting tools, and other auxiliary equip-
ment attached on motion axes. From the perspective of productivity, a machining
process also controls all process variables relevant to a material removal rate (RMR ).
RMR is defined as the volume of the removed materials in the unit time; RMR is
evaluated based on four process variables. These process variables are illustrated in
Fig. 4.92 and also have been explained.
(1) Cutting speed (v) is the relative speed at which a cutting tool passes over the
part, and the cutting speed is measured at the contact of the cutting tool and
the part. RMR is affected greatly by the cutting speed v. The higher a cutting
speed v is, the larger RMR is, and the better the productivity is. Note that the
life span of a cutting tool also depends on RMR . Therefore, tool manufacturers

Fig. 4.91 Main variables in

a machining process

Fig. 4.92 Main variables in

a machining process
4.9 Computer-Aided Machining Programming 297

usually make the recommendation of appropriate cutting speeds based on the

properties of tooling and part materials.
(2) Spindle speed (N) is an angular velocity of the rotational axis, which is usually
expected in the number of revolutions per minute (RPM). When the main
motion of a machining process is from a spindle, the spindle speed N is
correlated to the cutting speed v as
v
N= (4.12)
π·D

where v is the cutting speed and D is the diameter of the part at which the contact is
made between the cutting tool and the part.
(3) The depth of cut (d) refers to the depth that the cutting tool penetrates into the
part in the material removal process.
(4) The feed rate (f ) is the velocity to feed the cutting tool along the cutting path.
It is measured by a moving distance of the cutting tool in each revolution of
the spindle. The unit of a feed rate f can be either of inches per revolution or
millimeters per revolution.
RMR can be computed from the cutting speed v, the depth of cut d, and the feed
rate f as

RM R = v · f · d (4.13)

4.9.1 Procedure of Machining Programming

Machining programming begins with the collection of the information about

machined features on parts, machine tools, cutting tools, and geometric and dimen-
sional tolerances. Figure 4.93 shows the procedure of generating a control program
for a given machining process.
The procedure in Fig. 4.93 consists of eight steps, and the main tasks involved in
these steps are described in Table 4.9.

4.9.2 Standards of Motion Axes

A program consists of legal statements that are written by following a set of the prede-
fined rules called syntax in combining symbols and words. To program a machine
tool, the nomenclatures for the motions and coordinate systems are standardized by
the Electronic Industries Association (EIA) as EIA267-C standards.
298 4 Computer-Aided Manufacturing (CAM)

Fig. 4.93 The procedure of programming a machining process

Table 4.9 Main steps and tasks in machining programming

In the EIA267-C standards, 14 axes can be used to define the positions and move-
ments of parts, cutting tools, or other objects involved in machining processes. Note
that nine axes should be enough to describe the motions of conventional machine
tools, while some advanced machine tools are equipped with more auxiliary motion
axes. Without loss of generality, the following discussion considers a machine tool
that has less than nine motion axes, i.e., three primary linear axes (X, Y, and Z); three
primary rotary axes (A, B, and C); and three secondary linear axes (U, V, and W).
As shown in Fig. 4.94, a motion is defined in a given coordinate system, and the
positive directions of motion axes in a coordinate system follow the right-hand rule
as.
(1) Three fingers (i.e., the thumb, index, and middle fingers) are placed mutually
perpendicular with each other, and the origin O(0, 0, 0) is set at their inter-
section. Each finger points to one direction of the translation, i.e., the thump
4.9 Computer-Aided Machining Programming 299

Fig. 4.94 The right-hand

rule in a coordinate system
(CS)

finger, the index finger, and middle finger point to the positive X-, Y-, and
Z-axes, respectively. The coordinate system is defined as {O − XYZ}.
(2) The motion axis of the main spindle is defined as Z-axis, and the positive
direction is toward the spindle.
(3) The longest travel slide is designated as X-axis, and it is perpendicular to the
Z-axis.
(4) The positive rotation around an axis is the clockwise direction of the axis.
Using the right-hand rule, the motion axes of conventional machine tools are
described in Table 4.10.

4.9.3 Default Coordinate Systems and Planes

A toolpath consists of line, arc, or curve segments in a plane or 3D space. A 2D

toolpath can be usually defined in one of three default planes (XY, XZ, and YZ)
in the Cartesian coordinate system. The default Cartesian planes are defined based
on the given coordinate axes X, Y, and Z. Figure 4.95 shows the examples of three
default planes in a mill machine. In the default XZ-plane, a central reference point
O is used to measure the coordinates along X- and Z-axes, and the coordinates of O
is set as O (0, 0).
As shown in Table 4.10, a turning machine has only two motion axes, i.e., a
primary axis (Z-axis) and a secondary axis (X-axis). The dimension along Y-axis is
given by the depth of cut. As a result, Fig. 4.96 shows a default two-axis CS. Using
default CSs helps to transfer CNC programs or measurements among different lathes.
It should be noted that the cutter on a CNC lathe is different from the cutting tools
on conventional lathes. The cutter is usually placed on the top or side of the machine
tool; alternatively, it can be mounted at the front side horizontally on a conventional
lathe.
A CNC lathe makes rotational parts; therefore, the machined features are dimen-
sioned by radii or diameters. Diametrical programming uses the x-coordinate as a
diametrical dimension at the given Z-coordinate. For example, if a feature has a 5-in
outside diameter and absolute coordinates are used in programming, the x-coordinate
300 4 Computer-Aided Manufacturing (CAM)

Table 4.10 Motion axes of conventional machine tools

Type Description Illustration
4.9 Computer-Aided Machining Programming 301

Fig. 4.95 Default coordinate planes in a mill machine

Fig. 4.96 Default coordinate

system on a CNC lathe

of the motion commend becomes X5.0. Radial programming uses the x-coordinate
as a radical dimension at the given Z-coordinate. For example, if a feature has a 5-in
outside diameter and absolute coordinates are used in programming, the x-coordinate
of the motion commend becomes X2.5.
302 4 Computer-Aided Manufacturing (CAM)

4.9.4 Machine, Part, and Tool References

Working points in a machining process are described with respect to certain refer-
ences in a coordinate system, and three commonly used references are (1) machine
reference zero (M), (2) part reference zero (PRZ) (W), and (3) tool reference zero (R).
These references are used to define the coordinates of working positions as follows
(Figs. 4.97 and 4.98):
(1) The coordinates of a working point are measured from the machine reference
zero M.
(2) The points on workpiece are measured from the part reference zero W.

Fig. 4.97 Tool reference

zero (R), part reference zero
(W), and machine reference
zero (M)

Fig. 4.98 Absolute

coordinates from the origin
of coordinate system
4.9 Computer-Aided Machining Programming 303

(3) The PRZ (W) should be aligned to the lower left-hand corner on the top surface
of the part stock.

4.9.5 Absolute and Incremental Coordinates

The coordinates of the working points can be either of (1) absolute measured from the
origin or (2) incremental measured from a pre-existing location. Figure 4.99 shows
the scenario in which absolute coordinates are specified to define working points.
The absolute coordinates of points A, B, and C are measured from the origin O of
the coordinate system {O-XY}.
Example 4.7 Determine the absolute coordinates of working points shown in
Fig. 4.99a.

Solution The absolute coordinates are measured from origin O; therefore, the coor-
dinates of the working points (1, 2, …, 7) in Fig. 4.99a are determined as (1.0, 3.0),
(−1.0, 2.0), (−3.5, −0.5), (−3.0, − 1.0), (−4.0, −2.0), (1.0, −2.5), and (3.0, −2.0)
in Fig. 4.99b, respectively.
Absolute coordinates also apply to the default two-axis coordinate system for a
CNC lathe. For a lathe, absolute coordinates take the reference from origin (X0, Z0)
as shown in Fig. 4.100a. To find Z- and X-coordinates of a working point, project the
working point on Z-axis and X-axis, respectively. The absolute coordinates of the
working points P1 to P6 are determined in Fig. 4.100b. Note that the X-coordinates
can be either radical (X R ) or diametric (X D ).
Incremental coordinates are measured from existing points. A new working point
can be defined by specifying a relative distance of an existing point. Figure 4.101
shows an example of defining incremental coordinates for working points in a
toolpath.

Fig. 4.99 Example of absolute coordinates

304 4 Computer-Aided Manufacturing (CAM)

Fig. 4.100 Example of absolute coordinates in XZ-plane

Fig. 4.101 Working points

with incremental coordinates
in sequence

Example 4.8 Find the incremental coordinates of the working points shown in
Fig. 4.101a.

Solution The incremental coordinates of a new working point are measured from
an existing point one by one in sequence. Taking an example of A, its incremental
coordinates (−2.0, − 1.0) are measured from origin O; similarly, other working points
B, C, D, E, and F are measured from A, B, C, D, and E as shown in Fig. 4.102b,
respectively.
4.9 Computer-Aided Machining Programming 305

Fig. 4.102 Incremental coordinates for Example 4.8

4.9.6 Types of Motion Paths

A geometric feature of part is determined by the profile and motion path of the cutting
tool. The motion paths can be classified into four types at shown in Fig. 4.103. (1) A
point-to-point path consists of a set of discrete points where the machining operations
are performed; no machining operation occurs when the tool moves from one point

Features and Machined feature

working points and working points

Z
Z
Y
Y
X
X

(a). point to point motion (b). point to point straight

(drilling and boring) motion (frame milling)

Machined feature Machined feature

and working points and gridded points

Z
Z
Y Y

X
X
(c). 2-axis contouring with (d). 3-axis contouring with
switchable plane (2-D continuous path (3-D
contouring milling) contour milling)

Fig. 4.103 Four types of tool paths

306 4 Computer-Aided Manufacturing (CAM)

to the next point. A point-to-point path is used to specify the motion for a drilling or
boring operation. (2) A point-to-point straight path consists of a set of line segments,
each line segment connects two working points directly, and the machining operations
are performed continuously from one end to the other end of the line segment. A
point-to-point straight path is applicable to frame milling. (3) A two-dimensional
continuous path is a smooth curve for the connections of working points on a 2D
plane; it is mainly used in 2D milling. (4) A three-dimensional continuous path is
formed by following the nodes on the grid by zigging-zagging. It is used in 3D
contouring.

4.9.7 Programming of Machining Processes

Automated machine tools are controlled by computer programs, and three ways to
program machine tools are manual programming (NC), conversational programming
at shop floor, and offline programming.
Manual programming is to program a machine tool for simple tasks; the program is
tied to certain machine tool with the optimized performance of execution. However,
manual programming becomes very tedious and error-prone when the machining
operation is complex. Therefore, manual programs are mainly used when (1) a
machine tool is sophisticated to make parts with high volume and (2) the efficiency
of machining operations is prioritized.
Conversational programming uses graphic and menu-driven interfaces to create
programs for machining operations. Conversational programming allows users to
check inputs, review toolpaths, and simulate the machining operations. It is very
popular in small- or medium-sized enterprises (SMEs); since machine operators
in SMEs often take the full responsibility of setting up machine tools, fixtures,
and tooling, and creating, verifying, and running the programs. In comparison with
manual programming, conversational programming is more user-friendly and it can
reduce the programming time dramatically. Conversational programming provides
a convenient way to write programs for parts from the same type of machines;
sometimes, it is the only way of programming on a legacy machine tool.
Offline programming has gradually gained its popularity, and it supports program-
ming at a higher level than manual and conversational programing. Offline program-
ming has its advantages over programming techniques in (1) defining toolpaths auto-
matically based on given working points, (2) generating the programs which can be
customized to different types of machine tools, and (3) reusing functional modules
by establishing and sustaining the library for reusable modules and routines.
4.9 Computer-Aided Machining Programming 307

P4
Y

L3 L2

P3
P1 X L1 C1
P2
P0
0 1 2 3 4 5 6 7

Fig. 4.104 NC programming by APT

4.9.8 Automatically Programmed Tools (APT)

One of the most popular machine programming tools is Automatically Programmed

Tool (APT). APT was developed in 1965 by Douglas T. Ross to write the instructions
for NC machine tools. APT can be used to program the machine tool for complex
products. APT is still used in practice and it accounts for 5–10% of all machining
processes in the defense and aerospace industries.
An NC program from APT defines a series of lines, arcs, and points as the entities
to define the features of part to be machined. These entities are then used to generate
a cutter location (CL) file. The statements in a NC program from APT can be classi-
fied into four types: (1) geometric statements to describe part geometry, (2) motion
statements to describe toolpaths, (3) post-processor statements to specify feeds and
speeds, and (4) auxiliary statements to define tools and tolerances.
Example 4.9 Use APT to write a NC program to mill the part shown in Fig. 4.104.

Solution By using APT, the NC program is written and shown in Table 4.11. The
first part (line 1–3) is to label the program, specify machining type, and select the
tool. The second part (line 4–13) is to define all geometric entities. The third part
(line 14–16) is to set up operating conditions for cutting speed, feed rate, and coolant
supply. The fourth part (line 17–22) is for all motion statements. The last part (line
24–26) is to reset the tool position and terminate the machining operation.

4.9.9 Computer-Aided Machining Programming

APT is suitable for manual programming, but it is tedious and error-prone in defining
part geometry and working points. The technique of CAD-based programming
was introduced in 1980s, CAD-based programming uses CAD models to retrieve
the machined features and generate NC programs semi-automatically. Note that a
programmer should have the basic understanding of the machining processes to
308 4 Computer-Aided Manufacturing (CAM)

Table 4.11 NC program to mill the part in Fig. 4.103

NC Code Explanation
Line 1: PARTNO / EXAMPLE9-3 ; label the program as “EXAMLE9-4”
Line 2: MACHIN / MILL, 1 ; select the target machine and controller type
Line 3: CUTTER/ 0.5000 ; specifies the cutter diameter
Line 4: P0=POINT/-1.0, -1.0, 0.0
Line 5: P1=POINT/ 0.0, 0.0, 0.0
Line 6: P2=POINT/ 6.0, 0.0, 0.0
Line 7: P3=POINT/ 6.0, 1.0, 0.0
; geometry statement to specify the pertinent surface
Line 8: P4=POINT/ 2.0, 4.0, 0.0
of part
Line 9: L1=LINE/P1, P2
Line 10: C1=CIRCLE/CENTER, P3, RADIUS, 1.0
Line 11: L2=LINE/P4, LEFT, TANTO, C1
Line 12: L3=LINE/P1, P4
Line 13: PL1=PLANE/P1, P2, P3
Line 14: SPINDLE/573 ; set the spindle speed to 573 RPM
Line 15: FEDRAT/5.39 ; set the feed rate to 5.39 inch per minute
Line 16: COOLNT/ON ; turn the coolant on
Line 17: FROM/P0 ; gives the starting position for the tool
Line 18: GO/PAST, L3, TO, PL1, TO, L1 ; initialize motion, drive, part, and check surfaces
Line 19: GOUP/L3, PAST, L2
Line 20: GORGT/L2, TANTO, C1
; contour the part in clockwise direction
Line 21: GOFWD/C1, ON, P2
Line 22: GOFWD/L1, PAST, L3
Line 23: RAPID ; move rapidly once cutting is down
Line 24: GOTO/P0 ; return the tool to home position
Line 25: COOLNT/OFF ; turn off the coolant
Line 26: FINI ; terminate the program

specify feed rates, speeds, and depths in a NC program. A CAD-based programming

system is used to generate, edit, and simulate toolpaths, and predict the processing
time and cost of machining operations. In addition, the program system usually
includes a design library for standardized cutting tools and part materials. According
to EIA standards, a NC program is written based on the following assumptions:
(1) The cutting tool moves relatively to the workpiece without the consideration
of how such a relative motion happens. Taking an example of the motions on a
shaper, the cutting tool is actually fixed, and the workpiece moves toward the
cutting tool to have a relative motion with each other.
(2) The parts are positioned and oriented in a Cartesian coordinate system.
(3) The reference origin (0.0, 0.0, 0.0) is floating, so the program is flexible to
align with any coordinate system of interest when the setup of a machine tool
or part is changed.

(1). Motion of Cutting Tools

In a CNC program, the cutting tool is moved in one of the following ways:
(1) A rapid movement to a designated location without machining operation.
(2) A straight translation actuated by one or multiple axes.
(3) A circular motion in a planer surface.
(4) A mixed planar and linear motion that is commonly referred as a 2½-D motion.
A planar motion is implemented by two simultaneous motion axes in one plane.
4.9 Computer-Aided Machining Programming 309

This third motion direction is the feeding direction, which is perpendicular to

the plane formed by two motion axes.
(5) A complex motion that is combined by the aforementioned motions. For
example, the arc motion in an arbitrary plane can be modeled as a set of
straight lines and implemented by three simultaneous motion axes.
(6) A true 3D motion along any path and orientation in the workspace of the
machine tool. Due to new advancement in sensing and information tech-
nologies, most recent CNCs are able to generate true 3D motion with over
5 DoF.

(2). Block Structure

Each statement in a CNC program is called a block, and the program is a collection of
blocks. A block includes one or a few of instructions or words that are separated by
space or tab. A word is a character for a single function. For example, ‘X’ represents
a displacement along X-axis and ‘F’ represents for the feed rate.
A block to a program is a sentence to a language. Similar to the sentences that are
separated by periods in English, blocks are separated by the End-Of-Block (EOB)
in a program. An EOB character terminates a block to show the beginning of next
block. In running a program, EOB indicates to execute the commands specified in
the block. An EOB character can be inserted by pressing the Enter or Return Key
on keyboard. Sentences are assembled into an essay in a language, and blocks are
assembled into a program in programming.
The structure of a block is shown in Fig. 4.105. ‘/’ specifies if the following block
is skipped in running the program. ‘N’ is the identification of the block; the identi-
fication is only needed when it is referenced by other blocks. ‘G’ is the preparatory
functions for next operations. ‘X’, ‘Y ’, ‘Z’ are for the displacements in the primary
X-, Y-, Z-axes. ‘U’, ‘V ’, ‘W ’ are for the displacements in the secondary motion axes.
‘A’, ‘B’, ‘C’ are for the angular displacements in X-, Y-, Z-axes. ‘I’, ‘J’, ‘K’ are for

/ N**** G** X**.** Y**.** Z**.** A**.** B**.** C**.** F**.** S**.**

Preparatory Linear and angular Feed function

functions of G-codes position data
Speed function
Block identification

Block skip condition

T** D** M** N** ; *****

Block comment
Number of block repetitions, not ISO, corresponds to FAGOR 8055M
Miscellaneous or auxiliary functions
Tool offset number
Tool number

Fig. 4.105 The structure of blocks in a CNC program

310 4 Computer-Aided Manufacturing (CAM)

the distances to the arc center or thread leads which are parallel to X-, Y-, Z-axes,
respectively. ‘F’ is for the feed rate, ‘S’ is for the cutting speed, ‘T ’ is for the tool
number, ‘D’ is for the offset of tool, and ‘M’ is for the miscellaneous functions.
(3). G-Code

The programs for CNC machine tools are also called as G-code which has been stan-
dardized by several organizations. The US standard was RS274 version D developed
by the Electronic Industries Alliance (EIA) in 1979. The global standards were DIN
66,025 developed in Germany, PN-73 M-55256 and PN-93/M-55251 developed in
Poland, and ISO 6983 which was widely used by other countries (Wikipedia 2019).
However, all standards adopt similar words and the block structure in defining tool-
paths and specifying process parameters. Table 4.12 shows five types of words in
G-code.
G-words in Table 4.12 are used to control the tool motions, the commonly used
G-words are listed in Table 4.13.
In a CNC program, each G-word has its specific format, and the formats of some
common G-words are described as follows:

Table 4.12 Five types of words in G-code

Table 4.13 Commonly used G-words

4.9 Computer-Aided Machining Programming 311

(1) Rapid point-to-point motion (G00) takes the shortest path to move the tool to
the new position at the specified coordinates at a rapid feed rate:

Format: G00 X∗∗∗ Y∗∗∗ Z∗∗∗

where
X, Y, and Z are the words for the displacements along X-, Y-, and Z-axes,
respectively, and *** is the value of a displacement.

(2) Linear motion (G01) moves the tool to a new position with the specified coor-
dinates at the given feed rate; Fig. 4.106 shows an example of a straight line
generated by a G01 block:

Format: G01 X∗∗∗ Y∗∗∗ Z ∗∗∗ F∗∗∗

where
X, Y, and Z are the words for the displacements along X-, Y-, and Z-axes, respec-
tively; F is the word for the feed rate; and *** is the value of a
displacement or a feed rate.

(3) Closewise arc motion (G02) moves the tool along an arc in the clockwise
direction until it arrives to a new position at the given feed rate; Fig. 4.107
shows a closewise arc path in XY-, XZ-, and YZ-planes, respectively,

Format: G17 G02 X∗∗∗ Y∗∗∗ I∗∗∗ J∗∗∗ F∗∗∗ ; Arc on XY − plane
G18 G02 X∗∗∗ Z∗∗∗ I∗∗∗ K∗∗∗ F∗∗∗ ; Arc on XZ − plane
G19 G02 Y∗∗∗ Z∗∗∗ J∗∗∗ K∗∗∗ F∗∗∗ ; Arc on YZ − plane

Fig. 4.106 A straight-line Z

path defined by G01

New position (X1, Y1, Z1)

X
Current position (X0, Y0, Z0)
312 4 Computer-Aided Manufacturing (CAM)

Y Z End Z
position J
Start
Center position K
Center
End Center
J position K I Start
Start position
position I End
position
X X Y
(a). XY plane (b). XZ plane (c). YZ plane

Fig. 4.107 Clockwise arc paths by G02 in XY-, XZ-, and YZ-planes

where
G17, G18, or G19 represents the plane where an arc lies in;
X, Y, and Z are the words for the coordinates of the designated position;
I, J, and K are the words of the displacements from the starting position
to the arc center along X-, Y-, Z-axes, respectively;
F is the word for the feed rate; and.
*** is the value of a coordinate, displacement, or feed rate.
‘M’ stands for ‘machine’ and M words are used to activate or deactivate miscel-
laneous machine functions. Table 4.14 shows the functions of some commonly used
M words.
Example 4.10 Write the G-code for the motion control of the milling operation on
the part shown in Fig. 4.108a.

Solution The milling operation is to create the profile of workpiece. The milling tool
starts its motion from the home position S, and then it passes a series of working
points in the order of S → A → B → C → D → E → F → G → H → I →
A → S. The toolpath includes two arc motions. The first arc is B → C → D in
the counterclockwise direction, and the second arc is G → H is in the clockwise
direction. The G-code for the motion along the toolpath is written and shown in
Fig. 4.108b.

Table 4.14 Commonly used M words

4.9 Computer-Aided Machining Programming 313

Y
F G G90
E M G00 X-1.0 Y-1.0 ; (S)
G01 X0 Y0 F7.5 ; (A)
H Y2.314 ; (B)
D N
G03 X0.0 Y3.866 I-0.5 J0.866 ; (C)
C X0. Y3.866 I-1. J0 ; (D)
O
B G01 Y5.5 ; (E)
G02 X0.5 Y6.0 I0.5 J0 ; (F)
G01 X4.5 ; (G)
X
G02 X6.0 Y4.5 I0 J-1.5 ; (H)
A (0, 0) I
G01 Y0 ; (I)
S(-1, -1) G01 X0 ; (A)
G00 X-1.0 Y-1.0 ; (S)
(a). Contouring operation (b). NC programming

Fig. 4.108 G-code example with arc motions

Computer-aided manufacturing (CAM) tools support the offline programming, so

that CNC programs can be generated automatically based on part models. Here, the
SolidWorks HSMWorks is used as an example of the CAM tools. HSMWorks supports
all conventional strategies of machining operations such as contour, parallel, pocket,
radial, scallop, spiral, and pencil. It is capable of generating optimal toolpath with less
machining time and better surface finish, and it provides the simulation environment
to verify CNC programs. Figure 4.109 shows the main features of HSMWorks to
program CNCs for turning and milling operations.

Fig. 4.109 The main features of HSMWorks for turning and milling operations
314 4 Computer-Aided Manufacturing (CAM)

Fig. 4.110 Programming interface of HSMWorks

Figure 4.110 shows the interface of HSMWorks, which is integrated as an adds-

in CAM tool in SolidWorks. A part model can directly be imported, the machined
features can be automatically identified, but the operation for each machined feature
is programmed, respectively. The toolpath for certain feature can be created auto-
matically, and it can be visualized and simulated at any time as shown in Fig. 4.110b.
The HSMWorks includes design libraries for programmers to select process types,
machines, and cutters as shown in Fig. 4.110a. When the CNC is specified, the
complete CNC program can be generated by post-processing, and the program can
be downloaded to the machine to control machining operations. In addition, the
performance of a CNC program and corresponding machining operation can be
evaluated automatically as shown in Fig. 4.110c.

4.10 Summary

Design of manufacturing processes takes into consideration many factors including

materials, machines, tools, and numerous operating parameters; in addition, manu-
facturing processes are evaluated against some conflict criteria such as cost, accuracy,
productivity, flexibility, and adaptability. Conventional ways for design of manufac-
turing processes have their limits in dealing with the complexities of modern products
and manufacturing processes. It becomes critical for engineers at SMEs to formulate
4.10 Summary 315

manufacturing process design problem as a multidisciplinary engineering problem,

and seek the engineering solution via modern computer-aided techniques.
Many computer-aided manufacturing (CAM) tools are available to deal with
various engineering problems in design of manufacturing processes. This chapter
covers the theories and tools for designs of composite materials, molds and dies,
fixtures, and machining programs. Engineers should be trained to use these CAM
tools in designing manufacturing processes for complex products and systems.
Design Problems

Problem 4.1 Create a model of the object with composite materials and determine
the stress distribution subjected to the loads shown in Fig. 4.112. The object has
a rectangle shape with the width and the height of 1 inch. The load along y-axis
on the top edge is 2000 lb/in, and the loads along x-axis and y-axis on the right
side are 1000 lb/in. The displacements of X, Y, Z, Ry are restrained on the left side
(Fig. 4.111).

Table 4.15 shows the layup of the composites. It is composed of graphite/epoxy

tape, the angles shown are relative to the global axis, i.e., the 0 degree ply 1 has its
fibers running along the Y-direction. 90 degree ply 4 has its fibers running along the
X-direction. The composite plies are graphite/epoxy tape with a thickness of 0.0054
in. The properties of graphite are modulus of elasticity E = 6.96 × 105 psi, Poisson’s
ratio ν = 0.28, density ρ = 0.0809 lb/in3 , S ut = 1.46 × 104 psi, and yield strength
is S y = 1.75 × 104 psi.

