Springer Series in Advanced Manufacturing

Wesley A. Salandro
Joshua J. Jones
Cristina Bunget
Laine Mears
John T. Roth

Electrically
Assisted
Forming
Modeling and Control
Springer Series in Advanced Manufacturing

Series editor
Duc Truong Pham, Birmingham, UK
More information about this series at http://www.springer.com/series/7113
Wesley A. Salandro · Joshua J. Jones
Cristina Bunget · Laine Mears · John T. Roth

Electrically Assisted Forming

Modeling and Control

13
Wesley A. Salandro John T. Roth
Joshua J. Jones Behrend School of Engineering
Cristina Bunget Penn State Behrend
Laine Mears Erie, PA
International Center for Automotive USA
Research
Clemson University
Greenville, SC
USA

ISSN  1860-5168 ISSN  2196-1735  (electronic)

ISBN 978-3-319-08878-5 ISBN 978-3-319-08879-2  (eBook)
DOI 10.1007/978-3-319-08879-2

Library of Congress Control Number: 2014944715

Springer Cham Heidelberg New York Dordrecht London

© Springer International Publishing Switzerland 2015

This work is subject to copyright. All rights are reserved 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. Exempted from this legal reservation are brief excerpts
in connection with reviews or scholarly analysis or material supplied specifically for the purpose of
being entered and executed on a computer system, for exclusive use by the purchaser of the work.
Duplication of this publication or parts thereof is permitted only under the provisions of the Copyright
Law of the Publisher’s location, in its current version, and permission for use must always be obtained
from Springer. Permissions for use may be obtained through RightsLink at the Copyright Clearance
Center. Violations are liable to prosecution under the respective Copyright Law.
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.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

Abstract

This book encompasses a compilation of work by the authors on the manufacturing

technique, Electrically Assisted Forming (EAF), whereby an electric current is passed
through a metal during the forming process. The importance of improved metal
­deformation within manufacturing is described, and the need for novel enhanced
metal forming techniques is presented. EAF has shown promising experimental form-
ing results on many lightweight metals, and within this book, macro-scale compres-
sion and tension modeling is presented. Bringing the technique even further towards
industrialization, strategies for controlling the applied electric current during EAF are
described. Furthermore, the sensitivities and impacts of EAF on intrinsic and extrin-
sic material properties are covered. Concluding the text, an explanation on designing
an Electrically Assisted manufacturing process is presented, and real-world potential
EAF applications are introduced.

v
Acknowledgments

The authors would like to recognize the academic institutions and local companies
that have supported the research included in this book:
• Penn State—Erie, The Behrend College is where the benefits of EAF where
experimentally proven on many lightweight metals and using various forming
methods.
• Clemson University—International Center for Automotive Research (CU-
ICAR) is where further EAF development took place, by way of creating EAF
models, introducing an EAF control strategy, and explaining the formability-
enhancing mechanisms of EAF.
• Aggressive Grinding Services Inc. has provided centerless and surface grind-
ing services to support with specimen and tooling fabrication.
• Insulfab Plastics Inc. has provided countless pieces of reinforced plastics that
are used as insulation components for fixtures and dies.
• Ionic Technologies Inc. has heat treated or hardened many fixtures for the EAF
research.
• Fuchs Lubricants has provided several of the top metal forming lubricants to
be used in comparison studies with EAF.
• Koyo Bearings has provided testing instrumentation and has supported with
material testing.

vii
Contents

1 Deformation of Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Relevant Background on Automotive and Aerospace Industries. . . 2
1.2 Present Forming Technologies. . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Hot Working. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.2 Incremental Forming . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.3 Superplastic Forming. . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2.4 Tailor-Welded Blanking. . . . . . . . . . . . . . . . . . . . . . . . 7
1.3 Limitations of Current Technologies. . . . . . . . . . . . . . . . . . . . . . 8
1.4 Plastic Deformation of Metals. . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4.1 Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.2 Dislocations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4.3 Crystalline Structures. . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.4.4 Lattice Defects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.5 Metrics of Formability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.1 Formability in Sheet Metals. . . . . . . . . . . . . . . . . . . . . 17
1.5.2 Additional Forming Metrics. . . . . . . . . . . . . . . . . . . . . 18
1.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2 Introduction to Electrically Assisted Forming. . . . . . . . . . . . . . . . . . . 23

2.1 Electrically Assisted Forming . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2 EAF Literature Review. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.1 EAF Theory and Modeling. . . . . . . . . . . . . . . . . . . . . . 26
2.2.2 Significant EAF Modeling Variables
from Experimentation. . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.2.3 Relation to Crystal Structure and Resistivity. . . . . . . . 32
2.2.4 Electroplasticity and Electromigration. . . . . . . . . . . . . 33
2.3 Broader Impacts of EAF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.3.1 Automotive and Aerospace Industries. . . . . . . . . . . . . 33
2.3.2 Potential Early Adopters of EAF Modeling. . . . . . . . . 34
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

ix
x Contents

3 The Effect of Electric Current on Metals. . . . . . . . . . . . . . . . . . . . . . . 37

3.1 Electrical Current Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2 Previous Electroplastic Theories . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2.1 Localized Heating. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2.2 Electron Wind Effect . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3 Comprehensive Electroplastic Theory Explanation. . . . . . . . . . . 43
3.3.1 Electrical Current Without Metal Deformation. . . . . . 43
3.3.2 Electrical Current with Metal Deformation. . . . . . . . . 46
3.3.3 Electrical Current Effects on Formability . . . . . . . . . . 48
3.3.4 Supporting Experimental Results. . . . . . . . . . . . . . . . . 49
3.4 Electroplastic Theory Conclusions. . . . . . . . . . . . . . . . . . . . . . . . 51
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

4 Macroscale Modeling of the Electroplastic Effect. . . . . . . . . . . . . . . . 55

4.1 Mechanical-Based Approach to Determining the EEC. . . . . . . . 55
4.1.1 Experimental Setup and Procedure . . . . . . . . . . . . . . . 56
4.1.2 Mechanical-Based EEC Determination Procedure . . . 57
4.1.3 Mechanical-Based EEC Conclusions. . . . . . . . . . . . . . 58
4.2 Thermal-Based Approach to Determining the EEC. . . . . . . . . . . 58
4.2.1 Building a Thermal Model. . . . . . . . . . . . . . . . . . . . . . 59
4.2.2 Experimental Setup and Procedure . . . . . . . . . . . . . . . 63
4.2.3 EEC Thermal-Based Determination. . . . . . . . . . . . . . . 65
4.2.4 Thermal-Based EEC Conclusions . . . . . . . . . . . . . . . . 70
4.3 Comparison Between the Different EEC Determination 
Approaches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.3.1 SS304 Electroplastic Effect Coefficient Profiles. . . . . 71
4.4 EEC Profile Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.5 Empirical Modeling Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.5.1 Non-constant Current Density. . . . . . . . . . . . . . . . . . . 73
4.5.2 Constant Current Density. . . . . . . . . . . . . . . . . . . . . . . 78
4.6 Macroscale Modeling of the Electroplastic Effect Conclusions. . . 82
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

5 Compressive Electrically Assisted Forming Model. . . . . . . . . . . . . . . 83

5.1 Analytical Modeling of Compression Forming Processes. . . . . . 83
5.1.1 Definition of an EAF Modeling Strategy. . . . . . . . . . . 83
5.1.2 Analysis of an Electrically Assisted
Compression Process. . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.1.3 Effective Stress and Strain—Classical
Compression Test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5.1.4 Effective Stress and Strain—Electrically
Assisted Compression Test. . . . . . . . . . . . . . . . . . . . . . 86
5.1.5 Current Density Relationship During
Electrically Assisted Compression. . . . . . . . . . . . . . . . 88
Contents xi

5.1.6 Strain and Temperature Effect on Resistance

and Current. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.1.7 Analytical Model for Electrically Assisted
Compression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.1.8 Overall Solution Schematic. . . . . . . . . . . . . . . . . . . . . 90
5.1.9 EAF Modeling Approach Summary. . . . . . . . . . . . . . . 91
5.2 Simplified EAF Forging Model. . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.2.1 EAF Forging Stress–Strain Model. . . . . . . . . . . . . . . . 91
5.2.2 Modeling Strategy Overview. . . . . . . . . . . . . . . . . . . . 92
5.2.3 Coupled Thermo-Mechanical Modeling . . . . . . . . . . . 92
5.2.4 Assumptions of the Thermo-Mechanical Model. . . . . 94
5.2.5 Experimental Setup and Procedure . . . . . . . . . . . . . . . 96
5.2.6 Experimental and Modeling Results . . . . . . . . . . . . . . 97
5.2.7 Electrical Efficiency Analysis . . . . . . . . . . . . . . . . . . . 100
5.2.8 EAF Forging Model Conclusions. . . . . . . . . . . . . . . . . 101
5.3 Specific Heat Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
5.4 Heat Transfer Modes Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.5 EEC Profile—Material Sensitivity Comparison. . . . . . . . . . . . . . 107
5.6 EAF Modeling—Sensitivities and Simplifications
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

6 Tensile Electroforming Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

6.1 Thermal Modeling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
6.1.1 Model Development. . . . . . . . . . . . . . . . . . . . . . . . . . . 114
6.1.2 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
6.1.3 Results and Discussion. . . . . . . . . . . . . . . . . . . . . . . . . 124
6.1.4 Thermal Model Conclusions . . . . . . . . . . . . . . . . . . . . 132
6.2 Mechanical Modeling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
6.2.1 Deformation/Strength Model Derivation. . . . . . . . . . . 134
6.2.2 Deformation/Strength Model Solution Method. . . . . . 137
6.2.3 Deformation/Strength Model Results. . . . . . . . . . . . . . 140
6.2.4 Mechanical Modeling Conclusions . . . . . . . . . . . . . . . 152
6.3 Thermo-Mechanical Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
6.3.1 Thermo-Mechanical EAF Model Overview
and Solution Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . 153
6.3.2 Thermal Expansion Stress . . . . . . . . . . . . . . . . . . . . . . 154
6.3.3 Model Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
6.3.4 Division of Thermal Expansion, Thermal
Softening, and Direct Electrical Effects. . . . . . . . . . . . 157
6.3.5 Thermo-Mechanical Modeling Conclusions . . . . . . . . 158
6.4 Tensile Electroforming Model Conclusions. . . . . . . . . . . . . . . . . 159
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
xii Contents

7 Control of Electrically Assisted Forming . . . . . . . . . . . . . . . . . . . . . . . 161

7.1 Constant Force Forming. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
7.1.1 Benefits of Constant Force Forming. . . . . . . . . . . . . . . 168
7.2 Constant Stress Forming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
7.2.1 Benefits and Opportunities of Constant
Stress Forming. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
7.3 Constant Current Density Forming . . . . . . . . . . . . . . . . . . . . . . . 172
7.3.1 Benefits of Constant Current Density Forming. . . . . . 174
7.4 Model-Based Control Feasibility. . . . . . . . . . . . . . . . . . . . . . . . . 175
7.5 Process Control Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

8 Microstructure and Phase Effects on EAF. . . . . . . . . . . . . . . . . . . . . . 179

8.1 Grain Size Effect on EAF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
8.1.1 Specimen Preparation and Resulting Grain Sizes . . . . 180
8.1.2 Experimental Grain Size Testing. . . . . . . . . . . . . . . . . 181
8.1.3 EAF/Grain Size Conclusions. . . . . . . . . . . . . . . . . . . . 183
8.2 Prior Cold Work Effect on EAF. . . . . . . . . . . . . . . . . . . . . . . . . . 183
8.2.1 Importance of Percent Cold Work
on EAF Effectiveness. . . . . . . . . . . . . . . . . . . . . . . . . . 184
8.2.2 Specimen Preparation. . . . . . . . . . . . . . . . . . . . . . . . . . 184
8.2.3 Experimental Setup and Procedure . . . . . . . . . . . . . . . 185
8.2.4 Results and Discussion. . . . . . . . . . . . . . . . . . . . . . . . . 186
8.2.5 EAF/Percent Cold Work Conclusions . . . . . . . . . . . . . 191
8.3 Microstructure Analysis Under Tensile Loading. . . . . . . . . . . . . 192
8.3.1 As-Received Material Microstructure . . . . . . . . . . . . . 192
8.3.2 Summary of Statistical Analysis of Micrographs. . . . . 195
8.3.3 Room Temperature Deformation Microstructure. . . . . 198
8.3.4 EAF Microstructure. . . . . . . . . . . . . . . . . . . . . . . . . . . 201
8.3.5 Microstructure Analysis Conclusions. . . . . . . . . . . . . . 209
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210

9 Tribological and Contact Area Effects . . . . . . . . . . . . . . . . . . . . . . . . . 211

9.1 Contact Area Effect on EAF Effectiveness . . . . . . . . . . . . . . . . . 211
9.1.1 Specimen Preparation (Surface Ground). . . . . . . . . . . 212
9.1.2 Specimen Preparation (Enhanced Asperities) . . . . . . . 212
9.1.3 Post-forming EAF Roughness Examination . . . . . . . . 214
9.1.4 Experimental Setup and Procedure . . . . . . . . . . . . . . . 216
9.1.5 Thermal Analysis of EAF Based on Contact Area. . . . 217
9.1.6 Voltage–Resistance Contact Area Model. . . . . . . . . . . 223
9.1.7 Mechanical Analysis of EAF Based
on Contact Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
9.1.8 EAF/Contact Area Conclusions. . . . . . . . . . . . . . . . . . 230
Contents xiii

9.2 Tribological Effect on EAF Effectiveness. . . . . . . . . . . . . . . . . . 231

9.2.1 Effects of Electricity on Tribological Conditions. . . . . 232
9.2.2 Experimental Setup and Procedure
(Ring Tribo-Tests) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
9.2.3 Determining Friction Calibration Curves. . . . . . . . . . . 235
9.2.4 Testing Procedures. . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
9.2.5 Candidate Metal Forming Lubricants. . . . . . . . . . . . . . 237
9.2.6 Experimental Results and Discussion . . . . . . . . . . . . . 238
9.2.7 Lubricant Evaluation (Reduction
in Forming Load). . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
9.2.8 Temperature Measurements. . . . . . . . . . . . . . . . . . . . . 241
9.2.9 EAF/Tribology Conclusions. . . . . . . . . . . . . . . . . . . . . 243
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

10 Design of an Electrically Assisted Manufacturing Process. . . . . . . . . 245

10.1 Energy Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
10.1.1 Conventional Cold Forming. . . . . . . . . . . . . . . . . . . . . 245
10.1.2 Thermally Assisted Forming . . . . . . . . . . . . . . . . . . . . 247
10.1.3 Electrically Assisted Forming . . . . . . . . . . . . . . . . . . . 249
10.1.4 Energy Comparison. . . . . . . . . . . . . . . . . . . . . . . . . . . 250
10.2 AC Versus DC Current. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
10.2.1 Energy Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
10.2.2 Skin Effect. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
10.3 Additional Process Design Considerations . . . . . . . . . . . . . . . . . 252
10.3.1 Power Supply. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
10.4 EAF Process Design Conclusions . . . . . . . . . . . . . . . . . . . . . . . . 253

11 Applications of Electrically Assisted Manufacturing. . . . . . . . . . . . . . 255

11.1 EAF Bending Application and Model . . . . . . . . . . . . . . . . . . . . . 255
11.1.1 Analysis of an EA Bending Process. . . . . . . . . . . . . . . 257
11.1.2 Assumptions of the EAB Model . . . . . . . . . . . . . . . . . 257
11.1.3 Classical Bending Process (Force and Springback). . . 258
11.1.4 Analytical Modeling of EAB. . . . . . . . . . . . . . . . . . . . 260
11.1.5 EAB Solution Schematic. . . . . . . . . . . . . . . . . . . . . . . 263
11.1.6 Experimental Setup and Procedure . . . . . . . . . . . . . . . 264
11.1.7 Thermal Measurements in EAB. . . . . . . . . . . . . . . . . . 268
11.1.8 Validation of the Model via Experiments. . . . . . . . . . . 269
11.1.9 Effects of Electricity in Bending . . . . . . . . . . . . . . . . . 270
11.1.10 EAB Model Conclusions. . . . . . . . . . . . . . . . . . . . . . . 276
11.2 Electrically Assisted Machining. . . . . . . . . . . . . . . . . . . . . . . . . . 277
11.2.1 Observations in Low-Strain-Rate EA Machining . . . . 277
11.2.2 High-Strain-Rate Process Modeling
and Experimental Testing for EA Machining. . . . . . . . 279
11.2.3 EA Machining Conclusions. . . . . . . . . . . . . . . . . . . . . 283
xiv Contents

11.3 Electrically Assisted Friction Stir Welding . . . . . . . . . . . . . . . . . 284

11.3.1 Electrically Assisted Friction Stir
Welding Background. . . . . . . . . . . . . . . . . . . . . . . . . . 284
11.3.2 EAFSW Experimental Setup. . . . . . . . . . . . . . . . . . . . 285
11.3.3 Results and Discussion. . . . . . . . . . . . . . . . . . . . . . . . . 287
11.3.4 Conclusions and Future Work . . . . . . . . . . . . . . . . . . . 291
11.4 Experimental Findings for Alternative EAF Processes . . . . . . . . 291
11.4.1 Compression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
11.4.2 Tension. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
11.4.3 Non-uniform Deformation
(E.G. Channel Formation) . . . . . . . . . . . . . . . . . . . . . . 300
11.4.4 Springback Reduction Using EAF. . . . . . . . . . . . . . . . 301
11.4.5 Electrically Assisted Micro-Forming. . . . . . . . . . . . . . 302
11.5 Overhead Transmission Line Design Using EAF . . . . . . . . . . . . 303
11.5.1 The Electricity Transmission Grid. . . . . . . . . . . . . . . . 303
11.5.2 Transmission Line Structures and Setups. . . . . . . . . . . 303
11.5.3 Commercial Conductors and Sizing. . . . . . . . . . . . . . . 304
11.5.4 Conductor Sag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
11.5.5 Effect of Temperature on Transmission
Line Longevity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
11.5.6 Applying EAF Modeling to OHTL
Sag Calculations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
11.5.7 Future Work to Determine EEC Values
for OHTL’s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309

Appendix A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

Appendix B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327

Appendix C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Abbreviations

AISI American Iron and Steel Institute

AO Analog Output
APF Atomic Packing Factor
ASTM American Society for Testing and Materials
BCC Body-Centered Cubic
CAE Computer-Aided Engineering
CAFE Corporate Average Fuel Economy
CCD Constant Current Density
CGA Circle Grid Analysis
CI Confidence Interval
cRIO CompactRIO
DOT Department of Transportation
EA Electrically Assisted
EAF Electrically Assisted Forming
EA-Forging Electrically Assisted Forging
EAM Electrically Assisted Manufacturing
EDM Electrical Discharge Machining
EEC Electroplastic Effect Coefficient
EPA Environmental Protection Agency
FCC Face-Centered Cubic
FE Finite Element
FEA Finite Element Analysis
FFT Fast Fourier Transform
FLC Forming Limit Curve
FLD Forming Limit Diagram
FLIR Forward-Looking Infrared
GBS Grain Boundary Sliding
GHG Greenhouse Gas
GUI Graphical User Interface
HCP Hexagonal Close Packed
IF Incremental Forming

xv
xvi Abbreviations

LDH Limiting Dome Height

LVDT Linear Variable Differential Transformer
MBC Model-Based Control
NCCD Non-Constant Current Density
NI National Instruments
OEM Original Equipment Manufacturer
PID Proportional-Integral-Derivative
PLC Portevin–Le Chatelier
PS Parameter Set
SCR Silicon Controlled Rectifier
SMDI Steel Market Development Institute
SPF Superplastic Forming
TWB Tailor Welded Blank
UHSS Ultra-High Strength Steel
About the Authors

Dr. Wesley A. Salandro  is from Latrobe, Pennsylvania,

where he grew up as an industrial mechanic and machinist.
In 2009, Salandro received a B.S. in Mechanical Engineer-
ing from the Pennsylvania State University. During this
time, Salandro began experimentally researching Electri-
cally Assisted Manufacturing (EAM). In 2012, Salandro
received a Ph.D. in Automotive Engineering from the
Clemson University—International Center for Automo-
tive Research (CU-ICAR), where he was named the 2012
CU-ICAR Outstanding Ph.D. Student. Dr. Salandro’s
graduate research consists of modeling and predicting en-
ergy-assisted manufacturing processes. Wes has authored
over 15 technical publications on manufacturing process modeling, experimentation,
and design. Wes has received several international recognitions for his works, including
the 2012 Institution of Mechanical Engineers (IMechE) SAGE Best Paper Prize, as well
as the 2012 IMechE George Stephenson Gold Medal.
Email: [email protected]

Dr. Joshua J. Jones  is currently employed with Koyo

Bearings North America as a senior process develop-
ment engineer in their metal forming group. Prior to
joining Koyo, he received his Ph.D. degree in Automo-
tive Engineering from Clemson University in 2012 and
his B.S. in Mechanical Engineering from The Pennsyl-
vania State University. During his time at Clemson and
Penn State, his primary research focus was on the mod-
eling and control of Electrically Assisted Manufacturing
(EAM). He has five journal articles, eight conference
publications, and numerous presentations at technical conferences on this topic.
He received first place at the 2010 NAMRI/SME Student Research Presentation
Contest for presenting modeling work for Electrically Assisted Forming.

xvii
xviii About the Authors

He also has received several scholarships and grants during his time as student from
organizations such as SME, Clemson University, NSF, PA SciTech, and Penn State.
Aside from EAM, other research interests include lightweight material integration,
manufacturing process design, and physical process modeling/simulation.

Dr. Cristina Bunget  received her B.S. in Mechanical

Engineering, particularly machining and designing of
machine tools, from Polytechnic University of Bucha-
rest, Romania. After gaining practical experience from
working as a design engineer in a locomotive factory,
Cristina decided to continue her studies, and received
M.S. and Ph.D. majored in Mechanical Engineering
and minored in Mathematics from North Carolina State
University. During the time spent there, she was ex-
posed to various metal forming processes, and looked
for non-conventional ways to overcome some of the
challenges brought by newly emerged technologies, such as microforming and ul-
trasonically-assisted forming. Later, during the postdoctoral training at the Interna-
tional Center for Automotive Research (ICAR, part of Clemson University), her
exposure enlarged by working on various projects, from machining of superalloys
to metal forming applications (accumulative roll bonding on titanium sheets and
Electrically Assisted forming).

Dr. Laine Mears P.E., is an Associate Professor in

Automotive Engineering at Clemson University, teaching
and carrying out projects at the Clemson International
Center for Automotive Research. He teaches modeling
and analysis of automotive manufacturing processes, and
has performed research in Intelligent Machining Sys-
tems, manufacturing process design and control, and
manufacturing equipment diagnostics. He is the recipient
of the NSF CAREER award, SAE Teetor Educational
Award, and together with Drs. Salandro and Bunget,
­recipient of the IMechE George Stephenson Gold Medal.
Dr. Mears  has previously held industry positions with Hitachi Automotive
and SKF Bearings both as Manufacturing Engineer and Engineering Manager in
a high-­volume precision manufacturing environment. Applicable work in industry
includes: power optimization of hard machining processes, multi-spindle turning
analysis and startup of a bulk deformation rolling process. Dr. Mears has a B.S. in
Mechanical Engineering from Virginia Tech and M.S. and Ph.D. degrees in Me-
chanical Engineering from Georgia Tech. He is a member of the American Society
of Mechanical Engineers and a Senior Member of both the Society of Manufactur-
ing Engineers and the American Society for Quality. He is an ASQ Certified Quality
Engineer (CQE), BMW Lean Six Sigma Black Belt, and a licensed Professional
Engineer.
About the Authors xix

Dr. John T. Roth is professor of Mechanical

Engineering at Penn State Erie, The Behrend College
and associate director of research and technology trans-
fer in the School of Engineering. He received his B.S.
and Ph.D. degrees in Mechanical Engineering at Michi-
gan Technological University.
Dr. Roth  has an active research program in the ar-
eas of signal processing, material removal, cryogenics,
and electrical manufacturing. Some of the sponsors of
his research program include: Ford Motor Company,
COLDfire Technology, Pennsylvania Infrastructure Technology Alliance (PITA),
McInnes Rolled Rings, Spinworks, STERIS Corporation, Saegertown Manufactur-
ing Corporation, and The Lincoln Electric Company. He also has worked on various
joint projects with Carnegie Mellon University.
His teaching interests include: Manufacturing Processes, Materials Science,
Dynamic Systems, Adaptive Signal Processing, and Computer-Aided Design/
­
Manufacturing.
Dr. Roth  is a member of the American Society of Mechanical Engineering—
Manufacturing Engineering Division (ASME-MED), Society of Manufacturing
Engineers (SME), North American Manufacturing Research Institute (NAMRI),
and the American Society for Engineering Education (ASEE). He is currently the
Technical Program Co-Chair of the International Conference on Manufacturing
Science and Engineering (2007) and will be the Chair in 2008. He currently serves
as the chair of the Quality/Reliability Technical Committee of ASME-MED and
also serves on both the Manufacturing Equipment Technical Committee of ASME-
MED and the Scientific Committee of NAMRI/SME. He is also a member of
ASM International.
Chapter 1
Deformation of Metals

Deformation of metals is one of the key manufacturing processes for adding value
in secondary operations. Deformation as a process evolved soon after metal refine-
ment in ancient times to shape tools and other functional parts. With the dawn of the
second Industrial Age, the availability of localized power sources allowed for the
development of more standardized automated processes for metal deformation, such
as forging, stamping, drawing, and extrusion. These processes gave rise to high-
value, standardized components for construction, vehicles, and consumer goods.
With the development of automated deformation processes, there came a need
for analysis of process behaviors and physics, in order to understand the process
limitations, feasible design regions, and “optimum” points. A number of key mod-
els have been developed over the past century to describe behavior, such as the
power law and Johnson-Cook models for flow stress, forming limit diagrams for
sheet metal behavior, and various shearing and punching force models. These have
been augmented with improved variables that better describe material changes
with environmental and process effects such as temperature and strain rate, and are
used to size equipment, plan processes, and define limitations.
Two key limitations of the deformation process are achievable strain (related to
ductility) and flow stress (related to material strength and strain hardening behav-
ior). These become significant limitations to processing as more brittle, higher
strength metal alloys are introduced to design. These limitations and the associ-
ated cost of higher capital investment to overcome them are the motivation for the
methods described in this text.
To begin, we will describe technologies used to form metals, the mechanisms
by which metals plastically deform, and details of the limitations to be addressed.

© Springer International Publishing Switzerland 2015 1

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_1
2 1  Deformation of Metals

1.1 Relevant Background on Automotive

and Aerospace Industries

US-based automotive Original Equipment Manufacturers (OEM’s) are an integral

part of the US economy, responsible for nearly one million automotive-related
jobs [1]. A key evolution in today’s automotive designs is reduction of mass in
order to meet more stringent legislation on allowable fuel economy and vehicle
emissions for this industry (see Fig. 1.1).
Vehicle energy consumption across a given driving cycle can be described by

1
E= (P)dt = mβ1 + CD ρAf β2 + mgCr β3
2 (1.1)
cycle

where the power P over time t can be decomposed into mass m effects (accel-
eration, increasing elevation), drag effects through the drag coefficient CD, fluid
density ρ, frontal area Af, and rolling resistance Cr effects by way of weighting
coefficients βi. This energy, and hence fuel consumption, can be reduced through a
number of possible means as shown in Table 1.1.
Of these, lightweighting to reduce mass-based energy consumption is an
approach that can be readily implemented and has significant and immediate
effects on fuel consumption. The concept of vehicle lightweighting is currently a
major focus for all US and foreign automotive manufacturing OEMs, due to the

80

70
2025
2024
Fuel Economy Target, mpg

60 2023
2025
2022
2021
50 2020
2019
2018
2017
40 2016
2015
2014
2012
30 2013
2012

20
30 35 40 45 50 55 60 65 70 75 80
Footprint, ft2

Fig. 1.1  Current and planned minimum fuel economy standards for US passenger vehicles [after
US Department of Transportation (2013). Fact Sheet: DOT and EPA Establish CAFE and GHG
Emissions Standards for Model Years 2017 and Beyond]
1.1  Relevant Background on Automotive and Aerospace Industries 3

Table 1.1  Potential Electrification/hybridization of powertrains

approaches to reduction of Development of low-resistance tires
fuel consumption
Use of synthetic oil to reduce engine friction
Integration of fuel-tracking devices
Improvements in engine efficiency
Lightweighting of vehicles

above-described regulatory standards for fuel consumption, and associated emis-

sions regulations. To reinforce the importance of this statement, the following are
quotes from several automotive OEM’s about their commitment toward vehicle
weight reduction:
• “The use of advanced materials such as magnesium, aluminum, and ultra-high
strength boron steel offers automakers structural strength at a reduced weight
to help improve fuel economy and meet safety and durability requirements.”—
Ford Motor Company [2]
• “Lightweight design can be achieved by engineering lightweight, manufacturing
lightweight and material lightweight design.”—BMW AG [3]
• “Lightweight design is a key measure for reducing vehicle fuel consumption,
along with power train efficiency, aerodynamics and electrical power manage-
ment.”—Volkswagen [4]
• “Excess weight kills any self-propelled vehicle…Weight may be desirable in a
steam roller but nowhere else.”—Henry Ford [5]
Since the early 1980s, US vehicle weights have increased significantly [6]. A state-
ment from General Motors claims that about 20–40 % of this weight increase is
due to the content increase (i.e., navigation, electronics, and accessories) in newer
vehicles [7]. Consumer trends toward larger vehicles have also contributed to this
upward trend. This increase in weight has led to a decrease in efficiency of these
vehicles, even as powertrain efficiencies have improved. Automotive OEMs realize
that consumers want more efficient vehicles without compromising desires; this
situation has led to a heightened focus on a number of lightweighting techniques.
Vehicle lightweighting is approached through two primary strategies: (1) By
using new alloys with high strength-to-weight ratios (e.g., magnesium, high-strength
steels, or titanium), or (2) By using creative design strategies (e.g., stainless steels
with ribbed designs, integrated material full vehicle designs). Some case studies of
creative design approaches are shown in Fig. 1.2. Some of the lightweighting metals
may not be as strong as the heavier metals that they are replacing, so a combination
of both lightweighting techniques will commonly be used. The following examples
show how different lightweight materials and creative design strategies were used to
reduce weight in a vehicle without compromising strength.
• For an automobile bumper, a 20 % weight reduction was achieved using stain-
less steel versus carbon steel, when the C1000 Stainless Steel design included
“ribs” for strengthening rather than a large cross section of the carbon steel [9].
4 1  Deformation of Metals

Fig. 1.2  Lightweight design strategies [8]. Using lightweight materials or creative design strate-
gies are two main methods for lightweighting. a A stainless steel ribbed bumper beam [9] design
saves weight compared with a carbon steel design, b The reduced thickness of a stainless steel
gas tank [10] decreased the weight and allowed for greater capacity compared with a plastic tank.
a Stainless steel bumper beam [9], b Plastic versus steel gas tank [10]

• For an automotive fuel tank, a 4 % capacity increase and a 20 % weight reduc-
tion were accomplished by using stainless steel as the tank material compared
with conventional plastic used in fuel tanks, because the tank wall thickness
could be decreased [10].
Aside from the two examples shown above, the overall material content used
in vehicle production as a whole has changed dramatically over the last century
[11]. Specifically, larger amounts of high/medium strength steels, aluminum,
polymers, and even some magnesium are being integrated into all components
of the vehicle (e.g., body panels, trim, engine) to reduce weight [11]. An exam-
ple of this integration is the BMW X6 Sport-Activity Vehicle, which received the
“Great Designs in Steel, Automotive Excellence Award” from the Automotive
Applications Council of AISI’s Steel Market Development Institute (SMDI) for
utilizing ultra-high strength steels (UHSS) to enforce strength without increasing
weight [12].
Overall, there is a greater amount of aluminum and magnesium being imple-
mented into the vehicles. With the limited formability of these metals compared
to current automotive metals, there is the need for an efficient metal-forming tech-
nique capable of making components from these and other comparable metals in
place of the current metals.

1.2 Present Forming Technologies

There are numerous techniques used in manufacturing plants today that help to
improve metal formability. Within this subsection, several of the most common
techniques are described and their advantages/disadvantages are discussed.
1.2  Present Forming Technologies 5

1.2.1 Hot Working

Hot working is defined as the deformation of a material at an elevated temperature.

As part of this process, the metal is heated above its recrystallization temperature,
thus increasing the formability of the material. Advantages to hot working include
decreased flow stress and increased ductility. This is one of the simplest manufac-
turing methods because all that is required is a heat source, such as a heater or fur-
nace. In many cases, however, these benefits come at the expense of part quality.
One key disadvantage includes lower dimensional accuracy, due to uneven ther-
mal expansion resulting from temperature gradients within the material. Moreover,
a rougher surface finish (resulting from an oxide layer developing on the outside
of the part) is another consequence of using this process. Also, as the size of the
workpiece increases, larger furnaces will be needed, proving to be more costly and
taking up a larger footprint on the shop floor. Further, energy use is much higher
for this technique. Regardless of the minor fluctuations in part quality and cost,
this relatively simple and effective process makes it a desirable choice when hold-
ing rough tolerances where secondary finishing operations will likely follow.
Using the stress versus strain graph in Fig. 1.3, the effects of hot working can
be compared to a room-temperature (i.e., cold forming) compression test when
forging Ti-6Al-4V. Due to hot working, the compressive flow stress was decreased
and the amount of achievable compressive displacement prior to fracture was
increased when compared to cold-working conditions.

1.2.2 Incremental Forming

Incremental forming (IF) is a type of manufacturing process in which a metallic

part is deformed in small steps with a minor heat treatment (i.e., a process anneal)

Fig. 1.3  Hot working. Parts are bulk heated to induce thermal softening, then formed at high tempera-
ture with lower forces. a Ring being hot rolled to increase diameter. b Stress–strain curve for Grade 5
Titanium showing cold working (upper curve) versus hot working (lower curve) stress profiles. a Hot
ring rolling [13], b Ti-G5 hot working [14] (Figure courtesy of McInnes Rolled Rings, Erie, PA, USA)
6 1  Deformation of Metals

Fig. 1.4  Incremental Forming (IF): Parts are deformed in small increments, with a minor heat
treatment usually performed off-line in between steps. a A cone is formed from a flat sheet of
metal in four forming steps (increments). b A stress–strain schematic of an IF process

performed in between steps [16]. Figure 1.4 shows a schematic diagram of the incre-
mental-forming process, along with a graphic of a blank at different stages of the
IF process. This type of manufacturing is especially beneficial when forming brittle
sheet metals and is used in the automotive and aircraft industries. The major advan-
tages to this process are the large amounts of deformation and the decrease in the
required deformation forces that can be obtained. These advantages are possible
because of the minor heat treatments performed in between the increments of defor-
mation. The treatments eliminate the effects of cold work or strain hardening by
causing recrystallization to occur during each process anneal, thus resulting in a new,
overall weaker material. Aside from the benefits, this process does have its disadvan-
tages. A big downfall is the potential for low-dimensional accuracy, since the part
must be continuously removed and re-fixtured before and after the heat treatments.
The decreased accuracy arises from the fact that the part may not be fixtured in the
exact fashion each time it is removed and re-installed. Also, this process can be very
time-consuming, depending on the variables such as the number of heat treatments
and their respective durations, as well as the depth of the desired deformations.
Using this technique, production times are greatly increased, hence, IF may not be
an optimum process for high-production or high-precision manufacturing, however,
materials can be formed to great distances and complex shapes can be achieved.

1.2.3 Superplastic Forming

Superplastic forming (SPF) involves heating a material to extremely high temper-

atures (roughly two-thirds of its melting temperature) when deforming, as seen in
Fig. 1.5. This process can produce tremendous elongations of up to 2,000 %, cou-
pled with greatly reduced flow stress [17]. Other advantages include being able
to form precise complex shapes in which minimal or no residual stresses are pre-
sent. Also, lower strength tooling and fixtures can be used since the required forces
1.2  Present Forming Technologies 7

Fig. 1.5  Superplastic forming. A part is heated to roughly two-thirds of its melting temperature

and formed to allow for great elongations and a formed part with very few residual stresses

for deformation are minimized. This process can be used to form complex shapes
because of the very low-forming forces; however, it also has its disadvantages.
First, the superplastic-forming process is only applicable for very fine-grained
alloys (less than 10–15 µm), such as some aluminum (5083-FG and 7475), titanium
(Ti-6AL-4V), and magnesium alloys (Mg-AZ31B). These small grains allow for
GBS (grain boundary sliding) to occur at elevated temperatures, which is the pri-
mary SPF deformation mechanism responsible for the huge elongation increases.
Another consequence of this process is that extremely slow strain rates must be used
(10−4 – 10−2 s−1). Similar to the incremental-forming technique, the superplastic
technique may not be practical for many high-production manufacturing applica-
tions and can be classified as a batch-forming process. Neglecting the limited num-
ber of applicable materials and the slow strain rate that is required, this process is
capable of producing precise complex geometrical parts with little or no finishing
operations needed. Vehicle manufacturers, such as Porsche and Aston Martin, have
used SPF to form components for their low-production exotic cars.

1.2.4 Tailor-Welded Blanking

When using tailor-welded blanks (TWB’s), different sheets of material (i.e., differ-
ences in material grade, thickness, or coating) are mechanically or automatically
welded together before the forming process [18]. This allows the manufacturers
8 1  Deformation of Metals

to produce custom blanks, where strong, lightweight materials are placed where
they are needed, while utilizing more formable steels in other areas, thus allowing
for a relatively strong and easily formable part. However, this process can be time-
consuming, costly and can result in reduced part accuracy because of all the extra
manufacturing steps and associates required to prepare the blanks.

1.3 Limitations of Current Technologies

The goal of any manufacturing engineer is to produce quality parts as efficiently

and cost-effectively as possible. For many common engineering metals, this can
be accomplished rather easily, however, it has proven challenging with some of
the stronger, more lightweight metals which are being incorporated into today’s
designs. These materials, such as high-strength aluminum-, steel-, magnesium-,
copper-, and titanium-based alloys, all possess high strength-to-weight ratios, but
their limited formability makes them impractical for use in many real-world appli-
cations that require complex part geometries. The main downfall in using these
materials to make complicated shapes is the fact that, with current technology, the
forming capability is insufficient, such that the forming process is extremely time-
consuming and some very complex shapes may not even be able to be formed at
all. In this case, numerous simpler parts must first be formed and then attached
using screws, rivets, or welds, which can significantly increase the overall cost and
useable lifecycle of the products.
High-production costs and poor part quality issues can result from attaching
smaller, simpler parts together, making the disadvantages of using these materials
outweigh their great strength-to-weight characteristics. To this end, formability-
enhancing techniques are used to increase the overall efficiency of the manufactur-
ing process, thus increasing the applicability of these materials and allowing more
complex part geometries to be formed from single blanks rather than attaching many
smaller components together. Formability-enhancing techniques must be devised
and employed on current manufacturing methods to make them more applicable
for forming lightweight metals. Experts say that extensive research, which couples
materials and manufacturing engineering, is the key toward further developing light-
weight engineering [19, 20]. Not only do novel manufacturing techniques need to
be created and proven, but computer-aided engineering (CAE), analytical modeling,
and simulations of these novel processes must be further developed in order to gain
industry acceptance for a specific formability-enhancing technique.

1.4 Plastic Deformation of Metals

Plastic deformation can be classified as permanent reshaping of a metal. In plastic

deformation by slip, dislocations move through the crystal structure of the metal,
breaking, and reforming metallic bonds. Dislocation motion (i.e., deformation) can
1.4  Plastic Deformation of Metals 9

be hindered by defects in the crystal structure of the metal. In this section on plas-
tic deformation, the following will be explained:
• Bonding
• Dislocations
• Crystal structure
• Lattice defects

1.4.1 Bonding

Any group of bonded atoms has an associated energy. The bonding force for metal-
lic bonds consists of the attraction forces, due to opposing charges of the atoms,
and the repulsion forces, that are due to the overlapping of the outer shells of the
electrons. The equilibrium spacing is achieved by balancing the attractive forces
and the repulsive forces. As these forces increase, the energy state of the bonded
atoms increases, and thus, it is more willing to find another atom to bond with that
will decrease the energy state. In metals, the ion nuclei (consisting of protons and
neutrons) exert a net positive charge. The valence (or free) electrons (with a nega-
tive charge) surround these ion cores, creating an attraction. Note that the mass of
an electron (9.11 × 10−31 kg) is much smaller than the mass of a proton or neu-
tron (1.67 × 10−27 kg) [21]. Each ion core carries a net positive charge because its
valence electrons were given up to create the “electron cloud” or “sea of electrons”
that is shared between all the ion cores. The positive charge of a particular ion core
and the same positive charge of neighboring ion cores lead to the ion cores hav-
ing repulsive forces between them. The negatively charged valence electrons are
attracted to the much larger positively charged ion cores. In doing this, the valence
electrons negate these repulsive forces from the neighboring ion cores. The valence
electrons create the spacing between the ion cores and absorb the repulsive forces
from the same charges. Depending on the magnitude of the charge of the ion cores,
a corresponding amount of valence electrons will be attracted to it.
Of the three main types of bonds (ionic, covalent, and metallic), metallic
bonds are the weakest. A bond generally consists of atoms or a core, and elec-
trons. Surrounding each atom are shells of electrons. Each shell has the maximum
number of electrons that it can hold in it. The electrons in the outer-most shell are
called valence electrons or free electrons because they have the ability move from
one outer shell to another. If a particular atom has open spaces for extra electrons
in its outer shell, then that particular atom is at a higher energy level as compared
with an atom with all of the open spaces in its outer shell occupied. If the atoms
have their outer shells completely full of electrons, then they will be at the lowest
energy state and will not be reactive to the other atoms (i.e., the noble gases at the
far right column of the periodic table). If there are openings available in the outer
shell, the atom is still at a higher energy state than what it could be. In addition to
missing valence electrons, the energy of the atoms could also be increased by an
applied external stress. As this external stress on the atom increases, it will create
10 1  Deformation of Metals

an energy state higher than what the atom can withstand and will force the atom to
break its bonds and reform new bonds and share valence electrons.
In essence, the classification of the bond type is dependent on how the valence
electrons are utilized by the material. In ionic bonds of NaCl, the valence electrons
are permanently transferred from a metallic element to the nonmetallic element.
In doing so, the two elements now will have equal and opposing charges. This is
a strong type of bonding because the valence electrons are not shared, but actually
transferred, and each atom exclusively owns their electrons.
Materials with ionic bonding are ceramics. These materials can also withstand
high temperatures, since their bonding strength is high and it would take a lot of
heat to increase the energy of the atoms to cause the bonds to break.
In covalent bonds, the valence electrons are shared between multiple atoms.
Methane (CH4) is an example of a covalently bonded material, since electrons are
shared between the carbon element and the four hydrogen elements. In this case,
each element needs the shared valence electrons to stay bonded. This type of bond
is typically not as strong as the ionic bond because the valence electrons are being
shared, rather than actually being transferred from one atom to another, so several
atoms own the electrons and each can have an effect on what happens to the total
electron count in the bond. In the case of CH4, if one out-of-the four total hydro-
gen elements (per CH4 molecule) breaks and reforms with another set of elements,
then the CH4 molecule is now left at a higher energy state and desires another
hydrogen element to share electrons.

1.4.2 Dislocations

Dislocation motion is required for plastic deformation by slip. A dislocation is sim-

ply a misalignment of the atomic structure in the lattice of a metal. There are three
types of dislocations (edge, screw, and mixed dislocations). In the edge dislocation
shown in Fig. 1.6, there is an extra set of atoms within the top half of the lattice. The
location of where the string of atoms is un-bonded at its end is the dislocation in
the lattice structure, and it is at a higher energy state as compared with the bonded
atoms. As a force is applied to the metal, the string of atoms that were previously
un-bonded at its end will now bond with the neighboring string of atoms in the lower
half plane. This will then leave the neighboring string of atoms in the top plane,
which are one atomic unit in the direction of the force, and un-bonded. Thus, the
dislocation or the bonding defect in the lattice moved one atomic unit. As the bonds
continue to break and reform, which is caused by the external force exerted on the
metal leading to the stress exerted on the dislocation, the bonding defect (disloca-
tion) will migrate through the metal’s lattice. This overall shifting of the dislocations
is termed plastic deformation. To clarify the difference between elastic and plastic
deformation, in elastic deformation, the bonds are only stressed, and the lattice goes
back to its original spacing once the metal is unloaded. Conversely, bonds must be
broken for plastic deformation to take place. Once bonds are broken, the metal can-
not go back to its original shape without re-breaking and reforming of the bonds.
1.4  Plastic Deformation of Metals 11

Fig. 1.6  Edge dislocation.
The bonding defect in the
center of the lattice, where
there is a string of un-bonded
atoms, is an edge dislocation

Fig. 1.7  Screw dislocation.
The screw dislocation is
created by a shear stress
that causes the lattice above
or below the shear line to
advance one atomic spacing
unit

The second type of dislocation is the screw dislocation, as seen in Fig. 1.7. In this
type of dislocation, there is a step or ramp shape that is created due to the external
forces. As seen in the figure, the external forces create a shear stress that moves the
front upper region of the lattice one atomic spacing unit past the front bottom region.
As you examine the depth of the lattice, a screw-like “ledge” is formed since the
front region shifted a complete atomic spacing unit and the rear region did not yet.
As a greater amount of shear force is exerted, the neighboring bonds near the front
region will become more stressed and will likely break and reform to enable this
shift to move its way along the depth of the crystal structure in the figure.
12 1  Deformation of Metals

Fig. 1.8  Mixed dislocations.
Mixed dislocations are a
combination of edge and
screw dislocations

The third type of dislocation is the mixed dislocation, as shown in Fig. 1.8.

Most of the dislocations in metals are mixed dislocations, since they may consist
of multiple lattice defects that are representative of both edge and screw disloca-
tions. In the figure, as the force is exerted at point A, the bonded atoms at the loca-
tion of where the force was applied break bonds with the aligned atoms and reform
bonds with the atoms one atomic spacing in the direction of the force to cause a
screw dislocation. This causes one set of bonded atoms in the top half of the unit
cell to now be un-bonded, which is shown by the edge dislocation at point C.
Regardless of the dislocation type, the number of dislocations within a metal
increases as the level of plastic deformation increases. This is because the disloca-
tions do not only move through the lattice, but new dislocations are created at lat-
tice defects. A dislocation line can be classified as an un-bonded string of atoms.
A defect in the lattice disrupts the equilibrium bonding of the lattice. If forces are
exerted on the lattice and bonds must consistently break and reform, this disruption
by the lattice defect can create bonding inconsistencies in the form of dislocations,
because of the extra energy needed to break and reform bonds around this defect.
There are certain “pathways” that dislocations can move throughout a metal’s lat-
tice. These pathways are called slip systems. The dislocations travel on slip systems,
which are comprised of slip planes and directions, and are specific to the particular
crystalline structure of the material. While traveling on these slip systems, the dislo-
cation motion can be hindered by different interfacial defects within the lattice. Such
defects include impurities, voids, grain boundaries, faults, and other dislocations. As
these obstacles hinder the dislocation movement, the dislocations begin to pileup,
thus increasing the forces needed to continue plastic deformation. This phenomenon
is known as strain hardening. In order for the dislocations to be able to surpass the
obstacles, additional energy is required to force the dislocation past the defect. In
order for the dislocations to continue moving, there must be enough energy to: (1)
1.4  Plastic Deformation of Metals 13

distort the lattice surrounding the dislocation, (2) move the dislocation past the lat-
tice defect, and (3) break and reform bonds within the lattice [21].
If dislocation motion is not continued, the metal will fracture. In this case, frac-
ture is a result of too many dislocation pileups. As the metal deforms, more dislo-
cations are created and there is a higher dislocation density within the metal. As
the density increases, the separation distance between the dislocations is reduced
and the repulsive forces between the dislocations will increase as the dislocation
density increases. To this end, if a metal is being deformed and dislocations can-
not move, more dislocations are being generated and are being moved until they
come across a stuck dislocation, or a pileup. The addition of dislocations contin-
ues until the repulsive forces between dislocations become too high and cause the
metal to break. One possible way to eliminate dislocations from a metal before
fracture would be to perform an annealing procedure, where the metal is heated to
a material-specific time and temperature setting, and a new lattice structure would
be created from recrystallization with a lower dislocation density.

1.4.3 Crystalline Structures

A crystalline material can be any material that has an actual pattern in which
atoms are situated over a given atomic level distance. The pattern of atoms is the
lattice. There are three main crystal structures for metals, and the separating char-
acteristics of these structures involve spacing, packing, and stacking patterns. For
each crystalline structure, several key characteristics will be discussed, including
the coordination number, the atomic packing factor (APF), and the number of slip
systems. Figure 1.9 shows the unit cells of the three primary crystalline structures.
The coordination number describes the number of neighboring atoms that a single
atom touches within a unit cell of the metal. This can be important for character-
istics such as heat transfer, since thermal conductivity would typically be higher
in metals where more atoms are contacting each other. The APF shows the ratio
of the volume of the atoms to the total volume of a unit cell. This can describe the

Fig. 1.9  Unit cells of the primary crystalline structures [21]. a FCC, b BCC, and c HCP
14 1  Deformation of Metals

amount of excess space within a unit cell of a metal. The higher the APF, the less
excess space there is for any other atomic level particles to fit through. The num-
ber of slip systems that a metal has is important for plastic deformation, because
these are the only methods by which dislocations can travel in order to facilitate
deformation (ignoring nano-deformation mechanisms).

1.4.3.1 Face-Centered Cubic (FCC)

The coordination number of a FCC metal is 12, which is the number of atoms a
single atom touches within a unit cell. In addition, the APF (i.e., the volume of
the atoms in a unit cell compared to the total volume of the unit cell) is 0.74. FCC
metals have 12 slip systems. As stated above, these are like “roadways” for the
dislocations to travel on in order to allow for plastic deformation. Examples of
some FCC metals are copper, aluminum, and some stainless steels [21, 22].

1.4.3.2 Body-Centered Cubic

The number of neighboring atoms contacting each atom in a Body-Centered Cubic

(BCC) metal is 8 (the coordination number). Additionally, the packing factor for
BCC metals is 0.68, which is lower than the FCC metal. The BCC metals can have
12 or 24 slip systems, depending on the different slip plane/direction combina-
tions. Iron is an example of a BCC metal [21, 22].

1.4.3.3 Hexagonal Close-Packed

The coordination number and the APF for Hexagonal Close-Packed (HCP) metals
are the same as for FCC metals, which are 12 and 0.74, respectively. The HCP met-
als have the lowest number of slip systems, with three; therefore plastic deformation
by way of dislocation motion can be difficult and makes this class of metals act brit-
tle. These metals often twin to facilitate deformation, due to the lack of active slip
systems. Twinning is when a mirror representation of a particular lattice arrange-
ment is created across a given plane (i.e., a twin plane). As the shear force within the
metal increases, it will generate a mirror representation of a particular lattice struc-
ture. This newly generated lattice structure will provide reoriented slip systems that
will offer the dislocations potential pathways for which to travel on and continue
deformation. Examples of HCP metals are titanium and magnesium [21, 22].

1.4.4 Lattice Defects

A characteristic of a crystalline material is a repetitive atomic structure or pattern

across its lattice. However, there can be several different types of defects through-
out the lattice of a crystalline material. The importance of lattice defects is that
1.4  Plastic Deformation of Metals 15

they can affect dislocations that are traveling through the different slip systems
within the lattice. Some examples of lattice defects are voids, impurities or alloy-
ing components, grain boundaries, phase boundaries, and other dislocations.
• A vacancy is an empty location in the lattice structure where an atom would
normally appear. At a vacancy site, the neighboring atoms either cannot fully
bond or must bond with atoms further away, both of which will increase the
energy state of the bonding.
• An impurity can be any element in the atomic structure that is not normally
included in the lattice. Each element has atoms which have a specific atomic
radius. To this end, if all the atoms in the lattice were the same size (i.e., a pure
material), the bond spacing between the atoms would be equal. However, if an
original atom is replaced with an atom of a different size atomic radius, the bond
spacing at and near this impurity will now be affected. There can be two types of
impurities. A substitutional impurity is when the atom of a base metal is replaced
with an atom of another material. The difference in atomic size of the new atom
will provide for either a state of compression or tension on the surrounding
bonds. The second type of impurity is the interstitial impurity, where an atom
from a new element is aligned in the free space between two bonded atoms of
the base metal. As is the case with the substitutional impurities, the interstitial
impurities also distort the bonds in the local region.
• Grain boundaries mark the end of one lattice in a crystal structure and the begin-
ning of another. The lattice structure consists of slip planes oriented in particular
directions with respect to the local grain. Once into another grain boundary, the
slip planes on either side of the grain boundary will be oriented with respect to
the particular grain they are in, which will be different compared with the previ-
ous or next grain. In dual-phase materials, the different phases act the same as
the grain boundaries, in that they break up the consistency of the lattice pattern.
• Dislocations can also be considered lattice defects because they represent por-
tions of un-bonded atoms within a lattice.
Regardless of the type of lattice defect, any flaw in the lattice of the material will
have an effect on the dislocations that are moving through the metal. In the case
of a perfect lattice, the dislocations will move by way of breaking and reforming
bonds due to external forces that are exerted, as was described in the previous sec-
tions. However, when there is something in the structure of the lattice that is not
normally there (i.e., a defect), the dislocations now have a harder time advancing
past this point, unless extra energy is provided to do so. The halting or hinder-
ing of dislocation motion leads to dislocation pileups. If the dislocations cannot
move through the metal, the metal cannot continue deforming (since a collection
of many dislocation motions result in deformation at the bulk level). Continuing
this example, if the metal was being plastically deformed and dislocation pileups
were beginning to occur, the force required to continue deforming the metal would
increase, and this additional force is required to provide the dislocations with the
extra force they need to surpass the defects within the lattice. In general, this is the
basis for strain hardening. Different types of lattice defects can cause dislocation
pileups in different ways.
16 1  Deformation of Metals

• Impurities can be intentionally added to a material for strengthening purposes

(i.e., alloying in aluminum or stainless steel alloys). These impurities can be
intended to increase the strain hardening of the metal in order to make it stronger
for a particular application. As the number of impurities in a metal increases, the
magnitude of strain hardening increases.
• Grain boundaries are the locations where the orientation and direction of the
lattice change, and different grains are spread throughout the microstructure of
the metal. Each time a dislocation comes to a grain boundary; it cannot progress
through the boundary. The dislocation creates a stress at this boundary such
that a new dislocation is created from the stress at the grain boundary inside the
new grain, and then, this dislocation can begin moving through the new grain’s
different oriented slip systems. For the same bulk dimensions, a metal with a
smaller grain size would have more grain boundaries and a greater grain bound-
ary surface area for the dislocations to travel through so they would need excess
energy each time they reached a grain boundary in order to change direction/
orientation, or to create stress to produce a new dislocation. Therefore, a metal
with a smaller grain size is typically stronger than a metal with a larger grain
size. In addition, the number of slip systems also plays a role in the strain hard-
ening because a metal with fewer slip systems will have a less likely chance of
the slip planes lining up well at the grain boundaries, and the dislocations will
have a higher probability to need to change directions to a larger degree.
• The dislocations are also a defect, but they are not stationary, like the impuri-
ties and the grain boundaries. If a moving dislocation comes into contact with
another dislocation, they could either cancel each other out or they could repel
each other. If the first dislocation cannot move, because it is being held up
(repelled) by another dislocation, then the pileup will continue until enough
energy is now supplied such to provide the dislocation at the front of the pileup
with enough energy to break the bonds in front of it and travel through the lattice.

Aside from alloying, potential lattice defects can also be intentionally created
(strengthening mechanisms) or removed (weakening mechanisms). For strengthening
mechanisms, heat treating the metal to produce a small grain size could strengthen the
metal so there are more grain boundaries for the dislocation to pass through. Also, the
metal could be pre-worked so that the dislocation density within the metal is higher
than in the annealed state and these dislocations dispersed through the metal are all
additional defects. Dislocation density is presented in terms of the total of all the dislo-
cation lengths within a unit volume [mm/mm3], or it could be presented as the number
of dislocations intersecting a 2-D unit area of a material [#/mm2]. For a polycrystal-
line metal, the dislocation density at an annealed state can be about 107 mm/mm3 and
the dislocation density in a worked state can be about 1012 mm/mm3, which means
that the overall dislocation density between annealed and worked conditions could be
about 100,000 times different [23]. Heat treating could also be performed to weaken
the metal, by way of allowing for recrystallization and grain growth to occur. The
larger grains would equate to a fewer number of grain boundaries that a dislocation
must pass through and change its direction, hence less external force would be needed.
1.5  Metrics of Formability 17

1.5 Metrics of Formability

The formability of a material can be generally defined as the ability of a mate-

rial to plastically deform to a desired shape without the presence of defects. When
considering the formability of a material, there are many considerations or inputs
to the forming process. These can be broken down into three main categories:
process parameters, material attributes, and strain boundary criteria. These three
groups are detailed in Table 1.2.

1.5.1 Formability in Sheet Metals

For sheet stamping, work by Ford Motor Company has developed forming indices
that classify possible defects resulting from a forming process [25]. The defects
that are the basis of the indices are as follows:
1. Splits in the stamping—mechanical damage on the stamping surface that devel-
ops into a local neck that eventually creates a split with continued deformation.
2. Splits on the stamping edge—burrs on the sheet edge under tensile stresses
become regions where splitting can occur.

Table 1.2  Factors that Formability

influence formability [24] Process
Deformation mode
Applied stress state
Strain rate
Temperature
Lubrication regime
Material
Normal anisotropy
Yield strength
Fracture toughness
Strain hardening
Strain rate sensitivity
Grain size and shape
Dislocation density
Texture
Crystal structure
Strain boundary criteria
Wrinkling
Surface roughness
Springback
Strain localization
Tearing
18 1  Deformation of Metals

3. Wrinkling—a local area in the sheet suffers a compressive stress which leads to
plastic instability and sheet waviness.
4. Shape change—during unload of the sheet elastic recovery or springback can
occur.
5. Low stretch—low deformation of flat areas of the sheet can have less work
hardening than other areas and suffer in dent resistance.
6. Soft surface—an uneven amount of deformation in a flat area of the sheet can
cause unbalanced residual stresses such that a small force disturbance on the
sheet after forming can cause elastic instability (i.e., oil canning).
The first two of these defect types are hard failures while the remaining defects are
unacceptable during sheet forming. Six formability index terms analogous to the
above failure modes have been defined:
1. Anti-fracturability
2. Anti-edge-fracturability
3. Anti-wrinklability
4. Shape-fixability
5. Stretchability, and
6. Anti-buckability

1.5.2 Additional Forming Metrics

Additionally, other formability parameters can include the following: the normal
anisotropy, the limiting dome height (LDH), the hole expansion ratio, the form-
ing limit diagram (FLD), and the uniaxial tensile test data. These aforementioned
parameters are described below.

1.5.2.1 Normal Anisotropy

The normal anisotropy coefficient (rn) can be used as a measure of formability as

it defines the resistance of the material to thinning. It can be defined by the ratio
of the strain in the width direction to the strain in the thickness direction. Thus, it
is desirable to have a higher r value so that the material will be more resistive to
thinning and delay material failure. The normal anisotropy relates to formability
as the resistance to thinning is critical in sheet forming as the material fails from
local thinning (necking).

1.5.2.2 Limiting Dome Height

The limiting dome height (LDH) is a measure of formability of sheet metal in that
it collectively quantifies parameters such as the n-value, the m-value, and achievable
elongation. This value is limited to the stretching mode and provides more realistic
1.5  Metrics of Formability 19

data as it includes die contact and friction. For the testing, various sheet specimen
widths are tested (get varying strain paths) and the maximum height of the dome is the
LDH value. Thus, the higher the LDH value, the higher the formability of the material.

1.5.2.3 Hole Expansion Ratio

The hole expansion ratio (λ value) is a formability measure that describes the form-
ability of sheet metal near the edges of the part. This ratio is useful for determining
how a material will form in relation to the edges of the blank to be formed. A larger
ratio value is desired as there will be a higher capability for reaching higher strain
without material failure. This ratio and the results are very sensitive to the quality of
the edge and the microstructure of the material, thus testing should probably replicate
the same quality and microstructure that will be used for the actual forming process.

1.5.2.4 Forming Limit Diagram

The forming limit diagram (FLD) is a common conceptual tool used to represent
the forming behavior for a given material while describing the onset of sheet neck-
ing for varying loading configurations. On an FLD, the forming limit curve (FLC)
represents the maximum achievable major principle strains for a given minor prin-
ciple strain (Fig. 1.10). The strain paths shown in the figure from left to right are
pure shear, uniaxial tension, plane strain, stretching, and balanced biaxial stretch-
ing. This concept was introduced by Keeler and Backofen in 1964 [26] for the
positive minor strains and by Goodwin in 1968 [27] for negative minor strains.

1.5.2.5 Uniaxial Tensile Test

The uniaxial tensile test is the most common test for determining the flow characteris-
tics and basic properties used for engineering analysis. Properties that are commonly
derived from this test include Young’s modulus, yield strength, tensile strength, uni-
form elongation, total elongation, strain hardening value, and normal anisotropy.

Fig. 1.10  Schematic of a
Forming Limit Diagram
(FLD)
20 1  Deformation of Metals

1.6 Conclusions

Metal forming is one of the most ancient manufacturing methods. Only in the
last century have the atomic deformation mechanisms such as crystallographic
arrangement of metallic atoms, preferential slip systems, dislocation motion, and
crystal twinning been well modeled and understood. We now have both experi-
mental and theoretical descriptions of plastic deformation, and its limitations as a
process in the areas of forming power and achievable strain before fracture.
We can therefore begin to explore enhancements to the forming process to over-
come these limitations. Traditional enhancements include heating the metal to a sig-
nificant fraction of its absolute melting temperature to improve formability, thermally
annealing material to “reset” its properties and allow further forming, and alloying to
allow improved formability while maintaining desired properties in the end product.
In this work, we examine the Electrically Assisted Forming (EAF) process,
whereby electric current is passed through a part during forming. The added electri-
cal energy works to lower the flow stress, to improve the achievable elongation, and
to relieve residual stress in situ, and at much higher rates-of-change of material prop-
erties than can be achieved with thermal energy. In this book, we seek to empirically
describe what can be achieved with this process and to investigate the atomic mecha-
nisms underlying the benefits. The remainder of Section 1 (Chaps. 1–3) describes the
EAF approach, and how electricity affects metal deformation. Section 2 (Chaps. 4–7)
explores various ways to model the observed effects both empirically and through the
use of first physical principles. Different architectures, both model-based and model-
free, are examined for control of the process with current as a new control variable.
Finally, Section 3 (Chaps. 8–11) gives further considerations of process design for
EAF and the interaction effects related to tribology and microstructure evolution.