Fig. 4.111 Composite part Top side with the load of 2000 lbf/in
in Problem 4.1
(MSCsoftware 2020)
Right side with the load of (1000, 1000) lbf/in
Left side fixed

1-in by 1-in
rectangle made of
graphite/epoxy
composite

Bottom side free

316 4 Computer-Aided Manufacturing (CAM)

Fig. 4.112 Part model for

Problem 4.2

B1

A1

B2

A2

(a) Part model with GD&T (b) Assembly model

Table 4.15 Layup of

Ply no. Materials Thickness (inch) Direction
example composites in
(degree)
Problem 4.1
1 Graphite/epoxy 0.0054 0
2 0.0054 45
3 0.0054 −45
4 0.0054 90
5 0.0054 90
6 0.0054 −45
7 0.0054 45
8 0.0054 0

Problem 4.2 Consider a cube with the specified dimension in Fig. 4.112a. For the
assembly in Fig. 4.112b, what will be the dimension and tolerance (1) between the
plane of A1 and A2 and (2) between the plane of B1 and B2 ?
Problem 4.3 Figure 4.113 shows the dimensions of a machined part. Use the AutoDi-
mensionScheme tool in the Solidworks DimXpert to annotate GD&T information
automatically.
Problem 4.4 Figure 4.114 shows the dimensions of a machined part. Use the AutoDi-
mensionScheme tool in DimXpert to create both of plus/minus dimensioning and
geometric dimensioning automatically.
Problem 4.5 For Example 4.6, let L 1 = 100 mm, L 2 = 250 mm, and
H = 150 mm, and assume that clamping forces can be adjusted as
Fx [−2.0 k N , 0 kN]; Fy [−4.0 kN, 0 kN]. Determine the optimized clamping forces
to minimize the maximum displacement of workpiece in operation.
Problem 4.6 For Example 4.6, let L 1 = 70 mm, L 2 = 275 mm, H = 175 mm, F y
= −2.5 kN, F y = − 0.25 kN, and assume that the cutting forces are changed as
4.10 Summary 317

Fig. 4.113 Part model for Problem 4.3

Fig. 4.114 Part model for Problem 4.4

Ft_x [−2.0 kN, 0 kN]; Ft_y [−1.5 kN, 0 kN]. Determine the maximum displacement
of workpiece in the operation.

Problem 4.7 Assume that the materials of part and cutter are low-carbon steel and
high-speed steel (HSS), respectively. The thickness of parts is 0.500 in and the toler-
ance is 0.020 inch for all dimensions, unit. Write the programs for the parts shown in
Fig. 4.115. For each part, write two programs for (1) cutting holes and (2) contouring
milling.

Design Projects
Project 4.1. (1) Create a plastic part you can reach such as the one shown in Fig. 4.116;
(2) use SW Plastics package, select injection locations and filling settings; (3) run
318 4 Computer-Aided Manufacturing (CAM)

(a). (b). (c).

Fig. 4.115 Write CNC programs for parts in Problem 4.2

(a). bucket (b). Hairdry part (c). Stool

Fig. 4.116 Part example for mold filling analysis in Project 4.1

injection simulation to predict injection molding time; (4) predict defects including
air traps, welding lines, and wraps; and (5) document your process and result.
Project 4.2. Create a mold assembly for injection molding of a plastic part such as
a part in Fig. 4.116.
Project 4.3. Find some machined parts such as the examples shown in Fig. 4.117.
Use the SolidWorks integrated system to design its machining processes and create
machining programs for the part by following the steps given below:
(1) Create CAD model by reverse engineering.
(2) Design fixtures for all of machined features.
(3) Create an assembly model for fixturing systems.
(4) Create CNC program(s) for all of machined features.
(5) Simulate the CNC program to obtain basic statistics of machining (number of
setups, machining tools, the number of tools, cost….).
(6) Post-process the CNC programs to convert them into G-code for a generic
three-axis Hass machine tool.
(7) Analyze the part to predict deflection (tolerance).
(8) (Optional for bonus) make the part if a generic three-axis Hass machine tool
is available.
4.10 Summary 319

Fig. 4.117 Examples of machined parts for design Project 4.3

References

Allsup T (2009) How to speel GD&T a new way to learn GD&T. https://www.anidatech.com/How
ToSpellGDT.pdf
Basha VR, Salunke JJ (2013) An advanced exploration of fixture design. Int J Eng Res Appl 6(5),
part 3, pp 30–33
Bi ZM, Hinds B, Jin Y, Gibson R, McToal P (2009) Drilling processes of composites-the state of
the art. In: Drilling of composite materials, Nova Science Publisher, ISBN: 978–1–60741–163–5,
pp137–171
Bi ZM, Mueller D (2016) finite element analysis for diagnosis of fatigue failure of composite
materials in product development. Int J Adv Manuf Technol 87(5):2245–2257
Bi ZM, Zhang WJ (2001) Flexible fixture design and automation: review, issues and future
directions. Int J Product Res 39(13):2867–2894
Cimatron E Inc (2017) CAD/CAM Solution for Mold Making From Quoting to Deliver. https://
www.cimatron.com/SIP_STORAGE/files/0/1760.pdf
Eastman (2019) Medical devices processing guide. https://www.eastman.com/Literature_Center/S/
SPMBS3689.pdf
Kailas SV (2020) Chapter 3 imperfections in solids. https://nptel.ac.in/content/storage2/courses/
112108150/pdf/Lecture_Notes/MLN_03.pdf
Kreculj D, Rasuo B (2018) Impact damage modelling in laminated composite aircraft structures,
sustainable composites for aerospace applications, Woodhead Publishing Series in Composites
Science and Engineering, pp 125–153
MSCsoftware (2020) Making a composite model. https://www.mscsoftware.com/exercise-modules/
making-composite-model
Pilkey WD, Pilkey DF, Bi, ZM (2020) Petersons stress concentration factors, the 4th version,
ISBN-13: 978–1119532514, ISBN-10: 1119532515, Wiley
Science Notes (2020) Periodic table of the elements. https://sciencenotes.org/printable-periodic-
table/
320 4 Computer-Aided Manufacturing (CAM)

Steeves M (2016) Before there was MBD, there was MBD and more! https://techday2016dotcom.
files.wordpress.com/2016/04/16-q2-mi-before-there-was-mbd-there-was-mbd-and-more.pdf
Sun CC (2009) Materials science tetrahedron—useful tool for pharmaceutical research and
development. J Pharm Sci 98(5):167–1687
Tuttle M (2020) Predicting failure of multiangle composite laminate: micromehanics failure analysis
vs macromechanics failure analysis. https://courses.washington.edu/mengr450/LamFailures.pdf
United States International Trade Commission (2002) Tools, dies, and industrial molds: competitive
conditions in the United States and selected foreign markets, Investigation No. 332–435. https://
www.usitc.gov/publications/332/pub3556.pdf
Vorburger TV, Raji J (1990) Surface finish metrology tutorial. https://www.nist.gov/system/files/
documents/calibrations/89-4088.pdf
Wang H, Rong Y, Li H, Shaun P (2010) Computer aided fixture design: recent research and trends.
Comput Aided Des 42:1085–1094
Chapter 5
Computer Integrated Manufacturing
(CIM)

Abstract A manufacturing system is an organization to make products; a manu-

facturing system consists of various functional units for designing, manufacturing,
assembling, transporting, and marketing and sales. This chapter discusses the
design and operation of manufacturing systems; it covers some enabling technolo-
gies including cellular manufacturing, discrete event dynamic simulation, lifecycle
assessment, and cost analysis; and it also discusses how enabling technologies can
be integrated for effective coordination and interaction of hardware and software
systems from different vendors. Moreover, the evaluation of manufacturing systems,
products, and processes is discussed from the perspective of sustainability.

Keywords Group technology (GT) · Cellular manufacturing (CM) · Discrete event

dynamic system (DEDS) · Petric nets · Lifecycle assessment (LCA) · Computer
integrated manufacturing (CIM) · Cost analysis · Design for sustainability

5.1 Introduction

Figure 5.1 shows a description of a manufacturing system. A manufacturing system

is to transform raw materials into finished products through a series of manufacturing
processes. A manufacturing system requires various resources to perform manufac-
turing processes. These resources can be classified into (1) fixed assets such as plants,
machines, tools, fixtures, software tools, and human resources and (2) flow assets as
the system inputs such as raw materials, power supplies, and capital investments at
one end, and finished products, profiles, and the supports in the lifecycles of prod-
ucts at the other end. The manufacturing system implements its business goals by
delivering value-added products to customers. In addition, manufacturing system
operations are more and more influenced by the dynamics, the uncertainties, and the
disturbance of the business environment, the boundary of the manufacturing system
and the business environment becomes vague.
The complexity of a manufacturing system depends on the products to be
manufactured in the system. Due to the growing complexity of modern products,
manufacturing systems become more and more complex due to the increases of
system components and variants, the interactions of system components, and their

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 321
Z. Bi, Practical Guide to Digital Manufacturing,
https://doi.org/10.1007/978-3-030-70304-2_5
322 5 Computer Integrated Manufacturing (CIM)

Business Environment

Inputs Outputs
Dynamics, uncertainty, and disturbance
(Flow assets) (Flow assets)

Materials Products
Orders, A manufacturing system is to transform raw Customers’
predictions materials into final products, through a series of satisfaction
Power, manufacturing operations and processes. Profits
utilities
A manufacturing system implements its economic Employees’
Cash, satisfaction
capital and social objectives by producing and delivering
value-added products to meet customers’ needs. Maintenance,
Supplies services
Standards Wastes,
recycling
Fixed asset˖plants, machines, tools, fixtures,
software tools, human resources, material-
handling systems…

Fig. 5.1 Description of a manufacturing system

dynamic changes over time. Manufacturing systems are generally complex engi-
neered systems. To design and operate a manufacturing system, engineers should
understand basic resource types and their interactions in manufacturing. As an intro-
duction of manufacturing systems, some basic concepts relevant to manufacturing
systems are discussed in this section.

5.1.1 Continuous and Discrete Manufacturing Systems

A continuous manufacturing system transforms raw materials to finished products

continuously through a series of processes. In a continuous system, raw materials
are usually uncountable, which are in a state of powder, gas, or fluid. Manufacturing
processes are continuous transformations such as chemical reactions or the changes
of properties subjected to mechanical, heat, or other types of loads. The materials
are processed continuously in motion as shown in Fig. 5.2. Since manufacturing
processes are followed one after another, no inventory is needed in continuous manu-
facturing. Continuous manufacturing systems are widely used in chemical factories,
food processing plants, and oil refineries. They are also used to produce raw mate-
rials such as metal sheets, rolls, wires, powders, plastic resin, and cement for discrete
manufacturing. The advantages and disadvantages of continuous manufacturing are
summarized in Table 5.1 (Knowledgiate 2017). In this chapter, continuous manufac-
turing systems are not discussed since their layouts are the sequential connections of
processing machineries, which are relatively stable and simple in contrast to discrete
manufacturing systems.
5.1 Introduction 323

Process 1
Raw Materials

Process 2

Process i

Process n
Final Products

Fig. 5.2 Example of a continuous manufacturing system (HIBA 2019)

Table 5.1 Continuous manufacturing: advantages and disadvantages

Advantages Disadvantages
● The product quality is consistent since the product goes through ● Continuous manufacturing is rigid
the same sequence of processes and machineries. in sense that when one machine is
● The production can be easily automated to reduce direct labor, malfunctioned, the whole process
and the system control can be simplified due to sequential is affected.
processes. ● It requires to avoid piling up of
● No inventory is needed for sequence balancing. work or any blockage on the line.
● The need of material handling is minimized due to the set pattern ● Unless the malfunctioned machine
of continuous manufacturing. can be fixed immediately, it will
● Due to high-volume production, the overhead cost per unit can be force the preceding as well as the
reduced since the fixed costs of specialized equipment are shared subsequent stages to be stopped
by a large volume of output. Accordingly, there is quick return on completely.
investment (RoI).

Typical Applications: chemical factories, food processing plants, oil

refineries, and the suppliers of conventional industrial materials.

A discrete manufacturing system is characterized by making distinct products

such as automobiles, computers, cellphones, toys, aircraft, and furniture items. In
a discrete manufacturing system, manufacturing processes perform manufacturing
operations on discrete parts. Each manufacturing process can be begun or termi-
nated, respectively, and it is unnecessary for different machineries to have the same
cycle time of manufacturing processes. A discrete manufacturing system may involve
assembling processes, in which final products are assembled from parts and compo-
nents. A discrete manufacturing system often needs a variety of raw materials to
make different parts and components. For example, building a computer requires
mainboards, CPUs, storages, keywords, monitors, and many other accessories, and
these components are made from different materials or from different suppliers.
324 5 Computer Integrated Manufacturing (CIM)

Figure 5.3 shows an example of discrete manufacturing systems which was built
to manufacture aero-engines (Fang et al. 2020). The differences of continuous and
discrete manufacturing were discussed by many researchers (Pritchett et al. 2000;
Al-Habahbah 2015; Andrew 2019), and Table 5.2 gives a summary of the differences
in throughouts, quality measures, control variables, units, and characteristics.

Fig. 5.3 A discrete manufacturing system for aero-engines (Fang et al. 2020)

Table 5.2 Continuous versus discrete manufacturing

Aspects Continuous Manufacturing Discrete Manufacturing
Measured by attributes such as Measured by types, models, and numbers of
Throughout
weight and volume products
Consistency, concentration, free of
Dimensions, tolerances, surface finishes,
Quality indicator contaminants, and conformance to
free of defects, reliability, and lifespans.
specifications.
Relevant to shaping, assembling, or
Relevant to recipes and formulas
Typical control property enhancement such as position,
such as ingredients, temperature,
variables velocity, path, acceleration, force,
volume flow rate, time, and pressure.
temperature, heat, and power.
Units of measurement Lot, grades, potency, and shelf-life Pieces, counts, and numbers.
Making ‘ stuff ‘such as milk and milk Making ‘ things’ such as cars, furniture, and
Characteristic
powder computers.

Examples

Oil refining Milk production Car production Computer assembly

5.1 Introduction 325

5.1.2 Variety, Quantity, and Quality

Discrete products in a manufacturing system are by variety (V ), quantity (Q), and

quality. The product quantity Q affects the way how the resources, peoples, facilities,
and procedures are organized in a manufacturing system. Accordingly, the produc-
tions in manufacturing systems are classified into 1) low productions (Q = 1–100
units), medium productions (Q = 100–10,000 units), and high productions (Q >
10,000 units).
The product variety V refers to the number of product variants, which have differ-
ences in materials, features, quality, or manufacturing processes; but they are made
in the same manufacturing system. The number product variety V depends on how a
difference is defined for two products. For example, El-Sherbeeny (2016) suggested
classifying the products into soft product varieties and hard product varieties. The
difference of products in the first catalogue can be handled by planning, controlling,
and scheduling; however, different hardware equipment is needed to deal with the
difference of products in the second catalogue. An example of soft product varieties
is a car product family in the same brand; the products have a common platform and
they can be made in the same production line. An example of hard product varieties
is the difference between passenger cars and pickup trucks, different production lines
are used to build cars and trucks, respectively. Finally, product quality describes how
well a product is made.

Quantity (Q)

( , )

Mass Mass production

( , )

Medium

Cellular Manufacturing

Flexible Manufacturing System

Small
Process-focused Manufacturing

Few More Many Large Variants (V)

Fig. 5.4 Product variety and quantity in selecting a system paradigm

326 5 Computer Integrated Manufacturing (CIM)

Figure 5.4 shows that product variety (V ) and quantity (Q) are two criterial factors
in designing a manufacturing system. Assume the total capacity of a manufacturing
system is given, it is impractical to maximize both product variety and quantity;
therefore, some trade-offs must be made in selecting an appropriate paradigm for
a manufacturing system. The products with a low variety and a high quantity are
made in mass production; the products with a high variety and a low quantity are
made in process-focused manufacturing; the products with a medium variety and
medium quantities are made in cellular manufacturing, flexible manufacturing, or
mass customization. Design of a manufacturing system is a complex engineering
problem. To optimize a system for a given design criterion, a design variable should
be changed in a favorable direction (e.g., f (Q, t) and f (V, t)). However, when a
number of conflicting criteria are involved, some methods, such as the Pareto set in
Fig. 5.4, should be deployed to make the trade-offs for overall system performances.

5.1.3 Decoupled Points in Production

A manufacturing system makes final products to meet customer’s needs, and manu-
facturing businesses are driven by customers’ orders. To reduce lead times, parts,
components, or even products can be made before customers’ orders arrive. For
complex products, manufacturers usually customize products at the assembly stage
using pre-made parts and components. Therefore, designing a manufacturing system
requires to decouple a production line into two stages that are for the required busi-
nesses and activities before and after customers’ orders, respectively. Accordingly,
manufacturing systems can be classified into three types, i.e., make-to-stock (MTS),
make-to-order (MTO), and a combination of two (MTO-MTS).
In MTS, products are made before customers’ orders arrive. MTS is a push system
paradigm. It is effective when the products have a large volume, small variety, and the
manufacturing costs are relatively low in comparison with the costs of raw materials.
MTS requires the inventory to store materials, semi-finished and finished products.
It is challenging to determine the right volumes of products in inventory; on the one
hand, a reduced number of premade products reduces the inventory cost; on the other
hand, it exposes the risk of stock out when customers’ orders increase.
In MTO, the production begins when customers’ orders arrive. MTO is a pull
system paradigm. It is effective when products are highly diversified and with low
volumes, and the manufacturing costs are relatively high in comparison with the
costs of raw materials. It is critical to match manufacturing capabilities and product
demands in MTO. On the one hand, reducing production capabilities helps to increase
the utilization rates of manufacturing resources thus will reduce product unit costs.
On the other hand, limited product capabilities expose the risks of losing customers’
orders due to the need for long lead times.
In MTO-MTS, the ‘push’ and the ‘pull’ system paradigms are combined. The
production lines in the manufacturing system are decoupled. The ‘push’ strategy is
adopted for the premade parts or components, and the ‘pull’ strategy is adopted to
5.1 Introduction 327

Fig. 5.5 Comparison of MTO, MTS, and ATO

customize the products to meet customers’ needs. Products can be customized by the
technique so-called assemble to order (ATO). In ATO, different parts or components
are selected and assembled for different products. MTO-MTS is effective to balance
the strategic goals of reducing inventory costs and shortening product lead times
(Kaminsky and Kaya 2009).
Figure 5.5 shows a comparison of MTO, MTS, and ATO in terms of inventory
level, lead time, and the difficulty of planning and scheduling, rigidity, and cost. The
overall performance of a manufacturing system should be optimized by minimizing
the aforementioned five indicators simultaneously. On the one hand, MTS minimizes
product lead times but demands the highest inventory level of products, and MTS
has poor flexibility since the system is planned and controlled based on the long-
term prediction. On the other hand, MTO minimizes the inventory level of products,
and it increases the system flexibility since system planning and controlling is based
on customers’ orders. However, products have the maximum product lead times. In
contrast, ATO splits the supply chain in system planning and scheduling: before the
decoupling point, parts and components are made before customers’ needs arrive;
after the decoupling point, the manufacturing and assembling processes are planned
and scheduled after customers’ orders arrive. The overall performance of the manu-
facturing system can be improved since all five indicators of system performance are
328 5 Computer Integrated Manufacturing (CIM)

Table 5.3 Comparison of MTS, MTO, and ATO at other aspects

Aspect MTO MTS MTO -MTS (ATO)

Data for production Volumes, varieties, and Volumes, varieties, and Customer’ orders and
specifications of products specifications of products configuration
by customers by predication management
Assumption of production Engineering capacities Projected inventory levels Volumes and varieties,
planning lead-times of products
System control Adjust engineering Assure the levels of Make products in
capacities to meet customer services specified lead times
customer’ needs
Sales and operations Predicted demands and Predicted demands for Predicated demands for
performed the designs of designed products and all alternatives in
products and manufacturing processes. configuration
manufacturing processes management
Assumption of master Actual demands Projected demands The combination of
production scheduling prediction and actual
demands
Lead-time of products Begin with design stage Available until next A short lead time for the
and specify delivery time. inventory replenishment assemblies of existing
parts and components into
products.

optimized simultaneously. Table 5.3 shows the difference of MTS, MTO, and ATO
at other aspects (Vollmann et al. 2004; Cruz-Mejia and Vilalta-Perdomo 2018).

5.2 Manufacturing System Architecture

A manufacturing system involves various manufacturing resources that are closely

interacted with each other. To run a manufacturing system in a controlled manner,
the complexity of system must be manageable for the smoothness of business oper-
ations. Enterprise architecture (EA) aims to manage system complexity by defining
the structure and operation of a manufacturing system. An enterprise architecture can
be defined from different aspects such as functions, processes, businesses, informa-
tion, and technology changes. The functional requirements (FRs) of a manufacturing
system architecture include to (Gao 2001).
(1) define the mission, strategies, methods, and functions, and utilize them to plan
and operate the system;
(2) regulate the communications among functional units with standardized vocab-
ularies.
(3) be open to adopt emerging technologies for technological advancement and
upgrading;
(4) ensure the consistency, integrity, availability, promptness, and security of data
sharing and information integration;
(5) seek the solution to enhance system flexibility, adaptability, and efficiency at
an affordable cost;
5.2 Manufacturing System Architecture 329

(6) support resource sharing for a high utilization rate of manufacturing resources
in system;
(7) expend the lifespan of EA by practicing continuous improvement (CI).
Numerous EAs have been proposed for different applications, and the most influ-
ential EAs are (1) Open System Architecture for Computer Integrated Manufacturing
(CIMOSA) from the European CIM Architecture Consortium, (2) GRAI Integrated
Methodology (GRAI-GIM) from the GRAI Laboratory in France, (3) Purdue Enter-
prise Reference Architecture from the Industry-Purdue University Consortium, and
(4) Enterprise Architecture from the National Institute of Standards and Technology
(EA-NIST) (Williams 1994).
As an example, Fig. 5.6 shows the EA-NIST which was used to represent the
organization of a manufacturing system (Wikipedia 2020a). In the EA-NIST, system
elements are divided into layers and domains. The architecture consists of the busi-
ness, information, applications, data, and technology layers. The system has its
boundaries to tell if relevant businesses are internal and external, but the influence
of the business environment on the manufacturing system is reflected in external
discretionary and non-discretionary standards.

External discretionary and non-

discretionary standards and requirements

Organizations,
customers, business
functions, standards
policies, procedures

Information flow,
internal and internal Enterprise
content, format, discretionary and non-
presentations discretionary standards
Feedback and requirements
Automated data
processing, manual
systems, procedures,
logical data structure

Physical databases
design, file structure,
data dictionaries

Computers,
communication
networks, facilities,
security
infrastructures

Fig. 5.6 Enterprise architecture by NIST (EA-NIST)

330 5 Computer Integrated Manufacturing (CIM)

The system complexity can be measured by entropy. Entropy represents the

amount of information for the operations and interactions of system elements. There-
fore, system entropy is evaluated by the number of system elements and the types
and changes of system interactions over time. For system interactions in Fig. 5.6,
the elements at the information-layer are driven by the elements at the business-
layer, and the elements at the application-layer are prescribed by the elements at
the information-layer. The elements at the application-layer are used to define the
elements at the data-layer. Finally, the elements at the data-layer are supported
by the hardware, software, and communication at the technology-layer. The exec-
utive information flow is from the top layer to the bottom layer, and the feedback
information flow is from the bottom layer to the top layer in the opposite direction.
In implementing system architecture, system elements are usually modularized.
In other words, system elements are separable, and the elements at certain layers
and domains are sophisticated in managing, contextualizing, and producing data for
respective tasks independently. EA can benefit from modularization for its continuous
improvement (CI): EA should not be affected by the changes occurring to system
elements at module levels.
In a manufacturing system, the differences in manufacturing operations can be
examined from different aspects. Figure 5.7 shows the classification of manufacturing
businesses from the perspectives of structural layer, information integration, and life-
cycle (SAC 2018). System elements can be distinguished from one to another based
on the layers they are in EA; the layers corresponding to EA-NIST in Fig. 5.6 are the
layers of equipment, controls, workshops, enterprises, and corporations. Information
integration distinguishes system elements based on their roles in dealing with data;
typical activities relevant to data are collection, processing, utilization, communica-
tion, and integration, and the capabilities of system elements can be enhanced by the
information integration for self-sensing, self-learning, self-decision, self-execution,
and self-adaptation; therefore, system elements are classified based on the scope of
data processing into device, interconnection and interworking, information fusion,
incorporation, enterprise, and enterprise alliances. System elements play their roles
at different stages of product lifecycles. Therefore, they can be classified based on
the stages where they execute as design, manufacture and assemblies, logistics, sales
and service, and recycling.

5.3 Production Facilities

Production facilities make physical contacts with materials, parts, components, and
products. As shown in Fig. 5.8, typical production facilities are plants, machine
tools, utility equipment, fixtures, molds, dies and tools, gauge inspecting tools, and
material handling equipment, and system layouts where all manufacturing resources
are arranged as a system.
5.3 Production Facilities 331

Structural Layer

Device

Interconnection

Information
fusion

Enterprise

Enterprise
alliance Lifecycle

Information
Design

Integration
Manufacturing

Logistics

marketing
Sales and

recycling
Services and

Fig. 5.7 Manufacturing system architecture

5.3.1 Machine Tools

Figure 5.9 shows that two important specifications of machine tools are (1) applicable
manufacturing processes and (2) the strategies to pay back the initial costs. Machine
tools are the core facilities to perform manufacturing processes, and machine tools
can be classified based on the types of applicable manufacturing processes such as
power metallurgies, metal sheet forming, injection molding, machining, and non-
conventional machining.
A machine tool usually involves significant investment cost, which should be
recovered from the sales of the products that have a manufacturing process performed
on the machine tool. The more products a machine tool can make, the lower the
product unit cost is, the better the return of investment (ROI) of the machine tool is,
and the higher value of the machine tool contributes to products.
332 5 Computer Integrated Manufacturing (CIM)

System layout
Human Utility
resources equipment

Material- Assembling
handling equipment
Machine tools

Controllers Operators Gauges and

Inspecting
tools
Raw Plants, Finished
Materials workstations, Molds, dies, Products
fixtures and tools

Fig. 5.8 Typical production facilities in manufacturing

Machine tools for turning, milling, drilling,

broaching, planning, gear shaping, hobbling , honing,
threading, sawing, and grinding ……

Metal forming machines for rolling, forgoing,

Types of extruding…
machining processes
Plastic forming machines for injection molding,
thermoforming, blow molding, ……

Non-traditional machines for chemical, electrical,

electronic, physical processes such as Electro-
Types of machine discharge machining (EDM), ……
tools
General purpose machine tools for a wide range of
parts with a low-level productivity

Production machine tools for a limited range of parts

Levels of with a medium-level productivity
productivity
Special purpose machine tools for mass production
and specific products

Single purpose machine tools for highly automated

processes with a high-level of productivity.

Fig. 5.9 Typical types of machine tools

Assume that a machine tool has an initial cost (IC); it is designed to serve for
M product varieties and with the volumes of N i (i = 1, 2… M) for each product
variety. In addition, the added value by the machine tool is vi (i = 1, 2… M). The
total number of products (NP) made by the machine tool is a sum of the volumes of
all product varieties as
5.3 Production Facilities 333

Table 5.4 Classification of machine tools based on ROC models

Types Variety (M) Volume (Ni) Added value (vi) Initial cost (IC)
General purpose machine tools with
High low Low Low
a low-level productivity
Flexible machine tools with a limited
Medium Medium Medium– High High
productivity
Special purpose machine tools with a
Low High Low – High High
high-level productivity
Single purpose machine tools with a
One High Low – High Low-Medium
high-level productivity


M
NP = Ni (5.1)
i=1

Accordingly, the total added value (VP) of the machine tool to these products are


M
VP = Ni · vi (5.2)
i=1

Assume that each part takes an average machining time on the machine tool, the
unit cost (UC) to a product relevant to the machine tool can be estimated as

IC
UC = (5.3)
NP
The ROI of the machine tool is the difference of the total added value (VP) and
the initial cost of


M
RO I = Ni · vi − I C (5.4)
i=1

The design variables in Eq. (5.4) can be used to describe the characteristics of
different machine tools shown in Table 5.4. A machine tool can be one of the following
types, i.e., a general-purpose, flexible, special purpose, or single-purpose machine
tool.

5.3.2 Material Handling Tools

A product has multiple features which are made on different machine tools, and the
product is transported over the production line to receive the services at machine
tools. Material handling (MH) tools are used to move materials, parts, or tools in a
manufacturing system; a material handling system consists of all of manufacturing
resources which are used to keep and transport materials, parts, components, and
334 5 Computer Integrated Manufacturing (CIM)

Fig. 5.10 Main components of a material handling system

products. Figure 5.10 shows the schematic of a material handling system which
consists of the following five types of material handling tools.
(1) Transportation devices are used to transport objects in a manufacturing system.
Parts should be transported to workstations to receive the services over machine
tools. If a machine tool is not immediately available, the part should be
temporarily stored until the machine tool becomes available for the next manu-
facturing process. Common transportation devices are conveyors, cranes, rails,
trucks, and forks. It is also very common that human operators handle materials
manually such as loading/unloading parts at workstations.
(2) Positioning devices are used to position objects for manufacturing operations
such as machining, inspecting, assembling, or storing. Differing from a trans-
portation device to tackle the movement at multiple locations, a positioning
device locates the part at a single workstation.
(3) Loading/unloading devices are used to load or unload parts to machine tools
before or after manufacturing processes. Some machine tools are equipped with
sophisticated loading/unloading devices; while operators or industrial robots
are used to load/unload parts for other machine tools.
(4) Storages are used to keep or buffer parts temporarily in a manufacturing system.
Two common types of storage are carousels and buffers. In many cases, only
storage floor and space are needed to store parts without an additional storage
device. Storages can be integrated into a transportation system as an automated
storage and retrieval system (AS/RS).
(5) Identification and control systems are used to track objects and machine tools
and monitor and control material handling systems. For some simple systems,
parts are located and tracked, and MH devices are operated manually.
5.3 Production Facilities 335

5.3.3 Fixtures, Molds, Dies, and Tools

Fixtures, molds, dies, and tools are the type of production facilities that make direct
contact with parts in manufacturing operations.
Fixtures are used to locate and hold parts in performing manufacturing processes.
A fixturing system should support a smooth and fast transition for a batch of products,
simplify the system setup, and sustain the consistency of product quality (Wikipedia
2020b). Fixturing system designs have been discussed in Sect. 4.8 extensively; and
Fig. 5.11 shows five types of fixture elements: (1) a tool body is used as a frame
to mount all fixture elements together as a system. A tool body should be designed
to minimize the deformations of parts and fixturing system which are subjected
to external loads; (2) a support element supports an object by direct contacts. A
support element can be adjustable or fixed depending on if the contacts should be
adjusted or fixed; (3) a locator aims to position an object when a manufacturing
process is applying on object; locators and supports work together to confine the
motions in all directions; (4) a clamping system is to secure an object in the place
in a manufacturing process. A fixturing system may need other accessories such as
lifting devices, stoppers, and ejectors; these accessories are used in the fixture setup.
Other than fixtures, molds, dies, or tools make direct contacts with parts as well.
The geometry and the relative motion of a mold, die, or tool determines the part
geometry. For example, the geometry of casting is determined by the formed cavity
in a mold, the geometry of metal formed part is determined by a pair of punch and
dire, and the geometry of a machined part is determined by the profile and toolpath
of the cutter.

Fixture Elements

Tool Body Supports Locators Clamps Accessories

Cast bodies Solid supports Internal locators Lever type (1st, Spring-stop
Welded bodies Adjustable Large holes (screwed and 2nd, and 3rd buttons,
Built-up bodies supports doweled, press fit, classes), hinges, spring locating
Drill bushing (Threaded, threaded, and press fit sliders, latches, pins, ejectors,
spring, push plus threaded) screws, swings, lifting devices
types) Pins (round, bullet, plain, hooks, wedges,
Equalizing conical dowel, diamond) toggles, power
supports External locators screws, and non-
Nesting rings mechanical
Vee locators clamps
Fixed-stop locators
Dowel pins (Normal,
split, grooved)
Adjustable stops (lock nut,
locking screw, set screw,
variable torque)

Fig. 5.11 Typical elements used in a fixturing system

336 5 Computer Integrated Manufacturing (CIM)

(a). Assembled gripper (b). exploded assembly (c). Bills of materials (BoM)

Fig. 5.12 Example of assembled products

5.3.4 Facilities for Other Manufacturing Operations

Depending on the complexity of products, a manufacturing system includes other

value-added or non-value-added manufacturing operations processes such as assem-
bling, inspecting, prototyping, and packaging. These manufacturing operations can
also be mechanized or automated in the solutions of intelligent manufacturing (IM).
Products are usually assembled from parts and components. Figure 5.12 shows an
example of an assembled gripper which is assembled from 30 parts with a total of 10
different parts. Assembling facilities are used to fit and join parts and components
together as a new large component or finish product.

5.3.5 Manufacturing System Layouts

A manufacturing plant or factory is an industrial site that is built to run manufacturing

businesses. A plant or factory consists of buildings, machinery, material-handling
tools, capitals, and other resources that are needed to make products. The layout
of a plant or factory refers to the organization of manufacturing resources and the
arrangement of production methods (Kiran 2019).
A layout is designed to maximize the utilization of machinery and minimize non-
value-added activities to reduce the overall cost of system operations. The layout
design of a manufacturing system must take into consideration of the complexity
of products and manufacturing processes. As shown in Fig. 5.13, the layout of a
manufacturing system can be either rigid or flexible. The system elements in a rigid
layout are not reconfigurable; while the system elements in a flexible layout are
configurable in production. Rigid layouts include job shops, flow shops, project shops,
and continuous processes. Flexible layouts include flexible manufacturing systems
(FMSs), cellular manufacturing systems (CM), distributed manufacturing systems
(DMS), and virtual manufacturing systems (VMS).
5.4 Cellular Manufacturing 337

Layout types of
manufacturing systems

Continuous Types of rigid

Job Shop Flow Shop Project Shop layouts
process

Cell Flexible Distributed Virtual Types of flexible

Manufacturing Manufacturing Manufacturing Manufacturing layouts

Fig. 5.13 Layout types of manufacturing systems

5.4 Cellular Manufacturing

Cellular manufacturing (CM) has a flexible system layout, and it is a hybrid solution
of job shops and flow lines. Therefore, CM has the advantages of (1) a job shop
for the flexibility of making a wide scope of products and (2) a flow line for the
high productivity of making products in a production flow. A cellular manufacturing
system (CMS) consists of a number of workcells which are logically linked based on
the sequences of manufacturing processes. Figure 5.14 shows an example of CMS.
In each workcell, machine tools are arranged in a production flow; however, the

Mill Lathe Lathe Mill Mill

Mill Drill CELL 2 Drill

Mill Lathe
Assemly Grind Drill
CELL 1

Lathe Lathe Lathe Mill Drill

Lathe Assembly
CELL 3 Drill

Receiving raw materials

Shipping final products Assemly Grind Grind

Fig. 5.14 Cellular manufacturing layout

338 5 Computer Integrated Manufacturing (CIM)

Table 5.5 Characteristics of cellular manufacturing (Weber 2004)

Advantages Disadvantages
● Cells are designed to make products for a family; it ● It reduces the manufacturing
reduces setup time since no change is needed for flexibility,
machines and tools within cells to process similar parts. ● It is challenging to balance cells.
● With reduced setup times, the amount of work in ● High mixes of low volume
progress can be reduced. production can make cells
● Each part is processed in a single cell which reduces part impractical
travelling distance & time. No effort is wasted to store, ● Job rotation is common in CM but
protect and control materials. it can cause the problems due to
● Single machine can be used to manufacture one or more changes
products in each cell for high machine utilization. ● Hard for operators to adapt cells,
● Lead time can be used due to the reduction of setup there will be some resistance from
time, work in process, and the increase of machine operators since no rest period when
utilization. parts are made
● Manufacturing processes become simple, and incentive ● Easily underestimate training needs
and simplified processes boost worker morale.
Examples: U-shaped, inverted u-shaped, or straight-line cells for making discrete parts

system elements in a workcell can be reconfigured, changed, or retooled for making

different products as long as products are in the same family of products. Manufac-
turing operations such as loading/unloading, tooling changing, and transportations,
can be fully automated or manually assisted. Table 5.5 explains the advantages and
disadvantages of CMSs.

5.4.1 Design of Cellular Manufacturing System

System layouts in Fig. 5.13 were classified into two groups. A rigid layout such
as a job shop, flow shop, project shop, or a continuous process does not involve in
reconfiguration, while the system elements in a flexible layout such as a CMS, FMS,
DM, and VM should be reconfigured over time. The reconfiguration of an FMS, DM,
or VM is mainly at the software side, and it is implemented by control software or
workflow composition (Viriyasitavat et al. 2019a, b). From this perspective, CMS is
unique in sense that both hardware and software reconfigurations are needed to make
different products in one or a few product families. In this section, the formation of
CMS is discussed.
A cellular manufacturing system is designed based on two formations, i.e., part-
family formation and machine-cell formation. In a part-family formation, the parts
are grouped based on the similarities of part geometries and processing requirements.
In a machine-cell formation, different machine tools are grouped to manufacture one
or a few of part families. Note that both the part-family formation and machine-
cell formation are non-deterministic polynomial (NP)-complete problems. Efficient
clustering algorithms are expected to group parts and machines effectively. In the
following section, Group Technologies (GT) is discussed to group parts based on the
similarities of geometries, shapes, and processing requirements.
5.4 Cellular Manufacturing 339

5.4.2 Group Technology (GT)

A system layout aims to optimize the overall performance of the manufacturing

system. CMS is expected to make products in mass customization; therefore, the
most critical tasks in designing a CMS are to identify the similarities of parts and
machines, group parts as part families, and group machines into workcells to improve
the utilizations of manufacturing resources. Here, the similarities of products are
discussed, and group technology (GT) is introduced to identify the similarities of
products.
GT is used to analyze products and manufacturing processes and identify their
similarities to define product families. The outcome of GT is a set of workcells which
are dedicated to make product families. GT is an ideal choice to make the prod-
ucts with medium variants and volumes that were conventionally made in batches.
In contrast to traditional batch productions, GT is able to reduce downtimes and
changeovers of machines. Moreover, if the workcell from GT becomes a long-
term solution to a product family, such a solution can be referred to as a flexible
manufacturing system.
Table 5.6 shows the similarities of products should be explored at all aspects that
affect the selections and operations and organizations of machine tools. Especially the
similarities of design and manufacturing attributes should be taken into consideration.
Depending on the complexity of product variants, GT can be implemented manu-
ally or automatically. Three practical techniques to define product families are visual
inspection, classification of products, and production flow analysis.

5.4.2.1 Visual Inspection

Visual inspection is implemented manually. The materials, features, geometries of

products are visually inspected to identify the similarities and group the products with
the maximum similarities into product families. A product family is a set the products
that exhibit similarities in materials, features, geometries, sizes, and manufacturing
processes.

Table 5.6 Design and manufacturing attributes for similarities

Design Attributes Manufacturing Attributes

■ Major dimensions ■ Major process
■ Length/diameter ratio ■ Operation sequence
■ Basic external shape ■ Batch size
■ Basic internal shape ■ Annual production
■ Material type ■ Machine tools
■ Part function ■ Cutting tools
■ Tolerances ■ Material type
■ Surface finish
340 5 Computer Integrated Manufacturing (CIM)

(1) Milling (2) Turning (3) Forming

(a). Unorganized products (b). Product families

Fig. 5.15 Example of GT based on types of machining processes

Note that the similarities of product variants must be evaluated comprehensively.

A similarity at certain aspects does not quarantine meaningful product families. For
example, it does not make sense to group products with the same geometry but
from different materials as a product family since the machines to process different
materials are different. In a product family, the similarities of products must be
significant enough to identify a set of machines that are applicable to all products. In
addition, product volumes also make the difference in GT; for example, it is unlikely
to find an economic solution of the machines which are able to make the products
with (1) a volume of 1,000,000 units annually for a tolerance of ±0.010 inch and (2)
a volume of 100 units annually for a tolerance of ±0.001 inch in one product family.
To practice visual inspection, engineers should analyze the features of products
to be machined, and understand how these features can be made on what types of
machines; then one or a few of the main criteria are used to identify the similarities
of products. As shown in Fig. 5.15, the types of machined features are prioritized
to group the unorganized products in Fig. 5.15a into three product families from
milling, turning, and forming in Fig. 5.15b.
The machined features are usually the important measures in defining product
families; however, the similarities are also critical if the same set of machines can be
used to make products even though their geometries are fairly different. Figure 5.16a,
b show two examples of product families that exhibit the similarities in geometries
and machining processes, respectively.

5.4.2.2 Product Classification

In product classification, the coding scheme is defined to assign the codes to products,
and the products are analyzed to determine the similarities based on codes. Codes
can be assigned to products manually or automatically; however, a clustering process
to identify product families based on the codes should be performed by computer
programs. One critical task in coding products is to determine what and how the
5.4 Cellular Manufacturing 341

(a). A group of products with similar geometries

(b). A group of dissimilar products from milling operations

Fig. 5.16 Product families with the similarities in geometries and machining processes

features on products are coded since no universal rule exists to a wide scope of prod-
ucts. The common practice is to classify products based on their geometries into
rotational and non-rotational products. Moreover, industrial products are surveyed
to rank the parameters and features of products based on their influences on manu-
facturing processes. Figure 5.17 shows an example of the coding scheme for sheet
metal products where major shapes, materials, and material specifications are highly
ranked (Zeng 2009).
Generally, a coding scheme is required to (1) be flexible to represent and classify
existing and potential future products, (2) be specific enough to identify required
types of machines, and (3) be able to distinguish products by critical manufacturing
attributes including materials, tolerances, and processing types.
When a coding scheme is only used in individual companies, the manufacturing
attributes related to other important criteria should be taken into consideration; for
examples, (1) types and capacities of manufacturing processes, (2) types and number

Fig. 5.17 A coding scheme for sheet metal products

342 5 Computer Integrated Manufacturing (CIM)

Methods of Product
Classification

Hierarchical structure method Chain‐‐ type structure

(monocode) (polycode)
Interpretation of each successive Interpretation of each symbol is
digit depends on the value of the always the same and there is no
preceding digit dependence on previous digits

Mixed‐ code structure –

Combination of
hierarchical and chain‐
type structures

Fig. 5.18 Types of code schemes (Askin and Standridge 1993)

of tool changes, (3) types and number of machine setups, and (4) balance of utiliza-
tion rates of machines in a production line. In manufacturing companies, GT does
not usually lead to a permanent solution; since not all new products can be clas-
sified appropriately by an existing coding scheme. Therefore, the coding system
should be continuously improved to eliminate old products and admit new products
in workcells.
A cumbersome code may demand more resources for data collection and compu-
tation; therefore, a coding scheme should be concise as long as the main design
and manufacturing features of products are represented. Figure 5.18a code scheme
can be hierarchical, chain-type, or hybrid. A hierarchical scheme uses monocode,
a chain-type scheme uses polycode, and a hybrid scheme uses both monocode and
polycode.

5.4.2.3 Monocodes

Monocode is for a hierarchical code scheme, and Fig. 5.19 shows a monocode
example that uses four digits to represent the main attributes of products uses a
tree-like hierarchical structure. Each digit represents a main attribute; for example,
the third digit describes the type of the driving mechanism; it can be one of mechan-
ical, hydraulic, or electrical systems. The fourth digit represents the main shape in
the first branch and the functions of the driving mechanism in the third branch.
In a hierarchical code scheme, the next digit amplifies the information by previous
digits. Therefore, a monocode is capable of distinguishing a total number of (n1 ×
n2 × … × nn ) product variants; note that ni is the number of possible choices at
the i-th digit (i = 1, 2,…n) and n is the number of digits in a monocode. Monocode
5.4 Cellular Manufacturing 343

3211 Rotational
Mechanical 321X
3212 Prismatic

32XX Hydraulic 322X

3231 Power

Electrical 323X 3232 Transmission

3233 Receiving

Fig. 5.19 Monocode for hieratical code structure

is effective to differentiate products in terms of geometries, materials, and sizes.

However, monocode exhibits its limitations in representing the information relevant
to manufacturing processes.
Figure 5.20 shows a code scheme where monocode is used to classify sheet metal
products. The first three digits are for the information of raw materials; the next four
digits are for the attributes of machined features, and the last digit is for the special
requirement.

Example 5.1 A code scheme is defined as shown in Fig. 5.20a, determine the
monocode for the product variant shown in Fig. 5.23b.

Solution. The code scheme in Fig. 5.21a has a tree-like structure with four levels.
The digit at the first level is for an overall geometric shape, i.e., ‘0’ for cylindrical and
‘1’ for block. The digits at the second and third levels are for the dimensional ratios
of main dimensions. The digit at the fourth level is for the tolerance requirements.

1 2 3 4 5 6 7 8 Digital values
Code value depends on chosen
alphabet code
Material
A following digit
Material form
amplifies information of
Material property previous digit; the number
Bending method of product variants are n1
× n2 × …× nn
Tolerance Monocode is not easy to
be understood by humans
Surface finish the meaning of the next
Test method digit depends on the
values of preceding digits
Special requirement

Fig. 5.20 Monocode scheme for sheet metal products

344 5 Computer Integrated Manufacturing (CIM)

Cylindrical Block
0 1

L/D ≤ 1 1<L/D ≤ 2 L/D > 2 A/B <3 A/B ≥ 3

0 1 2 0 1

L/D < 0.5 L/D ≥ 0.5 B/C < 1<3 B/C ≥ 1<3
0 1 0 1
Inside Tolerance
<0.0001 ≥0.0001
0 1

(a). Example of the monocode (b). Drawing of example product

Fig. 5.21 Example of using monocode to distinguish products

By analyzing the attributes of the product in Fig. 5.21b, the digits at four levels of
the code scheme can be determined as,
(1) The first digit is ‘0’ for the cylindrical body,
(2) The second digit is ‘0’ for the ratio of L/D = 1/1.25, which is 0.8 less than 1.0,
(3) The third digit is ‘1’ for the ration of L/D = 0.8, which is 0.8 larger than 0.5,
and
(4) The fourth digit is ‘0’ is for the tolerance of 0.00005 less than 0.0001.
Therefore, the monocode for the product in Fig. 5.21b becomes ‘0010’.

5.4.2.4 Polycodes

A polycode is for chain-type code scheme, and a value at the specific digit of the
code has the same meaning to all product variants no matter what values at other
digits are. A polycode is easy to use but not very efficient since the polycode must
have the digits to represent all possible attributes, while a product only has a part of
these attributes; some digits in a polycode become unnecessary to those products.
Table 5.7 shows an example of polycode scheme. A value at the given digit has the
same meaning as any product variant; therefore, the meanings of a polycode are
easily understood. However, when a product family involves in a large number of
attributes, the polycode becomes very lengthy and excessive.

Table 5.7 Example of polycode scheme

5.4 Cellular Manufacturing 345

Level 1

Level 2 1 2 3 4 5 6 7 8 9 10 11

Level 3 Level 1 Level 2 Level 3

(b). Polycode
(a). Polycode scheme

Fig. 5.22 A polycode scheme and corresponding digital representation

Figure 5.22 shows an example of writing a polycode scheme with three levels into
a one-dimensional (1D) polycode. Since the digit at a certain place has been clearly
explained in the code scheme, a polycode is easily understood the corresponding
code scheme such as Table 5.7 is available.
Example 5.2 Table 5.8 shows the code scheme of Nissan passenger car family, write
out the polycode for the car shown in Fig. 5.23 (Kirby 2019).

Solution. The code scheme in Table 5.8 includes six digits, which are used to repre-
sent type, package, color, interior finish, radio, and tire size. Based on the description
of the car model in Fig. 5.23, the digits assigned to this car model are ‘1 1 3 4 1 1’.
Different code schemes may be used for the same product families. However,
the capacities to represent product variants may be quite different. To compare the
differences of monocode and polycode, it is assumed that both code schemes include

Attribute
Digit Description
1 Type of car 1 for “Maxima”
2 Car package 1 for “GXE”
3 Car color 3 for “Gold”
4 Interior 4 for “Leather”
5 Radio 1 for “AM/FM”
6 Tire size 1 for “15 inch”
(a). Car model (b). Code for car model (a)

Fig. 5.23 Ploycode for a car model example

Table 5.8 Code scheme of Nissan passenger cars

Size Package Color Interior Radio Tire size
(Digit 1) (Digit 2) (Digit 3) (Digit 4) (Digit 5) (Digit 6)
1 for “Maxima” 1 for “GXE” 1 for “White” 1 for “Black” 1 for “AM/FM” 1 for 15”
2 for “Altima” 2 for “XE” 2 for “Black” 2 for “Gray” 2 for “CD” 2 for “17”
3 for “Sentra” 3 for “SE” 3 for “Gold” 3 for “Brown” 3 for “CD changer”
4 for “GLE” 4 for “Blue” 4 for “Leather” 4 for “Premium”
5 for “Red”
4 for “Dark Gray”
346 5 Computer Integrated Manufacturing (CIM)

six digits (i = 6), and each digit

6 has ai set of possible values from 0 to 9. The capacity
of a monocode is found as i=1 10 = 1,111,110; while the capacity of a polycode
is found as the number of product variants in a polycode is 10 × (i) = 60.

5.4.2.5 Hybrid Codes

A hybrid code is for a hybrid code scheme where both monocode and polycode fully
utilize their advantages. Figure 5.24 shows an example of a hybrid code scheme that
integrates hierarchical and chain structures. The structure of a hybrid code can be
customized to represent the attributes of products and manufacturing processes effi-
ciently. The majority of product families are represented by hybrid codes. Popular
hybrid code schemes are Opitz, Brisch System, CODE, CUTPLAN, DCLASS, Multi-
Class, and the Part Analog System (Khan 2013). In the next section, the Opitz is
introduced as an example of hybrid code schemes.
Optiz Code Scheme
Opitz was developed at the Technical University of Aachen in Germany (Haworth
1968). It was widely used as a hybrid code scheme to represent formed and machined
product families. Opitz included the digitals for the information of products and
manufacturing processes. As shown in Table 5.9, the Opitz code scheme includes
three groups of digits, and the meanings of these digits are explained in Table 5.10
in detail.
The code scheme of Opitz is shown in Table 5.11 where the form code is monocode
(digits 1–5) and the supplementary code is polycode (digits 6–9), and Table 5.12
shows the rules to assign values at these digits based on the given product attributes.

Example 5.3 Determine the five-digit Opitz code for the part shown in Fig. 5.25.

Fig. 5.24 Example of hybrid

code for product families

(a). Polycode (b). Monocode (c). Polycode

Table 5.9 Opitz code scheme with three groups of digits

Digits in Opitz Classification
1 2 3 4 5 6 7 8 9 A B C D
Form Code Supplementary Code Secondary Code
5.4 Cellular Manufacturing 347

Table 5.10 The meanings of digits in an Opitz code scheme

Digit Description
1 rotational or non-rotational shapes; rotational shapes are further classified by length-
to-diameter ratios and non-rotational shapes are further classified by length, width, and
thickness.
2 features on external shapes.
3 internally machined features such as holes, threads, and other turning features.
4 machined surfaces such as flats and slots
5 special features such as auxiliary holes, gear teeth and others
6 overall dimensions of products
7 working materials such as steels, aluminums, and irons
8 original shapes of starting raw materials
9 tolerances of manufacturing processes.