References

1. United States Department of Labor (2011) Motor vehicle and parts manufacturing. Bureau of
labor statistics. http://www.bls.gov/oco/cg/cgs012.htm. Accessed 02 Oct 2011
2. Keith D (2010) HSS, AHSS and aluminum jockey for position in the race to cut auto curb
weight. American Metal Market Monthly. February 1
3. Pfestorf M (2009) Multimaterial lightweight design for the body in white of the new BMW
7 series. BMW Group. Innovative Developments for Lightweight Vehicle Structures.
Wolfsburg. May 26–27
4. Krinke S (2009) Life cycle assessment and recycling of innovative multimaterial applica-
tions. In: Conference Proceedings of Innovative Developments for Lightweight Vehicle
Structures. Volkswagen AG, Wolfsburg. May 26–27
5. Ford H (1924) My life and work. In: collaboration with S. Crowther. Kessinger Publishing
Co., Montana, 2003
6. U.S. Environmental Protection Agency (U.S. EPA) (2009a) Light-duty automotive technol-
ogy and fuel economy trends: 1975 through 2008. EPA420-R-08-015
7. Glennan TB (2007) Strategies for managing vehicle mass throughout the development pro-
cess and vehicle lifecycle. Society of Automotive Engineers. 2007-01-1721
References 21

8. Kleiner M, Geiger M, Klaus A (2003) Manufacturing of lightweight components by metal

forming. CIRP Ann Manufact Technol 52(2):521–542. doi:10.1016/S0007-8506(07)60202-9,
ISSN 0007-8506
9. Friesen F, Schwarz D, Cunat P-J (2002) Lightweight design with stainless steel. In:
Proceedings DYNAmore LS-DYNA Forum, Bad Mergentheim, D
10. Friebe E et al (2001) Hydro mechanical deep drawing of passenger car fuel tanks. In: Siegert K
(ed) Hydroforming of tubes, extrusions, and sheets, vol 2, MAT-INFO, Frankfurt, D, pp 193–213
11. Taub A, Krajewski P, Luo A, Owens J (2007) The evolution of technology for materials pro-
cessing over the last 50 years: the automotive example. JOM 59(2):48–57
12. American Iron and Steel Institute: 2009 Press Releases. http://www.steel.org/
13. www.mckinnesrolledrings.com
14. Ross CD, Kronenberger TJ, Roth JT (2009) Effect of DC on the formability of Ti-6AL-4V. J
Eng Mater Technol 131(3):11
15. Ji YH, Park JJ (2008) Formability of magnesium AZ31 sheet in the incremental forming at
warm temperature. J Mater Pro Technol 201:354–358
16. Golovashchenko SF, Krause A, Gillard AJ (2005) Incremental forming for aluminum auto-
motive technology. In: ASME 2005 International Mechanical Engineering Congress and
Exposition, IMECE2005-81069, p 7
17. Kalpakjian S (1997) Superplastic Forming. Manufacturing Processes for Engineering
Materials, Third Edition, pp 421–422
18. Kinsey B, Cao J (2003) An analytical model for tailor welded blank forming. J Manuf Sci
Eng 125:344–351
19. Center for Automotive Research (2011) Automotive technology: greener products, changing
skills, lightweight materials & forming report. Driving Change Project
20. Bandivadekar A, Bodek K, Cheah L, Evans C, Groode T, Heywood J et al (2008) On the
road in 2035: reducing transportation’s petroleum consumption and GHG emissions.
Massachusetts Institute of Technology
21. Callister WD (2007) Materials science and engineering: an introduction, 7th edn. Wiley, New
York
22. Askeland DR, Phule PP (2006) The science and engineering of materials, 5th edn.
Addison-Wesley
23. Kalpakjian S (1997) Manufacturing processes for engineering materials, 3rd edn. Springer
Science & Business Media, 2000
24. Banabic D, Bunge HJ, Pohlandt K, Tekkaya AE (2000) Formability of metallic materials.
Springer, London
25. Xu Y (2004) Universal formability technology and applications. Mater Process Technol
151:119–125
26. Keeler SP, Backofen, WA (1964) Plastic instability and fracture in sheets stretched over rigid
punches. ASM Trans Q 56:25–48
27. Goodwin GM (1968) Application of strain analysis to sheet metal forming problems in the
press shop. SAE Technical Paper
Chapter 2
Introduction to Electrically Assisted
Forming

2.1 Electrically Assisted Forming

Electrically assisted forming (EAF) is a recently introduced metal-forming technique

capable of enhancing a metal’s formability during deformation and reducing spring-
back after deformation. In this technique, electricity is applied to a metal blank while
it is deformed, without stopping the deformation. Electrically assisted manufactur-
ing (EAM) is a general term used for the technique of applying electricity to any
manufacturing process. EAF is a specific type of EAM, where electricity is applied
to metal-forming processes (i.e., bulk deformation or sheet metal forming). A sche-
matic of an EAF (forging) test setup can be seen in Fig. 2.1. The key components of
the test setup include a controllable power supply to generate the electricity, a DAQ
system to collect mechanical data, a thermal camera to collect thermal data, and
insulation to isolate the die/workpiece and machinery from the applied electricity.
The multiple benefits generated from the applied electricity are collectively
known as the electroplastic effect. The three main benefits of the electroplastic effect
are as follows:
• A reduction in the flow stress required to continue plastic deformation
• Increased achievable part deformation prior to failure
• Reduction/elimination of springback effects in formed parts
Figure 2.2 illustrates the electroplastic effect, by showing EAF’s ability to transform
strong and brittle Ti-6Al-4 V into a highly formable material. In Part a, a specimen
compressed under conventional conditions, with no applied electricity, is shown.
After a minimal amount of deformation, the HCP material failed due to brittle shear
fracture. Part b shows a specimen compressed under EAF conditions, where the
electricity was applied for the duration of the forging operation. Due to EAF, the
part was able to be completely formed to its desired height without failure.

© Springer International Publishing Switzerland 2015 23

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_2
24 2  Introduction to Electrically Assisted Forming

p Thermal
camera
Press
Power Insulation Temperature
source Die measurements
+ Workpiece

_~ V DAQ
System

A Position and Force

measurements
p

Fig. 2.1  Schematic of an EAF test setup [1]. During EAF, electricity is applied to a metal while it is
deformed, thus increasing the overall formability of the metal and producing the electroplastic effect

Fig. 2.2  EAF formability improvement (Ti-G5) [1]. EAF increases the achievable amount of part
displacement before failure. a Conventional forming of a Ti-G5 slug led to almost instantaneous shear
failure, b EAF forming of the same material enabled the slug to be deformed to its desired distance

However, like all processing techniques, EAF has some disadvantages and chal-
lenges that exist within the process and its implementation toward industrial use.
First, in order for the electricity to reach the part, there must be some type of electri-
cal applicator in contact with the conductive workpiece at all times. For manufactur-
ing processes where the workpiece is stationary, like forging or stamping, this can
be done rather easily. Conversely, when implementing the EAF technique on man-
ufacturing processes with workpieces of relative motion, like friction welding two
workpieces together, an applicator system which is in continuous contact with the
workpiece while not becoming entangled in the part must first be designed. In addi-
tion, since the workpiece is subjected to electrical flow, all personnel and machine
components should be insulated from the electricity. It must be noted that extreme
caution should always be used when working with electricity. Isolating the electric-
ity, however, can prove to be a challenging task in some cases, since most machin-
ery components are comprised of conductive metals and the common insulating
2.1  Electrically Assisted Forming 25

materials (e.g., nylons, rubbers, ceramics, or plastics) cannot withstand the mechani-
cal demands placed on these components (i.e., too soft or too brittle). Outweighing
these issues, in most cases, are the energy reduction benefits, decreased flow stress,
increased ductility, and reduced springback, that can be accomplished using the EAF
method. Some tooling designs devised to overcome issues with electric current inte-
gration to the forming process are described in Chap. 10.

2.2 EAF Literature Review

Research investigating how electricity affects materials can be traced back to the
mid-twentieth century in Russia. Toward the later part of this century, this research
slowly began in the United States. Now, there are an increased number of universi-
ties and national laboratories which have begun to focus on some portion of the
EAF technique. An in-depth explanation into the history of EAF research will be
provided below.
In 1959, Machlin et al. [2] examined electricity’s effect on group 1A salts (NaCl),
determining that an applied electric current significantly affected the material’s
ductility, flow stress, and yield strength. Later, Nabarro [3] discussed electricity’s
effect on metals as part of his book in 1967. In 1969, Troitskii et al. [4] studied how
electrons influence dislocation motion and reproduction in different alloys of zinc,
tin, lead, and indium, concluding that pulsed electricity could lower the flow stress
within the materials. Years later, in 1982, Klimov et al. [5] explained that the effects
on a metal’s structure from electricity are unrelated to those caused by Joule heating.
Moving forward, in 1988, a microstructure analysis was conducted by Xu et al. [6],
and it was discovered that a continuous electric current in titanium materials caused
the recrystallization rate and the grain size of the materials to increase. Next, Chen
et al. [7, 8] developed a relationship between electric flow and the formation of
intermetallic compounds (Sn/Cu and Sn/Ni systems). Afterward, in 2000, Conrad et
al. [9–11] determined that very high-current density/short-duration electrical pulses
can affect the plasticity and phase transformations of metals and ceramics. In 2005,
Heigel et al. [12] examined the microstructural alterations in Al 6061 resulting from
direct current.
Within the past few years, much experimental research has been performed to
establish how electricity affects the mechanical behavior of different metallic alloys.
In 2007, Andrawes et al. [13] was able to conclude that electrical current can signifi-
cantly reduce the energy needed for uniaxial tensile deformation of Al 6061-T6511
without greatly heating the workpiece. Perkins et al. [14] studied the effects of a
continuously applied electric current on various alloys undergoing an upsetting pro-
cess and found that the electricity increased the amount of allowable compressive
deformation prior to fracture and lowered the required compressive forces. Again
in 2007, Ross et al. [15] examined the application of a continuously supplied elec-
tric current on tensile specimens, only to conclude that, although deformation forces
were reduced, the achievable elongation was decreased, leading to premature failure.
26 2  Introduction to Electrically Assisted Forming

The problem of decreased elongation in EAF-tensile processes was overcome in

2008, when Roth et al. [16] achieved elongation increases of nearly 400 % by apply-
ing square wave pulsed (rather than continuous) current to Al 5754 tensile speci-
mens. Following this, Salandro et al. [17] examined the effect of pulsed electricity
on three different heat treatments of two 5xxx Aluminum Alloys (5052 and 5083).
Moreover, in 2009, Salandro et al. [18] discovered a linear relationship between
current density and pulse duration in Mg AZ31B-O tensile specimens that could
be used to reliably achieve intended elongations for a variety of pulsing conditions.
Research by McNeal et al. [19] examined microstructural alterations in the same
Mg AZ31B-O tensile specimens. Green et al. [20] determined that springback in Al
6111 sheet specimens could be completely eliminated with a single high-current,
short-duration electrical pulse. From work by Jones and Roth [21], achievable com-
pressive displacements of the same Mg alloy were increased by over 400 %, and the
electricity even led to strain weakening effects. Additionally, in 2009, Salandro and
Roth [22] found that, by applying electric pulses to Al 5052 while undergoing highly
localized channel formation, the achievable channel depth could be increased while
reducing the required machine forces. Siopis et al. [23] examined how different
microstructure properties affect the effectiveness of EAF in micro-extrusion experi-
ments. Specifically, it was concluded that a finer-grained material, with more grain
boundaries, enhanced the electroplastic effect, whereas a larger-grained material,
with fewer grain boundaries, lessened the effect. Another work by Siopis et al. [24]
determined that the effectiveness of EAF increased as the dislocation density within
the metal also increased, as a result of cold-working prior to EAF experiments. A
work by Dzialo et al. [25] examined the effect of current density and zinc content
during electrical-assisted forming of copper alloys. A more in-depth overview of the
development of EAF can be found in [26]. Additionally, several recent EAF patents
were found as a part of the EAF literature review [27–31].
Overall, the effort and number of researchers studying EAF in the USA have
increased tremendously since Roth began experimentally analyzing EAF in the
mid-2000s. Shown in Fig. 2.3 is a timeline displaying both the researchers and
universities that are involved in some type of EAF research (note the exponentially
increasing trend).

2.2.1 EAF Theory and Modeling

Due to the lack of knowledge about the electroplastic effect, past research-
ers have been unsuccessful in accurately modeling and predicting EAF effects
for process control. However, from the previous work in this field, a multi-part
postulated theory can be explained. At the microstructure level, metals are held
together by metallic bonds, consisting of clouds of electrons, which surround ion
cores containing protons and neutrons. Because of this, it is realistic that the appli-
cation of electricity (i.e., the application of flowing electrons) to any metal will
have noticeable effects. Specifically, when electricity is applied to a metal during
2.2  EAF Literature Review 27

Machlin, Nabarro,
Klimov, Kravschenko
Kronenberger, Bunget
,
Salandro, Jones

S. Wagner
Modeling Early Work
J. Zhou
Springback Microstructure J. Jesweitt

Perkins, Green Xu, Chen, Conrad,

J. Cao
EAM Researchers

Formability Heigel, McNeal

B. Kinsey
Experiments
T. McNeal
T. Kronenberger
Andrawes, Perkins, C. Green
Ross, Roth, Salandro, C. Ross
Jones J. Jones
J. Roth
C. Bunget
Troitskii T. Perkins
L. Mears
Nabarro Xu Conrad Andrawes
W. Salandro
Machlin Kravchenko Klimov Chen Heigel

1950 1960 1970 1980 1990 2000 2005 2010

Time Period

Fig. 2.3  EAF research timeline. The number of researchers and universities investing time and
money into different aspects of EAF research has increased exponentially over the past decade

deformation, a few phenomena occur simultaneously, thus transforming the mate-

rial into an easier-to-deform state, known as the electroplastic effect. This effect
has been attributed to the following aspects:
• Localized atomic-level resistive heating effects that are enhanced by the resistivity
of the material (i.e., electrons scatter off of interfacial defects within the lattice,
such as voids, impurities, grain boundaries) [16]. It is important to remember that
this heating occurs on the atomic level (within the metal’s lattice), and although
this contributes toward the overall heating of the workpiece, this temperature
increase is not the same as the bulk temperature increase that is witnessed at the
part’s surface (known as global or bulk heating). Specifically, the bulk tempera-
ture of a metal is the result of all the atomic-level heating locations. This effect
expands the local lattice and allows for easier dislocation motion (i.e., plas-
tic deformation) by way of enhanced diffusion. The resistive heating effects are
dependent upon resistivity; hence, a material with a greater resistance will expe-
rience larger amounts of localized resistive heating and will potentially achieve
greater formability benefits when the EAF technique is applied.
• Direct dislocation–electron interaction takes place when the flowing electrons
impact the dislocation lines, assisting in “pushing” the dislocation lines and fur-
ther enhancing plastic deformation and material ductility [32]. Kravchenko
[33], in his explanation of electroplasticity, succinctly stated this effect when he
explained that, if there is an electric current flowing and the electrons are trave-
ling at a faster rate than the dislocations within the lattice, the energy from the
electrons is transferred to the dislocations, thus making the plastic flow easier.
The overall impact of this effect can be significant or minimal, depending on the
28 2  Introduction to Electrically Assisted Forming

direction of the flowing electrons and the direction of deformation. This aims to
explain why the temperature of an EAF test, where electricity is applied during
deformation, is less than a stationary electrical test, where the electricity is applied
when no deformation takes place. In the EAF test, some of the energy is used to
assist plastic deformation, instead of fully contributing toward resistive heating.
• The addition of excess electrons to the metal’s microstructure is an important aspect.
Since the electron clouds control how strongly a metal is bonded and essentially
act as the “glue” which holds a metal together, the excess electrons (obtained from
applying the electricity) will assist in breaking and reforming bonds by reducing the
bond strength between electrons. As the metallic bonds are able to break and reform
easier, the ductility of the metal is improved; hence, it becomes more workable [34].
Figure  2.4 displays a schematic describing the three main pillars of the electro-
plastic effect explaining the EAF theory.
As previously stated, there have been several unsuccessful attempts at modeling
EAF; however, these attempts have helped to bring a better understanding to EAF.
In [36], a finite element (FE) model, capable of accurately predicting resistive
heating and isothermal forming effects, was considerably unsuccessful in simulat-
ing material behavior in an EAF compression test, as shown in Fig. 2.5.
Additionally, in [37], isothermal compression tests were run at temperatures
above the maximum temperature reached during EAF tests, concluding that the
isothermal tests accounted for about only 10 % of the formability improvement
witnessed in the EAF tests (Fig. 2.6). While these works helped to disprove the
common misbelief that EAF’s effect is due solely to temperature, they emphasize
the fact that the EAF theory is not fundamentally understood and EAF effects can-
not be effectively predicted.

2.2.2 Significant EAF Modeling Variables from

Experimentation

The previous experimental works on EAF highlight the important material- and
process-related parameters for EAF. The following lists of important variables are

Fig. 2.4  EAF research
theory summary [35]. Electroplastic Effect
From the previous research
performed on EAF, the
three effects listed above
are theorized to produce the
electroplastic effect in metals

Localized Electron-Dislocation Addition of

Resistive Heating Interaction Electrons
2.2  EAF Literature Review 29

Fig. 2.5  FEA modeling
of EAF [36]. In previous
research, generic FEA
modeling of EAF proved
unsuccessful. a The FEA
program was capable of
predicting resistive heating
temperature profiles, b the
FEA model proved capable
of predicting a stress–strain
profile for an isothermal test,
c the FEA model, which
accounted solely for resistive
heating, was shown to be
highly inaccurate when trying
to predict an EAF stress–
strain profile

derived from the experimental works on EAF, as well as from conversations with
experts in related fields who have shared their opinions on the electroplastic effect.
The material properties of importance to an EAF process are as follows: 1
thermal conductivity (k), 2 density (ρ), 3 specific heat (Cp), 4 heat transfer coef-
ficient (h), 5 starting strength coefficient of the material (C), 6 strain hardening
exponent of the material (n), 7 resistivity (r), and 8 the initial grain structure of the
metal. Each of the material-based properties is significant when modeling any heat
30 2  Introduction to Electrically Assisted Forming

Fig. 2.6  Experimental EAF modeling [37]. A previous research work proved that the stress–
strain profiles for an EAF test, and an isothermal test run at the maximum temperature reached
during the EAF test, were considerably different. Additionally, this confirmed that the effects of
EAF were not solely contributable to resistive heating or thermal softening

transfer or thermodynamic phenomenon, or performing any mechanical modeling

[35]. The effects/relations of these inputs are described below:
• The thermal conductivity (k), density (ρ), specific heat (Cp), heat transfer coefficient
(h), and resistivity (r) all affect the heat transfer and ultimately the stress–strain
characteristics of the EAF process. Additionally, each of these variables change as a
function of temperature, so depending on the temperatures reached during an EAF
process, these variables could have weighted effects. These intrinsic properties are
not only important for the workpiece, but it is also critical to know these properties
for the forming dies as well.
• The strength coefficient (C) and strain hardening exponent (n) are intrinsic
properties that determine the magnitude and shape of the forming load profile
of an EAF test (and any forming test in general). Further, both are affected by
the temperature of the workpiece and forming dies in the process. Any type
of metal deformation modeling would need to include the effects of both the
strength coefficient and strain hardening exponent at a minimum.
• The initial grain structure (i.e., grain size, grain direction) of a material can affect
the heating and mechanical characteristics of a workpiece during EAF. The grain
size dictates how often moving dislocations must pass through grain boundaries
which cause dislocation pile-ups and can limit achievable deformation. In addi-
tion, the applied electrons must also pass through the grain boundaries and the
grain size (dictating the number of boundaries) will potentially cause the work-
piece to become hotter (more boundaries) or cooler (less boundaries).
2.2  EAF Literature Review 31

Fig. 2.7  Electrical threshold 70
versus material resistivity

Electrical Threshold (A/mm 2)

60
comparison for several Al 6061-T6511 (BCC)
lightweight metals [35]. 50 MG AZ31B-O (HCP)
The figure depicts that, as
304 SS (BCC)
the material resistivity is 40
Ti-6AL-4V (HCP)
increased, the electrical 30
threshold decreases
20

10

0
0.0 5.0 1.0 1.5 2.0
0E 0E 0E 0E 0E
+0 +0 +0 +0 +0
0 7 6 6 6
Resistivity (Ω m)

• The resistivity has a direct correlation with the electrical threshold current den-
sity, as seen in Fig. 2.7. For the same die speed, metals with a higher resistivity
require a lower electrical threshold to produce significant formability improve-
ments. This could be related to the first part of the electrical theory, where the
flowing electrons scatter off of the lattice obstacles and cause localized atomic
heating. A material with a higher resistivity will have a greater number of lattice
obstacles and will result in a greater amount of localized heating around these
obstacles, which ultimately lowers the electrical thresholds of these metals.
The process-related variables to be presented are as follows: (1) initial dimensions
of the workpiece (ro and ho), (2) deformation speed (i.e., die speed), (3) current
density (current per normal area), (4) applied voltage (V), (5) workpiece/die con-
tact area, (6) electrical application method, and (7) initial percent cold work. These
additional effects are detailed below:
• The initial dimensions of the workpiece determine the magnitude of current
needed for EAF. It was determined that the electroplastic improvements are a
function of current density and not current magnitude, so the current density
will determine the appropriate current magnitude to use.
• The deformation speed is important because the EAF technique is strain rate-
dependent, and therefore, the electrical application parameters (starting current
density) must be adjusted if the die speed is increased or decreased.
• The current density and applied voltage make up the applied electrical power to
the process. In an EA-forging process where the electricity is applied continu-
ously, these variables must be adjusted to produce a desired amount of electrical
power.
• The effect of the actual contact area between the workpiece and die was not
previously explored experimentally; however, since the dies and workpiece are
separate parts, are composed of different materials, and must both transfer elec-
tricity during the EAF process, this variable is to be explored. Additionally, the
roughness between surfaces is a widely studied topic in the field of electrical
connectors.
32 2  Introduction to Electrically Assisted Forming

• Electricity can be applied to a deformation process in many ways. The work by

Ross et al. [15] showed that the electricity must be applied differently for com-
pressive and tensile processes.
• The percent cold work within a metal generally determines the dislocation density
within that metal. As the dislocation density is increased, there are more disloca-
tion pile-ups and the achievable deformation can become limited. It is theorized
that the flowing electrons directly affect the dislocations within the metal’s lattice.

2.2.3 Relation to Crystal Structure and Resistivity

As seen in the work by Perkins et al., for a specific die speed, each material has
a specific electrical threshold (i.e., a current density where significant formabil-
ity improvements due to the applied electrical power are observed) [14]. Table 2.1
shows several lightweight material properties (crystalline structure and resistiv-
ity), along with the electrical thresholds, which were experimentally determined
using data from works by Perkins et al. [14] and Jones and Roth [21]. Figure 2.7
shows the relationship between the material resistivity and the electrical threshold
current density. The calculated electrical thresholds for each of these materials are
from [14] and [21], where the specimens were deformed at 25.4 mm/min. It can
be noted that, as the material resistivity is increased, the electrical threshold cur-
rent density decreases (see Fig. 2.7). This supports the theory of localized heating
from electrons scattering off of lattice obstacles and allowing the lattice to expand
easier. Also noted in the figure are the crystalline structures of the specific materi-
als. It can be seen that, for these four metals, there is not a relation between the
crystalline structure and the electrical threshold.

Table 2.1  Lightweight materials and respective electrical thresholds [38–41]

Material Electrical Crystalline Resistivity Threshold Specimen Def. rate
threshold structure (Ωm) reference size (mm) (mm/min)
(A/mm2)
Al 60 BCC 3.99E-08 Perkins et 9.5 × 6.4 25.4
6061-T6511 al. (dia)
Mg 30 HCP 9.20E-08 Jones et al. 12 × 7.5 25.4
AZ31B-O (dia)
304 SS 18 BCC 7.20E-07 Perkins et 9.5 × 6.4 25.4
al. (dia)
Ti-6Al-4 V 20 HCP 1.78E-06 Perkins et 9.5 × 6.4 25.4
al. (dia)
2.2  EAF Literature Review 33

2.2.4 Electroplasticity and Electromigration

The above-described effects are shown experimentally to be due not just to bulk
thermal influences. There is an effect on material flow stress that appears to be due
solely to direct electrical influence, beyond what would be expected from tempera-
ture effects. This electroplasticity effect is one of the key phenomena associated with
EAF, and what we attempt to describe in the modeling chapters in later chapters.

2.3 Broader Impacts of EAF

The applications (or industries) which could use EAF will be discussed first, fol-
lowed by potential users of the EAF thermo-mechanical predictive models.

2.3.1 Automotive and Aerospace Industries

Like other formability-enhancing manufacturing techniques, EAF is not the opti-

mum technique to use for all metals or all part designs. Specifically, during EAF,
excess electrical power needs to be supplied to the workpiece, which does not
always make it the most efficient process. However, the results from using EAF
are significant and it may be one of the very few techniques that allow efficient
forming of particular metals. For this reason, EAF should be used to form metals
and alloys which are currently not able to be formed to great lengths or require
excessive heating or annealing. The EAF technique would act as a gateway for
these metals to be used in industry. Two metals whose formability improves tre-
mendously using EAF are magnesium and titanium, which are targeted by the
automotive and aerospace industries, respectively.
With the rising fuel and operational costs, the automotive and aircraft industries
are becoming more weight-, performance-, and efficiency-focused. One way to
achieve all three variables is by lightweighting. In this technique, lighter and stronger
materials are used instead of the heavier carbon steels mainly used today. Magnesium
is a desirable material for the automotive industry, where it is currently used in mainly
cast components due to its very low formability. On the other hand, titanium is popu-
lar for use in aerospace applications, but the manufacturers constantly struggle with
the poor formability and high required forming forces of this material. Again, previ-
ous research has shown that EAF significantly improves the formability of titanium,
which will increase the number of potential aerospace applications for the alloy.
34 2  Introduction to Electrically Assisted Forming

2.3.2 Potential Early Adopters of EAF Modeling

Now that the main industries that could benefit from EAF have been identified, it is
also important to explain how the predictive model could be used. There are two main
potential “Early Adopters” that could be interested in a modeling concept for EAF.

2.3.2.1 Simulation Software Companies

• These simulation software companies already produce software that is able to

predict outcomes of many current manufacturing processes. They already have
all of the general algorithms/methodologies needed for conventional forming.
By integrating the main algorithms generated by this research into their soft-
ware package, these companies would be able to sell EAF-predictive software.
• There are different applications of simulation. The main simulation applicabil-
ity would be for metal forming; however, there is also the potential to simulate,
or model, alternative EA processes. Such process could be EA machining, EA
bending, or EA joining.

2.3.2.2 Metal-Forming Companies

• The predictive model can be used for EAF process design, where the speed,
electrical settings, and die design will be optimized. The Tier I and Tier II metal-
forming suppliers will also be probable early adopters because the EAF technique
may be their chosen formability-enhancing technique for forming Mg and Ti, as
explained earlier. This is where the “heart” of metal forming is and each metal-
forming supplier wants to ensure that they are not overtaken by new technology of
a competitor. Of note is that the automotive/aircraft OEM’s would not be consid-
ered early adopters because they want something immediately and that is 100 %
dependable. It is safer to market to the suppliers because they are more likely to
work with some “growing pains” of a new manufacturing process.

References

1. Salandro WA, Bunget C, Mears L (2011) Thermo-mechanical investigations of the electroplas-

tic effect. In: International manufacturing science and engineering conference, MSEC2011-
50250, p 10
2. Machlin ES (1959) Applied voltage and the plastic properties of “brittle” rock salt. J Appl
Phys 30(7):1109–1110
3. Nabarro FRN (1967) Theory of crystal dislocations. Chapter IX
4. Troitskii OA (1969) Electromechanical effect in metals. Pis’ma Zhurn Experim Teoret Fiz,
No 10, pp 18
5. Klimov KM, Novikov II (1982) The “electroplastic effect”. A.A. Baikov Institute of Metallurgy,
Academy of Sciences of the USSR, Moscow. Translated from Problemy Prochnosti, No 2, pp
98–103
References 35

6. Xu ZS, Lai ZH, Chen YX (1988) Effect of electric current on the recrystallization behavior
of cold worked alpha-ti. Scr Metall 22:187–190
7. Chen SW, Chen CM, Liu WC (1998) Electric current effects upon the Sn/Cu and Sn/Ni inter-
facial reactions. J Electron Mater 27:1193
8. Chen SW, Chen CM (1999) Electric current effects on Sn/Ag interfacial reactions. J
Electronic Mat 28:902
9. Conrad H (2000) Effects of electric current on solid state phase transformations in metals.
Mater Sci Eng A287:227–237
10. Conrad H (2000) Electroplasticity in metals and ceramics. Mat Sci Eng, pp 276–287
11. Conrad H (2002) Thermally activated plastic flow of metals and ceramics with an electric
field or current. Mat Sci Eng A322:100–107
12. Heigel JC, Andrawes JS, Roth JT, Hoque ME, Ford RM (2005) Viability of electrically treat-
ing 6061-T6511 aluminum for use in manufacturing processes. Trans North Am Manuf Res
Inst SME 33:145–152
13. Andrawes JS, Kronenberger TJ, Roth JT, Warley RL (2007) Effects of DC current on the
mechanical behavior of AlMg1SiCu. Mater Manuf Proc 22(1):91–101
14. Perkins TA, Kronenberger TJ, Roth JT (2007) Metallic forging using electrical flow as an
alternative to warm/hot working. J Manuf Sci Eng 129(1):84–94
15. Ross CD, Irvin DB, Roth JT (2007) Manufacturing aspects relating to the effects of DC cur-
rent on the tensile properties of metals. J Eng Mater Technol 129(2):342–347
16. Roth JT, Loker I, Mauck D, Warner M, Golovashchenko SF, Krause A (2008) Enhanced
formability of 5754 aluminum sheet metal using electric pulsing. Trans North Am Manuf Res
Inst SME 36:405–412
17. Salandro WA, Jones JJ, McNeal TA, Roth JT, Hong ST, Smith MT (2008) Effect of electrical
pulsing on various heat treatments of 5xxx series aluminum alloys. International manufactur-
ing science and engineering conference, MSEC 2008-72512, p 10
18. Salandro WA, Khalifa A, Roth JT (2009) Tensile formability enhancement of magnesium
AZ31B-O alloy using electrical pulsing. Trans North Am Manuf Res Inst SME 37:387–394
19. McNeal TA, Beers JA, Roth JT (2009) The microstructural effects on magnesium alloy
AZ31B-O while undergoing an electrically-assisted manufacturing process. International
manufacturing science and engineering conference, MSEC 2009-84377, p 10
20. Green CR, McNeal TA, Roth JT (2009) Springback elimination for Al-6111 alloys using
electrically-assisted manufacturing (EAM). Trans North Am Manuf Res Inst SME,
37:403–410
21. Jones JJ, Roth JT (2009) Effect on the forgeability of magnesium AZ31B-O when a con-
tinuous DC electrical current is applied. International manufacturing science and engineering
conference, MSEC 2009-84116, p 10
22. Salandro WA, Roth JT (2009) Formation of 5052 aluminum channels using electrically-
assisted manufacturing (EAM). International manufacturing science and engineering confer-
ence, MSEC 2009-84117, p 9
23. Siopis MS, Kinsey BL (2009) Experimental investigation of grain and specimen size effects
during electrical-assisted forming. International manufacturing science and engineering con-
ference, MSEC2009-84137, p 6
24. Siopis MS, Kinsey BL, Kota N, Ozdoganlar OB (2010) Effect of severe prior deformation
on electrical-assisted compression of copper specimens. International manufacturing science
and engineering conference, MSEC2010-34276, p 7
25. Dzialo CM, Siopis, Kinsey BL, Weinmann KJ (2010) Effect of current density and zinc
content during electrical-assisted forming of copper alloys. CIRP Ann—Manuf Technol
59(1):299–302
26. Salandro WA, Roth JT (2010) Ch19. electrically-assisted manufacturing. In: Zhang W (ed)
Intelligent energy field manufacturing: interdisciplinary process innovations. CRC Press,
Boca Raton
27. Bunget C, Salandro WA, Mears L (2011) Thermal mechanical predictive algorithm for elec-
trically-assisted manufacturing processes (Provisional patented filed on May 25, 2011)
36 2  Introduction to Electrically Assisted Forming

28. US 7,302,821—Techniques for manufacturing a product using electric current during plastic
deformation of material
29. US 7,516,640—Method and apparatus for forming a blank as a portion of the blank receives
pulses of direct current
30. Electrically Assisted Single-Point Incremental Forming (EA-SPIF), (non-provisional applica-
tion, disclosure #3484)
31. Electrically Assisted Metal Forging Process, (non-provisional application, disclosure #3314)
32. Yao L, Hong C, Yunquo G, Xinbin H (1996) Effect of electric current pulse on superplastic-
ity of aluminum alloy 7475. Trans of Nfsoc 6(1):77–84
33. Kravchenko V (1966) JETP (USSR) 51:1676
34. Antolovich SD, Conrad H (2004) The effects of electric currents and fields on deforma-
tion in metals, ceramics, and ionic materials: an interpretive survey. Mater Manuf Processes
19(4):587–610
35. Salandro WA (2012) Thermo-mechanical modeling of the electrically-assisted manufacturing
(EAM) technique during open die forging. PhD dissertation, Clemson University
36. Kronenberger TJ, Johnson DH, Roth JT (2009) Coupled multifield finite element analysis
model of upsetting under an applied direct current. J Manuf Sci Eng 131:031003
37. Ross CD, Kronenberger TJ, Roth JT (2009) Effect of DC on the formability of Ti-6AL-4 V. J
Eng Mater Technol 131(3):11
38. MatWeb Aluminum 6061-T6; 6061-T651. MatWeb Material Property Data,
www.matweb.com. Accessed 01 July 2012
39. MatWeb Magnesium AZ31B-O, Annealed Sheet. MatWeb Material Property Data,
www.matweb.com. Accessed 01 July 2012
40. MatWeb 304 Stainless Steel. MatWeb Material Property Data, www.matweb.com. Accessed
01 July 2012
41. MatWeb Titanium Ti-6Al-4 V (Grade 5), Annealed. MatWeb Material Property Data,
www.matweb.com. Accessed 01 July 2012
Chapter 3
The Effect of Electric Current on Metals

This chapter describes the fundamentals behind electroplasticity in metals.

Specifically, it focuses on electrical current flow, previous electroplastic theories,
and an overall explanation of the electroplastic effect on metals. This overall theory
will be supported with experimental results, and electroplastic conclusions will be
drawn at the end of the chapter.

3.1 Electrical Current Flow

When an electric field is applied to a material, there is a force exerted on the free
electrons (i.e., valance electrons) such that they experience acceleration in the
direction opposite to the electrical field as a result of their negative charge. Ideally,
the electrons would continuously accelerate such that the current would always
increase over time. However, internal friction forces (i.e., electron collisions with
ion cores) within the material limit electron acceleration, which settles at some
constant current value. These collisions in the lattice make up the electrical (vol-
ume-specific) resistivity of the material. The electrical resistivity of a material is
characterized by the atomic structure, spacing, and bonding. However, the electri-
cal resistivity is increased by the number of dislocations, point defects, and inter-
facial defects (e.g., grain boundaries, cracks, voids) within the lattice. The total
electrical resistivity can be described by Matthiessen’s Rule:
ρe_total = ρo + ρi + ρd (3.1)
where ρe_total is the total electrical resistivity, ρo is the ideally pure and perfect
crystal resistivity which includes the influence of thermal vibration contribution, ρi
is the contribution due to impurities in the lattice, and ρd is the contribution from
plastic deformation [1]. Also, it is assumed that the scattering mechanisms act
independently within the material.

© Springer International Publishing Switzerland 2015 37

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_3
38 3  The Effect of Electric Current on Metals

As the electrical field is applied, the electrons accelerate and scatter off the
above-listed defects and the vibrating ion cores themselves. The localized scatter-
ing on the defects causes the electrons to lose kinetic energy and to change their
direction. Yet, the electrons still have a net movement (i.e., current) in the opposite
direction of the applied field. This net movement can be described by the electron
drift velocity, which is the average electron velocity in the direction of the applied
force. The electron drift velocity is given as
I
vd = (3.2)
n|e|A
where I is the current magnitude, n is the number of valence electrons per unit
volume, e is the charge of an electron, and A is the cross-sectional area that the
current passes through [1]. The drift velocity is on the order of a few mm/s for the
electrical current magnitudes during EAF.
The concept of electrons scattering off the material defects and the ion cores is
known as Joule or resistive heating. As the electrons are accelerated by the electric
field, they accelerate and only reach a velocity that is usually below the Fermi veloc-
ity (~1,800,000 m/s) and well below the speed of light (300 million m/s), as a result
of collisions with the material lattice. The Fermi velocity is the fastest possible veloc-
ity of an electron in a metal that is cooled to near zero Kelvin. Thus, at zero Kelvin,
the Fermi velocity of an electron is derived from the kinetic energy equal to the Fermi
energy. During the collisions, the electrons transfer kinetic energy to the ion cores,
which increases the ion cores vibrational energy. This increase in vibrational energy
causes the material to increase in temperature. Thus, when considering a larger mass
of the material, ion cores around the material defects will have greater vibrational
energy (i.e., greater temperature) due to lattice distortions and a greater frequency of
electron/ion core interaction. There is a greater frequency of electron/ion core interac-
tion due to the misalignment of the ion cores. In comparison, the defect-free lattice
regions will have a smaller vibrational energy increase due to the same applied elec-
tron flux through the lattice. Although the flux is the same, the defect region will not
incur as many electron/ion core interactions due to the aligned lattice structure. As a
result, the energy increases will be less. In addition, the vibrational energy gained in
each of the regions (i.e., defect and defect free) will provide or gain energy from its
neighboring ion cores. This creates vibrational energy gradients or thermal gradients
at the ion core level. From a lattice perspective, this translates to an average vibra-
tion energy of the individual ion cores within a grain (i.e., mean grain temperature).
Overall, the collection of mean grain temperatures and heating at grain boundaries
relates to the macro-observed temperature. This macro- or bulk temperature is what is
typically measured during experimental testing; however, there are higher peak tem-
peratures (i.e., vibrational energy) around defects within the material lattice.
When comparing Joule heating to raising the temperature of a material by con-
vection (e.g., in an oven), the average vibrational energy of the ion cores in the
lattice would increase. However, the vibrational energy would not have areas with
greater amounts of energy around the defect sites as there is not a direct inter-
action as with the electrical flow. Thus, heating by convection will provide a
3.1  Electrical Current Flow 39

Fig. 3.1  Edge dislocation represented as a cylindrical dislocation core surrounded by a defect-

free lattice

transient wave of vibrational energy from the exterior surface. However, once the
material is completely heated and soaked at the elevated temperature, the vibra-
tional energy will be uniform within the material’s lattice. This is one beneficial
aspect to using EAF over conventional elevated temperature forming.
Aside from using bulk observations to quantify this phenomenon [2–4], this
chapter introduces physics-based models to determine the significance of the
present electroplastic theories. Specifically, the transient energy provided to the
dislocation core and that transferred to the surrounding lattice are compared and
quantified. A schematic is shown in Fig. 3.1 where an edge dislocation is repre-
sented by a cylindrical dislocation core. The core geometry is characterized by a
right circular cylinder with a diameter (D) and length (L). The diameter used in
this work provides an equivalent area to the actual dislocation core area which is
represented by an elliptical cross section.

3.2 Previous Electroplastic Theories

The two primary theories for electroplasticity are localized heating at lattice defects
[5–7] and the electron wind effect [8–10]. The most recent work on electroplastic
theory suggests that the two phenomena occur simultaneously when an electric cur-
rent is applied during deformation [4]. This work compares the energy magnitude of
these two as related to the movement of a dislocation core in a metal’s lattice.

3.2.1 Localized Heating

The localized heating is a result of increased scattering at defects, which creates

areas of greater atomic vibrations or “hot spots” (i.e., the Joule heating effect
increased at defect sites), whereas the electron wind effect is based on actual
40 3  The Effect of Electric Current on Metals

Table 3.1  Magnesium ρo (defect-free lattice) 4.101 µΩ cm

material and lattice
ρd (dislocation core) 28.707 µΩ cm
parameters [1, 11–13]
ρ 1740 kg/m3
D 0.587 nm
A 0.271 nm2
L 6000 nm
V 1,623.743 nm3
n 4.309 × 1028 Atoms/m3
Ncore 69,921 Atoms
Q 135 KJ/mol
Z* −2
e −1.602 × 10−19 C
b 3.209 × 10−10 m

momentum transfer to the dislocation core. In the following two sections, a case
study is performed for each theory assuming that a current density (J) of 100 A/
mm2 is applied to a pure magnesium metal such that the magnitude of influence
can be determined. Critical material and lattice parameters are given in Table 3.1.
The localized energy provided to the dislocation core due to the Joule heating
effect is given by
EJoule = J 2 ρd V �t (3.3)
where J is the current density, ρd is the electrical resistivity of the dislocation core,
V is the core volume, and t is the time the current is applied. It is hypothesized
that these local “hot spots” ease dislocation movement through the lattice and
allow for them to pass by other un-movable lattice defects. Using a current density
of 100 A/mm2 applied for one second, this results in 4.657 × 10−15 J or 29,069 eV
of energy being applied to the dislocation core. The dislocation core electrical den-
sity was determined to be approximately six to eight times the electrical resist-
ance of the defect-free lattice [14]. This was determined by Kino et al. [14] where
careful experiments were performed to measure samples with varying dislocation
densities. For this analysis, a factor of seven was used during the calculations as
it is the mean of the results published by Kino et al. The importance of a quantifi-
able amount of energy applied to the core is that it can be compared to the activa-
tion energy (Q) for lattice diffusion in magnesium. The lattice diffusion activation
energy is the required energy to move an ion core from one lattice site to another
during deformation. Thus, the activation energy for magnesium is approximately
1.4 eV/atom and this equates to an activation energy of 97,867 eV for an entire
dislocation core. Therefore, the calculated additional energy due to Joule heating
at the dislocation core is slightly less than 1/3 the total activation energy required
to move the dislocation core one atomic distance. From a magnitude standpoint,
this would have a significant effect on the mobility of the core and reduce the
mechanical stress required to displace the dislocations in the material’s lattice. It
3.2  Previous Electroplastic Theories 41

should be noted that the entire dislocation core does not move all at one time, but
regions of the core advance through the lattice over time. This does not affect the
results presented as they are examining the magnitude between Joule heating and
the electron wind effect.
In addition, the localized Joule heating effect can be quantified as a temperature
rise by

J 2 ρe
�T = �t (3.4)
ρc

where T is the temperature rise, J is the current density, ρe is the electrical resis-
tivity of the area of interest, T is the change in time, ρ is the density, and c is the
specific heat. Due to the difference in electrical resistivity between the defect-free
lattice and dislocation core, the temperature magnitude is linearly scaled from the
variation in resistivity (i.e., dislocation core is seven times hotter than the defect-
free lattice) as per Eq. (3.4). To characterize this effect, a simplified model which
includes transient conductive heat transfer from a 2D nodal mesh is shown in
Fig. 3.2.
This model was produced to understand the heat generation and dissipation
during Joule heating around a dislocation core. As seen, the dislocation core in the
center generates the greatest heat and is dissipated outward from the core center.
Future use of this model could allow for the inclusion of additional defects (dis-
locations, point defects, and interfacial defects) such that a mean or bulk tempera-
ture could be calculated over a larger area. This bulk temperature from the model
would allow for a comparison to experimental thermal results. Additionally, as
deformation is imposed, newly formed dislocations could be incorporated into the
model with some dislocation distribution to quantify the addition of new disloca-
tions on the heat transfer response.

Fig. 3.2  Snapshot of
transient response of joule
heating as a result of the
greater dislocation core
electrical resistance
42 3  The Effect of Electric Current on Metals

3.2.2 Electron Wind Effect

Conversely, the energy imparted to the dislocation core from the electron wind is
Ewind = Z ∗ eρe JNcore b (3.5)
where Ewind is the electron wind energy imparted to the dislocation core, Z ∗ is
the effective charge number, e is the charge of an electron, Ncore is the number of
equivalent atoms per dislocation core, and b is an atomic distance [12]. It is sug-
gested that the momentum transfer from the electric field directly assists the dislo-
cation movement within the metal’s lattice. This model is evaluated by examining
the added energy to the dislocation core as a result of momentum transfer. Using
the values in Table 3.1, the total energy imparted on an individual atom for a cur-
rent density of 100 A/mm2 is 1.842 × 10−8 eV. This equates to 1.29 × 10−3 eV
per dislocation core. This result is significantly less than that calculated due to
Joule heating on the dislocation core (225 million times less). As a result, this
effect is extremely small as compared to the energy required to allow for the dis-
location core to diffuse in the material’s lattice. Overall, it is concluded that this
effect is not substantial in aiding dislocation movement within the metals lattice.
It should be noted that the electron wind force or electromigration has been seen
in semiconductor interconnects and integrated electric circuits. However, this
involves larger current densities (~1,000 A/mm2) and only individual ion cores
are moved over time such that voids form in the circuit [12]. The void formation
then opens the circuit which causes the interconnect to fail. A failed interconnect
is shown in Fig. 3.3 where void growth and stress caused the interconnect to fail.
Hence, from the prior analysis, it is concluded that the main contribu-
tion toward the observed electrical effects is due to localized areas of increased

Fig. 3.3  Failed interconnect
by electromigration [15]
3.2  Previous Electroplastic Theories 43

vibrational energy from electron scattering (i.e., Joule heating) and not a direct
momentum effect on the dislocation cores themselves for EAF. Therefore, the
concept of localized areas of greater vibrational energy is discussed below for the
interaction of these “hot spots” and dislocations without and with deformation
being imposed on the material. This proposed theory is different in that it only
attributes the observed effect (flow stress reduction and improved formability) to
one single phenomenon and that the theory is described from a material science
viewpoint.

3.3 Comprehensive Electroplastic Theory Explanation

In the following section, the electroplastic theory is explained from a material sci-
ence view such that a single physical contribution due to Joule heating explains
the observed effects of an applied current on the mechanical properties of a metal.
The influence the applied current has on the material with no deformation, during
deformation, and the current’s influence on material formability is presented.

3.3.1 Electrical Current Without Metal Deformation

When a metal is stationary (i.e., no deformation being imposed) and an electric

current is applied, Joule heating occurs which creates local regions of greater
atomic vibrations (i.e., temperatures) around defects within the material as com-
pared to defect-free regions. These areas with greater temperature can be called
“hot spots” within the material’s lattice. Thus, it is theorized that the “hot spots”
allow for a rapid decrease in the stored energy of the material by facilitating dis-
location annihilation. The motion to reduce the number of dislocations is a direct
result of the enhanced atomic diffusion due to the “hot spots.” This method of
reducing the residual stress by removing dislocation is expected to be faster than
by the conventional stress relief anneal using an oven. This is due to the difference
in bulk convection heating and heating locally with an electric current field. The
bulk convection heating requires energy input that has to heat the entire material
to allow for the dislocations to diffuse to a sink (e.g., grain boundary), whereas the
electric current provides a greater amount of energy directly at the dislocation pre-
sumably in a faster manner. As a result of the removal of residual stresses and the
annihilation of dislocations, the yield strength is expected to decrease. The yield
strength is classically related to the grain size by the Hall–Petch relationship:
1
σ = σo + kd − 2 (3.6)

where σ is the yield strength, σo is the frictional stress (i.e., equivalent to yield
stress of very course-grained polycrystal), k is constant which can be thought of as
44 3  The Effect of Electric Current on Metals

the source strength for dislocations, and d is the mean grain size [16]. Assuming
that a dislocation moves on average x̄ per unit strain, then
x̄ = βd (3.7)
where β is a fraction of the grain size (β < 1). Then, the strain can be said to be

�
ε = ρ bx̄ (3.8)

where ρ is the dislocation density and b is the Burger’s vector. Thus, the disloca-
tion density is
ε ε
ρ= = (3.9)
�
bx̄ �
bβd
The dislocation density relation to strength was derived by Kocks [17] as
1
� 2
σ = σo + αGbρ (3.10)

where α is a constant and G is the shear modulus. Upon substitution of Eq. (3.9)

into Eq. (3.10), this yields
 1
ε 2
σ = σo + αGb� (3.11)
�
bβd
Upon comparing Eq. (3.11) to the Hall–Petch equation [Eq. (3.6)], it can be seen
that the constant k is equivalent to
 1
ε 2
k = αGb� (3.12)
�bβ

which shows the interaction between grain size and dislocation density.
In summary, the material strength can be described as a function of grain size
or dislocation density, where a smaller grain size or larger dislocation density
provides greater strength. As a result, with a decrease in dislocation density due
to local “hot spots” from the electrical current, the yield strength is expected to
decrease. More information can be found in Appendix B, but a few key results are
shown below.
The electrical pre-treating tests (Fig. 3.4) showed variation in their mechanical
response with no difference in the microstructure as compared to the as-received
material. This is in agreement as no recrystallization occurred from electrically
treating the material, but the mechanical response has some variation. Since the
as-received material was annealed from processing, it is expected that the disloca-
tion density was fairly low and there was no cold work imposed on the material.
Thus, the low amount of lattice strain coupled with the local “hot spots” at dislo-
cations did not have a sufficient energy to drive a large quantity of dislocations to
3.3  Comprehensive Electroplastic Theory Explanation 45

Fig. 3.4  Experimental
flow stress results of room
temperature behavior
for preceding stationary
electrical testing

Fig. 3.5  Experimental EAF
flow stress results for single
pulse application of electrical
current during incremental
forming

sinks. As a result, this was seen by the material yield strength not being signifi-
cantly affected by electrical pre-treating; however, some difference was observed
once dislocation motion began.
Additionally, the EAF incremental tests also agree with this effect. As shown
in Figs. 3.5 and 3.6, the application of current had a significant effect on the yield
strength of the material without affecting the microstructure. This directly aligns
with the theory of localized “hot spots” which allow for dislocation annihilation as
a result of enhanced diffusion directly surrounding the dislocation. This is in con-
trast to the electrical pre-treating where there was little effect on the material yield
point. For the EAF incremental tests, there was a much greater driving force for
dislocation annihilation due to the increased amount of lattice strain present. Thus,
a larger amount of dislocations were removed and the stress of the material was
reduced which equates to the reduced yield point.
46 3  The Effect of Electric Current on Metals

Fig. 3.6  Experimental flow
stress results comparing RT
and EAF incremental forming

3.3.2 Electrical Current with Metal Deformation

A similar theory is presented for the application of an electrical current during the
deformation of a metal. As the current is applied during deformation, the local “hot
spots” created from greater electron scattering at defects significantly enhance the
vibrational energy in the surrounding area of the dislocation. This greater energy
surrounding the dislocation allows for enhanced mobility along the slip plane as
it can pass by lattice obstacles with less resistance. Thus, the dislocation has a
greater quantity of energy and can move under a lower required stress (i.e., exter-
nal required force for deformation is reduced). The lower required stress is what
is observed on a macroscale when forming using an applied electric current. Also,
for the other defects within the material (point and interfacial defects), they have
an increased vibrational energy surrounding them as a result of larger amount of
electron scattering. As a result, if dislocations interact or become piled up at these
defects, this additional energy from scattering may allow the dislocation to pass by
the obstacle, where it otherwise would have remained pinned.
Aside from the local “hot spots” at dislocations and defects, the surrounding
defect-free lattice and the overall material temperature is rising. This overall bulk
temperature rise translates to traditional elevated temperature effects on material
deformation (i.e., thermal softening).
The reduction in material strength was observed in the square wave EAF tests
(Appendix B, Figs. B.2–B.5) where the material stress was significantly reduced
during the application of the current. From the theory, it is proposed that the main
effect is a result of localized “hot spots” which significantly increase the mobility
of the dislocations. Additionally, some dislocation annihilation may occur during
the time the current is applied. From the microstructure analysis that is presented
in Chap. 8, it was noted that the EAF tests had a reduced amount of twinning
3.3  Comprehensive Electroplastic Theory Explanation 47

Fig. 3.7  Experimental EAF flow stress results (Parameter Set 8) and effect of specimen cooling
during EAF

(Figs. 8.23–8.25). This may be from the applied current providing excess energy
to dislocation or pinned dislocations such that they could continue the process of
slip. As a result, this reduced the necessity for twin-boundary formation, which
was necessary for the room temperature deformation test to continue. Hence, the
current application supplied energy directly to the regions of high stress or the
areas with very high dislocation densities. Also, for the tests where cooling was
compared to non-cooling (Fig. 3.7), a small difference in flow stress reduction was
noted while the cooled test quickly increased in strength after the application of
the current. This is in agreement with the “hot spots” improving the mobility of
the dislocations, while the remainder of the material was not at such an elevated
temperature. Once the current was discontinued, the electrical energy imparted to
the dislocations was removed and the strength quickly increased as shown experi-
mentally (e.g., Fig. 3.8).
Moreover, this theory is also in agreement with Ross et al. where isothermal
tests were performed at temperature greater than that reached in the electrical tests
[18]. The results from this work showed that the isothermal testing did not create
near the flow stress reductions or the increases in fracture strain as compared to
the electrical tests. The results are given in Fig. 3.9.
This work directly speaks to the aforementioned difference between heating by
external convection and with a direct electrical current, where the convection does
not allow for localized “hot spots” within the lattice. Additionally, early works in
EAF using very short-duration pulses produced large flow stress reductions with
very small bulk temperature increases [19, 20]. Thus, this work also coincides
with the theory in that the short-duration pulse allowed for high local temperatures
at defects, while the bulk of the lattice remained at a reduced temperature.
48 3  The Effect of Electric Current on Metals

Fig. 3.8  Experimental EAF flow stress results (Parameter Set 1)

Fig. 3.9  Isothermal versus EAF testing of Grade 5 Titanium [18]

3.3.3 Electrical Current Effects on Formability

With regard to material formability, ductile fracture is usually transgranular such

that failure occurs through the grains. Ductile fracture begins by the nucleation,
growth, and coalescence of micro-voids. The micro-voids are formed when a high
3.3  Comprehensive Electroplastic Theory Explanation 49

Fig. 3.10  Ductile fracture stages. a Initial Necked Region. b Formation of Micro-Voids. c

Coalescence to Larger Void

stress induces separation of the material at grain boundaries or at small impurity

particles (Fig. 3.10b). As the local stress in the material increases, the micro-voids
grow and coalesce into larger voids (Fig. 3.10c). Over time, crack initiation begins
at the void and the crack grows till the material ultimately fractures.
The high stresses within the material that causes micro-voids to form can be
a result of pinned dislocations within the lattice. As a result of the applied cur-
rent providing energy to the dislocations, the added energy can allow for pinned or
stuck dislocations to continue moving again. Consequently, this reduces the local
stresses within the material’s lattice and delays the process of void formation and
fracture. This theorized ability of the electric current to supply sufficient energy to
allow for pinned dislocations to be mobile can explain the observed effects seen in
experimental testing [6, 21].

3.3.4 Supporting Experimental Results

There has been various experimental works performed, which support the new
electroplastic theory. Specifically, the experimental conclusions will discuss the
heating effects of applied electricity and will address the increased atomic-level
vibrations generated by the increase in temperature at lattice “hot spots.”
Heating and increased atomic-level vibrations
• Threshold versus resistivity relation [22]
• Dislocation density versus temperature relation (stationary–electrical tests) [4]
• Springback elimination in sheet bending [23]
• Dislocation density versus EEC (EAF tests) [4]
50 3  The Effect of Electric Current on Metals

Experimental work supports the theory of localized resistive heating due to the
applied electricity. From this previous work, a threshold current density value has
become apparent which is specific to each metal tested [22]. This threshold value
is the current density value (for a specific metal/deformation speed combination)
whereby significant formability improvements are witnessed, by way of decreased
flow stress or increased achievable elongation as compared to conventional testing.
Upon further investigation, as the resistivity of metal is increased, the threshold
current density, where increased formability effects occur, is decreased (Fig. 2.7).
This supports the theory of resistive heating. The metals with a higher resistivity
have a greater number of lattice obstacles, and there are more sites for the flowing
electrons to scatter off. The increased number of locations of local lattice heating
allows for easier dislocation movement through the lattice.
Another work examined the thermal profiles of stationary–electrical tests
of specimens with different amounts of cold work in them (up to 50 %) [4]. In
these specimens, as the percent cold work was increased, the dislocation density
within the metal was increased. After a constant current was applied to each of
the specimens, the maximum temperature value increased by 85 % as the level
of cold work was increased from 10 to 40 % CW. This can be seen in Fig. 3.11.
Additionally, the voltage potential across the top and bottom dies was larger when
comparing a worked specimen to an annealed specimen of the same dimensions
at different levels of % CW. This indicated that the resistance was higher in the
worked specimen.
An interesting note is that although the dislocation density (number of disloca-
tions) was the only variable that changed with the cold work, and the dislocations
are the only moveable lattice defect, the temperature of the specimens during a sta-
tionary–electrical test still increased significantly with cold work, and the part did
not deform just due to the forces from the electrons. When considering deforma-
tion without electrical assistance, the dislocations receive the motivation to move
due to a stress exerted on them by the external forming force. This shows that the
extra energy imparted into the dislocations from the flowing electrons is lower in

Fig. 3.11  Temperature 400
versus percent cold work Temp. (%CW) = 0.0875(%CW) 2 + 1.275(%CW) + 176.25

relationship for SS304 350

[4]. The temperatures of
stationary–electrical tests 300 Worked
Temperature (ºC)

at the same current density

Annealed
increased as the initial 250 Poly. (Worked)
percent cold work was
increased
200

150

100
0 10 20 30 40 50
Percent Cold Work (%CW)
3.3  Comprehensive Electroplastic Theory Explanation 51

magnitude than the mechanical stresses. Additionally, this energy from the electrons
cannot solely cause dislocation motion or plastic deformation, but rather it acts as
a supplement to the mechanical stresses on the dislocations. When the electricity is
applied by itself without being coupled with deformation, the energy from the elec-
trons to the dislocations will assist in stress relaxation by breaking and reforming
bonds in the immediate area around the dislocations. This effect is similar to a pro-
cess anneal, where heat is added to a material to assist in breaking and reforming of
bonds to lower their energy states. This will result in elastic relaxation, or spring-
back reduction. The work by Green et al. [23] showed that a short, single pulse of
electricity was able to eliminate elastic springback in sheet specimens that were
already formed around a die. The electrons were able to accommodate breaking and
reforming of the bonds to eliminate the tension in the bonds above the neutral axis
and to eliminate the compression in the bonds below the neutral axis.
As the electrons impact the dislocations, they increase the energy (i.e., increase
the temperature) around the dislocations and this can provide the energy needed
to assist the dislocations past the lattice obstacles that are holding them up.
Kravchenko [8], in his explanation of electroplasticity, stated this effect when he
explained that, if there is an electric current flowing, the energy from the electrons
is transferred to the dislocations, thus making the plastic flow easier.
Percent cold work research shows that as the amount of preexisting disloca-
tions within the metal’s lattice increase (by way of cold-working), the efficiency of
the applied electrical power is increased [4]. This is because the larger number of
dislocations within the lattice allows for the applied electrons to impact them and
cause regions of localized heating. The increased temperatures of these regions
also increase the energy of the atoms in these regions, thus allowing for disloca-
tion motion to occur much easier than at lower temperatures.

3.4 Electroplastic Theory Conclusions

The main conclusions drawn from this chapter are as follows:

• The flow of electrical current or the movement of valance electrons within the
material is limited by the electrical resistivity. The resistivity is derived from the
atomic structure, bond type, atomic spacing, and the material defects present in the
lattice. As a result of electron scattering within the lattice, areas of greater vibra-
tional energy exist around defects due the increased amount of electron/ion core
interaction. Additionally, the defect-free lattice has some resistance to electric flow
and the entire material begins to heat. This phenomenon is known as Joule heating.
• Joule heating differs from conventional convection heating of a material (i.e., in an
oven). This is due to the convection heating only providing a uniform increase in
vibrational energy throughout the lattice (i.e., both defect and defect-free regions).
In contrast, Joule heating creates areas of increased vibrational energy at defects
as compared to the defect-free region. Thus, the energy is more directed to the
critical areas (i.e., dislocations and defects) in the lattice for material deformation.
52 3  The Effect of Electric Current on Metals

• Two primary theories for electroplasticity are compared by examining the mag-
nitude of energy they impart on a dislocation core. The first theory is based on
Joule heating and the localized heat generated at the dislocation core. The sec-
ond theory analyzed is from direct momentum transfer on the dislocation core
due to the electron wind effect. To perform the comparison, a current density
of 100 A/mm2 is applied to a pure magnesium metal and the energy transferred
to the core is calculated. The importance of a quantifiable amount of energy
applied to the core is that it can be compared to the activation energy for lattice
diffusion in magnesium. The activation energy to move an ion core is approxi-
mately 1.4 eV/atom, and this equals an activation energy of approximately
98 keV for the entire core. Of note is that the entire dislocation core does not
move all at one time, but regions of the core advance throughout the lattice over
time. Nevertheless, this does not affect the results as they examine the magni-
tude between Joule heating and the electron wind effect. From the analysis, the
Joule heating creating local “hot spots” at defects was shown to provide a sig-
nificant amount of energy (~29 keV) to the core which would have a significant
effect on the dislocations mobility. Also, this amount of energy would signifi-
cantly help to reduce the mechanical stress required to displace the dislocation.
In contrast, the electron wind effect produced a very small amount of additional
energy to the dislocation core (1.29 × 10−3 eV).
As a result, it is expected that the electron wind effect will have little effect in
aiding or increasing the mobility of the dislocation. In conclusion, the contribu-
tion toward the observed electrical effects is due to localized areas of increased
atomic vibration from electron scattering (i.e., Joule heating) and not solely due
to direct momentum transfer on the dislocation cores themselves.
• In the case of stationary electrical current application (i.e., no deformation), the
local areas of increased atomic vibration allow for a rapid decrease in the stored
energy of the material by facilitating dislocation motion and annihilation. The local
“hot spots” provide the driving energy to allow the dislocations to move to a sink
such that the overall lattice energy is reduced. In addition, if the material has been
worked (i.e., additional lattice strain present), this increases the driving force for
the movement of the dislocations. Thus, larger effects on the dislocation density
are expected. This theory was supported by the observed mechanical and micro-
structure effects seen by the electrical pre-treating and incremental EAF tests.
• For electrical current applied during deformation, the local “hot spots” created
from greater electron scattering at defects significantly enhance the vibrational
energy in the surrounding area of the dislocation. This greater energy surround-
ing the dislocation allows for enhanced mobility along the slip plane as it can
pass by lattice obstacles with less resistance. Thus, the dislocation has a greater
quantity of energy and can move under a lower required stress (i.e., external
required force for deformation is reduced). Also, for the other defects within
the material (point and interfacial defects), they have an increased vibrational
energy surrounding them as a result of larger amount of electron scattering. As a
result, if dislocations interact or become piled up at these defects, this additional
energy from scattering may allow the dislocation to pass by the obstacle, where
3.4  Electroplastic Theory Conclusions 53

it otherwise would have remained pinned. This theory is supported by the EAF
square wave and cooling versus non-cooling tests in this work. In addition, prior
EAF tests with large currents coupled with short pulse durations and isothermal
tests are also in agreement with this theory.
• With respect to increased elongation before failure, it is theorized that the
applied current provides a sufficient energy to pinned or stuck dislocation
within the lattice such that it allows for the dislocations to continue moving
again. As a result, this reduces the local stress within the material’s lattice and
delays tvhe process of void formation and fracture.

References

1. Callister WD Jr (2000) Materials science and engineering an introduction, 5th edn. Wiley,
New York
2. Jones JJ, Mears L (2010) Empirical modeling of the stress-strain relationship for an upsetting
process under direct electrical current. Trans North Am Manuf Res Inst SME 38
3. Salandro WA, Bunget C, Mears L (2011) Thermo-mechanical Investigations of the electro-
plastic effect. In: ASME international manufacturing science and engineering conference,
Corvallis, OR, p 10
4. Salandro WA (2012) Thermo-mechanical modeling of the electrically-assisted manufacturing
(EAM) technique during open die forging. PhD dissertation, Clemson University
5. Salandro WA, Jones JJ, McNeal TA, Roth JT, Hong ST, Smith MT (2008) Effect of electrical
pulsing on various heat treatments of 5xxx series aluminum alloys. In: ASME international
manufacturing science and engineering conference, Evanston, IL, 2008, p 10
6. Jones JJ, Roth JT (2009) Effect on the forgeability of Magnesium AZ31B-O when a continu-
ous DC electrical current is applied. In: ASME international manufacturing science and engi-
neering conference, West Lafayette, IN, 2009, p 10
7. Roth JT, Loker I, Mauck D, Warner M, Golovashchenko SF, Krause A (2008) Enhanced
Formability of 5754 Aluminum Sheet Metal Using Electric Pulsing. Trans North Am Manuf
Res Inst SME 36:405–412
8. Kravchenko V (1966) Influence of electrons in delaying dislocation in metals. JETP (USSR) 51
9. Conrad H (2000) Electroplasticity in metals and ceramics. Mater Sci Eng A287:276–287
10. Conrad H (2002) Thermally activated plastic flow of metals and ceramics with an electric
field or current. Mater Sci Eng A322:100–107
11. Askeland DR, Phule PP (2003) The science and engineering of materials, 4th edn. Brooks/
Cole, Australia
12. Suo Z (2003) Reliability of interconnect structures, pp 265–324. In: Gerberich W, Yang W
(eds) Volume 8: Interfacial and nanoscale failure. Milne I, Ritchie RO, Karihaloo B (Editors-
in-Chief) Comprehensive structural integrity. Elsevier, Amsterdam
13. Seth RS, Woods SB (1970) Electrical resistivity and deviations from Matthiessen’s rule in
dilute alloys of Aluminum, Cadmium, Silver, and Magnesium. Phys Rev B 2(8)
14. Kino T, Endo T, Kawata S (1974) Deviations from Matthiessen’s rule of the electrical resis-
tivity of dislocations in Aluminum. J Phys Soc Japan 36(3)
15. Hau-Rieg CS (2004) An introduction to Cu electromigration. Microelectron Reliab 44
16. Hall EO (1951) Proc Phys Soc B64
17. Kocks UF (1966) A statistical theory of flow stress and work hardening. Phil Mag 13
18. Ross CD, Kronenberger TJ, Roth JT (2009) Effect of DC on the formability of Ti-6AL-4V. J
Eng Mater Technol 131(3):11
19. Troitskii OA (1969) Electromechanical effects in metals. Pis’ma Zhurn Experim Teoret Fiz
(10):118
54 3  The Effect of Electric Current on Metals

20. Okazaki K, Kagwa M, Conrad H (1978) A study of the electroplastic effects in metals. Scr
Metall 12:1063–1068
21. Salandro WA, Khalifa A, Roth JT (2009) Tensile formability enhancement of Magnesium
AZ31B-O alloy using electrical pulsing. Trans North Am Manuf Res Inst SME 37
22. Perkins TA, Kronenberger TJ, Roth JT (2007) Metallic forging using electrical flow as an
alternative to warm/hot working. J Manuf Sci Eng 129(1):84–94
23. Green CR, McNeal TA, Roth JT (2009) Springback elimination for Al-6111 alloys using
electrically-assisted manufacturing (EAM). Trans North Am Manuf Res Inst SME 37
Chapter 4
Macroscale Modeling of the
Electroplastic Effect

For successful implementation of EAF in manufacturing industries, one area that

needs to be addressed is the predictability or material response at a bulk level. This
chapter introduces the modeling strategy for the electroplastic effect at the macro-
scale, which are used for compressive modeling (Chap. 5) and tensile modeling (Chap.
6). For macro-level modeling of EAF in Chaps. 5 and 6, the models use a coupled
thermo-mechanical approach based on energy and displacement continuity. Additional
laws utilized in Chaps. 5 and 6 include: the first law of thermodynamics, Joule’s first
law for heat generation, Fourier’s law for conduction, and Newton’s law of cooling.
When modeling the electroplastic effect, the division of the total electrical
energy input can be portioned into energy that directly aids in deformation (elec-
troplastic) and energy that contributes toward bulk heating. One way to capture
this energy distribution is by a ratio of the amount of electricity contributing
toward plastic deformation versus the total amount of electricity applied to the
process. This ratio can be defined as the electroplastic effect coefficient (EEC).
The EEC profile can be determined by utilizing either the mechanical power pro-
files or the thermal profiles of EAF and non-EAF tests. This chapter is divided into
four sections. First, an explanation of the mechanical-based approach for deter-
mining the EEC is described. The EEC is not a constant value, but it is a function
of the time during the forming process. Next, the thermal-based EEC determina-
tion strategy is explained. After, a comparison between the two methods is pro-
vided. Last, empirical modeling strategies for EAF are presented.