Table 5.11 Opitze code scheme

Solution. The attributes and dimensions of the part in Fig. 5.25 are used to assign
the values of the digits in the Opitz code scheme as below.
(1) The overall length (L) and diameter (D) of the part are 1.50 and 2.50,
respectively; therefore, L/D = 1.0 and the first digit becomes ‘1’.
(2) The part is rotational and has the steps at both ends but with one thread;
therefore, the second digit is ‘5’.
(3) The part has an internal shape with a through-hole; therefore, the third digit is
‘1’.
(4) The part has no machined surface, therefore, the fourth digit is ‘0’.
(5) The part has no other holes or gear teeth, the fifth digit is ‘0’.
348 5 Computer Integrated Manufacturing (CIM)

Table 5.12 The rule to generate an opitz code

Therefore, the five-digit Opitz code of the part in Fig. 5.25 becomes ‘15100’.
Other than the first nine digits as the form code and the supplementary code, a
complete Optiz code scheme uses additional four digits to represent other manu-
facturing attributes. Each digit has the options of 10 different values. Therefore,
the Opitz code scheme has the capacity of representing a large number of product
families for a wide scope of applications.
5.4 Cellular Manufacturing 349

Fig. 5.25 Product drawing for Example 5.3

5.4.3 Production Flow Analysis

GT focuses on the classification of products; it would be better to consider products

and productions simultaneously to analyze the similarities of products and machines.
Production flow analysis (PFA) is used to model the correspondence of products and
machines as relational matrices, and the relational matrices can be then analyzed to
define workcells for product families. Each workcell consists of the machines for the
manufacturing processes of one product family.
Instead of using product models in GT, PFA uses the production route sheets to
group products and machines. The products in a family should have similarities in
their production route sheets.
In defining workcells for products and machines, PFA follows the steps as below:
(1) determine the sequence of manufacturing processes and the routes of machines
for products,
(2) group products and machines based on the similarities of the routes of manufac-
turing processes, the grouped products and machines are arranged as “packs”;
each pack is modeled as an incidence matrix for product-machine relations,
(3) cluster the packs into the groups with similar routes, and
(4) define a workcell for each group of the machines.
Two popular algorithms to group products and machines are the single-linkage
clustering (SLC) algorithm (SLCA) and Rank-Order Clustering (ROC) Algorithm.
In the following, ROC by King (1980) is introduced to group products and machines
based on a given product-machine matrix.
Assume that a product-machine relational matrix is given as [M]n×m ; where n and
m are the numbers of products and machines, respectively. ROC is used to sort rows
and columns in [M]n×m in Table 5.13.
350 5 Computer Integrated Manufacturing (CIM)

Table 5.13 Main steps in a rank-order clustering (ROC) algorithm

Example 5.4 Define the workcells for the given product-machine relational matrix
below.
Products
Mij 1 3 4 7 2 5 6 8
A 1 1 1
Machines

E 1 1
C 1 1 1 1
F 1 1
D 1 1 1 1
B 1 1 1

Solution. Step 1: for column i (i = 1, 2,…, n), the binary weight 2n−i is assigned
as (27 , 26 , . . . 21 , 20 , respectively. The decimal weight of Row A becomes 27 (1) +
26 (1) + 25 (0) + 24 (0) + 23 (1) + 22 (0) + 21 (0) + 20 (0) = 200, and those for other
rows are calculated as below:

Step 2: sort the rows in the order of decreasing decimal weight values as A, B, C,
D, E, and F.

Step 3: Repeat step 1 for row j (j = 1, 2,…, m), the binary weight 2m−i is assigned
as (25 , 24 , . . . 21 , 20 , respectively. The decimal weight of Column 1 becomes 25 (1)
5.4 Cellular Manufacturing 351

+ 24 (1) + 23 (0) + 22 (0) + 21 (0) + 20 (0) = 48, and those for other columns are
calculated as below:

The columns are then reordered in decreasing values from left to right as 3, 1, 2,
4, 5, 6, 7, and 8.
Step 4: Repeat steps 1, 2, and 3 to get the final result (no more switch when the
steps are repeated) as.

Finally, three workcells should be defined, the first one consists of machines A,
B, and C, for products 1, 2, and 3; the second one consists of machine C and D for
products 4, 5, and 6, and the third one consists of D, E, and F for products 7 and 8.
Note that ROC may not be able to generate workcells for some product-machine
relational matrices; since it is not uncommon that the iterative process in ROC leads
to an oscillation. This should be solved by introducing more machines of same
types. Another scenario is that the finished clusters have an outlier or void; an outlier
should be addressed by a machine replication, while no action is needed for a void;
the product just skips any operation on the corresponding machine.

5.4.4 Cellular Manufacturing

GT or PFA has been applied to group products as families and sorts the machines as
workcells. Moreover, the manufacturing resources should be well organized to plan,
schedule, and control manufacturing processes efficiently. Cellular manufacturing
is used to serve this purpose. The machines in a workcell should be organized at first,
and it is referred to as a cell formation problem (CFP) in cellular manufacturing.
CFP is generally a nondeterministic polynomial (NP) problem, and the solution to a
CFP is the workcells, which is composed of heterogeneous machines for designated
product families.
352 5 Computer Integrated Manufacturing (CIM)

The required manufacturing capabilities of a workcell are defined based on the

features of a composite product. A composite product can be real or imaginal; it is
modeled as a collection of the primitives and features of all products in a family.
Cellular manufacturing is designed to ensure that any feature on the composite
product can be processed by the machines in the workcell. When a new product is
assigned to the workcell, this product must show similarities with the composite part.
The layout of CM is organized based on the sequence of manufacturing processes
of the composite part.
Figure 5.26 shows an example of a composite product; it is an artificial product
model with the machined features of the products in the family. Figure 5.26a–c show
that the products in the family have the machined features of faces, chamfers, left
shoulders, right shoulders, and slots. The composite product model in Fig. 5.26d has
all of the aforementioned features.
A composite product model is used to select machines and tooling for a workcell.
If a machine or tool is available to make any feature on a composite product, the
workcell has the machines or tools for all the products in the family. Figure 5.27
shows an example of the product families (Fig. 5.27a), which can be derived from
a composite part model. If the workcell is equipped with all manufacturing tools
(Fig. 5.27c) to make the composite product, this workcell can be the solution to the
manufacturing processes of the product family.
The design of cellular manufacturing can be viewed as a multi-objective optimiza-
tion problem where the trade-offs are made among a set of conflicting design criteria.
Strategosinc (2019) described the procedure of designing a cellular manufacturing
system shown in Table 5.14; it begins with the determination of product families,
then follows by the selection of machines and the design of cellular manufacturing
system, and finally, the layout design of the system.
The procedure for the design of a cellular manufacturing system in Table 5.14 is
simplified. In practice, many other factors such as the types, varieties, and volumes
of products must be taken into considerations. By all means, the effectiveness of
cellular manufacturing systems has been proven in manufacturing products with
medium varieties and quantities. Cellular manufacturing can benefit companies to

(a). Product 1 (b). Product 2 (c). Product 3

(Chamber, slot, left shoulder) (Slot, left shoulder, right (Face, chamber, left shoulder)
shoulder)

(d). Composite product

(Chamber, slot, left shoulder, right shoulder, right face)

Fig. 5.26 Merging product features for a composite product

5.4 Cellular Manufacturing 353

(a). Product variants in a family

(b). Composite product with all (c). Toolset for machining of all
manufacturing feature of product family products/ features for a family

Fig. 5.27 Composite product for the determination of machines and tools

improve system performance by reducing transportation, throughout time, lead time,

and work-in-progress (WIP) and increasing the profit margin and quality of products
(Hyer and Wemmerlov 2001).

5.5 Discrete Event Dynamic Systems

Manufacturing systems are typical discrete event dynamic systems (DEDSs) where
state transitions are initiated by events that occur at discrete instants of time. An
event refers to the start or the termination of an activity. Taking a manufacturing
system as an example, an event can be the completion of a manufacturing process,
a breakdown of machine, and the beginning of transportation, and many others.
In addition, the intervals between two events might be deterministic or stochastic.
Therefore, the complexity of planning, scheduling, and controlling a manufacturing
system can be increased exponentially when the numbers of machines, tools, and
parts and the interactions of these manufacturing resources increase. Many theories
and computer-aided techniques have been proposed to model, analyze, and control
DEDSs. In this section, prevalent Petri nets are introduced to model a manufacturing
system at workcell level.
A DEDS consists of various system elements such as machines, tools, and prod-
ucts, the behaviors of these system elements must be controlled in way to optimize the
performance of system, and system performances can be measured by qualitative and
quantifiable indicators. Typical performance indicators of a manufacturing system
include throughput, delivery time, inventory, utilization ratio, and the probabilities
of malfunctions, breakdowns, and deadlocks. In a Petri net model, the behaviors
of system elements are represented by the states and the transitions of states, and
354 5 Computer Integrated Manufacturing (CIM)

Table 5.14 Procedure of designing a cellular manufacturing system (Strategosinc 2019)

Step Main Activities
Step 1. The products are analyzed to form
(1). Obtain forecasts and (2). Obtain current or
product families based on similarities. Each product profiles proposed roungs
product family corresponds to a group of
heterogamous machines. GT or PFA can be used (3). Prepare product
to analyze and group products. profiles and forecast 4). Process charts

The activities at this step yield the answers to

A
some basic questions such as, 1) how to sort (6). Less than 20 Yes (7). Process char for
products as groups? (2) how to achieve better products? each item or routing
utilization rates of machines? (3) is
manufacturing capability lacked or saturated? No Yes

(8). Less than 200 Yes (9). Less than ~20

products? unique routes?

No
(5). Analyze forecasts
(10). Prepare product
and profiles No B profiles and forecast
charts

(11). Code and classify

(12). Determine preliminary A

C part families
B

D
Step 2. Select manufacturing resources for D
product families
The features of products are analyzed to (13). Select preliminary process and
determine manufacturing solutions including machining resources

machines tools, and operations.

(14). Streamline the process with
The activities at this step yield the answers to value stream or process mapping
some basic questions such as 1) how to
manufacture a product? 2) what is the best (15). Estimate the times for setup,
sequence of manufacturing processes for a machine usage, operators, and
product? 3) what machine tools are needed for processing

identified processes? 4) any human operation is

(16). Calculate the requirements for
needed to perform certain process? machines, equipment and human
operators Yes

F (17). Select preliminary

external lot sizes
G

(18). Do the families No (19). Are other

appear viable? processes available?

Yes No
E C

(continued)

the quantifiable indicators can be evaluated based on the dynamic changes of the
properties of the states and the transitions.
Petri nets are featured with their capabilities in modeling concurrency, synchro-
nization, mutual exclusion, conflict behaviors, and the representations of the states
are more completely than analytical models but more structured and abstracted than
simulations. A Petri net provides a graphical representation of a DEDS, which is
indeed a mathematical model resembling the physical system. Moreover, Petri nets
are very easy to be implemented as computer programs for the simulation and control
5.5 Discrete Event Dynamic Systems 355

Table 5.14 (continued)

Step 3. Design a cellular manufacturing E H
system that is a collection of workcells and
accessary equipment such as storages and (20) .identify the required (i)
external containers, (ii) material No
(26). Is it compatible?
transportation equipment. handling, and (iii) product
The activities at this step yield the answers to controls.
Yes
some basic questions such as 1) what are the (27). Identify people
methods to transport materials from cell to cell? balance methods
2) how to balance the workloads on machines? 3) (21). Is it compatible?
No (28). Identify quality
how to plan, schedule, and control production? 4)
Yes G assurance methods
how to manage work-in-process workpieces? 5)
(22). Identify internal
how to assure the quality of products? 6) how to production control
motivate human operators in cellular No
(29). Is it compatible?
manufacturing system? (23). Identify internal lot sizes
Yes
(30). Assign people,
skills, and tasks
(24). Is internal lot size No
less than external?
(31). Identify (i) supervision
F methods and (ii)
Yes
compensation systems
(25). Identify (i) equipment
balance methods and (ii)
internal containers I
(32). Plan of operations

I
Step 4. Design the layout of the cellular
manufacturing system, which is logical follow-up (33). Define SPUs (32). Plan of operations
when the tasks in above three steps are
accomplished thoroughly. A cellular
(34). Rate affinities
manufacturing system includes 1) inter-cell
layouts to arrange the machines and minimize the
inter-cell movements of materials and 2) intra- (36). Plan of workcell
(35). Design space plane
operations
cell layout to arrange workcells and minimize the
transportation overall cost.

The activities at this step yield the answers to

two questions 1) how to optimize system
performance by arranging machines and tools?
(2) how to deal with external constraints and
changes? (3) how to integrate manufacturing
resources as a network?

of DEDSs. A Petri net represents a DEDS as

P N = { P, T, I, O, K, M} (5.5)

where
P: the places for states of resources,
T: the transitions for state changes,
I: the inputs for the places to fire transitions,
O: the outputs for the places after firing transitions, and
K: the capacities of the places, and
M: the number of the tokens in each place.
In this section, a flexible manufacturing system in Fig. 5.28 is shown as an example
of discrete event dynamic system (DEDS) to illustrate how the Petri net can be used
to model and analyze DEDSs.
356 5 Computer Integrated Manufacturing (CIM)

Workstation 1 Workstation 2

Parts/Jobs

Workstation 3

AGV 1

Buffers
Robot

Finished
parts AGV 2

Fig. 5.28 A flexible manufacturing system (FMS) as a DEDS example

An FMS consists of material flow and tool flow, and the components in an FMS
can be generally classified into two groups: resources to provide services and jobs
to receive services. Table 5.15 shows the classification of service resources based on
their primary functions, and Table 5.16 shows the classification of the workstations
based on their accessory functions.
The states and events in the FMS can be modeled by the places and transitions of
a Petri net. A place (P) represents the states of aggregated resources. For example, a
machine at its idle state is a 1D place, a job is receiving the service on a machine is a
two-dimensional (2D) place, a job is in transportation from one machine to another
is a three-dimensional (3D) place.

Table 5.15 Types of service resources in an FMS (Bi et al. 2001; Zhang et al. 2000)
Types Description Examples
Workstation (WS) A device to perform manufacturing Machining, cleaning, inspecting and
process in the material flow or a loading and unloading station in the
device to change the state of tools in material flow;
the tool flow The devices for tool changing and
inspection in the tool flow
Buffer (BF) A device to accommodate parts or Pallets for parts, and central or local tool
tools magazines for tools
Transportation (CD) A device to change the physical Autonomous guided vehicles (AGV), tool
location of parts or tools changer, or other transportation equipment
Holding and Placing (CT) A device to secure parts or tools in Fixtures, pallets, and tool holders.
service
Auxiliary Tooling (AT) A device used to assist completion of Manipulators, inspecting or cleaning tools.
a service

Table 5.16 Classification of workstations

Type 1 2 3 4 5 6 7
Local buffer No No No Yes Yes Yes Yes
Loading/unloading tool No Yes Yes No No Yes Yes
Machining operation Yes No Yes No Yes No Yes
5.5 Discrete Event Dynamic Systems 357

A transition (T) is an event occurring to one or a few system resources; for

example, a job arrives at the workstation, a machine completes one operation, and a
machine is malfunctioned. Transitions in an FMS can be classified into the following
four types,
one dimensional ( P) ⊕ one dimensional P)  two dimensional ( P)
one dimensional ( P) ⊕ two dimensional ( P)  thr ee dimensional ( P)
two dimensional ( P)  two dimensional ( P)
thr ee dimensional ( P)  thr ee dimensional ( P)
In addition, a transition can be active or passive. An active transition is fired from
left to right by the control command, and a passive transition is fired from right
to left when the system responds to the state change of system element. Upon the
satisfaction of the firing conditions, an active transition is fired by a command issued
by the FMS control system. Each active transition corresponds to a decision point
in the control system, the decisions on the selections of parts and manufacturing
resources are usually made based on predefined rules. Table 5.17 gives examples
of typical decision points and rules for decision-making supports in controlling an
FMS.
Figures 5.29 and 5.30 show the Petri net models for the material and tool flows of
the FMS illustrated in Fig. 5.28. The places and transitions in the respective models
are explained in the figures.
A Petri net model can be used to control an FMS or diagnose the malfunctions
involved in the system operation; inappropriate commands such as those to cause the
following scenarios can be identified in real-time operations (Bi et al. 2001).
(1) Illegal command to activate a transition: a transition can be fired only when all
pro-places have the required number of tokens; if a command is issued when
the firing condition is not satisfied, it is an illegal command.
(2) Violation to synchronization: the resources are shared by all transitions in an
FMS, and one resource may be associated with a number of transitions at the
same time; however, the state of the resource must be consistent with all the
transitions.
(3) Violation of designated rules at decision points: at a decision point, the deci-
sion is made to select a resource, part, or tool from a list of the candidates.
The selection is based on the decision-making rules shown in Table 5.17, a
command to fire a transition must be aligned with the specified priority of
manufacturing operations.

Table 5.17 Decision points and decision-making rules

Maximum Minimum
Rules Longest Shortest
number of number of Average Shortest First in
Fixed Random Priority processing processing Lead-time
remained remained used time path first out
Decision points time time
processes processes
Part enters system Yes Yes Yes Yes
Part chooses workstation Yes Yes Yes Yes
Part chooses buffer Yes Yes Yes
AGV chooses part Yes Yes Yes
Robot chooses tool Yes Yes
Machine chooses tool Yes Yes
Tool chooses AGV Yes Yes Yes Yes
Machine chooses part Yes Yes Yes Yes Yes Yes Yes Yes
358 5 Computer Integrated Manufacturing (CIM)

PF W0

t1
LJ T
t1’ LJ, UJ: The operation to load or unload a job
W0: The state where a part is not in system
W1 W1, W2: The state where a part is on a workstation or buffer
t2 PJ1, PJ2: The process where a pallet is picked from a
PJ1 t2 workstation or buffer
t2’
CJ1, CJ2: The operation where a pallet is placed to a
T CJ1 PJ1 workstation or buffer
t6 t2’ TJ1, TJ2: The process where a part is transported to a
TJ2 workstation or buffer
M CJ1 T
t6’ TW2: The state where a part is receiving service
W2 t7 PF, T, B, M: The free state of a fixture, tool, buffer, or worksation
t3 TJ1 t1, t1’: An event for beginning or completion of loading
PJ2 t2, t2’: An event for beginning or completion of picking a
t7’
t3’ M pallet from a workstation
t7’ t7’ t5 t5’ t3, t3’: An event for beginning or completion of picking a
CJ2 TJ1 W1 UJ PF pallet from buffer
B
t4, t4’: An event for beginning or completion of machining
T W0
t5, t5’: An event for beginning or completion of unloading
T t4
t6, t6’: An event for beginning or completion of
TW2 P M transporting a pallet to a buffer
t4’
t7, t7’: An event for beginning or completion of
transporting a pallet to a workstation

(a) Petri net model of the material flow (b) Places and transitions

Fig. 5.29 Petri net of material flow of FMS (Bi et al. 2001)

LU LT, UT: The operation to load or unload a tool

TW0: The state where a tool is not in system
TW1, TW2, The state where a tool is on a workstation, local or
t0 TW0 t7’ TW3: center magazine
PT1, PT2: The process where a tool is picked from a
workstation, local or center magazine
LT t0’ MAN t7 UT GT1, GT2, The operation where a tool is placed to a
GT3: workstation, local or center magazine
ST1, ST2, ST3: The process where a tool is transported to a
PT1 t1 TW1 ST1 t4
R t4’ workstation, local or center magazine
R
t1’ GT3 R, MAN, LU, The state where a tool is available at robot, man,
GT1 t1
t5 MP, HP: workstation, local or center magazine
t7’
LU ST1 PT1 t1’ LU t0, t0’: An event for beginning or completion of loading
PT3 t1, t1’: An event for beginning or completion of loading a
ST2 R t7
MP GT2 GT1 t6 HP tool from workstation
ST3 t6’ t2, t2’: An event for beginning or completion of picking a
t5’ PT2
PT2 tool from a local magazine
GT3 t3, t3’: An event for beginning or completion of picking a
t2’ ST2 PT3
t2 t5 t3’ tool from a center magazine
TW2 TW3
GT2 t5’ t3 t4, t4’: An event for beginning or completion of sending a
t5 MP HP t6’ tool to workstation
t2 t6 t5, t5’: An event for beginning or completion of sending a
ST3 t5’ t2’
ST3 tool to a local magazine
PT2 GT2
t6, t6’: An event for beginning or completion of sending a
tool to a center magazine
t7, t7’: An event for beginning or completion of unloading
R

(a) Petri net model of the tool flow (b) Places and transitions

Fig. 5.30 Petri net of tool flow of FMS (Bi et al. 2001)

(4) Illegal action against malfunctioned resources: when a resource is broken

down, any command to use the malfunctioned resource is illegal.
(5) Detection of deadlock: a deadlock is a scenario where two or more parts are
waiting for the resources, which are possessed by each other. A deadlock occurs
5.5 Discrete Event Dynamic Systems 359

to a pair of parts, which may be the major issue to be addressed in resource-

sharing systems such as DEDSs. A deadlock can be modeled and detected
based on the following concepts.
Definition of a waiting relation. In a Petri net model P N = { P, T , I, O, K , M},
assume t 1 , t 2 ∈ T are two transitions, s0 ∈ P is the preplace of transition t 1 and
the post-place of the transition t 2 , and the token at the place s0 is zero, i.e., t 1→ s0 ,
s0→ t 2 and m(s0 ) = 0. A waiting relation occurs to (t 1, t 2 ) under the mark M when the
number of tokens of other preplaces of t 2 satisfies the firing condition of t 2 except
m
for s0 , and it is denoted as t1 ← t2 and illustrated in Fig. 5.31a.
Definition of a waiting loop. In a Petri net model P N = { P, T , I, O, K , M},
assume there is a set of transitions ∀T n T , where T n = {t1 , t2 , . . . tn }, and the
transitions in T n satisfy the conditions of the waiting relations of,
m m
∃k ∈ (1, n − 1), tk ← tk+1 and tn ← t1 (5.6)

Then, T n is called a waiting loop under the mark M, and it is denoted as OT n .

Figure 5.31b shows a waiting loop with the minimum number of transitions. Note
that s0 ∈ P is the preplace of transition t 1 and the post-place of the transition t 2 , and
s1 ∈ P is the preplace of transition t 2 and the post-place of the transition t 1 the tokens
at s0 or s1 are insufficient to fire t 1 or t 2 .
Definition of deadlock field. In a Petri net model P N = { P, T , I, O, K , M},
assume that there are m waiting loops O T n 1 ,O T n 2 ,…O T n m , the set of m waiting
loops (T d = O T n 1 ∪O T n 2 …O T n m ) are called a deadlock field under mark M. The
concept of deadlock field can be adopted to detect the following cases:
(1) If T d = null, there is no deadlock under the mark M,
(2) If T d null but T d < T , there is a local deadlock under the mark M, and
(3) If T d = T , there is a whole deadlock under the mark M.

Example 5.5 Use the Petri net to model the testbed for gearbox assembly in Fig. 5.31;
the testbed was developed at the National Institute of Standards and Technology
(NIST) to evaluate the grasping capabilities of end-effectors (Falco et al. 2015; Koot-
bally et al. 2016). A series of operations are to (1) open a gearbox kit, (2) move around
parts, and (3) assemble them together as a functional gearbox. The testbed consisted
of robots A and B, vision systems A and B, force sensors, trays and the assembling
platform. Note that opening the gearbox kit and place parts on the platform are done

Fig. 5.31 A deadlock PI,2 PI,1 PI,2 PI,1

occurring to (t 1, t 2 )
t2 t1 t2 1

PO,2 0 PO,1 PO,2 1 0 PO,1

0

(a) Waiting relation (b) Waiting loop

360 5 Computer Integrated Manufacturing (CIM)

Table 5.18 Events for assembling processed of gearboxes (Bi et al. 2020)
Steps Events
1. Open gearbox kit and place parts on tray
2. Move robot B, identify the base part, and transport it to buffer, retract robot B.
3. Move robot A, identify the base part, and transport, and reposition it on assembly platform, retract
robot A.
4. Move robot B, identify the medium gear, and transport it to buffer, retract robot B
5. Move robot A, identify the medium gear, transport, and assembly it on the base; robot B holds the
base, retract robots.
6. Place shaft A on the base part
7. Move robot B, identify the large gear, and transport it to buffer, retract robot B.
8. Move robot A, identify the large gear, transport, and assembly it on the shaft A;
robot B holds the base, retract robots.
9. Place shaft B on the base part
10. Move robot B, identify the small gear, and transport it to buffer, retract robot B
11. Move robot A, identify the small gear, transport, and assembly it on shaft B;
robot B holds the base, retract robots.
12. Move robot B, identify the gearbox cover, and transport it to buffer, retract robot B.
13. Move robot A, identify gearbox cover, transport, and assembly it on sub-assembly;
robot B holds the base, retract robots
14. Robot A picks gearbox assembly, transport, place it to tray, retract robot A.

1st event type 2nd event type 3rd event type

manually. robot A was equipped with a force sensor to assemble parts; while robot
B was equipped with a vision system only to move and hold a part (Bi et al. 2020).

Solution. Table 5.18 shows the possible events which can be classified into three
basic types, i.e., (1) an event initialized by human, (2) an event performed by a robot
and a vision, (3) an event performed by a robot, a vision, and a force sensor.
Figures 5.32, 5.33, and 5.34 show the Petri net examples of three event types,
respectively.
The sub-Petri net models for different events can be selected and assembled as a
testing plan. In executing the testing plan, the states of resources and transitions are
tracked to evaluate the performance of the assembly system.

5.6 Simulation of Discrete Event Dynamic Systems

Computer simulation is the most effective technique for the design and optimization
of DEDSs. This section discusses how to use commercial software tools to simulate
various DEDSs. Numerous software tools, such as Anylogic, Enterprise Dynamics,
Delmia, FlexSim, and Plant Simulation, have been developed to simulate DEDSs
(Wikipedia 2020c). Here, the simulation modeling framework based on intelligent
objects (Simio) software by Simio LLC is used as an example for the simulation of
DEDSs.
5.6 Simulation of Discrete Event Dynamic Systems 361

(b) Available resources

(a) Unpacked gearbox kit (c) Assembled gearbox

Fig. 5.32 A testbed used to assemble a gearbox

P1-1: operator ready to take new task P9-2

P1-2: operator in operation to open kit P12-1
P12-1: gearbox kit is available
P12-2: gearbox kit unavailable T1-1
P9-2: tray is at home with closed kit
P9-3: tray is at home with opened kit
P12-2
P9-5: tray is unavailable P1-1 P1-2
P13-1: base part is available
P14-1: shaft A is available
P15-1: large gear is available T1-2
P16-1: medium gear is available
P17-1: shaft B is available P19-1
P13-1
P18-1: small gear is available
P19-1: cover is available P9-3 P18-1
P14-1
T1-1: Operator starts to open kit P17-1
P15-1 P16-1
T1-2: Operator finishes opening kit

Fig. 5.33 Event 1 (the 1st type): opening a gearbox kit

Simio is a graphic modeling tool that simplifies object representations and provides
the flexibility of defining manufacturing processes without the need for programming
(Simio 2020). Simio mainly uses the object-oriented modeling paradigm to define
objects and processes; however, it supports other modeling paradigms such as event-
based, process-oriented, or agent-based modeling techniques.
362 5 Computer Integrated Manufacturing (CIM)

P2-1
P2-1: robot A ready to move new positon
P2-2: robot A in motion T3-1
P2-3: robot A arrives at given position P2-2 P7-1 P13-1
P2-4: robot A holds at given positon
T3-2 T3-3
P2-5: robot A retract to home positon
P7-1: vision A ready to detect object P2-3 P7-2
P7-2: vision A searches object
P7-3: vision A finds object/position T3-4
P3-1: Robot A tool ready to pick an object P2-1 P7-3 P3-1
P3-2: Robot A tool holds an object
P3-3: An object is placed T3-6 T3-5
P13-1: base part is available
P13-2: base part is unavailable P2-2 P7-1 P3-2
T3-7 T3-14

T3-1: Robot A starts to move to buffer P2-3 P7-2

T3-2: Robot A arrives buffer
T3-8
T3-3: Vision A starts to search base part
T3-4: Vision A finds base part P7-3
P2-1
T3-5: Robot tool A starts to pick base part
T3-6: Robot A starts to move to assembly platform T3-9
T3-7: Robot A arrives assembly platform P2-2
T3-8: Vision A starts to search location for base part
T3-10
T3-14: Vision A finds location for base part
T3-9: Robot A starts to move identified position P2-3
T3-10 Robot A arrives identified position T3-11
T3-11: Robot tool A starts to place base part P2-1 P3-3
T3-12: Robot A stars to retract to home position
T3-12
T3-13: Robot A arrives home position
P2-2
T3-13

P2-3

Fig. 5.34 Event 3 (the 2nd type): transporting a part

5.6.1 Modelling Paradigms

The supports for graphic modeling and animation are essential to computer simula-
tion tools. A graphics-based approach simplifies the operations to model objects and
processes, and an animation-based approach helps users in reviewing, visualizing,
understanding, and verifying system behaviors efficiently.
Manufacturing systems are mostly discrete event dynamic systems (DEDSs).
The changes of DEDS are made by asynchronous and discrete events occurring
to objects over time. Therefore, early simulation tools such as Simulation Program-
ming Language (Simscript) (Markowitz et al. 2020) and GASP (Hooper and Reilly
2020) are event-oriented; a DEDS is modeled as a series of the events that change
system statuses dynamically. Accordingly, a process flow is modeled as a series of the
5.6 Simulation of Discrete Event Dynamic Systems 363

P2-1: robot A ready to move new positon

P2-2: robot A in motion P2-1
P2-3: robot A at given position
P2-4: robot A holds at given positon
T5-1
P2-5: robot A retract to home positon
P7-1: vision A ready to detect object P2-2 P7-1 P16-1
P7-2: vision A searches object/position
P7-3: vision A finds object/position T5-2 T5-3
P3-1: Robot A tool ready to pick an object
P3-2: Robot A tool holds an object P16-2
P2-3 P7-2
P3-3: An object is placed
P16-1: medium gear is available T5-4
P16-2: medium gear is unavailable
P2-1
P7-3 P3-1
P5-1: robot B ready to move new positon
P5-2: robot B in motion
T5-6 T5-5
P5-3: robot B at given position
P5-4: robot B holds at given positon
P5-5: robot B retract to home positon P2-2 P7-1 P3-2
P8-1: vision B ready to detect object
P8-2: vision B searches object/position T5-7 T5-8
P8-3: vision B finds object/position
P6-1: Robot B tool ready to pick an object P2-3 P7-2
P6-2: Robot B tool holds an object
P5-3 P8-1
T5-9
T5-1: Robot A starts to move to buffer
T5-11
T5-2: Robot A arrives buffer P7-3
T5-3: Vision A starts to search medium gear P5-1 P6-1 P2-1
P8-2
T5-4: Vision A finds medium gear T5-13
T5-5: Robot tool A starts to pick medium gear T5-12 T5-14 T5-10

T5-6: Robot A starts to move to assembly platform

P5-2 P8-3 P6-2 P2-2
T5-7: Robot A arrives assembly platform
T5-8: Vision A starts to search location for medium gear T5-15 T5-16
T5-9: Visoin A finds the position
T5-10: Robot A starts to move to identified position P5-3 P2-3
T5-11: Vision B starts to search sub-assembly
T5-17
T5-12: Robot B starts to move to sub-assembly
T5-13: Vision B finds sub-assembly P5-1 P6-1 P2-1
P3-3
T5-14: Robot B tool starts to hold sub-assembly
T5-20 T5-18
T5-15: Robot B arrives assembly platform
T5-16: Robot arrives identified position P5-2 P2-3
T5-17: Robot A and B in assembling
T5-18: Robot A starts to move to home position T5-21 T5-19
T5-19: Robot A arrives home position
P5-3 P2-3
T5-20: Robot B starts to move to home position
T5-21: Robot B arrives home position

Fig. 5.35 Event 5 (the 3rd type): Performing assembling operation

movements occurring to system elements, and each movement changes the statuses
of relevant system elements. From the perspective of processes, an event-based or
process-oriented modeling technique is efficient and flexible in representing discrete
events; however, the details of system elements are more or less ignored.
Simio has incorporated object-oriented modeling (OOM) in the event-based simu-
lation, so that system elements can be defined in detail and the efficiency and flex-
ibility of the modeling process can be preserved. In object-oriented modeling, both
system elements and events can be defined as objects with attributes and states. Note
that a manufacturing system consists of system elements such as machines, opera-
tors, robots, and transportation devices, and running manufacturing businesses corre-
sponds to the events such as machining, assembling, transporting, and delivering;
events are the results of the interactions of system elements.
364 5 Computer Integrated Manufacturing (CIM)

In Simio, any type of system elements, such as a machine, robot, product, tool, and
customer, can be defined as an intelligent object with given attributes and states. The
state change of an intelligent object can be animated graphically for visualization
purpose. Once a new intelligent object is defined, it would be stored in the design
library and can be reused in any other models. Simio includes a design library with
pre-defined objects for users.

5.6.2 Object Types and Classes

Other than the pre-defined objects in the design library, Simio provides the interface
for users to build custom objects, and the common types of Simio objects are listed
in Table 5.19 (Pegden 2020).
Objects in Table 5.19 can further be classified into six types shown in Fig. 5.36
(Thiesing and Pegen 2020). A fixed object has an immobile location in the system;
the examples of fixed objects are stationary equipment such as machines, fueling, and
stations; in Table 5.19, ‘Source’, ‘Sink’, ‘Server’, ‘Combiner’, ‘Separator’, ‘Work-
station’, and ‘Resource’ are fixed objects. An agent object can move in the space
freely. An entity is an instantiation from an agent class which can be moved from
one object to another in the system. Examples of entities are the customers who wait
in front of a service station and the workpieces in a production line. A link or node
object is used to build a network in which entities move around to receive services.
A link refers to a pathway for an entity to be moved from one place to another. A
node corresponds to a starting or ending point in a link. In Table 5.19, ‘Connector’,

Table 5.19 Common types of Simio objects (Pegden 2020)

No. Object Description

1 Source: A place where entities enter the system

2 Sink: A place where entities level the system
3 Server: A service provider with multiple channels and input/output
queues.
4 Combiner: A combiner of entities as batches.
5 Separator: A separator of entities from a batches.
6 Workstation: An operating place with three stages: setup, processing, and
unloading.
7 Resource: An object that can serve other objects.
8 Vehicle: An object that can transport entities
9 Worker: A human operator to move entities.
10 BasicNode: A simple intersection of links.
11 TransferNode: An intersection of links that is set as a destination or waiting
stop on transporter.
12 Connector: A connector of two nodes with no time delay.
13 Path: A path between two nodes where an object travels with a
given speed.
14 TimePath: A path between two nodes with a specified time of traveling.
15 Conveyer: An accumulating or non-accumulating conveyor device.
5.6 Simulation of Discrete Event Dynamic Systems 365

Object Class

Agent Fixed

Entity Link Node

Transporter

Fig. 5.36 Types of Simio objects

‘Path’, ‘TimePath’ and ‘Conveyor’ are link objects, and ‘Basic Node’ and ‘Transfer
Node’ are node objects. A Transporter is a type of ‘entity’ to pick, carry, or drop
object(s). The transportation equipment such as an AGV, crane, forklift, or truck can
be defined as a transporter. In Table 5.19, ‘Vehicle’ and ‘Worker’ are modeled as
transporter objects.
The behaviors of an object are governed by its properties. For example, a ‘source’
object has a property for the inter-arrival time, a ‘server’ object has a property to
specify a processing time. Therefore, the intelligence of an object is modeled as the
collection of the processes controlled by the defined properties.

5.6.3 Types of Intelligence

In defining an object, commonly used properties of an object are defined by clicking

the checkboxes in the properties window. A basic object has some pre-defined events
that are automatically triggered to change the status of object. Objects might need
other properties such as failures, status assignments, additional resource allocations,
financials, or custom process logics, these properties can be added from the fully
exploded list of properties.
To insert an object in a Simio model, the object type is selected, and the object
icon in the library is dragged and placed in the Facility View. The intelligence of
a fixed object is modeled as one or more processes. A process is executed by an
available token, and the token hold by the process is released when the process is
terminated. A process itself can be treated as a sequence of the process steps that
are triggered by events. Process steps are stateless; they have their input properties
but no output or response. This implies that the process steps can be shared by an
366 5 Computer Integrated Manufacturing (CIM)

arbitrary number of objects. If the logic of a process is changed, the change occurs to
any object with such process steps. The states for objects are recorded in elements.
Elements may have both input and output properties.
A process step is a simple process such as holding a token in a time duration,
holding or releasing a resource at a time moment, waiting an event to be activated,
updating state of object, or making a decision of alternative paths. Some process steps
such as ‘delays’ are common to any object including links, entities, transporters,
agents, and groups. Other process steps are customized to certain types of objects
such as ‘pickup’ or ‘drop-off’ steps to a transporter and ‘engage’ and ‘disengage’
steps to a link. Each object class has its own set of events. For example, a link object
usually has the following events (1) an entity enters or leaves the link, (2) an entity
fully merges to the link, and (3) an entity collides or separates from other entities in the
link. By defining the intelligence on objects, processes, and events, the movements
of the entities over the networked resources can be fully controlled.

5.6.4 Case Study

Simio provides many case studies of using Simio to simulate various DEDSs in
aerospace engineering, business processes, supply chains, transportation, and manu-
facturing. Here, the simulation of an assembly workcell has cited an example to
illustrate how a CAE tool such as Simio is used to model and simulate a DEDS
(Simio 2020).
The entities in the workcells are type-A parts, type-B parts, products assembled
from type-A and type-B parts. The objects include two inventories as the sources
of type-A and type-B parts, respectively; one preparation station, two assemble
stations, one repairing station, one packing station, and two docks to ship products.
The process flows of the assembly workcell include (1) a pallet with type-A parts is
picked by a forklift from one of two inventories for that type of parts; (2) a type-A
part is sent to the preparation station; (3) a type-A part and a type-B part is assembled
as one product; (4) a product with a defect is sent and repaired when the product is
identified to have defects; (5) a product is packed; (6) a packed product is transported
randomly to one of two docks for shipment. The simulation aims to predict the number
of packaged products and repaired products over time.
Type-A parts. Type-A parts are picked from two inventories. At the beginning of
the system operation, the pallet with part-A parts is taken from the first inventory;
while the number of pallets in the first inventory is less than 2, a new pallet becomes
randomly available at either of two inventories. This schedule is implemented by
monitoring the number of pallets (NumberInPrep) by a monitor element whose
threshold value is ‘2’ and the crossing direction is ‘negative’. When the state variable
NumberInPrep is less than 2, the monitor element triggers an event to create a type-A
part.
5.6 Simulation of Discrete Event Dynamic Systems 367

Pallets. Each pallet holds 10 type-A parts. The entities on a pallet are emptied at
a ‘separator’ located at the beginning of the continuous assembly line. The type-A
parts are sent by a ‘conveyor’, and a pallet is taken to a storage rack that is modeled
as a ‘sink’.
Preparation Station. The number of type-A parts at the preparation station is tracked
by a user-defined state variable; the number is updated when a part enters or leaves the
preparation station. When a part leaves, the designated assembly station is randomly
selected from two assembly stations.
Assembly Station I and II. Two assembly stations are used to assemble type-A and
type-B parts into products. One product has one type-A part (parent) and one type-B
part (child). The assembly stations are modeled as ‘combiners’. It is assumed that
type-B parts are available by the sources that are linked to assembly stations.
Packaging Station and Repairing Station. 100% of the products from assembly
station II and 70% of the products from assembly station I are directly sent to the
packaging station and 30% of the products from assembly station I are sent to the
repairing station, and they will be sent to the packaging station after they are repaired
by workers.
Shipping. Packed products are randomly transported to two trucks that are modeled
as sinks. The random selection is applied by (1) using ‘random selection goal at the
output of the package station, (2) setting the entity destination type as ‘select from
list’, and (3) setting the ‘node list name’ as a list of input nodes of two trucks.

5.7 Computer Integrated Manufacturing

A manufacturing system involves numerous tasks and operations which require

different information systems from different vendors for decision-making supports.
As shown in Fig. 5.37, computer integrated manufacturing is an integrated solu-
tion of using computers to plan, schedule, and control manufacturing businesses in
product lifecycles (Armagard 2020). The main sub-systems in CIM are computer-
aided design (CAD), computer-aided engineering (CAE), computer-aided manufac-
turing (CAM), computer-aided process planning (CAPP), computer-aided quality
assurance (CAQ), production planning and control (PPC), enterprise resources plan-
ning (ERP), and other high-level business systems such as supply chain management
(SCM), and customer relationship management (CRM).
As shown in Fig. 5.38 (Wikipedia 2020d), CIM serves as the hub of an integrated
solution for an enterprise to automate manufacturing businesses at all levels and
domains in its production system. CIM deals with the integration, coordination,
collaboration, and cooperation of functional units in an enterprise system. Therefore,
three critical requirements of CIM are identified as follows (Wikipedia 2020d):
368 5 Computer Integrated Manufacturing (CIM)

Fig. 5.37 Snapshot of DEDS simulation in Simio environment

Integration of sub-systems from different vendors. Different hardware and soft-

ware systems are developed using different platforms, programming languages, and
standards. For example, automated machines such as robots, AGVs, conveyors,
and CNC use different protocols for communication. CIM must deal with the
heterogeneity of sub-systems in communication and interactions.
Human factors in process control. CIM is an ideal solution for full automation. In
reality, human intervenes are needed to deal with the changes and uncertainties in
both machine operations and decision-making supports. CIM must be able to provide
appropriate interfaces to (1) assist human operators in manufacturing facilities in
the material flow and (2) to allow engineers to tackle the circumstances which are
unpredicted by the automated information systems.
Data integrity. The higher the degree of automation a system is, the more crucial
the data integrity is to system control. When data is exchanged and shared between
two systems, it usually requires some additional human efforts to ensure that there
are proper safeguards for raw data to make effective interactions.
Similar to a modular system, CIM can be viewed as a philosophy in which the system
complexity and uncertainties system are tackled by selecting existing technical solu-
tions and interacting with these solutions in different ways. Therefore, no universal
solution of CIM is available to manufacturing enterprises, especially for small- and
5.7 Computer Integrated Manufacturing 369

Customers
Relations
Product
Services, Design
Maintenance, and
Recycles
Process
Design

Product
Distribution
Computer
Integrated System Design
and Cost
Manufacturing Analysis
Marketing CIM)
and Sales

Purchas Raw
Materials
Quality
Assurances

Product Manufacturing
Assemblies Execution

Fig. 5.38 Computer integrated manufacturing for system integration

medium-sized enterprises (SMEs) that have the limited resources in the investment
of technological acquisition. Here, a case study is presented to show the manual
intervention for data integrity in system integration.
Case Study 5.1. A manufacturer is a tier II supplier of thermoformed products
in the automotive industry. The manufacturer takes the charge of designing and
implementing manufacturing processes for the products that are designed by tier-I
companies. The company needs to obtain product models from clients and using CAE
to analyze the models to verify if all the product features can be manufactured. If all
the product features can be manufactured, CAPP and CAM are used to set up fixtures,
design molds, and run machine tools. Otherwise, the recommended changes are
returned to clients for continuous improvement of products. As shown in Fig. 5.39, the
system integration needs manual efforts for the communication of product models.
The company faces difficulty communicating with clients in data exchanges. As
shown in Fig. 5.40, clients use different CAD tools such as Catia, Unigraphics, Creo,
Solidedge to create their product modules; however, only SolidWorks is available
at the company to review and edit CAD models. Note that product models must
370 5 Computer Integrated Manufacturing (CIM)

Markets, strategies and

enterprise missions
Strategical
plans
Customers’ Simulation,
orders verification,
Production and validation
plans
Results,
Project Performance,
Request to simulation
Proposals Planning costs, and
estimates
capacities

Technical data Enterprise

Programmable
CAD, CAE, information
Intelligent
CAM, CAPP, databases
Controller (PIC),
PPC, TQM…
ERP, SCM…

Monitoring What and when”

Quality
purchases
performance How? What and when?

Production system (material flow)

Products

Fig. 5.39 Computer-integrated manufacturing for system integration (Wikipedia 2020d)

Fig. 5.40 CAD data exchange between a manufacturer and client companies in case study 5.1
5.7 Computer Integrated Manufacturing 371

Table 5.20 List of standard file formats of CAD models

3D file format Type
STL Neutral
OBJ ASCII variant is neutral, binary variant is proprietary
FBX Proprietary
COLLADA Neutral
3DS Proprietary
IGES Neutral
STEP Neutral
VRML/X3D Neutral

be editable for the company to (1) obtain the dimensions relevant to thermoforming
processes, (2) add new features required by thermoforming, (3) utilize product models
to create molds and design fixtures, and (4) run CAE to ensure products can be
manufactured with the required strengths and quality.
A CAD model can be generated in different file formats. Other than generating
its native CAD format, a CAD tool is often compatible to a number of file formats
such as .STL, .OBJ, .STEP, and .IGES; Table 5.20 shows some standard file formats
for CAD models (Sketchfab 2020).
Even with the same geometric information, the CAD models in different file
formats have different contents. If one CAD model is transferred from one software
tool to another software tool, high-level information such as features, design intents,
and constraints will be missed. SolidWorks is used by the company, and SolidWorks
allows to import or export of over 30 different file formats (SolidWorks 2019) as
listed in Table 5.21.
Figure 5.41 shows a comparison of the workflows with and without system inte-
gration. In Fig. 5.41 unicate the design changes with clients, create molds and make
processing plans, and it led to a long lead time, high development time, and more
importantly, missed business opportunities due to a long quote time and the uncer-
tainty factors in quoting. In Fig. 5.40b, engineers are assisted greatly by the Solid-
Works packages in recovering design intents and knowledge from imported CAD

Table 5.21 Compatible file formats in SolidWorks

3D XML Autodesk Inventor IGES*
3DS Autodesk Mechanical JPG
3MF Desktop OBJ
ACIS CADKEY Parasolid
Adobe Illustrator CATIA Graphics ProEngineer
AMP CGR (CATIA Graphics) * Adobe Photoshop
PRC SWG PSD
PDF DXF Rhino
IFC HCG SAT
STEP* Unigraphics/NX* SolidEdge
STL VDA-FS ECAD (IDF, ProStep)
TIFF VRML AEC (IFC format)
372 5 Computer Integrated Manufacturing (CIM)

Part drawings and models

Part drawings and models from customers
from customers
Model conversion
and preparation of
parametric 3D models
Quoting based on design
experiences and rules
Virtual analysis
for manufacturability
for quoting
Is contract No
awarded?
Is contract
Yes awarded? No
Mold design, fixture design, Yes
process planning based on
existing models and experience Mold design, tool design
and design rules and process optimization

Build and test

Are
No
requirements
met?
Are
requirements
No Yes
met?
Build and test

Yes
Implementation Implementation in
production line
in production line

(a) without system integration (b) with system integration

Fig. 5.41 A comparison of workflows with and without system integration

models, communicate the suggested design changes with client companies unambi-
tiously and use the knowledge-based engineering method in CAM and CAPP. It is
expected to reduce the lead time and development cost significantly.
Figure 5.42 shows the dataflow when product models are imported and exported
from the information system at the company, and Table 5.22 shows the main steps to
ensure that the information about features, design intents, and knowledge embedded
in CAD models can be sustained to the maximum extent possible. Note that all of
the computer aide functions in SolidWorks can be fully utilized to accomplish the
tasks from step 2 to step 4.
5.8 Computer-Aided System Evaluation 373

Step 1 Step 2 Step 3 Step 4

SW Geometry Feature Parametric
Translator correction recognition Design

Neutral CAD
formats

Step 5
Neutral CAD formats such as
.step and .iges

Neutral CAD
formats

SW Geometry Feature Parametric

Translator correction recognition Design

Step 1 Step 2 Step 3 Step 4

Importing data flow Exporting data flow

Fig. 5.42 Dataflow for importing or export product models

Table 5.22 Importing and exporting processes in solidworks

Step Task
Use the SolidWorks translator to open a file generated by the CAD package at the client
1
company (such as Catia, Unigraphics/NX, .step, and .iges)
2 Fix, delete, suppress errors on edges, faces, references, and datums to clean solids
Importing
3 Run feature recognition to recover meaningful and correct features
Use the parametric design technique to modify solids, add new features, and conduct
4
engineering analysis
Use the SolidWorks translator to export a file in the format compatible to the CAD
Exporting 5
package at the client company

5.8 Computer-Aided System Evaluation

The role of evaluation metrics in system designs can never be overestimated. An

enterprise system requires the metrics to evaluate the performances of functional
units at different domains and levels.
With an increasing concern on global warming, the scarce of natural resources,
the pollution of eco-environment, enterprise values have been evaluated comprehen-
sively from the perspective of economy, society, and environment, and the scope
of manufacturing businesses has been expanded greatly to sustain manufacturing
systems. As shown in Fig. 5.43, traditional manufacturing processes such as design,
374 5 Computer Integrated Manufacturing (CIM)

Natural Energy
Resources
Customers'
Requirement
s Use
Raw Materials Design Manufacturing Assembly
Lease
Recycled
Use

Environmental
Materials/part
s Reconfiguration
Manufacturing

Impact
Wastes
Collection and
Redesign Reuse Assessment

Recycle Disassembly

Disposal Reverse Manufacturing

Non-recovery materials
Scope of traditional Scope of sustainable
manufacturing manufacturing
businesses businesses

Fig. 5.43 Description of sustainable manufacturing system (Bi 2011)

manufacturing, and assembly used to be economic activities which are evaluated

based on costs and profits; however, the scope of manufacturing processes was
expanded greatly to reuse, recover, redesign, recycle, and remanufacture, and the
impacts of manufacturing processes on economy, society, and environment have to
be taken into consideration in determining enterprise values. Various standards (e.g.
ISO 14,000 and ISO 14,064) have been developed to evaluate system sustainability
and costs. Here, the techniques for computer-aided system evaluation are introduced.

5.8.1 Sustainability of Manufacturing Systems

Figure 5.44 represents the impacts of manufacturing businesses on three aspects of

sustainability, i.e., economy, society, and environment. Manufacturing businesses
are associated with the products at certain phases of product lifecycles, i.e. ‘pre-
manufacture’, ‘manufacture’’, ‘use’’, and ‘post-use’. The impact of a specific manu-
facturing process on three aspects is evaluated individually, and the overall impact
of a manufacturing system on three aspects of sustainability is the sum of those of
all manufacturing businesses in the system. However, it can be seen from Fig. 5.44
that the relations of manufacturing businesses with sustainability are so complicated
to model them in analytical models; it would be more realistic to evaluate system
sustainability using the statistical methods based on existing data. Fortunately, since
5.8 Computer-Aided System Evaluation 375

Fig. 5.44 Evaluation of sustainability in a manufacturing system (Bi 2011)

any manufacturing business is associated with products, the sustainability of a manu-

facturing system can be evaluated statistically based on the products that the system
makes.

5.8.2 Main Indicators

Figure 5.43 showed that manufacturing businesses were expanded to the activities
involved in entire product lifecycles; therefore, system sustainability should be eval-
uated in product lifecycles, and system sustainability is quantified by four main
indicators as below.

5.8.2.1 Carbon Footprint

Carbon dioxide CO2 and other gases with equivalent emissions are generated when
natural resources such as fossil fuels burn in the atmosphere. CO2 and its equiva-
lence affect the average temperature of the earth, and continuous increase of average
temperature causes global warming effects such as loss of glaciers, extreme weather
conditions, and other environmental problems. The carbon footprint is defined to
measure the impact of products on environment in the unit of a ton of carbon dioxide
(CO2 ).
376 5 Computer Integrated Manufacturing (CIM)

5.8.2.2 Total Energy

Total energy is the total amount of direct energy consumption in making products,
the unit of energy is in mega-joules (MJ). The total energy is calculated from (1)
the upstream energy to acquire and process natural energy resources, (2) consumed
energy by releasing or burning natural energy resources, (3) electrical power to make,
transport, and use products. In addition, the efficiencies in energy conversion (e.g.
power, heat, steam) must be taken into consideration.

5.8.2.3 Air Acidification

Air acidification relates to acid rain which is caused by the amount of sulfur dioxide
(SO2 ), nitrous oxide (N2 O), and other acidic emissions. Acid rain is an adverse
environmental factor since it pollutes the land and water where plants and aquatic
animals live; in addition, air acidification dissolves manmade building materials such
as concrete slowly. Air acidification is measured in the unit of kilogram sulfur dioxide
equivalent (SO2 –e).

5.8.2.4 Water Eutrophication

Water eutrophication measures the impact on the water ecosystem due to the
overabundance of nutrients; this depletes the water of oxygen and results in the
deaths of plants and animals. Nitrogen (N) and phosphorous (PO4 ) in fertilizers
are the main sources of water eutrophication. The unit of water eutrophication
is kilogram phosphate equivalent (PO4 –e).