4.1 Mechanical-Based Approach to Determining the EEC

The EEC is a ratio that represents the difference between the power required for
a non-electrical baseline test and the power required for each EAF test using a
different current density. This fraction of power is assumed to be converted into

© Springer International Publishing Switzerland 2015 55

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_4
56 4  Macroscale Modeling of the Electroplastic Effect

Fig. 4.1  Mechanical
power profiles (12.7 mm/ 12
min) [1]. The mechanical
power profiles decrease in 10 CD15
magnitude as a higher current
CD10
density is applied 8

Power (Nm/s)
Baseline
6
CD25

4
CD20

0
0 5 10 15 20
Time (s)

mechanical work with respect to time, thus aiding the deformation (i.e., imparting
energy onto the dislocations and facilitating their movement by providing enough
energy to overcome lattice obstacles). The specific steps to solve for the EEC
using the mechanical-based method are as follows. Within the following sections,
these steps will be carried out using EAF compression tests on Al 6061-T6511.
1. Run a conventional and an EAF compression test. Note that these tests need
to be run at the same parameters (i.e., die speed, deformation stroke, initial
preload, and starting specimen size), with the exception of the applied current
for the EAF test. It is also recommended to run several replicates of each test to
verify that the results are repeatable.
2. Plot the mechanical power versus time profiles. In Fig. 4.1, four different start-
ing current densities were tested; therefore, four different sets of EAF tests and
one set of conventional compression tests were needed to construct the profiles.
3. Determine the difference between the two power profiles and normalize it with
respect to the conventional forming power profile.

4.1.1 Experimental Setup and Procedure

The experiments for this section involved running conventional compression tests
and EAF compression tests at several starting current densities. The experimen-
tal data used in determining the EEC are from EAF compression tests run on Al
6061-T6511 specimens, by Perkins et al. [2]. The initial billets had a diameter
of 6.4 mm and a height of 9.5 mm. The compression tests were deformed to a
maximum stroke of 6.4 mm at a speed of 25.4 mm/min. The material parameters
and the friction conditions were determined by solving the model for a classical
test. After fitting the data, the following values were determined: C = 348 MPa,
n = 0.04, and µ = 0.08.
4.1  Mechanical-Based Approach to Determining the EEC 57

4.1.2 Mechanical-Based EEC Determination Procedure

In determining the EEC using this method, the mechanical power profiles of
­conventional compression and EAF compression tests are needed. As mentioned,
the EEC is a ratio between the usable applied electrical power and the over-
all applied electrical power. Further, any portion of the overall applied electrical
power that is not utilized to assist plastic deformation is assumed to contribute
directly toward heating the workpiece. The EEC can be described in equation form
as Eq. (4.1).
(Pconv − PEAF )
ξ= (4.1)
Pconv
where Pconv is the mechanical power profile of a conventional compression test
and PEAF is a mechanical power profile of an EAF test. The difference between
the mechanical power profiles of a conventional and EAF test provides the mag-
nitude of electrical power that contributed toward aiding the plastic deformation.
Figure  4.2 shows the mechanical power profiles for the mentioned compression
tests, where Al6061 specimens (9.5 mm height × 6.4 mm diameter) were com-
pressed at 25.4 mm/min to a final displacement of 6.4 mm. Please note that the
“Baseline” and “EAM” tests in the figure have the exact same test parameters,
with the exception of the “EAM” test having a current density of 20 A/mm2
applied to the specimen during deformation. The Baseline/EAF mechanical power
profiles cannot be compared unless the following test parameters are the same:
stroke, die speed, starting specimen size, and preload.
The next step is to determine the actual EEC profile. Although the EEC is
complex, an initial method to simplify the relation was to use an average value
approximation. To determine an average EEC, a coefficient profile was cre-
ated by plotting the coefficient versus time for each respective current density, as
shown in Fig. 4.3. Once the coefficient profile was established, an average coef-
ficient for each current density was determined using a flat-line approximation.
This type of approximation is sufficient for initial model verification and is fairly
accurate except with the 60.8 A/mm2 test. To better characterize the EEC profile,

Fig. 4.2  Determination 16
Baseline
of the electroplastic effect 14
coefficient (EEC) [3]. In this 12
figure, the differences in the EAM
Power (Nm/s)

10
mechanical power profiles are
used to determine the ratio of 8
“useable” applied electricity, 6
quantified by the EEC 4
2
0
0 5 10 15 20
Time (s)
58 4  Macroscale Modeling of the Electroplastic Effect

0.6
27.7A/mm^2
Electroplastic Effect Coefficient (EEC) 35.8A/mm^2
0.5 54.6A/mm^2
60.8A/mm^2
0.4

0.3

0.2

0.1

0
0 2 4 6 8 10 12 14 16 18
Time (s)

Fig. 4.3  Electroplastic effect coefficient (EEC) profiles [3]. The EEC profiles are shown, along
with a straight-line approximation

a time-based function is more suitable. To create an EEC profile as a function of

time, the same EEC profiles in Fig. 4.3 can be used, where a power law trend line
can be fitted to each profile as a function of the testing time.

4.1.3 Mechanical-Based EEC Conclusions

The mechanical approach examines power required for conventional tests and
EAF tests to produce the EEC. The EEC is time based and can be approximated
with a linear or power law function. Also of note, previous work has highlighted
a “threshold effect” whereby electrical application benefits are negligible until
a critical current density is reached [2]. Figure 4.3 supports this threshold effect
since the 60.8 A/mm2 test has an EEC profile that is much higher than the other
three current densities. However, this does not mean that the electrical threshold is
60.8 A/mm2, but this was just one of the current densities tested. In this case, the
threshold is somewhere between 54.6 and 60.8 A/mm2.

4.2 Thermal-Based Approach to Determining the EEC

The objective of this section is to establish a methodology for quantifying the

EEC using thermal profiles. From the conservation of energy and empirical obser-
vations, an analytical model able to predict the temperature rise in the specimen
due to the electricity that was applied while subjected to deformation must be
4.2  Thermal-Based Approach to Determining the EEC 59

completed before being able to determine the EEC. The EEC is dependent on the
magnitude of current applied to the workpiece. This can be seen from the differ-
ences in the EEC profiles at different current densities in Fig. 4.3.
To determine the EEC, the variation of the temperature of the specimen in time
is solved from Eqs. (4.3) and (4.5) and combined with experimental thermal meas-
urements and with geometrical changes occurring during the deformation process.
The procedure is described below:
1. Solve the model for a stationary electrical test. This is a test when the part is
clamped between the dies and preloaded, and electricity is applied. The dies are
stationary, thus no plastic deformation occurs, and the dimensions of the part
are not changed.
2. Validate the model by comparing with thermal measurements from experiments
(i.e., strictly resistive heating).
3. Solve Eqs. (4.3) and (4.5) for an electrically assisted deformation test. Note that
just the temperature rise component due to the application of electricity is con-
sidered. Initially assume the EEC to be ξ = 0.
4. Compare the solution from the previous step with the thermal measurements.
At this step, the increase in temperature due to electricity is separated from
the increase due to plastic deformation by subtracting the temperature increase
recorded during conventional compression experiments. As will be shown in
the results section, there will be a significant thermal profile difference between
the experiment and the model. This difference proves the concept of the EEC.
5. Determine the EEC, ξ, in order to match the model with the experimental ther-
mal profile and propose a formulation for this coefficient for the different mate-
rials studied.

4.2.1 Building a Thermal Model

The EEC depends on specific material properties, such as heat capacity, electric
resistivity, electric field characteristics, and the current magnitude. The simplifying
assumptions that were used to facilitate the derivation of the model are as follows:
• The material is homogeneous and isotropic.
• The increase in temperature due to friction at the workpiece/die interface is
negligible. This assumption is valid for the small dimensions and low relative
speeds in the compression tests.
• The barreling effect due to friction is neglected when the model is solved.
• In the initial calculations, the material properties that are assumed constant with
temperature are Cp and r.
• The temperature distribution in the part is considered uniform. In reality, contact
resistance under low loading causes an increase in the temperature at the con-
tact points between the specimen and the dies at the beginning of application,
followed rapidly by homogeneous heat generation as the load increases and
60 4  Macroscale Modeling of the Electroplastic Effect

Fig. 4.4  Heating and cooling sequence of a stationary electrical experimental test. Throughout

the beginning of the test, the specimen absorbs the majority of the heat, but during the second
half, a lot of heat is conducted into the dies

contact resistance effect becomes negligible. The die/specimen interfaces will

also have a different temperature distribution, and these boundary conditions
will be taken into consideration when the solution for the model is calculated.
• The voltage in the specimen is assumed constant, and it is determined for the
initial dimensions of the specimen at room temperature. However, the voltage
will vary during the test, since the power supply used is a variable-voltage
source (i.e., constant current). On one hand, the decrease of the ratio of height
over cross-sectional area results in lower resistance, thus lower voltage, while
the increase in temperature will increase the resistivity, thus resistance and
voltage increase. Overall, the present work assumes that these effects cancel
each other.
The temperature rise and its distribution during an electrical test are illustrated
in Fig. 4.4 during a stationary electrical test, where electricity is applied without
plastic deformation. It can be seen that the specimen quickly heats up for about the
first 14 s with almost all the heat going exclusively into the workpiece. Then, the
heat from the specimen conducts into the compression dies for the remaining half
of the stationary electrical test.
When examining only the specimen, the specific temperature profile will also
change throughout the test. More specifically, the temperature of the die/workpiece
contact points at the beginning of the test will be much hotter than the rest of the
specimen (because the actual contact area will be less, leading to a higher overall
electrical power at these points). However, as the part is compressed and the asperi-
ties are crushed, the temperature profile inverts and the middle of the specimen is
the hottest (since the asperities crush and the contact area increases, leading to a
decrease in the electrical power at these points). In addition, the heat generation is
reduced to a point that it cannot keep up with conduction into the dies. The esti-
mated thermal profiles of an EAF test are schematically illustrated in Fig. 4.5.
For a better understanding of the electroplastic effect, the thermal aspects must
be isolated. When the electric current passes through the metallic workpiece, heat
is generated and the temperature of the part rises. The temperature increase is
determined from the energy balance, illustrated in Fig. 4.6.
4.2  Thermal-Based Approach to Determining the EEC 61

Fig. 4.5  Specimen thermal profiles at the beginning and end of an electrically assisted compres-
sion test [4]. The specimen is hottest at the die/workpiece interfaces at the beginning of the EAF
test, while the center of the specimen is hottest at the end of the EAF test. a Temperature profile
at the beginning of an electrically assisted compression test. b Temperature profile toward the
end of an electrically assisted compression test

Mech.
Force

Input from
electric energy

Die Conduction

Radiation Workpiece Convection

(Air)

Die Conduction

Mech.
Force

Fig. 4.6  Energy balance of the heat flux. The thermal model accounts for all three heat transfer
modes (conduction, convection, and radiation)

The general form of the energy conservation is expressed on a rate basis as Eq. (4.2):

Ėpart = Ėg − Ėout ,

(4.2)
Q̇ = Q̇e − Q̇cond − Q̇conv − Q̇rad ,
62 4  Macroscale Modeling of the Electroplastic Effect

where Ėpart is the rate of change of the energy content in the part analyzed (Q̇), Ėg
is the rate of heat generation due to the electric resistive heating (Q̇e), and Ėout is
the rate of energy leaving the part due to conduction into the dies in the contact
zone (Q̇cond), convection (Q̇conv) and radiation from the surface (Q̇rad). To clarify, Ėg
has only one contributor (Q̇e), whereas Ėout has three contributors (Q̇cond, Q̇conv, and
Q̇rad) that are the three forms of heat transfer. After developing each component, the
heat equation can be written in one dimension as Eq. (4.3):
∂T ∂T  
ρVv Cp = −As [h(T − T∞ )] − 2kAc − As εσSB T 4 − Tsurr
4
+ η(1 − ξ )VI
∂t ∂x
(4.3)
where ρ is the density of the material, Vv is the volume of the part, Cp is the
specific heat of the material, T is the temperature, t is time, As is the lateral sur-
face area of the part, h is the convection heat transfer coefficient, T∞ = Tsurr is
the surrounding temperature, k is thermal conductivity for the die material, Ac is
the cross-sectional area, x is coordinate, ε is radiative emissivity for the part, and
σSB is the Stefan–Boltzmann constant. The input from electricity was taken from
Eq. (4.4).
Pe = Pheat + Pdef = η(1 − ξ )VI + ηξ VI (4.4)
where Pe is the electric power (I·V), η is the efficiency, and ξ is the EEC. Pheat
represents the amount of electric power that will dissipate into heat through resis-
tive heating of the workpiece. Pdef is the electrical power component that will aid
the plastic deformation.
The changes in geometry are calculated assuming compression at constant
speed, as well as volume constancy, as given by Eq. (4.5):

∂h ∂D Vv u̇
= −u̇; = √ (4.5)
∂t ∂t π h h

where D and h are instantaneous dimensions of the workpiece, and u̇ is the com-
pression speed. The second
 equation was determined from the volume constancy
condition (Vv = πD2 h 4).
The change in temperature solely due to the portion of electrical power that did
not contribute toward plastic deformation can be shown in the integral in Eq. (4.6).

tf
ηVI
Tf − T o = (1 − ξ )dt (4.6)
mCp
0

where m is the mass of the workpiece. The EEC profile with respect to time can be
shown in Eq. (4.7).
ξ = ξ0 ∗ t b (4.7)
4.2  Thermal-Based Approach to Determining the EEC 63

where ξ0 is an initial value that is dependent on specific material properties and

the magnitude of the applied current, t represents time, and b is an exponential
term. The ξ0 value is similar to the C-term in the power law equation, where it
is dependent on the specific material (i.e., its strength, crystal structure, etc.).
Although the EEC does not simply vary in time, the term “time” is related to the
deformation parameters. Equation (4.7) can be rewritten in terms of strain, as
shown in Eq. (4.8), where ε is the material strain and ε̇ is the strain rate (i.e., the
deformation speed).
 ε b
ξ = ξ0 × (4.8)
ε̇

4.2.2 Experimental Setup and Procedure

The main experimental objective of this section was to measure the thermal pro-
file during the EAF test and use it to generate an EEC profile. Figure 4.7 displays
the experimental setup used for testing. An Instron Model 1332 hydraulic testing
machine was used to compress the specimens. Machined, hardened, and insulated
dies, made from A2 tool steel, were installed in the Instron machine (note that
insulation was used such to isolate the electricity from the test machine). For ther-
mal measurements, a FLIR A40 M thermal camera, with a temperature capacity
of 550 °C, resolution of 0.1 °C, and sample frequency of 50 Hz was utilized. All
force and position data were gathered using an onboard data acquisition system at
1,500 Hz sampling frequency.
The materials tested were two grades of titanium: Grade 2, which is a single
phase polycrystalline material and Grade 5 (Ti-6Al-4V), a dual-phase alpha–beta
alloy. The initial dimensions of the specimens are summarized in Table 4.1. A dis-
placement speed of 12.7 mm/min was used for all tests and a constant DC current
of 300 A was utilized for all the EAF tests. Please note that, since the cross-sec-
tional areas between the specimens of the three levels of deformation will be dif-
ferent, the resulting current densities will also be different. Considering the size

Fig. 4.7  EAF test setup.

The setup consists of the Upper Grip
mechanical testing machine,
dies, and a thermal camera
Electric Lead

Workpiece

Forging Die

FLIR Camera

Lower Grip
64 4  Macroscale Modeling of the Electroplastic Effect

Table 4.1  Specimen dimensions [5]

Def. level Diameter (mm) Height (mm)
L0 3.81 5.72
L1 4.32 4.45
L2 5.11 3.18

of the “as-is” specimens, the current density supplied to the workpieces is about
26 A/mm2, which is above the threshold current density of about 20 A/mm2 deter-
mined by Perkins et al. [2].
Two types of tests were performed in this work. First, deformation tests were
run, where the specimens were deformed at a constant die speed from an ini-
tial height of L0 to deformation levels of L1 or L2, with and without electricity
applied (note: electricity applied at 300 A). Stress–strain and thermal profiles
were created from these tests. Second, stationary electrical tests were per-
formed, where electricity was applied to a specimen without plastic deforma-
tion. In particular, the dies were in contact with the specimen such to prevent
hot spots at the contact interfaces and to allow electricity to flow, but without
causing plastic deformation in the sample. Thermal profiles were gathered from
these tests and were used in comparison with each other and with the thermal
profiles from the deformation tests, as will be shown in the following sections.
The different test combinations and corresponding specifications can be seen in
Table 4.2.

Table 4.2  Test combinations and specifications [5]

Grade Deformed/Stationary EAF/Conv. Displ. Prev. structure %CW
2 Def. EAF L0–L2 As-Is –
2 Def. Conv. L0–L2 As-Is –
5 Def. EAF L0–L2 As-Is –
5 Def. Conv. L0–L2 As-Is –
2 Def. EAF L1–L2 Init. Def. to L1 22
2 Def. Conv. L1–L2 Init. Def. to L1 22
2 Def. EAF L1–L2 Init. mach. to L1 –
2 Def. Conv. L1–L2 Init. mach. to L1 –
2 Stationary EAF L1 Init. Def. to L1 22
2 Stationary EAF L1 Init. mach. to L1 –
2 Stationary EAF L2 Init. Def. to L2 44
2 Stationary EAF L2 Init. mach. to L2 –
5 Stationary EAF L1 Init. Def. to L1 22
5 Stationary EAF L1 Init. mach. to L1 –
5 Stationary EAF L2 Init. Def. to L2 44
5 Stationary EAF L2 Init. mach. to L2 –
4.2  Thermal-Based Approach to Determining the EEC 65

4.2.3 EEC Thermal-Based Determination

Within this section, true stress–strain plots and temperature plots are used to
emphasize several EAF relations and effects. First, the reductions in flow stress
due to EAF are shown. Of note is that since this work focused on modeling EAF,
only one current was used, so the reductions in flow stress from EAF could be
much greater if larger currents were utilized. Second, the model prediction pro-
cess and accompanying experimental results are discussed. The steps required
for model prediction are outlined in the previous section. From the creation of
the model and analysis of several of the main input variables, EAF can be bet-
ter understood and more accurately predicted through this variety of influencing
factors.

4.2.3.1 Force Reduction Due to EAF

From many experimental works over the last several years, it has been shown
that EAF reduces a metal’s flow stress compared to conventional forming [6].
Figures 4.8 and 4.9 show the true stress–strain profiles of a conventional compres-
sion test (at room temperature) and an EAF test for Ti-G2 and Ti-G5, respectively.
A die speed of 12.7 mm/min was used and a constant current of 300 A was applied
during the EAF test [4]. Both figures show that the flow stress was reduced sig-
nificantly due to EAF. In Fig. 4.8, one can see that the difference in the flow stress
is offset by about 150–200 MPa throughout the test. Conversely, in Fig. 4.9, the
difference in flow stress increased slightly as the test progressed. Additionally,
the Ti-G5 in the conventional compression test failed before the end of the test.
However, in the electrically assisted compression test, the material was able to be
completely formed without failure. The results in Fig. 4.9 are comparable to EAF
compression test results of the same alloy run by [2]. In particular, a small amount
of strain weakening is apparent at the beginning of the test, and in both works, the
elongation of Ti-G5 is increased using EAF.

Fig. 4.8  Flow stress 1000

reduction due to EAF Conventional Test
Conventional Test
(Ti-G2) [4]. The flow 800
True Stress, [MPa]

stress was notably reduced

in comparison with the 600
conventional stress–strain
profile EAF Test
EAF Test
400

200

0
0 0.1 0.2 0.3 0.4 0.5 0.6
True Strain
66 4  Macroscale Modeling of the Electroplastic Effect

Fig. 4.9  Flow stress 1500

reduction due to EAF (Ti-
G5) [4]. The flow stress 1200

True Stress, [MPa]

was reduced to EAF, and Conventional Test
the specimen was able to
900
be completely formed to
the desired length without EAF Test
failure, unlike with the 600
conventionally formed test
300

0
0 0.1 0.2 0.3 0.4 0.5 0.6
True Strain

Although the displacement varies linearly in time due to a constant die speed, the
force for the EAF test showed some variation (“noise”). The frequency analysis of
the data indicated a frequency response at ~22.29 Hz and its multiples for G2, and
1.02 Hz and its multiples for G5, which are not found in the conventional tests. The
behavior observed may be due to a cyclic softening/hardening phenomenon present
during EAF. It is hypothesized that the input of electric energy may result in alternat-
ing of electroplastic softening of the material (thus, reduction in the forming load)
and hardening of the material due to deformation (thus, increase in load).

4.2.3.2 Model Prediction and Experiments

The first step in the model prediction process is to compare the model to a simple
stationary electrical test, to isolate the resistive heating effects and validate that the
heat transfer relationships within the model are correct.
Figure 4.10 displays both an experimental test and the model output for a stationary
electrical test on Ti-G2. The model tends to slightly underestimate the temperatures in

Fig. 4.10  Ti-G2 Stationary
electrical test at L0. The
stationary electrical tests
were run by applying
Experimental Output
electricity without deforming
(no def.)
the workpiece. The model
assumes 100 % of the
Model Output
electrical power contributes (no def.)
toward heating the workpiece
4.2  Thermal-Based Approach to Determining the EEC 67

600

500 CD 25

400 CD 20
Temperature, [C]

300
CD 15
200

CD 10
100

0
0 2 4 6 8 10
Time, [s]

Fig. 4.11  Stationary electrical modeling tests at different starting current densities for SS304.
The resistive heating thermal model was tested on various materials with different starting cur-
rent densities prior to integrating in the effect of deformation

the middle of the test, but is accurate at the beginning and end of the test. Additionally,
the heating profile of the model is slower than that of the experiments.
Figure 4.10 shows the modeled thermal profile of a stationary electrical test for
only one specific starting current density. Please note that prior to incorporating
plastic deformation into the model, the resistive heating-only model was verified
at different starting current densities. Figure 4.11 shows the model results of four
different starting current densities for SS304 (note that the modeling uses titanium,
however, the model was validated on several different materials, and this figure
displays its results on Stainless Steel 304). Now that the model can effectively
predict resistive heating effects, the effect of plastic deformation must now be
included to represent an actual EAF test. The temperatures recorded are presented
in Fig. 4.12 for Ti-G2. The EAF deformation test is compared with the stationary
tests for intermediate stages of deformation, i.e., 6 s for L1 and 12 s for L2 (the
test durations are list in Table 4.2). The temperatures generated during stationary
electrical tests due to resistive heating are greater than the EAF deformation tests.
The difference cannot be explained simply by the very small difference in the
cross-sectional area (the machined specimens do not have barreling), but by the
difference in the dislocation density. Moreover, during the EAF deformation test, a
significant portion of the electric power goes toward assisting plastic deformation,
rather than direct resistive heating. Knowing this, the EEC can be determined from
the difference in thermal profiles of the model and experimental EAF tests, after
deducting the temperature profile for the conventional forming test.
After experimenting with different EEC relations, the model and experimental
EAF thermal profiles were matched using a power law function (Eq. 4.7) to repre-
sent the EEC, as shown in Fig. 4.13.
68 4  Macroscale Modeling of the Electroplastic Effect

Fig. 4.12  Ti-G2 temperature
profiles. The temperature Stationary,
profiles for a conventional L 0 – 300A
compression, EAF
compression, and stationary Stationary,
electrical test are shown, with L1 – 300A
Stationary,
single points representing EAF – 300A
L 2 – 300A
the maximum temperatures
of the stationary electrical
Conventional
experiments at L1 and L2
compression

Fig. 4.13  Ti-G2 model
with EEC (L0–L2). The
EAF Model
EEC power law function
was optimized such that it
agreed with the experimental
temperature profile EAF Test
300A

The same temperature model prediction process procedure was followed for
Ti-G5. The model and experimental results for the stationary electrical test are
shown in Fig. 4.14. For this test, a specimen machined to L2 dimensions was used
in the experiments. The model and experiments matched closer than for Ti-G2.
As was done for Ti-G2, an EEC function was developed to match the model
and experimental EAF thermal profiles, shown in Fig. 4.15. For this material, the
model produced roughly the same peak temperature as the experiments, but the
model showed a slightly slower heating profile, as was also seen for Ti-G2.
Figure 4.16 displays the EEC profiles for an EAF test from L0–L2 (0–12 s). Of
note is that the Ti-G5 profile ends at 6 s, while the Ti-G2 profile extends to 12 s.
This is because the Ti-G5 specimens failed after L1 and were not able to be con-
ventionally formed to L2 without fracture. The coefficients were found to be non-
linear and were approximated using the power law.
From Fig. 4.16, the EEC for each Ti-grade is different, due to differences in the
electrical, thermal, and microstructure properties of the two materials. In particu-
lar, the Ti-G2 EEC increased rapidly in the first few seconds of deformation, hence
4.2  Thermal-Based Approach to Determining the EEC 69

Fig. 4.14  Ti-G5 stationary
electrical test at L2 [4, Max. FLIR camera temperature
5]. The thermal model
(assuming 100 % heating)
was compared to a stationary Model Output
electrical experiment (no def.)

Experimental Output
(no def.)

Fig. 4.15  Ti-G5 EAF test

and model with EEC (L0–L1).
The EEC power law relation
EAF Test
was used in the model
300A
EAF Model

Fig. 4.16  Ti-G2 and Ti-G5

EEC profiles (L0–L2). The G2 EEC (L0-L2)
profiles of the two grades of
titanium are different
Fracture

G5 EEC (L 0-L1)
70 4  Macroscale Modeling of the Electroplastic Effect

Fig. 4.17  Percent error
between model predictions
and experiments [5]. The
magnitude of the percent
error varied between the
material type and the test type

the larger ξ0 value, whereas the Ti-G5 EEC increased more consistently through-
out the length of the test, hence the larger b value.
The difference between model predictions and experiments was computed, and
the maximum and minimum differences as compared to the experimental meas-
urements of the temperature are presented in Fig. 4.17. For most of the tests, the
model predicted the temperatures with less than 12 % error, while one of the tests
underestimated some of the values with about −20 %. Overall, the model predic-
tions agree well with the measurements, but refinement will be needed for greater
accuracy.

4.2.4 Thermal-Based EEC Conclusions

This section analyzed the thermo-mechanical aspects of the electroplastic effect. A

thermal model, accounting for material properties specifically for titanium alloys
various levels of plastic deformation, was verified using experimental EAF tests
[4]. The conclusions from this section are as follows:
• EAF reduced the flow stress in both Ti-G2 and Ti-G5. Although just one current
was used, 300 A DC, based on previous work, similar results are expected for
different currents.
• To account for varying differences in thermal profiles between the conventional/
EAF tests, a new EEC was introduced, which was defined by a power law
relation.
• The EECs for the two Ti-grades were significantly different. This could be
because the grades of titanium were different. Ti-G2 is a single phase poly-
crystalline material, and Ti-G5 is a dual-phase alpha–beta alloy. For Ti-G2, the
electrons were able to travel easier because they only had to move through one
phase. This can be seen by noticing the steeper and higher EEC for Ti-G2 com-
pared to Ti-G5.
4.3  Comparison Between the Different EEC Determination Approaches 71

4.3 Comparison Between the Different EEC

Determination Approaches

Since the EEC determination methods are essentially the inverse of the other (i.e.,
one is based off the difference in mechanical power and the other is based off the
difference in temperature), the resulting EEC profiles should be equivalent. The
EEC profiles are calculated using both methods and then compared to each other
in the next section.

4.3.1 SS304 Electroplastic Effect Coefficient Profiles

In this section, EEC profiles derived from the previously described mechanical and
thermal methods will be presented, and the effects of different die speeds will also
be examined. Table 4.3 shows the different tests required to generate the EEC pro-
files using both of the methods. In particular, there was a batch of conventional com-
pression, and an EAF compression tests for each respective current density and die
speed combination. In addition, there was a stationary electrical test batch run for
each respective current density. All tests are described in the table. Figures 4.18 and
4.19 display mechanical and thermal EEC profiles for tests run at 12.7 and 25.4 mm/
min with starting current densities of 15 and 25 A/mm2. Because two different die
speeds are analyzed in each figure (and the tests are different time lengths), the
x-axis was chosen to represent the percentage of the compressive stroke completed
rather than time, to allow for ease of comparison between all four tests.
In Fig. 4.18, a starting current density of 15 A/mm2 is explored. Please note
that the x-axis on this figure and the following figure is in terms of “% Stroke
Completed.” This is due to the fact that there are two different deformation speeds,
and by using this unit for the x-axis, the EEC profiles for both speeds can be com-
pared. From the figure, both methods of determining the EEC are extremely accu-
rate with respect to each other, for both die speeds. The EEC profile for the slower

Table 4.3  Deformed and stationary electrical tests for determining EEC profiles for SS304 [1, 7]
Deformed/Stationary Conv./EAF Current density (A/mm2) Die speed (mm/
min)
Deformed Conv. – 12.7
Deformed EAF 15 12.7
Deformed EAF 25 12.7
Deformed Conv. – 25.4
Deformed EAF 15 25.4
Deformed EAF 25 25.4
Stationary EAF 15 –
Stationary EAF 25 –
72 4  Macroscale Modeling of the Electroplastic Effect

Fig. 4.18  EEC profiles 0.4

(CD15, 12.7 and 25.4 mm/ 0.35
min) [1]. The mechanical
0.3
and thermal EEC profiles are
Mech. (12.7)
similar for each respective 0.25
Therm. (12.7)

EEC
speed. Additionally, the 0.2
slower speed produced a
0.15
higher EEC value at this
particular current density 0.1
Mech. (25.4)
0.05
Therm. (25.4)
0
0% 25% 50% 75% 100%
% Stroke Completed

Fig. 4.19  EEC profiles 0.3

(CD25, 12.7 and 25.4 mm/
min) [1]. The faster 0.25
deformation speed produced
0.2
a higher EEC value when Mech. (25.4)
coupled with this particular
EEC

0.15 Therm. (25.4)

current density
0.1

0.05 Mech. (12.7)

Therm. (12.7)
0
0% 25% 50% 75% 100%
% Stroke Completed

die speed (12.7 mm/min) is higher than the faster die speed profile after about
25 % of the desired compressive stroke is completed. This can be expected, since
previous works have proven the EAF technique to be strain rate dependent [8].
The EECs for a starting current density of 25 A/mm2 are shown in Fig. 4.19. As
was the case in Fig. 4.18, both EEC calculation methods are consistent throughout
the entire compressive stroke for both the die speeds. For the case of current den-
sity of 25 A/mm2, the EEC for the die speed of 25.4 mm/min was on average 12 %
higher than for the case of 12.7 mm/min, with a maximum difference of 17 % at
the end of stroke. Please understand that the EEC is not solely based on the reduc-
tion in mechanical power, but that it is determined from the power reduction with
respect to the magnitude of electrical power input required to create that reduction.

4.4 EEC Profile Conclusions

The previous sections analyzed two methodologies for quantifying the EEC. An
experimental upper bound mechanical approach and a thermal analytical approach
were utilized to determine the EEC profiles. The conclusions from the previous
sections are as follows:
4.4  EEC Profile Conclusions 73

• EAF reduced the flow stress in the 304 Stainless Steel specimens for both the
die speeds.
• The EEC profiles determined using both methods were consistent to each other
throughout the entire stroke range for both die speeds.
• The mechanical-based approach to determining the EEC only required force versus
position data to generate the EECs, whereas the thermal-based approach required
there to be thermal data taken, which required the use of a thermal camera.
• The mechanical-based approach does not solely isolate the effects of thermal sof-
tening, since there will be some extent of thermal softening contributing to the
mechanical power reduction of the EAF tests due to the increased temperature.

4.5 Empirical Modeling Strategies

The following two sections present empirical modeling for compression EAF
forming. The first section contains a model that is only based off the nominal cur-
rent density where the second section takes into account the non-steady state cur-
rent density during forming.

4.5.1 Non-constant Current Density

For the development of the electrical current and flow stress model, previously
performed continuously applied stress versus strain data was analyzed. The fol-
lowing sections discuss the methodology used in creating the stress predictor func-
tion and the observed comparison between empirical results and actual testing.

4.5.1.1 Relationship Development

When examining the relationship between the flow stress and applied strain with
the addition of current, a range of the strain equivalent to the baseline fracture
strain was examined. This is shown in Fig. 4.20 for the different initial current
densities.
The next step was to determine the difference or ratio between the baseline and
each of the electrical tests. To perform this, a ratio was determined, which would
relate the electrical test back to the baseline. The stress ratio with respect to strain
for each nominal current density is shown in Fig. 4.21. After seeing the relation-
ship for each current density, an exponential fit (dashed line) was utilized as the
transfer function between the electrical test and the baseline. This type of equation
was utilized as it best represented the observed ratio functions. To better character-
ize the relationship, the different current density’s stress ratio converged to one at
an approximate strain value of 0.018, thus an additional offset parameter in the
function was included. This is mainly a result of the elastic limit of the material
74 4  Macroscale Modeling of the Electroplastic Effect

Fig. 4.20  Flow stress of EAF magnesium over the Baseline Strain. The flow stress is reduced
with increasing current density

Fig. 4.21  Stress ratio as a function of strain for various current densities. The exponential fit
dashed lines represent the transfer function between each electrical test and the baseline

and fixture compliance from the machine. Likewise, the exponential decay val-
ues (0.175, 1.605, 7.431, and 17.528) from Fig. 4.21 were plotted against their
respective nominal current density to determine a relationship between the indi-
vidual current density tests. The results from this are shown in Fig. 4.22 and a
power function trend line (dashed line) was fitted to this observed relation. Using
the two values (0.000075 and 3.36), these now provide a given characterization of
4.5  Empirical Modeling Strategies 75

Fig. 4.22  Power relationship
for exponential values as a
function of current density

the material. Combining the developed relationship with the traditional power law,
it is postulated to characterize the flow stress of the electrically deformed material
as a result of its initial nominal current density of over the range of strain tested.

4.5.1.2 Empirical Relationship

Using the methodology from the Relationship Development section, the resulting
model for the upsetting flow stress is predicted using:
 
σpredicted = K ′ εn ε̇m exp −(ε − c)B�A (4.9)

where, σpredicted is the predicted flow stress with the application of electricity, K′ is
the strength coefficient, ε is the strain, n is the strain hardening exponent, ε̇ is the
strain rate, m is the strain rate sensitivity exponent, Φ is the current density, and A,
B, and C are material specific constants for this model. Also, when considering the
current density, since the current varies as a function of strain, the current density
in this model was considered to be decreasing. Neglecting barreling, the instanta-
neous current density during the test can be described by:
Ih
= (4.10)
Ao ho
where Φ is the current density, I is the current, Ao is the initial area of the specimen,
ho is the initial height of the specimen, and h is the instantaneous height of the speci-
men during the testing. Thus, as the test progresses, the current density decreases
as the specimen decreases in height. One important fact to note is that the current
(I) during the test is assumed to be constant. However, since the current provider
for this setup supplies a constant voltage, the current is dependent on the resistance
of the system, simply using Ohm’s law. As a result, the resistance of the system is
dependent on the temperature, geometrical changes, and microstructure changes of
the specimen. As the temperature is increased, this causes the resistance to increase
76 4  Macroscale Modeling of the Electroplastic Effect

Fig. 4.23  Comparison of empirical model for magnesium over small strain values. The model
predicts the stress–strain response for all current densities

due to the increased level of atomic vibration. Also, as the specimen is compressed,
the area increases and the length decreases, thus reducing the resistance. Last, as
the specimen is deformed, the microstructure (number of vacant lattice sites, inter-
stitials present, the number of dislocations, and the number of stacking faults) may
increase or decrease the resistance of the material. Hence, the overall effect is not
fully mapped in this research; thus the current is assumed to be constant.
Using this model, the material behavior of this magnesium alloy was then plot-
ted for different current densities. The results are depicted in Fig. 4.23. As shown,
the empirical model predicts the relationship of the flow stress when the process
is subject to a continuous current. There are some more noticeable differences in
the empirical model and the actual test results at higher current densities; how-
ever, it still provides a fair approximation when considering the range of current
density covered. The final material specific constants that yielded the minimum
least squares error are: A = 3.38, B = 0.00009, and C = 0.018. These values differ
slightly from the derived values in the Relationship Development section, as the
original values were used along with the least squares method to reduce any error
present in the curve fitting (Fig. 4.21/Fig.  4.22). The average error for all of the
curves is 5.82 MPa, with the largest difference being 27.48 MPa.

4.5.1.3 Model Testing

To further the ability of this predictor to map the flow stress during an EAF com-
pressive process, the empirically based model was used to forecast the flow stress
over a larger strain region of Copper C11000. Once again, the modeled flow stress
will be compared to the experimental results of the testing.
4.5  Empirical Modeling Strategies 77

Fig. 4.24  Comparison of empirical model for copper over large strain values. The model is able
to predict the general true stress–strain profiles, with less precision at lower strains

The empirical model was tested and compared to previous experimental tests
of Copper C11000 under an EAF upsetting process [2]. Shown in Fig. 4.24 are
the results of the experimental testing over the deformation region of the test in
addition to the empirical prediction. As can be observed, the empirical model is
capable of providing a close approximation of the flow stress throughout the dura-
tion of the test for a given current density. Also, in comparison with the magne-
sium results presented in Fig. 4.23, the empirical model is utilized over a larger
strain area, thus proving the estimation ability of the stress predictor function.
For this material, the material constants were once again fit to the data using the
least squares regression to minimize the error. The values for A, B, and C are 2.23,
0.00015, and 0.0, respectively. Overall, the average error present for this material
is 6.81 MPa, with a maximum error of 25.32 MPa, occurring during the initial
region of the 60.9 A/mm2 test.

4.5.1.4 Discussion

When considering the empirical model, this model proved a reasonable estimate
of the flow stress during the deformation region for an EAF upsetting test. Yet,
one shortcoming of this model is that it was created based off the nominal current
densities (Figs. 4.21 and 4.22). To fully map the flow stress, the mathematical rela-
tionship should incorporate the instantaneous current density as function of true
strain as well. This 3D mapping surface is shown in Fig. 4.25, where the stress
ratio is a function of the instantaneous current density and true strain. Mapping
this surface will provide a more closely modeled mathematical descriptor for the
effect of electrical current during an upsetting test. Another factor limiting this
78 4  Macroscale Modeling of the Electroplastic Effect

Fig. 4.25  Magnesium surface of stress ratio as a function of instantaneous current and true strain

model is that it does not fully encompass the notion that the addition of electricity
lowers the yield point of the material. This is most prevalently seen in the Copper
C11000, where error is a maximum directly after yield. This error is more pro-
nounced in the Copper C11000 as compared to the magnesium, as the applica-
tion of electricity significantly decreased the yield point of the copper compared to
magnesium. Another aspect that should be considered to improve this model is the
field density throughout the specimen, thus possibly using barreling compensation
and incorporating frictional affects (especially for larger strains). An added factor
that may also play a crucial role in modeling the flow stress is the occurrence of
an electrical threshold [2]. This threshold causes a non-uniform stress distribution
between current densities for a given strain.

4.5.1.5 Conclusion for Non-constant Current

Density Empirical Modeling

This section developed an initial empirical relationship, which provided a rep-

resentation for the flow stress during an EAM compression test across different
materials. The preliminary results from this study show positive results for the
classification of flow stress over short and larger ranges of strain. Also, the results
obtained from this study have allowed review of the many aspects that affect the
flow stress in relation to the addition of electrical current.

4.5.2 Constant Current Density

Considering a model that would be capable of mapping the flows stress for an
electrically assisted forming process is highly useful for process planning and
parameter selection. The previous section presented an empirical model for
4.5  Empirical Modeling Strategies 79

Non-constant current density (NCCD) testing that mapped the material flow stress
for EAF. The model is shown in Eq. (4.11).
 
σpredicted = K ′ εn ε̇m exp −(ε − c)B�A (4.11)

where σpredicted is the predicted flow stress with the application of electricity, K′ is
the strength coefficient, ε is the strain, n is the strain hardening exponent, ε̇ is the
strain rate, m is the strain rate sensitivity exponent,  is the current density, and A,
B, and C are material specific constants for this model.
Using this model, the NCCD tests were fit using the stress predictor equation
to examine the relationship for 304 Stainless Steel and Grade 5 titanium. The
material specific constants were determined using least squares regression and
are shown in Table 4.4. The corresponding results for the 304 Stainless Steel and
Grade 5 titanium are shown in Figs. 4.26 and 4.27, respectively. As shown, the
model was fairly accurate at predicting the flow stress behavior for both materi-
als. On average, the stainless steel experimental flow stress curves were averaged
or approximated using this flow stress predictor. When considering titanium, the
large reductions that were seen at higher current densities were fairly approxi-
mated, however, from a possible threshold effect that causes a non-linear increase
in the flow stress reductions, the lower current density curves suffered in predict-
ability and were underestimated.

Table 4.4  NCCD model A B C
constants
SS304 1.39 0.007869 −0.1385
G5 titanium 3.83 7.95 × 10−6 0.00234

Fig. 4.26  SS304 NCCD empirical modeling

80 4  Macroscale Modeling of the Electroplastic Effect

Fig. 4.27  G5 titanium NCCD empirical modeling

Using this model for a constant current density (CCD) where the current den-
sity does not change during the test as described by Eq. (4.10), hence reduces the
flow stress predictor in Eq. (4.11) to an approximate linear flow stress modifier.
Thus, this does not allow for an effective estimator of the flow stress for CCD as
there are significant reductions in flow stress and great variation when increasing
from current density to current density. Using the present data, an alternate rela-
tionship is introduced and based off of the CCD data. The proposed relationship
for a CCD tests is presented in Eq. (4.12). The new predictor function is based off
of a power function relation that corresponds to the strain imposed in the material
and the CCD value.
D
σpredicted = K ′ εn ε̇m (A� + B)εC� (4.12)

where σpredicted is the predicted flow stress with the application of electricity, K′ is
the strength coefficient, ε is the strain, n is the strain hardening exponent, ε̇ is the
strain rate, m is the strain rate sensitivity exponent,  is the current density, and A,
B, C, and D are material specific constants for this model.
The resultant model material specific constants are shown in Table 4.5 and the
predicted flow stress for the SS304 and the Grade 5 titanium is shown in Figs. 4.28
and 4.29, respectively. Of note is the fact that the titanium results were only pre-
dicted until the baseline fracture.

Table 4.5  CCD model A B C D
constants
SS304 −0.018 1.101 −0.00061 1.592
G5 titanium −0.015 1.053 −0.00048 1.560
4.5  Empirical Modeling Strategies 81

Fig. 4.28  SS304 CCD empirical modeling

Fig. 4.29  G5 titanium CCD empirical modeling

Overall, when comparing the experimental data to the modeled data, the newly
introduced flow stress predictor does an accurate job at estimating the flow stress
for a CCD. It should be noted that for the NCCD and CCD flow stress predictors,
the variables K′, n, and m are temperature sensitive values; however, the room tem-
perature values were used as the modifying terms introduced in the model account
for the temperature sensitivity of these values.
82 4  Macroscale Modeling of the Electroplastic Effect

4.5.2.1 Conclusion for Constant Current Density Empirical Modeling

Using prior empirical models, the NCCD data were fitted to expand the current
capability of the original empirical model. Also, as a result of the CCD tests, a
newly developed flow stress predictor was introduced to predict the flow stress
for CCD tests as the prior empirically based model was not capable of being
employed. The results of the CCD predictor showed excellent mapping capability
of both the 304 Stainless Steel and Ti-6Al-4V materials.

4.6 Macroscale Modeling of the Electroplastic

Effect Conclusions

This chapter provided an introduction to the modeling approach taken in Chaps. 5

and 6. In particular, the concept of the EEC was presented. The EEC is defined as
the amount of electrically that contributes directly toward plastic deformation ver-
sus the total amount applied to the process. To establish the EEC, two methods were
presented that included a mechanical-based and thermal-based approach. After estab-
lishing the procedure with examples, the two approaches were compared and they
provided the same approximation for the EEC. Last, this chapter included a method-
ology and examples of empirical modeling for the bulk flow stress during EAF. The
empirical models were constructed for both NCCD and CCD experimental data.

References

1. Salandro WA, Bunget C, Mears L (2012) Modeling the electroplastic effect during electrically-
assisted forming of 304 stainless steel. In: ASME international manufacturing science & engi-
neering conference, MSEC2012-7241, p 10
2. Perkins TA, Kronenberger TJ, Roth JT (2007) Metallic forging using electrical flow as an
alternative to warm/hot working. J Manuf Sci Eng 129(1):84–94
3. Bunget C, Salandro WA, Mears L, Roth JT (2010) Energy-based modeling of an electrically-
assisted forging process. Trans North Am Manuf Res Inst SME 38:647–654
4. Salandro WA, Bunget C, Mears L (2012) A thermal-based approach for determining elec-
troplastic characteristics. In: Proceedings of the Institution of Mechanical Engineers, Part B:
Journal of Engineering Manufacture, doi:10.1177/0954405411424696
5. Salandro WA, Bunget C, Mears L (2011) Thermo-mechanical investigations of the elec-
troplastic effect. In: ASME international manufacturing science & engineering conference,
MSEC2011-50250, p 10
6. Roth JT, Loker I, Mauck D, Warner M, Golovashchenko SF, Krause A (2008) Enhanced form-
ability of 5754 aluminum sheet metal using electric pulsing. Trans North Am Manuf Res Inst
SME 36:405–412
7. Salandro WA, Bunget C, Mears L (2012) Electroplastic modeling of 304 stainless steel. Under
review for publication in the Journal of Manufacturing Science and Engineering
8. Jones JJ, Roth JT (2009) Effect on the forgeability of magnesium AZ31B-O when a continu-
ous DC electrical current is applied. In: ASME international manufacturing science & engi-
neering conference, MSEC 2009-84116, p 10
Chapter 5
Compressive Electrically Assisted
Forming Model

5.1 Analytical Modeling of Compression Forming Processes

The overall goal of this chapter is to develop a model for an electrically assisted
forging process and then use this model to highlight specific sensitivities and rela-
tionships within the EAF process. A modeling strategy will be defined, based on
existing conventional forging equations. These equations will then be modified to
account for the electroplastic effect in an EAF process.

5.1.1 Definition of an EAF Modeling Strategy

The first task in creating a thermo-mechanical model is to define how EAF could
be modeled, with the optimization of current metal forming equations by creating
new coefficients to capture electricity’s effects. As is explained below, an electroplas-
tic effect coefficient (EEC) is created to help account for the direct electrical effects
on the forming process [1]. The objective of this chapter is to establish a closed-
form solution describing the flow stress of a material, which is plastically deformed
under a compressive load while direct electric current is applied through the dies.
Equations predicting the effective stresses are derived from balancing the input and
output energies of the system. These energy-based equations are aimed at predicting
the effective stresses and strains of the system, and the forming loads for different
parameters of the applied current. It is expected that this model may be beneficial in
the preliminary determination of the feasibility of electrically assisted forging a part
based on formability improvement. In particular, the model would determine process
input requirements for given formability improvements, in order to support process
design decisions by the user. An analytical model brings a better understanding of the
electroplastic deformation mechanism and opens new avenues for derivation of solu-
tions for more complicated manufacturing processes.

© Springer International Publishing Switzerland 2015 83

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_5
84 5  Compressive Electrically Assisted Forming Model

The derivation begins by establishing expressions for the effective stresses

and strains acting on a plastically deformed billet without current in the classi-
cal cylindrical upsetting test using the upper bound analysis. This approach uti-
lizes key metal forming relations derived from works by Lange [2], Backofen [3],
Kobayashi et al. [4], Wagoner and Chenot [5]. The material is assumed to follow
the power law as shown in Eq. (5.1):
σ̄ = C ε̄n (5.1)
where σ̄ is the effective flow stress, C is the strength coefficient, ε̄ is the effective
strain, and n is the strain hardening exponent. After the equations for stress and
strain are established, the next step is to include the effects of the electrical current
applied to the part, which takes place in the following section. Two aspects are
considered: (a) the applied electrical energy and (b) the electroplastic effect on the
material’s behavior and on the efficiency of the process. The analytical model is
verified with results (force and position data) from EAM compression tests run on
Al6061-T6511 specimens, by Perkins et al. [6]. Finally, the potential applications
of the developed analytical model are discussed.

5.1.2 Analysis of an Electrically Assisted Compression

Process

An experimental schematic of an electrically assisted compression test can be seen

in Fig. 5.1. A power source provides electricity and a variable resistor controls the
magnitude of current flowing through the dies. Insulation is placed between the
dies and the machinery, thus to ensure all electricity flows through the workpiece.
Electrons flow through the workpiece from anode to cathode; material flow is

p Thermal
camera
Press
Power Insulation Temperature
source Die measurements
+ Workpiece

_~ V DAQ
System

A Position and Force

measurements
p
Fig. 5.1  EAM compression test schematic. Main components include a controllable power sup-
ply, a thermal camera, and insulation placed between the dies
5.1  Analytical Modeling of Compression Forming Processes 85

shown schematically in Fig. 5.1. Friction is present at the workpiece/die interfaces,

thus causing the part to bulge into a specific barrel shape. This barreling effect is
neglected in the initial EAM modeling analysis.
The following major simplifications and assumptions are used throughout the
derivations in this study:
• The material is homogeneous and isotropic.
• The elastic deformation is negligible and the power law is used for the flow
stress of the material.
• The von Mises yield criterion is applied.
• The friction of the workpiece/die interfaces follows Coulomb’s friction law.
Sliding friction conditions exist, and barreling is ignored.
• The specific heat/resistivity of the material is independent from temperature.
• The strain rate sensitivity is neglected since previous experimental investiga-
tions found that the temperature rise is not high enough for the process to be
considered hot forming.
• Volume is maintained.
• The lubrication regime is unaffected by the presence of electricity.

5.1.3 Effective Stress and Strain—Classical Compression

Test

The energy of a process is the total work needed to perform that process. The power
of a process relates the required work to a particular speed at which the work is
performed. The power needed to plastically deform a cylindrical billet is found by
employing the upper bound analysis method as presented by Avitzur [7]. The objec-
tive of this method is to determine the strain rate field that minimizes Eq. (5.2):

J ∗ = Ẇi + Ẇs + Ẇb + Ẇk + Ẇp + Ẇγ (5.2)

where J ∗ is the upper bound on power, Ẇi is the internal power of deformation,
Ẇs is the shear power given by the discontinuity of power to overcome opposing
external forces (back stress), Ẇk is the inertia velocity, Ẇb is the power due to the
inertia forces, Ẇp is the pore opening power, usually neglected under the volume
constancy assumption, and Ẇγ is the surface power that includes the rate of intro-
duction of new surfaces in the deformation process. This component can also be
assumed negligible in bulk deformation, but will be of more importance as this
method is extended to sheet metal forming. The power components were computed
by Avitzur [7] for the ideal case where no bulge is formed, and the shape of the bil-
let is assumed to remain cylindrical throughout the deformation process, unaffected
by the interface friction. Using the average pressure approximation, where the pres-
sure on the workpiece is assumed to be constant across the cross section, the follow-
ing equation for total power needed for compression can be derived as Eq. (5.3):
86 5  Compressive Electrically Assisted Forming Model

 
2µr0
J ∗ = πrinst
2
σ̄ u̇ 1 + = F u̇ (5.3)
3hinst
where rinst and hinst are the instantaneous radius and height of the workpiece, σ̄ is the
effective flow stress of the material, u̇ is the velocity of the compressive die, r0 is the
initial radius of the billet, and F is the forming load. When the applied load and die
velocity are known, the pressure exerted by the die is determined as Eq. (5.4):
 
F 2µr0
p= 2
= σ̄ 1 + (5.4)
πrinst 3hinst

The power law will be used to formulate the relation between the effective stress
and strain, thus the pressure will be given by Eq. (5.5):
 
2µr0
p = C ε̄n 1 + (5.5)
3hinst
By employing the slab analysis method and simple balance of forces in the axial
direction, along with the von Mises criterion, the compressive and effective
stresses are determined from Eq. (5.6):
σz = p, σθ = σr = p − σ̄ ,
εz = −2εθ = −2εr ,
  (5.6)
h0
ε̄ = −εz = ln ,
hinst

where σz , σθ , σr and εz , εθ , εr are the principal stresses and strains in axial, hoop,
and radial direction, respectively, and h0 is the initial height of the billet.

5.1.4 Effective Stress and Strain—Electrically Assisted

Compression Test

The flow stress required to deform a specimen to a desired strain is lowered by

applying electric current during the process. In this section, the effective electrical
energy utilized in reducing this stress is formulated.
When electricity is applied during deformation, a part of the electrical energy
is imparted into the mechanical deformation process. Thus, the total power con-
sumed by the process will be given by Eq. (5.7):
J ∗ = Jm∗ + Je∗ (5.7)
where Jm∗ is the mechanical component given by the product of the forming load and
the die velocity, and Je∗ is the effective (usable) electrical power, as shown in Eq. (5.8):
Je∗ = ξ · Pe (5.8)
5.1  Analytical Modeling of Compression Forming Processes 87

where ξ is defined as the EEC, and Pe is the power of the electrical current pass-
ing through the workpiece. It is known from previous experimental research that
the EEC is a new material property coefficient introduced here, and it reflects the
ratio of the electrical power that contributes toward plastic deformation to the total
input electrical power [1]. The EEC, ξ, depends on the material, applied current
density and time, and can be determined through the tests defined below. Electric
current will also increase the part temperature, thereby lowering the flow stress of
the material and contributing to decreased required work.
Through resistive heating, the electric current results in bulk thermal soften-
ing of the material, thereby decreasing flow stress as represented by the strength
coefficient C in Eq. (5.5). In our analysis, a stepwise approach is used, whereby
the material is strained by dε, the contribution of electrical energy to the deforma-
tion work calculated, the remaining electrical energy converted to heat, then the
material workpiece temperature rise and subsequent effect on strength coefficient
derived. The process is repeated to the final desired strain.
When the electric current passes through the metallic workpiece, heat is gen-
erated and the temperature of the part rises. Heat transfer and thermodynamics
knowledge was gathered from Cengel [8] and Cengel and Boles [9] to derive the
thermal equations for this section. The temperature rise can be determined from an
energy balance conforming to Eq. (5.9)
∂ρU ∂  
+ Q̇conv + Q̇cond = Qrad + (1 − ξ )Pe (5.9)
∂t ∂xj
 
where ∂ρU ∂
∂t is the rate of change of the internal energy of the part, ∂xj Q̇conv + Q̇cond
are the convective and conduction components of the heat flux, Qrad is the radiation
heat, and (1 − ξ )Pe is heat generated in the part from the electric energy dissipated.
Using constitutive equations for each component, the heat equation to be solved for
that determines the temperature rise for particular electric parameters is Eq. (5.10):

∂T ∂ 2T  
ρVv Cp = −As [h(T − T∞ )] − 2kAc 2 − As εσSB T 4 − T∞
4
+ (1 − ξ )VI
∂t ∂xj
(5.10)

where ρ is the density of the material, Vv is the volume of the part, Cp is the spe-
cific heat of the material, T is the temperature, t is time, As is the lateral surface of
the part, h is the convection heat transfer coefficient, T∞ is the surrounding tem-
perature, k is thermal conductivity of the die material, Ac is the cross-sectional
area, xj are coordinates, ε is radiative emissivity for the part, σSB is the Stefan–
Boltzmann constant, V is the electric voltage, and I is the intensity of the current,
given by the product of the current density and cross-sectional area.
The local heating of the workpiece influences the flow of the material. At room
temperature, the flow is given by the power law presented earlier, but as the tempera-
ture rises, the flow curve depends strongly on the temperature. If the temperature is
higher than the temperature at which recovery and recrystallization take place, then
the flow depends also on the strain rate of the process, conforming to Eq. (5.11)
88 5  Compressive Electrically Assisted Forming Model

σ̄ = C ε̄n ε̄˙ m (5.11)

where C is the strength of the material, ε̄˙ is the effective strain rate, and m is the
strain rate sensitivity.

5.1.5 Current Density Relationship During Electrically

Assisted Compression

Work by Ross et al. [10] shows that the electricity in an EAF process had to be
applied differently for different forms of metal deformation. In particular, Ross et
al. determines that applying EAF continuously for a tensile process caused detri-
mental effects. This is due to the increasing current density throughout a tensile
process, due to the shrinking cross-sectional area. Conversely, the cross-sectional
area of a workpiece in a compression-based process increases, leading to a
decreasing current density over the duration of the test.
The term “current density” will be used frequently throughout this chapter.
When used, it represents the starting current density, not the instantaneous cur-
rent density. For the forging experiments in this chapter, the current density will
decrease since the cross-sectional area of the workpiece will increase during
deformation. Equation (5.12) shows this relation as:
 
Ao
CDact = CDnom · (5.12)
Af

where CDact is the actual current density during the deformation process, CDnom is the
  Ao is the initial cross-sec-
nominal current density (i.e., the starting current density),
tional area, and Af is the final cross-sectional area. As Ao decreases, the current den-
Af
sity decreases as well. Equation (5.12) can be rewritten in terms of strain as Eq. (5.13):
 
1
CDact = CDnom · = CDnom · e−ε (5.13)
eε
where ε is the material strain.
Figure  5.2 shows the change in current density over the duration of an electri-
cally assisted forging test. The specific tests in the figures are of SS304 specimens
(7.154 mm × 4.761 mm diameter) that were deformed at dies speeds of 12.7 and
25.4 mm/min. From the figure, the ending current density is about 41 % of the start-
ing current density, hence there is a significant change in the current density over the
duration of the EAF tests if current is held constant. Note that derivations in this chap-
ter are based on current level, and not current density; these approaches are contrasted
in Chap. 4. The modeling includes electrical power, which is obtained by the current
(which is held constant in all tests). The starting current density value of any EAF test
is used purely as a classification mechanism and not as a modeling variable, because
it is dependent on the cross section of the part and this is constantly changing.
5.1  Analytical Modeling of Compression Forming Processes 89

Fig. 5.2  Current density 25
decrease during an EA
forging test. Since the cross-

Current Density (A/mm2)

20
sectional area increases in 12.7 mm/min
compression, the current
density will decrease over the 15
duration of the test
10
25.4 mm/min

0
0 5 10 15 20
Time (s)

5.1.6 Strain and Temperature Effect on Resistance and

Current

By virtue of strain and material heating, the resistance of the sample, and ulti-
mately the induced current, is affected. Referring to Fig. 5.1, the DC source is a
voltage source; the current is set through adjustment of the voltage and a variable
resistor as shown. During deformation, the resistance is affected by the workpiece
geometry and the material resistivity according to Eq. (5.14).
L
R[�] = ρ . (5.14)
A
For Al6061-T6, the resistivity (r) is given by Eq. (5.15):
r = 4 × 10−8 (1 + 0.0039T )[�m]. (5.15)
For specimens at different nominal applied current densities, the reduction in
resistance of the specimens is plotted in Fig. 5.3. We observe that the geometric
effects dominate and that the specimen resistance decreases up to 87 % in the test
data. However, the overall system resistance is dominated by the variable resistor,
so this effect is negligible.

5.1.7 Analytical Model for Electrically Assisted Compression

The proposed thermo-mechanical analytical model is based on the stress–strain

equations presented for the classical test, but adapted to take into consideration
the various effects of the electrical energy on the deformation mechanism. Thus,
Eq. (5.16) represents the amount of electrical energy contributing to the deforma-
tion, as presented in Eqs. (5.7) and (5.8), and introduced in Eqs. (5.3)–(5.6).
90 5  Compressive Electrically Assisted Forming Model

Fig. 5.3  Specimen resistance 100%

change with respect to 90%

Normalized Resistance [Ω/Ω]

strain and temperature 0
80%
[1]. Although the overall
70% 27.7 A /mm2
normalized resistances
35.8 A /mm2
change with the strain 60%
54.6 A /mm2
(because of the decreasing 50%
specimen height), there is 60.8 A /mm2
40%
no notable difference in
resistance change in relation 30%
to the specimen temperature 20%
(different current densities) 10%
0%
0 0.2 0.4 0.6 0.8 1 1.2
True Strain

 
∗ 2µr0 2
J = F u̇ + ξ · Pe = σ̄ 1 + πrinst (5.16)
3hinst
where the effective stress will be given by Eq. (5.11). Thus, Eq. (5.17) can be writ-
ten as follows:
 
n ˙m 2µr0 2
F u̇ + ξ VI = C ε̄ ε̄ 1 + πrinst (5.17)
3hinst
where V and I are the voltage and current intensity. The current is given by Eq. (5.18):
   Amps 
2 2
I = πrinst Cd [=] mm · (5.18)
mm2
with Cd being the current density (i.e., current/cross-sectional area) and π rinst 2

being the cross-sectional area of the workpiece. Equations (5.10), (5.17), and
(5.18) constitute the analytical model for an electrically assisted compression test.

5.1.8 Overall Solution Schematic

A MATLAB program was implemented to numerically solve the equations derived

for the model using an incremental approach. By imposing the material and ini-
tial dimensions of the part, the model is used to determine the effective strain
and stress for different current density values. The solution schematic is given in
Fig.  5.4. A mechanical power step, Pm, is preset such that Pm increases incre-
mentally throughout the test. In particular, the overall mechanical power pro-
file is divided into a given number of steps, at which mechanical forces will be
determined. More steps will produce higher modeling accuracy; however, it will
increase modeling time. At step i, Eq. (5.10) is solved to determine the temperature
rise and then used to find the corresponding strength coefficient, C. The new value
is used in Eq. (5.15) and the height of the billet is calculated. The stroke, effective
5.1  Analytical Modeling of Compression Forming Processes 91

Fig. 5.4  Solution scheme
for solving the analytical
model [1]. The EAF model
is set such that all important
variables are established
first, and then the desired
compression stroke is divided
into increments at which the
model will output data for
each increment

strain and stress, and the complete state of stress and strain can be determined if
desired. The iteration continues until deformation reaches the desired value of the
stroke, or a failure criterion, such as the fracture limit, is reached.

5.1.9 EAF Modeling Approach Summary

This chapter presents an energy-based approach to analyze electrically assisted

forming (EAF). The input electrical energy is separated into the useful energy that
assists the mechanical deformation process, while the remaining energy is con-
verted to heat, causing thermal softening by way of resistive heating. The EEC
is created to account for the ratio of “usable” electricity compared to the overall
magnitude of applied electricity.

5.2 Simplified EAF Forging Model

In this section, the modeling strategy and different methods for determining EECs
is combined to produce a simplified EAF relationship model. This model predicts
stress–strain profiles and will be compared to experimental conventional/EAF
compression tests to verify accuracy. In addition, sensitivities of the model, and of
EAF in general, will be explained with supporting experimental results.

5.2.1 EAF Forging Stress–Strain Model

In this section and preceding subsections, an explanation of the simplified EAF

forging model is described and compared to experimental results. In particular, this
section includes an overview of the simplified modeling strategy, an explanation of
92 5  Compressive Electrically Assisted Forming Model

the experimental setup and procedures used to calibrate the model for accuracy, a
discussion of the experimental/analytical results, an efficiency analysis comparing
the magnitude of applied electrical power to the decreases in the forming stress,
and finally conclusions on the EA forging model [11]. The resulting stress–strain
profiles, after the EEC values are fed into the thermo-mechanical model, will be
displayed and explained. Additionally, the efficiency of applying electricity to the
deformation process and the benefits gained will be provided.

5.2.2 Modeling Strategy Overview

As part of the modeling strategy, an EEC was previously created to represent the
efficiency of the applied electricity (i.e., how much of the applied electrical power
contributes toward plastic deformation vs. how much contributes toward resistive
heating). EAF tests were performed on SS304 specimens, while varying the die
speed and the starting current density (i.e., the current magnitude was held constant
throughout each test, and density varied according to the instantaneous cross-sec-
tional area). The “current density” values are based on the initial cross-sectional area
of the specimen. As the part is deformed, the cross section will increase. However,
in this chapter, the current magnitude (i.e., the current selected at the beginning of
the test in order to produce a particular starting current density) will remain constant
throughout the test. This means that the current density will decrease as deforma-
tion takes place and the workpiece diameter increases. In Chap. 4, the results from
these experiments were used to determine the EEC for stainless steel using the
mechanical- and thermal-based approaches, and the results from both methods were
compared. Now, the EECs will be fed into a thermo-mechanical model in order to
predict stress–strain profiles comparable to the experiments.
The hybrid modeling strategy introduced by Bunget et al. [1], accounts for the
thermal, mechanical, and coupled thermo-mechanical aspects of the direct electri-
cal effects witnessed during EAF. Further, the EEC was conceptualized to quantify
the “usable” direct electrical effects. The interrelations between the thermal and
mechanical aspects of EAM, and how they relate to the workpiece temperature
and force profiles are depicted in Fig. 5.5.