5.8.3 SolidWorks Sustainability

SolidWorks Sustainability allows engineers to estimate the environmental impacts of

products via the lifecycle assessment (LCA) method. LCA is the method to assess
the environmental impact of a product from the preparation of raw materials to
production, distribution, use, disposal, and finally to recycles. LCA supports the
comparison of optional materials in terms of environmental impacts, and LCA is
used to assess system sustainability in terms of (1) the consumed energy and raw
materials, (2) the generated emissions and wastes by manufacturing processes, (3) the
potential environmental impacts, and (4) the options of reducing the environmental
impact.
Figure 5.45 shows the framework of SolidWorks Sustainability. LCA assesses
sustainability from the perspective of carbon footprints, total energy, air acidification,
and water eutrophication, and the sustainability is quantified using the collected
5.8 Computer-Aided System Evaluation 377

Outputs
Inputs
Carbon
Materials SolidWorks Footprint
Product
Manufacturing lifecycle Total Energy
processes assessment Consumption
for
Manufacturing sustainability Air
Regions Acidification
Transportation Water
and uses Eutrophication

Compare similar Set baseline Simulation-based

materials design optimization

Fig. 5.45 SolidWorks Sustainability framework and tools

data of the environmental impacts by various materials in the library. Engineers can
use it to select and compare raw materials since LCA supports simulation-based
optimization.

5.8.3.1 Material Library

The impact of extraction and process of raw materials depends on the specific mate-
rials. Figure 5.45 shows the options of materials in the Material library. The materials
are organized in two levels, the first level in Fig. 5.45a shows the classes of materials
such as steel, iron, rubber, and woods. The second level in Fig. 5.45b shows the types
of materials such as AISI 1020, AISI 304, and alloy steel in the steel class. Note that
the types of materials correspond to the required treatments of materials.

5.8.3.2 Manufacturing Processes and Regions

The impact of manufacturing processes of products depends on the process types

and the geographic locations where the processes are performed. Figure 5.46 shows
the inputs for LCA to determine the impact of manufacturing processes on sustain-
ability. The region input determines the consumed energy in kWh for performing
such manufacturing processes.
378 5 Computer Integrated Manufacturing (CIM)

Material
Library

Classes of
materials
Types of
materials
in a class

(a). Material class

(b). Material type

Fig. 5.46 Classes and types of materials

5.8.3.3 Transportation and Use

Sustainability takes into account energy consumption by some non-value-added

activities, especially, transportation and some activities that are not accomplished in
production such as product uses. In addition, the energy consumptions of both these
factors are affected by the regions where those activities take place. Figure 5.47
shows the interface to input the information for regions, transportation, and use of
products.
Figure 5.48 shows an example of sustainability assessment of a product that
includes (1) the quantified impact on carbon footprint, consumed energy, air acidifi-
cation, and water eutrophication Fig. 5.48a and (2) the economic impact of products
as the aspect of materials, manufacturing, use, and disposal.

5.8.3.4 Material Comparison Tool

The SolidWorks Sustainability supports engineers to compare raw materials from

the perspective of the sustainability of products, a comparison is made between
a baseline material and alternative materials, and Fig. 5.49 shows the interface to
specify a baseline material for product. In comparison, the environmental impact
5.8 Computer-Aided System Evaluation 379

Fig. 5.47 Impact of

manufacturing processes,
regions, and transportation

Manufacturing
Region

Lifespan of
product

Types of
manufacturing
processes

Impact for
each process

Impact for
painting process

can be converted to other quantities that are more familiar to people; for example,
the carbon footprint can be interpreted into miles to drive a car, and this conversion
can be made by using the online tool.
SolidWorks Sustainability is applicable to assembled products. The total envi-
ronmental impact of product is the sum of those of individual parts in the product.
Figure 5.50 shows the assessment of a structure that is assembled from structural
elements. However, LCA of an assembly model is not supported by SolidWorks
Sustainability, not the light version SustainabilityXpress.
380 5 Computer Integrated Manufacturing (CIM)

Product use
region

Transportation
(train, bus, ship,
airplane)

End of product life

(recycle,
incinerated, land)

Fig. 5.48 Impact of transportation and use of product

(a). Environmental impact (b). Economic impact of product

Fig. 5.49 Example result of LCA

5.8 Computer-Aided System Evaluation 381

Fig. 5.50 Interface to

specify a baseline material
Save as Online tool
Set baseline Import baseline

5.8.3.5 Cost Analysis

Enterprises run manufacturing businesses for economic benefits; therefore, economic

factors must be taken into the consideration in designing and operating manufac-
turing systems. In sustainable manufacturing, the costs of a manufacturing system
must include direct and indirect costs, contingent liability, initial capitals, and tangible
costs relevant to environmental impacts. The cost analysis of LCA includes overhead
environmental costs. SolidWorks Costing supports the cost analysis with the consid-
eration of environmental factors. The costing tool can be accessed in the ‘Evaluation’
CommandGroup as shown in Fig. 5.51.
Figures 5.52 and 5.53 show the inputs for a cost analysis of product. In Fig. 5.52,
the processing types and materials are specified, and in Fig. 5.53, the detail of each
manufacturing process is provided. The details include the dimensions of raw mate-
rials and the quantity of products. Figure 5.54 shows an example result of the cost
analysis of a product.

Fig. 5.51 Example of LCA for an assembly model

382 5 Computer Integrated Manufacturing (CIM)

(a). Evaluation (b). Tool for cost analysis

command group

Fig. 5.52 SolidWorks Costing for cost analysis

Process type

Process template

Status of costing
estimation

Fig. 5.53 Inputs of cost analysis—process type and material

5.9 Summary

A manufacturing system involves in numerous activities which are performed by

different hardware and software units from different vendors. Each functional unit
may be optimized for its purpose, but the manufacturing system has system-level
objectives. Therefore, an enterprise system must integrate all the system resources
seamlessly for coordination, collaboration, and cooperation of functional units. From
this chapter, engineers should learn some enabling technologies for organization,
integration, analysis, and evaluation of manufacturing systems. Especially, engineers
5.9 Summary 383

Process details

Quantity of part

Status of costing
estimation

Fig. 5.54 Inputs of cost analysis—process details and quantity

are expected to gain understanding of computer integrated manufacturing (CIM),

cellular manufacturing, discrete event dynamic simulation, LCA and cost analysis
from the perspective of sustainability.
Design Problems

Problem 5.1 Given the product shown in Fig. 5.55, determine nine digits of ‘form
code’ and ‘supplementary code’ using the Opitz coding system.

Fig. 5.55 Example result of

cost analysis of product
Status of cost
estimation

Cost breakdowns
384 5 Computer Integrated Manufacturing (CIM)

Problem 5.2 Use Matlab or other languages to program the raking of clustering
(ROC) algorithm, and use the program to find the machine groups for the following
two machine-part incidence matrices. Alternatively, if you have limited programming
skills, use ROC by hand to find the solution of GT.

(a) case 1 (b) case 2

Problem 5.3 Consider a 6-part-and-9-machine problem shown in the following

table, use ROC to form the part family and machine group.
Product
Mij 1 2 3 4 5 6
A 1 1
B 1 1
C 1 1 1
Machine

D 1 1
E 1 1 1
F 1 1
G 1 1 1 1
H 1 1
I 1 1

Problem 5.4 Consider a 20-part-and-10-machine problem shown in the following

table, use ROC to form the part family and machine group.
Product
Mij 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
A 1 1 1 1
Machine

B 1 1 1 1 1
C 1 1 1 1
D 1 1 1
E 1 1 1

F 1 1 1 1
G 1 1 1 1
H 1 1 1 1 1
I 1 1 1 1 1
J 1 1 1 1

Problem 5.5 Create a composite product for a product family with the following
instances (Fig. 5.56).
5.9 Summary 385

(a). Dimension (b). 3D part view

Fig. 5.56 Product drawings for Problem 5.1

Problem 5.6 Create a similar model to analyze the operation of a 3D printing lab.
The lab is equipped with four 3D printers, i.e., Dreamer-Flashforge, SeeMeCNC,
Dreamel, and Dreamel digiLab. The lab provides 3D services for students in the
courses of ‘Solid Modelling’, ‘CAD/CAM applications’, ‘senior design projects’,
and ‘course projects’ in other mechanical courses. Each student is required to submit
one printing job for each course. The complexity of parts varies from one course
to another. The simulation aims to evaluate if available resources are sufficient to
meet students’ needs; in other words, the utilization rates of the 3D printers will
be evaluataed. Note that the frequency of 3D printing orders from a certain class is
proportional to the average enrolments in each class in a specified period of time.
Therefore, the needs of printing jobs are assumed in Table 5.23 (Fig. 5.57).

Design Project

Project 5.1 Use one of the products you modeled in Chap. 2 (e.g. Problems 2.5, 2.6,
2.7, and 2.8), use the SolidWorks Sustainability and SolidWorks Costing to evaluate
its environmental impact and cost, and modify the materials, processes, materials,
and business regions to reduce the environmental impact and cost.
 386

Table 5.23 Predicted needs of printing jobs for students in different courses at lab
Solid Modelling CAD/CAM Senior Design Other
Courses
(ME160) Applications (ME546) (ME487) Courses
Average enrolments
24 20 15 30% of 150
in semester
Probability (%) 22.86% 19.05% 14.29% 42.86%
Printing Time (hours) (1, 3) (5, 10) (5, 30) (5, 10)
5 Computer Integrated Manufacturing (CIM)
5.9 Summary 387

(a). chamber (b). through hole (c). through hole, (d). through hole (e). slot
step, and chamber and chamber

Fig. 5.57 Product variants for Problem 5.5

References

Andrew L (2019) What is the difference between discrete and process manufacturing? https://www.
sensrtrx.com/difference-between-discrete-and-process-manufacturing/
Al-Habahbah OS (2015) Chapter 5 industrial control systems. https://eacademic.ju.edu.jo/o.hab
ahbeh/My%20Documents/Industrial%20Automation%20By%20Dr%20Osama%20AlHabah
beh/9-%20Ch_5_Industrial%20Control%20Systems.pptx
Armagard (2020) Computer-integrated manufacturing explained clearly. https://www.armagard.
com/ip54/computer-integrated-manufacturing-explained-clearly.html
Askin RG, Standridge CR (1993) Modeling & analysis of manufacturing systems. Wiley. ISBN
0-417-51418-7
Bi ZM (2011) Revisiting system paradigms from the viewpoint of manufacturing sustainability. J
Sustain 3(9):1323–1340
Bi ZM, Zhang WJ, Li Q (2001) A software environment for flexible manufacturing system control
software testing. Proc Inst Mech Eng Part B: J Eng Manuf 215(B):339–352
Bi ZM, Miao ZH, Zhang B, Zhang WJ (2020) A framework for performance assessment of
heterogeneous robotic systems. IEEE Syst J. https://doi.org/10.1109/JSYST.2020.2990892
Cruz-Mejia O, Vilalta-Perdomo E (2018) Merge-in-transit retailing: a micro-business
perspective. Univers Empresa 20(34):83–101. https://doi.org/10.12804/revistas.urosario.edu.co/
empresa/a.5500
El-Sherbeeny am(2016) Introduction and overview of manufacturing. https://fac.ksu.edu.sa/sites/
default/files/1_introduction_sep07_13_ams_2.pdf
Falco J, Wyk KV, Liu S, Carpin S (2015) Facilitating replicable performance measures via
benchmarking and standardized methodologies. IEEE Robot Autom Mag 22(4): 125–136.
Fang W, Guo Y, Liao W, Huang S, Yang N, Lui J (2020) A parallel gated recurrent units (P-GRUs)
network for the shifting lateness bottlenech prediction in making-to-order production system.
Comput Ind Eng 140(2020):106246
GAO (2001) A pratical guid to dederal enterprise architecture. Chief Information Officer Council,
Version 1. https://www.gao.gov/assets/590/588407.pdf
Haworth EA (1968) Group technology—using the Optiz system. Prod Eng 47(1):25–35
Hyer N, Wemmerlov U(2001) Reorganizing the factory: competing through cellular manufacturing,
the 1st version. Productivity Press, Portland, USA, ISBN-10:1563272288.
Hyer N, Wemmerlov U (2002) Reorganizing the factory, competing through cellular manufacturing.
Productivity Press, Portland, OR, 2001. ISBN-10:1563272288
HIBA (2019) Operations management. https://www.slideshare.net/Joanmaines/process-and-layout-
strategies
Hooper J, Reilly K (2020) The GPSS—GASP combined (GGC) system. https://www.semantics
cholar.org/paper/The-GPSS%E2%80%94GASP-combined-(GGC)-system-Hooper-Reilly/ce6
a58307f1fb8052b28762a9db8216ff7b18242
Kaminsky P, Kaya O (2009) Combined make-to order/make-to-stock supply chains, IIE Trans
41(2009): 103–119.
Khan N (2013) Computer integrated manufacturing. https://www.slideshare.net/NoumanKhan2/9-
oct-2013-lec-13-1415161718
388 5 Computer Integrated Manufacturing (CIM)

King JR (1980) Machine-component grouping in production flow analysis: an approach using a

rank order clustering algorithm. Int J Prod Res 18(2):213–232
Kirby K (2019) Focused factories and group technology. https://web.utk.edu/~kkirby/IE527/Ch9.
pdf
Kiran DR (2019) Chapter 18: plant layout, production planning and control, pp 261–278. https://
doi.org/10.1016/B978-0-12-818364-9.00018-4
Knowledgiate Team (2017) Advantages and disadvantages of continuous production
systems. https://www.knowledgiate.com/wp-content/cache/wp-rocket/www.knowledgiate.com/
advantages-disadvantages-continuous-production-system/index.html_gzip
Kootbally Z (2016) Industrial robot capability models for agile manufacturing. Ind Robot, 43(5):
481-494.
Markowitz HM, Hausner B, Karr HW (2020) Simsript—a simulation programming language.
https://www.rand.org/pubs/research_memoranda/RM3310.html
Pegden D (2020) An introduction to Simio for beginners. https://www.simio.com/resources/white-
papers/Introduction-to-Simio/Introduction-to-Simio-for-Beginners.pdf
Pritchett AR, Lee S, Huang D, Goldsman D (2000) Hybrid-system simulation for national airspace
system safety analysis. In: Joines JA, Barton RR, Kang K, Fishwick PA (eds), Proceedings of the
2000 winter simulation conference. https://www.researchgate.net/publication/221526881_Hyb
rid-System_Simulation_for_National_Airspace_System_Safety_Analysis
SAC (2018) Alignment report for reference architectural model for industrie 4.0/ intelligent manu-
facturing system architecture. https://sci40.com/files/assets_sci40.com/img/sci40/Alignment%
20Report%20RAMI.pdf
Simio (2020) The story of Simio. https://www.simio.com/about-simio/
Strategosinc (2019) Design of workcelland micro layouts—cellular manufacturing, https://www.
strategosinc.com/celldesign.htm
Sketchfab (2020) 3D file format. https://help.sketchfab.com/hc/en-us/articles/202508396-3D-File-
Formats
Solidworks (2019) SMG export options. https://help.solidworks.com/2019/english/SolidWorks/sld
works/r_SMG_export_options.htm
Thiesing RM, Pegen CD (2020) Introduction to Simio. https://informs-sim.org/wsc13papers/inc
ludes/files/407.pdf
Viriyasitavat W, Xu L, Bi ZM (2019a) rmSWSpec: real-time monitoring of service workflow
specification language for specification patterns. IEEE Trans Ind Inf 15(7):4021–4032
Viriyasitavat W, Xu L, Bi ZM (2019b) The extension of semantic formalization of service workflow
specification language. IEEE Trans Ind Inf 15(2):741–754
Vollmann TE, Berry WL, Whybark DC, Jacobs RF (2004) manufacturing planning and control
systems for supply chain management, 5th edn. McGrawHill Professional, New York
Weber A (2004) The pros and cons of assembly cells. https://www.assemblymag.com/articles/
83136-the-pros-and-cons-of-cells
Williams TJ (1994) The Purdue enterprise reference architecture and methodology (PERA). https://
citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.194.6112&rep=rep1&type=pdf
Wikipedia (2020a) NIST enterprise architecture model. https://en.wikipedia.org/wiki/NIST_Ente
rprise_Architecture_Model
Wikipedia (2020b) Fixure (tool). https://en.wikipedia.org/wiki/Fixture_(tool).
Wikipedia (2020c) List of discrete event simulation software. https://en.wikipedia.org/wiki/
List_of_discrete_event_simulation_software
Wikipedia (2020d) Computer integrated manufacturing. https://en.wikipedia.org/wiki/Computer-
integrated_manufacturing
Zeng BC (2009) Group technology (GT) in manufacturing. https://www.me.nchu.edu.tw/lab/CIM/
www/courses/Computer%20Integrated%20Manufacturing.htm
Zhang WJ, Li Q, Bi ZM, Zha XF (2000) A generic petri net model for flexible manufacturing
systems and its use for FMS control software testing. Int J Prod Res 38(5):1109–1132
Chapter 6
Digital Manufacturing (DM)

Abstract This chapter discusses the applications of digital technologies in manu-

facturing systems: (1) the main functional requirements are discussed to construct
enterprise system architecture; (2) a new system architecture is proposed for the
applications of digital manufacturing; (3) two examples of digital technologies, i.e.,
reverse engineering (RE) and direct manufacturing (DM) are discussed in details;
and (4) some case studies are briefed to show the diversified applied researches in
using digital technologies in manufacturing.

Keywords Digital manufacturing (DM) · Digital twin (DT) · Reverse engineering

(RE) · Additive manufacturing (AM) · Direct manufacturing (DM) ·
Cyber-physical systems (CPS) · Cloud computing (CM)

6.1 Introduction

A digital twin (DT) refers to a computer representation of (1) potential and physical
assets, devices, peoples, places, processes, and systems and (2) the dynamics of how
the networked things operate and interact over time. The computer representation
for a DT focuses on the connections of physical and virtual models and how both
models are interacted to optimize the performances of physical systems. Therefore,
DT replies greatly on the cutting-edge information technologies (IT) such as Internet
of Things (IoT), artificial intelligence (AI), machine learning, cyber-physical systems
(CBSs), and big data analytics (BDA). The state of a virtual model in a DT system
is updated continuously based on the real-time data feedback from its physical twin
(Wikipedia 2020; Parrott and Warshaw 2020).
This chapter focuses on the application of digital technologies in manufacturing.
The idea of DT was proposed by Gelernter (1991) that the physical world was
modeled, simulated, and evaluated in the virtual reality. The DT concept was coined
by Grieves 11 years later, and it was introduced as an underlying principle for product
lifecycle management (PLM) (Geieves and Vickers 2006; Catapult 2018). However,
due to the high requirement of computing and network capabilities, early DT tech-
nologies were not prevalent tools until the General Electric (GE) managed to use DT
in designing and manufacturing aircraft structures, vehicles, and turbines (Glaessgen

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 389
Z. Bi, Practical Guide to Digital Manufacturing,
https://doi.org/10.1007/978-3-030-70304-2_6
390 6 Digital Manufacturing (DM)

and Stargel 2012; Tuegel et al. 2011; GE 2020). In this chapter, firstly, the functional
requirements (FRs) of modern manufacturing systems are discussed. Secondly, DT
is introduced as the solution to design and operation of manufacturing systems, and
DT architecture is presented to illustrate main system components and their relations.
Thirdly, the system architecture of DT is introduced and some critical enabling tech-
nologies are discussed. Fourthly, reverse engineering (RE) and additive manufac-
turing (AM) are introduced as the examples of computer implementation of digital
technologies. Finally, a number of case studies have been presented to illustrate the
ongoing applied studies in digital technologies.

6.2 Functional Requirements (FRs) of Digital Twin

In a complex manufacturing system, different information systems are used to support

the decision-making supports at different domains and stages such as designs, produc-
tions, assemblies, and marketing and sales. These information systems have to be
integrated as a cross-functional enterprise system to share data seamlessly, coordinate
business processes, and plan and schedule enterprise resources. DT is an implemen-
tation of an enterprise system; therefore, DT relies on various digital manufacturing
technologies.
Digital manufacturing technologies are the collection of computer-aided tools for
modeling, simulation, visualization, and optimization of products, processes, and
manufacturing systems. Early digital technologies, such as computer-aided design
(CAD), computer-aided manufacturing (CAM), product data management (PDM),
and total quality Management (TQM), are stand-alone systems; however, modern
enterprises prioritize the needs of integration, coordination, collaboration, and inter-
operation in a digital twin (Siemens 2020). The applications of digital technologies
have been widely explored in manufacturing industry. In fact, the industry evolution
is driven by the emerging digital technologies. For example, the fourth industrial
revolution (i.e., Industry 4.0) aims at the data transparency, interoperability, predic-
tivity, and adaptability via newly developed technologies such as the Internet of
Things (IoT), blockchain technology (BCT), and big data analytics (BDA) (Berger
2019).
Materials in a physical world are similar to the data in a virtual world. All of the
functional units in a physical manufacturing system are used to transfer raw materials
into finish products; similarly, all of the functional modules in virtual manufacturing
system (i.e., DT) are used to generate, collect, process, distribute, transfer, share,
mine, or utilize data. As shown in Fig. 6.1 (Bi and Wang 2020), with an increasing
complexity of a physical manufacturing system, the volume of data in DT will be
increased exponentially along the lifecycle of products. Other than its volume, the
data in DT can also be characterized by high heterogeneity and diversity since the
data can be sensed signals, events, methods, plans, commands, knowledge, rules,
or models for different manufacturing resources at different spatial and temporal
schemes.
6.2 Functional Requirements (FRs) of Digital Twin 391

Machine
tools

Auxiliary
Raw tools Data sharing, integration, and utilization over product lifecycle
materials

Qualification, After-sale
Manufacturing service and Recycle,
testing, and
processes maintenance redesign,
Assembly validation Marketing reuse…
processes and sales

Design
Data
Information
Decisions
Values

Fig. 6.1 The data growth in product lifecycle

In addition, the functions of DT are far beyond the data collection and sharing.
A digital environment may suffer from a data-rich but information-poor (DRIP)
syndrome. When the volume, velocity, and variety of data from design, manufacture,
assembly, and production are increased exponentially, the system becomes incapable
of processing and utilizing data for decision-making supports. Therefore, it is essen-
tial for DT to incorporate advanced digital technologies to process, utilize, and mine
data to reach right decision for right physical count-partners promptly. The digital
technologies in DT are expected to fulfill the main functional requirements (FRs) as
follows.

6.2.1 Data Availability, Accessibility, and Transparency

An information system counts on correct, reliable, abundant, and updated data. DT

must have the information infrastructure to provide right data at right time to right
place whenever the data is needed. All manufacturing resources at different levels
and domains should be modeled in the digital world to ensure the accessibility,
availability, and transparency of real-time data (Mäkiö-Marusik et al. 2019). For
example, when a CAM package is used to program a machining process, the part
CAD model must be available since the data of the machined features on the part
determines the types of machines and tools, cutting paths, and processing parame-
ters. Moreover, virtual models of physical systems make it possible to (1) integrate
external and internal manufacturing resources, (2) reconfigure systems when an enter-
prise’s mission is adjusted, and (3) incorporate new manufacturing technologies in
an existing information infrastructure (Buttner and Muller 2018).
392 6 Digital Manufacturing (DM)

6.2.2 Integration

Many technological solutions may be available to one functional unit, and an enter-
prise system usually consists of technological solutions from different vendors. As a
customized and holistic system, an enterprise system must integrate all of functional
modules seamlessly to support the interaction, interoperation, coordination, and
collaboration of physical systems. Therefore, system integration includes the integra-
tions of data, platforms, business processes, and hardware and software applications
at various levels and domains (Fenner 2003).
(1) Data integration. In an enterprise system, right decisions rely on reliable and
abundant data. Data integration ensures the data availability, so that it can be
utilized in various business processes in product lifecycles (Tao et al. 2019).
The data in an enterprise system must be shared, maintained, and integrated
through the network. When it is needed, data should be distributed and accessed
across different applications. Some mechanisms that are used to support data
integration include component object model (COM), distributed component
object model (DCOM), common object request broker architecture (CORBA),
enterprise date integration (EDI), Java remote method invocation (JavaRMI),
and extensible markup language (XML).
(2) Integration of platforms. A manufacturing system consists of different busi-
ness units for purchasing, designing, manufacturing, assembling, logistics,
human resources, marketing, and sales and an enterprise system integrates
the processes, software, and tools from different vendors and platforms.
These system resources must be integrated to support the interactions and
interoperation in the heterogeneous environment safely and efficiently.
(3) Integration of business processes. In an enterprise, the required business
processes are defined, and the methods for data exchanges of business processes
are specified. Business process integration includes process management,
process modeling, and process workflow. The solution to the integration
includes the tools to integrate the procedures, organizations, inputs and outputs,
and required tools. The integration of business processes will streamline
business operations.
(4) Integration of applications. Applications in an enterprise are implemented as
black boxes; however, applications have their interfaces to be integrated with
other applications for high-level system objectives. Both internal and external
applications must be considered. For example, an enterprise’s relations with
the manufacturing business environment must be taken into consideration; a
customer relationship management (CRM) system can be integrated with the
backend applications of the enterprise to support the interoperations cross the
boundaries of enterprises.
6.2 Functional Requirements (FRs) of Digital Twin 393

6.2.3 Coordination, Collaboration, and Cooperation

Functional units in the material flow of a manufacturing system are locally optimized.
When low-level units are integrated for system-level objectives, the enterprise system
must be able to provide decision-making supports on the coordination, collaboration,
and cooperation of system elements to optimize system performances. Convention-
ally, a steady hierarchical structure is used to organize system elements in layers
and domains, so that the scope of each decision-making task becomes manageable.
However, a steady structure is impractical when the business environment is highly
turbulent, and the boundaries of manufacturing businesses or the environment are
vague and dynamic.

6.2.4 Decentralization

Manufacturing businesses to produce parts, assemble parts into components and

products, and distribute products to customers are operated in a geographically
distributed environment. To achieve the flexibility and adaptability of the manufac-
turing system, decision-making on organizing, planning, scheduling, and controlling
of business processes are decentralized. Therefore, DT must support the decentral-
ized controls of manufacturing processes so that these businesses can be operated at
geographically distributed locations through the material and information flows.

6.2.5 Reconfigurability, Modularity, and Composability

In a dynamic environment, the optimization of a manufacturing system must take

sustainability into consideration; a manufacturing system must be reconfigurable
to accommodate the changes and uncertainties over time. Reconfigurability is the
capability of changing system configurations to meet new needs. Reconfigurability
and flexibility deal with the changes in different ways in sense that both of hardware
and software components are changed in a reconfigurable system, while only the
software components are changed in a flexible system. Therefore, reconfigurability
can be further characterized by modularity, compatibility, universality, mobility, and
scalability (Wiendahl et al. 2007; Buttner and Muller 2018).
Modularity is implemented by selecting different system elements, and assem-
bling them in different manners to build a system configuration for a set of specific
task requirements; a set of system elements can serve for different processes and
products. A modular system is a collection of independent, reusable system elements.
Composability is a measure for the capability of system elements to be connected
with others. Taking an example of one computer-aided design (CAD) package in DT,
its composability is measured by the number of compatible computer-aided tools that
can import and export data from the package directly.
394 6 Digital Manufacturing (DM)

6.2.6 Resiliency

Any system element has the probability of failure; in addition, an integrated manufac-
turing system is vulnerable to the uncertainties, changes, and disruptions occurring
to its supply chains. Resiliency is the measure of the system capability to sustain
system operations and recover the system from an abnormal state to a normal state.
The resiliency relates to adaptability, agility, redundancy, and learning capability of a
manufacturing (Kusiak 2019). When the manufacturing environment is dynamic, DT
should be resilient to cope with the unexpected disruptions in the dynamic business
environment.

6.2.7 Sustainability

Sustainability is the measure of the long-term survivability of a manufacturing

system. The sustainability is evaluated at the three aspects, i.e., economy, society, and
environment. The impact of taking the sustainability as one of system FRs has been
discussed in depth (Bi 2011). Simply speaking, this would greatly expand the scope
of manufacturing businesses since the activities in the whole lifecycles of products
have to be covered. DT should have the enabling technologies to provide decision-
making supports on expanded manufacturing processes; in addition, DT should be
able to evaluate system sustainability from the perspective of economy, society, and
environment (Gregori et al. 2017).

6.3 System Architecture

Making complex products involves numerous direct manufacturing processes and

indirect business operations; DT provides a platform to network and integrate all of
the computer-aided tools to support these manufacturing businesses. However, any
system resource has its limitation in dealing with the scale, the complexity, and the
uncertainty of system; system architecture aims to maximize the system capability
in dealing with the complexity and uncertainties of system. In the digital world, the
complexity of a system can be measured by entropy. Entropy is a thermodynamic
property; it refers to an amount of system energy that is not yet transferred into
works. The Shannon entropy is used to quantify the randomness, the disorder, and
the complexity of system (Sönmez and Koç 2015) as

  
1
H (X ) = p(x)log2 (6.1)
x∈χ
p(x)
6.3 System Architecture 395

where H(X) is the Shannon entropy for the measure of uncertainty or information; χ
is the set of all possible events in a system; p(x) is the probability of event x where
x ∈ χ ; and log2 is a logical operation over the probability p(x).
The higher H(X) is, the more complex the manufacturing system is, and the larger
amount of information or uncertainty the manufacturing system has. Equation (6.1)
shows that the system entropy is mainly determined by two factors. The first factor
is the number of possible events. A manufacturing system has many types of events,
for example, the interaction of two or more system elements is an event, the case
when a machine is malfunctioned is an event, and the state change of system element
is an event. The system entropy increases when the number of system elements is
increased. The second factor is the probability of having a certain event in a system.
The higher possibility of occurrence of an event, the less uncertainty or a lower
entropy this event contributes to the system.
Conventionally, system architecture is introduced to reduce the complexity and
uncertainty of manufacturing systems. As we discussed in Chap. 5, many system
architectures have been proposed. For example, some popular system architec-
tures include Open System Architecture for Computer-Integrated Manufacturing
(CIMOS), GRAI-Integrated Methodology (GRAI-GIM), Purdue Enterprise Refer-
ence Architecture (PEPA), and Enterprise Architecture by the National Institute
of Standards and Technology (EA-NIST) (Williams 1994). However, adopting a
conventional system architecture in enterprises has exposed the following major
drawbacks:
(1) Hierarchical or grid architecture may cause the delays for system elements to
respond to the changes. Taking an example of a hierarchical architecture, the
directions of the communication for plans, schedules, controls, and executions
are specified from high-level functional units to low-level functional units in
a sequence, and the directions of the communications for data collections and
feedbacks are specified from low-level to high-level functional units. Moreover,
the functional units at different levels are updated in different time durations;
this scarifies the flexibility of low-level functional units to accommodate the
changes promptly.
(2) System architecture represents system elements and their interactions; it actu-
ally brings some artificial barriers to the direct communications of system
elements at different levels or domains even though all system elements are
networked.
(3) To develop system architecture, the boundary for system elements and manu-
facturing businesses is clarified. However, the system boundary becomes
ambiguous when the enterprise needs to collaborate with business partners;
in addition, an enterprise has to adjust its business spectrum continuously to
adopt to the dynamic changes. Existing system architectures are not effective
when an enterprise deals with the changes and dynamics in the environment.
(4) System architecture is designed for given products and manufacturing
processes within a period. As shown in Fig. 6.2, it is evolved along with the
changes of the business environment continuously. System architecture will
396 6 Digital Manufacturing (DM)

Development of new enterprise architecture (EA)

Lifecycle of Lifecycle of
nd
the 1st EA the 2 EA

Continuous
improvement (CI)

Fig. 6.2 Evolution of system architecture in its lifecycle

be terminated when continuous improvement (CI) is not able to fill the gap of
system architecture and the mission of enterprises. This leads to an abrupt shift
to new system architecture.
Conventional system architectures reduce the complexity and uncertainties by
regulating the interactions of system elements by standards and rules. This limits the
capabilities of a manufacturing system at the aspects of the delays caused by indirect
communications, artificial barriers to respond changes, and discontinuities of updates
of enterprise systems. Existing system architectures were able to manage system
complexity by sacrificing system flexibility to certain degree. They are effective
only when the business environment is relatively steady.
The higher the system entropy is, the higher is uncertainty level of information,
and more possible interactions occur to system. From this perspective, a networked
but free system architecture can support the interactions of system elements to the
greatest extent. It enhances the flexibility and adaptability for a manufacturing system
to deal with uncertainties and changes promptly. Figure 6.3 shows a free architecture
of a digital twin (Berman and Bell 2007; Pati and Bandyopadhyay 2017). All of
functional units in the digital world are networked, and they can communicate and
interact directly with each other; this offers the flexibility, adaptability, and agility
to the physical twin. Based on their roles in system, digital tools can be classified
into five types, i.e., digital engineering, manufacturing operations, digital customer
engagement, business intelligence (BI), and digital technology platform.
(1) Digital engineering deals with engineering problems involved in the material
flow such as designs of products and manufacturing processes and selections
of manufacturing equipment. The solution to an engineering problem must
be verified in the digital world before it is released for implementation. The
computer-aided technologies discussed in previous chapters such as CAD,
CAE, CAM, and DEDS simulation are used for digital engineering.
(2) Manufacturing operations refer to the executions of manufacturing processes
such as the value-added processes such as metal forming, material removing
processes, injecting molding, and additive manufacturing and non-value-added
6.3 System Architecture 397

Fig. 6.3 Free system architecture of DT

processes such as inspecting and transportation. The executions of manufac-

turing operations are the outcomes of the direct interactions of digital and phys-
ical twins; the performances of a manufacturing system such as cost, delivery
time, and productivity depend on the enabling technologies for manufacturing
operations. The examples of the digital tools to support the decision-making
activities of manufacturing operations are FMS, MRP-I, MRP-II, MES, and
SCM.
(3) Customer engagement deals with the relations to customers since the busi-
nesses in an enterprise are driven by customers’ needs. Products must be
398 6 Digital Manufacturing (DM)

designed and manufactured to meet customers’ needs to the great extent.

Customer relation management (CRM) is used to integrate an enterprise with
suppliers and end users in the distributed environment through supply chains.
In addition, information and communication technologies (ICTs) allow end
users to be fully engaged in product lifecycle from the product design stages.
(4) Business intelligence deals with the strategical changes of enterprises for the
sustainability in the competitive environment. When everything is networked,
an enterprise is capable of fully utilizing the external resources over the Internet
to catch emerging businesses, and the enterprise can enhance its flexibility and
adaptability by forming workflows with business partners dynamics. Tradi-
tional digital tools for business intelligence are return of investment (ROI) and
quality function deployment (DFD).
(5) Digital technology platforms serve as the information infrastructure to inte-
grate heterogeneous digital technologies for various applications in the phys-
ical world. A number of advanced technology platforms have been developed
recently. For example, the Industrial Internet of Things (IIoT) supports both
machine-to-machine (MTM) and machine-to-man communications; Social,
Mobility, Analytics and Cloud (SMAC) provides the technology platform for
the implementation of digital strategies in organizations.

6.4 Example of Digital Engineering—Reverse Engineering

Figure 6.3 shows that so many digital technologies are applied in digital twins;
the introduction of all these technologies is certainly beyond the limit of coverage.
Some digital technologies such as CAD, CAE, CAM, and DEDS have been intro-
duced in previous chapters; in Sects. 6.4 and 6.5, reverse engineering (RE) and
direct manufacturing (DM) are discussed as the examples of digital engineering,
respectively.

6.4.1 Forward Engineering (FE) and Reverse Engineering

(RE)

Engineering is a process of accumulating knowledge and information data about

materials, parts, products, processes, and systems. Depending on the method by
which the knowledge and information is accumulated, an engineering process can
be either of forward engineering (FE) or reverse engineering (RE). As shown in
Fig. 6.4, forward engineering begins with a conceptual design at the abstract level
and concludes at a detailed physical solution. In contrast, reverse engineering begins
with a physical model and completes with a virtual model that represents the knowl-
edge and information of the physical model. Reverse engineering is an effective
digital technology to (1) create a virtual model of a physical object, product, process,
6.4 Example of Digital Engineering—Reverse Engineering 399

Fig. 6.4 Forward

engineering (FE) and reverse
engineering (RE)

or system; (2) explore geometric dimensions, information, and knowledge based

on physical objects; and (3) explore the structure (system elements and relations)
of an assembled product and system. A reverse engineering process can shorten
the technical gaps of an enterprise and business competitors in new technology
developments.
RE aims to recover, capture, and reuse knowledge, wisdom, and information
embedded in existing objects, products, or systems. However, RE is effective only
when the physical model to be reverse engineered shows the economic benefits
based on the return of investment (ROI). Since ROI depends on the expected product
volumes on markets, the cost of an RE project has to be analyzed to justify if
mimicking existing innovations is a right option for new product development. RE is
400 6 Digital Manufacturing (DM)

Reverse
Engineering by
Hardware

New
Reverse
Design and
Engineering by
Prototyping
Software

Office Chair System Level

Frame Seat Base Handle Back Assembly

Assembly Assembly Assemblies Assemblies Assembly Level

…… Adjustment Wheels Base frame …… Module Level

Fig. 6.5 Example of reverse engineering at different levels

applicable to any of design aspects; however, only the RE applications in geometric

designs of products are interested; computer-aided reverse engineering (CARE) tools
are introduced to (1) create virtual models from physical objects and (2) explore the
design intents of original physical objects such as design variables, constraints, and
features.
RE can be applied to assembled products, and Fig. 6.5 illustrates an RE process
used to create a digital model of an office chair. To create a virtual model, (1) the
chair is disassembled, and parts are scanned to acquire point clouds from exterior
surfaces; (2) the acquired data is processed to recover the surfaces of the virtual
model; (3) the solid model can then be defined using the boundary surfaces that are
defined at the previous step; (4) the solid model is then followed by a re-conception
process to discover design intents and generate a high-level parametric representation
of models; and (5) the part models are then assembled as a complete chair.

6.4.2 Procedure of Reverse Engineering

An RE project involves four critical steps as shown in Fig. 6.6. Firstly, a physical
object is scanned to acquire point cloud dataset from exposed surface or internal
solids. Commonly used scanning devices are laser scanners, cameras, coordinate
measuring machine (CMM), and computed tomography (CT). Secondly, the point
cloud dataset is processed to generate a polymesh as the boundary surfaces of
physical objects. Thirdly, the polymesh is cleaned to define a watertight volume
of solid. Fourthly, the solid model is analyzed to identify features, design intents,
6.4 Example of Digital Engineering—Reverse Engineering 401

CAD, CAE, CAM, Prototyping, and

Re-conception Process
Digital Modelling

Production
Existing physical Point clouds Parametrized Models, Analysis Report,
products polymesh feature-based models Prototypes

(a). The first and second steps to create (c). The post-processing phase to
(b). The third and fourth steps to
point cloud and polymesh from an utilized RE model for CAD,
create a parametric model of solids
existing physical object CAE, CAM, and prototypes

Fig. 6.6 Main tasks in a reverse engineering project

and constraints for a parametric model. Finally, the reverse engineered model can be
utilized in computer-aided engineering analyses or applications.

6.4.3 Reverse Engineered Modeling

One of the key tasks in an RE project is to construct a surface from point cloud.
To construct a watertight surface for a solid model, points are acquired by three-
dimensional sensors such as stereo cameras and laser sensors, and the point cloud is
then analyzed by computers to generate the boundary surfaces of solids. Figure 6.7
shows the surface reconstruction process, which consists of some main tasks
described below (Bi and Kang 2014).
(1) Data acquisition. Three-dimensional scanners are used to capture points over
visible surfaces of a physical object. When the object cannot be fitted in

Data acquisition Parametric modelling

Data Filtering Feature and applications
recognition
Data Surface
Simplification Data Reconstruction
Registration

Fig. 6.7 Critical tasks in surface reconstruction (Bi and Kang 2014)
402 6 Digital Manufacturing (DM)

workspace of the scanners or surface points are collected from different orien-
tations, the points should be acquired from different views and segments to
ensure the sufficient information for a watertight volume.
(2) Data Filtering. Raw data from sensors includes noise, redundant, distorted, or
even invalid data; it must be filtered and cleaned to remove noise and redundant
points for the collected data.
(3) Data Registration and Integration. When the data acquisition process involves
in a relative motion between the sensing system and physical object, the datasets
obtained from multiple views or continuous scanning paths should be registered
to merge datasets in a unified coordinate system. Registration is to determine
the transformation of a dataset between two views, and integration is to create
single part model from multiple sources of the datasets.
(4) Surface reconstruction. A point cloud from visible surface, or a volumetric
dataset, is analyzed to construct boundary surfaces of the object.
(5) Data simplification and smoothing. For a dataset with a large amount, or there
are some special requirements about the smoothness of a constructed surface,
surface data is simplified to reduce the size of the dataset and smooth the
boundary surface for better quality. Data simplification and smoothing are
extremely critical when a reverse engineered model is used in other engineering
designs such as finite element analysis and collision detection.
(6) Feature recognition. A part model consists of a number of features, which are
assembled by logical operators such as addition, subtraction, and union. It is
desirable to recognize the features, such as the machined features of holes,
bosses, and fillets, from a solid model.
(7) Parametric modeling and applications. An RE process aims to recover insights,
wisdom, and design intents of an engineering design from existing physical
objects. The parametrization of a solid model helps to catch the knowledge and
design intents. In addition, the obtained solid model can be utilized for other
engineering purposes such as rapid prototyping, mold design, and computer-
aided manufacturing.

6.4.4 Computer-Aided Reverse Engineering (CARE)

Many computer-aided tools have been developed for computer-aided reverse engi-
neering (CARE), for example, the commercial and open-source CARE tools that are
shown in Fig. 6.8. In comparison, commercial CARE tools offer better functionalities,
while open-source tools are flexible to be integrated with other computer programs
for continuous developments. Many software tools, such as Inventor, Sketchup, and
Blender, have commercial and open-source licenses.
A CARE tool is mainly used to accomplish two tasks (1) convert point cloud
into a polymesh model and (2) create a watertight solid model, and explore design
features and design intents for a parametric model. Here, the Autodesk Recap and
 Commercial Tools Open-Source Tools

FreeCA
SketchUp D

Autodesk SolidWorks
OpenSCAD

Autodesk Inventor

Tinkercad
6.4 Example of Digital Engineering—Reverse Engineering

Rhino 5 3D Studio Max

Blender

Fig. 6.8 Examples of computer-aided reverse engineering (CARE) tools

403
404 6 Digital Manufacturing (DM)

the ScanTo3D are used to illustrate how the CARE tools are applied to accomplish
these tasks.

6.4.4.1 Creating Polymesh Model

The raw data acquired by the triangulation, time-of-flight, or interferometry method

is in the format of point cloud, and a hardware system with such a capability is usually
equipped to produce point clouds directly. However, the raw data from photogram-
metry is 2D images, some sophisticated tools, such as Autodesk Recap Pro, are
required to convert 2D images into polymesh models. As shown in Fig. 6.9a, Recap
accepts three types of raw data, point cloud, image files from mobile devices, and
photos. The data processing service can be accessed online as shown in Fig. 6.9b; the
outcome of image processing is the polymesh model, which can be exported as in the
format of .rcs, .obj, .rcm, .fbx, or .ipm. Figure 6.10 shows three steps of using Recap

Three types of inputs

Folders for completed

projects

(a). Create an RE project (b). Interface for cloud-computing

Fig. 6.9 Cloud reverse engineering service by Autodesk Recap Pro

Create an object
from photos

(1). Prepare a number of photos, which (2). Create a new project in ReCap Pro, select (3). Allow the cloud service to complete surface
are taken from different locations, the mesh quality, the export file types, reconstruction, download a polymesh model
directions, and heights of the object. upload the photos, and confirm the in the .rcm format. The generated model can
submission. be viewed online and exported to other
computer-aided tools.

Fig. 6.10 The steps to create a polymesh model in Recap Pro

6.4 Example of Digital Engineering—Reverse Engineering 405

(a). Tools for editing a polymesh model (b). Exporting a polymesh model

Fig. 6.11 Editing and exporting a polymesh model in Recap Pro

Pro to convert a set of image files into a polymesh model. Recap also provides some
basic functions to analyze, edit, and refine a polymesh model as shown in Fig. 6.11a
before the model is exported in a format of .obj, .stl, or .fbx (see Fig. 6.11b). Create
an object from photos
The quality of the polymesh model depends on the completeness and resolution of
photos. In photogrammetry, the location and orientation of a camera are determined
by matching pixels in a group of photos. Ideally, 80% of the pixels of one photo
should be covered by other photos. The following guides should be applied to prepare
high-quality photos for a reverse engineering project (Autodesk 2020):
(1) Lighting conditions. Lights should be placed to avoid shadows in photos. In
an indoor environment, flash should not be used, and diffuse lights should be
used without a shadow on object. In an outdoor environment, photos should
not be taken under direct lights to avoid a strong contrast in photos.
(2) Cameras. Use a lens that produces a good sharpness and avoid the focal change
when the photos are taken. Stabilize the camera using support and clicker,
expand the wide depth of field to reduce amount of light, and use a camera
with a high optical resolution as possible.
(3) Photoing strategies. Place the object completely in the view window but let
the object occupy the window as large as possible. Avoid a background with
complex textures rather than monochromatic ones. Avoid moving any objects
in the scene in catching photos. Take photos from all directions of object to
ensure a feature of interest is covered by multiple photos. At the height, take
the photos around the object at each of 5–15°.

6.4.4.2 Generate Parametric Models

A polymesh model only includes the geometric information, while the ultimate goal
of a reverse engineering process is to explore features, intents, and knowledge of phys-
ical products. Therefore, polymesh models should be further analyzed and processed
to create parametric models with identified parameters, features, and design intents.
Figure 6.12 shows the procedure where the SolidWorks ScanTo3D was used to create
406 6 Digital Manufacturing (DM)

Import point clouds or

polygonal model

Run mesh preparation: align,

eliminate noise, simplification,
smooth, and fill holes.

Create section & Create 2D and 3D Automatic Guided surfaces

boundary curves or sketch curves solid create creation wizard
import discrete curves against mesh wizard
using the curve wizard

Build surfaces Edit feature Trimming and

from curves lines knitting

Feature Solid Model

Recognition

Fig. 6.12 Parametrize a solid model using Solidworks ScanTo3D

a parametric model from a point cloud or polymesh model. Firstly, a data file is
imported as an application of ScanTo3D. Secondly, the ‘mesh preparation’ tool is
used to correct, simplify, align, and smooth data; if no watertight boundary surface
is available, further trim extra faces and fill holes. Thirdly, the parametric modeling
tools are used to define sections, edges, surfaces, and generate a legal boundary
surface for a solid model. Fourthly, the ‘feature recognition’ tool is used to identify
possible features such as holes, cylinders, extrusions, and fillets.

6.5 Example of Digital Engineering—Direct

Manufacturing

Direct manufacturing provides an alternative to make unique products or prod-

ucts with a low volume. Direct manufacturing has been identified as one of the
emerging digital technologies that will have a great impact on manufacturing industry.
Direct manufacturing is a type of additive manufacturing in which physical products
are made directly from their digital models without custom manufacturing tools.
6.5 Example of Digital Engineering—Direct Manufacturing 407

Figure 6.13 shows three basic types of fabrication processes in creating geometric
shapes of parts. A formative process in Fig. 6.13c uses the cavity in a mold assembly
to define part geometry. When the material is added into the cavity, it is solidified or
sintered into a desired shape. A subtractive manufacturing process in Fig. 6.13a cuts
away unwanted materials to reshape a part by the cutting tool that moves along the
part; the part geometry is determined by the cutting tool and its motion with respect to
the part. A subtractive process is performed at a machine tool such as a lath, drill, or
mill. An additive manufacturing (AM) process in Fig. 6.13a makes a part by adding
the materials to a solid layer by layer gradually. A machine and programming tool
of additive manufacturing processes can be used to make any geometries of parts
without using customized tools.
Many technologies were developed for additive manufacturing, and Fig. 6.14
shows that there are seven basic types of additive manufacturing: vat photo-
polymerization, powder bed fusion, binder jetting, material jetting, sheet lamina-
tion, material extrusion, and directed energy deposition. The characteristics of these
technologies have been briefly explained in the figure.
A power bed fusion process in Fig. 6.15 is used as an example to illustrate the
additive manufacturing approach. The power particle is supplied by a supply mecha-
nism on the left layer by layer. Powder particles at each layer are fused and melted at
a high degree close to the melting point of power materials, and the thermal energy
is typically from a laser or electron beam. The portion of the heated powder particles
are sintered and fused as the bonded solid over the part. Once the additive process
is finished, unheated and loose powder particles can be blown or blasted away. The

Desired
Part
Geometry Specific tooling for
Generic tooling
individual part
for any part

Specific program for

individual part
Specific design of tooling for
individual part

(a). Additive process

(c). Forming process

(b). subtractive process

Fig. 6.13 Three basic methods to fabricate parts in manufacturing

408 6 Digital Manufacturing (DM)

Vat Photopolymerization: liquid photopolymer resin is cured when

it is exposed to light, and it initiates polymerization and be converted
a solid part.

Powder bed fusion (PBF): powder is consolidated selectively after it

is melted by a heat source. Surrounding powder acts as support
materials for overhanging features. .

Binder jetting: liquid bonding agents are applied selectively onto

thin layers of powder to build a part layer by layer .

Additive Material jetting: the materials in droplets are deposited layer by

manufacturing layer to build a part. Applicable materials are jettin and phoycurable
(AM) techniques resin which can ben cured by UV lights.

Sheet lamination: the sheets of material are stacked and laminated to

form objects. Layers can be bonded by adhesives or chemical,
ultrasonic welding, or brazing. After the AM process, Unwanted
materials are cut away in post-processing.

Material extrusion: the materials are extruded through nozzles or

orifices, and these materials are then combined into multi-player
models.

Direct energy deposition (DED): poser or wire is fed into a melt

pool on part surface; the melt material can adhere to the part by a
laser or electron beam.

Fig. 6.14 Types of additive manufacturing

materials that are suitable to a power bed fusion process include metal, metal alloys,
plastics, ceramic powders, and sand.
Despite of the fact that additive manufacturing techniques are highly diversified,
the procedure of making products by direct manufacturing is straightforward and
Fig. 6.16 shows seven main steps of a direct manufacturing process. In addition, the
hardware and software resources in accomplishing the tasks in these steps are also
introduced in the figure.

6.5.1 Prepare Digital Models

Direct manufacturing begins with a virtual model of product. A virtual model may
be for a newly designed product from scratch or an existing product for reverse engi-
neering. If the target is an existing physical object, reverse engineering techniques in
Sect. 6.4 are used to collect data from the physical object, process data, and construct it
into a sold model. If the target is a new product, it is assumed that the design is detailed
6.5 Example of Digital Engineering—Direct Manufacturing 409

X-Y movable tooling

Sintering
Powder supply
area Application of
mechanism
powder layer

Sintered part Un-sintered part

Platform

Frame

Lift mechanism

Fig. 6.15 Powder bed fusion (PBF) process

enough to determine geometric shapes of products. In addition, design for manufac-

turability has to be considered since not all of the features, materials, and functions
can be made with the required accuracy appropriately. Many computer-aided and
manufacturing tools such as commercial SolidWorks, SolidEdge, Unigraphics, and
free computer-aided packages such as FreeCAD, 3D Builder, and LibreCAD can be
used to create virtual models of products.