5.2.3 Coupled Thermo-Mechanical Modeling

As stated, the electroplastic effect is the fraction of electrical power imparted into
the plastic deformation process, as shown below in Eq. (5.19):
Pe = Pheat + Pdef = η(1 − ξ )VI + ηξ VI (5.19)
where Pe is the electric power (I · V), η is the efficiency, and ξ is the EEC [12].
Pheat represents the amount of electric power that will dissipate into heat through
5.2  Simplified EAF Forging Model 93

Mechanical
Electrical Effects Thermal
Deformation
Modeling Modeling
Modeling

Mechanical Force Profile Thermal Profile

Fig. 5.5  Coupled relations of EAM modeling [13]. The following modeling aspects specifically
contribute to the mechanical force profile, the thermal profile, or both. It is assumed that, from
the entire amount of electricity applied to the process, a portion contributes toward plastic defor-
mation, and the remainder contributes toward resistive heating, which increase the temperature of
the part and also contributes toward thermal softening

resistive heating of the workpiece. Pdef is the magnitude of the reduction of the
required mechanical power to deform, due to the total applied electrical power.
Figure  5.6 shows that the magnitude of the thermal softening effect is highly
dependent on the specific metal being formed, as well as the actual temperatures
reached during the manufacturing process. Of note, in the thermo-mechanical
model, the material strength coefficient, C(T ) is adjusted to one constant value or
to a temperature-dependent profile representative of the temperatures reached for
specific EAF processes. As a result, the decrease in the force needed for the same
deformation as in a conventional test can be considered through two mechanisms:

Fig. 5.6  Tensile strength 1200

versus temperature for
various metals [14, 15]. 1000
Tensile strength (MPa)

Tensile strength versus Ti-G5

temperature profiles show 800
that the strength is reduced
by different magnitudes as 600
the temperature is increased 304 SS
for Al6061-T6511, SS304, 400
and Ti-G5
200
Al6061-T6511

0
0 200 400 600 800 1000
Temperature (oC)
94 5  Compressive Electrically Assisted Forming Model

(i) thermal softening, thus lower flow stress, and (ii) facilitating deformation due
to the energy provided by the electrons to the dislocations.
Within Eqs. (5.20) and (5.21), the strength coefficient of the material is
included in the mechanical modeling of the EAF technique. Figure 5.6 below
shows the relationship between the strength and temperature for several metals
that were tested with EAF throughout the research (Al6061-T6511, SS304, and
Ti-G5). From the figure, each metal’s strength has a different response as tempera-
ture is increased. Additionally, there are temperature regions where the strength is
unaffected by the change in temperature (i.e., 200–400 °C for Ti-G5), and other
regions where the strength significantly decreases as temperature is increased (i.e.,
150–200 °C for Al6061-T6511). With this said, the sole impact of increased tem-
perature (without direct electrical effects) has the potential to have an impact on
the overall mechanical profile of the process, depending highly on the material and
specific temperatures reached during the process. However, the presence of elec-
tricity is significant over bulk heating effects.
Equations (5.20) and (5.21) are used for calculating the flow stress of a conven-
tional and an EAF test, respectively.
   n  
n 2µro hinst 2µro
σconv = Cε 1+ = C · ln · 1+ [MPa] (5.20)
3hinst ho 3hinst

� �
� � ηξ IV
2µro 106
σEAF = Cεn 1 + − 2 u̇
[MPa] (5.21)
3hinst πrinst

where is C is the strength coefficient, ε is the material strain, n is the strain harden-
ing exponent, µ is the friction coefficient, ro is the starting specimen radius, ho is
the starting specimen height, hinst is the instantaneous specimen height, ξ is the
EEC, I is the electrical current, V is the electrical voltage, and u̇ is the die speed.

5.2.4 Assumptions of the Thermo-Mechanical Model

The following are the major assumptions for the model [13] (more general
assumptions can be found in [1] and [12]):
• The material is homogeneous, isotropic, and the density is uniform throughout
the specimen.
• The increase in temperature due to friction at the die/workpiece interfaces can
be neglected.
• Barreling effect due to friction is neglected.
• Specific heat and resistivity are assumed constant with temperature.
• Die conduction length is estimated from thermal camera images.
5.2  Simplified EAF Forging Model 95

Mechanical Load

e- e- e- e- e- e-
Top Die

Lc Zone 1, R1

LTotal Lw Zone 2, R2

Workpiece
Lc
Zone 3, R3

Bottom Die
e- e- e- e- e- e-

Mechanical Load

Fig. 5.7  Die and workpiece zones [13]. The dies/workpiece were divided into three zones where the
top and bottom zones accounted for contact area/interface obstacles, while the center zone did not

• The workpiece and dies are divided into three zones, as shown in Fig. 5.7, where:
– Zone 1 is equivalent to Zone 3.
– The real contact area <80 % of the apparent contact area for Zones 1 and 3
and has a length of Lc, which depends on the surface aspect (asperity peaks).
– The real contact area is equal to the apparent contact area in Zone 2.
• One voltage value is used in the thermal/mechanical calculations, but it is com-
prised of proportioned voltages from each zone (Vz1, Vz2, and Vz3), as shown below.
The value used for Lc in Eq. (5.22) was 1.27 mm. This was about 20 % of the total
specimen height. This was an estimation based on the thermal videos, where exces-
sive heating was apparent at these locations when electricity was applied.
        
Lc Lw Lc
VTotal = · VZ1 + · VZ2 + · VZ3 (5.22)
LTotal LTotal LTotal

Equation (5.22) is a simplified approach to the more complicated problem of

thermal contact resistance. In reality, Zones 1 and 3 are regions where the cur-
rent flow between the dies and specimen is constricted to the narrow intermetallic
contacts. A thermal contact resistance calculation should include the resistance of
96 5  Compressive Electrically Assisted Forming Model

Fig. 5.8  Thermal profiles for 200

partial contact and sufficient 180
contact between the dies and 160

Temperature (°C)
workpiece [13]. If there is not 140
sufficient contact between 120
the dies and the workpiece, 100
then sparking could occur 80 Partial Contact
and generate steep spikes at 60 Sufficient Contact
the beginning of the thermal 40
profile for that test 20
0
0 5 10 15 20
Time (s)

the air pockets and the resistance of the asperity peaks of both surfaces in contact.
Therefore, the real contact area depends on the surface finish of the dies and the
workpiece. During a stationary electrical test, there is no significant change in the
asperities aspect, while during the deformation process, the asperities are flattened
to some extent, and thus, the real contact area increases. The increase in contact
area results in a decrease of the resistance of the zone, thus a decrease in voltage
drop across the junction point. A more in-depth calculation of the real contact area
and its influence is a point of future research for this process.
The change in the real contact area can have an effect on the workpiece tempera-
ture profile throughout the EAF tests. Figure 5.8 displays representative temperature
profiles along the longitudinal axis of the workpiece at the beginning and toward the
end of an EAF test. Of note is that, at the beginning, where the real contact area is
lower, the die/workpiece interfaces will be hottest, since the electrical power dissipa-
tion will be greatest at these points. Conversely, the center portion of the workpiece
will be hottest at the end of the test, since the real contact area will be larger at the
interfaces than at the beginning, thus lowering the electrical power. Figure 5.8 helps
to illustrate this phenomenon. In this figure, there are two EAF tests: (1) a test with
sufficient contact (i.e., a large real contact area) and (2) a test with partial contact
(i.e., a small real contact area). The test with partial contact at a preload of about
670N shows a spike at the beginning of the test where the power at the die/work-
piece interfaces is extremely high until sufficient pressure is applied to the part. The
other thermal profile, where a much larger preload of about 1780N was used, did not
show this thermal spiking phenomenon at the start of the electricity application.

5.2.5 Experimental Setup and Procedure

The experimental objectives of this section are to: (1) compare and validate the
thermo-mechanical model via comparing experimental results to modeling results,
and (2) discuss the efficiency of the EAF tests performed in relation to applied elec-
trical power versus stress reductions. This is approached through design of experi-
ments testing using current density and die speed as the input variables, with flow
5.2  Simplified EAF Forging Model 97

Table 5.1  Test combinations
Deformed/Stationary Conv./EAF Current density (A/mm2) Die speed (mm/min)
Deformed Conv. – 12.7
Deformed EAF 10 12.7
Deformed EAF 15 12.7
Deformed EAF 20 12.7
Deformed EAF 25 12.7
Deformed Conv. – 25.4
Deformed EAF 15 25.4
Deformed EAF 25 25.4
Stationary EAF 10 –
Stationary EAF 15 –
Stationary EAF 20 –
Stationary EAF 25 –

stress per strain as the response. All specimens in this research were made from the
same 4.61 mm-diameter rod of 304 Stainless Steel. An EDM process was used to cut
the specimens to a height of 7.154 mm and to maintain surface finish consistency.
Deformation and stationary electrical tests were run as part of this work. The
deformation tests were run at speeds of 12.7 and 25.4 mm/min, coupled with start-
ing current densities of 10, 15, 20, and 25 A/mm2. Further, stationary electrical
tests, where a small static load is applied to the specimen while an electrical cur-
rent is applied, were also run at the same starting current densities as deforma-
tion tests. Three repetitions were run on all test configurations. Of note is that, all
of these tests were consistent with each other; therefore, only a single test will
be plotted which is representative of the three tests for each test combination. All
experimental test combinations can be seen in Table 5.1.
Figure 5.9 displays the experimental setup for both the deformation and stationary
electrical testing (specimen area is magnified in the upper-left corner of figure). An
Instron Model 1332 hydraulic testing machine was used to compress the specimens
at the desired die speeds. Machined, hardened, and insulated dies, made from A2 tool
steel, were fitted to the Instron machine (insulation was used to isolate the electricity
from the test machine). For thermal measurements, a FLIR A40 M thermal camera,
with a temperature capacity of 550 °C and a resolution of 0.1 °C, was utilized. All
force and position data was gathered using an Instron on-board data acquisition system.

5.2.6 Experimental and Modeling Results

This section contains two subsections. First, several stress–strain profiles will be
shown (conventional and EAF) to illustrate that the various test parameters are
indeed in the electroplastic region, where formability improvements are apparent.
Second, the EEC profiles, obtained using both methodologies, will be compared to
each other.
98 5  Compressive Electrically Assisted Forming Model

Upper Grip

Electric Lead

Workpiece

Forging Die

FLIR Camera

Lower Grip

Fig. 5.9  Experimental test setup. The close-up view of the dies and specimen is shown in the
top left-hand corner

5.2.6.1 Force Reduction Due to EAF

Figures 5.10 and 5.11 display the stress–strain profiles of baseline and EAF tests run
at die speeds of 12.7 and 25.4 mm/min, respectively. In both figures, when a higher
current density is applied, the reduction in material flow stress becomes greater. Of
note is that, irrespective of the die speed, the 25 A/mm2 EAF test reduced the over-
all flow stress compared to the baseline test by approximately 30 % in both cases.
This significant reduction in flow stress signifies that the electrical parameters that
were run were well within the “electroplastic region” with respect to this particular
material. Additionally, when comparing these results with results from Perkins et al.

Fig. 5.10  EAF flow stress 1400

CD15
reduction (12.7 mm/min) CD10
1200 Baseline
[13]. The reduction in flow
True Stress (MPa)

stress is proportional to the 1000

magnitude of the applied
electricity 800 CD20

CD25
600

400

200

0
0 0.2 0.4 0.6 0.8 1
True Strain
5.2  Simplified EAF Forging Model 99

Fig. 5.11  EAF flow 1400

stress reduction (25.4 mm/
1200
min) [13]. The flow stress

True Stress (MPa)

reductions are still related 1000
to the amount of applied
800
electricity CD15 CD25
600
Baseline
400

200

0
0 0.2 0.4 0.6 0.8 1
True Strain

[6] where the same metal composition was tested with EAF, the same general trends
were witnessed. Additionally, from both figures, it can be seen that the baseline tests
for 12.7 and 25.4 mm/min die speeds are very similar, with the 25.4 mm/min die
speed producing only a slightly higher stress–strain profile. To this end, the strain
rate sensitivity of this material was minimal when conventionally formed at the two
die speeds tested. However, the behavior is unknown at higher die speeds.

5.2.6.2 Experimental Versus Modeling Stress–Strain Results

The EEC profiles that were determined are fed into Eqs. (5.20) and (5.21) to generate
a corresponding stress–strain profile. This predicted stress–strain profile can then be
compared to the experimental stress–strain profile for each current density/die speed
combination. Of note is that, the EEC profiles determined from the thermal method
will be used to generate the EAF stress–strain profiles (since they are very similar for
all tests, the thermally derived EEC profiles can be used to represent both profiles).
Figure  5.12 displays the experimental and analytical stress–strain profiles for
the baseline test and EAF tests (10 and 20 A/mm2) for a die speed of 12.7 mm/
min. The modeled profiles (shown with dashed lines) predict a slightly lower over-
all stress–strain profile, when compared to the experimental profiles. However,
the predicted profiles are within 10 % of the experimental results. Additionally,
the model at a current density of 20 A/mm2 was even able to predict some strain
weakening at strains between 0.2 and 0.4.
Predicted and experimental stress–strain profiles for a baseline test and an EAF
test at 25 A/mm2, for a die speed of 25.4 mm/min, are shown in Fig. 5.13. From
the figure, the model is able to predict the general stress–strain profile of the EAF
test, but not to the degree of accuracy as with the slower die speed tests in Fig. 5.12.
Additionally, the predicted profile shows an overall higher stress–strain profile com-
pared to experiments. Also, the model did not predict as much strain weakening as
the experimental test at 25 A/mm2. Nevertheless, the model is still within 20 % accu-
racy compared to experimental results throughout the entire strain range.
100 5  Compressive Electrically Assisted Forming Model

Fig. 5.12  Flow stress model 1400

(12.7 mm/min) [13]. The
1200
model and experiments

True Stress (MPa)

generally agree for all of the 1000
tests, with a percent error of
800
less than 10 %
CD20- Exp
600 CD10- Exp
Base- Exp CD20- Model
400 CD10- Model
200 Base- Model

0
0 0.2 0.4 0.6 0.8 1
True Strain

Fig. 5.13  Flow stress 1400

model (25.4 mm/min) [13].
1200
The model is not able to
True Stress (MPa)

predict the strain weakening 1000

to the same extent as the
800
experiments, but to within a CD25 -
percent error of about 20 % 600 Baseline - Exp
Exp
400 CD25 -
Baseline -
Model
200 Model

0
0 0.2 0.4 0.6 0.8 1
True Strain

5.2.7 Electrical Efficiency Analysis

Within this subsection, the efficiency of the EAF tests run and how the EEC can
be interpreted are discussed [13]. As previously mentioned, the EEC represents the
efficiency of the applied electrical power on the manufacturing process. To this
end, the magnitude of the EEC profile is determined on the effect that the applied
electrical power has, and not just on the amount of applied electrical power. More
specifically, if excess electrical power is added to the process and it cannot be uti-
lized for plastic deformation assistance, it will contribute only to resistive heating
and will result in a lower EEC.
Figure  5.14 displays EEC values at 50 and 100 % of the total compressive
stroke for both die speeds of 12.7 and 25.4 mm/min, at the current densities
tested for each die speed. The percentage of the total compressive stroke was used
instead of time such that the two different die speeds could be compared to each
other. From the figure, for both die speeds, the EECs were higher at 100 % stroke
compared to 50 % stroke. This signifies that the efficiency of the applied electri-
cal power (which is represented by the EEC) increases throughout the test and it
increased at a higher rate with the slower die speed. With the EAF technique being
5.2  Simplified EAF Forging Model 101

Fig. 5.14  EEC efficiency 0.6 50% Red. (12.7)

comparison [13]. The EEC 100% Red. (12.7)
values associated to the lower 0.5 50% Red. (25.4)
die speed are the greatest, 100% Red. (25.4)
0.4
except when coupled with the

EEC
highest current density value 0.3

0.2

0.1

0
10 15 20 25
Starting Current Density (A/mm2)

strain rate sensitive, in this case, the slower die speed may have allowed for the
applied electricity to continue improving the formability, whereas the faster die
speed does not allow for that.
Another notable factor is that the EEC values for the 12.7 mm/min die speed
were consistently above the EEC values for the 25.4 mm/min die speed, except
for the current density of 25 A/mm2. This supports the time dependency of the
EEC and signifies that a current density as high as 25 A/mm2 is not efficient with
a slower die speed of 12.7 mm/min because it provides excess electrical power to
the system that cannot be utilized, and it results in more wasted electricity by way
of heating. It is more efficient when coupled with a faster die speed of 25.4 mm/
min where a greater amount of the applied electrical power can be utilized, as
shown by the higher EEC values in the figure at a current density of 25 A/mm2.
A similar finding was determined from uniaxial EAF tension testing of Mg sheet
metal specimens [16]. In previous EAF works, a threshold effect, by which signifi-
cant formability improvements were witnessed above a particular current density,
was discovered [6]. This threshold appeared to be a function of die speed and cur-
rent density magnitude. For the other three lower current densities (10, 15, and
20 A/mm2), the slower die speed allowed for more of the electrical power gener-
ated to be utilized compared to the faster die speed, hence the slower die speed
produced greater EEC values than the faster die speeds for these current densities.

5.2.8 EAF Forging Model Conclusions

This section utilized EEC profiles, which were determined from mechanical power
profiles and thermal profiles, and utilized them to construct stress–strain profiles of
the EA forging process. The conclusions are as follows:
• The thermo-mechanical model predicted the stress–strain profiles to within
20 % of experimental results.
• The model was more accurate with the slower die speed.
102 5  Compressive Electrically Assisted Forming Model

• The contribution that thermal softening had on reducing the mechanical profile
of the process was dependent on the specific material (since each metal has a
different strength vs. temperature relationship) and the specific temperatures
reached during the EAF process.
• Factors such as die speed and current density significantly affect the efficiency
of the EAM technique. The concept was highlighted in the previous experimen-
tal work for tensile processes. This particular section shows that the current den-
sity and die speeds also need to be “balanced” in order to optimize results in
compression as well.

5.3 Specific Heat Sensitivity

Although the heating effects from EAF are not a major contributor to the formability
enhancements (as highlighted from the work by Ross et al.), these effects are utilized
in the thermal-based EEC determination method, where differences in temperatures
between stationary electrical and EAF tests were used to quantify the EEC profile.
Therefore, it is important to understand the effect that the temperature has on the spe-
cific heat of the material, and ultimately on the EEC profile (i.e., the efficiency of the
applied electrical power). In this analysis, specific heat sensitivities for both SS304
and Ti-G5 are evaluated at three different temperatures respective of the temperatures
reached during three different tests (conventional forming, EAF tests, and stationary
tests). Figure 5.15 displays the relationship between the specific heat values and the
temperatures for each metal (note that the maximum temperature that was reached
during stationary testing at 25 A/mm2 for each material is denoted by a black dot).
Additionally, Table 5.2 lists the exact specific heat values used for each test
description in the model per material [17]. The maximum temperature observed in

750
304 SS
700
Specific Heat (J/kg*°C)

Ti - G5
650

600

550

500

450

400
0 100 200 300 400 500 600 700 800 900
Temperature (°C)

Fig. 5.15  Specific heat versus temperature profiles (SS304 and Ti-G5) [17]. Relationships
between the specific heat values and temperatures for SS304 and Ti-G5
5.3  Specific Heat Sensitivity 103

Table 5.2  Specific heat Material Description Temperature (°C) Cp (J/kg

values at temperatures °C)
reached during testing
SS304 Conv. 30 475
(SS304 and Ti-G5)
EAF 250 503
Stationary 500 530
Ti Conv. 30 544
EAF 450 650
Stationary 800 690

Fig. 5.16  Specific heat
sensitivity (Ti-G5) [11].
Thermal profiles of stationary
electrical tests with specific
heat values for particular
temperatures Conv. (30oC)

EAF (450oC)

Stationary (800oC)

each type of tests was considered, in order to observe the maximum variability in
the results due to variation of Cp with temperature.
To analyze the sensitivities of the specific heat value, stationary electrical tests,
where a non-deforming static load is applied to the workpiece and a constant elec-
trical power is applied, were simulated using the previously developed thermo-
mechanical predictive model [1]. Figure 5.16 displays the thermal profiles for
stationary electrical tests with specific heat values representative of maximum tem-
peratures witnessed during conventional, EAF, and stationary electrical tests for
Ti-G5. Additionally, Fig. 5.17 displays the same information for Stainless Steel 304.
When comparing the two figures, one can see that the specific heat of the Ti-G5
material was more dramatically affected, and this resulted in a larger variation in
the specimen thermal profiles when using the three different specific heat values.
At one point, the difference in the temperature profiles of the three tests for Ti-G5
was about 50 °C, whereas the maximum difference between the three tests for
SS304 was about 15 °C. One possible reason for this is the intrinsic specific heat
property of the material and its relationship with temperature (Fig. 5.15), coupled
with the temperatures reached during testing. The black dots in the figure repre-
sent the maximum temperature reached for each material. The temperature range
reached for the Ti-G5 (0–800 °C) produced a 30 % increase in the metal’s specific
heat value as the temperature was increased. In comparison, the temperature range
reached for the SS304 (0–500 °C) produced a 12 % increase in the specific heat
104 5  Compressive Electrically Assisted Forming Model

Fig. 5.17  Specific heat
sensitivity (SS304) [11].
Thermal profiles of stationary
electrical tests with specific
heat values for particular
temperatures
Conv. (30oC)

EAF (250oC)

Stationary (500oC)

value as the temperature increased. In this case, the Ti-G5 material had a slightly
steeper specific heat versus temperature relationship, and the maximum temper-
ature reached during the Ti-G5 testing was about twice the temperature reached
during the SS304 testing. In summary, the specific heat of the material needs to
be considered when modeling EAF. It is highly dependent on the temperature
range reached during forming (between room temperature and the hottest electri-
cal temperature). As the specific heat increases, the energy required to increase the
temperature of the workpiece also increases, and this can lead to an overall lower
thermal profile. This is shown in Figs. 5.16 and 5.17, where the thermal profiles
using the highest specific heat values are the lowest.

5.4 Heat Transfer Modes Analysis

As electrical power is input into a system, as in the EAM technique, some level
of heating will take place. To this extent, heat will be lost by way of the three heat
transfer modes (conduction, convection, and radiation). The temperature increase
of the workpiece can be represented by Eq. (5.23):

Ėpart = Ėg − Ėout ,

(5.23)
Q̇ = Q̇e − Q̇cond − Q̇conv − Q̇rad ,
 
where Ėpart is the rate of change of the energy content in thepart analyzed Q̇ , Ėg is
the rate of heat generation due to the electric resistive heating Q̇e , and Ėout is the rate
of energy leaving the part due to the sum of conduction into the dies in the contact zone
(Q̇cond), convection (Q̇conv), and radiation from the surface (Q̇rad). To clarify, Ėg has only
one contributor (Q̇e), whereas Ėout has three contributors (Q̇cond, Q̇conv, Q̇rad) that are the
three forms of heat transfer. The energy is the total amount of heat lost or gained, and
the power is the rate of heat lost or gained over time. Using constitutive equations for
each component, the heat equation that determines the temperature rise for particular
electric parameters is shown in Eq. (5.24):
5.4  Heat Transfer Modes Analysis 105

∂T ∂ 2T  
ρVv Cp = −As [h(T − T∞ )] − 2kAc 2 − As εr σSB T 4 − T∞ 4
+ (1 − ξ )VI
∂t ∂xj
(5.24)
where ρ is the density of the material, Vv is the volume of the part, Cp is the spe-
cific heat of the material, T is the temperature, t is time, As is the exposed sur-
face of the part, h is the convection heat transfer coefficient, T∞ is the surrounding
temperature, k is thermal conductivity of the die material, Ac is the cross-sectional
area, xj are coordinates, εr is radiative emissivity for the part, σSB is the Stefan–
Boltzmann constant, V is the electric voltage, and I is the intensity of the current,
given by the product of the current density and cross-sectional area.
The goal of this subsection is to determine the significance of each of the three
heat transfer modes. To determine this, Eq. (5.24) was modified in the model to
represent the following heat transfer mode combinations and then compared to
experimental results:
• All heat transfer modes except radiation (Eq. 5.25)
• All heat transfer modes except radiation and convection (Eq. 5.26)
• All heat transfer modes except conduction (Eq. 5.27)

∂T ∂ 2T
ρVv Cp = −As [h(T − T∞ )] − 2kAc 2 + (1 − ξ )VI (5.25)
∂t ∂xj

∂T ∂ 2T
ρVv Cp = −2kAc 2 + (1 − ξ )VI (5.26)
∂t ∂xj

∂T  
ρVv Cp = −As [h(T − T∞ )] − As εr σSB T 4 − T∞
4
+ (1 − ξ )VI (5.27)
∂t
The resulting thermal profiles for all of the combinations are displayed in Fig. 5.18.
From the figure, the experimental temperature profile and the profiles neglecting

Fig. 5.18  EAF heat transfer

modes analysis [11]. Thermal
profiles for stationary w/out cond.
electrical tests were
calculated without particular Experiments
All H/T modes
heat transfer modes to w/out rad. and conv
identify the most significant
mode (conduction)
106 5  Compressive Electrically Assisted Forming Model

both convection and radiation are nearly identical. Conversely, the thermal profile
neglecting conduction heat transfer is nearly three times hotter at the maximum
temperature than the experimental thermal profile. The fact that there is a large
discrepancy between the experimental thermal profile and that without conduction
signifies that conduction heat transfer is a major contributor to the magnitude and
overall shape of the thermal profile of an EAF test. Furthermore, the fact that the
profiles neglecting radiation and convection are very similar to the experimental
profile signifies that the effect of these two heat transfer modes is minor. Overall,
when modeling EA forging, the all-inclusive heat transfer model could be simpli-
fied by neglecting both the radiation and convection heat transfer (in Eq. 5.26).
Note that this simplification is only valid for EA forging. Alternative EA processes,
such as EA stretch forming or EA bending, may require including the other forms
of heat transfer since the exposed surfaced area is typically greater in comparison
with the total volume of the workpiece. An accurate thermal profile can be gener-
ated by assuming only conduction heat transfer is present. For the results presented
here, Cp was assumed constant. As presented in an earlier section, the variability of
Cp with temperature may results in a difference, more significant for Ti.
Figure 5.19 displays a thermal image of the workpiece and the upper/lower dies
after electricity has been applied during an EAF test. The figure helps to highlight
the notable amount of heat that is transferred from the workpiece into the die by
way of conduction, as the dies act as thermal sinks in the system. The fact that
conduction is the primary heat transfer mode is important in manufacturing pro-
cess design because the size of the contact area between the dies and the work-
piece will dictate the magnitude of heat transfer into and out of your system.
Now that conduction was determined to be the primary heat transfer mode for
EAF compression, it is critical to verify any specific sensitivity. One variable that
can change the thermal profile of the workpiece is the die conduction length, or the

Top die

Cond. heat transfer

Specimen (heat gen.)

Cond. heat transfer

Bottom die

Fig. 5.19  EAF thermal camera conduction heat transfer image [11]. This image depicts how the
heat generated at the specimen center is conducted though the dies
5.4  Heat Transfer Modes Analysis 107

Fig. 5.20  Stainless Steel
304 die conduction length Ldc=0.0065m
sensitivity [11]. Different die
conduction lengths were used
in the model to determine the
significance and sensitivity
of the die conduction length
value Ldc=0.0085m

Ldc=0.0105m Exp.

length of the die that contributes toward drawing heat from the workpiece (Ldc).
Figure  5.20 displays workpiece thermal profiles for several different die lengths,
along with the actual experimental data for EAF tests run at a starting current den-
sity of 20 A/mm2 and a die speed of 12.7 mm/min. Currently, the authors use the
thermal images to estimate the die conduction lengths, where the excessive heating
was apparent when electricity was applied. The die conduction length used for the
model in this case was 0.0085 m. This was determined visually from the thermal
camera profiles of the EAF tests. From the figure, where the conduction lengths
are moved in increments of 2 mm, the overall thermal profile is extremely sensi-
tive to the die conduction length value. By varying the conduction length by about
0.004 m, there is about a 40 °C difference in the maximum temperatures within
the thermal plots. Overall, when modeling the electroplastic effect, it is critical to
devise a consistent and accurate method of approximating the conduction length.

5.5 EEC Profile—Material Sensitivity Comparison

In this subsection, the general shape of the EEC profiles for the Stainless Steel 304
and Ti-G5 specimens will be discussed. Figures 5.21 and 5.22 display the EEC
profiles for Ti-G2 and SS304, respectively (both tests run at a starting current den-
sity of 25 A/mm2 with a die speed of 12.7 mm/min) [11]. The EECs are plotted
here as a function of stroke. Other parameters could be used as well, e.g., strain
or displacement, but the overall observations are still valid. From the two figures,
it can be seen that the percentage of “usable” electricity increases throughout the
test; however, the increase is different for each material. In particular, the majority
of the EEC profile increase for the Ti-G2 material occurs in the initial part of the
stroke (about 50 % of the maximum EEC value occurs over about the first 5 % of
the compressive stroke), whereas the majority of the EEC profile increase for the
SS304 material occurs at the end of the stroke (about 50 % of the maximum EEC
value occurred in the final 25 % of the stroke). Both of these EEC profiles were
approximated using the power law, and the value of the exponential term is what
108 5  Compressive Electrically Assisted Forming Model

Fig. 5.21  Ti-G2 example 1
EEC profile (v = 12.7 mm/
min) [11]. Depicted is a 0.8
general shape of an EEC
profile for Ti-G2 at this 0.6

EEC
particular speed ξ = 0.46 · S%
0.24

0.4

0.2

0
0.00 20.00 40.00 60.00 80.00 100.00
% Stroke Completed

Fig. 5.22  Stainless Steel 0.6

304 example EEC profile
(v = 12.7 mm/min) [11]. 0.5
Depicted is a general shape 0.4
of an EEC profile for SS304
EEC

at this particular speed 0.3

2.2
ξ = 0.00075·S%
0.2

0.1

0
0.00 20.00 40.00 60.00 80.00 100.00
% Stroke Completed

determines the overall shape of the profile. In particular, if the exponential term
is less than one, the profile will be convex-shaped, and if the exponential term is
greater than one, the profile will be concave-shaped.
The concave/convex shapes of the profiles can also be explained visually by
examining the mechanical power profiles of the EAF tests run on each of the materi-
als. Figure 5.23 shows the mechanical power profiles of a conventional compression
test and an EAF compression test for the Ti-G2 material. The difference between
the baseline and EAF power profiles, at discrete points, becomes slightly smaller or
remains the same as the test progresses. This explains the profile in Fig. 5.21, where
minimal increase is seen toward the end of the test, when compared to the increases
at the beginning. Figure 5.24 displays the mechanical power profiles for EAF tests
on the SS304 material. Contrary to the Ti-G2 material, the difference in the baseline
and EAF power profiles increases as the tests progress. This leads to the profile in
Fig. 5.22, where the largest EEC increase is witnessed toward the end of the tests.
Two properties of the metals can be responsible for the difference in the shapes
of the EEC profiles: the effect of temperature increase on the strength of the
metal, and the effect of alloying. The strength of the titanium begins to immedi-
ate decrease as temperature is increased, whereas the strength of the stainless steel
does not show much of sensitivity to temperature until higher temperatures are
5.5  EEC Profile—Material Sensitivity Comparison 109

Fig. 5.23  Ti-G2 example
mechanical power profiles
[11]. Depicted are example
mechanical power profiles for
EAF tests on Ti-G2

Fig. 5.24  Stainless Steel 304

12
example mechanical power
profiles [11]. Depicted are
10 CD15
example mechanical power
profiles for EAF tests on CD10
SS304 8
Power (Nm/s)

Baseline
6
CD25

4
CD20

0
0 5 10 15 20
Time (s)

reached [14, 18]. The thermal softening could take place at a faster rate for the
titanium due to the temperatures reached during the EAF tests coupled with the
sensitivities of this metal’s strength with temperature. The amount by which each
metal was alloyed could also have an effect on the EEC profile. The SS304 had up
to 20 % chromium alloyed into the base iron metal. The chromium was weaker
than the iron base metal, and this allows for high ductility of the stainless steels.
Also, the titanium was more than twice as resistive as the stainless steel, which
contributed to the faster heating profile and higher EAF temperatures.

5.6 EAF Modeling—Sensitivities and Simplifications

Conclusions

This section analyzes the effects of specific heat values and various heat transfer
modes on the effectiveness of EAF [13]. Additionally, observations are provided
for the different EEC profiles for certain materials. Also, a frequency analysis was
110 5  Compressive Electrically Assisted Forming Model

performed on the force position data acquired from the testing machine. The fol-
lowing conclusions can be drawn from the research:
• The specific heat of a material must be considered when modeling EAF. The
effect that the specific heat has on the ability to model EAF is proportional to
the sensitivity of the specific heat value with respect to temperature.
• When modeling EA forging, the all-inclusive heat transfer model could be sim-
plified by neglecting both the radiation and convection heat transfer. This sim-
plification is only valid for EA forging. Alternative EA processes, such as EA
stretch forming or EA bending, may require including the other forms of heat
transfer since the exposed surfaced area is typically greater in comparison with
the total volume of the workpiece.
• Different metals produce different EEC profiles when the EAF technique
is used on them during a forging process. The differences include the overall
shape of the profile (the profile could be concave, convex, etc.); this shape deter-
mines the overall efficiency of the applied electrical power throughout the dura-
tion of the process.
• The frequency analysis indicated that there are specific frequencies dependent
on the magnitude of the current density applied. The frequency increases with
the current density, thus with energy input. This may be an indication of a cyclic
softening/hardening phenomenon present during EAF.
• The frequency observed for titanium is higher than that for SS304. This could
be due to any differences in the crystalline structure or specific alloying compo-
nents within the metals.
• The quality of the electrical power output during an EAF process can vary
depending on the capabilities of the power source. This can also have an impact
on the mechanical profiles of the same parts undergoing the same EAF process.

References

1. Bunget C, Salandro WA, Mears L, Roth JT (2010) Energy-based modeling of an electri-

cally-assisted forging process. Transactions of the North American Manufacturing Research
Institute of SME, 38
2. Lange K (1985) Handbook of metal forming. McGraw-Hill, New York
3. Backofen WA (1972) Deformation processing. Addison-Wesley Educational Publishers Inc,
Boston
4. Kobayashi S, Oh S-I, Altan T (1989) Metal forming and the finite-element method. Oxford
University Press, New York
5. Wagoner RH, Chenot JL (1996) Fundamentals of metal forming. Wiley, Hoboken
6. Perkins TA, Kronenberger TJ, Roth JT (2007) Metallic forging using electrical flow as an
alternative to warm/hot working. J Manuf Sci Eng 129(1):84–94
7. Avitzur B (1980) Metal forming: the application of limit analysis. Dekker, New York
8. Cengel YA (2007) Heat and mass transfer: a practical approach, 3rd edn. McGraw-Hill
9. Cengel YA, Boles MA (2008) Thermodynamics: an engineering approach, 6th edn.
McGraw-Hill
10. Ross CD, Irvin DB, Roth JT (2007) Manufacturing aspects relating to the effects of DC
­current on the tensile properties of metals. J Eng Mater Technol 129(2):342–347
References 111

11. Bunget C, Salandro W, Mears L (2012) Sensitivities when modeling electrically-assisted

forming. In: International manufacturing science and engineering conference, MSEC2012-
7334, p 9
12. Salandro WA, Bunget C, Mears L (2011) Thermo-mechanical investigations of the elec-
troplastic effect. In: International manufacturing science and engineering conference,
MSEC2011-50250, p 10
13. Salandro WA, Bunget C, Mears L (2012) Modeling the electroplastic effect during electri-
cally-assisted forming of 304 stainless steel. In: International manufacturing science and
engineering conference, MSEC2012-7241, p 10
14. Allegheny Ludlum Corp., product data sheet (304SS)
15. Kaufman JG (ed) (1999) Properties of aluminum alloys, Pub. ASM and the Aluminum
Association of America (Al6061)
16. United States Department of Defense (2003) Department of defense handbook: metallic
material and elements for aerospace vehicle structures, MIL-HDBK-5J
17. Salandro WA, Bunget C, Mears L (2011) A thermal-based approach for determining electro-
plastic characteristics. In: Proceedings of the institution of mechanical engineers, part B. J
Eng Manuf (5 Jan 2012). doi:10.1177/0954405411424696
18. Dedman HE, Wheelahan EJ, Kattus JR (1950) WADC technical report 58-440, part 1 (Ti-G5)
Chapter 6
Tensile Electroforming Model

This chapter contains modeling work for sheet metal subject to an applied direct
electrical current in uniaxial tension. The modeling is divided into a thermal and
mechanical approach which is combined to a thermo-mechanical model to pre-
dict the deformation behavior during EAF. Additional experimental testing is per-
formed to validate the introduced models and to study varying factors for EAF
of sheet metals in uniaxial tension. These tensile testing results are provided in
Appendix B along with other relevant mechanical tests. This chapter examines the
thermo-mechanical response of Mg AZ31 as a result of its industrial use, low den-
sity, and overall low formability in common sheet-forming processes. However,
the models and testing presented are still directly applicable to other materials.

6.1 Thermal Modeling

One aspect of understanding how an applied direct electrical current influences the
flow stress and deformation behavior is to determine the thermal profile response
and resistive heating during the process. Thus, the objective of this section is to
provide a model of the thermal response of sheet metals subject to EAF. The pre-
sented model predicts the thermal behavior of the sheet metal under the assump-
tion that all of the input electrical energy is converted to heat generation through
Joule heating. As a result, the response can be compared to experimental EAF
results to determine whether a portion of the applied energy aided in deformation.
The work presented in this chapter varies from other thermal research in EAF
as it incorporates geometrically larger specimens which have large thermal gradi-
ents present. Therefore, this work cannot consider a lumped mass approach in the
analysis of the component subject to EAF.

© Springer International Publishing Switzerland 2015 113

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_6
114 6  Tensile Electroforming Model

6.1.1 Model Development

In the following two sections, the development of the thermal stationary model for
process variable identification and the thermal deformation model for verification
and understanding of the electroplastic effect are presented.

6.1.1.1 Stationary Model

The stationary model was created for a stationary test (i.e., no deformation) of
an ASTM tensile specimen subjected to an applied direct electrical current. This
model is used for process variable identification and overall modeling methodol-
ogy and construction validation. The key process variables to be identified were
the heat transfer convection coefficient, initial component and clamping die tem-
peratures, the power supply efficiency and its associated losses, and effective
clamping die conduction length in the clamped region. As this model is the basis
for the subsequent thermal deformation model, it is described in detail.
The constructed 1D transient finite-difference thermal model is of a standard
12.5-mm-wide ASTM tensile specimen [1] that is to be used for the uniaxial test-
ing of EAF, and the model contains equally spaced nodes along the length of the
specimen (Fig. 6.1).
The model accounts for:
1. Heat conduction throughout the specimen
2. Heat conduction to the dies in the clamping region
3. Heat convection to the environment in the testing region
4. Joule heating of the specimen and dies when the direct electrical current is applied
(corresponds to varying wave shapes, magnitudes of current, and duty cycles)
5. Temperature-dependent material properties of the sheet material (density, elec-
trical resistivity, thermal conductivity, and heat capacity) [2]
The model is based on the following assumptions:
1. The temperatures across the width and thickness of the specimen are uniform,
and the temperature only varies along the length of the specimen (i.e., it is

Fig. 6.1  Stationary thermal model schematic

6.1  Thermal Modeling 115

assumed that the thermal gradient in the thickness and width direction is rela-
tively small as compared to the gradient along the length of the specimen).
2. The material is homogeneous and isotropic.
3. The geometric model does not incorporate the fillets which link the testing
region to the clamping region of the specimen.
4. The clamping dies are analyzed as a lumped mass to conserve the 1D nature of
the model.
5. The radiation heat transfer is incorporated into the convection coefficient (i.e.,
hcombined and it is just denoted as h in the below equations for simplicity).
6. The electrical resistivity and specific heat of the clamping dies are not tempera-
ture dependent (there is not a large temperature change of the clamping die so
this is an accurate assumption).

The general expression to characterize the balance of energy for the system given
in terms of power is as follows:
 ESystem
Q̇ + Ėgeneration = (6.1)
t
All Sides

where Q̇ is the respective rate of heat transfer into all of the system sides depend-
ing on the boundary conditions (i.e., conduction or convection), Ėgeneration is the
heat generated within the specimen from resistive heating, ESystem is the change
of internal energy associated with the system, and Δt is the change in time.
Considering only one element,
 EElement
Q̇ + Ėgen,element = (6.2)
t
Element Sides

In constructing the model, there are three separate areas which had varying bound-
ary conditions. These include the interior nodes in contact with the clamping die
interface, the two nodes at each end of the specimen, and the exterior nodes in the
testing region exposed to the environment.
For an interior node in contact with the clamping die interface, the power bal-
ance can be written as follows:
 i   i    i 
kA1 Tm−1 − Tmi kA1 Tm+1 − Tmi ka2 A2 Tdie − Tmi
+ +2
�x �x Ldie
(6.3)
Tm − Tmi
i+1
+ ėgen,clamp A1 �x = ρA1 �xc
�t
where k is the thermal conductivity of the sheet, A1 is the element conduction
area of the sheet in the clamping die region, x is the nodal spacing, Tm−1
i is the
present temperature at the node to the left of the node being analyzed, Tmi is the
temperature of the node being analyzed, Tm+1i is the present temperature at the
node to the right of the node being analyzed, ka2 is the thermal conductivity of
116 6  Tensile Electroforming Model

the clamping die made from A2 Steel, A2 is the element conduction area into the
i is the present temperature of the clamping die, L
dies, Tdie die is the die conduction
length, ėgen,clamp is the heat generation per unit volume for the sheet in the clamp-
ing region, ρ is the density of the sheet metal, c is the heat capacity of the sheet,
Tmi+1 − Tmi is the temperature change of the node being analyzed from the present
time to the future time, and t is the time step.
For the two nodes at each end of the specimen,
 i    i 
kA1 Tm+1 − Tmi ka2 �x w T − T i  
die m
+2 2
+ htw T∞ − Tmi
�x Ldie
  �x �x Tmi+1 − Tmi
+ htw T∞ − Tmi + ėgen,clamp A1 = ρA1 c
2 2 �t
(6.4)

where w is the sheet width in the clamping region, h is the convection coefficient, t
is the sheet thickness, and T∞ is the atmospheric temperature.
For a node in the testing region exposed to the environment,
 i   i    
kA11 Tm−1 − Tmi kA11 Tm+1 − Tmi
+ + 2 hA22 T∞ − Tmi
�x �x
(6.5)
T i+1 − Tmi
+ ėgen,test A11 �x = ρA1 �xc m
�t
where A11 is the element conduction area of the sheet in the testing region, A22
is the element convection area in the test region, and ėgen,test is the heat genera-
tion per unit volume for the sheet in the testing region. And the material properties
were updated at each time step as a function of temperature,
 
k, ka2 , ρ, c = f Tmi (6.6)

in all the above equations [2].

Using an explicit solution approach, Eqs. (6.3)–(6.5) can be solved to deter-
mine the new nodal temperature after a given time step.
Thus, for an interior node in contact with the die interface,

i
kTm−1 i
i+1 i 2ka2 A2 Tdie
Tm = Tm + �t +
�x 2 ρc Ldie ρA1 �xc
   (6.7)
kT i
2k 2ka2 A2 i m+1 ėgen,clamp
− + T + +
�x 2 ρc Ldie ρA1 �xc m �x 2 ρc ρc

For the nodes at each end of the specimen,

6.1  Thermal Modeling 117

  
i
kTm+1 k 2ka2 w htw
Tmi+1 = Tmi+ �t − + + Tmi
�x 2
ρc �x 2
ρc Ldie ρA1 c ρA1 �x
2 c
2 2
 (6.8)
i
2ka2 wTdie ėgen,clamp
htwT∞
+ + �x
+
Ldie ρA1 c ρA1 2 c ρc

For the nodes in the testing region,

  
i
kTm−1
i+1 i 2k 2hA22
Tm = Tm + �t − + Ti
�x 2 ρc �x 2 ρc ρA11 �xc m
 (6.9)
i
kTm+1 2hA22 T∞ ėgen,test
+ 2 + +
�x ρc ρA11 �xc ρc

As the electrical current will be passing through the die and heat will be trans-
ferred from the sheet metal to the die, the die temperature will also change as a
function of time. To conserve the 1D nature of this analysis, the die is considered
to be a lumped mass with a uniform temperature. This is an accurate assumption
as the Biot number is less than 0.1 for the die geometry and heat transfer proper-
ties [3].
Thus, the power balance for the dies is given by
 
  2ka2 A3 Tavg,mg i
clamp − T i
die
i
hAs,die T∞ − Tdie +
Ldie (6.10)
i+1 i
Tdie − Tdie
+ ėgen,die Vdie = ρa2 Vdie ca2
�t
where As,die is the die surface area, Tdie
i is the present die temperature, A is the full
3
conduction area between the sheet and the die, Tavg, i
mg clamp is the average temper-
ature at the clamping region for the sheet, Vdie is the volume of the die, ρa2 is the
density of the die material (A2 Steel) which is a function of temperature, ca2 is the
i+1
heat capacity of the die material, and Tdie i is the temperature change of the
− Tdie
die from the present time to the future time.
Additionally, the average temperature at the clamping region for the sheet is
defined by

1 i
b
i
Tavg,mg clamp = Tm,clamp region,b (6.11)
b
1

where b is the number of nodes in the clamping region.

Using an explicit solution approach, Eq. (6.10) can be written as follows:
118 6  Tensile Electroforming Model

 i
i+1 i hAs,die T∞ 2ka2 A3 Tavg,mg clamp
Tdie = Tdie + �t +
ρa2 Vdie ca2 ρa2 Vdie ca2 Ldie
   (6.12)
hAs,die 2ka2 A3 i ėgen,die
− + Tdie +
ρa2 Vdie ca2 ρa2 Vdie ca2 Ldie ρa2 ca2
Thus, Eqs. (6.7)–(6.9) and (6.12) are the generic nodal solutions which constitute
the thermal model.
To characterize the resistive heating, Joule’s first law is used:

Ėgen = ėgen V (6.13)

Thus,
Ėgen I 2R I 2R
ėgen = = = (6.14)
V V Lwt
where V is the volume of interest, I is the current in amps, R is the resistance in
ohms, and L, w, and t are the length, width, and thickness of interest, respectively.
For a rectangular cuboid geometry, the resistance can be defined by
L L
R = ρe = ρe (6.15)
A wt
where ρe is the electrical resistivity of the material of interest and A is the area of
interest.
Thus, by substituting Eq. (6.15) into Eq. (6.14),
I 2 ρe
ėgen = (6.16)
w2 t 2
Then, using Eq. (6.16) along with the corresponding electrical resistivity, geomet-
ric width, and geometric thickness, the magnitude of the heat generated per unit
volume can be determined for each element. Equation (6.16) is therefore used to
calculate the magnitude of the heat generation per unit volume of each element in
the sheet metal in the clamping region, the testing region, as well as the die using
the appropriate dimensions and material properties. Applying this magnitude with
varying wave shapes and duty cycles represents the present testing methods for
EAF in tensile applications. The solution schematic for solving the model is given
in Fig. 6.2.

6.1.1.2 Deformation Model

A model was created for a deformation test subject to an applied direct electrical
current, and this can used to determine whether a portion of the energy from the
applied electric current went directly into aiding deformation and not to Joule heat-
ing. For the deformation modeling, two variants were created. The first considers
6.1  Thermal Modeling 119

Fig. 6.2  Stationary model solution schematic [4]

the deformation to be uniform; therefore, the elements in the test region are equal
in size and change shape as deformation is imposed. The second accounts for dif-
fuse necking (i.e., non-uniform deformation) during forming, and in this section,
the diffuse neck is predicted using circle grid analysis (CGA) results obtained from
experimentation. The diffuse model was created as in prior EAF testing, a diffuse
neck was found to be apparent during uniaxial tension [5, 6].
For the case with uniform deformation, there are several variables that do not
remain constant as in the case of stationary testing. The factors that had to be
accounted for were the shape change of the elements in the testing region of the
sheet and the heat generation per unit volume as a result of specimen shape change
[note: it is a function of sheet width and thickness as described in Eq. (6.16)].
The deformation in the length direction of the specimen (�s) can be calculated
as follows:
s = tṡ (6.17)
where t is the present time in the test and ṡ is the platen speed.
120 6  Tensile Electroforming Model

Therefore, as there are a fixed number of elements (Elementstest) in the test

region, the displacement of each element can be determined by
s
sElements = (6.18)
Elementstest
The strain in the elements in the length direction can be calculated using
 
�s + Linitial,test
εL = ln (6.19)
Linitial,test

where Linitial,test is the initial length of the entire test region and when assuming
isotropy, the width and thickness strain is
εw = εt = −0.5εL (6.20)
The instantaneous size of the elements is determined using
xuniform = sElements + x (6.21)

tuniform = teεt (6.22)

wuniform = wtest eεw (6.23)

where xuniform is the length of the element, tuniform is the thickness of the sheet
under uniform deformation, t is the initial thickness of the sheet, wuniform is the
width of the sheet under uniform deformation, and wtest is the initial width of the
sheet in the testing region. As a result of the uniform deformation assumption, all
the elements have equal lengths, widths, and thicknesses. Additionally, the con-
duction areas and convection areas are determined from the new element sizes by
x
A11_uniform = A11 (6.24)
xuniform

A22_uniform = xuniform wuniform (6.25)

where A11_uniform is the new conduction area in the test region of the sheet and
A22_uniform is the new convection area in the sheet for each element.
Last, the heat generated from the applied current now varies as a function of
deformation as the resistance of the sheet increases with elongation. The heat gen-
erated can now be given as follows:

I 2 Runiform
ėgen_uniform =   (6.26)
tuniform wuniform s + Linitial,test

where
 
�s + Linitial,test
Runiform = ρe (6.27)
tuniform wuniform
6.1  Thermal Modeling 121

Thus,

I 2 ρe
ėgen_uniform = 2 2 (6.28)
tuniform wuniform

For the case with diffuse deformation, the model considers that the length, width,
and thickness of the elements in the testing region are non-uniform. The present
element size is calculated using experimental strain measurements at failure which
were assumed to be a linear function from zero strain to the fracture strain. Thus,
the length of the element and width of the element can be calculated using
�xm,diffuse = �xeεm,L (6.29)

wm,diffuse = weεm,w (6.30)

where εm,L and εm,w are, respectively, the length and width strains for each
element.
The strain in the thickness direction is calculated by
 
εm,t = − εm,L + εm,w (6.31)
Thus, the sheet thickness is given as follows:
tm,diffuse = teεm,t (6.32)
The heat transfer dimensions of the model can be calculated for each element as
follows:
A11,m,diffuse = tm,diffuse wm,diffuse (6.33)

A22,m,diffuse = xm,diffuse wm,diffuse (6.34)

where A11,m,diffuse is the new conduction area in an element and A22,m,diffuse is the
new convection area for an element.
Finally, the heat generated from the applied current now varies as a function of
the element size as the resistance of the element increases with elongation. The
heat generated can now be given as follows:

I 2 Rm,diffuse
ėm,gen_diffuse = (6.35)
tm,diffuse wm,diffuse xm,diffuse

where
�xm,diffuse
Rm,diffuse = ρe (6.36)
tm,diffuse wm,diffuse
Thus,
I 2 ρe
ėm,gen_diffuse = 2 2
tm,diffuse wm,diffuse (6.37)
122 6  Tensile Electroforming Model

For determination of the strain present in each element at a certain time (for dif-
fuse model), experimental analysis of the local strains at failure (summarized in
Table  6.1 for the center of the specimen) was linearized from time zero to the
time at which the fracture occurred. Figure 6.3 (left) shows the averaged strain in
the length, width, and thickness directions from the experimental testing results
at failure for Parameter Set 4, which is defined in Table 6.2 of the subsequent
“Experimental Setup” section. In Fig. 6.3 (right), the corresponding strain input
for the length strain is depicted as an example.

Table 6.1  Fracture strains for specimen center

Center of specimen Final length strain Final width strain Final thickness strain
Parameter set 1 0.35 −0.27 −0.08
Parameter set 2 0.43 −0.32 −0.11
Parameter set 3 0.30 −0.22 −0.08
Parameter set 4 0.39 −0.26 −0.13

Fig. 6.3  Linearized experimental strain data for parameter set 4 at failure (left) and correspond-
ing input length strain surface (right) [4]

Table 6.2  Tensile testing conditions

Parameter Current Initial current Pulse Pulse Duty Wave
set magnitude (A) density (A/mm2) duration (s) period (s) cycle (%) shape
0 0 0 n/a n/a n/a n/a
1 800 64 0.3 60 0.50 Square
2 800 64 0.5 60 0.83 Square
3 500 40 0.5 60 0.83 Square
4 500 40 1.0 60 1.67 Square
6.1  Thermal Modeling 123

6.1.2 Experimental Setup

To validate and examine the results from the derived models, experimental tests
were carried out with varying test conditions using a square wave input as this
will create transient thermal periods for robust model validation. The testing con-
ditions are listed in Table 6.2. As can be seen, two current densities were examined
(increased during test), three pulse durations were used, and the pulse period was
held constant. For the conditions in Table 6.2 (Parameter Sets 1–4), both stationary
(i.e., no deformation) and deformation tests with a platen speed of 2.54 mm/min
(corresponds to an initial true strain rate of 0.0004 s−1) were performed with two
replications. As a result of the repeatability of the test results (force/thermal), this
replication number was deemed sufficient. The maximum thermal difference was
less than 3 °C for all thermal tests, and the average force difference was 5 MPa.
Parameter Set 0 represents conventional room temperature forming.
The experimental setup of the testing is shown in Fig. 6.4. An Instron hydrau-
lic testing machine with specialized tensile grips to isolate the electricity from the
testing equipment was used to deform the tensile specimens. For the mechanical
response, force and displacement (resolution of 0.0254 mm) were collected and
this allowed for the mechanical strain and stress to be calculated. The tensile
specimens started as 20 × 200 mm sheet strips and were prepared according to
ASTM Standard B557M [1]. A thin layer of ceramic paint was applied on one side
to reduce emissivity issues for thermal response measurements, while the other
side of the specimen was acid-etched with a strain grid for (CGA). The material
tested in this study was Mg AZ31B warm-rolled sheet, and the test region had a
cross section of 1 mm thick by 12.5 mm wide per ASTM Standard B557M [1].
To measure the thermal response during the test, a FLIR A40M thermal camera
(upper temperature capacity 550 °C, temperature resolution 0.1 °C, and frame rate
12.5/s) was used (not shown in Fig. 6.4).

Fig. 6.4  Experimental testing setup [4]

124 6  Tensile Electroforming Model

Fig. 6.5  Stationary test
thermal sequence over one
period for parameter set 4 [4]

An example thermal response of a stationary test over one period is shown in

Fig. 6.5 (settings correspond to Parameter Set 4). For this test case, the maximum
temperature is reached at 1 s as this is where the applied current is discontinued and
the remaining three profiles (20, 40, and 60 s) display the cooling of the tensile spec-
imen. Also by observation, the thermal gradient in the width direction is small as
compared to the length direction (less than 3 °C), and it is assumed that the thickness
gradient is even smaller; therefore, the assumption of a 1D model is sufficient.

6.1.3 Results and Discussion

In the following section, the stationary and deformation models (uniform and
diffuse) are compared to the experimental results obtained from the experimen-
tal testing. Additionally, the diffuse thermal model sensitivity to the experimental
6.1  Thermal Modeling 125

strain input is examined. For experimental results, the average of the test replica-
tions is displayed.
It should be noted that when using the explicit solution approach, the time step
should be chosen carefully as the solution is not unconditionally stable. Hence, if
the time step is not sufficiently small, the solutions may oscillate uncontrollably or
diverge from the actual solution. To choose a time step, a stability criterion can be
established based on the second law of thermodynamics. For this, the coefficients
of the prior time step must be greater or equal to zero for all nodes when explicitly
solved for the next time step. As a result of different nodes having varying bound-
ary conditions, the most restrictive time step should be used. For this analysis, the
most restrictive condition was the nodes in the test region of the specimen. Taking
into consideration the thermal conductivity, density, specific heat, conduction/
convection coefficients, and the mesh properties, the upper limit on the time step
was 0.30 s for the uniform deformation case. In this analysis, a time step of 0.01 s
was used and there were no issues with the solution stability. For the deformation
cases, the same time step of 0.01 s was used, and from analysis of the results, there
was maintained stability of the solution and the second law of thermodynamics
was not violated during analysis of the data. Therefore, even with fast temperature
changes and changing element shapes, the time step used in these simulations was
deemed appropriate.

6.1.3.1 Thermal Model Comparison

For the stationary model, one parameter set was used to define/establish the key
process variables (heat transfer convection coefficient, initial component and die
temperatures, the power supply efficiency and its associated losses, and effective
die conduction length in the clamped region) and to validate the overall model
construction. Once the variable identification was complete, the key process vari-
ables were held constant for the remaining simulations for both the stationary
and deformation models. Hence, when comparing the stationary models to the
experimental results, the only variable input was the electrical input parameters
(Table  6.2). In the following results, Parameter Set 4 was chosen to display in
detail the model results as compared to experimental data; all the developed trends
were similar for all of the testing conditions.
To evaluate the stationary model, there are two major criteria to be assessed.
These include the maximum temperature observed with respect to test time and
the distribution of the temperature along the length of the specimen as a function
of time. Figure 6.6 displays the maximum temperature profile for the stationary
experimental and model results.
As can be seen, there is agreement with the experimental data and the model
such that the maximum error is 3 °C. As shown, the temperature rises quickly dur-
ing the current application and cools fairly fast as the specimen acts like a large fin
(illustrated in Fig. 6.5).
126 6  Tensile Electroforming Model

Fig. 6.6  Stationary
maximum temperature
response of experimental and
model results for parameter
set 4 [4]

Fig. 6.7  Stationary axial
length temperature profile
of experimental and model
results for parameter set 4 [4]

Another metric to consider in the successful prediction of the stationary

response is the temperature along the length of the specimen. Figure 6.7 shows
the temperature along the length of the specimen for varying times during one
pulse sequence. As can be seen, there is a small variation in the model result at
1 s where the model overpredicts the experimental data. Also, in the fillet regions
of the specimen, there is a larger thermal profile difference in the experimental
and model results. This can be explained from the model assumption that the fillet
region is not included in the model, but may have had an effect on the temperature
distribution (see Fig. 6.5 near die interface). Thus, as the fillet region has a larger
cross section than assumed, the model predicts this region to have a slightly higher
heat generation and thus a higher temperature.
6.1  Thermal Modeling 127

Fig. 6.8  Stationary maximum temperature response of experimental and model results for

remaining parameter sets [4]

The remaining stationary maximum temperature profiles for the other param-
eter sets are summarized in Fig. 6.8 to display the validity of the established pro-
cess variables that were held constant while only modifying the electrical input
parameters. As can be seen in the figure, the model was capable of matching the
experimentally observed thermal response (maximum error is less than 2 °C).
Also, the maximum temperatures reached can be observed for the varying electri-
cal conditions, and the higher current with the middle pulse duration produced the
largest temperature.
Upon validation of the stationary model, the thermal response for the case with
deformation was examined. The maximum thermal profile for an experimental sta-
tionary and deformation test is displayed in Fig. 6.9. As can be seen, the stationary
curve maintains a constant maximum temperature, whereas the deformation curve
response increases as time progresses. This increase in the maximum temperature
is a result of the elongation and shrinking cross-sectional area which modifies the
resistance of the sheet and thus the heat generation per unit volume. Additionally,
with deformation, the heat transfer conduction area decreases and convection
area increases which influence the cooling during and after the applied electrical
current.
The models created to account for deformation are compared to the experimen-
tal results in Fig. 6.10 for Parameter Set 4. As can be seen, the uniform model
underpredicts the thermal response as a result of the assumption that only uniform
deformation occurs during the process even though it assumes that all of the input
128 6  Tensile Electroforming Model

Fig. 6.9  Experimental
maximum temperature
response of stationary and
deformation results for
parameter set 4 [4]

Fig. 6.10  Maximum
temperature comparison
of deformation models to
experimental results for
parameter set 4 [4]

electrical energy goes into resistive heating. However, in EAF uniaxial testing, dif-
fuse necking is present as a result of the specimen geometry and the high cooling
rate into the die regions, which leaves the center of the specimen at higher elevated
temperatures for a longer duration. The diffuse model which assumes that all of
the input electrical energy goes into resistive heat and which uses the actual strain
from experimental testing overpredicts the thermal response of the experimental
data. This overprediction of temperature is suggestive that some energy of the
applied electrical current may have gone directly into aiding in deformation (i.e.,
electroplastic effect) and not toward Joule heating.
The experimental deformation model response over the length of the speci-
men is displayed in Fig. 6.11 for the sixth pulse and the subsequent cooling
period before the seventh pulse. The results in this figure more clearly portray the
6.1  Thermal Modeling 129

Fig. 6.11  Axial comparison of deformation model to experimental results for parameter set 4 [4]

Fig. 6.12  Thermal response surface for deformation models and experimental data as a function
of time [4]

overprediction of the diffuse deformation model where the variation in the thermal
response would represent the applied electrical energy that went directly into aid-
ing deformation and not Joule heating.
Figure  6.12 depicts the thermal response of the deformation models and
experimental data over the entire data set to see the complete response shape. As
observed, the uniform model’s overall temperature increases with time; however,
it does not show an increased temperature at the specimen center as the diffuse
model and the experimental data.
130 6  Tensile Electroforming Model

6.1.3.2 Thermal Model Sensitivity

To examine the sensitivity of the experimental strain measurements from the CGA
to the output temperature of the diffuse deformation model, several model simula-
tions were performed where the original experimental strain data were mathemati-
cally transformed. To transform the strain data, it was reduced by multiplying it by
a percentage. This was performed to verify that the difference from the all-heating
thermal model was not greater or equal when compared to the experimental data
as a result of strain input sensitivity. To compare the results, the center tempera-
ture of the specimen is examined (corresponds to the maximum temperature of the
specimen). For the analysis, Parameter Set 4 was examined to show the general
trends. The simulation runs performed are given in Table 6.3.
An example of the decrease in strain is shown in Fig. 6.13 where the maximum
40 % decrease is depicted.

Table 6.3  Strain sensitivity runs performed

Percentage of decrease in strain Percentage of original experi-
(%) mental data (%)
Diffuse deformation model 0 100
Diffuse deformation model 10 90
(10 %)
Diffuse deformation model 20 80
(20 %)
Diffuse deformation model 40 60
(40 %)

Fig. 6.13  Original and modified (40 % decrease) linearized experimental strain data for param-
eter set 4 at failure [4]
6.1  Thermal Modeling 131

Fig. 6.14  Strain input
sensitivity for diffuse
deformation model versus
experimental results at first
application of current (60 s)
[4]

The following figures show an enlarged portion of the thermal response at sev-
eral times during the test to depict the varying thermal responses from the model
results as compared to the experimental data. Figure 6.14 shows the different model
responses and experimental results during the first application of electricity. As
seen with a decreasing strain input to the diffuse model, this correlates with a lower
thermal profile. This is expected as the cross-sectional area is greater and it has an
inverse correlation with the heat generated within the specimen [i.e., larger cross-
sectional area means less heat generation as described in Eq. (6.16)]. As shown, the
experimental results are still less than all of the models, and this is a result of the
deformation still being quite uniform for the experimental test at this point.
In Fig. 6.15, the same results are seen; however, the experimental results are equiva-
lent to the diffuse deformation model at a 40 % decrease. This is a result of the experi-
mental test having a greater amount of diffuse deformation. For the experimental test,
the deformation along the axial length is diffuse and may not be linearly distributed
from one end of the specimen to the center or at the center point. Thus, the peak tem-
perature may not increase linearly. However, for the model, this is an assumption that
was used when taking the fracture strain measurements from the CGA.
Similar results are presented in Fig. 6.16, but the experimental result is now at
the diffuse deformation model at 20 % reduction in strain.
The overall conclusion from this analysis is that the sensitivity of the strain
input is not the reason for seeing a larger thermal response from the model that
considers all of the electrical energy is transformed into joule heating. This is
since at the end of the deformation test (i.e., experimental result is near failure),
the experimental result is only at the diffuse deformation model at 20 % reduction
in strain which is a large variation from the original experimental measured strain
data. Thus, this confirms that the observed variation in the thermal responses from
the diffuse deformation model to the experimental result is due to energy going
directly into aiding in deformation (i.e., electroplasticity).
132 6  Tensile Electroforming Model

Fig. 6.15  Strain input
sensitivity for diffuse
deformation model versus
experimental results at fifth
application of current (300 s)
[4]

Fig. 6.16  Strain input
sensitivity for diffuse
deformation model versus
experimental results at
seventh application of current
(420 s) [4]

6.1.4 Thermal Model Conclusions

The thermal response is an important aspect to consider during EAF as this is a cou-
pled thermal–mechanical process. As a result, this section examined and successfully
modeled the thermal response of sheet metal subject to EAF and established mod-
els that can predict the response for stationary tests as well as deformation tests for
varying electrical testing conditions. It was shown that there is good agreement with
the model as compared to the stationary response with an applied electrical current.
However, for consideration of deformation during the process, the diffuse model sug-
gests that a portion of the applied current goes directly into aiding deformation and
not into Joule heating. Overall, the output of this section is the creation of accurate
thermal response models which can be coupled with mechanical modeling for EAF.
6.2  Mechanical Modeling 133

6.2 Mechanical Modeling

This section introduces the mechanical modeling portion of the thermo-mechanical

model for electrically assisted forming in uniaxial tension. The results presented in
the previous thermal modeling section are used as an input to the model to pro-
vide the thermal variation axially along the specimen and with respect to time. The
output of this modeling portion is the local material strain, stress, and force of the
sheet metal.
The model considers both a structural and geometric non-homogeneity as a
result of a non-uniform temperature distribution (structurally related) and diffuse
necking (geometrically related) observed during EAF testing. A schematic is pre-
sented in Fig. 6.17 where the tensile sample is divided into a number of elements,
each capable of having different strength properties and dimensional sizes. The
basis for the model is that force equivalency must be maintained throughout the
tensile sample, but the stress and area can vary along the length of the specimen.
Also, the model is solved incrementally, where the input is a given displacement
(d) which in turn creates plastic deformation. As the displacement is imposed, the

Fig. 6.17  Deformation
model schematic
134 6  Tensile Electroforming Model

other continuity condition is that the summation of the individual element dis-
placements adds to the total imposed displacement.