6.5.2 Prepare STL Files

A rapid prototyping machine accepts a product model only in the tessellated format.
Tessellation is a method to manage polygon datasets, and the .stl format is special
tessellated type for rapid prototyping. The abbreviation of ‘.stl’ stands for stere-
olithography, standard triangle language, and standard tessellation language. A .stl
file can be ‘binary’ or ‘ASCII’. A binary file is more concise and commonly used
unless the virtual model has to be reviewed and changed manually. Note that a virtual
model in .stl includes the information of boundary surfaces of an object only.
A .stl file should be exported only when the product design is finalized since
different formats correspond to different contents even though the geometric infor-
mation is the same. A .stl file cannot be changed easily. Figure 6.17 shows a compar-
ison of a SolidWorks .sldprt file and .stl file for a part with the same geometry.
Figure 6.17a is parametrized in which both of the 2D sketch and the revolving axis
can be changed easily at the feature level. However, the .stl file in Fig. 6.17b is a
triangulated surface model; it only includes a list of triangles and associated vertices.
410 6 Digital Manufacturing (DM)

Target a newly designed or (1). Create

existing physical product product model

CAD tools (SolidWorks, Autocad,

Inventor, Unigraphics,
SolidEdge….)
(2) Generate Iterations of tasks
tessellated model in design process

3DprinterOS, Astroprint, CraftWare,

Cura, IceSL, IdeaMaker,
KISSlicer,…. (3). Slicing
planning
Cura, MatterControl,
3DPrinterOS, Slic3r,
SliceCrafter, IceSl,…. (4). Machine
setup
VAT Photopolymerisation,
powder bed fusion, binder jetting,
material jetting, material
extrusion, direct energy (5). Build
deposition, and sheet lamination, Process

Automated or manual (6). Post

operations processing

Visual inspection, machines for

material properties testing, and (7). Verification?
machines for quality and
functional testing

Finish product

Fig. 6.16 Procedure of AM processes

solid RevolveFeature
facet normal -1.871077e-02 -9.115928e-01 -4.106682e-01
outer loop
vertex 8.656340e+01 5.184039e+00 6.051220e+01
Revolving axis 2D sketch for A long list of vertex 8.593156e+01 8.314227e+00 5.359267e+01
vertex 8.918127e+01 5.184039e+00 6.039293e+01
dimensioning triangular faces endloop
endfacet
facet normal -1.871067e-02 -9.115928e-01 -4.106683e-01
outer loop
vertex 8.593156e+01 8.314227e+00 5.359267e+01
vertex 8.918127e+01 8.314227e+00 5.344461e+01
vertex 8.918127e+01 5.184039e+00 6.039293e+01
endloop
endfacet
facet normal 1.871077e-02 -9.115928e-01 -4.106683e-01
outer loop
vertex 8.918127e+01 8.314227e+00 5.344461e+01
vertex 9.179913e+01 5.184039e+00 6.051220e+01
vertex 8.918127e+01 5.184039e+00 6.039293e+01
endloop
endfacet
facet normal 0.000000e+00 -1.000000e+00 -0.000000e+00
outer loop
vertex 8.918127e+01 5.184039e+00 7.929943e+01
vertex 8.656340e+01 5.184039e+00 6.051220e+01
vertex 8.918127e+01 5.184039e+00 6.039293e+01
endloop
endfacet
……

(a) Native SolidWorks model (b). STL surace model (ASCII)

Fig. 6.17 The same geometry with different contents in .sldprt and.stl formats
6.5 Example of Digital Engineering—Direct Manufacturing 411

Vertices are specified by the coordinates in a Cartesian system. Before the model
is ready for a rapid prototyping process, engineers are encouraged to preserve as
the high-level information (such as features and parameters) as possible to improve
designs in an iterative process.
Due to the importance and popularity of .stl format to additive manufacturing,
nearly all CAD packages are capable of exporting native solid models as the models
in a .stl format. As an example, Table 6.1 shows the interface, the steps, and the
settings in SolidWorks when a .sldprt model is exported as a .stl is sile.
When a product is involved in some curvy features, the resolution determines the
smoothness of curvy features unless the resolution of a .stl file exceeds that of the
rapid prototyping process. The resolution also affects the file size of a virtual model.
Figure 6.18 gives an example of a solid model as well as two .stl files with coarse

Table 6.1 The interface, steps, and settings to export a.stl file from SolidWorks

Steps in exporting a .stl file

(1). Click File menu choose Save As
(2). Select file type as .stl
(3). Click the Options button in the
Save dialog box.
(4). Select Output as Binary and select
the desired units (inches or
millimeters)
(5). Set Resolution to Fine
(recommended for most parts)
(6). If you want to preview the stl
model before saving, check the
“Show stl info before saving” box
(7). Name and Save your STL file

Note: STL files can be created from

SolidWorks part or assembly models.
Check the box at the bottom of the Export
Options dialog box to save the assembly
as one .stl file or individual .stl files.

(a) Solid model (b) Coarse STL model (c) Fine STL model
(size 19.5 MB) (65,126 triangles (327,667 triangles
size 18.4 MB) size 92.8 MB)

Fig. 6.18 Resolution affects the smoothness of curvy features and file sizes
412 6 Digital Manufacturing (DM)

Fig. 6.19 Visualize a rapid

prototyping process in
Simplify3D

and fine resolutions. The file size can be increased greatly with an increase of the
number of triangles in the model.

6.5.3 Slicing Algorithms and Visualization

A rapid prototyping process builds a part layer by layer. A slicing algorithm is used
to decompose a volume into a set of layers with a specified resolution. At each
layer, the algorithm determines the moving paths to cover the cross-sectional area.
The slicing algorithm generates the results of (1) a set of toolpaths for the given .stl
model, (2) estimated processing time and material consumption based on the specified
percentage of infills, and (3) the constructions of supporting materials when one or a
few of features on part are not self-supported (All3DP 2019). It is desirable that the
results from slicing algorithms are visualized and reviewed graphically to ensure that
the toolpaths or supports are reasonable. As shown in Fig. 6.19, many software tools,
such as Simplify3D, can be used to visualize printing paths and layered structures
from a rapid prototyping process.

6.5.4 Setup Machine

The quality of a rapid prototyped product is determined not only based on the tool-
paths from a slicing algorithm, but also on various design factors relevant to an
additive manufacturing process such as machine type, material type, and process
parameters. A control program has to be tailored to certain RP machine. Figure 6.20
shows a list of major design parameters involved in a fused deposition modeling
(FDM) process (Ha 2016). Even though the RP machine can be run to generate a
control program using the default settings with few manual intervenes, the optimized
quality of part relies on users’ understanding on various process parameters and the
experience in setting up the machines.
6.5 Example of Digital Engineering—Direct Manufacturing 413

Z
Concept Part Build
Model Materials Orientation
Y
Topology Melt and cool X-direction
optimization
X
Density Y-direction
STL file Color Z-direction
Key Performance
Part interior style Indicators (KPI)
Temperature Model build temperature Part fill style
Part angle
Humidity Envelop temperature Layer thickness
Raster width and gap
Machine calibration Part shrinkage
Contour width and gap
Nozzle diameter
Environmental FDM Machine Operating
Factors Setup Parameters

Fig. 6.20 Main design factors of a rapid prototyping process (Bi and Wang 2020)

6.5.5 Building Process

A control program is a set of instructions to prepare for tooling, move and follow
a toolpath for operating, and operate auxiliary tools based on the specified settings
of the process parameters. After the slicing algorithm processes the .stl file and
generates a control program, the RP machine is calibrated and setup appropriately,
and the rapid prototyping process can be performed to build a part layer by layer.
Depending on the part geometry, the operating speed of end-effector tools, and the
effectiveness of the slicing algorithm, the machining time varies from one part to
another.

6.5.6 Post-processing

When the building process is terminated, the part can be taken from the rapid proto-
typing machine. The post-processing is often needed to clean debris and remove
a supportive structure from the part. Occasionally, secondary processes such as
machining are used to create some machined features or improve geometric and
dimensioning tolerances on existing features.
Rapid prototyping offers some significant advantages in reducing the cost of
tooling and shortening the time of product development. However, RPs have their
disadvantages on quality and strengths of products due to (1) the applicable raw
materials having limited strengths and (2) staircase effects in a layer-by-layer opera-
tion. Table 6.2 shows the ranges of surface quality and minimum possible thickness
from different additive manufacturing. The surface roughness could be as poor as
(9–40 µm) from fused deposition modeling (Campbell et al. 2002; Kumbhar and
Mulay 2018).
414 6 Digital Manufacturing (DM)

Table 6.2 Surface roughness from different AM processes (Campbell et al. 2002; Kumbhar and
Mulay 2018)
Minimum layer Surface roughness
No. Process type
thickness (mm) (Ra) in m
1 Stereolithography (SLA) 0.100 2 ~ 40
2 Selective laser sintering (SLS) 0.125 5 ~ 35
3 Fused deposition modelling (FDM) 0.254 9 ~ 40
4 Material extrusion (3D printing) 0.175 12 ~ 27
5 Laminated object manufacturing (LOM) 0.114 6 ~ 27
6 Material jetting 0.100 3 ~ 30

The surface finish of a RP part may be improved in two ways: (1) optimizing
orientation of toolpath and reducing thickness between layers and (2) applying
secondary processes such as milling, laser surface finishing operations, or abrasive
flow machining.

6.5.7 Verification and Validation

Most of the rapid prototyped parts are for verification and validation purpose. As the
last step, physical parts are tested and evaluated to see if continuous improvement
is necessary. An iterative process applies until the performance of the prototyped
part meets the design requirements satisfactorily. Simple validation such as part
defects can be manually inspected; some quantifications such as surface roughness,
hardness, and tensile strengths need sophisticated testing machines. Recent efforts
on verification and validation were emphasized on micro-level structures, residual
stresses, fatigue strengths, and mechanical properties (Kim et al. 2015).

6.6 Studies in Customizing Digital Manufacturing

Technologies

There is no universal solution of DT, which is applicable to any manufacturing enter-

prises. Some customized efforts of research and development are usually required to
tailor general digital manufacturing solutions to certain applications. In this section,
some case studies of the research and development are introduced to show the
diversity of the applied studies in digital manufacturing.

6.6.1 Ubiquitous Sensing

Digital manufacturing is data driven, and the smartness and intelligence of the manu-
facturing system rely on the abundance and sufficiency of data collected from system
6.6 Studies in Customizing Digital Manufacturing Technologies 415

(a). various grippers (b). Cost-effective force sensing

Fig. 6.21 Force sensors for smart grippers (Bi et al. 2018)

elements. From this aspect, more sensors are used in a manufacturing system, the
better chance the right data is collected, and the faster and better decisions can be
made for the operations of manufacturing systems. A large amount of research and
development is needed to have cost-effective instrumenting solutions for smart things
in manufacturing systems.
Figure 6.21 shows an example of developing smart force sensors for robotic
grippers. As smart things, industrial robots are widely applied in manufacturing.
However, the controls of industrial robots are mostly based on kinematic models,
and this poses the challenge to (1) optimize the performance of a robot since any
motion is caused by force and (2) use them in complex environment such as a human-
machine coexistence environment. To implement dynamic control for robots, the
external forces occurring to system elements should be measured in real time and
the instrumenting solutions are demanded for robotic modules such as grippers in
Fig. 6.21a. A cost-effective sensing solution was developed to measure gripping
forces, and the collected data can be utilized to support the human-robot interaction
in open working environment.

6.6.2 Holistic Multi-sensing Solution for Real-Time Controls

Many manufacturing operations require the coordination, collaboration, and coop-

eration of multiple system elements. Taking an example of an automated guided
vehicle (AGVs) system in an assembly plant, the information of delivering tasks, the
states of AGVs, jobs, and relevant manufacturing resources are collected and fed to
the control system as inputs to plan, schedule, and control AGVs. The system may
416 6 Digital Manufacturing (DM)

Fig. 6.22 Holistic sensing solutions for real-time decision-making supports (Bi et al. 2017)

be complicated when the controls are decentralized since system elements coordi-
nate, collaborate, and cooperate with each other dynamically. In addition, the data
about relevant environmental factors should be collected and made available to these
system elements, and the real-time data will be fused, processed, and utilized to
support decision-making activities. Such types of research are highly demanded in
distributed and decentralized manufacturing systems; enterprise systems should be
able to collect environmental data from multiple sources, fuse, process, and utilize
data to achieve a high-level intelligence of decision-making units.
Figure 6.22 shows an example of decentralized control systems in which system
components coordinated and collaborated to accomplish respective tasks for opti-
mized system performances. It was a football robot team consisting of different types
of robots equipped with various sensors including range sensors, vision sensors, and
force sensors. With the real-time data collected from robots, robot members coor-
dinated and collaborated at system level. In developing such systems, engineers
need to select right sensors to detect events or measure physical quantities, and the
information systems should be able to utilize data for decision-making.

6.6.3 Methods of Coping with Big Data

To shorten the development time, virtual design must deal with the ever-increasing
volume, velocity, and variety of big data. In general, a big data means that the required
computation to an engineering problem is far beyond the capability of available
computing resources. With an increase of system complexity, more and more engi-
neering problems face the challenges of limited computing resources; data analysis
6.6 Studies in Customizing Digital Manufacturing Technologies 417

(a). actual body model (b). Simplified model for propagation simulation

Fig. 6.23 Developing of cost-effective solution for ubiquitous sensing (Särestöniemi et al. 2019)

must be innovative to narrow down big data into the dataset with a manageable size,
so that a feasible solution can be found in a reasonable timeframe.
Figure 6.23 gives an example of using an innovative approach to deal with the big
data in simulating the signal propagation in human body (Särestöniemi et al. 2019). A
human body is formed by skin, fat, bones, and sterns, and these organic matters have
different dielectric properties that affect signal propagating. Using a realistic body
model shown in Fig. 6.23a leads to the scenario of big data in numerical simulation.
However, the models should be greatly simplified by approximating the volumes
of different material properties as parametrized bodies. Accordingly, the computing
time for an acceptable solution was reduced significantly.

6.6.4 Methods of Data Mining

In an IoT-based system network, numerous smart things have their local control
units to optimize the performances. However, the coordination and cooperation of
these functional units are required to optimize system-level objectives. System-level
decisions are made based on the real-time data collected from smart things. However,
certain data may be relevant or irrelevant, and interpretable or uninterpretable to a
system-level objective such as variety, quality, lead time, and cost. Effective methods
are demanded to sort and process data and make a useful dataset manageable in
decision-making.
Figure 6.24 shows that the axiomatic design theory (ADT) is used to process the
data collected from low-level functional models and map the data into the category
related to system-level objectives to make bid data manageable (Cochran et al. 2017).
418 6 Digital Manufacturing (DM)

Fig. 6.24 Axiomatic design

theory (ADT) for data
mining (Cochran et al. 2017) Variety Quality Lead-time Cost

(a). Instrumented objects in IoT (b). Graphic user interfaces (GUIs) for data visualization

Fig. 6.25 Data visualization of human-machine interaction (Bi et al. 2016)

6.6.5 Data Visualization Methods for Human-System

Interaction

Humans still play important roles in digital manufacturing; in particular, humans are
usually needed to oversee the operations of manufacturing systems in identifying
abnormal situations and taking immediate actions. Effective interactive tools are
needed for humans to understand the system states using the real-time data collected
from manufacturing plants. Figure 6.25 shows an integrated surveillance system,
which is capable of accessing, retrieving, and visualizing real-time data based on the
selection criteria by users.

6.6.6 Data-Driven Decision-Making Units

Conventionally, the decision-making units at one-level of system architecture get the

inputs from a upper level unit and send the outputs to lower level unit(s). The decision-
making units do not require the flexibility to deal with the changes and uncertainties
beyond fixed interactions. However, IoT allows the direct interactions of functional
units, and the capabilities of these functional units should be expanded in terms
of flexibility and adaptability to deal with changes and uncertainties. Figure 6.26
6.6 Studies in Customizing Digital Manufacturing Technologies 419

Data driven functional

units at various domains
and levels

Fig. 6.26 Data-driven functional units (Bi and Wang 2013)

describes a scenario of an intelligent system, in which all of functional units are data
driven, so that a functional unit in the system can directly interact and respond to
another unit at any level and domain (Bi and Wang 2013).

6.6.7 Methods for Workflow Compositions

IoT brings the possibilities for small- and medium-sized companies to utilize external
manufacturing resources to catch emerging business opportunities. A virtual enter-
prise can be formed to execute an engineering project. However, a virtual enterprise
differs from physical enterprises in sense that manufacturing resources are across
the boundaries of physical enterprises, and additional businesses are involved in
selecting partners to configure workflows; executing engineering projects collec-
tively; and ensuring the privacy, security, safety, and reliability of business opera-
tions. Figure 6.27 shows a research example on the algorithms for the selection and
composition of services for virtual enterprises.

6.6.8 Standardization of Specifications

In the material flow of a manufacturing system, the complexity of products or

manufacturing processes is mainly tackled by selecting different types of manu-
facturing resources and assembling them in different ways. Manufacturing resources
are selected if their capabilities satisfy the functional requirements in applications.
It becomes critical to standardize the measures and descriptions of manufacturing
resources, so that a manufacturing system can be configured optimally. As shown
420 6 Digital Manufacturing (DM)

(a). Workflow composer (b). Petri net for representation of composition formulas

Fig. 6.27 Service selection and composition in virtual enterprise (Viriyasitavat et al. 2019a, b, c)

Fig. 6.28 Standardization of functional requirements of robotic modules (Bi et al. 2020)

in Fig. 6.28, Bi et al. (2020) discussed the challenges in standardizing functional

requirements of robotic systems in testing and measurement.

6.7 Summary

The idea of digital twin (DT) I discussed to have the first-time-right mapping from
a virtual twin and physical twin. DT emphasizes the mirror relation of a digital twin
and physical twin. However, the chapter focuses on the digital technologies to narrow
down the solutions of digital twins in vast design space. To this end, the functional
requirements of DT are discussed at all main aspects of system decision-making, a
free architecture is proposed, and a number of emerging digital technologies have
been discussed. From this chapter, engineers are expected to know how to (1) use RE
in digitizing physical objects; (2) use DM to create physical objects directly based
6.7 Summary 421

on digital models; and 3) understand main areas of applied researches in digital

manufacturing.
Design Projects
Project 6.1. Find a malfunctioned machine or machine elements you can reach
such as the broken dehumidifiers shown in Fig. 6.29, disassembly the machine to
investigate the parts and the assembly, and understand how the machine should
function, then create a digital model for the machine, demonstrate how the mechanical
system is supposed to work, and make your recommendation how the machine could
be fixed.
Project 6.2. Find a physical object you can reach, such as some decorative objects
shown in Fig. 6.30; create its solid model by using reverse engineering technique;
determine its physical properties such as mass, volume, and inertial properties; and
run feature recognition on the solid model to see if some parameters and features can
be recovered from constructed model. The following steps can be used as the guide
in using a cellphone or digital camera to collect data and using Autodesk Recap Pro
and Solidworks ScanTo3D to reconstruct a solid model and feature recognition.
(1) Select a physical object to be modeled in your RE project. Place an object on a
flat surface. Make sure the object can be taken photos from all sides and angles.
(2) Use a cellphone or digital camera to take over 50 photos from different positions
and angles.
(3) Create a user account at https://www.autodesk.com, download and install
ReCap Pro, create an RE project under in ReCap Pro, upload all photos, convert
them into a mesh file, download .rcm file, and use ReCap to export .rcm file as
an .obj file or .stl file for its use in Solidworks.

(b)
(a) (c)

Fig. 6.29 Examples of used products in design project 6.1

422 6 Digital Manufacturing (DM)

(a) (b) (c)

(d) (e) (f) (g)

Fig. 6.30 Examples of physical objects in design Project 6.2

(4) Use Solidworks—ScanTo3D to clean the imported mesh file, fix the mesh,
create a surface model, fix the surface model, and generate a solid model from
the watertight surface model.
(5) For the solid model, assign the materials and evaluate mass and volume
properties.
(6) Document your project idea, procedure, the challenges and solution, and result
of reverse engineering.
(7) 3D print a physical model using the solid model.

References

All3DP (2019) Want to get the best results from your 3D printer? https://all3dp.com/1/best-3d-sli
cer-software-3d-printer/#what
Autodesk (2020) What makes photo good for photogrammetry? https://blogs.autodesk.com/recap/
what-makes-photos-good-for-photogrammetry/
Berger R (2019) Digital factories, the renaissance of the U.S. automotive industry. http://secure.rbj.
net/collaborative-design-and-planning-for-digital-manufacturing-1st-edition.pdf
Berman SJ, Bell R (2007) Digital transformation—creating new business models where
digital meets physical. https://www-07.ibm.com/sg/manufacturing/pdf/manufacturing/Digital-
transformation.pdf
Bi ZM (2011) Revisiting system paradigms from the viewpoint of manufacturing sustainability. J
Sustain 3(9):1323–1340
Bi ZM, Kang B (2014) Sensing and responding to the changes of geometric surfaces in flexible
manufacturing and assembly. Enterprise Inf Syst 8(2):225–245
References 423

Bi ZM, Liu YF, Krider J, Buckland J, Whiteman A, Beachy D, Smitch J (2018) Real-time force
monitoring of smart grippers for Internet of Things (IoT) applications. J Indus Inf Integration
11:19–28
Bi ZM, Miao ZH, Zhang B, Zhang WJ (2020) A framework for performance assessment of
heterogeneous robotic systems. IEEE Syst J. https://doi.org/10.1109/JSYST.2020.2990892
Bi ZM, Wang G, Xu L (2016) A visualization platform for internet of things in manufacturing
applications. Internet Res 26(2):377–401
Bi ZM, Wang G, Xu LD, Thompson M, Mir R, Nyikos J, Mane A, Witte C, Sidwell C (2017) IoT-
based system for communication and coordination of football robot team. Internet Res 27(3):162–
181
Bi ZM, Wang L (2013) Manufacturing paradigm shift towards better sustainability in cloud manu-
facturing. Springer, London. ISBN: 978-1-4471-4934-7, http://doi.org/10.1007/978-1-4471-493
5-4_5
Bi ZM, Wang XQ (2020) Computer aided design and manufacturing (CAD/CAM). John Wiley &
Sons, Inc, Hoboken, NJ USA. IBSN-13:9781119534211
Buttner R, Muller E (2018) Changeability of manufacturing companies in the context of
digitalization. Procedia Manuf 17(2018):539–546
Campbell RI, Martorelli M, Lee HS (2002) Surface roughness visualization for rapid prototyping
models. Comput Aided Des 34(2002):717–725
Catapult (2018) Feasibility of an immersive digital twin: the definition of a digital twin and
discussions around the benefit of immersion. https://www.amrc.co.uk/files/document/219/153
6919984_HVM_CATAPULT_DIGITAL_TWIN_DL.pdf
Cochra D, Ki YS, Fole J, Bi ZM (2017) Use of the manufacturing system design decomposition for
comparative analysis and effective design of production systems. Int J Product Res 55(3):870–890
Fenner J (2003) Enterprise application integration technique. http://www-icm.cs.ucl.ac.uk/staff/W.
Emmerich/lectures/3C05-02-03/aswe21-essay.pdf
GE (2020) GE digital twin—analytic engine for the digital power plant, https://www.ge.com/dig
ital/sites/default/files/download_assets/Digital-Twin-for-the-digital-power-plant-.pdf
Geieves M, Vickers J (2006) Digital twin: mitigating unpredictable, undesirable emergent
behavior in complex systems (excerpt). https://www.researchgate.net/publication/307509727_
Origins_of_the_Digital_Twin_Concept/link/57c6f44008ae9d64047e92b4/download
Gelernter DH (1991) Mirror worlds: or the day software puts the universe in a shoebox—how it
will happen and what it will mean. Oxford University Press, Oxford, New York. ISBN 978-
0195079067. OCLC 23868481
Glaessgen EH, Stargel DS (2012) The digital twin paradigm for future NSA and US air force
vehicles. https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20120008178.pdf
Gregori F, Papetti A, Pandolfi M, Peruzzini M, Germani M (2017) Digital manufacturing systems: a
framework to improve social sustainability of a production site. Procedia CIRP 63(2017):436–442
Ha S (2016) 3D printing/process parameters. https://worldmaterialsforum.com/files/Presentations/
WS1-1/WMF%202016%20-%20WS%201.1%20-%20Sung%20Ha%20Final.pdf
Kim DB, Witherell P, Lipman R, Feng SC (2015) Streamlining the additive manufacturing digital
spectrum: a systems approach. Additive Manuf 5(2015):20–30
Kumbhar NN, Mulay AV (2018) Post processing methods used to improve surface finish of products
while are manufacturing by additive manufacturing technologies: a review. J Instit Eng (India)
Ser C 99(4):148–487
Kusiak A (2019) Fundamentals of smart manufacturing: a multi-thread perspective. Annual Rev
Control. https://doi.org/10.1016/j.arcontrol.2019.02.001
Mäkiö-Marusik E, Colomboa AW, Mäkiö J, Pechmann A (2019) Concept and case study for teaching
and learning industrial digitalization. Procedia Manuf 31(2019):97–102
Parrott and Warshaw (2020) Industry 4.0 and the digital twin. https://www2.deloitte.com/us/en/ins
ights/focus/industry-4–0/digital-twin-technology-smart-factory.html
Pati A, Bandyopadhyay PK (2017) Digital manufacturing: evolution and a process oriented approach
to align with business strategy. Int J Econ Manag Eng 11(7):1746–1751
424 6 Digital Manufacturing (DM)

Särestöniemi M, Pomalaza-Ráez C, Bi ZM, Kumpuniemi T, Kissi C, Sonkki M, Hämäläinen M,

Iinatti J (2019) Comprehensive study on the impact of sternotomy wires on UWB WBAN channel
characteristics on the human chest area. IEEE Access 7(1):74670–74682
Siemens (2020) Digital manufacturing—a holistic approach to the complete product lifecycle.
https://www.plm.automation.siemens.com/global/en/our-story/glossary/digital-manufacturing/
13157
Sönmez OE, Koç VT (2015) On quantifying manufacturing flexibility: an entropy based approach.
In: Proceedings of the world congress on engineering 2015 Vol II WCE 2015, July 1–3 2015.
London
Tao F, Zhang M, Nee AYC (2019) Digital twin and big data in digital twin driven smart
manufacturing, pp 183–202
Tuegel EJ, Ingraffea AR, Eason TG, Spottswood (2011) Reengineering aircraft structural life
prediction using a digital twin. Int J Aerospace Eng. Article ID 154798
Viriyasitavat W, Xu L, Bi ZM, Hoonsopon D (2019a) Blockchain technology for applications
in internet of things—mapping from system design perspective. IEEE Internet of Things J
6(5):8155–8168
Viriyasitavat W, Xu L, Bi ZM (2019b) rmSWSpec: real-time monitoring of service workflow
specification language for specification patterns. IEEE Trans Indus Inf 15(7):4021–4032
Viriyasitavat W, Xu L, Bi ZM (2019c) The extension of semantic formalization of service workflow
specification language. IEEE Trans Indus Inf 15(2):741–754
Wiendahl HP, ElMaraghy HA, Nyhuis P, Zäh MF, Wiendahl HH, Duffie N, Brieke M (2007)
Changeable manufacturing - classification, design and operation. CIRP Ann 56(2):783–809
Williams TJ (1994) The Purdue enterprise reference architecture and methodology (PERA). http://
citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.194.6112&rep=rep1&type=pdf
Wikipedia (2020) Digital twin. https://en.wikipedia.org/wiki/Digital_twin