6.2.1 Deformation/Strength Model Derivation

This section contains the derivation of the deformation/strength model. The deri-
vation is written for a time step (i). Thus, at a given time step i, force equivalency
gives
F = Fm = Fm+1 for m = {1 . . . (m − 1)} (6.38)
where F is the force and m is the number of nodes/elements in the model.
The force can be written in terms of stress and area by
σm Am = σm+1 Am+1 (6.39)
or
σm tm wm = σm+1 tm+1 wm+1 (6.40)
where t and w are the thickness and width, respectively.
Knowing that
t = to eεt (6.41)

w = wo eεw (6.42)
where to is the initial thickness, εt is the thickness strain, wo is the initial width, and
εw is the width strain. This yields
σm to,m wo,m eεt,m eεw,m = σm+1 to,m+1 wo,m+1 eεt,m+1 eεw,m+1 (6.43)
or
σm to,m wo,m eεt,m +εw,m = σm+1 to,m+1 wo,m+1 eεt,m+1 +εw,m+1 (6.44)
Since volume is conserved,
εt + εw + εL = 0 or εt + εw = −εL (6.45)
where εL is the incremental strain in the length direction (i.e., along axial length).
Incremental strain is the strain in each individual element over one time step. The
incremental strain accrues over time to the accumulative strain of each element.
This gives
σm to,m wo,m e−εL,m = σm+1 to,m+1 wo,m+1 e−εL,m+1 (6.46)
or
σm Ao,m e−εL,m = σm+1 Ao,m+1 e−εL,m+1 (6.47)
6.2  Mechanical Modeling 135

Manipulating
σm Ao,m e−εL,m+1
= −ε (6.48)
σm+1 Ao,m+1 e L,m

σm Ao,m
= e(εL,m −εL,m+1 ) (6.49)
σm+1 Ao,m+1
   
ln σm Ao,m − ln σm+1 Ao,m+1 − εL,m + εL,m+1 = 0 (6.50)
where the stress (σ) is defined by a modified power law relation that is a function
of temperature and strain for Mg AZ31:
  nm
σm εL,m,total , Tm = Km εL,m,total eεL,m,total sm (6.51)

  nm+1
σm+1 εL,m+1,total , Tm+1 = Km+1 εL,m+1,total eεL,m+1,total sm+1 (6.52)

and

i−1
εL,m,total = εL,m,acc + εL,m (6.53)

i−1
εL,m+1,total = εL,m+1,acc + εL,m+1 (6.54)

where εL,m,total and εL,m+1,total are the total accumulative strain in element m and
i−1 i−1
m + 1 developed during forming at time step i, respectively. εL,m,acc and εL,m+1,acc
represent the total accumulative strain in element m and m + 1 from the start of
forming to the prior time step (i − 1). εL,m and εL,m+1 are the incremental strains
occurring during time step i for elements m and m + 1, respectively.
Such that K, n, and s are as follows:

 T ≤ 25 ◦ C : K = KRT
K := 25 ◦ C < T < 150 ◦ C : K = K1 T + K2 (6.55)

T ≥ 150 ◦ C : K = K3 exp (K4 T )


 T ≤ 25 ◦ C : n = nRT
n := 25 ◦ C < T < 150 ◦ C : n = n1 T + n2 (6.56)

T ≥ 150 ◦ C : n = n3 T + n4


 T < 120 ◦ C: s=0
s := 120 ◦ C ≤ T ≤ 327 ◦ C : s = s1 T 2 + s2 T + s3 (6.57)

T > 327 ◦ C : s=0
where the coefficients are given in Table 6.4.
136 6  Tensile Electroforming Model

Table 6.4  Conventional Constant Value

room and elevated
KRT 457.72
temperature model
coefficients for Mg AZ31B K1 −1.9529
K2 500.68
K3 931.06
K4 −0.01
nRT 0.1818
n1 −0.0004
n2 0.1909
n3 −0.0009
n4 0.2713
s1 0.00008
s2 −0.0357
s3 3.1162

Thus,
   
nm nm+1
ln Km εL,m,total eεL,m,total sm Ao,m − ln Km+1 εL,m+1,total eεL,m + 1,total sm+1 Ao,m+1
(6.58)
− εL,m + εL,m+1 = 0
or
     
Km nm nm+1
ln + ln εL,m,total − ln εL,m+1,total + εL,m,total sm
Km+1
  (6.59)
Ao,m
− εL,m+1,total sm+1 + ln − εL,m + εL,m+1 = 0
Ao,m+1

Finally,
     
Km Ao,m i−1
ln + ln + nm ln εL,m,acc + εL,m
Km+1 Ao,m+1
 
i−1
− nm+1 ln εL,m+1,acc + εL,m+1
(6.60)
i−1
+ εL,m (sm − 1) − εL,m+1 (1 − sm+1 ) + εL,m,acc sm
i−1
− εL,m+1,acc sm+1 = 0

Using Eq. (6.60), this leads to a system of m − 1 equations for m = {1 . . . (m − 1)}

where m is the number of nodes/elements.
It should be noted that the terms Ao,m and Ao,m+1 change independently from
each other from the time step i to i + 1, thus simulating a varying cross-sectional
area along the specimen length.
This leads to a system of m − 1 implicit nonlinear equations with m unknowns.
Thus, the final condition required is
6.2  Mechanical Modeling 137


m
d = Lm (6.61)
1

where d is the imposed input displacement and Lm is the change in length of an
element. This expression states that the summation of the element displacements is
equal to the overall imposed displacement.
Knowing
 
Lo,m + �Lm
εL,m = ln (6.62)
Lo,m

or
 
�Lm = Lo,m eεL,m − 1 (6.63)
Hence,

m
 
�d = Lo,m eεL,m − 1 (6.64)
1

Finally,


m
Lo,m (eεL,m − 1) − �d = 0 (6.65)
1

It should be noted that Lo,m can vary along the length of the specimen per each ele-
ment and it is the initial length of the element from the prior time step (i − 1) as a
result of d being defined as an increment (i.e., constant).
Therefore, using Eqs. (6.60) and (6.65), the strain in each element in the axial
direction can be determined by solving the system of equations. These strains
can then be used to determine the strain in the other directions for each element
using an associated flow rule, and the new dimensions of the elements can be
determined. Since the corresponding stress and area in each element relate to an
overall equivalent force, this force can be determined and used with the displace-
ment given to produce a force and displacement profile (just as would occur for an
experimental test).

6.2.2 Deformation/Strength Model Solution Method

There are several numerical methods to find the roots of this nonlinear implicit
system of equations. The Newton–Raphson method is efficient and converges
quickly given good initial approximations of the variables. The method is given by
Chapra and Canale [7] for a system of nonlinear equations:
138 6  Tensile Electroforming Model

f1 (x1 , x2 , . . . , xn ) = 0
f2 (x1 , x2 , . . . , xn ) = 0 (6.66)
f3 (x1 , x2 , . . . , xn ) = 0

For each equation, a multivariable first-order Taylor series expansion is written.

For example, for the kth equation,

  ∂fk,i   ∂fk,i
fk,i+1 = fk,i + x1,i+1 − x1,i + x2,i+1 − x2,i
∂x1 ∂x2
  ∂fk,i (6.67)
+ · · · + xn,i+1 − xn,i
∂xn

where the first subscript (k) represents the equation of unknown and the second
subscript denotes whether the value or function is at the present value (i) or at the
next value (i + 1).
By setting fk,i+1 to zero, this means that we are looking for the roots of the sys-
tem of equations. This then gives

∂fk,i ∂fk,i ∂fk,i

− fk,i + x1,i + x2,i + · · · + xn,i
∂x1 ∂x2 ∂xn
∂fk,i ∂fk,i ∂fk,i (6.68)
= x1,i+1 + x2,i+1 + · · · + xn,i+1
∂x1 ∂x2 ∂xn

By examining Eq. (6.68), the only unknowns are the xk,i+1 terms on the right-hand
side of the equation as all other quantities are known at the present value (i). This
now provides a system of linear equations that can be solved. Matrix notation is
used to simplify the expression. The partial derivatives can be expressed by
 ∂f1,i ∂f1,i

dx1 ··· dxn
 . .. 
[J] = 
 ..
..
. 
.  (6.69)
∂fn,i ∂fn,i
dx1 ··· dxn

where J is commonly called the Jacobian matrix.

The initial and final values in vector form are as follows:
 
{Xi }T = x1,i . . . xn,i (6.70)

and
 
{Xi+1 }T = x1,i+1 . . . xn,i+1 (6.71)
The function values in vector form are as follows:
 
{Fi }T = f1,i . . . fn,i (6.72)
6.2  Mechanical Modeling 139

Thus, the linear system can be expressed in the standard form:

Ax = b (6.73)
or
[J]{Xi+1 }T = −{Fi }T + [J]{Xi }T (6.74)
This system can be solved using a technique such as Gauss elimination. The process can
be repeated iteratively to obtain refined estimates of the unknown variables. For the sim-
ulations performed in this work, the Newton–Raphson method was repeated until the
norm (i.e., vector length) of the root vector was very small (i.e., less than 1 × 10−10).
Once the strains in each element are known in the axial direction, the strain in
the width and thickness direction can be determined using a material flow rule.
Presently, it is proposed to assume the von Mises yield criterion and material isot-
ropy (note: other yield criterion or anisotropy could be applied here as well). This
results in the Levy–Mises equations:
dε1 dε2 dε3
= = (6.75)
2σ1 −σ1 −σ1
where dεp is the incremental strain in the primary three directions and σ1 is the
stress applied along the length axis.
This results in
1
εw,m = εt,m = − εL,m (6.76)
2
After determining the strain in each direction (length, width, and thickness), the
new element geometry can be determined by
i+1
Lm = Lo,m eεL,m
i+1
wm = wo,m eεw,m (6.77)
i+1 εt,m
tm = to,m e
where i+1 , wi+1 , and t i+1 are
Lm m m the new length,
 width,
 and thickness of the element,
respectively. The new cross-sectional area Ai+1
m of the elements can be calculated
using the new width and thickness by
Ai+1 i+1 i+1
m = wm tm (6.78)
The stress for each element can be given by
  nm
σm εL,m,total , Tm = Km εL,m,total eεL,m,total sm (6.79)

where the constants are calculated from Eqs. (6.55)–(6.57) and the total strain in
the axial direction of the element is used. The force is calculated (equal for each
element) as follows:
F = Ai+1
m σm (6.80)
140 6  Tensile Electroforming Model

The force at each time step can be paired with the imposed displacement at that
time step to produce a corresponding force and displacement curve. This profile
can be converted to true strain and stress by the assumption of uniform deforma-
tion for comparison with experimental results.

6.2.3 Deformation/Strength Model Results

In the following sections, the model results are presented for the application of
both uniform temperature distributions and EAF temperature distributions during
tensile deformation.

6.2.3.1 Uniform Temperature Distribution

The uniform temperature distribution inputs are summarized in this section.

Detailed results (e.g., incremental strain, accumulative strain, stress, and force) are
presented for the room temperature (22 °C) test, and the other temperatures are
summarized by comparing the force output.

6.2.3.2 Results at 22 °C (Room Temperature)

The incremental strains to predict the deformation behavior of magnesium sheet

are shown in Fig. 6.18 for the simulation performed at 22 °C. The incremental
strain is equal for each element due to the temperature distribution applied to each
element being equal and constant throughout the entire simulation.
The incremental strains are small for each time step (i.e., displacement incre-
ment) as a result of the time step being small (0.5 s) with a deformation rate of
2.54 mm/min. Thus, at each time step, there is not a significant amount of strain
imposed to the entire material. The incremental strain decreases over time as a
result of the ratio between the final length of each element and the initial length
of each element decreasing with time. This can be seen in Eq. (6.81) where L
remains constant for each time step and Lo increases with each time step as defor-
mation is imposed. Thus, the quantity in the natural logarithm decreases, which
causes the incremental strain to decrease over time.
   
Lf �L
εL = ln = ln 1 + (6.81)
Lo Lo
From the incremental strain solution, the accumulation of these small increments
over time provides the total accumulative strain in each element. The total accu-
mulative strain for each element is given in Fig. 6.19.
Figure 6.20 shows the element length change with respect to the total imposed
displacement. As seen, the length of each element increases in length by 2 mm
6.2  Mechanical Modeling 141

Fig. 6.18  Incremental strain results for a uniform temperature input of 22 °C

Fig. 6.19  Accumulative strain results for a uniform temperature input of 22 °C

during the simulation. Again, the length increase for each element is equivalent as
the temperature of each element is the same throughout the simulation.
Corresponding to the increase in length, Fig. 6.21 displays the reduction in
area for each element based on conservation of volume. The area is reduced from
12.5 mm2 to approximately 9 mm2.
The stress required for deformation during the simulation is given in Fig. 6.22.
As seen, the true stress required to deform the sample increases with imposed
142 6  Tensile Electroforming Model

Fig. 6.20  Element length results for a uniform temperature input of 22 °C

Fig. 6.21  Element area results for a uniform temperature input of 22 °C

displacement. This is a result of the material strain hardening at this temperature

(22 °C).
The forming force from the simulation is given in Fig. 6.23.
As seen, the force increases and reaches a maximum force at approximately
21 mm of displacement, and then, the force begins to decrease. The decrease in
force indicates a material instability point where the material does not have perfect
6.2  Mechanical Modeling 143

Fig. 6.22  True stress results for a uniform temperature input of 22 °C

Fig. 6.23  Force results for a uniform temperature input of 22 °C

uniform elongation (i.e., indication of localized necking). However, localized

necking was not incorporated in the model and that is why the incremental strain
solutions were equal throughout the entire simulation. To better visualize the insta-
bility point, the force profile is plotted in Fig. 6.24.
Using the classical tensile instability approach by Considère, the predicted
point of instability for a uniform temperature input of 22 °C with this flow stress
model should occur at
144 6  Tensile Electroforming Model

Fig. 6.24  Force results plot

for a uniform temperature
input of 22 °C

Fig. 6.25  True stress plot for

a uniform temperature input
of 22 °C (experiment vs.
simulation)

n 0.1818
ε∗ = = = 0.1818 (6.82)
1−s 1−0
where the strain hardening exponent (n) is equal to 0.1818 and the softening coef-
ficient (s) is zero at 22 °C.
The instability strain can be converted to a displacement by
 
Linstability = Lo e∗ − 1 = (105 mm)e0.1818 − 105 mm = 20.93 mm (6.83)

where the gauge length (i.e., region where deformation occurs) is 105 mm. Thus,
comparing this result to Fig. 6.24, it is shown that the model predicts that instability
occurs at the same displacement (i.e., maximum force corresponds to 20.93 mm).
To verify the output of the model, Fig. 6.25 shows the true stress/strain
response versus the experimental data. As seen, the model accurately predicts the
6.2  Mechanical Modeling 145

experimental result. The model extends past the fracture point of the experimental
result as the model does not incorporate any failure criteria that would indicate
that the point at which the material will fail.

6.2.3.3 Results at Elevated Temperatures

The results from additional simulations are provided in Fig. 6.26 where the force
is proved as a function of total imposed displacement. As seen, the model predicts
the required forming force to decrease with increasing temperature.
Additionally, the prediction of instability was calculated using Eq. (6.82), and
the results are provided in Table 6.5. As seen, the model results from the simula-
tion were capable of predicting the instability point (i.e., maximum force) for each
temperature simulated.
Figure 6.27 compares the elevated temperature responses (i.e., 200 and 250 °C)
from the model to the actual experimental data. As shown, the model accurately pre-
dicts the response until greater strains (>0.3) are reached. This deviation is a result
of severe non-uniform elongation in the gauge region of the experimental test. This
non-uniform elongation is where the material begins to fail, and as a result of the
model not incorporating any failure criteria, this response is not modeled.
Thus, this section shows the ability of the model to predict the strain in
each element and the required force to deform all of the elements at varying

Fig. 6.26  Force results plot

for a uniform temperature
input of 22, 70, 130 and
200 °C

Table 6.5  Summary of instability strain and displacement predictions

22 °C 70 °C 130 °C 200 °C
Strain hardening exponent (n) 0.1818 0.1654 0.1436 0.0913
Softening coefficient (s) 0 0 −0.1728 −0.8238
Instability strain (mm/mm) 0.1818 0.1654 0.12244 0.05006
Instability displacement (mm) 20.93 18.89 13.68 5.39
146 6  Tensile Electroforming Model

Fig. 6.27  True stress plot for a uniform temperature input of 22, 200, and 250 °C (experiment
vs. simulation)

temperatures. Additionally, the simulations using the derived models matched the
experimental data.

6.2.3.4 Non-uniform Temperature Distribution Input

The previous thermal modeling section has shown that large thermal gradients can
exist along the length of a tensile sample during forming during EAF. As a result,
the capability of this deformation/strength model allows for the prediction of local
strain in each element and the overall forming force to deform the material.

6.2.3.5 EAF Diffuse Thermal Model Results (Parameter Set 4)

To simulate EAF forming conditions, the temperature distribution output from the
diffuse deformation model in the previous thermal modeling section is used to be
the input to the deformation/strength model in this section. The results below are
presented for Parameter Set 4 (500 A for 1 s every 60 s) to maintain consistency
with the results presented in the thermal modeling section. The thermal input to
the model is given in Fig. 6.28 where the temperature increases quickly as the cur-
rent is applied and then cools after the current is discontinued. Each of the tem-
perature rises on the figure represents an applied electrical pulse. Also, since the
thermal model assumed a diffuse geometry from experimental measurement, the
6.2  Mechanical Modeling 147

Fig. 6.28  Temperature distribution (EAF diffuse thermal model output for PS 4)

Fig. 6.29  Incremental strain results for EAF diffuse thermal model input for PS 4

temperature is greater for the center elements over time as a result of a reduced
cross-sectional area in these elements.
The solution to the simulation is presented in Fig. 6.29, which displays the
incremental strain results.
As seen, during a temperature rise, the incremental strain can now vary from
element to element. This is seen during the current application as the elements
with a greater temperature (i.e., near center) have a greater incremental strain as
148 6  Tensile Electroforming Model

Fig. 6.30  Accumulative strain results for EAF diffuse thermal model input for PS 4

compared to the elements at a lower temperature (i.e., near ends). Also, as the
material cools after the electrical pulse, it is shown that the incremental strain for
each element approaches each other as the strength properties become similar.
As a result of varying incremental strain results from the simulation, the accu-
mulative strain over time will not be equal for each element. The accumulative
strain for each element is presented in Fig. 6.30. As seen, the center elements with
greater incremental strain have a larger amount of strain imposed over time (i.e.,
accumulative strain). Also, the simulation solution does not begin to show a sig-
nificant amount of localized straining until near the end of the simulation. This can
be attributed directly to the input temperature distribution where there is a much
larger thermal gradient along the length at the end (i.e., >15 mm of total imposed
displacement).
Due to the diffuse accumulative strain, the length and area of each element will
vary at a given total input displacement. These results for element length and area
are presented in Figs. 6.31 and 6.32, respectively. The element length is directly
derived from the accumulative strain; thus, the overall profile shape is the same.
For the area, the elements with more strain (i.e., center) have a smaller cross-
sectional area due to the length of the element being greater.
The stress response from the simulation is given in Fig. 6.33. As seen, during
the application of current (i.e., temperature rise), the true stress of the material sig-
nificantly decreases as a result of the material being in a weaker state.
The stress response for each element at a given time step may not be equal due
to the elements having different cross-sectional areas. This is seen where the ele-
ments with a smaller cross-sectional area (i.e., center) have a greater stress than
the elements with a larger cross-sectional area (i.e., ends).
6.2  Mechanical Modeling 149

Fig. 6.31  Element length results for EAF diffuse thermal model input for PS 4

Fig. 6.32  Element area results for EAF diffuse thermal model input for PS 4

The force required for deformation is given in Fig. 6.34 where the force is
equal in each element. As the current is applied, the force required for deformation
is reduced.
To better visualize the force response, Fig. 6.35 displays the forming force
versus the imposed overall displacement. The force is reduced during the cur-
rent application and then increases as a result of the material decreasing in
150 6  Tensile Electroforming Model

Fig. 6.33  True stress results for EAF diffuse thermal model input for PS 4

Fig. 6.34  Force results for EAF diffuse thermal model input for PS 4

temperature. Also, it is shown that the decrease in force increases with respect to
displacement. This is a result of the material being deformed and the cross-sec-
tional area continuously decreasing with imposed displacement. As the cross-sec-
tional area is reduced, this results in greater temperatures which induce a greater
amount of material softening. Hence, the decrease in force increases over time.
6.2  Mechanical Modeling 151

Fig. 6.35  Force results plot for EAF diffuse thermal model input for PS 4

Fig. 6.36  True stress plot for

EAF diffuse thermal model
input for PS 4 (experiment
vs. model simulation)

The simulation of the model for Parameter Set 4 versus the experimental data is
given in Fig. 6.36.
As seen, the model accurately predicted the true stress/strain response until the first
application of electrical current. During the first application of current (i.e., first pulse)
and the subsequent applications, the model underpredicted the flow stress reduction.
The temperature input to the model assumed that the entire quantity of electrical energy
went to only Joule heating. Thus, the drop in stress is due to bulk thermal softening
and the electroplastic effect. For this model, it does not separate direct electrical effects
152 6  Tensile Electroforming Model

Fig. 6.37  Stress reduction difference during current application between experiment and model
simulation for PS 4

against bulk heating effects as a result of the diffuse thermal output from the ther-
mal modeling section being used. The division of bulk heating and the electroplastic
effect will be accounted for in the subsequent thermo-mechanical modeling section.
Additional error may be due to the initial linear strain input used in the diffuse defor-
mation thermal model to produce the thermal response applied in this modeling sec-
tion. To physically quantify the difference in flow stress reduction, Fig. 6.37 displays the
difference in the flow stress reduction between the experimental result and the model
result. The difference in the flow stress reduction is nearly constant for each electrical
pulse with an average of 27 MPa. Thus, it is assumed to be a result of thermal expansion
stress which was not considered in this model. The flow stress reduction due to ther-
mal expansion is incorporated in the thermo-mechanical model in the next section. As
a result of thermal expansion not being incorporated in the model, the simulation pre-
dicts a greater material flow stress as compared to the experimental result. The model
exceeds the point at which the experimental result failed due to the lack of failure crite-
ria applied to the model simulation.

6.2.4 Mechanical Modeling Conclusions

The flow behavior in uniaxial tension is one key aspect to understanding the
macro-material properties of sheet metal subject to electric current flow during
deformation. As a result, this section examined the local material strain and stress
by incorporating both a structural and geometric non-homogeneity in the sheet
material. The model in this section was capable of predicting the flow behavior
and instability strain of sheet forming at room temperature and at constant elevated
temperatures. For non-uniform temperature distributions as in EAF, the model
was capable of predicting the incremental strains throughout the sheet sample and
6.2  Mechanical Modeling 153

produced a resultant force that underpredicted the experimental EAF response.

This variation is due to the model not incorporating thermal expansion effects.

6.3 Thermo-Mechanical Model

In this section, the local material strain, flow stress, and thermal profile are pre-
dicted for EAF of sheet metal in uniaxial tension. This is accomplished by com-
bining the thermal model with the deformation/strength model. Additionally,
thermal expansion effects are incorporated as they also introduce stress to the
material. The division of electrical energy applied to the workpiece during EAF is
also divided between bulk thermal softening and direct electrical effects.

6.3.1 Thermo-Mechanical EAF Model Overview and

Solution Scheme

The thermo-mechanical EAF model incorporates bulk thermal softening effects, direct
electrical effects (i.e., electroplasticity), and thermal expansion effects to predict the
local material strain, flow stress, and thermal response of sheet metal during EAF. The
model solution scheme is given in Fig. 6.38. First, the initial conditions are set for the

Fig. 6.38  Multi-physics
EAF model solution scheme
154 6  Tensile Electroforming Model

sheet metal, which include the geometric and strength properties. Again, the sheet metal
in uniaxial tension is divided into elements along the length such that the geometry and
strength of the elements can vary spatially and with time. The simulation runs by deter-
mining whether the desired total displacement to the sheet is applied. When not fully
deformed to the desired amount, a displacement of Δd is imposed on the sheet metal.
Using the relationships from the mechanical modeling [Eqs. (6.60) and (6.65)], the
incremental element strains are calculated. Using the incremental strains, the element
shapes are updated corresponding to the amount of imposed strain in each element.
After, the element temperatures are calculated by using Eqs. (6.7)–(6.9) and (6.12). The
time step is compared with the electrical current application sequence during this step to
determine whether there is local heat generation due to Joule heating. After the tempera-
tures are determined, the element temperatures are stored and each element has its tem-
perature updated for the next iteration. As the temperatures are updated, new strength
and thermal properties are calculated for each element. Following, the process repeats
until the desired total displacement of the sheet is reached.

6.3.2 Thermal Expansion Stress

As there is a temperature distribution during EAF both along the specimen and as
a function of time, the elevated temperatures can impact the observed force due to
thermal expansion. As a result, this effect is incorporated into the model to deter-
mine its corresponding effects.
For an element m at time i, the coefficient of thermal expansion can be given by
  1
B
CTE = A Tmean [=] (6.84)
K
where
1  i−1 
Tmean = Tm + Tmi [=]K (6.85)
2
and such that for Mg AZ31B [8],

A = 8.7218 × 106
(6.86)
B = 0.17979

The modulus of elasticity can be given by

3 2
E = xTmean + yTmean + zTmean + w[=]GPa (6.87)
And, such that for Mg AZ31B [8],

x = 2 × 10−6
y = −0.0022
(6.88)
z = 0.8791
w = −68.205
6.3  Thermo-Mechanical Model 155

Thus, assuming a fixed–fixed end with no buckling, the thermal stress developed
in the element is given by
 
�σTE = 1,000(E)(CTE) Tmi − Tmi−1 [=]MPa (6.89)

This thermal stress is applied to the element during each time step such that the
effects of thermal expansion are incorporated.

6.3.3 Model Results

The model simulation results are presented in this section for Parameter Set 4
using the derived EAF thermo-mechanical model. The incremental strain (left) and
accumulative strain (right) are given in Fig. 6.39.
As shown, the incremental strains are greater for elements that are at higher
temperatures as compared to the elements with lower temperatures. The accumula-
tive strain increases with time, and the elements in the center have a greater overall
strain due to the elements having greater temperatures.
From the accumulative strains, the element length and element areas are cal-
culated. These results are presented in Fig. 6.40, where increased element length
results in a smaller cross-sectional area.
As a result of the EAF multi-physics model incorporating a thermal aspect, the
temperature distribution is calculated at each time step. The temperature distribu-
tion is given on the left of Fig. 6.41, and the maximum temperature with respect
to time from the model and experimental results is displayed on the right of
Fig. 6.41.
As seen, the thermal response increases during the application of the electri-
cal current and decreases once the current is discontinued. Also, the temperature
increases over time, and the element with the most strain (i.e., center element) is at

Fig. 6.39  Incremental strain (left) and accumulative strain (right) results from EAF multi-phys-
ics model for PS 4
156 6  Tensile Electroforming Model

Fig. 6.40  Element length (left) and element area (right) results from EAF multi-physics model
for PS 4

Fig. 6.41  Temperature distribution (left) and maximum temperature versus experimental results

(right) for PS 4

the highest temperature. The maximum temperature response from the model shows
good agreement with the experimental result; however, the last three pulses of elec-
trical current produce a greater temperature rise for the experimental result. This is a
result of the actual experimental specimen probably having a larger amount of local-
ized necking at the center of the sample as compared to the prediction of the model.
The maximum thermal error is on the last pulse and is approximately 20 °C.
The force to deform the sample is also provided from the EAF thermo-mechanical
model output. The force is given in Fig. 6.42, and it is seen that the current applica-
tion is predicted by a decrease in the material force. The increase in force increases as
a function of displacement due to the cross-sectional area of the sample decreasing.
The most valuable output from the model is the accurate prediction of the mate-
rial flow stress. The material flow stress output with no thermal expansion effects
is given by the left plot in Fig. 6.43.
As shown, the model displays similar results to the output from the mechani-
cal modeling section where the full reduction in stress during current application
is not modeled. However, including the stress reduction from thermal expansion
(right in Fig. 6.43) allows for a greater prediction of the material flow stress when
6.3  Thermo-Mechanical Model 157

Fig. 6.42  Force results plot from EAF multi-physics model for PS 4

Fig. 6.43  True stress plot for the EAF multi-physics model for PS 4 where left does not include
thermal expansion and right includes thermal expansion effects

comparing the model and simulation result. Thus, it is shown that a good portion
of the stress reduction is due to thermal expansion effects. This is studied in more
detail in the next section.

6.3.4 Division of Thermal Expansion, Thermal Softening,

and Direct Electrical Effects

One of the main outcomes from this chapter is the understanding of the actual mech-
anisms which reduce the material strength when applying an electric current field.
Thus, what portion of the electrical energy reduces the material flow stress from
thermal expansion, bulk thermal softening, and the electroplastic effect. The ther-
mal expansion stress effect could be directly calculated [Eq. (6.89)] and compared to
the overall predicted reduction in material flow stress. This results in a percentage or
158 6  Tensile Electroforming Model

Fig. 6.44  Division of
thermal expansion, bulk
thermal softening, and the
electroplastic effect for EAF

contribution to the overall stress reduction. It was found that the thermal expansion
stress accounts for approximately 30 % of the total reduction. The main question is
what portion is from direct electrical effects (i.e., electroplastic effect) and from bulk
thermal softening. To answer this question, the thermal response was analyzed for
each flow stress reduction due to the applied current. Using the thermal response and
the corresponding material strain at that time, the flow stress reduction due to purely
thermal effects was calculated. The calculation was performed using a constitutive
equation that predicts the material stress response at varying temperatures for this
material. Once the stress reduction due purely to bulk thermal heating was deter-
mined, it was compared to the result in Fig. 6.43 (left). The percentage of the total
reduction was found to be approximately 60 %. Thus, including thermal expansion,
this leaves the remaining 10 % to direct electrical effects or electroplasticity. These
calculations are summarized in Fig. 6.44. As seen, there is some variation, but the
overall trend attributes the most to bulk thermal softening and then to thermal expan-
sion. The smallest contribution is due to direct electrical effects.

6.3.5 Thermo-Mechanical Modeling Conclusions

The thermo-mechanical model introduced in this section combined the thermal

and mechanical model to predict the local material strain, flow stress, and thermal
6.3  Thermo-Mechanical Model 159

profile for EAF of sheet metals in uniaxial tension. The model does not utilize
experimental results for prediction of the material flow stress, and this model also
incorporates thermal expansion effects. The model results show very good agree-
ment with experimental testing, and analysis of the flow stress reduction during
current application is approximately 60, 30, and 10 % for bulk thermal softening,
thermal expansion, and the electroplastic effect, respectively.

6.4 Tensile Electroforming Model Conclusions

This chapter covered the modeling of tensile forming for sheet metals subject to
an applied electrical current field. As this process is coupled from a thermal and
mechanical standpoint, the modeling approach considered these two areas individ-
ually and then combined the individual models into a thermo-mechanical model.
Both the individual models were capable of predicting the desired response of
the material in uniaxial tension. The introduced thermo-mechanical model which
also included thermal expansion effects was able to predict the local material
strain, flow stress, and thermal profile for sheet metal in uniaxial tension during
EAF. Analysis of the flow stress reduction during current application showed that
approximately 60 % of the reduction is from bulk thermal softening, 30 % is from
thermal expansion, and 10 % is due to the electroplastic effect. Last, the mechani-
cal experimental testing displayed various current application methods and tech-
niques for utilizing EAF in sheet metal forming.

References

1. ASTM B557M—10 (2010) Standard test methods for tension testing wrought and cast alu-
minum- and magnesium-alloy products (metric)
2. United States Department of Defense (1998) Department of defense handbook: metallic mate-
rials and elements for aerospace vehicle structures, MIL-HDBK-5H
3. Cengel YA (2007) Heat and mass transfer: a practical approach, 3rd edn. McGraw Hill, Boston
4. Jones JJ, Mears L (2013) Thermal response modeling of sheet metals in uniaxial tension dur-
ing electrically-assisted forming. J Manuf Sci Eng 132:021011-1–021011-11
5. Salandro WA, Jones JJ, McNeal TA, Roth JT, Hong ST, Smith MT (2009) Formability
of Al 5xxx sheet metals using pulsed current for varying heat treatments. J Manuf Sci Eng
132:051016-1–051016-11
6. Salandro WA, Khalifa A, Roth JT (2009) Tensile formability enhancement of magnesium
AZ31B–O alloy using electrical pulsing. Trans North Am Manuf Res Inst SME 37:387–394
7. Chapra SC, Canale RP (2006) Numerical methods for engineers, 5th edn. McGraw Hill,
Boston
8. International Magnesium Association (1998) McLean, VA, and MIL-HDBK-5H, 1 Dec 1998,
pp 4–11
Chapter 7
Control of Electrically Assisted Forming

In this chapter, several control approaches are described for forming a metal under
an electrical current field. In addition, the approaches are demonstrated and poten-
tial applications for these control schemes are discussed. The specific examples
where closed-loop control is used to determine the process output are for constant
force forming, constant stress forming, and constant current density (CCD) form-
ing. Last, the feasibility toward model-based control (MBC) is discussed for the
models developed in Chaps. 4–6.
For most direct current sources, the control of the output current is based off of
a supplied feed voltage to the unit. Thus, the first step in controlling current to the
process is to define the relationship between the signal voltage to the current sup-
ply and resultant output current. An example is shown in Fig. 7.1.
The relationship can be used to derive the required signal voltage to elicit a
desired current:
I + 47.83
Vfeed = (7.1)
809.74
where Vfeed is the feed voltage required by the power supply and I is the desired
direct current output.
To control the feed voltage that is supplied to the power supply, a reconfigur-
able embedded control and acquisition system such as a National Instruments (NI)
CompactRIO (cRIO) should be used to provide precise control during the forming
process. An example of GUI is given in Fig. 7.2 for constant force forming.
To allow for material forming under a constant force, a control system that
provided continuous current during EAF testing was modified to incorporate a
proportional-integral-derivative (PID) control block, where the I and D gains
were set to zero. Thus, only the proportional component was used. The GUI is
shown in Fig. 7.2, where material dimensions are set along with the PID control-
ler gains and the desired set point. Additionally, limits were imposed such that a

© Springer International Publishing Switzerland 2015 161

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_7
162 7  Control of Electrically Assisted Forming

Fig. 7.1  Feed voltage and

current output relationship.
Offset and gain are calculated
for this linear relationship

maximum amount of current was allowed to be applied to the process for material
flow alteration. The limit in this research was set to allow a maximum of 300 A,
and the P control gain was set at a value of 3.00. The P control gain was exper-
imentally varied until an appropriate response of force modification by electri-
cal current was achieved (i.e., fast response in force change relative to overall
process time, and minimal overshoot). The control block diagram is shown also
in Fig. 7.2 where the PID block was used in Sect. 7.2 of the flat time sequence.
Last, it should be noted that there was a correction factor used to remove the ini-
tially observed steady-state error or droop. It is known that this additional cor-
rection factor could have been unnecessary if the use of the integral term was
used in the control of the process. However, for this work, it was not necessary
as the goal was not to perfectly tune the system but to demonstrate EAF control
architectures.

7.1 Constant Force Forming

The concept for constant force forming was realized from experimental test-
ing where the current was manually modulated such that the forming force could
be regulated to some extent. Thus, a formal control strategy was envisioned that
could regulate or maintain the force during forming at a specific set point value.
To achieve this, a block diagram (Fig. 7.3) was first constructed to understand the
flow of information and relationships. Fdesired is the desired force set point, Force
is the force feedback from the process, ΔF is the force error, and Vfeed is a feed
voltage that the current source uses to output current I to the process.
To realize the goal of constant force forming, a Darrah silicon controlled recti-
fier (SCR) with a current output of 0–4 kA was used to supply the process with
direct electrical current. To control the power supply, an external remote was built
using a NI cRIO-integrated controller/chassis containing various I/O modules
7.1  Constant Force Forming 163

Fig. 7.2  User interface and block diagram for current control. The user can set controller gains
and system parameters above; program execution is according to the block diagram

Fig. 7.3  Block diagram for constant force forming. Force is directly measured and controlled
through a reference tracker by modulating electric current [1]

programmed with NI LabVIEW software. A control schematic is presented in

Fig.  7.4, where a linear variable differential transformer (LVDT) provides dis-
placement data (d), and a load cell provides the force data (F) to the analog input
164 7  Control of Electrically Assisted Forming

Fig. 7.4  Control schematic using cRIO. The position, load, and current are acquired using an
analog input, and control is made using a signal voltage to the power supply through an analog
output [1]

(AI) on the cRIO. Additionally, the measured current (Imeasured) is collected using
the AI on the cRIO. The cRIO interfaces with a computer, which also records ther-
mal data (T). The cRIO controls the power supply output (I) by applying a feed
voltage (Vfeed) from the analog output (AO).
For the current supply to provide the correct output to the process, a feed volt-
age with a linear relationship was established. The feed voltage was used to com-
municate to the power supply to set the desired current output. Once the power
supply control was established, the LabVIEW software was modified to specifi-
cally include a PID control block. A PID controller is a general controller that cal-
culates an error between the set point value and the measured result. Using this
calculated error, the controller adjusts the process inputs to try to minimize the
error. The controller has three terms: P, I, and D, where P is dependent on the
present error between the actual state and set point, I is on the summation of past
errors, and D is the current gradient of error, estimated from the last two data
points. For constant force forming, three force set points were tested to show the
robustness of the control application.
The force results for constant force forming at 1,334 N (300 lb), 1,779 N
(400 lb), and 2,224 N (500 lb) are presented in Fig. 7.5. As the control system is
turned on just after the material’s yield point, the applied current quickly drives
the force to the desired set point. After reaching the desired set point value, the
controller is capable of accurately modulating the applied current to maintain con-
stant force forming until the specimen fractures. The physical reason for the oscil-
lations in the force response is not presently known. The oscillations may be a
result of the cyclic behavior of the applied current from the control algorithm, a
result of the material hardening and softening, or from an AC current component
imposed on the large DC signal as the power source does not provide a purely DC
signal.
7.1  Constant Force Forming 165

Fig. 7.5  Constant force forming at varying set points (no force correction). Three different force
set points are achieved through modulation of electric current [1]

The conversion of the constant force results to stresses is presented in Fig. 7.6

where the true stress increases linearly as a result of the force maintaining a constant
value. This calculation was performed assuming uniform deformation as given by:

Fig. 7.6  Stress response for constant force forming. The estimated stress response for constant
force control for the three different cases is given [1]
166 7  Control of Electrically Assisted Forming

(Force)(L)
σtrue = (7.2)
A0 L0
where Force is the instantaneous measured force, L is the instantaneous gauge
length, A0 is the initial cross-sectional area, and L0 is the initial gauge length.
Thus, the presented stress represents a mean stress state for the material. Due to
diffuse deformation during the process (as a result of specimen geometry and ther-
mal conditions), regions in the test sample will experience both lower and higher
stress states. Also of note is that the overall achievable elongation is near equiva-
lent to the room temperature behavior of the material. It was expected that form-
ing under a constant load would improve the amount of deformation the material
could withstand before fracture. However, this did not occur as a result of interac-
tion of thermal gradients generating localized areas of deformation which lead to
regions in a much high stress state, ultimately causing fracture.
The current applied during the process is summarized in Fig. 7.7 where a maxi-
mum allowable current was set (300 A).The results show the filtered response, but
the current actually increases to 300 A and then shortly decreases as the forming
force is reduced. After this initial spike, the current is modulated by the controller
such that a constant force is maintained. As seen, lower forming force set points
require a greater current input to maintain the desired force level. To compare this
profile to constant force forming under a compressive loading, it is expected that
the current will again have an initial spike to reduce the forming load to the desired
force set point. Once the force set point is reached, the current would presuma-
bly increase during the forming process due to the increasing cross-sectional area
and material strain hardening. Also, for compressive operations, this may require
higher maximum allowable current levels to achieve the desired force set point.

Fig. 7.7  Current application during constant force forming. Absolute current decreases with
elongation due to reduction in area [1]
7.1  Constant Force Forming 167

Fig. 7.8  Thermal response for constant force forming tests. The surface temperature increases
rapidly at first, then decreases with time and deformation displacement [1]

In addition, the thermal surface response of the tests was recorded and the
maximum temperature of the sample with respect to time is given in Fig. 7.8. As
the current is applied, the temperature drastically increases, and then, the rate of
change of temperature begins to decrease as the material reaches the desired force
set point (i.e., lower current level to maintain force). As the test continues, the
temperature follows the same trend as the electrical current which decreases until
the specimen fractures. The thermal response is presented here as this could repre-
sent another possible area for control. Specifically, the temperature during forming
could be controlled by modulating the electrical current applied if real-time tem-
perature data were available. A similar approach has been presented for stationary
heating using an electrical current before performing a Kolsky Bar test [2], but not
for forming during deformation.
As discussed previously, thermal gradients generated from the forming
boundary conditions (i.e., dies) create areas that undergo greater amounts of
deformation. Thus, these regions that are subject to greater amounts of defor-
mation are in a higher stress state. The stress profile presented in Fig. 7.6 is
therefore a mean stress prediction of the material based off of Eq. (7.2). The
strain and thermal gradient can be observed by examining the specimen geom-
etry and temperature as depicted in Fig. 7.9. The current was applied just after
33 s and was discontinued at 414 s. Also, the peak temperature exists in the
center of the specimen and is lowest near the die regions, thus the deforma-
tion is highly localized in the center of the specimen. For a practical usage of
this technique, die design is a critical area to be further investigated such that
any thermal gradients would be minimized from the workpiece to the dies.
Overall, the complex mechanical and thermal relations for EAF required more
advanced models if it is desired to understand the localized stress and thermal
gradients.
168 7  Control of Electrically Assisted Forming

Fig. 7.9  Thermal profile for constant force forming test (1,334 N) [1]

7.1.1 Benefits of Constant Force Forming

The significance of constant force forming allows the forming force to now be
specified as a control parameter and not just monitored as a process output. As a
result, this technique could allow for lower capacity (i.e., smaller force) machines
which often have smaller capital investments to form high-strength materials.
Additionally, with having the capability to form a greater range of material on a
lower capacity machine, this reduces the number of individual machines that a
company may require.
7.2  Constant Stress Forming 169

7.2 Constant Stress Forming

Constant stress forming was also performed using a similar method as described
for the constant force forming. The block diagram for constant stress forming is
presented in Fig. 7.10 where Force is the measured force, σtrue is the calculated
true stress [Eq. (7.2)], σdesired is the desired set point, and �σ is the stress error.
Using this equation to calculate stress assumes that the material deformation is
uniform along the length of the sample. However, in EAF, it is common to observe
diffuse deformation based off of the specimen geometry and thermal conditions
along the length of the specimen which would result in areas with localized stress.
Thus, this presented architecture is a simplified approach to constant stress forming
that provides a mean constant stress, and more advanced models can be incorporated.
The force results are shown in Fig. 7.11 for the constant stress forming tests. As
seen, the force is immediately reduced with the application of electrical current to
the desired stress level, and the force decreases linearly over the length of the test
to maintain a constant flow stress. Again, the physical reason for the oscillations in
the force response is not presently known.

Fig. 7.10  Block diagram for constant stress forming. The stress is calculated from force, dis-
placement, and initial part dimensions [1]

Fig. 7.11  Force response for constant stress forming. Force decreases with time for constant
stress forming [1]
170 7  Control of Electrically Assisted Forming

Fig. 7.12  Constant stress forming at varying set points. Stress is controlled to approximately

constant value through modulation of current [1]

The flow stress results are given in Fig. 7.12, and the true stress during forming
is maintained at the correct set point values of 100, 150, and 200 MPa. Again, the
conversion of force to stress is based on the assumption of uniform deformation.
Thus, from a bulk standpoint, the stress is a mean representation as discussed for
the constant force forming results. Also, the achievable elongation was comparable

Fig. 7.13  Current application during constant stress forming. The current application again
decreases with displacement stroke [1]
7.2  Constant Stress Forming 171

Fig. 7.14  Constant stress forming sample. The magnesium sample exhibits diffuse necking [1]

to the room temperature testing. This can be explained by certain regions have
areas of localized deformation causing the material to fail.
The filtered current supplied to the process is summarized in Fig. 7.13 for the
three test cases performed (100, 150, and 200 MPa). Although not shown clearly
by the filtered data, the current quickly increases to the maximum allowable cur-
rent (300 A) once the controller is activated and quickly decreases at the point
where the material reaches the desired stress state. Once the stress state is reached,
the current slowly decreases until the specimen fractures. Again, lower set points
require greater current levels. When considering the current response in general,
it is expected that a process using this control architecture in compression will
exhibit similar behavior; however, the current input would increase over time to
counter the cross-sectional area increase and material strain hardening.
For the constant stress forming results, uniform strain was assumed for the
entire test length. However, as a result of the testing setup, there is a thermal gra-
dient within the test samples which causes diffuse necking during the test (see
Fig. 7.14). Due to the diffuse necking, this modifies the actual local stresses within
the material due to the presence of an area gradient along the sample length.
Consequently, the presented response is an averaging of the true stress within the
sample, and it can be seen that the experimental response decreases slightly near
the end of the tests due to larger amounts of diffuse necking present just prior to
fracture.

7.2.1 Benefits and Opportunities of Constant Stress Forming

With the introduction of constant stress forming, this opens additional areas
of research for determining the desired or optimal material flow stress response
during forming for a given material/process combination. Additionally, this
172 7  Control of Electrically Assisted Forming

demonstration also leads to the opportunity for present forming machine architec-
tures/designs to be modified with the goal of becoming more flexible, which is
highly desirable in industry.

7.3 Constant Current Density Forming

CCD forming has been performed for uniaxial compression upsetting (Fig. 7.15)
for both 304 Stainless Steel and Grade 5 titanium (Ti-6Al-4V). The internal block
diagram for the overall control of the current scheme is shown in Fig. 7.16.
In the block diagram, the initial state of the force and load is set to use as a
reference point. Then, depending on the desired process (NCCD or CCD), the
controller calculates the feed voltage that gets supplied to the current source. The
current source reads the feed voltage and outputs a corresponding electrical cur-
rent to the forming process. This process is repeated until the desired amount of
material strain is reached. For the NCCD tests, the supplied current is constant
and the shape change makes the current density decrease with time (compression
tests). For the CCD tests, the supplied current is increased with time to maintain a
CCD irrespective of specimen shape change.
The variation in the material flow stress by taking into consideration compo-
nent shape change during the forming process is shown in Figs. 7.17 and 7.18 for
304 Stainless Steel and Grade 5 titanium, respectively. As seen, by modulating
the current during the test, the material flow response is altered as compared with
using only a nominal current value. As a note, for compression forming, the cur-
rent is applied continuously during the entire forming process. These results pro-
vide a better representation of the actual material stress due to an applied current
as specimen shape is not a factor.

Fig. 7.15  Constant current
density forming setup. The
specimen is pressed between
two platens
7.3  Constant Current Density Forming 173

Fig. 7.16  Internal block diagram for control of NCCD and CCD processes. For the CCD case,
voltage is calculated to maintain constant concentration of current [3]

Fig. 7.17  Flow curves for 304 Stainless Steel: comparing non-constant current density (NCCD)
versus constant current density (CCD). CCD case generally decreases with strain as compared
with the NCCD case [3]
174 7  Control of Electrically Assisted Forming

Fig. 7.18  Flow curves for Grade 5 titanium: comparing non-constant current density (NCCD)
versus constant current density (CCD). Again, the CCD case decreases stress compared with the
NCCD case [3]

Fig. 7.19  Theoretical
and experimental data of
electrical current output for
constant current density
forming

Last, the measured current during the process was compared to the theoreti-
cal current to maintain a CCD during forming, and this is shown graphically in
Fig. 7.19.

7.3.1 Benefits of Constant Current Density Forming

In summary, this work introduced the first CCD tests to examine the actual mate-
rial flow stress while subject to an EAF process. As a result of CCD testing, the
shape change of the specimen during the process did not influence the effect that
7.3  Constant Current Density Forming 175

the electrical current had on the material strength. Accordingly from CCD testing, it
was shown that the flow stress is further reduced as a result of the increased current
applied to the material with increasing strain. Moreover, increases in flow stress reduc-
tions were increased by approximately 30 % in certain cases. Additional experimental
results and data-driven modeling for both NCCD and CCD profiles are given in [3].
From an applications standpoint, this type of control technique could be used
to help maintain consistent material flow and strength levels through components
with varying cross sections such that the resulting output is a formed part with
more uniform strain/strength properties.

7.4 Model-Based Control Feasibility

MBC is a control method where the control system incorporates a process model in
the control algorithm. Within MBC, there have been numerous approaches devel-
oped and this work focuses on model predictive control (MPC). In MPC, the model
of the process is used to estimate the response of the system to apply control action
instead of waiting for feedback from the process. Specifically in MPC, a weighted
objection function is defined, the response of the system to the inputs is predicted
over a finite time horizon, the performance of the system is optimized with respect
to the objective function using design variables as system inputs, and the system
is driven toward the optimized state [4]. This type of strategy has two main advan-
tages over traditional control in that it (1) betters the performance as a result of an
understanding of the system physics instead of reactive compensation, and (2) the
process output can be optimized to any parameter(s) while the underlying model
may contain uncertainty [5]. A general MPC architecture is shown in Fig. 7.20.
When considering this control strategy for EAF, the previous sections used
a PID controller which employed a compensation strategy instead of predictive
action. Additionally, the desired state was directly measurable or capable of being
directly calculated from the actual state of the process. For advanced control of EAF

Fig. 7.20  General MPC block diagram. The MPC controller optimizes the objective function
with information from the process model. Output is the control input to the process [1]
176 7  Control of Electrically Assisted Forming

Fig. 7.21  MPC block diagram for temperature control during EAF, an example of application in
temperature control [1]

processes, the incorporation of MPC and physics-based models could allow for
immeasurable process outputs to be controlled by the use of measurable processes
feedback. From the models created in this text, the required force or stress, local
strain, and temperature profile of the tensile sample could be calculated. As a result,
one strategy using the thermo-mechanical process model for EAF developed in this
work could allow for the temperature of the formed tensile sample to be controlled.
Although the temperature is a measurable output, there are difficulties in measuring
the entire thermal response (i.e., large thermal gradients during EAF sheet forming)
as a result of image/data processing. Hence, real-time feedback may be limited to
point measurements on the tensile sample. The forming process could be controlled
such that the temperature does not exceed a certain value or the part is formed in
a certain temperature range. In addition, the input electrical energy to the process
could be minimized while still maintaining the constraints for temperature. The block
diagram is shown in Fig. 7.21, where the process measurements could include tem-
perature (most likely point measurements), current, force, and displacement. The
thermo-mechanical process model would allow for temperature prediction such that
the control actions could be set before the actual feedback or past output measure-
ments are provided. Again, the MPC is shown, providing a feed voltage (Vfeed) which
the current source translates to direct electrical current to the physical process.
Additional strategies could include maximizing the elongation before failure or
providing a desired elongation while minimizing the amount of electrical energy
applied to the component. Also, with further work in microstructure analysis of
EAF samples, this could allow for grain size control using current and the defor-
mation rate as the control variables.

7.5 Process Control Conclusions

The main conclusions drawn from this chapter are as follows:

• Several control approaches were envisioned, created, and demonstrated for
forming using an electric current field.
7.5  Process Control Conclusions 177

• The first examples of constant force forming and constant stress forming using
modulation of electric current flow through the workpiece were demonstrated
and successful at maintaining the desired set points.
• The constant force-forming control approach allows for the forming force to be
a specified process input and not just an output of the process. This can allow
for lower capacity machines to be used on a wider range of materials with vari-
ous strength properties.
• The constant stress forming was successfully demonstrated for three flow stress
set points. With this introduced capability, there are now additional areas where
future research could be performed. For example, the desired or optimal stress
response when forming a material using a certain process could be a possible
area. Additionally, this also leads to the opportunity for forming machine archi-
tectures/designs to be modified to allow for more flexibility in material defor-
mation which is highly desirable in industry.
• The CCD forming introduced the first material response under electrical current
irrespective of material shape change. As a result, the true response of the mate-
rial due to the electrical current field was visualized. This technique is applica-
ble in that it could allow for more consistent material flow and strength levels in
components with varying cross sections.
• MBC has potential applications for controlling EAF processes using derived
physics-based models. Specifically, this work concentrated on MPC which
is a subset of MBC. In MPC, the model of the process is used to estimate the
response of the system to apply control action instead of waiting for feedback
from the process. One control application was presented using the thermo-
mechanical process model for EAF such that the temperature during forming
could be controlled.

References

1. Jones JJ, Mears L (2013) Control of tensile forming force and stress using direct electric cur-
rent modulation Accepted in J. of Manuf. Science and Eng., doi:10.1115/1.4025566
2. Mates SP, Rhorer R, Whitenton E, Burns T, Basak D (2008) A pulse-heated Kolsky bar tech-
nique for measuring the flow stress of metals at high loading and heating rates. Exp Mech
48:799–807
3. Jones JJ, Mears L (2011) Constant current density compression behavior of 304 Stainless Steel
and Ti-6Al-4V during electrically-assisted forming. In: ASME international manufacturing
science and engineering conference, Corvallis, OR, p 9
4. Mehta P, Mears L (2011) Model based prediction and control of machining deflection error in
turning slender bars. In: ASME international manufacturing science and engineering confer-
ence, Corvallis, OR, p 9
5. Mears L, Mehta P, Kuttolamadom M, Montes C, Jones JJ, Salandro W, Werner D (2012)
Manufacturing process modeling and application to intelligent control. J S C Acad Sci 10:6
Chapter 8
Microstructure and Phase Effects on EAF

This chapter evaluates the microstructure and its impact on EAF. This chapter
is divided into three sections where Sect. 8.1 discusses the impact that different
starting grain sizes have on the thermal and mechanical profiles of EAF in com-
pression. Section 8.2 evaluates differences in starting dislocation density (from
specimen pre-working) and how they impact mechanical EAF profiles. Section 8.3
provides analysis for post-formed microstructure of tensile sheet specimens using
statistical methods. To illustrate these concepts, the chapter contains experimental
testing and analysis.

8.1 Grain Size Effect on EAF

The grain size of a metal has a significant impact on formability parameters of the
material, such as yield, ultimate strength, and percent elongation prior to fracture.
Previous researchers have noted these relationships between the grain size and form-
ing characteristics of various metals [1–3]. During deformation, the grain size has a
direct influence on the material flow stress due to the motion of dislocations which
interact with grain boundaries. As dislocations move through the crystal lattice, they
may encounter a grain boundary which often leads to dislocation buildup as a result
of dislocations being unable to move to the adjacent grain’s slip plane. This buildup of
dislocations within the lattice causes the amount of force to continue deformation to
increase (a phenomenon known as the strain hardening effect). Thus, a material with
a smaller grain size will have more grain boundary area for dislocation interaction and
the potential for increased amount of strain hardening over a larger grain material.
During EAF, the electrons that pass through the metal will also have an inter-
action with the grain boundary areas and will experience an increased amount of
scattering due to the lattice misalignments. Consequently, it is expected that sam-
ples of the same material with different grain sizes should react differently during

© Springer International Publishing Switzerland 2015 179

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_8
180 8  Microstructure and Phase Effects on EAF

deformation with an applied electrical current. In this section, conventional com-

pression, stationary electrical, and EAF compression tests are run on the same
material with several different grain sizes to determine the grain size/EAF effec-
tiveness relationship. Specifically, experimental testing is performed on 6.35-mm-
diameter 304 Stainless Steel specimens which were heat-treated, using three
different temperature and time combinations, to create three different size grain
structures. Conventional compression, stationary electrical, and EAF tests were
run on the different specimens in an attempt to quantify the effects of the different
grain sizes. This section will consist of (1) an explanation of the heat treating and
resulting grain sizes, (2) a description of the microstructure specimen preparation
grain size measurement procedure, and (3) an experimental grain size EAF investi-
gation through the types of tests referenced above.

8.1.1 Specimen Preparation and Resulting Grain Sizes

All specimens were produced from the same initial rod of annealed 304 Stainless
Steel, and once the compression slugs were sized appropriately, they were heat-treated
to produce different grain sizes. Specifically, the cycle at 1,066 °C for 0.5 h produced
the nominal grain size, the cycle at 1,010 °C for 1 h produced the smallest grain size,
and the cycle at 900 °C for 1.5 h produced the largest grain size. Table 8.1 summarizes
the time and temperature of each heat treatment and the actual resulting grain size.

8.1.1.1 Microstructure Preparation/Grain Size Measurement

In order to determine the average grain size for each respective heat treatment,
a Zeiss Axiovert 25 inverted optical microscope, equipped with a uEye camera,
and Buehler OmniMet software were used to generate micrographs to evaluate the
microstructure. For the micrographs, the specimens were cut parallel to and per-
pendicular to the longitudinal axis and then cold-mounted. A clearly defined pol-
ishing and etching procedure for SS304 was followed from [4].
The grain size measurement method consisted of measuring as many full grains
as possible within each micrograph (Fig. 8.1). Additionally, each grain was meas-
ured using a four-measurement method, where the grain is measured horizontally
(0°), vertically (90°), and on both diagonals (45° and 135°). The values found in
Table 8.1 are the average of all four measurements per grain.

Table 8.1  Summary of the Part # Temperature (°C) Time (h) Average grain size
average starting grain sizes (µm)
[5]
1 1,066 0.5 22
2 1,010 1 15
3 900 1.5 27
8.1  Grain Size Effect on EAF 181

Fig. 8.1  Example of the four-measurement method for determining the average grain size [5].
Each complete grain in the micrograph was measured at four different angles (0°, 45°, 90°, and
135°) to account for any differences in grain orientation

8.1.2 Experimental Grain Size Testing

The following results present the testing performed on the stainless steel sam-
ples with the different grain sizes produced from the heat treatments. Figure 8.2
shows the thermal profiles of stationary electrical tests run at a current density of
15 A/mm2 at a static load of 2,700 N for the three parts. Two replicates were run
for each grain size, and the replicates were consistent for each respective grain
size. The figure shows that the largest grain size has a thermal profile consistently
higher than the other two grain sizes (about 15 °C hotter at the end of the test).
This is uncharacteristic, because it was expected that the specimen with the small-
est grain size would have the hottest temperature profile because there would be
more localized scattering at grain boundaries. However, the difference between the
three grain sizes is minimal.
Figure  8.3 shows the thermal profiles for EAF tests run at a starting current
density of 15 A/mm2 and a deformation speed of 0.5 in/min. As was the case with
the stationary electrical tests in the prior figure, the specimens with the largest
grain size had the highest temperature response (again, about 15 °C hotter at its
maximum temperature).
182 8  Microstructure and Phase Effects on EAF

Fig. 8.2  Thermal profiles of 350 27.20µm

stationary electrical tests run 15.08µm
300 21.99µm
at CD = 15 A/mm2 [5]. There
was no significant difference
250

Temperature (°C)
in the thermal profiles of
the stationary electrical tests 200
run on the different heat
treatments 150

100

50

0
0 5 10 15 20
Time (s)

Fig. 8.3  Thermal profiles of 250

EAF tests run at CD = 15 A/
mm2 [5]. As was the case 200
with the stationary electrical 27.20µm
Temperature (°C)

thermal profiles, there is no

significant difference in the 150
EAF thermal profiles as well
15.08µm, 21.99µm
100

50

0
0 5 10 15 20 25 30
Time (s)

Fig. 8.4  True stress–strain 1400

profiles for conventional
compression tests [5]. When 1200
testing the different heat-
True Stress (Mpa )

1000
treated specimens under
conventional compression, 800
there was no difference in the 15.08 µ m, 21.99 µm,
and 27.20 µm
conventional compression 600
stress–strain profiles
400

200

0
0 0.2 0.4 0.6 0.8 1 1.2
True Strain

Figure 8.4 displays the true stress–strain profiles of all the parts during a con-
ventional compression test at a deformation speed of 0.5 in/min. The stress–strain
curves are identical, regardless of the grain size.
8.1  Grain Size Effect on EAF 183

Fig. 8.5  True stress–strain 1200

profiles for EAF tests
at CD = 15 A/mm2 [5]. 1000
As was the case with the

True Stress (Mpa)

conventional compression 800
stress–strain profiles, the
stress–strain profiles for the 600
15.08µm, 21.99 µm,
EAF test also showed no and 27.20 µm
difference 400

200

0
0 0.2 0.4 0.6 0.8 1 1.2
True Strain

Figure 8.5 depicts true stress–strain profiles for EAF tests with a starting cur-
rent density of 15 A/mm2. All of the profiles were the same, thus signifying no
relation with the different grain sizes tested.

8.1.3 EAF/Grain Size Conclusions

The differences between the grain sizes did not have an effect on the stress–strain
profiles of the conventional compression and the EAF compression tests under
these conditions. Additionally, there was only about a 15 °C difference in the ther-
mal profiles (stationary electrical and EAF compression) between the largest grain
size and the other two grain sizes; however, this was a very minor difference and
this did not have any effect on the mechanical performance during the EAF tests.
Work by Siopis et al. examined the effect of different grain sizes on EAF compres-
sion at the micro-level [6]. In this work, specimens with a diameter of up to 2 mm,
with grain sizes from 9 to 276 µm, were compressed, while electrical power was
applied. This work showed that the difference in the grains, at the micro-level, did
affect the EAF technique, in that as the grain size increased, the electrical thresh-
old current density increased.

8.2 Prior Cold Work Effect on EAF

Within this section, the relationship between dislocation density and efficiency of
EAF is explored. Since the starting grain sizes in the previous section were not dif-
ferent enough to produce experimental trends, the focus was shifted toward exam-
ining the effect of prior cold work in the specimens. As the cold work is increased
in a metal, the dislocation density within that metal will increase, thus leading
to a greater amount of obstacles which moving dislocations will have to surpass
184 8  Microstructure and Phase Effects on EAF

in order to continue plastic deformation. With this being said, this investigation
should be able to provide the same type of information as the grain size research
was intended to provide, such as how the different dislocation densities affect the
EAF technique. The specimen preparation, the experimental setup and procedure,
and the results and discussion of thermal profiles and stress–strain profiles of
worked/non-worked specimens are described.

8.2.1 Importance of Percent Cold Work on EAF

Effectiveness

The interactions between electrons and dislocations within a metal are responsible
for the electroplastic effect. The percent cold work, or prior plastic deformation
imposed on a metal before forming it again, has a correlation with the dislocation
density within a material. Depending on the material, as it is pre-worked to a greater
degree, the dislocation density within the metal increases potentially up to 1012 mm/
mm3 [7]. More specifically, as the percent cold work within a metal is increased, the
dislocation density within that metal is also increased. Since the flowing electrons
interact with dislocations, different amounts of dislocations within a metal should
have an effect on the efficiency of the EAF technique. Depending on the number of
preexisting dislocations, the flowing electrons will have more or less of an effect on
reducing the flow stress and extending elongation of the workpiece.

8.2.2 Specimen Preparation

All of the specimens used in this research work were derived from the same
9.5-mm-diameter rod of 304 Stainless Steel. This rod was sectioned off into
approximately 125-mm-long increments. These increments were then centerless
ground to 6.35 mm diameter, along with five other larger diameters which are rep-
resentative of the diameters which would be reached at various levels of cold work
(10 to 50 % in increments of 10 %—see Table 8.2). Now, having all the different
diameter rods, two types of specimens could be made: one without any cold work
(i.e., in the annealed state) and one with cold work (i.e., in the worked state). The
worked specimens were fabricated by simply taking a 6.35-mm-diameter speci-
men and physically compressing it to 10, 20, 30, 40, and 50 % cold work. The
non-worked specimens were created using the other rods which were centerless
ground to the same diameters as the cold-worked specimens at each cold work
percent (neglecting the barreling effect). The length of these specimens was sized
using an EDM operation. In the end, cold-worked and non-worked specimens
were fabricated from the same rod to the same dimensions and were ready for test-
ing and evaluation. Table 8.2 displays the length and diameter changes required to
produce each level of cold work.
8.2  Prior Cold Work Effect on EAF 185

Table 8.2  Initial percent cold work specimen dimensions [5]

h (mm) Δl (mm) %CW Dfinal (mm) ΔD (mm) Exp. verified Dia.
(mm)
9.525 0 0 6.350 0 6.350
8.573 0.925 10 6.693 0.343 6.706
7.620 1.905 20 7.100 0.750 7.163
6.668 2.858 30 7.590 1.240 7.696
5.715 3.810 40 8.198 1.848 8.230
4.763 4.763 50 8.890 2.630 8.915

8.2.3 Experimental Setup and Procedure

To truly determine the effect that cold work has on the EAF technique, three dif-
ferent experimental tests were run. Conventional compression tests were per-
formed to quantify the differences in the thermal and stress–strain profiles of the
different amounts of cold work without applying any electricity. Then, stationary
electrical tests were performed to determine the differences in the heating profiles
of specimens at a static load of 2225 N, with different levels of cold work where
no deformation was taking place. Finally, EAF compression tests were run, with
an electrical input of 800 A, to identify the differences in the thermal profiles and
stress–strain profiles of the different cold-worked specimens during deformation
and during the application of the EAF technique. Table 8.3 below shows the differ-
ent test-type/cold-worked specimen combinations performed for this work.

Table 8.3  Test-type/specimen combinations for %CW research [5]

Specimen type Test type Preload (N) Current (A) Amount of
deformation
(mm)
CW 10 % annealed, worked Stationary electrical 2,225 354, 531 0
CW 20 % annealed, worked Stationary electrical 2,225 403 0
CW 30 % annealed, worked Stationary electrical 2,225 466 0
CW 40 % annealed, worked Stationary electrical 2,225 532, 1,064 0
CW 50 % annealed, worked Stationary electrical 2,225, 400 0
62,300
CW 30 % annealed, worked Stationary electrical 42,275 466 0
CW 40 % annealed, worked Stationary electrical 50,063 532 0
CW 50 % annealed, worked Stationary electrical 62,300 625 0
CW 20 % annealed, worked Conv. comp. 2,225 0 4.445
CW 20 % annealed, worked EAF comp. 2,225 806 4.445
CW 40 % annealed, worked Conv. comp. 2,225 0 2.54
CW 40 % annealed, worked EAF comp. 2,225 806 2.54
186 8  Microstructure and Phase Effects on EAF

8.2.4 Results and Discussion

Figures  8.6 and 8.7 display the thermal profiles of stationary electrical tests per-
formed on 10 % CW/annealed specimens at starting current densities of 10 and
20 A/mm2, respectively. One can see that the thermal profile of the cold-worked
specimen in each figure is higher, possibly due to the fact that there are a greater
number of dislocations or lattice obstacles that work to resist the flow of electrons
and produce heat. Specifically, the maximum temperatures for the 10 A/mm2 tests
are 160 °C (annealed) and 200 °C (worked), and the maximum temperatures for
the 20 A/mm2 tests are 330 °C (annealed) and 425 °C (worked). Also of note is
that the 10 A/mm2 specimens reached a saturation temperature during the dura-
tion of the test, whereas the 20 A/mm2 specimens did not reach thermal saturation
in the short duration of the test. The remainder of the thermal results is given in
Appendix C.

Fig. 8.6  10 %CW/annealed 250
thermal profiles (stationary
10% CW – 2225N
electrical tests at CD = 10 A/
200
mm2) [5]. The thermal
Temperature (°C)

profile for the cold-worked

specimens was consistently 150
about 40° higher than the 10% Anneal – 2225N
annealed specimens
100

50

0
0 2 4 6 8 10
Time (s)

Fig. 8.7  10 %CW/annealed 450
thermal profiles (stationary 400
electrical tests at CD = 20 A/ 10% CW – 2225N

mm2) [5]. The difference in 350

Temperature (°C)

the thermal profiles increases 300

slightly toward the end of the
250
stationary electrical test to 10% Anneal – 2225N

about 80 °C 200

150
100
50
0
0 2 4 6 8 10
Time (s)
8.2  Prior Cold Work Effect on EAF 187

Table 8.4  Summary of %CW and annealed stationary electrical test results [5]

Stat. elec. test description Maximum Thermal profile description
temp. (°C)
10 %CW (CD = 10), 2,225 N 200 Thermal profiles are almost flat, signifying that
10 %Ann (CD = 10), 2,225 N 160 thermal saturation was almost reached
10 %CW (CD = 20), 2,225 N 425 Thermal profiles very curved, signifying that
10 %Ann (CD = 20), 2,225 N 330 the specimens did not reach thermal saturation
20 %CW (CD = 10), 2,225 N 230 Curved thermal profiles with a constant
20 %Ann (CD = 10), 2,225 N 200 temperature difference over the test duration
30 %CW (CD = 10), 2,225 N 300 Linear-type thermal curvature with
30 %Ann (CD = 10), 2,225 N 200 a constant 100 °C difference over the test
40 %CW (CD = 10), 2,225 N 365 Temperature difference increased
40 %Ann (CD = 10), 2,225 N 200 from 75 to 165 °C
30 %CW (CD = 10), 42,275 N 110 Linear-shaped thermal profiles with
30 %Ann (CD = 10), 42,275 N 85 increasing temperature difference to 25 °C
40 %CW (CD = 10), 50,063 N 117 Linear-shaped thermal profile with
40 %Ann (CD = 10), 50,063 N 82 a small temperature difference (20 to 35 °C)
50 %CW (CD = 10), 62,300 N 120 Linear-shaped thermal profile with
50 %Ann (CD = 10), 62,300 N 80 a small temperature difference (25 to 40 °C)

Table  8.4 summarizes the findings from the thermal results of the station-
ary electrical tests. For the same respective current density (10 A/mm2), the tem-
perature increased as the %CW increased (to an overall temperature difference of
165 °C). For the annealed specimens, there was only a 40 °C difference in the 10 %
annealed specimen temperature compared to the 40 % annealed thermal profile.
From the data in Table 8.4 and Fig. 8.8 could be created which shows the rela-
tionship between the maximum temperature of the stationary electrical tests and
the initial percent cold work. In this figure, the maximum temperature (which con-
sistently occurred at the end of the tests) increased as the initial percent cold work

Fig. 8.8  Temperature 400
Temp. (%CW) = 0.0875(%CW) 2 + 1.275(%CW) + 176.25
versus percent cold work
relationship for SS304 350
[5]. The temperatures of
Temperature (°C)

stationary electrical tests 300 Worked

at the same current density Annealed
increased as the initial 250 Poly. (Worked)
percent cold work was
increased 200

150

100
0 10 20 30 40 50
Percent Cold Work (%CW)
188 8  Microstructure and Phase Effects on EAF

was increased due to the fact that the greater amount of cold work had a larger
number of interfacial defects within the lattice. With the greater number of dislo-
cations, the flowing electrons had more obstacles to scatter off, and thus, the heat-
ing of the workpiece was increased.
In addition to analyzing the thermal profiles of stationary electrical tests, the
thermal profiles of worked and non-worked specimens undergoing conventional
compression and EAF tests were also evaluated. In the thermal plots for EAF com-
pression tests, the temperature will decrease at some point in the test, where the
maximum temperature will not be at the end of the test, as was the case with the
stationary electrical tests. This is due to the increasing cross-sectional area on the
specimens as they are compressed. Specifically, the current used to make the start-
ing current density is held constant, and therefore, the current density decreases
over the duration of the test. This causes the specimen temperature to peak at some
point in the test and then reduce for the remainder of the test.
Figures 8.9 and 8.10 show the thermal profiles of 20 % worked/non-worked spec-
imens undergoing conventional compression and EAF compression, respectively. In

Fig. 8.9  20 %CW/annealed 90
thermal profiles (conventional 80
compression) [5]. The
cold-worked conventional 70
20% CW
Temperature (°C)

compression thermal profile 60

was no greater than 10 °C
higher than the thermal 50
20% Anneal
profile of the annealed 40
specimen
30

20

10

0
0 2 4 6 8 10
Time (s)

Fig. 8.10  20 %CW/annealed 300
thermal profiles (EAF tests at 20% CW

800 A) [5]. As was the case 250

with the stationary electrical 20% Anneal
Temperature (°C)

tests, the cold-worked 200

specimen temperature profile
for the EAF tests was higher 150
than the annealed specimen’s
temperature profile. The 100
largest recorded difference
throughout the tests was 50
about 30 °C
0
0 2 4 6 8 10
Time (s)
8.2  Prior Cold Work Effect on EAF 189

the conventional compression tests in Fig. 8.9, the worked specimen was about 10 °C
hotter than the non-worked specimen. From the profile, the temperatures were identi-
cal at the beginning of the test. Then, the largest difference between tests was near
the middle of the tests. Toward the end of the tests, the temperatures of the worked
and non-worked specimens became close to one another. The same trend was fol-
lowed for the EAF tests in Fig. 8.10, where the worked and non-worked specimen
temperature profiles were the same at the beginning, they reached their maximum
difference near the middle of the test, and then, the difference between the temper-
atures reduced slightly as the tests came to an end. This divergent then convergent
thermal behavior is a result of the initial dislocation densities in the parts. With the
higher starting dislocation density, the %CW thermal profile increases at a greater
rate in the beginning of the test, whereas the annealed specimen (with a lower dis-
location density) showed a linear increase in temperature. In the middle of the test,
where the %CW specimens were the hottest, conduction heat transfer could be
responsible for the convergence of the %CW and annealed thermal profiles.
The engineering stress–strain plots for the 20 % worked/non-worked tests are
shown below. Specifically, the conventional compression tests are in Fig. 8.11 and
the EAF compression tests are in Fig. 8.12. Looking at the conventional compres-
sion tests in Fig. 8.11, one can see that the worked specimens can be distinguished
from the annealed specimens due to the large difference in the flow stress yield
point. However, this difference in flow stress is depleted about halfway through the
testing, and the ending stress for both specimen types is the same. The EAF com-
pression tests in Fig. 8.12 show the same starting difference in stress at the begin-
ning. The difference decreases through the tests, but there is still a recognizable
difference between the stresses of the worked and annealed specimens undergo-
ing these tests. This signifies that the difference in dislocation densities due to the
20 % cold-working was enough to minimally affect the effectiveness of the EAF
technique. The remainder of the mechanical results and discussion is continued in
Appendix C.
From the thermal results, it can be seen that the worked specimens consistently
produced higher thermal profiles compared to the annealed specimens of the same

Fig. 8.11  Stress–strain 7000
profiles (20 %CW/annealed,
conventional compression) 6000
Engineering Stress (Mpa)

[5]. The stress–strain profiles

5000
of the worked specimens are
higher at the beginning of 4000
the test, and by the end of Conv - 20% CW
the tests, both stress–strain 3000
profiles are at the same level
2000
Conv - 20% Anneal
1000

0
0 0.1 0.2 0.3 0.4 0.5 0.6
Engineering Strain
190 8  Microstructure and Phase Effects on EAF

Fig. 8.12  Stress–strain 4500
profiles (20 %CW/annealed, 4000
EAF at 800 A) [5]. The EAF

Engineering Stress (Mpa)

stress–strain profiles show 3500
the same differences as the 3000
conventional compression
2500
tests; however, the stress– EAF - 20% CW
strain profile of the worked 2000
specimens is still slightly 1500
higher than the annealed EAF - 20% Anneal
1000
profile at the end of the tests
500
0
0 0.1 0.2 0.3 0.4 0.5 0.6
Engineering Strain

dimensions. As the percent cold work is increased, the metal has a higher disloca-
tion density (i.e., a greater number of dislocations per area of lattice). With more
lattice obstacles, it is understandable that the worked specimens would be hotter
because there are more lattice obstructions for the flowing electrons.
In an attempt to correlate the different levels of cold work with different dis-
location densities, the following was performed. First, a static load of the magni-
tude that it took to deform the specimens to the desired height was applied to each
specimen, while 300 Amps of electrical current was run through the specimen. A
value of 300 A was chosen such that there was sufficient current to cause a heat-
ing effect, but not so much current to resistively heat the specimen to tempera-
tures above the maximum temperature of the thermal camera. While the current
was applied, the voltage between the top and bottom dies was recorded using a
multimeter. This procedure was performed for each level of cold work, and the
resulting voltages are shown below in Table 8.5. In addition, Fig. 8.13 provides a
way to compare the voltages of the annealed and worked specimens at the same

Table 8.5  Stationary Height (%) Designation Static load (N) Voltage

electrical voltage (mV)
measurements of worked and
10 Worked 22,250 70
annealed specimens [5]
Annealed 57
20 Worked 32,040 58
Annealed 30
30 Worked 42,275 31
Annealed 17
40 Worked 50,063 24
Annealed 15
50 Worked 62,300 15
Annealed 8
8.2  Prior Cold Work Effect on EAF 191

Fig. 8.13  Voltage versus 80
percent cold work [5]. The
70
voltages resulting from

Measured Voltage (mV)

running a small amount 60
of electricity through Worked specimens
the specimens worked to 50
Annealed specimens
different levels are shown 40

30

20

10

0
0% 10% 20% 30% 40% 50%
Height Reduction Amount (%)

height reduction levels. From the figure, it can be seen that the annealed specimen
voltages were consistently lower than the voltages of the worked specimens. This
supports the belief that the worked specimens had much higher dislocation densi-
ties compared to the annealed specimens, which lead to more resistive heating and
overall higher temperature profiles for the worked specimens.

8.2.5 EAF/Percent Cold Work Conclusions

The amount of percent cold work in a material designates the dislocation density
within the material. Within this section, stationary electrical tests were run where
temperature was recorded, and EAF tests were run where temperature and stress–
strain profiles were recorded. The conclusions from this section are as follows:
• For stationary electrical tests, the temperature profile will increase as the start-
ing percent cold work in the specimen is increased.
• The recorded voltages in the stationary electrical tests increased as the percent
cold work increased and the higher resulting dislocation densities caused the
resistance of the specimen to increase. The measured voltages of the %CW and
annealed specimens decreased at the same rate, but the annealed voltage values
were consistently below the %CW voltage values.
• The amount of induced percent cold work within a specimen has an effect on
the ability of the applied electricity to reduce the overall flow stress. The overall
reduction of the flow stress between conventional and EAF testing of annealed
specimens may be greater than the difference between cold-worked specimens.
• The stationary electrical test thermal profile curvatures are dependent on the
electrical application parameters (current density) and the material parameters
(%CW or annealed). Sometimes, this profile is curved and still increasing, and
sometimes, it can flatten out, thus representing that thermal saturation has taken
place.
192 8  Microstructure and Phase Effects on EAF

Fig. 8.14  Sample mounting
locations and orientations

8.3 Microstructure Analysis Under Tensile Loading

Bulk properties such as strength or ductility of a metal can be empirically obtained

from material testing. However, to fully understand these observed macro-prop-
erties, the underlying material microstructure must be examined. Therefore, this
section examines the microstructure of Mg AZ31B sheet metal in the as-received
condition and under various EAF conditions. As the microstructure may vary in
different orientations or spatially due to processing, several locations and orienta-
tions on the ASTM tensile samples were examined. The possible locations and ori-
entations examined are shown in Fig. 8.14 where L1 to L4 represent the possible
mounting locations (select areas were chosen depending on the desired compari-
sons) and 1–3 represent the different orientations at a given location.
Direction 1 is oriented such that the micrograph is showing a plane perpen-
dicular to the rolling direction where the roll direction is parallel to the specimen
length axis. Direction 2 is oriented to provide a micrograph that examines a plane
parallel to the rolling direction. Direction 3 is oriented orthogonal to direction 1
and direction 2 such that the micrograph is in plane with the top surface of the
sheet. In order to specify these orientations during mounting, different color speci-
men clips were used. Thus, in the below results which label the sample, a capi-
tal letter “R” denotes orientation 1, “B” or “K” denotes orientation 2, “NONE”
denotes orientation 3, and additional letters and numbers represent the image
number. As an example, “S1Rb2” means Sample 1 showing orientation 1 and the
image number is b2.

8.3.1 As-Received Material Microstructure

The as-received material was first examined to evaluate grain size and to under-
stand the distribution of the grain size measurements. To quantify the grain size,
approximately 100 grains were each fit by an ellipse for each micrograph. From
the ellipse geometry, the area of each measured grain could be calculated. Using
the calculated area, the equivalent circular grain diameter was determined and this
value is presented to represent the grain size in the micrographs.
8.3  Microstructure Analysis Under Tensile Loading 193

Fig. 8.15  3D microstructure
of as-received material
(sample 1) showing
orientations 1, 2, and 3

Figure 8.15 displays a 3D representation of the as-received material microstruc-

ture along with the three material orientations.
For orientation 1, the average grain size is 6.34 µm with a standard deviation
of 2.46 µm and orientation 2 has an average grain size of 6.35 µm with a standard
deviation of 2.15 µm. Again, since this sheet was warm-rolled, there is no sign of
residual cold work in these orientations as verified by the lack of twin boundaries.
The grain size of the as-received material in orientation 3 is 4.93 µm with a
standard deviation of 1.71 µm which is typical for this Mg alloy and for a warm-
rolled sheet. Also, since this sheet was warm-rolled, there should not be an abun-
dance of cold work instilled into the material. This is visually verified as the
micrographs do not appear to display much or any significant amount of twinning.
A twin boundary would appear similar to that of a grain boundary, but it would be
slicing through the grain linearly.
The numerous measurements taken on each micrograph allow for a distribution
of measurements to be generated as shown in Fig. 8.16. The data presented in this
figure use the as-received material in orientation 3 as an example. From examining
the distribution, the data set appeared to follow a lognormal distribution. This type
of distribution was verified by performing a lognormal distribution test, and the
results are presented in Fig. 8.17. As shown, a p-value of 0.961 resulted from the
test which is highly suggestive that the data set of grain sizes is represented well
by a lognormal distribution.
In addition to examining the distribution of the data set, several micrographs
were taken on the same sample in the same orientation to ensure that the results
194 8  Microstructure and Phase Effects on EAF

Fig. 8.16  Lognormal
histogram of equivalent
circular grain size for the
as-received material in
orientation 3

Fig. 8.17  Lognormal
probability plot of equivalent
circular grain size for the
as-received material in
orientation 3

were equivalent. This was performed as each micrograph is only displaying a

small portion of the actual mounted sample. This analysis was conducted, and it
was concluded that the micrograph image location for a given orientation did not
have a significant influence on the measured distribution estimation. The analysis
was conducted on the as-received material. The average equivalent circular grain
sizes for images a, b, and c of orientation 3 are shown in Fig. 8.18 along with a
95 % lognormal confidence interval (CI) as an example.
To specifically determine whether there is a statistical difference between the
data sets, two different test methods were used. To perform the tests, the data were
first transformed to normal by taking the natural logarithm of the data sets such
that the tests performed would be valid (since the statistical tests require assump-
tion of normally distributed data sets). After performing the tests, it can be inferred
from the results if the original data sets were equivalent.
8.3  Microstructure Analysis Under Tensile Loading 195

Fig. 8.18  Average and 95 %

lognormal confidence interval
of equivalent circular grain
size at three locations for
the as-received material in
orientation 3

The first was a two-sample t test with unequal sample sizes and variances
(Welch’s t test) to verify whether the population means are equal. The test pro-
vides a p-value which can be compared to a level of significance. The p-value for
this test is calculated by using the Student’s t distribution.
The second test is to examine the variance of two data sets to verify that they
are equal. To perform this, both the Levene test and Bartlett test were performed
and the average result is provided. The Bartlett test was designed for a nearly
normal distribution where the distribution does not affect the result of the Levene
test [8].

8.3.2 Summary of Statistical Analysis of Micrographs

This section provides a statistical comparison between all the samples studied in
this section, and a summary of the samples examined is provided in Table 8.6. As
seen, there are a total of 18 samples which contain microstructure images for the
as-received material, room temperature deformation, EAF square wave testing,
EAF continuous wave testing, and incremental EAF tests. This section only con-
tains a representative analysis of these samples. For a full summary and analysis,
refer to work by Jones [9]. To compare the micrographs, the two-sample t test is
used to determine whether the measured means are statistically equivalent and the
Levene/Bartlett tests are used to conclude whether the grain size measurements
have equal variances. The results from these tests are summarized for all sam-
ples although not all samples are directly comparable (i.e., specific comparisons
are made for select samples in the below discussion). When comparing the micro-
graphs of the samples, orientations 1 and 2 were only examined as orientations
2 and 3 have the same axial elongation (i.e., orientations 2 and 3 are equivalent).
Hence, orientation 1 and orientation 2 are compared across all samples separately.
The average equivalent circular grain size for all samples in orientation 1
is given in Fig. 8.19 along with a 95 % lognormal confidence interval. As seen,
Table 8.6  Summary of samples examined
196

Sample # Orientation Location Deformation Wave shape Current Pulse Pulse Notes
1 2 3 magnitude duration period
Sample 1 Rb Bc Noneb n/a No n/a n/a n/a n/a As-received material
Sample 2 Rb2 Bc2 n/a L1 Yes n/a n/a n/a n/a Room temperature deformation
Sample 3 Ra Bc n/a L1 Yes Square 500 A 1 s 60 s EAF square wave (500 A/1 s/60 s)
Sample 4 Rc Bb n/a L4 No Square 500 A 1 s 60 s Stationary electrical test
Sample 5 Rb2 Bb2 n/a L1 Yes n/a n/a n/a n/a Electrically treated deformation test
Sample 6 Rc2 Ba n/a L1 Yes Continuous 200 A n/a n/a Continuous EAF (200 A)
Sample 7 Ra Bc n/a L2 Yes Continuous 200 A n/a n/a Continuous EAF (200 A)
Sample 8 Ra Bb n/a L3 Yes Continuous 200 A n/a n/a Continuous EAF (200 A)
Sample 9 Ra Bc n/a L1 Yes n/a n/a n/a n/a Room temperature deformation
to 5.08 mm
Sample 10 Rc2 Bb n/a L1 Yes Square 1000 A 1 s n/a Room temperature deformation
to 5.08 mm and single pulse
Sample 11 Rb2 Bb2 n/a L1 Yes Continuous 150 A n/a n/a Continuous EAF (150 A)
Sample 12 Ra2 Ba2 n/a L2 Yes Continuous 150 A n/a n/a Continuous EAF (150 A)
Sample 13 Ra Ba n/a L3 Yes Continuous 150 A n/a n/a Continuous EAF (150 A)
Sample 14 Rb Bc n/a L1 Yes Square 800 A 0.5 s 60 s EAF square wave (800 A/0.5 s/60 s)
Sample 15 Rb Bc n/a L2 Yes Square 800 A 0.5 s 60 s EAF square wave (800 A/.5 s/60 s)
Sample 16 Rc2 Bb2 n/a L3 Yes Square 800 A 0.5 s 60 s EAF square wave (800 A/.5 s/60 s)
Sample 17 Ra2 Kc2 n/a L1 Yes Continuous 200 A n/a n/a Continuous EAF (200 A) with cur-
rent discontinuation
Sample 18 Rb Kc n/a L1 Yes Square 500 A 0.5 s 60 s EAF square wave (500 A/.5 s/60 s)
8  Microstructure and Phase Effects on EAF
8.3  Microstructure Analysis Under Tensile Loading 197

Fig. 8.19  Average and 95 % lognormal confidence interval for all microstructure samples in ori-
entation 1

there is not a great difference in the average grain size (3.82–7.48 µm) or variance
(2.41–8.45 µm2).
The results for the statistical tests are given in Tables 8.7 and 8.8 for orientation
1. The values from the tests that are lower than 5 % are highlighted to show the
samples which do not provide statistical equality assuming a 95 % confidence level.
The average equivalent circular grain size for all samples in orientation 2 is
given in Fig. 8.20 along with a 95 % lognormal confidence interval. Again, there is
not a great difference in the average grain size (4.18–7.01 µm) or variance (1.66–
5.77 µm2) for all the samples examined in this work.
The results for the statistical tests (t test and Levene/Bartlett tests) are given in
Tables 8.9 and 8.10 for orientation 2. The values from the tests that are lower than

Table 8.7  Two-sample t test results for all microstructure samples in orientation 1

198 8  Microstructure and Phase Effects on EAF

Table 8.8  Average of Levene and Bartlett results for all microstructure samples in orientation 1

Fig. 8.20  Average and 95 % lognormal confidence interval for all microstructure samples in ori-
entation 2

5 % are highlighted to show the samples which are not statistically equivalent,
assuming a 95 % confidence level.

8.3.3 Room Temperature Deformation Microstructure

The microstructure of the room temperature deformation (Sample 1) was exam-

ined in orientations 1 and 2 at the fracture location (L1). The room temperature
test had approximately 22 % strain imposed at material failure. The amount of
8.3  Microstructure Analysis Under Tensile Loading 199

Table 8.9  Two-sample t test results for all microstructure samples in orientation 2

Table 8.10  Average of Levene and Bartlett results for all microstructure samples in orientation 2

strain imposed is significant on the final grain distortion from the original state.
The micrographs are presented in Figs. 8.21 and 8.22 for orientations 1 and 2,
respectively. As shown in Fig. 8.21, the micrograph shows signs of material defor-
mation by the change in the grain shape and the abundance of twin boundaries.
The average grain size of the room temperature deformation microstructure in ori-
entation 1 is 6.73 µm with a standard deviation of 2.29 µm. A few of the twin
boundaries are denoted in Fig. 8.21. Statistically, the mean grain size and variance
are the same compared to the as-received material (Tables 8.7 and 8.8). This can
be physically explained by the deformation altering the shape of the grains to be
less equiaxed, but the grain area is still the same. Hence, the equivalent circular
grain size and variance are statistically equal.
200 8  Microstructure and Phase Effects on EAF

Fig. 8.21  Microstructure
of room temperature
deformation (sample 2) in
orientation 1 at L1

Fig. 8.22  Microstructure
of room temperature
deformation (sample 2) in
orientation 2 at L1

For orientation 2, it is noted that the grain shape is more changed, which is
represented by the jagged appearance of the grains and the grains being less equi-
axed. Also, since this orientation is in line with the direction of loading, there is
a significantly greater amount of twin boundaries present. A few of the bounda-
ries are shown in the figure for reference. The average grain size of the room tem-
perature deformation microstructure in orientation 2 is 6.15 µm with a standard
deviation of 2.33 µm. Again, statistically, these values are equivalent in terms of
equivalent grain size and variance to the as-received material. This can physically
be explained by the grains still having the same area after deformation.
Overall, the microstructure of the room temperature deformation sample shows
a less equiaxed grain structure with a significant amount of twinning. The presence
of twinning is due to the limited number of slip systems active at room tempera-
ture for this material.
8.3  Microstructure Analysis Under Tensile Loading 201

8.3.4 EAF Microstructure

In the following sections, various EAF micrographs are compared to the as-
received material and to room temperature deformation micrographs in order to
gain a better understanding of material deformation during EAF. The mechani-
cal results that generated these specimens are detailed in Appendix B and will be
noted in the below sections. Additional micrographs for the below results can be
seen in Appendix C and will be noted in the text.

8.3.4.1 Stationary Electrical Square Wave Current Application

To examine the influence of an applied electrical square wave without any defor-
mation, a test was performed where 500 A was passed through the sample every
60 s for the duration of 1 s. This square wave was applied to the specimen for nine
minutes which equates to nine electrical pulses (2.25 × 106 J/Ω). The micrograph
for orientation 1 is given in Fig. C.12 of Appendix C where a visual analysis sug-
gests no alterations to the microstructure when it is compared to the as-received
material.
This sample was compared to the as-received material as no deformation was
imposed. The average grain size for Sample 4 in orientation 1 is 6.78 µm with a
standard deviation of 2.13 µm. From the statistical analysis, the visual results are
confirmed in that the mean grain size and variance are equal.
The micrograph for orientation 2 is shown in Fig. C.13 where the average grain
size is 7.01 µm with a standard deviation of 2.22 µm.
From a visual inspection, there appears to be no difference as compared to the
as-received material in this orientation; however, the statistical analysis suggests
that the means are not equal. However, the Levene/Bartlett tests suggest that the
variances of the images are equivalent. The grain size in orientation 2 is slightly
larger as compared to the as-received material, and the reason for this may be due
to a temperature rise of the material which allowed for a slight amount of grain
growth which could have been favored in this orientation. Yet, there is no indica-
tion of a direct electrical effect on the material’s grain size under stationary test-
ing. From the mechanical testing of these samples after electrical treatment, it
was concluded that stationary electrical treatment did not significantly alter the
mechanical response. This is confirmed from this microstructure analysis as the
material after electrical treatment is mostly equivalent to the as-received mate-
rial at this level of analysis. However, some dislocation annihilation may have
occurred to the as-received material due to the applied electrical current. This
type of analysis would need an additional study to quantify the dislocation density
before and after electrical treatment. As a result of the grain sizes being mostly
equal (orientation 2 is slightly larger in equivalent average grain size), any would
be attributed to the annihilation of dislocations or a change in dislocation density
within the material’s lattice.
202 8  Microstructure and Phase Effects on EAF

After completing this test, another sample was tested where the same electrical
treatment was performed, and then, the sample was deformed at room temperature
to failure (refer to Fig. B.7 in Appendix B). This result is compared to the room
temperature deformation sample (Sample 2). The micrographs for the electrically
treated deformed sample in orientations 1 and 2 are shown in Figs. C.14 and C.15,
respectively.
From a visual comparison of orientation 1, the grain size of the electrically
treated sample deformed at room temperature appears to have a smaller average
grain size with approximately an equal amount of twinning. From the statisti-
cal analysis, it is said that the means are not equivalent, but the variances are the
same. The average equivalent grain size of the stationary electrical test in orienta-
tion 1 is 5.43 µm with a standard deviation of 1.74 µm as compared to 6.73 µm
with a standard deviation of 2.29 µm for the deformation test of the as-received
material in orientation 1.
For orientation 2, a visual analysis suggests that grain size is slightly larger and
the grains have become more elongated. The statistical analysis concluded that
the means are not equal, but the variances are the same. The average equivalent
grain size of the stationary electrical test in orientation 2 is 6.78 µm with a stand-
ard deviation of 2.13 µm as compared to 6.15 µm with a standard deviation of
2.33 µm for the deformation test of the as-received material in orientation 2. The
variation in the mean and grain shape as compared to the as-received deformation
sample (Sample 2) may be a result of the additional strain that was imposed on the
material before failure (approximately 24 %) which is greater than that of Sample
2. This additional deformation could explain the greater amount of grain elonga-
tion and directional alignment shown in orientation 2 (Fig. C.15). As a result of
the grains in orientation 2 becoming elongated more, the grains in orientation 1
reduced in grain size to maintain approximately the same grain volume. Thus, this
additional amount of deformation could explain the observed micrograph differ-
ences despite the stationary electrical pre-treating not altering the grain size/shape
as described above.

8.3.4.2 EAF with Square Wave Current Application During

Deformation

To examine the influence on an applied electrical current during EAF testing with
a square wave current application, different electrical conditions were tested and
analyzed. This section varies from the prior section in that the electrical square
wave is applied during deformation.
The test summarized in this section applied 800 A at intervals of 60 s for the
duration of 0.5 s, and this application was continued until the specimen fractured
(refer to Fig. B.3 of Appendix B). In addition to examining the microstructure
at the fracture location (L1), this section examines a section at the middle of the
specimen (L2) and a section from the region near the specimen fillet (L3). This
was performed to determine whether a microstructure gradient is present as a
8.3  Microstructure Analysis Under Tensile Loading 203

result of non-uniform local strains present during EAF. The resultant micrographs
for orientation 1 are given in Figs. 8.23, 8.24, and 8.25 for L3 to L1, respectively.
From visual examination, there is a significant change from location L3 to L1
where the grain size decreases and is smallest at the fracture location (L1). This
is in agreement with the prior results where the grain size decreases with greater
amounts of strain. This gradient in microstructure is observed here as the amount
of local strain increases from L3 to L1, and in this orientation, the grain decreases
in overall diameter.
The average grain size for the EAF (800 A; 0.5–60 s) test in orientation 1 at
L3 is 5.80 µm with a standard deviation of 1.96 µm. For the same test at L2, the
average grain size is 5.66 µm with a standard deviation of 1.55 µm, and at L3, the
average grain size is 5.41 µm with a standard deviation of 1.74 µm. Although the

Fig. 8.23  Microstructure of
EAF (800 A; 0.5–60 s) test
(sample 16) in orientation
1 at L3

Fig. 8.24  Microstructure of
EAF (800 A; 0.5–60 s) test
(sample 15) in orientation
1 at L2
204 8  Microstructure and Phase Effects on EAF

Fig. 8.25  Microstructure of
EAF (800 A; 0.5–60 s) test
(sample 14) in orientation
1 at L1

two-sample t test equates these three average grain sizes, it is clearly visible that
the shape varies along the length of the test specimen. Again, this is physically due
to the thermal gradients present during EAF, which results in localized strains that
vary along the length as compared to uniform room temperature deformation. It is
also noted that there is a lower number of twin boundaries present in these micro-
graphs as compared to the room temperature deformation micrograph.
Upon comparing these microstructure results (orientation 1) to the test
deformed at room temperature in orientation 1, it is statistically observed that the
mean is not equivalent even when comparing the results at L1. This is mainly a
result of the difference in strain of these regions. In the test deformed at room tem-
perature, the material strain is uniform throughout its gauge section until localized
necking occurs, whereas the EAF test displays diffuse necking. Last, it is noted
that the variances were statistically equivalent to the test deformed at room tem-
perature (Sample 2) even though the grain size was not. Last, the EAF (800 sA;
0.5–60 s) test displayed a reduced amount of material twinning which may be a
result of the applied electrical current aiding in deformation by a temperature rise
(Joule heating) and/or by direct electrical effects.
A similar analysis was conducted for orientation 2, and the micrographs are
presented in Figs. C.16, C.17 and C.18 for the EAF (800 A; 0.5–60 s) test. From
a visual analysis, it is evident that a microstructure gradient exists within the test
sample for this orientation as well. This is seen where the grains become less equi-
axed from L3 toward L1 and the grain boundaries are more jagged indicating a
greater amount of strain. The average grain size for the EAF (800 A; 0.5–60 s)
test in orientation 2 at L3 is 5.7 µm with a standard deviation of 1.66 µm. For the
same test and orientation at L2, the average grain size is 5.69 µm with a standard
deviation of 1.79 µm, and L1 has a grain size of 6.96 µm with a standard devia-
tion of 2.40 µm. From the statistical analysis, the average grain size of L2 (Sample
15) and L3 (Sample 16) was equivalent, but L1 (Sample 14) was different from
both L2 and L3. However, the variances for L1–L3 were equal. Again, it is noted
8.3  Microstructure Analysis Under Tensile Loading 205

that this orientation had a reduced amount of twin boundaries as compared to the
micrographs of the room temperature deformation test.
To compare the EAF (800 A; 0.5–60 s) test to the room temperature deforma-
tion test (Sample 2) in orientation 2, it is statistically shown that the average grain
size is different at the fracture region (L1). Again, this is a result of the strain dif-
ference due to the room temperature deformation having uniform strain with local
necking and the EAF test having diffuse necking with non-uniform strain.

8.3.4.3 EAF with Continuous Wave Current Application During

Deformation

In this section, the microstructure of EAF tests with a continuous wave is com-
pared. Specifically, two current magnitudes (150 and 200 A) are examined, and the
results are given at L1, L2, and L3 for each orientation (mechanical response in
Fig. B.13 of Appendix B).
The micrographs are presented for the 150 A continuous test in orientation 1
at L3 to L1 in Figs. 8.26, 8.27, and 8.28, respectively. From a visual examination,
the grain size appears to decrease from the fillet region (L3) toward the fracture
region (L1). From the statistical analysis, it is also confirmed that the grain sizes
are different from L3, L2, and L1. For orientation 1, the average grain size at L3
is 7.49 µm with a standard deviation of 2.91 µm. For L2, the average grain size is
5.68 µm with a standard deviation of 1.87 µm, whereas L1 has an average grain
size of 3.82 µm and a standard deviation of 1.64 µm.
This microstructure gradient is suggested to be a result of the temperature
and deformation gradient along the specimen length. A summary of the tempera-
ture results for the 150 A test at L1, L2, and L3 with respect to time is shown
in Fig. 8.29. The temperature is greatest for L1 and is least for L3 as a result of
the heat transfer of the testing setup. For the 150 A continuous test, the stress

Fig. 8.26  Microstructure of
continuous EAF (150 A) test
(sample 13) in orientation
1 at L3
206 8  Microstructure and Phase Effects on EAF

Fig. 8.27  Microstructure of
continuous EAF (150 A) test
(sample 12) in orientation
1 at L2

Fig. 8.28  Microstructure of
continuous EAF (150 A) test
(sample 11) in orientation
1 at L1

and strain response of the test is suggestive of dynamic recrystallization due to

the decreasing stress response after a peak stress is reached. This type of behavior
has been noted in works performing elevated temperature forming of magnesium
[10, 11]. The difference is that the temperature rise is a result of Joule heating
within the material as compared to external heating in a chamber. Also, unlike
external heating methods, a temperature gradient is present along the length of the
specimen. Consequently, this allows for a variation in deformation behavior along
the specimen length which corresponds to the observed variation in microstruc-
ture. The localized deformation behavior is illustrated in Fig. 8.30 where there is
an extreme amount of localized deformation at L1. From a visual perspective, it
appears that L3 and L2 only have a small amount of deformation and the defor-
mation is concentrated at L1. Also, it appears that dynamic recrystallization has
8.3  Microstructure Analysis Under Tensile Loading 207

Fig. 8.29  Temperature
response for continuous
current application at
L1, L2, and L3

Fig. 8.30  Continuous EAF
tensile specimens tested at
150 and 200 A

occurred during forming at L1. This is a result of the higher temperature and addi-
tional strain which increases the internal energy of the lattice. Recrystallization is
a process where strain-free grains are nucleated and grain growth continues until
the boundaries are impinged. The orientation 1 micrograph at L1 is suggestive of
dynamic recrystallization due to a large quantity of smaller grains surrounded by a
few coarse grains.
For orientation 2 of the 150 A test, similar results are seen for the three loca-
tions. Again, L3 (Fig. C.19) and L2 (Fig. C.20) have larger grains with a small
amount of deformation, whereas L1 (Fig. C.21) displays a recrystallized structure.
From the statistical analysis, this is shown where Sample 11’s average grain size
is different from Sample 12 and Sample 13. The average grain size of Sample 13
(L3) and Sample 12 (L2) in orientation 2 is 6.16 µm with a standard deviation of
2.02 µm and 6.00 µm with a standard deviation of 2.31 µm, respectively. In com-
parison, Sample 11 (L1) in orientation 2 has an average grain size of 4.18 µm with
a standard deviation of 1.29 µm.
Since this micrograph orientation depicts the microstructure along the deforma-
tion axis, a non-recrystallized structure would be extremely elongated in this axis
(e.g, Fig. C.18). However, this is not the case and the grain structure is refined.
The micrographs are presented for the 200 A continuous test (Fig. B.13 of
Appendix B) in orientation 1 at L3 to L1 in Figs. C.22, C.23 and C.24, respec-
tively. The average grain size for Sample 8 (L3) and Sample 7 (L2) in orienta-
tion 1 is 5.65 µm with a standard deviation of 1.89 and 5.84 µm with a standard
208 8  Microstructure and Phase Effects on EAF

deviation of 2.18 µm. Sample 6 (L1) has an average grain size of 6.10 µm with a
standard deviation of 2.16 µm in orientation 1.
As seen, similar trends are seen from this set of micrographs as compared to the
150 A test (Samples 11–13). However, it is noted that the material has not recrys-
tallized at L1 as small equiaxed grains are not present. Although the temperature is
greater (Fig. 8.29), it is reached for a shorter period of time and there is less local-
ized deformation at L1 (Fig. 8.30). However, there appears to be regions where
the recrystallization process has started. This can be seen in the top left corner of
Fig. C.24; however, the material has not completely recrystallized. It is also noted
by the box containing an enlarged view. From the statistical analysis, the average
grain size and variance are all equivalent in this orientation for the 200 A test. This
is somewhat expected due to the low amount of strain imposed on the material
(approximately 12 %). This low amount of strain is a result of the test being ter-
minated by the testing control system as the lower stress limit set by the controller
was reached.
The results for orientation 2 of the 200 A tests at L3 to L1 are given in Fig. C.25,
C.26 and C.27, respectively. The average grain size for Sample 8 (L3) and Sample 7
(L2) in orientation 2 is 5.43 µm with a standard deviation of 1.67 and 6.06 µm with
a standard deviation of 1.91 µm, respectively. Sample 6 (L1) has an average grain
size of 6.12 µm with a standard deviation of 2.03 µm in orientation 2.
The results are similar to the 150 A test in orientation 2, but the micrograph at
L1 is not fully recrystallized. This is in agreement with the results presented for
orientation 1 at 200 A at L1. As shown in Fig. C.27, there are regions where the
material is beginning to recrystallize, but the material is not fully recrystallized as
seen in the 150 A test. This again is a result of the lower amount of strain imposed
in addition to being subject to elevated temperatures for a shorter duration.

8.3.4.4 EAF with Incremental Current Application (in Situ Current

Annealing)

This section compares two samples where both are conventionally formed at room
temperature to a given displacement (5.08 mm); however, one sample is supplied
a pulse of current (1,000 A) for 1 s prior to the sample’s force being unloaded. The
other sample is just unloaded. From mechanical analysis, it was seen that the sam-
ple with the applied current had a modified yield point after the sample was loaded
with force (Fig. B.20 of Appendix B). Thus, this section examines these two sam-
ples to determine whether microstructural alterations (grain size/shape) were the
cause of the reduced yield point.
The micrographs of the sample with only 5.08 mm of deformation are given
in Figs. C.28 and C.29 for orientation 1 and orientation 2, respectively. The aver-
age grain size in orientation 1 for Sample 9 is 5.63 µm with a standard deviation
of 2.09 µm, whereas the grain size for Sample 9 in orientation 2 is 5.54 µm with a
standard deviation of 1.74 µm. From a visual analysis, twin boundaries are present
in both micrographs with a greater amount present in orientation 2.
8.3  Microstructure Analysis Under Tensile Loading 209

The micrographs for the deformation test (5.08 mm) with a single pulse of
1000 A for 1 s are given in Figs. C.30 and C.31 for orientation 1 and orientation 2,
respectively. The average grain size of Sample 10 in orientation 1 is 5.66 µm with
a standard deviation of 1.63 µm, whereas orientation 2 has an average grain size of
5.51 µm with a standard deviation of 1.89 µm. As seen, the results are very simi-
lar to the deformed test (Sample 9) with no electrical current application (grain
size and amount of twinning). This is also statistically verified by the two-sample
t test where the grain size means are equivalent for both orientations. However,
the variances are only equivalent for orientation 2. Thus, it is suggestive that the
reduced yield point is a result of the interaction of heat generation/electroplasticity
within the material lattice and not an alteration of the material grain size/shape.
This interaction with the lattice is suggested to be a modification of the dislocation
density within the material.

8.3.5 Microstructure Analysis Conclusions

The main conclusions drawn from the analysis in this section are as follows:
• The as-received material had an average grain size that is in agreement with
literature. Also, this material was free from twin boundaries as a result of the
material being warm-rolled.
• The room temperature forming microstructure showed a deformed structure
(non-equiaxed grains) and the presence of a large number of twin bounda-
ries which were a result of the limited number of slip systems active at room
temperature.
• The stationary electrical square wave (electrical treatment) showed no indica-
tion of a direct electrical effect at the grain level. Consequently, the observed
variation in the mechanical response is suggested to be a result of the applied
current altering the dislocation density of the material.
• The square wave electrical tests all showed similar microstructures to the room
temperature forming test, but the amount of twinning appeared to be reduced.
This could be a consequence of the electrical current allowing for pinned dislo-
cations to be freed, thus reducing the necessity of twin boundary formation for
continued deformation. Also, there was a microstructure gradient present due
to the diffuse necking of the specimen (non-uniform strain). This non-uniform
strain is a result of the thermal gradient in the specimen due to forming setup
geometry. Overall for these tests, the grain size was affected by the amount of
deformation imposed on the material and the current magnitude and pulse dura-
tion did not appear to modify the grain size to any significant level. Thus, this
suggests that the microstructure is not affected by pulsed current, but the acting
mechanism (thermal/electroplastic) for force reduction is occurring only at the
atomic level (in the material’s lattice).
• The continuous EAF tests also displayed a microstructure gradient along the
specimen length due to non-uniform strain. Also, dynamic recrystallization was
210 8  Microstructure and Phase Effects on EAF

present at some locations and was dependent on the amount of strain, the tem-
perature, and time at the elevated temperature.
• The EAF applied in an incremental manner is in agreement with the square
wave electrical tests where the single pulse did not alter the grain size. As a
result, the observed mechanical variation in the yield point is suggested to be
from heat/electroplastic contributions on the dislocation density of the material.
Overall, this study showed that EAF with various electrical conditions can alter
the final microstructure, but it was more related to the local strain present as a
result of the Joule heating causing thermal distributions within the sample. Thus,
it is concluded that the observed bulk forming relations seen with EAF (reduced
flow stress/increased elongation) are more connected with the interaction of dislo-
cations which were not visible at the present scale examined. Although this section
only considered one initial grain size, it is expected that starting initial grain size
will influence the heat generated (Joule heating contribution) and thus the flow
stress reduction by thermal softening. A reduced grain size will create a greater
quantity of interactions between electrons and grain boundaries (i.e., greater elec-
trical resistance), thus creating regions of greater atomic vibration or more heat
generation.

References

1. Hall EO (1951) The deformation and ageing of mild steel: III discussion of results. Proc Phys
Soc B 64:747–753
2. Hansen N (2004) Hall-Petch relation and boundary strengthening. Scripta Mater 51:801–806
3. Meyers MA, Ashworth E (1982) A model for the effect of grain size on the yield stress of
metals. Phil Mag 46(5):737–759
4. Struers Application Notes, Metallographic preparation of stainless steel. http://www.strue
rs.com/resources/elements/12/101820/Application%20Notes%20Stainless%20Steel%20
English.pdf. Accessed 1 May
5. Salandro WA (2012) Thermo-mechanical modeling of the electrically-assisted manufacturing
(EAM) technique during open die forging. PhD dissertation, Clemson University
6. Siopis MS, Kinsey BL (2009) Experimental investigation of grain and specimen size effects
during electrical-assisted forming. In: ASME international manufacturing science & engi-
neering conference, MSEC2009-84137, p 6
7. Kalpakjian S (1997) Manufacturing processes for engineering materials, 3rd edn. Addison
Wesley Longman Inc, Menlo Park
8. NIST-SEMATECH (2012) Levene and Bartlett tests. Engineering Statistics Handbook
9. Jones JJ (2012) Flow behavior modeling and process control of electrically-assisted forming
(EAF) for sheet metals in Uniaxial Tension. PhD dissertation, Clemson University
10. Al-Samman T, Gottstein G (2008) Dynamic recrystallization during high temperature defor-
mation of magnesium. Mater Sci Eng 490:411–420
11. Ion SE, Humphreys FJ, White SH (1982) Dynamic recrystallization and the development
of microstructure during the high temperature deformation of magnesium. Acta Metall
30:1909–1919
Chapter 9
Tribological and Contact Area Effects

In order for the electroplastic effect to take place, the applied electricity must be
able to flow from the dies and through the workpiece. Because of this, the inter-
faces between the dies/workpiece are critical. The contact area effects and the
tribological effects on EAF will be examined in this chapter. Contact area varies
depending on the surface roughness of the specimen contact faces and the magni-
tude of the applied force, which will then vary the resistance of the system, hence
the temperature of the system [1–4]. Even at high forming loads, there will still be
some level of differences between the apparent contact area and the actual con-
tact area [5], and it must be determined at which point the differences in the con-
tact areas have a noticeable effect on the effectiveness of the EAF technique. In
addition to contact area, the tribological effects study investigates several common
metal forming lubricants and compares how well they perform when used with the
EAF technique.

9.1 Contact Area Effect on EAF Effectiveness

Within this chapter, the examination explores several aspects of the impact contact
area has on EAM effectiveness. In addition, unique compression test specimens
were fabricated to experimentally prove the impacts of contact area. The subsec-
tions on this work include: (a) an explanation of how the surface ground speci-
mens and enhanced asperity specimens were produced, (b) a post-forming EAF
roughness profile examination, (c) a description of the experimental setup and pro-
cedures, (d) a thermal analysis, (e) an explanation of the voltage-contact resistance
contact area model, (f) a mechanical analysis of EAF based on contact area, and
(g) conclusions on contact area effects.

© Springer International Publishing Switzerland 2015 211

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_9
212 9  Tribological and Contact Area Effects

9.1.1 Specimen Preparation (Surface Ground)

Specimens were prepared in two different ways. Stainless Steel 304 specimens
were cut from 6.35-mm- and 4.76-mm-diameter rods. Then, the overall lengths of
the specimens were sized using different surface grinding wheels for different spec-
imens to create specimens with specific surface roughnesses on their contact faces.
The grinding wheels used were 320, 600, and 800 grit. A Zygo New View 7200 was
used to quantify the different surface roughnesses produced by the grinding wheels.
To accomplish this, roughnesses were measured on both sides of each specimen,
and two locations were measured on each side. Further, at each location one hori-
zontal and two diagonal measurements were taken to remove any directional bias
from the measurements. Figure 9.1 shows an example of this measurement strategy.
Table 9.1 displays the different roughness profiles which were determined.
Magnified pictures of the surface ground specimens from the profilometer can
be seen in Fig. 9.2. Additionally, surface roughness profiles of the same specimens
are shown in Fig. 9.3.

9.1.2 Specimen Preparation (Enhanced Asperities)

In addition to using the different surface grinding wheels to produce different sur-
face roughness, it was quickly determined that the difference in the surface rough-
ness was not large enough to show any trends in the experimental thermal data.

0° measurement 45° measurement 135° measurement

Fig. 9.1  Surface roughness measurement strategy [6]. Measurements were taken at two loca-
tions on each side of the specimen. At each location, several roughness profiles were determined
along imaginary lines drawn at various angles (0°, 45°, and 135°)

Table 9.1  Wheel grit versus surface roughness relationship [6]

Wheel Grit Avg Ra (µm) Avg RMS (µm) Size − x (mm) Size − y
(mm)
320 0.668 0.909 2.83 2.12
600 0.539 0.729 2.83 2.12
800 0.217 0.288 2.83 2.12
As a higher grit grinding wheel is used, average roughness value is decreases (i.e., the surface
becomes smoother)
9.1  Contact Area Effect on EAF Effectiveness 213

320 Grit 600 Grit 800 Grit

Fig. 9.2  Initial Zygo surface profiles [6]. Shown are magnified images of the contact surfaces of the
Stainless Steel 304 test specimens when they were surface ground using different grit grinding wheels

320 Grit

600 Grit 800 Grit

Fig. 9.3  Zygo surface roughness measurements [6]. Shown are the height versus distance plots
of a particular area on the contact surface of each specimen prepared using a different grit grind-
ing wheel

Therefore, an EDM was used to machine “enhanced asperity” profiles on the faces
of the specimens, as shown below in Fig. 9.4. These “enhanced asperities” pro-
vided a significant difference in the contact area, so that experimental trends can
be easily recognized. Let it be known there were three different enhanced asperity
specimens which were made (i.e., Large CA, Med. CA, and Small CA), however,
the Large CA specimens were the only ones used in this research.
Equation (9.1) and Fig. 9.5 show the equation used to determine the apparent
contact area of a specimen with exaggerated asperities:

π 2 
CAapparent = D − 2 · (L1 · w) + 2 · (L2 · w)
4   (9.1)
+ 2 · (L3 · w) + 2 · (L4 · w) − 16 · w2
214 9  Tribological and Contact Area Effects

Small CA Med. CA Large CA

Fig. 9.4  Enhanced asperity specimens [6]. The profiles for the enhanced asperities were cut using
an EDM. There are three types of enhanced asperities (small, medium, and large contact area)

Fig. 9.5  Determination
of apparent contact area
for enhanced asperity
specimens [6]. This scheme
accompanies the equation
to determine the apparent
contact area of enhanced
asperity specimens

where D is the outer diameter of the specimen, L1 is the length of the first cut, L2
is the length of the second cut, L3 is the length of the third cut, L4 is the length of
the fourth cut, and w is the width of each cut.
Table  9.2 shows the designations (surface ground, Large CA, Med CA, and
Small CA) and the corresponding contact areas.

9.1.3 Post-forming EAF Roughness Examination

In this subsection, post-formed surface roughness values were determined for

4.76-mm-diameter specimens formed using conventional compression and EAF
tests (at current densities of 20 and 30 A/mm2). The same measuring strategy
as discussed in the previous subsection was used. The total displacement stroke
9.1  Contact Area Effect on EAF Effectiveness 215

Table 9.2  Surface designations and contact areas [6]

Surface Specimen Specimen Contact Area Apparent Contact area
description OD (in) OD (mm) area removed contact ratio (%)
(mm2) (mm2) area (mm2)
Surface 0.1875 4.76 17.81 0.00 17.81 100 
ground
Large CA 0.1875 4.76 17.81 8.65 9.16 51
Med CA 0.1875 4.76 17.81 10.86 6.96 39 
Small CA 0.1875 4.76 17.81 12.76 5.06 28 
Surface 0.25 6.35 31.67 0.00 31.67 100
ground
Large CA 0.25 6.35 31.67 12.04 19.63 62 
Med CA 0.25 6.35 31.67 18.41 13.26 42 
Small CA 0.25 6.35 31.67 25.49 6.18 20

Table 9.3  Average Ra-values for post-formed compression tests (conventional and EAF) [6]
Deformation level Ra-values
Conv CD20 CD30
L0 2.000 2.000 2.000
L1 0.782 0.858 0.903
L2 0.603 0.714 0.838
L3 0.318 0.421 0.658
L4 0.205 0.358 0.400
The average roughness values decreased as the deformation level increased. However, the
average roughness increased as the current density was increased

was divided into five different levels of displacement (L0 to L4), with L0 being
no deformation. The surface roughness values were determined at each step.
Descriptions of each deformation level are as follows:
L0 No deformation
L1 0.95 mm of deformation
L2 1.91 mm of deformation
L3 2.86 mm of deformation
L4 3.81 mm of deformation
Please note that all of the specimens used in this subsection were sandblasted
to an average Ra-value of 2 µm prior to machining the enhanced asperities and
testing. Table 9.3 displays the average Ra-values for the conventional compression
test and the two EAF tests at each deformation level.
This same data is shown graphically in Fig. 9.6. From the figure and the table,
one can see that the post-formed roughness is consistently higher for the EAF
tests compared to the conventional compression tests. Additionally, the EAF tests
216 9  Tribological and Contact Area Effects

2.0
1.8

Ra Roughness Value (µm)

1.6
1.4
Conv
1.2
CD20
1.0
CD30
0.8
0.6
0.4
0.2
0.0
L0 L1 L2 L3 L4
Deformation Level (L0, L1, L2, L3,L4)

Fig. 9.6  Post-forming surface roughness at different deformation levels [6]. The post-forming

roughness value is increased as the magnitude of applied electrical power is increased for each
respective deformation level

with the greater current density (30 A/mm2) consistently produced a rougher post-
formed surface. This helps to prove that the EAF technique can have detrimental
impacts on the tribological conditions compared to cold forging. These impacts
will be discussed later in this chapter.

9.1.4 Experimental Setup and Procedure

Figure  9.7 shows the experimental testing setup used for this investigation. An
Instron testing machine ran stationary electrical, conventional compression, and
EAF tests on surface ground parts and enhanced asperity parts (4.76-mm-diameter

Fig. 9.7  Experimental test
setup for contact area effects
investigation [6]. This testing
setup includes a multi-meter
that reads the voltage drop
across the top and bottom
dies at the locations where
they contact the test specimen
9.1  Contact Area Effect on EAF Effectiveness 217

specimens). The part temperature, the voltage across the height of the specimen,
the load/position data (when deforming), and Zygo images were all recorded.

9.1.5 Thermal Analysis of EAF Based on Contact Area

As part of the thermal analysis subsection, the method of obtaining thermal data,
along with the comparisons of that thermal data will be presented. The thermal
measuring strategy consisted of using a FLIR A40 M infrared thermal imaging cam-
era to record the specimen temperature profile for each test. Then, ThermaCAM
software was used to generate thermal data for specific locations on the part. In this
work, thermal data was generated for three locations on the workpiece:
1. A single-point measurement at the top die/specimen interface (the top die is
stationary on the Instron machine).
2. A single-point measurement in the middle of the specimen.
3. A maximum of the whole specimen/die area (this is to catch any instantaneous
sparking phenomenon during the first initiation of the electricity)
Figure  9.8 displays the locations of the three areas of thermal data gathering for
this work. It was decided that the thermal data used for comparison in this work
should be taken from the center of the specimen, since the hottest area of the spec-
imen quickly and consistently becomes the center through testing, which will be
shown in the following few figures.
Shown below in Figs. 9.9, 9.10, and 9.11 are heating and cooling sequences
during an EAF compression test for specimens with enhanced asperities in

Fig. 9.8  Locations of thermal data gathering at the specimen and die interfaces [6]. The thermal
data was taken at points at the center of the specimen, at the top of the specimen, and the maxi-
mum temperature was recorded for the whole area surrounding the specimen and the dies. The
temperature at the center of the specimen was used
218 9  Tribological and Contact Area Effects

Fig. 9.9  Heating and cooling sequences for enhanced asperities at the top of the specimen [6].
In the case of the enhanced asperities being at the top of the specimen, the most heat is generated
at the top of the specimen very briefly, and then this heat moves to the center of the specimen,
ultimately making this location the hottest

Fig. 9.10  Heating and cooling sequences for enhanced asperities at the bottom of the specimen
[6]. In the case of the enhanced asperities being at the bottom of the specimen, the most heat is
generated at the bottom of the specimen very briefly, and then this heat moves to the center of the
specimen, ultimately making this location the hottest
9.1  Contact Area Effect on EAF Effectiveness 219

Fig. 9.11  Heating and cooling sequences for enhanced asperities at the top and bottom of the
specimens [6]. In the case of the enhanced asperities being at the top and bottom of the speci-
men, the heat is generated at the top and bottom of the specimen very briefly, and then this heat
moves to the center of the specimen, ultimately making this location the hottest

particular locations. Figure 9.9 has the enhanced asperities on the top face, and
from the thermal profiles, this is the hottest area at the beginning of the test (notice
the white color signifying the hottest temperature). Figure 9.10 has the enhanced
asperities at the bottom face, and this area is the hottest at the beginning of the
test. Figure 9.11 has the enhanced asperities at the top and bottom faces and both
of these areas are approximately equally the hottest locations at the beginning of
the test. These three experiments show that, where there is a significant change
in the contact area (larger than the micro-level), the temperature will be affected.
However, this change in temperature is only noticed briefly at the beginning of the
test, and then the center of the specimen becomes the hottest location for all three
tests. The heat transfer during EAF can be affected by many different variables
(i.e., different forms of heat transfer, die volume, etc.) and the specific effects of
these variables will be discussed in Chap. 3.
Figure  9.12 shows stationary electrical thermal profiles for specimens which
were surface ground using a 320-grit wheel to an average roughness value of
0.539 µm. It can be seen that, as the static load is increased, the temperature also
decreases by a small extent. Please note that all of these static loads were within
the elastic region of the specimens, so the apparent contact between the dies/speci-
men did not change.
Figure  9.13 displays the thermal profiles for stationary electrical tests run on
specimens with a “Large CA” surface roughness. From the figure, the tempera-
tures at static loads of 1,335 and 2,000 N are much higher than the temperatures at
the other, heavier loads, which were into the plastic regime of the asperity portion
of the specimens. Of note is that the 1,335 and 2,000 N loads are still within the
elastic region of the “Large CA” region, but the other three loads are toward the
220 9  Tribological and Contact Area Effects

350
1335N
2000N
300

250

Temperature (°C) 200

2670N
3338N
4000N
150

100

50

0
0 5 10 15
Time (s)

Fig. 9.12  Temperature versus time plots for stationary electrical tests at various static loads
(320 grit) [6]. The temperatures for each of the stationary electrical tests decrease as the static
loads applied to the part while the test is underway increase

600
1335N
500
2000N
Temperature ( ° C)

400

300
2670N
3338N
4000N
200

100

0
0 5 10 15
Time (s)

Fig. 9.13  Temperature versus time plots for stationary electrical tests at various static loads (Large
CA-enhanced asperity parts) [6]. The temperatures of the stationary electrical tests decreased as the
static load increased. Additionally, once the static load had surpassed the elastic–plastic threshold,
the difference in temperature between the tests was unaffected by the different loads

plastic regime. With this said, a clear difference between the elastic and plastic
regimes can be noticed. Specifically, when the parts are still in the elastic regime,
the temperature and applied static load have an inverse relationship, where a heav-
ier elastic load causes a cooler temperature than a lighter elastic load. However,
when the parts are within the plastic regime, the specimen temperature is relatively
unaffected due to the increase or decrease in the static load.
Now data will be taken from Figs. 9.12 and 9.13 to be able to compare tem-
peratures of the 4.76-mm-diameter specimens at different discrete times over the
different static loads. Table 9.4 shows the specimen temperatures at 5-second inter-
vals, as well as the percent of the baseline temperature for all of the static loads run
9.1  Contact Area Effect on EAF Effectiveness 221

Table 9.4  Stationary Force Time Temperature ΔT (% of baseline

electrical test temperatures (N) (s) (°C) Temp at 1,335 N)
(4.76-mm-­diameter surface
1,335 5 288.1 100.0
ground specimens) [6]
10 290.0 100.0
15 294.6 100.0
2,000 5 268.4 93.2
10 267.3 92.2
15 272.8 92.6
2,670 5 261.3 90.7
10 265.9 91.7
15 270.0 91.7
3,338 5 257.4 89.4
10 256.6 88.5
15 262.8 89.2
4,000 5 233.7 81.1
10 239.5 82.6
15 244.1 82.9

Fig. 9.14  Stationary 100
electrical temperatures
(4.76-mm-diameter surface All loads in the
% of Baseline Temperature

95 elastic region.
ground specimens) [6]. There
is a clear linear decreasing
at 1335 N Load

trend for the percent of the 90

baseline temperature and
static load because the static 85
loads tested did not surpass
the elastic–plastic threshold After 5s
for these particular specimens 80 After 10s
After 15s

75
0 1000 2000 3000 4000 5000
Static Load (N)

on the specimens with a 320-grit surface ground roughness. Additionally, Fig. 9.14

shows this data plotted as a “percent of the baseline temperature at 1,335 N” versus
the static load after 5, 10, and 15 s of testing. The figure illustrates a clear declining
trend in the percent of the baseline temperature as the static load is increased.
Table 9.5 shows the specimen temperatures at 5-second intervals, as well as the
percent of the baseline temperature for all of the static loads run on the 4.76-mm-
diameter specimens with a “Large CA” roughness profile. Additionally, Fig. 9.15
shows this data plotted as a “percent of the baseline temperature at 1,335 N” ver-
sus the static load after 5, 10, and 15 s of testing. Unlike the previous figure, this
figure shows a very steep declining trend in the percent of the baseline temperature
with increased static load for only the first two static loads (note these two loads
were in the elastic regime). Then, there is a level relationship between the percent
of the baseline temperature and the load for the remaining three static loads. The
222 9  Tribological and Contact Area Effects

Table 9.5  Stationary Force Time Temperature ΔT (% of baseline

electrical test temperatures (N) (s) (°C) Temp at 1,335 N)
(4.76-mm-diameter Large CA
1,335 5 496.7 100.0
specimens) [6]
10 529.0 100.0
15 536.0 100.0
2,000 5 410.8 82.7
10 430.2 81.3
15 440.3 82.1
2,670 5 312.5 62.9
10 314.6 59.5
15 331.3 61.8
3,338 5 313.3 63.1
10 324.9 61.4
15 325.0 60.6
4,000 5 303.4 61.1
10 303.3 57.3
15 314.1 58.6

Fig. 9.15  Stationary 100
electrical temperatures 95 Elastic region Plastic region
(4.76-mm-diameter Large
% of Baseline Temperature

90
CA specimens) [6]. There
is a clear decreasing trend 85
at 1335 N Load

between the temperature 80

and the static load, until the
75
elastic–plastic threshold
is reached. Then, the 70
After 5s
temperature does not change 65 After 10s
with an increase in static load After 15s
60

55
0 1000 2000 3000 4000 5000
Static Load (N)

three loads which show a leveling relationship are within the plastic regime of the
“Large CA” region of the specimens.
Table 9.6 shows the specimen temperatures at 5-second intervals, as well as the
percent of the baseline temperature for all of the static loads run on the 6.35-mm-
diameter specimens with a roughness profile created using a 320-grit grinding
wheel. Additionally, Fig. 9.16 shows this data plotted as a “percent of the base-
line temperature at 1,335 N” versus the static load after 5, 10, and 15 s of testing.
Like the 4.76-mm-diameter specimens with the same surface roughness, there is
a clear declining trend in the percent of the baseline temperature as the static load
increases. This trend is apparent at all static loads, and all of the static loads are
within the elastic region of the part.
9.1  Contact Area Effect on EAF Effectiveness 223

Table 9.6  Stationary Force Time Temperature ΔT (% of baseline

electrical test temperatures (N) (s) (°C) Temp at 2,376 N)
(6.35-mm-diameter surface
2,376 5 418.7 100.0
ground specimens) [6]
10 507.7 100.0
15 541.8 100.0
3,564 5 393.0 93.9
10 472.5 93.1
15 505.2 93.2
4,753 5 386.6 92.3
10 464.2 91.4
15 496.7 91.7
5,941 5 350.6 83.7
10 427.7 84.3
15 463.1 85.5
7,129 5 331.9 79.3
10 405.1 79.8
15 438.3 80.9

Fig. 9.16  Stationary 100
electrical temperatures
(6.35-mm-diameter surface All loads in the
% of Baseline Temperature

95 elastic region.
ground specimens) [6].
at 2376 N Load

The linear–inverse relation

between temperature and 90
static load is also apparent
for the larger specimens due 85
to all of the static loads only
After 5s
being in the elastic region
80 After 10s
After 15s

75
0 2000 4000 6000 8000
Static Load (N)

9.1.6 Voltage–Resistance Contact Area Model

The voltage (in mV) was recorded for each of the stationary electrical tests using
a digital multi-meter, as shown in Sect. 9.1.4. Since the voltage is related to the
actual contact area, the voltage measurements can provide a dynamic confirma-
tion of the asperity crushing due to an increase in the static load. Figure 9.17
below depicts the voltage measurements obtained during stationary electrical tests
under various static loads for 4.76-mm-diameter surface ground specimens and
for 4.76-mm-diameter “Large CA” specimens. The “Large CA” specimens show
a notable decrease in the measured voltage at the first two static loads, and then
show the same level effect as the surface ground specimens (although the “Large
CA” specimens still have a higher voltage value). The higher voltage value is to
224 9  Tribological and Contact Area Effects

Fig. 9.17  Voltage versus 450

static load stationary 400
electrical tests (Large
CA and surface ground 350
4.76-mm-diameter 300

Voltage (mV)
specimens) [6]. There is more
250
of a change in the voltage
of the Large CA specimens 200
at the lower static loads in 150
comparison with the surface
100
ground specimens Surface ground (4.76mm)

50 Large CA (4.76mm)

0
1000 1500 2000 2500 3000 3500 4000 4500
Static Load (N)

Fig. 9.18  Voltage versus 360

stationary electrical
340
tests (surface ground
All loads in the
6.35-mm-diameter 320
elastic region.
specimens) [6]. The voltage
Voltage (mV)

300
in the 6.35-mm-diameter
surface ground specimens 280
shows a decreasing linear
trend with the increase in 260
static load 240

220 220 Surface ground (6.35mm)

200
1500 2500 3500 4500 5500 6500 7500
Static Load (N)

be expected since there is less contact area (because the EDM profiled it out) and
there is the same electrical current running through these parts as is with the sur-
face ground parts.
Figure  9.18 displays the voltage versus static load relationship for the
6.35-mm-diameter specimens. There is a very steep decrease in the voltage as the
static load is increased, which correlates to the same decrease in specimen tem-
perature as the static load was increased, in Fig. 9.16 above.
Now, utilizing the voltage measurements, the actual contact area can be pre-
dicted. The basic relation for electrical power can be described as Eq. (9.2):

P = IV = I 2 R (9.2)
where P is electrical power, I is current, V is voltage, and R is the resistance. Then,
the voltage can be written as Eq. (9.3):
r · Lpart
V = IR = I · (9.3)
Acontact
9.1  Contact Area Effect on EAF Effectiveness 225

where r is the electrical resistivity, Lpart is the length of the workpiece, and Acontact
is the actual contact area at the specimen and die interface. The relation for the
actual contact area can be re-written as Eq. (9.4):
 
rpart · Lpart
Acontact = V
(9.4)
I
This relation describes electrical resistance, but it does so assuming that the work-
piece and the top and bottom dies are essentially all one piece. However, there is
also some type of contact resistance between the specimen and dies since there is
a difference in the parts and in the die/workpiece materials. The total resistance
in the system (including electrical resistance and contact resistance) is shown as
Eq. (9.5):
Rtotal = Relectrical + Rcontact (9.5)
The electrical contact resistance was described above, but the contact resistance
can be written as Eq. (9.6):
rpart + rdie rpart + rdie
Rcontact = + (9.6)
4 · nasp · aasp D
where rpart is the resistivity of the part, rdie is the resistivity of the die, n is the
number of asperities, a is the radius of the contact points, and D is the diameter
of the region where the asperities are located. When including both the electrical
resistance and the contact resistance, the equation for the total resistance can be
written as Eq. (9.7):
rpart · L rpart + rdie rpart + rdie
Rtotal = Relectrical + Rcontact = + + (9.7)
Acontact 4 · nasp · aasp D
The final equation to be used to back-solve for the contact area, based on the
measured voltage, is Eq. (9.8):
 
rpart · Lpart
Acontact =  V rpart +rdie rpart +rdie
 (9.8)
I − 4·nasp ·aasp − Dpart

For 4.76-mm-diameter “Large CA” specimens in which stationary electrical tests

were run, Table 9.7 summarizes the recorded voltages, the apparent contact area
(i.e., the area calculated using Eq. (9.1), the contact areas calculated using Eq. (9.4)
(without the contact resistance), and the contact area calculated using Eq. (9.8)
(with the contact resistance). Figure 9.19 displays the contact areas with and with-
out the contact resistance accounted at the various static loads, along with a dashed
line representing the apparent contact area created from the EDM process dur-
ing the fabrication of the specimens. Please note that the area calculated with the
contact resistance is higher than the actual apparent area. This is possibly due to
the fact that the loads are in the plastic region and the total specimen height and
asperity height/width (i.e., cross-sectional area) has changed in this region. Of all
226 9  Tribological and Contact Area Effects

Table 9.7  Actual contact area values based on voltage measurements from stationary electrical
tests [6]
Test Specimen Preload Voltage Area (w/o A (w/ Apparent
# conditions (N) (mV) contact contact area (mm2)
resistance) resistance)
mm2 mm2
11 Large CA 4,005 204 9.03 15.51 9.15
12 (4.76 mm) 3,338 210 8.77 14.77 9.15
at 356 Amps
13 3,338 215 8.57 14.20 9.15
14 2,670 219 8.41 13.77 9.15
15 2,003 290 6.35 9.00 9.15
16 2,003 276 6.68 9.66 9.15
17 1,335 390 4.72 6.05 9.15

18
Calculated Contact Area (mm 2)

16

14

12

10

6
Contact Area (w/out contact resistance)
4
Contact Area (w/ contact resistance)
2 Apparent Area

0
1000 1500 2000 2500 3000 3500 4000 4500
Static Load (N)

Fig. 9.19  Contact areas calculated from the voltage measurements (4.76-mm-diameter Large

CA specimens) [6]. The calculated contact area without the contact resistance is much closer
to the actual contact area than the calculated contact area which considered contact resistance.
However, the plastic deformation was unaccounted for in the model and could affect the results

the research papers reviewed in which the actual contact area is predicted, all of the
loadings were localized and were within the elastic regime, not the plastic regime.
As the material enters the plastic regime, the overall dimensions of the potential con-
tact areas are continuously changing with the deformation and new contact surfaces
are continuously forming with differing roughness profiles, thus making the predic-
tion of actual contact areas under plastic loadings much more difficult.

9.1.7 Mechanical Analysis of EAF Based on Contact Area

Within this subsection, stress–strain plots are shown for 4.76-mm-diameter speci-
mens with surface ground and “Large CA” roughnesses, as well as for 6.35-mm-
diameter specimens with surface ground and “Large CA” roughnesses. In addition
9.1  Contact Area Effect on EAF Effectiveness 227

Fig. 9.20  Stress–strain 1400
4.76mm, Conv.
profiles (4.76-mm-diameter (x2 tests)
surface ground specimens) 1200
[6]. The stress–strain profiles
1000

True Stress (Mpa)

show that the EAF tests 4.76mm, EAF
produced a lower overall 800
(x2 tests)
stress–strain profile
600

400

200

0
0 0.2 0.4 0.6 0.8
True Strain

Fig. 9.21  Mechanical power
profiles (4.76-mm-diameter
surface ground specimens)
[6]. The mechanical power
4.76mm, Conv.
profile for the EAF test is
lower than the power profile
for the conventional test
4.76mm, EAF

to the stress–strain profiles, mechanical power profiles are also plotted, and these
will then be used to calculate the EEC profiles for each of the specimens described
above. Finally, the EEC profiles can be compared and analyzed.
Figures  9.20 and 9.21 show the stress–strain profiles and mechanical power
profiles for conventional compression and EAF tests at CD = 20 A/mm2 on
4.76-mm-diameter specimens with a surface ground roughness. Two stress–strain
profiles for each condition were shown in Fig. 9.20 to prove the repeatability of
the experiments. The EAF stress–strain profile is lower than the conventional com-
pression profile, thus signifying that the electrical parameters which were chosen
are in the electroplastic region for this material.
Figures  9.22 and 9.23 show the stress–strain profiles and mechanical power
profiles for conventional compression and EAF tests at 20 A/mm2 on the
6.35-mm-diameter specimens with a surface ground roughness. Again, duplicates
were shown for the stress–strain profiles to signify the repeatability. The EAF
stress–strain profile is much lower than the conventional compression profile,
when compared to the 4.76-mm-diameter specimen stress–strain profiles above.
228 9  Tribological and Contact Area Effects

Fig. 9.22  Stress–strain
profiles (6.35-mm-diameter
surface ground specimens) 6.35mm, Conv.
[6]. The stress–strain profiles (x2 tests)

for the EAF tests are much

lower than the conventional
tests. Additionally, the
difference between the
6.35mm, EAF
conventional and EAF tests
(x2 tests)
increases as the strain is
increased

Fig. 9.23  Mechanical power
profiles (6.35-mm-diameter
surface ground specimens)
[6]. The difference in the 6.35mm, Conv.
mechanical power profiles
of the conventional and EAF
tests increases as the tests
progress

6.35mm, EAF

Fig. 9.24  Stress–strain
profiles (4.76-mm-diameter
Large CA specimens) [6]. 4.76mm, Conv.
The stress–strain profile for
the EAF test is slightly lower
than the conventional stress– 4.76mm, EAF

strain profile

Figures  9.24 and 9.25 show the stress–strain profiles and mechanical power
profiles for conventional compression and EAF tests at 20 A/mm2 run on the
4.76-mm-diameter specimens with a “Large CA” surface roughness.
9.1  Contact Area Effect on EAF Effectiveness 229

Fig. 9.25  Mechanical power
profiles (4.76-mm-diameter
Large CA specimens) [6]. 4.76mm, Large CA, Conv.
The mechanical power profile 4.76mm, Large CA, EAF
for the EAF test is slightly
lower than the conventional
test, which is consistent with
the surface ground specimens
in Fig. 9.21

Fig. 9.26  Stress–strain
6.35mm, Conv.
profiles (6.35-mm-diameter
Large CA specimens)
[6]. The difference in the
stress–strain profiles of the
6.35mm, EAF
conventional and EAF tests
increases, as it did with the
surface ground specimens

Fig. 9.27  Mechanical power
profiles (6.35-mm-diameter
Large CA specimens)
[6]. The difference in the
6.35mm, Large CA, Conv.
mechanical power profiles is
the same as the difference in
the stress–strain profiles for
these particular specimens

6.35mm, Large CA, EAF

Figures  9.26 and 9.27 show the stress–strain profiles and mechanical power
profiles for conventional compression and EAF tests at 20 A/mm2 run on the
6.35-mm-diameter specimens with a “Large CA” surface roughness.
230 9  Tribological and Contact Area Effects

Fig. 9.28  EEC
profiles (4.76-mm- and EEC (6.35mm)
6.35-mm-diameter surface
ground specimens) [6].
The EEC profiles for both
specimen sizes show an
increase and then a decrease
toward the bottom of the
EEC (4.76mm)
test. These profiles are not
consistent with other material
EEC profiles because the
compression speed used to
generate these profiles was
much slower

Fig. 9.29  EEC
profiles (4.76-mm- and EEC (6.35mm, Large CA)
6.35-mm-diameter Large
CA specimens) [6]. The
EEC profile for the larger
specimen was much higher
compared to the smaller
specimen
EEC (4.76mm, Large CA)

Figures 9.28 and 9.29 display the EEC profiles for 4.76-mm- and 6.35-mm-diam-
eter specimens that have been prepared by surface grinding and by enhanced asper-
ities, respectively. Of note is that the general EEC profiles for the surface ground
and enhanced asperities surface conditions are very similar for both specimen sizes
(when comparing the top and bottom EEC profile in one figure to the same profile
in the second figure). With this being said, it can be concluded that that the overall
starting contact area is insignificant to the effectiveness of the EAF technique, unless
only the elastic region is considered. Furthermore, when considering the contact area
for modeling EAF in manufacturing processes, it could be neglected since the part
will primarily be in the plastic region (since it will be deformed).

9.1.8 EAF/Contact Area Conclusions

Several of the contact area references found, focused on trying to predict the
roughness or bearing area of the material [7–12]. Of note is that all of these papers
concentrated on elastic deformation regime of the material, under low localized
9.1  Contact Area Effect on EAF Effectiveness 231

loads. This examination focused on the surface roughness under both elastic and
plastic loadings, over the entire apparent area of the specimen. The following are a
list of conclusions about the EAF/contact area research:
• The post-formed surface roughnesses at the die/workpiece interfaces become
worse as a higher magnitude of applied electricity is used.
• There is a clear trend in that, as the applied static load is increased for a given
amount of current, the overall temperature of the specimen will decrease.
However, this was found to only be valid within the elastic regime on the
specimen. For the enhanced asperity specimens, where the plastic regime was
reached in the asperity region, the temperature from the stationary electrical
tests did not change.
• The EEC profiles from the same EAF tests run on surface ground specimens
and specimens with enhanced asperities were very similar, thus indicating that
the contact area does not have much of an effect on effectiveness of EAF out-
side of the elastic region. Hence, this is possibly why there is such a focus on
contact area for electrical connectors, where all components are in the elastic
region.
• When attempting to model the EAF technique for a forging process, the contact
area is insignificant because the manufactured part will always be in the plastic
region since it is being deformed.

9.2 Tribological Effect on EAF Effectiveness

This section will contain an exploration into the effects of applied electricity on
tribology within an EA-Forging process [13, 14]. Specifically, (a) the general
effects of the electricity on tribological conditions will be explained, (b) the exper-
imental setup and procedure for this section will be described, (c) the method by
which the friction calibration curves were determined will be given, (d) test pro-
cedures will be stated, (e) candidate metal forming lubricants will be introduced,
(f) experimental results and discussion will be detailed, (g) an evaluation of the
lubricants based on the reduction in the forming load will be noted, (h) tempera-
ture measurements will be provided, and (i) the section will be summarized with
conclusions.
Because most metal forming operations incorporate some type of lubricant, it
is necessary to understand how different lubricant compositions/types react while
used in an EAF compression test. Although there is no true record that traces
lubricant use to a particular time period, it can be assumed that lubricants were
being used to assist metal deformation as early as fifth century B.C. Until the
mid-1940s, lubricants remained relatively unrefined, consisting of seed or animal
oils/greases. Around the later part of the twentieth century, increasing demands
from the aerospace, chemical, and electronic industries led to the development of
new exotic metals with which ordinary lubricants were not sufficient. These new
232 9  Tribological and Contact Area Effects

metals, along with increased production and heavier material reductions, signifi-
cantly increased the severity of the conditions on the lubricant. Hence, much effort
has supported tribology research in the last half century [15]. One of the directions
in lubricant research is to replace the widely used zinc–phosphate-based lubricant
which, despite its excellent qualities, presents the disadvantage of being polluting.
Depending on the type of lubricant, type of manufacturing process, and the
severity of the forming, metal forming lubricants can serve different purposes.
Overall, the main functions of any metal forming lubricant are to reduce or con-
trol the friction at the workpiece/die interfaces and control the temperature of the
workpiece. Different lubricant compositions use various lubrication regimes to
accomplish this feat. By controlling friction, a manufacturer can achieve reduced
tool/die wear and improved surface finish. Additionally, by controlling the temper-
ature of the workpiece, the lubricant can produce an overall cooler or hotter part
(i.e., the lubricant can insulate heat to keep the part hot during forming, or it can
help to keep the part cool by transferring the heat rapidly). Some form of lubrica-
tion is used in almost every metal deformation process, and the continuous effort
to produce better state-of-the-art lubricants confirm the critical nature of their role
in the metal deformation process.
Lubricants are used in almost all metal forming operations and, before EAF can
be used to potentially replace a widely used forming technique such as hot working,
IF, or TWB’s, its impact and effect on metalworking lubricants must first be evalu-
ated. The objective of this work was to investigate the influence of the presence of
electricity on the tribological conditions at the die/workpiece interface. Ring com-
pression tribo-tests (explained in Sect. 9.2.2) are conducted and the performance of
several lubricants is studied. By combining the experiments and finite element simu-
lation results, friction coefficients and field effects on friction can be estimated, and
possible lubricant candidates to be used in EAF processes identified.

9.2.1 Effects of Electricity on Tribological Conditions

Friction is present at the workpiece/die interfaces, thus lubricants play an impor-

tant role in forming. They reduce the frictional forces, thus the power consump-
tion, and also improve the formability of the material and surface finish of the final
part. In conventional forming, an effective lubricant exhibits a high level of adhe-
sive strength in order to follow the surface expansion without breaking down. In
addition, the lubricant must withstand high pressures (up to 2,500 MPa in con-
ventional forming), and the interface temperature may reach 200 °C [16, 17]. But
when electricity is applied through the dies and workpiece, it may influence the
tribological conditions. Careful attention to the effects of the electricity is needed
when selecting the lubricants for EAF due to their complex interaction (Fig. 9.30).
One effect of the electricity is the decrease in forming load, thus the decrease of
the interface pressure, resulting in a better lubrication regime. At the same time, a
higher temperature at the workpiece/interface decreases the lubricant viscosity and
9.2  Tribological Effect on EAF Effectiveness 233

Effects of electricity in
EAF

Decrease flow stress Increase formability Localized heating

Decrease forming load Larger strains achieved

and interface pressure

Better tribological More severe tribological

conditions conditions

Fig. 9.30  Tribological effects of EAF [13, 14]. As the electricity is applied to the manufacturing
process, it increases the formability of the part; however, it can also have negative effects on the
tribological effects at the die/workpiece interfaces

favors evaporation and chemical decomposition of the lubricant. Thus, the reduced
quantity of lubricant and possible modified chemical attributes are not able to con-
tinue providing sufficient separation. Metal-to-metal contact between the asperi-
ties of the dies and workpiece occurs, resulting in higher friction, die wear, and
poor surface quality. In conclusion, one requirement for a good metal forming
lubricant for EAF is the capacity of retaining its lubricity at elevated temperatures
and in the presence of electric current fields. The objective of this subsection is to
present the challenges brought by this process, and to screen the performance of a
few selected lubricants through electrically assisted ring compression tests.

9.2.2 Experimental Setup and Procedure (Ring Tribo-Tests)

To evaluate the performance of the three selected lubricants, the ring compression
tribo-test was used. This is a relatively simple test that can mimic the medium-to-
severe deformation occurring in an actual forming operation. The testing setup for
these experiments is shown in Fig. 9.31. The insulated dies, consisting of a combi-
nation of A2 steel and reinforced plastics, are securely gripped by the top/bottom
platens of the 250-kN-capacity Instron hydraulic testing machine. Electrical cables
are fastened to the top and bottom dies, and the insulation ensures electricity flows
through the dies/workpiece in a closed loop. The electricity is applied continu-
ously from the beginning to the end of each test.
The ring compression specimens used in this work were manufactured from a
rod of 304 Stainless Steel. The rod was received as precision ground to a diameter
of 4.763 mm, and cut to a height of 2.10 mm using a wire EDM. Then, a drilling
operation was performed using a CNC lathe to create the 2.38 mm inner diameter.
Careful measures were taken to ensure the repeatability and flatness of all speci-
men surfaces in contact with the dies.
234 9  Tribological and Contact Area Effects

Top Platen

Insulation

Top Die
Specimen
Bottom Die
Insulation
Electric
Cable
Bottom
Platen

Fig. 9.31  Test setup for EAF tribological evaluation [13, 14]. This test setup included insulated
top and bottom dies that compressed ring compression samples, while thermal and mechanical
data were collected

Initial specimen

Deformed specimen –Low friction

Deformed specimen –High friction

Fig. 9.32  Effect of friction on the deformation pattern [13, 14]. If there is low friction, the inside
diameter of the ring compression specimen will increase throughout the test. Conversely, if there
is high friction, the inside diameter will decrease over the duration of the test

The purpose of the ring compression test is to establish a correlation between

the friction conditions at the workpiece/die interfaces and the ratio of the reduction
of the inner diameter, and use it to quantify the existing friction conditions. The
internal diameter of the ring is very sensitive to the friction conditions. If the fric-
tion is low, than the internal diameter increases, but if the friction is high, the inner
diameter reduces (Fig. 9.32). The shear friction for the lubricant can be estimated
when the inner diameter reduction corresponding to a certain height reduction is
measured. This value is superimposed on friction calibration curves determined
using finite element analysis (FEA) [18, 19].
9.2  Tribological Effect on EAF Effectiveness 235

9.2.3 Determining Friction Calibration Curves

For this study, the shear friction is considered, which is more applicable for bulk
forming than Coulomb’s Law, which uses the coefficient of friction. The friction
shear law is written as Eq. (9.9):
√
τ = mσ̄ 3 = mkshear (9.9)

where τ is the shear stress, σ̄ is the flow stress, kshear is the shear strength, and m
is the friction factor, which ranges between 0 and 1. Friction calibration curves,
determined through FEA, are used to estimate the friction factor of the tested
lubricant. The simulations were run for 304 Stainless Steel with initial dimensions
as specified earlier, for different friction factors, m.
The electric energy applied does not go exclusively into resistive heating, but
also into aiding deformation. The existing finite element software cannot take into
consideration the complex effect of the electricity on the deformation process or
on the lubricant, but it can implement the thermal effect, which results mainly in
lowering the flow stress. The electrical current applied contributes to a temperature
rise in the billet proportional to the current density and application time. Since the
temperature distribution in the workpiece and dies cannot be accurately predicted
analytically yet, preliminary electrically assisted tests were carried out and tempera-
tures were measured. In order to replicate the conditions of the real test, non-iso-
thermal finite element simulations were carried out by considering heated dies and
workpiece at a temperature observed in the EAF test. Two types of tests were simu-
lated for different friction coefficients: (i) compression at room temperature and (ii)
compression at elevated temperature, specifically 200 and 300 °C for the dies and
ring specimen, respectively. All the other experimental conditions, such as the mate-
rial, dimensions, and compression speed were incorporated into the simulations.
Although the temperature was observed to have a significant influence on the form-
ing load, due to decreasing the material flow stress, the effect on the inner diameter
reduction was insignificant. The comparison between the two conditions is shown
in Fig. 9.33. In conclusion, for ease of comparison, the calibration curves used fur-
ther for superimposing the experimental data will be the set determined at 20 °C.

9.2.4 Testing Procedures

For this work, the effectiveness and compatibility of three metal forming lubri-
cants with EAM is assessed using tribo-tests. Prior to testing, specific measures
were taken to ensure consistency. The ring compression specimen was submerged
into the lubricant. After excess lubricant was allowed to drip off, the evenly coated
specimen was placed between the dies. Before deformation began, a preload of
300 N was placed on the specimen to ensure contact between workpiece/dies,
thus providing a path for the electricity to travel and avoid generation of electrical
236 9  Tribological and Contact Area Effects

80%
Top and bottom dies m=0.5
temperature:200ºC 20ºC
Thermal test
Inner diameter reduction [%] 60%

40% m=0.3

Ring temperature:
~300ºC
20%
m=0.2

0%
0% 10% 20% 30% 40% 50% 60% 70% 80%

m=0.1
20%-
Ring height reduction [%]

Fig. 9.33  Friction calibration curves for tests at room temperature and at an elevated tempera-
ture [13, 14]. The friction calibration curves were tested at room temperature (20 °C) and at an
elevated temperature that was respective of what would be reached during an EAF test (300 °C)

Initial Stroke 1 Stroke 2

Fig. 9.34  Initial and deformed ring compression samples [13, 14]. Shown is the initial unde-
formed ring compression sample, as well a sample that has been deformed to the two different
strokes used for testing

sparks. After this, the ring specimens were compressed at a rate of 6.35 mm/min
to strokes of 1 mm and 1.3 mm. Figure 9.34 displays a ring specimen prior to
deformation, deformed to a stroke of 1 mm (Stroke 1), and deformed to a stroke of
1.3 mm (Stroke 2).
For each stroke, three electrical conditions were evaluated: (i) conventional
forming (no electricity), (ii) a current density of 25 A/mm2 (333 Amps), and (iii)
a current density of 35 A/mm2 (467 Amps). The voltage across the specimens was
measured at the beginning of the tests. Specifically, the voltage for a current den-
sity of 25 A/mm2 was 110 mV, and the voltage for a current density of 35 A/mm2
was 160 mV. From Perkins et al. [20] where the same stainless steel alloy was
tested under electrical conditions, the “electrical threshold” (i.e., minimum current
density where enhanced formability effects are first observed) was determined to
9.2  Tribological Effect on EAF Effectiveness 237

be around 18 A/mm2. Hence, with the current densities and deformation speeds
used in this work, the threshold will be met and exceeded.
For the tests where electricity is present, current is applied at the start of defor-
mation and is ceased when the desired stroke is displaced. Note that the current
densities mentioned here are initial values. A constant current (variable voltage)
power supply was used to produce the electricity and maintain a constant current
throughout the duration of each test. Specifically, as the specimen deformed, the
resistance decreased and therefore the voltage was decreased to maintain the spec-
ified current. However, while the current is constant, the specimen area increases,
thus the current density decreases during a test. For each combination of lubricant,
stroke, and current density, a number of three-to-five replicates were run to ensure
repeatability.

9.2.5 Candidate Metal Forming Lubricants

There are four groups of lubricants used in metal forming processes: water-based,
oil-based, synthetic, and solid film. The choice of the lubricant for one specific
process takes into consideration a multitude of process variables that may influ-
ence the effectiveness of the lubricant, such as: interface pressure and tempera-
ture, sliding velocity, surface expansion, tool and workpiece material properties
and geometries, and surface topography [17]. For EAF, there are two more aspects
to consider: dielectric permittivity and higher temperatures at the interface, which
may require the presence of pressure additives (EP). For this study, three lubri-
cants were selected: (i) Renoform OL569NE (oil-based lubricant), (ii) TufDraw
1919 (water-based lubricant), and (iii) SynDraw 1310D (synthetic lubricant).
Table 9.8 summarizes some characteristics of the lubricants.

Table 9.8  Metal forming lubricants analyzed with EAF [13, 14]

Lubricant Lubricant characteristics
Renoform Petroleum-based
OL569NE Light duty
Exhibits high lubricity from a lard oil additive
Viscosity at 40 °C: 11.5 cSt
TufDraw 1919 Water dilutable petroleum-based
Heavy duty
Has chlorinated paraffin extreme pressure additives and a proprietary
binder/lubricity additive, used 1: 1 with water
Viscosity at 40 °C: 400 cSt
SynDraw 1310D Water-based synthetic, non-petroleum
Medium duty
Has phosphate ester extreme pressure agent and fatty ester lubricity
additives
Viscosity at 40 °C: 110 cSt
238 9  Tribological and Contact Area Effects

9.2.6 Experimental Results and Discussion

9.2.6.1 Lubricant Evaluation Based on Shear Friction Factor

The reduction of the inner diameter was determined for all the specimens, and the
values were superimposed on the friction calibration curves determined earlier
for the corresponding ring height reduction. The friction factor was quantified for
each lubricant for conventional forming and EAF. Figure 9.35 shows in (a–c) the
measurements used to interpolate m.
Figure 9.36 shows the evolution of an average current density-dependent friction
factor for each lubricant. Conventional forming was represented by a current density
of 0 A/mm2. The oil-based lubricant exhibited an average friction factor of m = 0.18
in conventional forming. However, when the electricity was applied, the friction fac-
tor went to 0.25 and 0.28, respectively. The synthetic lubricant initially exhibited a
slightly lower friction factor, but it reached the same values when the electricity was
applied. The poorest performance was exhibited by the water-based lubricant, which
in conventional forming had the highest average friction factor (m  = 0.24), and
increased up to m = 0.34 for a current density of 35 A/mm2. This evolution observed
with the current density suggests that the dielectric permittivity of the lubricant has
less effect on the lubricant performance under electricity, than the temperature rise.
The heat generated will result in more severe evaporation for a water-based lubricant.
Note that the error bars depicted in the figure indicate some scattering, which
is to be expected when forming microscale specimens. At this scale, small differ-
ences in material structure, specimen dimensions, or applied current may result in
variation of the results.
The effective increase in friction with current density can be observed in the flow
pattern of the material and increase in the barrel shape. Figure 9.37 shows magnified
images taken on a section of the ring, for a deformed specimen at 0, 25, and 35 A/
mm2. The lubricant used was TufDraw 1919. The specimens were carefully cut on
a CNC milling machine. Then, sectioned rings were mounted in resin for the metal-
lographic preparation. Grinding, polishing, and etching were performed on the speci-
men, in order to have a clear delimitation of the grains. Thus, the specific “x” shape
of the higher localized strain could be observed with an increase in current density.
To re-emphasize the results discussed in Fig. 9.6, an increase in friction can
also be observed from the increased surface roughness of the ring specimens.
Preliminary measurements indicated that the roughness increased from 0.89 µm in
conventional forming to 1.24 µm at 25 A/mm2 and 1.4 µm at 35 A/mm2.

9.2.7 Lubricant Evaluation (Reduction in Forming Load)

When electric current is superimposed on the deformation, the forming load is

reduced. Figure 9.38 shows the compressive force versus stroke displacement
9.2  Tribological Effect on EAF Effectiveness 239

(a) 100% m=0.7

Conventional Forming
80%

Inner diameter reduction [%]

m=0.5
60%

40% m=0.3
m=0.25
20% m=0.2
m=0.15
0%
0% 10% 20% 30% 40% 50% 60% 70% 80%
20%
- m=0.1
Renoform OL569NE
40%
- TufDraw 1919
m=0.05
SynDraw 1310D
60%
-
Ring height reduction [%]
(b) 100% m =0.7
EAF, Current density: 25 Amps/mm2
80% + m =0.5
Inner diameter reduction [%]

60% ~
_
40% m =0.3
m =0.25
20% m =0.2

0% m =0.15

0% 10% 20% 30% 40% 50% 60% 70% 80%

m =0.1
- 20%
Renoform OL569NE
- 40% Tuf Draw 1919
m =0.05
SynDraw 1310D
-60%
Ring height reduction [%]

(c) 100% m =0.7

EAF, Current density: 35 Amps/mm 2
80% + m =0.5
Inner diameter reduction [%]

60% ~
_
40% m =0.3
m =0.25
20%
m =0.2

0% m =0.15
0% 10% 20% 30% 40% 50% 60% 70% 80%
- 20% m =0.1
Renoform OL569NE
- 40% Tuf Draw 1919
m =0.05
SynDraw 1310D
- 60%
Ring height reduction [%]

Fig. 9.35  Friction calibration curves and friction factors for the three lubricants tested [13, 14].
The friction factors were calibrated for several different applied current densities: a Conventional
forming, b EAF at 25 A/mm2, and c EAF at 35 A/mm2
240 9  Tribological and Contact Area Effects

0.4
Renoform OL569NE
0.35 Tuf Draw 1919

Friction Factor (m)

SynDraw 1310D
0.3

0.25

0.2

0.15

0.1
0 25 35
Current Density (Amps/mm2)

Fig. 9.36  Average friction factors [13, 14]. The average friction factors in relation to the applied
current densities are shown. TuffDraw 1919 consistently had the highest average friction factor

Fig. 9.37  Sections of deformed ring compression samples [13, 14]. The deformation patterns
are different for the different applied current densities: a Conventional forming, b 25 A/mm2, and
c 35 A/mm2 (Mag 50X). Lubricant: TufDraw 1919

plots for Renoform OL569NE. The plot contains the load profile for a conven-
tional test, a test at 25 A/mm2 and a test at 35 A/mm2.
A comparison between the reductions in forming load recorded for all three
lubricants is shown in Fig. 9.39. The reduction was calculated using Eq. (9.10):
9.2  Tribological Effect on EAF Effectiveness 241

Fig. 9.38  Forming load

Compressive Force [kN]

40
profiles for Renoform 35
Renoform OL569NE Conventional
OL569NE [13, 14]. As the 30
applied electrical power is 25 A/mm2
25
increased, the overall forming 20 35 A/mm2
load profiles decreases 15
10
5
0
0 0.5 1 1.5
Stroke Displacement [mm]

Fig. 9.39  Reduction in the 30
forming load versus current Renoform OL569NE
densities for all lubricants 25 TufDraw 1919
tested [13, 14]. Renoform
Reduction in Load [%]

SynDraw 1310D
OL569NE showed the lowest 20
overall reduction in the
forming load, whereas the
15
other two lubricants showed
comparable reductions in the
forming load 10

0
0 10 20 30 40
2
Current Density [Amps/mm ]

Fconv − FEAF
Reduction = (9.10)
Fconv
where Fconv is the average forming load recorded in the conventional test, represented
in the figure by 0 A/mm2, and FEAF is the forming load recorded in the EAF test.
For Renoform OL569NE, the current densities of 25 and 35 A/mm2 were able
to reduce the final forming force by about 8 and 22 % compared to the conven-
tional test. Tuf Draw 1919 further decreased the compressive loads to about 16
and 29 %, compared to its respective conventional test. Similarly, SynDraw 1310D
was able to show force decreases of 15 and 28 %. When coupled with EAF, Tuf
Draw 1919 and SynDraw 1310D produced greater force reductions compared to
their respective conventional forming test.

9.2.8 Temperature Measurements

A thermal camera measured the temperature distribution on specimens as they

compressed throughout the test. The thermal measurements were taken using a
FLIR Systems ThermoVision A40 camera, calibrated for a temperature range from
242 9  Tribological and Contact Area Effects

0 to 550 °C, and checked at room temperature, with a spatial resolution (IFOV) of

1.3 mrad. The temperature was measured for three tests: (i) stationary tests, when
electricity is applied through the specimen, but when there is no deformation, just
preloading to ensure contact and no electrical spark, (ii) an EAF compression test,
and (iii) a conventional compression test. In order to measure the temperature rise
during the tests, no lubricant was used during these tests. The specimens were
cylindrical, with the same diameter as the outer diameter of the ring samples, and
a height of 4 mm. The same material and same test conditions as for rings were
used. For a higher emissivity factor and more accurate readings, the specimens
were painted black. The thermal profiles were calculated using ThermaCAM soft-
ware. Specifically, the software allowed for the selection of an area on the speci-
men. Then the maximum value in the area was recorded at a frequency of 1/250 s
until the user ended the recording.
For the stationary test and a current density of 25 A/mm2, the temperature
increased rapidly up to approximately 310 °C (Fig. 9.40). The second test, which
was an EAF compression test, exhibited a different behavior. The temperature
increased initially up to 105 °C, and then it started to decrease, before turning off
the electricity. During the conventional compression test, an increase of maximum
10 °C was recorded. A few conclusions are reached from these tests:
The temperature measurements during the EAF test agreed with other studies
that indicated a temperature between 100 and 200 °C [20]. The low temperature
rise during the conventional compression test suggests that the increase in tem-
perature in the electrical tests is due to the resistive heating. The large difference
in temperature rise between stationary and EAF tests can be partially attributed to
the change in specimen surface area during the deformation, which affects the heat
transfer through convection and conduction. Note though that the change in total

350
300
Temperature [ºC]

Stationary test
250
EAF test
200
Conventional
150
test
100
50
0
0 5 10 15
Electricity applied Time [sec]
(25 Amps/mm 2) for
Stationary/EAF tests

Fig. 9.40  Temperature profiles of stationary electrical and EAF tests [13, 14]. The stationary
electrical test at the same current density is much hotter than the EAF test. This supports the idea
that the electrons aiding with the plastic deformation of the material
9.2  Tribological Effect on EAF Effectiveness 243

area of the specimen was less than 1.5 %, which cannot explain the drastic drop in
temperature between the two tests. This leads to the conclusion that a significant
part of the electric energy is actually imparted by the flow of electrons to the mate-
rial and this aids the deformation, as described earlier, and the remaining energy
is converted into bulk heating [21]. Thus, the higher temperatures recorded at the
beginning of the test may result in rapid evaporation of the lubricant, and poten-
tially change in a less favorable lubrication regime.

9.2.9 EAF/Tribology Conclusions

The advance of EAF for larger-scale applications depends on identifying a well-

performing lubricant. Three lubricants chosen from three different groups were
tested under different current densities and their performance was evaluated. The
ring compression tribo-test was used and a friction factor was determined from the
friction calibration curves generated through finite element analysis. The following
conclusions were drawn from this study:
• The use of electricity during deformation processes has beneficial effects in
reducing flow stress, thus the energy required to reach the same level of defor-
mation as in conventional forming is lower. Also, larger strains are reached for
the same mechanical load.
• The lubricants analyzed indicated more severe tribological conditions with the
application of electricity and with the increase of current density. The higher
temperatures resulted in rapid evaporation of the lubricant and chemical decom-
position, thus the lubrication mechanism changed to a less favorable regime.
• Although the forming load, thus the interface pressure, decreased in all cases
when the electricity was applied, there is potential for even more reduction if the
lubricant withstands the effects of electricity. The higher temperatures suggest that
pressure additives are needed in order to maximize the efficiency of the process.
• The oil-based/synthetic lubricants performed better than water-based lubri-
cants. Thus, it is recommended that the development of the lubricants for EAF
should take into consideration the temperature rise, rather than just dielectric
permittivity.

References

1. Holm R (1976) Electric contacts, theory and applications. Springer, Berlin

2. Bunget C, Ngaile G (2008) Ultrasonic microforming. VDM, Saarbrucken. ISBN
978-3-639-00055-9
3. Timsit RS (2001) Connector lubricants enhance performance. www.interconnectionworld.com.
Accessed 06 Jun 2011
4. Bowden FP, Tabor D (1939) The area of contact between stationary and between moving sur-
faces. Proc R Soc Lond A 169(938):391–413
244 9  Tribological and Contact Area Effects

5. (2006) MS thesis, ‘Microforming and Ultrasonic Forming’, North Carolina State University
6. Salandro WA (2012) Thermo-mechanical modeling of the electrically-assisted manufacturing
(EAM) technique during open die forging. PhD dissertation, Clemson University
7. Uppal AH, Probert SD (1972) Mean separation and real contact area between surfaces
pressed together under high static loads. Wear 23:39–53
8. Pullen J, Williamson JBP (1972) On the plastic contact of rough surfaces. Proc R Soc Math
Phys Eng Sci 327:159–173. doi:10.1098/rspa.1972.0038
9. Kragelsky IV, Demkin NB (1960) Contact area of rough surfaces. Wear 3:170–187
10. Stewart M (1990) A new approach to the use of the bearing area curve. In: International con-
ference on honing technologies and applications, FC90-229, pp 11
11. Uppal AH, Probert SD (1972) Considerations governing the contact between a rough and a
flat surface. Wear 22:215–234
12. Boyer L (2001) Contact resistance calculations: generalizations of greenwood’s formula
including interface films. IEEE Trans Compon Packag Technol 24:50–58
13. Bunget C, Salandro WA, Mears L (2010) Tribological aspects in electrically-assisted form-
ing. In: International manufacturing science and engineering conference, MSEC 2010-34249,
p8
14. Bunget C, Salandro WA, Mears L (2010) Evaluation of lubricants for electrically-assisted
forming. Proc Inst Mech Eng Part B: J Eng Manuf. doi:10.1177/0954405411401267
15. Schey JA (1983) Tribology in metalworking. American Society for Metals
16. Altan T, Ngaile G, Sheng G (2004) Cold and hot forging; fundamentals and applications.
ASM International
17. Gariety M, Ngaile G, Altan T (2007) Evaluation of new cold forging lubricants without zinc
phosphate precoat. Int J Mach Tools Manuf 47:673–681
18. Ngaile G, Botz F (2008) Performance of graphite and boron-nitride-silicone based lubricants
and associated lubrication mechanisms in warm forging of aluminum. J Tribol 130:7
19. Ngaile G, Cochran J, Stark D (2007) Formulation of polymer-based lubricant for metal form-
ing. Proc IMechE Part B: J Eng Manuf 221:559–568
20. Perkins TA, Kronenberger TJ, Roth JT (2007) Metallic forging using electrical flow as an
alternative to warm/hot working. J Manuf Sci Eng 129(1):84–94
21. Bunget C, Salandro WA, Mears L, Roth JT (2010) Energy-based modeling of an electrically-
assisted forging process. Trans North Am Manuf Res Inst SME 38:647–654
Chapter 10
Design of an Electrically Assisted
Manufacturing Process

10.1 Energy Analysis

In previous chapters, we have explored the underlying physics behind electrically

assisted manufacturing (EAM), how to model the phenomenon and its effect on
deformation process forces, power, and resultant microstructure. In this chapter,
we aim to explore the feasibility of implementing EAM on a commercial scale,
primarily specific design requirements, cost, and sustainability impact on the man-
ufacturing system as a whole. We begin with an analysis of feasibility, and then
supply metrics of sustainability and cost. The chapter closes with practical design
considerations for processing using EAF.
Sustainability for a manufacturing process is defined in terms of the burden on
society for future generations. In the context of EAF, we address sustainability
through energy analysis and comparison with competing processes.

10.1.1 Conventional Cold Forming

Let us first consider the energy required for cold forming of the open-die forged
case example part given in Fig. 10.1.
The energy required is purely mechanical, and is equivalent to the area under
the true stress-true strain curve, given experimentally in Fig. 10.2.
Note that the equivalent strain for this example is 0.3, beyond the fracture limit
for this material. We will calculate the theoretical work required to deform to this
strain, then explore cases in which it can realistically be achieved.
The stress-strain curve at room temperature is ideally modeled through the
power law for strain-hardening materials:
σf = Kεn (10.1)

© Springer International Publishing Switzerland 2015 245

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_10
246 10  Design of an Electrically Assisted Manufacturing Process

Fig. 10.1  Open-die forged (a) (b)

titanium rod. a Initial raw
material, b final part after
forging, 0.3 final strain

Material: Mg AZ31B
Mass: 0.19 g
Diameter: 4.34 mm Diameter: 5.04 mm
Height: 7.26 mm Height: 5.38 mm

Fig. 10.2  Stress-strain curve
for Mg AZ31B. Shown in
this figure is the experimental
forming curve, as well as the
power law simulation model

where, σf is the material flow stress, ε is the strain, and K and n are the strain-hard-
ening coefficient and exponent, respectively. For Mg AZ31B, K  = 458 MPa and
n = 0.1818. We now calculate the specific energy of deformation (per unit volume):

εf 0.3 0.3

Kεn+1 
u= σf dε = Kεn dε = (10.2)
n + 1 0
0 0

This is evaluated, and confirmed empirically with the above test curve, with the
result:
uconv = 93.4 J/mm3 (10.3)
The total energy of deformation is calculated by:
Econv = uconv V (10.4)
The total energy required is:
Econv = 10, 032 J (10.5)
10.1  Energy Analysis 247

10.1.2 Thermally Assisted Forming

As explained in Chap. 1, the limitations of cold forming for lightweight metals

have been addressed through raising temperature either before forming or during
intermediate stages for annealing. In this case, we must augment the ideal model
for the effects of thermal softening.

10.1.2.1 Material Flow Model for Elevated Temperature

For the room temperature and elevated temperature data, a phenomenological

constitutive equation was developed to predict the flow stress behavior under con-
ventional forming conditions at varying temperatures. This was derived such that
the effects of temperature could be related to material flow stress and bulk ther-
mal softening in the models created in this work. The expression is given gener-
ally in Eq. (10.6) and the constants as a function of temperature are given in Eqs.
(10.7)–(10.9).
σf = Kεn exp (εs) (10.6)


 T ≤ 298 K: K = KRT
K := 298 K < T < 423 K: K = K1 T + K2 (10.7)

T ≥ 423 K: K = K3 exp (K4 T )


 T ≤ 298 K: n = nRT
n := 298 K < T < 423 K: n = n1 T + n2 (10.8)

T ≥ 423 K: n = n3 T + n4


 T < 392 K: s = 0
s := 392 K ≤ T ≤ 600 K: s = s1 T 2 + s2 T + s3 (10.9)

T > 600 K: s = 0
where, σf is the flow stress, K is a strength coefficient, ε is strain, n is the strain-
hardening exponent, and s is a softening term. The term exp(εs) represents the sof-
tening potential of the material for higher temperatures. The strength coefficient
is constant at room temperature and increases linearly to 423 K (150 °C) where it
then decays exponentially with an increasing temperature, n has a linear decrease
with temperature as related to two temperature ranges, and s has a parabolic form
such that at lower temperatures and higher temperatures there is no softening, but
is a maximum at a moderate temperatures. The lower effect of s at a higher tem-
perature may be a result of other possible deformation mechanisms or dynamic
recrystallization and at lower temperatures there is not much material softening.
The best fit coefficients for Mg AZ31B given in Table 10.1 were determined using
least squares regression of experimental data.
248 10  Design of an Electrically Assisted Manufacturing Process

Table 10.1  Conventional Constant Value

room and elevated
KRT 457.72
temperature model
coefficients for Mg AZ31B K1 −1.9529
K2 1033.8
K3 14276
K4 −0.01
nRT 0.1818
n1 −0.0004
n2 0.2867
n3 −0.0009
n4 0.517
s1 0.00008
s2 −0.0794
s3 18.825

The model results are compared to the experimental results in Fig. 10.3. To vis-
ually depict the response of this function over a range of temperatures, Fig. 10.4
displays the corresponding true stress over varying temperatures and strains.

10.1.2.2 Energy Requirement Calculation

Consider the case where the part is bulk heated in an oven of ηt = 70 % thermal
efficiency (i.e., 70 % of the energy put into the heating process directly works to
raise the temperature of the part). The energy of deformation is composed of the
mechanical deformation work and the heat input:
ETAF = ETAF,mech + ETAF,heat
εf
mcp (T − T∞ )
ETAF = σTAF dε +
ηt
0 (10.10)
εf
mcp (T − T∞ )
ETAF = K εn eεs dε +
ηt
0

The first term results in an exponential integral, and is evaluated numerically.

For the thermally assisted case, we will raise the temperature of the part to
473 K (200 °C). This results in thermal softening modeled by Eq. (10.6) with
K = 126 MPa, n = 0.0912, and s = −0.833. Using a specific heat value of 960 J/
kg-K, the thermally assisted case is evaluated to find:

ETAF,mech = 2939 J
ETAF,heat = 47 J (10.11)
ETAF = 2986 J
10.1  Energy Analysis 249

Fig. 10.3  Conventional room and elevated temperature flow stress model comparison

Fig. 10.4  Conventional flow stress response surface for room temperature to elevated testing
temperatures over varying strain levels

10.1.3 Electrically Assisted Forming

In the EAF case, direct electric current is applied to the workpiece, imparting
softening of the material. Though previously discussed as a primarily resistive
heating phenomenon at lattice defect sites, in this analysis, we calculate the total
electrical energy imparted as a function of current and applied voltage:
250 10  Design of an Electrically Assisted Manufacturing Process

EEAF = EEAF,mech + EEAF,elec

εf 
VIdt
EEAF = σEAF dε + (10.12)
ηe
0

The model of Eq. (4.9) is used to predict the stress state:

 
′ n
 a
 VIdt
EEAF = K ε exp −(ε − c)BΦ dε + (10.13)
ηe

Using an efficiency of 90 %, resistivity of 9.8 × 10−8 Ohm-m, constants of

a = 3.38, B = 0.00009, c = 0.018, and applying 40 A/mm2; this is evaluated with
the result:

EEAF,mech = 1552 J
EEAF,heat = 83 J (10.14)
EEAF = 1635 J

10.1.4 Energy Comparison

The summary of energy use for each of the test cases is given in Fig. 10.5.
We see that energy-assisted forming (thermal or electrical) can reduce the
required mechanical work significantly. However, there is a considerable differ-
ence between thermal softening (at the bulk material level) and electrical softening
(at specific locations in the lattice structure). For thermal softening, the bulk work-
piece must be brought to temperature as a whole. However in EAF, the electric-
ity works to soften the material directly in the zones of deformation. In the given
example, the total mass of the part was relatively insignificant to thermal softening
energy. However, in larger mass parts such as vehicle body panels, the difference
between thermal and electrical energy to achieve the same reduction in flow stress
should be more significant.
Electrically assisted softening significantly reduces the amount of deformation
energy required at a quite low augmentation energy cost. Though current is very

Fig. 10.5  Energy use in
forming Mg-AZ31B-O
sample using various
methods
10.1  Energy Analysis 251

high due to low resistance, the overall power (and therefore energy delivered in a
given time) is low.

10.2 AC Versus DC Current

All previous analyses in this book have been for application of direct current to
assist forming, but alternating current (AC) is worthy of consideration, as it is
hypothesized to produce the same effect. Additional considerations with the use of
AC are discussed below.

10.2.1 Energy Analysis

Energy use for the AC case with power factor cos (φ) can be evaluated using the
following relationship (φ is the phase angle between current and power). This
equation uses an expression for average power, relevant to AC circuit analysis.

EEAFAC = EEAFAC,mech + EEAFAC,elec

εf 
EEAFAC = σEAFAC dε + VI cos (φ)dt
0
tf
εf sin2 ωtdt (10.15)
0
EEAFAC = σEAFAC dε + Vm Im cos (φ)
tf
0
εf
Vm Im cos (φ)
EEAFAC = σEAFAC dε +
2
0

where Vm and Im are the nominal voltage and current values, respectively.

10.2.2 Skin Effect

One consideration for use of AC power is the skin effect, a phenomenon observed
in AC conductors where the current density is higher at the surface of the conduc-
tor than at the center. Though not typically an issue in well-designed AC circuitry,
when AC is applied to EAF, there can be process behavior not captured in the pre-
viously described models.
252 10  Design of an Electrically Assisted Manufacturing Process

Recall from Chap. 2 that the electroplastic effect that softens metals in an elec-
tric field is proportional to the current density. In EAF, the workpiece itself is the
conductor, so varying current density through the orthogonal cross section trans-
lates to varying material properties as well, with greater softening occurring near
the outer surface of the specimen. The skin effect is modeled by
 
−d
J = Js exp (10.16)
δ
where J is the instantaneous current density, Js the current density at the surface
of the conductor, d the distance into the conductor, and δ the skin depth, a depth at
which the density has fallen to Jes . This skin depth is approximated by

2ρ
δ= (10.17)
ωµ

where ρ is the resistivity of the material, ω the angular frequency of the current,
and µ the magnetic permeability of the material. Examining the relationships, the
following general conclusions are reached:
• Higher-resistivity materials have both
– more responsiveness to the electroplastic effect
– higher skin depth
• Higher angular frequency results in lower skin depth, which implies a higher
concentration of current density at the surface.
• Higher permeability also reduces the skin depth. This is especially pertinent
for irons and steels, which typically cannot be used as AC conductors for this
reason.

10.3 Additional Process Design Considerations

10.3.1 Power Supply
The key characteristics for specification of the power supply to be used in process
augmentation are:
• High-power capacity
• Adequate current capacity for the (current density × expected cross-sectional
area) for the process
• High-dynamic controller to deal with current transient suppression.
Some potential power supply designs are:
• Battery-based system. This is an expandable direct-current (DC) system that can
scale up to practically any power level. It should be integrated with a current
controller.
10.3  Additional Process Design Considerations 253

• Isolated-gate bipolar transistor (IGBT) system. This is a high-power, high-

efficiency, fast-switching device and can be used for current modulation.
• Power metal-oxide-semiconductor field-effect transistor (power MOSFET)-
based system. This is typically for lower-voltage (not necessarily lower-power)
applications.
• Silicon-controlled rectifier (SCR). Also termed a thyristor, this device is typi-
cally used to control high-current, high-voltage applications, and can be rapidly
switched for high-dynamic behavior, which makes this type of power supply
good for the EAF application.

10.4 EAF Process Design Conclusions

An analysis of the energy required to form a sample lightweight alloy was per-
formed. In addition to achieving greater elongation, the energy-augmented
processes (thermal and electrical) reduce the required mechanical work of defor-
mation without significant energy penalty. The benefit of electrical forming assis-
tance as compared with thermal assistance is noted in that, as part mass increases,
EAF becomes a more preferable option.
The analyses of this work all use DC application. Some considerations are out-
lined for use of AC, as well as considerations for various types of power supplies.
Chapter 11
Applications of Electrically Assisted
Manufacturing

Within this book, a modeling strategy for the EAF technique is explained for both
compression and tension. Both strategies separate the thermal softening effects
from the direct electrical effects and thus produce temperature and force profiles
for their respective processes. However, in the real world, manufacturing pro-
cesses are rarely exclusively compression or tension. Therefore, within this chap-
ter, manufacturing processes that can be applicable to EAF will be explained.
These include bending, stretch forming, machining, friction stir welding, and
miscellaneous other EAF-industrialization research by researchers other than the
authors. In addition, this chapter will include experimental EAF findings for com-
pression, tension, channel formation, springback, and various types of forming.
Furthermore, for example, industrial applications for EAF will be described.

11.1 EAF Bending Application and Model

The tension/compression EAM modeling strategies explained in previous chapters

can be applied to other EAM-applicable processes. Specifically, this same mod-
eling approach was instituted to predict forming loads in an electrically assisted
bending (EAB) process [1, 2]. Within this section, an analysis of how to model an
EAB process will be explained, assumptions of the model will be provided, a clas-
sical bending process will be described, an analytical piece-wise model for EAB
will be presented, the solution schematic will be described, the experimental setup
and procedure will be explained, thermal measurements for an EAB process will
be illustrated, the EAB model will be validated using experiments, the effects of
electricity on a bending process will be given, and conclusions will be summarized
(Fig. 11.1).
The overall objective of the EAB model is to establish a closed-form solution
that describes the bending force specific to a certain deformation, while pulsed

© Springer International Publishing Switzerland 2015 255

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2_11
256 11  Applications of Electrically Assisted Manufacturing

Fp Thermal camera
Press Temperature
Insulation measurements
Punch
Workpiece
Die
Position and
Force
Insulation
measurements

Fp DAQ System
A
V

~ _
+

Power source

Fig. 11.1  Schematic of an EAB test [1, 2]. The setup consists of an insulated and adjustable
bending die, along with insulation to guide the electricity into the workpiece

current (not continuously applied) is applied through the dies to the workpiece.
Since air bending is stretching the metal (i.e., the material cross section decreases),
the pulse electrical application technique will be used where the electricity is
applied in pulses with a set duration and spaced over a given period of time. The
model may be used for determination of the state of stress and strain during the
deformation and its dependence on different parameters of the applied current
(e.g., current density, pulse duration, and pulse period). The derivation begins
by establishing equations for the effective stresses and strains acting on a plasti-
cally deformed sheet metal part in classical bending based on the first principle of
mechanics. The material is assumed to follow the power law, σ̄ = C ε̄n where σ̄ is
the flow stress, C is the strength coefficient, ε̄ is the effective strain, and n is the
strain hardening exponent. The bending is assumed to be under plane strain condi-
tions (no strain in the width direction), and a transverse force acts on the flat strip
of sheet metal through the punch. The forces acting on the material are shown in
Figs.  11.2 and 11.3. Using equilibrium of force and moments, the equations for
bending force are established for the classical process, and the deformation energy
can be computed.
Using this model as a starting point, the next step is to include the effects of
the electrical current applied to the workpiece during the process: energy effect,
expressed by the EEC, and the temperature rise effect (e.g., 1-EEC). The ana-
lytical model is validated via experiments for classical (non-electrical) and EAB
tests run on 304 Stainless Steel sheet specimens. The effects of the electric current
applied on the bending forces and springback reduction achieved are also investi-
gated experimentally for different electrical parameters (i.e., current density, pulse
11.1  EAF Bending Application and Model 257

frequency, and pulse duration). It is expected that both, model predictions and
experiments, will bring a better understanding of the electroplastic deformation
mechanism in general, and of EAB in particular.

11.1.1 Analysis of an EA Bending Process

Due to the punch motion in a bending process, the geometry, the bending
moments, and force directions change continuously; thus, the bending is a non-
steady-state process, and the stresses and strains on the material change through-
out the process. At the beginning of the process, the deformation is elastic. As the
punch advances, the stresses in the outer fiber at mid-span increase and exceed the
elastic limit; thus, the plastic deformation starts and propagates toward the inner
fiber and into the rest of the material. The punch motion and the amount of mate-
rial springback will dictate the final bending angle.
In an electrically assisted test, a power source provides electricity and the cur-
rent flows through the dies. Figure 11.1 shows an experimental schematic of an
EAB test. Insulation is positioned between the dies and machinery to guide the
flow of the electricity into the workpiece to assist in deformation.

11.1.2 Assumptions of the EAB Model

The following major simplifications and assumptions are used throughout the deri-
vations in this study:
• The workpiece material is homogeneous and isotropic, and the thickness and
width of the sheet are uniform. This is due to the fact that the sheet metal used
for this research was in the annealed state.
• The width of the specimens is at least 10-times larger relative to the thickness;
thus, plane strain conditions exist.
• In the moment that plastic deformation starts, the elastic deformation is negligi-
ble and the power law is used for determining the flow stress of the material.
• The strip is subdivided into three sections: two linear and a circular section. The
bending line is a circular arc with the same radius as the punch itself. The sec-
tions normal to the sheet surface are assumed to remain plane at all times.
• The friction at the workpiece/die interfaces follows Coulomb’s friction law. The
lubrication regime is assumed unaffected by the presence of electricity.
• The specific heat and resistivity of the material are assumed independent of
temperature.
• The strain-rate sensitivity is neglected since previous experimental investiga-
tions found that the temperature rise is not high enough for the process to be
considered as hot forming.
• Volume constancy is valid throughout the deformation process; thus, the change
in sheet thickness during the bending process can be neglected.
258 11  Applications of Electrically Assisted Manufacturing

11.1.3 Classical Bending Process (Force and Springback)

The geometry of the bending tooling and the coordinate system are shown in
Fig. 11.2. The model development starts with an analysis of the state of strain and
stress in the bending zone under plane strain conditions, using the volume con-
stancy assumption and flow rules. The width of the workpiece is assumed to not
change; thus, the strain in the y-direction is zero. The strains and stresses are given
by Eq. (11.1).
εx ; εy = 0; εz = −εx ,
σx (11.1)
σx ; σy = ; σz = 0.
2
The effective strain and stress are determined using von Mises yield criterion as
given by Eq. (11.2).
√
2 3
ε̄ = √ εx ; σ̄ = σx . (11.2)
3 2

The strain in the x-direction is solved

 by assuming the exact position of the neutral

axis given by Hill et al., rNA = rp rp + tb . Thus, the strain can be computed
using the engineering strain, given by the change in length of the outer fiber as
compared to the neutral fiber [3]. Equations (11.3) and (11.4) give the strain in the
outer fiber, where α is the bending angle.

(rp + tb )α − rNA α tb
ex = = 1+ − 1; εx = ln(1 + ex ). (11.3)
rNA α rp

Fp
Punch

Workpiece
rNA
w rp
tb tb
rp

Die y
wd x
z

Fig. 11.2  Geometry of an air bending test [1, 2]. All of the notations that will be used in the fol-
lowing bending equations are provided in the figure
11.1  EAF Bending Application and Model 259

 
tb
εx = ln 1+ . (11.4)
rp

The entire section will be stressed, except for an elastic core near mid-plane,
which will shrink in thickness as the deformation progresses. For this analy-
sis, the elastic core is neglected after plastic deformation starts. Note that at the
beginning of the deformation, elastic deformation based on Hooke’s Law is con-
sidered. To calculate the bending moment needed to produce  this bend, M, it is
assumed that there is no net external force in the x-direction ( Fx = 0). However,
the internal force, dFx, acting on any incremental element of cross section, w dz,
is dFx = σx w dz. The contribution of this element to the bending moment is
dM = z dFx = σx wz dz. The total bending moment can be obtained by integrating
from –t/2 to t/2, as shown in Eq. (11.5). To facilitate the mathematical develop-
ment, the material is assumed perfectly plastic (σx = ±σ̄ ), and the flow stress is
evaluated for the highest stress achieved in the bending process, which is the stress
in the outer fiber [4–6].

t/2   n
1 2 1 2 tb
M= wσx z dz = σ̄ wtb = Kwtb ln 1+ . (11.5)
4 4 rp
−t/2

The bending moment will be used further to determine the punch force, by
employing an equilibrium of moment and forces. Figure 11.3 presents the simpli-
fied model used in the punch force analysis. The reaction force in the die shoul-
ders and the friction force at the die/workpiece interfaces are shown. The free
body diagram summarizes the bending moment and forces acting on the linear
zone. For simplification, it is assumed that the neutral axis does not change during
the process, but it stays at the middle of the section.
The following Eqs. (11.6)–(11.10) are derived from [5]. The equilibrium of
forces in the vertical direction results in an expression of the normal force as a

µN µN
/2
B
Fp
N µN
N /2
tb B

A N
wd
A
M

Fig. 11.3  Simplified bending force analysis model [1, 2]. This simplified model figure displays
the notations that will be used throughout the chapter
260 11  Applications of Electrically Assisted Manufacturing

function of the punch force Fp, bending angle, α, and friction coefficient, μ
[Eq. (11.6)].

Fp
N=  . (11.6)
2 cos α2 + µ sin α2

The equilibrium of moments about point A leads to Eq. (11.7).

M
N= . (11.7)
µ t2b + a

The length of the arm, a, is determined geometrically using Eq. (11.8).

α
wd − 2(rp + tb ) sin 2
a= (11.8)
2 cos α2

After substituting Eq. (11.8) into Eq. (11.7), the resultant equation is combined

with Eq. (11.6), and the punch force is resolved as follows in Eq. (11.9).
 
4 cos α2 cos α2 + µ sin α2
Fp = · M. (11.9)
µtb cos α2 + wd − 2(rp + tb ) sin α2

Then, Eq. (11.4) is substituted into Eq. (11.9). Thus, the bending force needed to
deform the material to the bending angle α and to overcome the friction at die/
workpiece interfaces can be obtained from Eq. (11.10).
    n
wtb2 cos α2 cos α2 + µ sin α2 tb
Fp = K · · ln 1+ . (11.10)
µtb cos α2 + wd − 2(rp + tb ) sin α2 rp

Total power for plastic bending is derived mechanically as Eq. (11.11):

Jm∗ = Fp u̇ (11.11)
where u̇ is the velocity of the die/punch.

11.1.4 Analytical Modeling of EAB

Previous research indicated that the application of electricity through a deforming

workpiece results in reduced required flow stress to reach the same deformation as
in the classical process. When analyzing an EAF process, two aspects have to be
considered: (i) the applied electrical energy, and (ii) the electroplastic effect on the
material behavior and on the energy efficiency of the process.
For EAB, the electricity was applied in pulses, since constantly applied elec-
tricity would lead to negative effects because the cross section of the specimen is
11.1  EAF Bending Application and Model 261

Fig. 11.4  Electrical

Current density, Cd
parameters varied in EAB
investigation [1, 2]. The t
electrical parameters that
can be varied are the current
magnitude, the pulse
duration, and the time period 0
between electrical pulses p p

Time, t

decreasing over the test, and therefore, the current density is increasing. The pulse
parameters that were varied during the tests are current density, Cd; pulse duration,
Δt; and pulse period, p (Fig. 11.4).

11.1.4.1 Energy Effect

When electricity is applied during plastic deformation, a part of the electrical

energy is imparted into the mechanical deformation process. The total power con-
sumed by the deformation process is given by Eq. (11.12):
∗
Jtotal = Jm∗ + Je∗ = Jm∗ + ξ Pe (11.12)
where Jm∗ is the mechanical component related to the applied forming load applied
and die velocity, Je∗ is the effective electric power that aids deformation, Pe is the
electrical power (I · V) passing through the workpiece, and ξ is the electroplastic
effect coefficient (EEC). The EEC depends on the material, applied current den-
sity, time, and strain rate and can be determined through tests. The remaining elec-
tric energy, (1 − ξ )Pe, is converted into heat through resistive heating.

11.1.4.2 Temperature Rise Effect

The global heating from the electricity will result in some degree of thermal sof-
tening, thereby lowering the flow stress of the material and contributing to the
decreased required work. At room temperature, the flow stress is given by the
power law presented earlier, but as the temperature rises, the flow curve depends
strongly on the temperature. The investigations conducted by previous researchers
indicated that the temperature is lower than the temperature at which recovery and
recrystallization take place; thus, the strain-rate dependency may be neglected, but
the influence of the temperature on the strength of the material, given by the tem-
perature-dependent coefficient function, C(T), is still significant. The flow stress
is given by equation in Eq. (11.13). When considering all metals, the strength
coefficient is reduced as they are heated. However, each metal has a specific
strength–temperature relationship and has a specific temperature range where this
relationship is most prevalent, which is shown in Fig. 5.6.
262 11  Applications of Electrically Assisted Manufacturing

σ̄ = C(T )ε̄n (11.13)

The temperature rise is obtained from the energy balance in Eq. (11.14).
∂ρU ∂  
+ Q̇conv + Q̇cond = Qrad + Qe (11.14)
∂t ∂xj
 
where ∂ρU ∂
∂t is the rate of change of the internal energy of the part, ∂xj Q̇conv + Q̇cond
are the convective and conduction components of the heat flux, Qrad is the radiative
heat, and Qe is heat generated in the part from the electric energy dissipated. Using
constitutive equations for each component, the heat equation to be solved to deter-
mine the temperature rise for particular electric parameters is shown in Eq. (11.15):

∂T   ∂ 2T  
4 4
ρVv Cp = −As [h(T − T∞ )] − k 2Ad + Ap − A s εσSB T − T ∞ + VI,
∂t ∂xj2
(11.15)

where ρ is the density of the material, Vv is the volume of the part, Cp is the spe-
cific heat of the material, T is the temperature, t is time, As is the lateral surface
of the part, h is the convection heat transfer coefficient, T∞ is the surrounding
temperature, k is thermal conductivity for the die and punch material, Ad and Ap
are the cross-sectional area of the dies and punch, xj are coordinates, ε is radia-
tive emissivity for the part, σSB is the Stefan–Boltzmann constant, V is the electric
voltage, and I is the intensity of the current, given by the product of the current
density and cross-sectional area.
After each pulse, when the current is not applied any longer, the temperature
decreases and can be calculated from Eq. (11.16):

∂T ∂ 2T  
ρVv Cp = −As [h(T − T∞ )] − k(2Ad + Ap ) 2 − As εσSB T 4 − T∞
4
,
∂t ∂xj
(11.16)

which is obtained from Eq. (11.15) after taking out the heat source.

11.1.4.3 Summary of Electrical Effects

The analytical model for EAB uses the energy-based approach to incorporate
the effects of electricity on the bending force model from the classical bending
process. The same punch displacement, s, and the same die velocity, u̇, are con-
sidered for both classical and EAB processes. It is assumed that the same total
energy is required for deformation in both cases. Thus, the force needed in the
EAB process can be determined and compared with the force in the classical
11.1  EAF Bending Application and Model 263

process. Using Eqs. (11.17) and (11.18), a relation between the mechanical

forces is obtained.
∗ ∗
Jclassical = JEAF (11.17)

Fpclassical u̇ = FpEAF u̇ + ξ VCd wtb (11.18)

where Fpclassical is the mechanical load needed in classical bending, FpEAF is the
mechanical load required in the EAF process, V is the voltage applied, Cd is the
current density, and wtb is the cross-sectional area of the sheet. The difference
in required force, as shown in Eq. (11.19), depends on the material (i.e., EEC),
dimensions of the workpiece, current density, and die velocity.
ξ VCd wtb
�Fp = (11.19)
u̇
A reduction in force can also be calculated as in Eq. (11.20):
�Fp ξ VCd wtb
Fred = = (11.20)
Fpclassical u̇Fpclassical
Equations (11.10) and (11.15)–(11.20) constitute the analytical model for the forces
required for EAB. Since the electricity will be pulsed, the model allowed the user to
input the length of the pulses (pulse duration) and the time gap in between each suc-
cessive pulse (pulse period). These two variables are also experimentally evaluated.

11.1.5 EAB Solution Schematic

The analytical model can be solved numerically by implementing a MATLAB

program. The solution schematic is given in Fig. 11.5. A step-wise approach is
used; the position of the punch moves by an increment, ds. Since the velocity u̇
is constant, a time increment, dt, is also computed. Because the electricity is pul-
sated, the model will alternate between classical and EAF bending equations.
When the electricity is applied, the contribution of electrical energy to the defor-
mation work is calculated, and the remaining electrical energy is converted into
heat. The material workpiece temperature rise and subsequent effect on strength
coefficient are derived. This is followed by a period of cooling down. The iteration
continues until the desired bending angle is reached, which will correspond with a
pre-imposed displacement of the die. The number of pulses is notated by N. The
forming punch force is determined for each step and then compared to the clas-
sical process. Note that the same model can be used for the case when the same
mechanical forming load is used, but the punch displacement is different; thus, the
bending angle differs between the classical and EAF process.
264 11  Applications of Electrically Assisted Manufacturing

Material, process and current parameters:

K, n, C(T), , tb, w, L, µ, u, V, C d , p, t

Set ds and dt

Step i
s=i· ds; t=i·dt

Yes
s = s(desired) i = i+1

No
Compare t with the electricity
application (Fig. 4)

t< p N·p t N·p + t N·p t < N·p + t

Calculate temperature Calculate temperature

C= K increase and the new C(T) decrease and the new C(T)

Det. Fp from Det. Fp from Det. Fp from Eqn.

Eqn. (9) Eqn. (16) (9) using C(T)

Print results

Fig. 11.5  Solution scheme for solving the EAB model [1, 2]. A piece-wise approach was used
to solve the EAB model. Specifically, the model switched from conventional bending to EAB
when electrical pulses were applied

11.1.6 Experimental Setup and Procedure

The experiments have two aims: (i) validate the analytical model and (ii) inves-
tigate the effects of electricity in air bending. The aspects considered for evalua-
tion and comparison are related to the bending load and the amount of springback
after each test. The specimens used for this portion of the research were fabricated
from a single sheet of 304 Stainless Steel. They were sheared into dimensions of
12.38 mm wide × 0.864 mm thick × 127 mm long. All shearing was performed
along the roll direction and care was taken when shearing to ensure dimensional
repeatability [since any change in the cross-sectional area will affect the current
11.1  EAF Bending Application and Model 265

Table 11.1  EAB testing parameters [1, 2]

Parameters Bend diameter (mm) Current density Pulse duration (s) Pulse period (s)
(A/mm2)
Values 38.1 20 2 15
20
50.8 30 3 30
60

density (Amps/mm2 or A/mm2)]. Specifically, a tolerance of ±0.127 mm was

held while shearing. Four uniaxial tensile tests were performed, and the material
parameters were determined as being K = 1,275 MPa and n = 0.51.
To fully examine the effects of electricity on an air bending process, many dif-
ferent parameters were explored, while holding others constant (Table 11.1). For
all tests, a single displacement speed of 6.35 mm/min and a constant displacement
depth of 15 mm were used. A variable bending fixture was used to test two dif-
ferent bending diameters (38.1 and 50.8 mm). Additionally, two current densities
of 20 and 30 A/mm2 were tested. Preliminary testing resulted in the selections of
these current densities since they were above the “electrical threshold” (i.e., the
current density in which formability enhancement occurs). Pulse durations of 2
and 3 s were tested in this work, since preliminary testing revealed that these dura-
tions were most beneficial. Pulse period (i.e., time in between electrical pulses)
was also varied, and pulse periods of 60, 30, 20, and 15 s were tested.

11.1.6.1 Experimental Setup

The EAB experimental testing setup can be seen in Fig. 11.6. In the figure, a
manually adjustable bending fixture is utilized to produce a variety of die widths.
An insulated punch with a diameter of 15.9 mm was fabricated. Electrical pulses
were supplied to the bending fixture by a Lincoln R3R-500 constant current source
welding unit. Beneath the bending fixture, a Sensotec 22,250 N capacity load cell
with a signal conditioning box was used to record instantaneous bending force
data. The resolution of this system was ±4.45 N, which was able to read the low
bending forces as well as help illustrate the flow stress drop-offs due to application
of the electricity. This instrument was used because the resolution of the load cell
on the testing machine was too low to recognize such low bending forces.
Insulation was used in the construction of the bending fixture for two main rea-
sons: (i) to ensure that all electricity flows through the sheet metal test specimens
and (ii) to prevent the high electrical current from harming any sensitive electronic
devices. Fiberglass-reinforced plastic was used for the insulating material, due to its
strong/rigid compressive properties. This material was used in the construction of
the punch (to prevent electricity from traveling into the Instron testing machine), was
placed between the contact points and the bottom of the bending fixture, and was
inserted between the bottom of the bending fixture and the load cell to ensure all
components were completely isolated from electricity.
266 11  Applications of Electrically Assisted Manufacturing

Top Platen

Punch with insulation

Sheet Specimen
Electrical Cable
Adjustable Bending Die
with insulation

Insulation

Load Cell

Bottom Platen

Fig. 11.6  EAB test setup [1, 2]. All key components are listed, and the insulation component are
comprised of fiberglass-reinforced plastics for their compressive strength

There were two possible configurations for placing the insulation and directing
the electrical flow. The chosen configuration was to apply the electricity on one side
of the bending fixture and have it travel through the specimen to the other side of
the fixture. The top punch was completely insulated except for the tip. Another con-
figuration could have been to apply the electricity through the top punch. In this
case, the electricity enters the center of the specimen and then about half of the
electricity flows through one side of the bending fixture and the other half flows
through the other side. This configuration was not optimum in this case because
now the electricity was flowing in parallel and the voltage was reduced by half.
When first testing the bending fixture, some fundamental design flaws were real-
ized in the points of the fixture which contacted the sheet metal specimen and trans-
mitted the electrical current into the specimen. Figure 11.7a shows the initial design
of these components. Every time an electrical pulse was applied, the sheet specimen
would weld itself to the fixture and this would ultimately increase the bending force
after each electric pulse, until the small weld broke. This can be observed in the force/
displacement plot. Careful evaluation revealed that the large radii on these contact
points caused the line of contact between the fixture and the specimen to constantly
move with bending displacement, and the relative velocity between the dies and the
part is small. Due to the presence of asperities, until the contact is re-established, it is
possible that gaps form, thus allowing for welding to occur. As a solution to this, a sec-
ond design, with a very small radius, was tested. Because the radii on this design were
much smaller, the contact point between the fixture and the specimen did not change
as much. This design, coupled with a thin film of dielectric grease placed on the fixture
contact points, primarily prevented any welding. The result is force/displacement pro-
files as shown in Fig. 11.7b, where the electricity lowers the bending force.
11.1  EAF Bending Application and Model 267

(a)
120

Average Load [N]

100
80
60
40
20
0
0 5 10 15
15.9mm Dia. Round Position [mm]

(b)
120
100
Average Load [N]

80
60
40
20
0
0 5 10 15
Smaller Corner Radius Position [mm]

Fig. 11.7  Bending die design variants [1, 2]. A bending die with rounded edges allowed for
tiny micro-welds to form as deformation and electrical application took place. This caused steep
spikes in the mechanical load profile. This was alleviated by using sharper edges with smaller
corner radii. a Poor design—led to a varying contact point and welding. b Acceptable design—
led to a single contact point and no welding

Fig. 11.8  304 SS sheet
bending specimens [1, 38.1mm Die Classical
2]. Sheet metal bending Width / 91.5º EAF
specimens are shown with
two different die widths for Classical
classical and EAB tests
EAF

50.8mm Die Width / 109º

Examples of classical and EAB specimens for both die widths (38.1 and
50.8 mm) can be found in Fig. 11.8. From the figure, one can notice the difference
in springback (bend angle) between the classical and EA specimens. This will be
quantified later in this section.
268 11  Applications of Electrically Assisted Manufacturing

Fig. 11.9  EAB temperature 450

measurements [1, 2]. For all 400 1
EAB tests, the specimens

Temperature [°C]
350
were the hottest at location 1, 3
300 2
and the coolest at location 3 3 (30 Amps/mm 2 ) wd
250
200
1 (20 Amps/mm 2 )
150
2 (20 Amps/mm 2 )
100
50
3 (20 Amps/mm 2 )
0
0 10 20 30 40
Time[seconds]

11.1.7 Thermal Measurements in EAB

An important aspect of modeling EAB is the thermal/heat transfer of the process.

To help quantify the magnitude of heating, Fig. 11.9 displays thermal profiles
that were recorded using a FLIR infrared camera, from the time when an electri-
cal pulse was applied until the specimen temperature cooled significantly (note the
thermal profiles are for a single pulse only). Holding the die width (50.8 mm) and
pulse duration (2 s) constant, thermal profiles for current densities of 20 and 30 A/
mm2 were produced. The three specimen locations where temperature was recorded
were: (1) the die contact point (where electricity entered/exited the specimen), (2) a
point between the punch interface and die contact point, and (3) the punch interface
(i.e., the bending zone). The die contact point for the 20 A/mm2 test was almost
400 °C, and this was due to the minor sparking from the electricity entering the
specimen here. The bending of the part takes place at the top punch/specimen inter-
face; therefore, the temperature in this region will be the temperature of the metal
as it is being deformed. The most important location to monitor the temperature
is at the punch interface because this is where the actual bending takes place. The
maximum temperatures in this region for the 20 and 30 A/mm2 tests were about
115 and 335 °C, respectively. The temperatures resulting from these different cur-
rent densities were about 200 °C apart. Additionally, for the 20 A/mm2 test, temper-
ature profiles were considerably different for different locations on the specimens.
This shows that, with a thin sheet metal part, an increase in applied electrical power
could cause a large amount of global heating in the workpiece (as shown here
where there was a 200 °C difference from a 10 A/mm2 difference in starting current
density). This also shows that heat transfer is an important aspect of a bending pro-
cess, since there is not a large surface-to-surface contact area between the punch/
specimen and the bottom bending fixture/specimen.
11.1  EAF Bending Application and Model 269

Fig. 11.10  Model 120
verification for the classical model 38.1mm Die width / 91.5° Die angle

Average Load [N]

100
case [1, 2]. The model
experiment
and experimental profiles 80
are shown for the case of 60
classical bending for the two model 50.8 mm Die width /
different die widths 40 experiment 109° Die angle
20
0
0 2 4 6 8 10 12 14
Position [mm]

11.1.8 Validation of the Model via Experiments

The analytical model developed earlier was solved by following the solution scheme
from Fig. 11.5, and the bending force needed to reach a desired stroke or bending
angle was determined. The model was initially solved for the classical bending case
and its predictions were compared with experimental results for two different die
widths (Fig. 11.10). A friction coefficient of 0.25 was used, as recommended by pre-
vious research [5]. The predictions agree well for the two cases. The differences can
be attributed to the simplifications and assumptions of the model.
Next, the model for EAB was solved and compared with experimental results
for two current densities, 20 and 30 A/mm2. The variation of the strength, C, with
the temperature was determined by using the data from Fig. 5.6 [7], and this was
integrated into the model. The next step is to determine the EEC. Average EEC
values are taken from experimental results of previous research specific to the
same material, die speed, and current density [8]. The coefficients were approxi-
mated as ξ  = 0.18 for Cd  = 20 A/mm2, and ξ  = 0.35 for Cd  = 30 A/mm2. For
compression tests, the overall EEC profile with respect to time can be used, which
is more accurate than an approximated EEC value. However, in tension, the elec-
tricity is only applied for several seconds and an EEC profile as a function time
was not able to be determined for this case. These values were included in the
numerical solution. Figures 11.11 and 11.12 compare the model prediction with

Fig. 11.11  Model 120
verification for EAB at
Average Load [N]

100
CD = 20 A/mm2 [1, 2]. The
piece-wise model for the 80
38.1 mm die width and a
60
current density of 20 A/mm2 model
experiment
40
20
Cd = 20 Amps/mm 2
0
0 2 4 6 8 10 12 14
Position [mm]
270 11  Applications of Electrically Assisted Manufacturing

Fig. 11.12  Model 120
verification for EAB at model

Average Load [N]

100
CD = 30 A/mm2 [1, 2]. The
piece-wise model for the 80
38.1 mm die width and a
60
current density of 30 A/mm2
40
experiment
20
C d = 30 Amps/mm 2
0
0 2 4 6 8 10 12 14
Position [mm]

experiments for current densities of 20 and 30 A/mm2, respectively, at a pulse

duration of 2 s, and a pulse period of 30 s, for a die width of 38.1 mm.
It can be observed again that the predictions agree well with the experimen-
tal results for both cases. The differences can be attributed to the simplifications
and assumptions done in the model, but also to the approximation of the material
strength, C, and EEC, ξ. Future work should include refinement of the model to
account for the factors that were neglected, i.e., the strain-rate sensitivity, and also
to conduct specific experiments targeting accurate measurements of material prop-
erties included in the analytical model.

11.1.9 Effects of Electricity in Bending

The aspects considered for evaluation and comparison were related to the bending
load and amount of springback after testing. In the next subsections, bending load
versus position plots will be used to examine the effects of different pulsing param-
eters (current density, pulse duration, and pulse period), while changes in the bend
angle will be used to quantify electricity’s effectiveness at reducing springback.

11.1.9.1 Effect of Current Density

Figures  11.13 and 11.14 display experimental EAB tests and modeled load ver-
sus position profiles in which die width, pulse duration, and pulse period are held
constant, while two different current densities of 20 and 30 A/mm2 are evaluated.
Figure  11.13 is a plot of the specimens tested with a die width of 38.1 mm and
Fig. 11.14 represents tests performed on a die which is 50.8 mm wide. The pulse
duration for both plots was 2 s, and the pulse period was 30 s.
It can be seen in both figures that the electric pulses resulted in instantaneous
flow stress “drop-offs,” which were proportional to the magnitude of the current
density. Specifically, for the 38.1 mm die width, the flow stress drop-offs of the
30 A/mm2 test were up to four times greater than the 20 A/mm2 drop-offs. For the
11.1  EAF Bending Application and Model 271

Fig. 11.13  Forming load 120

recorded for wd = 38.1 mm Classical

Average Load [N]

100
[1, 2]. Displayed are the
forming loads for the 80
classical and two EAB tests 60
parameters at a die width of
38.1 mm 40
C d = 20 Amps/mm 2
20
C d = 30 Amps/mm 2 w d = 38.1 mm
0
0 2 4 6 8 10 12 14
Position [mm]

Fig. 11.14  Forming load 75
Classical
recorded for wd = 50.8 mm
Average Load [N]

[1, 2]. Displayed are the

forming loads for the 50
classical and two EAB tests
parameters at a die width of
50.8 mm 25
C d = 20 Amps/mm 2
C d = 30 Amps/mm 2 w d = 50.8 mm
0
0 2 4 6 8 10 12 14
Position [mm]

50.8 mm die width, the 20 A/mm2 drop-offs were more noticeable and the 30 A/
mm2 drop-offs were only about twice as large. In Fig. 11.13, the drop-offs for the
30 A/mm2 test were larger than those of the 20 A/mm2 test. However, the aver-
age bending load profile throughout the test remained relatively the same between
the two current densities (note that the load profile was reduced by about 20 N in
comparison with the respective non-pulsed classical test). In Fig. 11.14, the drop-
offs were again more significant for the higher current density. The bending load
profile produced from the 20 A/mm2 test showed no notable reduction compared
to the respective classical test, whereas the 30 A/mm2 test posted a bending load
profile that was about 15 N lower than the conventional load profile. When com-
paring the two figures, the 38.1 mm die width allowed for greater reductions in
the load profile, where both current densities lowered the load profile by 20 N.
However, with the 50.8 mm die width, there may be a type of threshold reached
since the 20 A/mm2 profile was not reduced compared to the conventional profile,
but the 30 A/mm2 profile was reduced by about 15 N.
The amount of springback after the test, SB, was determined from Eq. (11.21).
β − βr
SB = · 100 [%] (11.21)
β
where β is the angle at which the specimen was bent (β = 180° − α), and βr is the
real angle, measured after the test. The application of electricity during the defor-
mation process resulted in lower springback (up to about a 20 % overall reduction
272 11  Applications of Electrically Assisted Manufacturing

in springback). However, the springback reduction magnitudes will be far less than
those achieved by Green et al. because of differences in when the electricity was actu-
ally applied [9]. Specifically, Green et al. applied the electricity via a single pulse
for a specified pulse duration after deformation, but before removal from the die.
This method is effective at eliminating a considerable amount of springback effects
because deformation has already ceased and there will be no more dislocation pile-
ups after the pulse is applied. Conversely, in this research, the electricity was applied
during deformation. In this work, the formability increase due to electricity was the
main objective to analyze, and the reduction in springback was a beneficial side effect.
Table 11.2 shows an overview of the EAF springback works by Green et al. and
Salandro et al. Aside from the fact that Green et al. applied the electrical pulse after
deformation, there are also other differences. Specifically, the materials, the specimen
dimensions, the bending diameters, and the electrical pulse parameters were differ-
ent. The two different materials used have different thresholds, where the electric-
ity begins to improve the formability (60 A/mm2—Green, 18 A/mm2—Salandro) [8].
The specific goal of Green et al.’s research was to determine parameters by which
springback could be eliminated; however, the research by Salandro et al. examined
springback as a side effect to the applied electricity that reduced the bending forces.
When applying the electricity during deformation, the electrical application parame-
ters (current density, duration, and period) become important, since deformation will
continue. For example, if pulse periods are spaced such that the last pulse is right at
the end of deformation, then it will have a greater impact on springback than a pulse
period that allowed for a given amount of deformation after the last pulse. To this
end, it is hard to directly compare both of the works. However, both works show that
applied electricity has an effect on reducing the springback in formed parts. A further
in-depth analysis of EAF and springback reduction will come later in this chapter.
In order to have a better evaluation of the effect of electricity, a springback
reduction was computed by comparing the EAB tests with the classical test (or
Cd = 0 A/mm2), as shown in Eq. (11.22).
SBclassical − SBEAF
RedSB = · 100 [%] (11.22)
SBclassical
Figure  11.15 displays the percent of springback that is still in the sheet specimens
after bending, for both die widths, pulse duration of 2 s, a pulse period of 30 s, and
current densities from 0 to 30 A/mm2. From this, it can be seen that the percent of
springback in the specimens is reduced from the application of electricity dur-
ing deformation. Also, when just focusing on the tests run at 0 A/mm2 (classical
tests), the die width plays a notable role in the amount of springback in the part.
Specifically, a die width of 38.1 mm results in about 14 % of springback in the part,
while a die width of 50.8 mm results in about 12 % springback. Figure 11.16 still
represents the same tests as in Fig. 11.15, but now presents them in terms of the per-
cent of springback reduction (or how much springback was removed from the parts
compared to the springback in the respective classical tests). This figure also confirms
that, as current density is increased, the percent reduction of springback will also
increase. Further, a smaller die width will help to amplify the springback reduction.
Table 11.2  EAF application parameters and resulting springback effects [1, 2, 9]
Researcher Matl. Specimen dims. (mm) Bending dia. (mm) Electrical application descriptions Results
Green et al. Al 6111 250 × 3,750 × 0.91 101.6 A single electrical pulse was Springback completely eliminated with
applied after deformation a 1.5 s pulse at 120 A/mm2
Single-pulse duration of 1.5 s
Current densities of 70–135 
11.1  EAF Bending Application and Model

A/mm2
Salandro et al. 304 SS 12.38 × 127 × 0.864 38.1 and 50.8 Multiple electrical pulses applied 27 % springback red (see CD subsection)
over distinct periods during
deformation
Two current densities 80 % springback red (see pulse duration
(20 and 30 A/mm2) subsection)
Two pulse durations (2 and 3 s) 50 % springback red (see pulse period
Four pulse periods (15, 20, 30, subsection)
60 s)
273
274 11  Applications of Electrically Assisted Manufacturing

15%

Percent of Springback
w d = 38.1mm
Series1
w d = 50.8mm
Series2
10%

5%

0%
0 20 30
Current Density [Amps/mm 2]

Fig. 11.15  Recorded springback for EAB and conventional bending [1, 2]. Percent of spring-
back recorded for a classical bending test and for the different EAB tests

30
Percent of Springback

wd = 38.1mm
25
Reduction [%]

wd = 50.8mm
20
15
10
5
0
0 5 10 15 20 25 30 35
2
Current Density [A/mm ]

Fig. 11.16  Springback reduction due to electricity [1, 2]. Percent of springback reduction in

comparison with the amount of electrical current applied

11.1.9.2 Effect of Pulse Duration

This subsection evaluates the isolated effects of pulse duration on the bending load
and springback of the 304 Stainless Steel sheet specimens. Figure 11.17 presents
bending load versus position profiles for tests with a constant die width (38.1 mm),
current density (30 A/mm2), and pulse period (60 s), while having pulse dura-
tions of 2 and 3 s. From the graph, the bending load was slightly lowered in both

Fig. 11.17  EAB forming 120

Classical
loads for different pulse
Average Load [N]

100
durations [1, 2]. Forming
load profiles with electrical 80
pulse durations of 2 and 3 s 60
(constant pulse period, die t = 2 sec
width, and current density) 40
t = 3 sec
20
C d = 30 Amps/mm 2
0
0 2 4 6 8 10 12 14
Position [mm]
11.1  EAF Bending Application and Model 275

Fig. 11.18  Springback

Percent of Springback
80
reduction for various pulse

Reduction [%]
duration [1, 2]. Percent of 60
springback reduction in
relation to the duration of 40
each electrical pulse (constant w d = 38.1mm
pulse period, die width, and 20
c d = 30 Amps/mm 2
current density) 0
0 1 2 3
Pulse Duration [s]

electrical tests. Also, there was only about a 5–10 N difference in flow stress drop-
offs between the 20 and 30 A/mm2 tests, with the 30 A/mm2 test producing the
larger drop-offs.
The percent of springback reduction as a function of pulse duration can be
shown in Fig. 11.18. For this graph, a constant current density of 30 A/mm2, die
width of 38.1 mm, and pulse period of 30 s were used. The percent reduction in
springback is significantly dependent on the length of the pulse duration. Of note
is that, using a pulse duration of 3 s, and pulsing every 30 s during deformation,
about 77 % of the total springback can be removed from the specimens.

11.1.9.3 Effect of Pulse Period

The following figures help to illustrate how different pulse periods (time in
between successive electrical pulses) can influence bending load and springback.
Figure  11.19 displays plots of load versus position data where a constant current
density of 30 A/mm2, a constant pulse duration of 2 s, and constant die width of
38.1 mm were maintained while pulse periods of 20 and 30 s were utilized. The
flow stress was reduced by about 5–10 N for the 30-s pulse-period test and by about
15–20 N for the 20-s pulse period test, when compared to the classical test. The
more frequent pulses of the 20-s-period test provided for a slightly lower flow stress
profile compared to the 30-s-period test. Additionally, the drop-off magnitudes of

Fig. 11.19  EAB forming 120

Classical
loads for different pulse
100
Average Load [N]

periods [1, 2]. Forming load

profiles are presented which 80
represent a 20 and 30 s pulse
60
period in between successive
electrical pulses (constant 40
p = 20 sec
pulse duration, die width, and
20
current density) p = 30 sec C d = 30 Amps/mm 2
0
0 2 4 6 8 10 12 14
Position [mm]
276 11  Applications of Electrically Assisted Manufacturing

Fig. 11.20  Springback 60

Percent of Springback
reduction for various pulse 50

Reduction [%]
periods [1, 2]. The percent 40
of springback reduction in
30
relation to the period of time
in between each successive 20 w d = 38.1mm
pulse is presented (constant 10
c d = 30 Amps/mm2
pulse duration, die width, and 0
current density) 0 10 20 30 40 50 60 70
Pulse Period [s]

each electrical test were very similar, leading to the fact that pulse period has little
effect on drop-offs and more of an effect on the overall flow stress profile.
Figure  11.20 displays the amount of springback reduction as a function of
pulse periods of 20, 30, and 60 s, while holding constant the die width (38.1 mm),
current density (30 A/mm2), and pulse duration (2 s). The shortest pulse period,
allowing for the most electric pulses, reduced the springback the greatest, by
about 50 %. There was no notable difference between the percent of springback
reduction between the 30- and 60-s-period tests. For the given current density, die
width, and pulse duration combination, a pulse period of 20 s appears to be opti-
mal; however, this will change as the other pulse parameters are changed.

11.1.10 EAB Model Conclusions

These previous subsections in this chapter presented a preliminary approach to the

analysis of electrically assisted air bending by employing equilibrium of forces and
moments, and also utilizing the same EEC-based methodology as was used for an
EA forging operation. Two effects of the applied electrical energy were taken into
consideration: (i) the electroplastic effect of the electricity helping the deformation
by assisting the dislocations and (ii) thermal softening as a result of resistive heating.
The bending force needed for a certain deformation can be determined for both
classical and electrically assisted tests. For the second case, the pulse parameters
can be varied, and the effect on the deformation process can be estimated, in terms
of force reduction. The following conclusions were drawn from this study:
• The experimental results indicated that the electricity can assist the deformation
process due to beneficial effects in reducing the flow stress; thus, the forming
force and mechanical energy required for reaching the same level of deforma-
tion as in classical forming are reduced.
• Another benefit of using the electricity during bending was observed in the
notable springback reduction, up to 77 %, depending on the electric parameters
and pulse characteristics.
• The analytical model was in good agreement with the experiments, and it was
able to predict the bending forces within a difference of 10–15 % (in most cases).
11.1  EAF Bending Application and Model 277

• The analytical model coupled with the experimental observations can be used to
estimate the pulse parameters required for certain reductions in the forming load.
• A model was created to account for the reduction of springback effect in EAF
bending as a function of process parameters.

11.2 Electrically Assisted Machining

The EAM technique has not previously been applied to machining; however, based
on the results seen in EAF, the potential exists to alter the machinability and manu-
facturing processes employed for difficult-to-machine materials. Only brief pre-
liminary testing has been conducted in this area, in which strain rates much lower
than typical machining rates were used. These preliminary electrically assisted tests
yielded improved surface quality of the workpiece and reduced machining forces.
The application of electrically assisted machining holds opportunities to improve
the conventional machining method through the possibility of enabling higher mate-
rial removal rates, reduced cutting forces, modified tool wear, and improved work-
piece surface properties.

11.2.1 Observations in Low-Strain-Rate EA Machining

Initial testing has been conducted in this area using strain rates much lower than
typically found in machining. The electrically assisted machining approach was
applied to a skiving-type machining operation while applying direct current
flow from the cutting tool to the metallic workpiece. The results of the electrical
machining test were compared against a baseline model, without electrical current
flow. The electrically assisted test yielded improved surface quality of the work-
piece and reduced machining forces. Below is a brief description of the experi-
mental setup, results, and findings from this research.

11.2.1.1 Experimental Setup

Preliminary electrically assisted machining tests were performed to investigate the

effect of electrical current on machining. As shown in Fig. 11.21, direct current (DC)
was applied directly through the cutting tool into the workpiece across the cutting
zone. Relative motion between the tool and workpiece is achieved by linear motion.
The feasibility prototype was constructed for a skiving-type operation of an
A2 Tool Steel workpiece, where the tested surface speed was 16.93 mm/s, which
is significantly lower than typical machining surface speeds. However, this was
an initial test and its purpose was to simply explore possibilities for electrically
assisted machining. The forces were measured with and without electric current
applied, and the resultant surfaces examined for topography.
278 11  Applications of Electrically Assisted Manufacturing

Fig. 11.21  Schematic and prototype of electrically assisted machining by skiving [10]

11.2.1.2 Results

For this experimental analysis, a baseline test (i.e., no electricity) as well as tests
with the application of current through the tool into the cutting region (EAM) was
examined. The preliminary results indicate a reduction in the machining force and
improved surface quality of the workpiece. Figure 11.22 shows the steady-state
force profile for the baseline and EAM tests. It can be observed that the test using
the application of current resulted in an approximate 25 % decrease in cutting
force over a large range of the skiving stroke.
Figure 11.23 depicts a comparison of the resultant surfaces for both tests as meas-
ured using scanning white light interferometry. The baseline test exhibits a rough
surface, similar to a chatter condition, while the EAM test shows an elimination

Fig. 11.22  Steady-state
tooling forces for A2 tool
steel workpiece
11.2  Electrically Assisted Machining 279

Fig. 11.23  Cutting channel
surface profiles in A2 tool
steel workpiece

of this chatter condition. This can also be observed in the force profiles for the two
tests, where the EAM force profile shows less fluctuation in force than the baseline.
The chatter condition was a result of the processing parameters and was not actually
desired during the testing.

11.2.2 High-Strain-Rate Process Modeling and

Experimental Testing for EA Machining

With the observation of lower forces in EA machining, a bulk force model is

developed to predict average process behavior. The approach is to augment a tra-
ditional orthogonal machining force model to capture the modification to material
strength due to the applied electricity.

11.2.2.1 Empirical Relationship

The classical Merchant orthogonal machining model is modified by changing the

shear stress, τ, to account for electrical effects. The shear stress under electrical
effects was related to the forming stress under electrical effects, as found by previ-
ous EAF work. The shear stress is assumed, through Tresca material failure theories,
to be one half of the normal stress. From the EAM work of Jones et al. in 2011, the
normal flow stress of a material under electrical effects is predicted using:
 
σEA = K ′ εn ε̇m exp −(ε − C)B�A (11.23)

where σEA is the predicted flow stress with the application of electricity, K′ is
the strength coefficient, ε is the strain, n is the strain hardening exponent, ε̇ is the
strain rate, m is the strain-rate sensitivity exponent,  is the current density, and A,
B, and C are material-specific constants for this model [11].
When considering the current density, since the current varies as a function of
contact area with the shear zone, the contact area due to the formation of chips could
have an effect on the instantaneous current density. Neglecting the effect of dynamic
chip formation, the instantaneous current density during the test can be described by:
280 11  Applications of Electrically Assisted Manufacturing

I
= (11.24)
wt0
where  is the current density, I is the applied current, and wt0 is the contact area
of the tool with the workpiece during the testing. Thus, the current density is a
function of both the width of the cut as well as the depth of cut.
The combined model for machining force under an electric current is given by:
 
K ′ εn ε̇m exp −(ε − C)B�A wt0 cos (β − α)
FC,EA = (11.25)
2 sin (ϕ) cos (ϕ + β − α)

where τ is the average shear stress along the shear plane, w is the width of the cut,
t0 is the undeformed chip thickness, ϕ is the shear angle, β is the friction angle,
and α is the rake angle.

11.2.2.2 Experimental Setup

To validate and examine the results from the derived models, experimental tests
were performed with varying test conditions. The testing conditions are listed in
Table  11.3. As can be seen, four current magnitudes were examined (remained
constant during test) where Parameter Set A represents conventional room tem-
perature machining without applied current.
The machining operation was performed by a vertical milling machine, in
which the workpiece was rotated by the spindle and the cutting tool was manu-
ally fed into the rotating workpiece by the traversing table. The current was pro-
vided by a Darrah 4-kA power supply to a plate surrounding the workpiece. A
LabVIEW virtual instrument (VI) was used to control the applied current. For
the tests in which current was applied, the VI control system applied the specified
current magnitude for a specified number of seconds once the cutting thrust force
reached a 400 N threshold. The current followed a path through the support of the
workpiece, through a tapered roller bearing infused with conductive grease, to the
workpiece itself, and out through the tool. The workpiece and tool were insulated
from the milling machine by an FR4 dielectric fiberglass laminate epoxy resin
substrate. The machining forces during testing were measured by a six-axis force
dynamometer. This setup is illustrated in Fig. 11.24.

Table 11.3  Testing Parameter set Applied current

conditions for EA Machining magnitude (A)
A 0
B 100
C 300
D 500
11.2  Electrically Assisted Machining 281

Mill Head

Pos. (+)

Neg. (-)

Fig. 11.24  Electrically assisted machining experimental setup. Schematic (above), physical

detail of test sample and cutting tool (below) [10]

The disk-like workpiece specimens were produced from rod stock of grade 5 tita-
nium to have a diameter of 38 mm and thickness of 4 mm. The disk specimen was
then attached to a 1,018 steel shaft, which was supported by a plate, to which the cur-
rent source was connected. To ensure testing consistency, the specimens were pro-
duced from the same rod stock and were held to tolerances of 0.1 mm. The cutting
tool was made of uncoated tungsten carbide (WC). The rotational speed of the work-
piece during testing was 350 rpm and the feed rate averaged 0.015 mm/revolution.

11.2.2.3 Results and Discussion

The Merchant orthogonal machining model was used to predict the machining
cutting force experienced in empirical conventional machining tests. Baseline
machining tests were run to solve for the material-specific machining parameters
(e.g., coefficient of friction and shear strength). These variables were determined
by applying the machining cutting force model to the empirical average cutting
force for baseline tests and performing least squares regression to minimize the
error between the model and empirical results.
The coefficient of friction process parameter found empirically from the base-
line tests was then applied in the model for tests undergoing applied DC during the
machining operation. The average cutting force results from the experimental tests
were used to solve for the modified material shear strength for the respective mag-
nitude of applied current using least squares regression. The results for the mate-
rial shear strength are shown in Fig. 11.25.
As shown, the material shear strength was significantly reduced with an increas-
ing applied current magnitude. Baseline tests with no applied current demonstrated
282 11  Applications of Electrically Assisted Manufacturing

Fig. 11.25  Shear stress
versus applied current
magnitude

Fig. 11.26  Mean model
error for predicted versus
experimental machining force

material shear strength of 585 MPa, while 100, 300, and 500 A demonstrated reduced
material shear strength of 476, 355, and 350 MPa, respectively. As shear strength of
the material directly relates to the machining cutting force required to machine the
material, the machining cutting force under applied current of magnitudes 100, 300,
and 500 A demonstrate a reduction in machining cutting force. The results of the effect
of electricity on material shear strength obtained in this testing indicate that there may
be an asymptotic limit to the achievable force reduction enabled by the application of
increasing current magnitude. Further testing could be performed to better define the
surface of the curve relating current magnitude to material shear strength modification
as well as to determine a possible limit to the achievable strength reduction.
The error between the cutting force model incorporating a modified material
shear strength process parameter and the average cutting force found empirically
is given in Fig. 11.26.
As shown in Fig. 11.26, the coefficient of friction and shear strength deter-
mined for the grade 5 titanium workpiece material fit the empirical tests with less
than 6 % error for the baseline tests. There was less error present for the tests with
applied current. Friction was held constant in this testing between baseline tests
and tests with applied electric current to quantify the effect of electric current on
material shear strength. However, it is possible that the application of electricity
and its resistive heating effects could affect the tool-chip friction experienced in
the machining operation, as well.
11.2  Electrically Assisted Machining 283

Fig. 11.27  Normalized cutting force with respect to average depth of cut for 0 and 500 A

It should be noted that the feed, which relates directly to the machining thick-
ness or cutting depth, was not precisely controlled during the machining tests and
therefore does not remain uniform throughout the test. The average depth of cut
for each respective test was used for all cutting force analysis. As the machining
cutting force is dependent upon the depth of cut, each parameter set was normal-
ized with respect to average feed rate, or cutting depth, during the test for com-
parison. This allowed the force results to be compared across different current
application levels. Figure 11.27 is an example that compares the normalized cut-
ting force at 0 A to the normalized cutting force at 500 A. As seen, the 500 A
results portray an approximate 40 % reduction compared to the baseline tests with
no current applied.

11.2.3 EA Machining Conclusions

The classical orthogonal machining model was augmented using empirically

found material strength relationships. The typically observed material strength
reduction in electrically assisted forming is also present in the tested architecture
of electrically assisted machining. Material shear strength; therefore, machin-
ing cutting force was observed to experience up to 40 % reduction. Additionally,
forces are reduced from traditional machining and exhibit lower fluctuation, indi-
cating that this process may be amenable for high-strength materials or processes
where dynamic chatter phenomena are a concern.
284 11  Applications of Electrically Assisted Manufacturing

11.3 Electrically Assisted Friction Stir Welding

The process of friction stir welding involves high tool forces and requires robust
machinery; hence, the forces involved make tool wear a predominant problem. As
a result, many alternatives have been proposed in decreasing tool forces, such as
laser-assisted friction stir welding and ultra-sound-assisted friction stir welding.
However, these alternatives are not commercially successful on a large scale due
to scalability and capital/maintenance costs.
In an attempt to reduce forces in a cost-effective manner, electrically assisted fric-
tion stir welding (EAFSW) is studied in this work. EAFSW is a result of applying
the concept of electrically assisted manufacturing (i.e., passing high direct electri-
cal current through a workpiece during processing) to the conventional friction stir
welding process. The concept of EAFSW is a relatively new adaptation of conven-
tional frictional stir welding, which is well established. The expected benefits are
reduction in the feed force and torque, which allow for improved processing produc-
tivity as well as the possibility for deeper weld penetration.
The effect of passing direct electric current through Aluminum 6061 during the
welding process is studied in this section with respect to the feed force and torque.
From this study, it is shown that the feed forces are reduced by 58 % on average
in the EAFSW process compared to a conventional frictional stir welding process.
Also, a decrease in torque at the start of the feed is present in the EAFSW process
as compared to conventional friction stir welding.

11.3.1 Electrically Assisted Friction Stir Welding

Background

Electrically assisted friction stir welding (EAFSW) is a variant of traditional FSW,

in which the EAM processing methodology is integrated into the FSW process
by passing electricity through the workpiece during the welding process [12]. As
a result of the high forces associated with conventional friction stir welding, the
process of EAFSW could elevate these excessive forces and reduce the required
machine size and stiffness. Additionally, as most heat is generated by the tool
flange, this limits the processing depth of the traditional FSW process. With the
incorporation of EAFSW, this limited penetration depth can be eliminated as the
flow of electrical current creates both a temperature rise and direct material soften-
ing. The implementation of this process will require additional safety and insula-
tion measures; however, the cost is expected to be lower than that for comparable
technologies such as laser-assisted friction stir welding.
The application of EAM to FSW has only started very recently. In 2005, Long
et al. performed finite element modeling of an electrically enhanced FSW process
and concluded that plunge force can be considerably decreased by passing electric
current and also stated that the temperature profile can be achieved in half the time
11.3  Electrically Assisted Friction Stir Welding 285

compared to conventional FSW [13]. Though they applied the concept of EAM to
FSW process, they did not perform any experimental work to validate their model
results. In 2008, Ferrando experimentally studied the EAFSW process and con-
cluded that a significant decrease in Z-axis force is achievable in an EAFSW process
compared to a conventional FSW process [14]. This work also stated that there may
be a decrease in tool wear by using EAFSW. Yet, there was no data reported on the
feed forces and torque during the EAFSW process, which are important measures
of process productivity. In 2010, Pitschman et al. [15] reported that there is a scope
to increase the weld speed and decrease the power consumption by using EAM in
FSW; however, the feed force reduction was not quantified in this work. As there is
a limited amount of research that has been performed for EAFSW, this work is the
first to examine and quantify the effects that an applied electrical current has on the
feed force and torque. The feed force and torque are significant process parameters
as they dictate the achievable weld speed and allowable weld penetration.

11.3.2 EAFSW Experimental Setup

The specimens for this study consist of two Aluminum 6061 plates with dimen-
sions of 12.7 mm × 3.175 mm and an aluminum block of 16.38 mm wide
by 8.25 mm thick. These specimens are welded together to make a T-lap joint.
Figure 11.28 shows the specimens used for this study prior to welding.
A conventional vertical milling machine with an electric motor for automatic
feeding of the table is used in this study. The speed of the spindle is kept constant at
350 rpm for all cases, and the feed rate is also kept constant at a rate of 101.6 mm/
min. The tool is made of hardened A2 tool steel. The tip of the tool is 6.35 mm both
in length and diameter and threaded for better material flow. The diameter of the
shoulder of the tool is 19.05 mm. The tool is completely insulated from the machine
to prevent the flow of current from the workpiece into the machine. Figure 11.29
presents an image of the tool used in this study along with a joined sample.
Figure 11.30 presents the experimental setup used in this study, and Fig. 11.31
displays a schematic of the experimental setup. Copper wires from a 4-kA DC
current supply were used to supply electrical flow into the workpiece. The copper

Fig. 11.28  T-shaped
specimen for EAFSW [16]
286 11  Applications of Electrically Assisted Manufacturing

Fig. 11.29  Hardened A2 steel tool used in this study with welded sample [16]

Fig. 11.30  EAFSW experimental setup [16]

wires were joined to the copper clamps attached to the workpiece by mechanical
fastening. The workpiece specimens were held in a vice for welding, and the vice
was electrically insulated from the workpiece.
The forces involved during the process were measured using a 6-axis Kistler
force dynamometer. The force dynamometer was attached to the vice holding the
workpiece. A sampling frequency of 3,000 Hz was used to record the process results.
To begin the weld, a hole is drilled into the workpiece to minimize the forces of
plunging on the tool. The drilling operation is performed on the same conventional
11.3  Electrically Assisted Friction Stir Welding 287

Fig. 11.31  Schematic diagram of the EAFSW experimental setup [16]

vertical milling machine using a drilling tool. The FSW tool rotating at 350 rpm is
manually fed downward during the plunging stage and locked when the shoulder
of the tool touches the top surface of the specimens. As soon as the shoulder of the
tool is in complete contact with the workpiece, current was allowed to flow through
the workpiece for the EAFSW tests. A dwell period of 5 s is used in both the con-
ventional FSW and EAFSW tests. The start of the dwell period is defined as when
the shoulder of the tool is in complete contact with the workpiece and the end of
dwell period is when the feed is stared. It should be noted that, for the EAFSW
tests, the start of the dwell period and the start of the current are at the same time.
In the EAFSW tests, the current was discontinued at the end of the feed.
Continuous DC current of 1,500 A, which results in a current density of
37.2 A/mm2, is passed through the workpiece for the EAFSW tests. A program in
LabVIEW software was developed to monitor the flow of current into the work-
piece. Both the conventional FSW and EAFSW tests were performed in a random
order and replicated three times to ensure consistency of the results.

11.3.3 Results and Discussion

The data collected using the force dynamometer are analyzed by plotting of the
feed force and torque with respect to time. The results are filtered with a mov-
ing average of 1,000 data points, and the results from the same type of tests
(conventional FSW or EAFSW) are averaged to obtain a single process curve
(three tests per test type). In the following sections, the process curve of con-
ventional FSW is compared to EAFSW through the use of the process feed,
force, and torque.
288 11  Applications of Electrically Assisted Manufacturing

11.3.3.1 Comparison of Feed Force

Figure  11.32 shows the comparison of feed forces in conventional FSW and
EAFSW with respect to time. The curves are aligned such that the time zero on the
abscissa is when the feed was started and the error bars represent ± one standard
deviation. From the figure, it can be concluded that the overall feed force in an
EAFSW process is less than that in a conventional FSW process. The decrease in
feed force on average is approximately 58 %.
The large increase in force at the start of the welding process indicates the
beginning of the tool feed. After the initial peak at the start of the feed, the magni-
tude of feed force decreases slowly until the feed is stopped.
In the conventional FSW process, the material is subjected to relatively cold
material flow at the start of the feed and the transportation of the material by the
tool involves high frictional force, as there is little thermal softening of the mate-
rial. Again, only a small dwell time was used before the feed was started. Due to the
limited amount of thermal softening, which creates high resistance to material flow,
a higher feed force is required at the start of the feed. As the weld progresses, heat
is generated from the friction between the tool and the workpiece and from material
deformation. The heat generated from the friction between the pin of the tool and
workpiece and due to material deformation is small compared to the heat generated
due to the friction between the shoulder of the tool and workpiece. Hence, the design
of the tool flange is critical in FSW processing. As the tool traverses the weld line,
the feed force decreases from the peak force and nearly reaches a steady-state force
as the heat transfer begins to equalize. This means that the heat generated is equal to
the heat loss from the system; thus, the temperature during the weld is stabilizing.

Fig. 11.32  Feed force versus time

11.3  Electrically Assisted Friction Stir Welding 289

Conversely, the electrical test shows reduced feed forces during the entire weld-
ing cycle. From the addition of electrical current, the material has additional heat
generation due to the electrical current scattering off of defects within the material
lattice. With the scattering of electrons, this increases the vibrational energy of the lat-
tice, which is observed on a macroscale as a temperature rise. Due to the influence of
Joule heating being an instantaneous effect all throughout the material in comparison
with the processing cycle time, the addition of current has great potential to reduce
the required heat needed from the flange/workpiece interaction. It is shown that there
is a larger reduction at the beginning of the test as the baseline test has not generated
much heat; however, the addition of current helps with heat generation and reduces the
feed force. As the test progresses, the heat generated due to friction becomes so great
that the addition of electrical current is not as significant. This may be a result of the
additional electrical energy not having a large effect due to the material lattice being
at such an elevated vibrational state (i.e., lattice energy becomes saturated). Therefore,
the experimental results are unable to quantify the electroplastic effect for EAFSW or
determine whether there is any interaction that results in the observed decrease in force.

11.3.3.2 Comparison of Torque

The torque required by the tool during the welding process is compared between
the conventional FSW and EAFSW processes in Fig. 11.33. The time zero on the
abscissa indicates the start of the feed.
From Fig. 11.33, it is noticed that the torque required at the start of feed for
the EAFSW is less compared to the conventional FSW process. For the rest of the

Fig. 11.33  Torque versus time

290 11  Applications of Electrically Assisted Manufacturing

process, the torque required by the tool is approximately equal for both the con-
ventional FSW and EAFSW processes.
For the baseline test, the torque required at the start of the feed is higher as
the material undergoes relatively cold material flow due to the lack of heat gener-
ated by the flange at that time. After the start of the feed, the temperature of the
material rises due to friction between the shoulder of the tool and the material,
which causes the material to thermally soften. So, the torque required to transfer
the material drops slightly from approximately 10–20 s. As the welding process
continues, the temperature of the material rises significantly, and this causes the
material to thermally expand. The thermal expansion of the material increases
with temperature rise and the tool needs more and more torque to rotate as the
material pushes back against the tool flange. The sudden drop in torque after the
continuous period of increasing torque indicates the end of the welding process.
The torque needed at the start of the feed in the EAFSW process is less than
the conventional FSW process. The additional of the electrical current induced
some thermal softening such that the material flowed with less resistance; how-
ever, it appears that the temperature did not rise a significant amount in order to
create a great amount of thermal expansion. This is seen by the EAFSW torque
remaining relatively constant until 15 s where excessive heating from fric-
tion causes the material to increase in temperature. The EAFSW test’s torque
increased just as the baseline test, as a result of thermal expansion. Again, the
thermal expansion applied a greater pressure on the tool flange which increased
the required torque. It can be seen that the torque of the EAFSW tests increases
earlier in the process as compared to the baseline test. This can be explained
by the fact that the influence from Joule heat combined with heat generation
by friction caused a faster rise in temperature which equates to greater thermal
expansion, thus greater torque.
Again, the influence of electroplasticity which is observed during deformation
by slip is not directly quantifiable in these results.

11.3.3.3 Post-Weld Analysis

In this section, the welded cross section is examined and the results are presented in
Fig. 11.34. The cavity in the aluminum workpiece is due to the flow of the material (in
the axial direction of the tool) to the surface from the tool being threaded. Thus, the
cavity present is purely a result of tool design. Also, it can be observed that the weld
cavity is larger at the start of the weld (left) as compared to a section in the middle of
the welded length (right). This is a result of a larger amount of material not being pre-
sent at the beginning of the weld due to the pre-drilled starter hole for the FSW tool.
The cavity was observed for both the conventional FSW and EAFSW processes.
During the welding process, the material is transferred from the leading edge to
the trailing edge by the pin of the tool. Aside from the cavity due to the tool geom-
etry, excellent material flow is present throughout the cross section creating one
bulk workpiece.
11.3  Electrically Assisted Friction Stir Welding 291

Fig. 11.34  Cross section of welded specimen with (left) start of the weld and (right) middle of
the weld [16]

11.3.4 Conclusions and Future Work

This work showed the first experimental modification of the welding feed force and
torque during the process of EAFSW. As a result, these preliminary results show
great potential for improved welding speed and penetration depth. Additionally,
these results suggest the possibility of reduced tool wear during plunging and weld-
ing due to enhanced material softening. Also, as the application of current is capa-
ble of reducing the material strength, significantly stronger or thicker materials may
be welded with the EAFSW process where they are prohibited by traditional FSW.
Moreover, with the reduction in forces (on average 58 %) and torque at the begin-
ning of the weld, smaller capacity machines could be used or adapted to perform
FSW with electrical assistance. This could reduce the capital cost for manufactures
that use this type of welding process.
Further work is examining the interactions of processing conditions, tool geom-
etry, and electric current magnitude on the material flow and heat generation during
the welding process. Also, the microstructure evolution during the process should be
analyzed to compare the traditional FSW process to the EAFSW process.

11.4 Experimental Findings for Alternative EAF Processes

Throughout the last decade, many works have been published which experimen-
tally examine the effect of EAM on different processes. In the following sections,
experimental works involving EAM’s impact on uniaxial compression, uniaxial
tension, non-uniform deformation, and springback will be discussed. Included in
these discussions will be examples of work by the authors, as well as work by
other EAM researchers.
292 11  Applications of Electrically Assisted Manufacturing

11.4.1 Compression

Explanations of two key EAM compression works will be discussed. In the first
work, EAM compression tests will be run on various types of metals, and using
several size specimens [8]. The second research work focuses on the impact that
manufacturing speed has on Mg compression specimens [17].

11.4.1.1 EAM Compression Results on Various Metals

In the work by Perkins et al., EAM compression tests were run on various types of
metal alloys, including aluminum alloys (Al6061-T6511, Al7075-T6, Al2024-T4,
Al2024-T351), copper-based alloys (C11000, 360 brass, 464 brass), ferrous-based
alloys (304 Stainless Steel, A2 steel), and titanium alloy [8]. In addition, this work
explored two different specimen sizes for several of the metals mentioned above.
The specimens were compressed 6.4 mm (a 60 % height reduction) with a platen
speed of 25.4 mm/min, while various starting current densities were applied to
the specimens constantly throughout each test. Figure 11.35 shows stress–strain
curves for the titanium alloy at different starting current densities. From this work,
an “electrical threshold” effect was discovered. Specifically, there is a particular
current density in which a significant reduction in forming load is witnessed, or a
significant improvement in formability is observed.
Conclusions from this work are as follows:
• There is an “electrical threshold” where electroplastic effects are produced, which
is specific to each material.
• The electrical threshold is a function of the starting current density, die speed,
and material.

Fig. 11.35  Electrical threshold effect (Ti-G5) [8]. Once a particular current density level is
reached, which is specific to each metal alloy, a significant force/stress reduction due to EAM
will take place. This current density level is denoted as the “electrical threshold” level at the par-
ticular electrical settings
11.4  Experimental Findings for Alternative EAF Processes 293

• At the same die speed, some metals in this work expressed a low-threshold
current density, while others showed a higher-threshold current density (up to
4-times higher than others).
• The EAM technique produced similar results on two different size specimens,
leading to the conclusion that the effects are related to current density and not
current amplitude.

11.4.1.2 Effects of Different Die Speeds on EAM Effectiveness

Work by Jones and Roth [17] determined that this electroplastic effect is strain-
rate-dependent. More specifically, in this work, Mg AZ31B-O specimens were
compressed, while a constant current was applied. As the die speed was increased,
while the current remained constant, the magnitude of the electroplastic effect
decreased, as shown in Fig. 11.36.
The final geometries for the same test series are depicted in Fig. 11.37. The
25.4, 19.05, and 12.70 mm/min tests fractured by shear, whereas the 6.35 mm/min
test only had cracking and was able to be formed to a much greater extent. As seen
for the specimens which failed by shearing, the fracture once again occurred on a
45° plane. As mentioned in the previous subsection, the “threshold” effect can also
be categorized as a significant improvement in formability. From Fig. 11.37, one
can see that the “electrical threshold” for this particular current density is between
die speeds of 6.35 and 12.70 mm/min.
Conclusions from this work are as follows:
• The higher current densities produced greater formability improvements, how-
ever, the same benefits could be produced from lower current densities if the
platen speed was slowed down.
• A threshold current density was noticed where the formability was considerably
improved compared to the other lower current densities.
• The EAM technique in compression was found to be strain-rate dependent.

Fig. 11.36  EAM die speed

dependence (CD = 20 A/
mm2) [17]. For the same
electrical settings, as the
die speed is increased, the
amount by which EAM
has an effect on the metal’s
formability becomes reduced
294 11  Applications of Electrically Assisted Manufacturing

Fig. 11.37  Upsetting specimens compressed at 20 A/mm2 with varying platen speeds [17]

11.4.2 Tension

There has been a lot of experimental work on EAM in tension over the past decade.
The first work will discuss the first attempts of applying a constant electrical cur-
rent to a uniaxial tensile process, in which it was determined that electrical appli-
cation parameters played a role in formability improvement [18]. The second work
describes a technique of applying electricity to tension in a “pulsing” form, rather
than being applied continuously, which generated better results [19]. Further, an
experimental work is explained, where different aluminum alloys with different start-
ing microstructures were tested with various electrical application parameters [20].
Finally, the last work explains a relationship between some of the electrical appli-
cation parameters and balances between each of these parameters that will result in
optimum formability for a uniaxial tension process on a magnesium alloy [21].

11.4.2.1 EAM Application Methods for Different Deformation

Processes

In 2009, Ross et al. [18] studied the effect of a continuous current on the same metals
as Perkins, now for uniaxial tension tests. It could be concluded that the yield strength,
flow stress, and elastic modulus were all reduced. Also, because of strain weakening,
the overall energy of deformation was decreased. However, one important finding of
this research was that, due to the continuously applied current, the achievable elon-
gations of the specimens decreased (an opposite effect of using continuously applied
current in compression), as shown in Fig. 11.38. This led to experimentation with
pulsed electricity, rather than leaving it applied continuously, for tensile applications.
11.4  Experimental Findings for Alternative EAF Processes 295

Fig. 11.38  Al2024-T4
EAM stress–strain tensile
profiles [18]. When the
electricity is applied to a
tensile deformation process
continuously (as is done
with compression tests), the
larger amounts of electrical
power will produce decreased
elongations. Hence, this is
why electrical pulsing is
performed on EAM tensile
processes

The conclusions from this work are:

• If electricity is applied continuously during a tensile deformation process, the
forces will be reduced; however, the achievable elongation will also be reduced.
• The difference in the specimen conditions throughout tensile and compressive
tests explains the reasoning for the continuously applied electricity causing a
detrimental effect on the tensile process. Specifically, the cross-sectional area
of the workpiece will increase in compression (the current density will decrease
over the test), and the cross-sectional area of the workpiece in tension will
decrease (which will increase the current density of the specimen over the test).

11.4.2.2 Benefits of Pulsed Electricity in Tensile-Based Processes

The previous experimental EAM work indicated that, when continuous DC current
is applied to a part during tensile or compressive deformation, the required flow
stress is reduced [8–18]. However, the effect of the continuous current on the max-
imum achievable elongation of the part was significantly different for compres-
sive and tensile loading. When performing compression-based deformation, it was
found that the application of a continuous current dramatically increased the mate-
rial’s deformability. This was not true when performing tensile-based deformation,
however. When continuous electricity was applied during tensile-based deforma-
tion, it was noted that the deformability of the material decreased (Fig. 11.38).
In subsequent studies, however, it was discovered that, when a pulsed current
(rather than continuous) was applied during tensile deformation, the maximum
achievable elongation was increased (rather than decreased as was true for contin-
uous). In a recent study conducted by Roth et al. [19], the effects of a pulsed DC
current on 5754 aluminum alloy, a widely used alloy for body panels in the auto-
motive industry, were investigated. This research involved developing an optimal
pulsing parameter in order to minimize the flow stress and maximize the material’s
overall achievable elongation, thereby placing the material into its most workable
296 11  Applications of Electrically Assisted Manufacturing

Fig. 11.39  Al5754 with decreasing pulse periods lead to approximately 400 % elongation

improvement in uniaxial tension [19]

state. The current density at which the specimens were pulsed was held constant at
90 A/mm2, while the pulse duration and period between pulses were varied. The
electric pulsing lead to extreme increases in elongation over non-pulsed baseline
tests as shown in Fig. 11.39, where the pulsing period was steadily decreased.
About a 400 % increase in elongation and a notable decrease in flow stress due to
electrical pulsing can be seen from the figure.

11.4.2.3 Effects of Heat Treatments on EAM in Tension

In 2008, the effect of different heat treatments on EAM for two different Al alloys
were tested [20, 21]. Two aluminum alloys (Al5052 and Al5083) were electri-
cally pulsed using two different electrical parameter sets (i.e., combinations of cur-
rent density, pulse duration, and pulse period settings). Additionally, specimens
from each alloy were subjected to three different heat treatments (As Is, 398 °C,
and 510 °C). The analysis focuses on establishing the effect the electrical pulsing
had on the aluminum alloy’s various heat treatments by examining the displace-
ment of the material throughout the testing region of the dog bone-shaped speci-
mens. Figure 11.40 shows the experimental setup. The results from this research
show that pulsing significantly increased the maximum achievable elongation of
the aluminum (when compared to the baseline tests conducted without electrical
pulsing), as well as reduced the engineering flow stress. The electrical pulses also
caused the aluminum to deform non-uniformly, such that the material exhibited a
diffuse neck where the minimum deformation occurs near the ends of the speci-
men (near the clamps) and the maximum deformation occurs near the center of the
specimen (where fracture ultimately occurs). This diffuse necking effect is similar
to what can be experienced during superplastic deformation, but on a smaller scale.
11.4  Experimental Findings for Alternative EAF Processes 297

Fig. 11.40  Experimental EAM setup [20, 21]. This EAM testing setup includes conductive
metal dies with rigid insulation materials placed in between the dies and the testing machine, to
ensure that the electricity only flows through the workpiece

Figures 11.41 and 11.42 display the general force reductions, elongation improve-
ments, and diffuse necking effects compared to the non-pulsed baseline tests.
The conclusions from this work are as follows:
• The effectiveness of the pulsed electricity is dependent on both the alloy and its
heat treatment.
• The size of the diffuse neck resulting from electrical pulsing was found to be
irrespective of the material pulsed, but dependent on the parameter set used for
pulsing.
• In any aluminum alloy/heat treatment combination, the greater pulsing fre-
quency of Parameter Set 2 developed a larger diffuse neck compared to the
greater magnitude of pulses in Parameter Set 1.
• The formability of both aluminum alloys was improved by EAM, but Al5083
had the greatest elongation increases and flow stress reductions due to EAM.

11.4.2.4 Effects of Electrical Application Parameters on Formability

Improvements

This research work explored the effects of particular EAM pulsing parameters
(current density and pulse duration) on the formability of Mg AZ31B-O speci-
mens undergoing uniaxial tensile deformation [22]. As a part of investigating
these effects, various current density and pulse duration combinations (i.e., high
current/short duration or low current/long duration) are examined. An important
outcome of this work was the development of a linear inverse relationship between
298 11  Applications of Electrically Assisted Manufacturing

Fig. 11.41  EAM force
reduction and elongation
increase [20, 21]. The
application of electrical
pulses (i.e., the steep drop-
offs in the EAM profile)
proved to reduce the overall
engineering stress–elongation
profile and increase
elongation compared to the
non-pulsed baseline test

Fig. 11.42  Diffuse necking
effect (baseline vs. EAM)
[20, 21]. The two EAM tests
produced much more defined
diffuse necks than the non-
EAM baseline test

current density and pulse duration, which allows accurate elongation predictions
to be made from respective parameter combinations (Fig. 11.43). This relation-
ship could also be represented in 3-D as a “formability ridge,” specifying expected
elongation for different current density and pulse duration combinations, as shown
in Fig. 11.44.
Conclusions from this section are as follows:
• EAM produced the most optimum formability improvements with high cur-
rent density/short pulse duration and low current density/long pulse duration
combinations.
• In cases where this inverse relationship was violated, such as low current/short
duration, the formability remained unchanged.
11.4  Experimental Findings for Alternative EAF Processes 299

Fig. 11.43  Inverse linear
relationship between current
density and pulse duration
[22]. This chart depicts a
line of optimum elongation.
Specifically, to maintain
optimum elongations, longer
pulse durations must be
coupled with lower current
densities and shorter pulse
durations must be paired with
higher current densities

Fig. 11.44  3-D EAM formability ridge for Mg AZ31B-O. This is a 3-D representation that
shows that pulse duration and current density magnitudes must be “balanced” to reach maximum
allowable elongation levels. Not meeting this requirement will result in reduced elongation lev-
els, away from the peaks of the ridge

• In parameters with high current/long durations, the specimens failed prema-

turely due to the overly powerful combination of electricity, leading to extensive
and undesirable heating effects.
• The electrical power input into an EAM process needs to be considered because an
under powerful or overly powerful combinations of input parameters (starting cur-
rent density, pulse duration, and pulse period) can decease the formability of the part.
300 11  Applications of Electrically Assisted Manufacturing

• The efficiency of the applied electrical parameters must be considered because

the same amounts of formability improvements may be possible from less pow-
erful electrical input parameter combinations.

11.4.3 Non-uniform Deformation (E.G. Channel Formation)

The majority of EAM research has been conducted on processes where uniform and
uniaxial deformations were taking place. However, this work researched the effect
of the EAM technique on a non-uniform channel formation process using Al5052-O
sheet strip specimens [23]. The experimental setup can be seen in Fig. 11.45, includ-
ing the custom insulated EAM channel die supplied by the Ford Motor Company.
The EAM variables which were examined in this work are current density,
pulse duration, pulse period, and die speed. Specifically, the Al5052 sheet strip
specimens were formed into channels using the EAM technique and experiments
were run with different combinations of the variables mentioned above.
Engineering stress versus elongation plots examining the effects of either pulse
period or pulse duration can be seen in Fig. 11.46 (parts a and b). In part a for this
particular current density magnitude, it can be seen that the length of the pulse
period had little effect on the increased elongation. However, in part b, a longer
pulse duration (2 s) was proved to be overly powerful and led to premature failure
compared to non-EAM baseline channel forming.
Conclusions from this work are as follows:
• EAM parameter combinations above the threshold can improve the part elonga-
tion and reduce the forming forces.
• Additionally, the magnitude of each of the electrical variables must be propor-
tionally “balanced” for best results.

Ceramic Top Plunger

Specimen

Clamp
Bottom Die
Bottom Insulation

Fig. 11.45  EAM channel formation experimental test setup [23]. This setup included a 2-part
channel die, composed of a top ceramic plunger and a metal bottom portion with insulation inte-
grated into the base structure
11.4  Experimental Findings for Alternative EAF Processes 301

(a) 100A/mm 2 – Effect of Pulse Period (b) 120A/mm 2 – Effect of Pulse Duration

Fig. 11.46  EAM channel formation formability results [23]. EAM was not as effective with
channel forming as uniaxial tension, because channel forming caused localized deformation
regions, whereas uniaxial tension consisted of uniform deformation. a Engineering stress–elon-
gation profiles for CD = 100 A/mm2 with varying pulse periods. b Engineering stress–elongation
profiles for CD = 120 A/mm2 with varying pulse durations

• When dealing with localized strain regions, such as with channel forming, it is
no longer optimum to apply the electricity throughout the entire part, but rather
to devise a method to divert the electricity to primarily the regions of localized
strain, which need it most.
• When examining the effects of die speed, it can be concluded that there is a
relationship between the strain-rate sensitivity of the Al5052-O alloy and the
electricity’s sensitivity.
• As the current density is increased, the longer durations and more frequent
pulse periods can be detrimental to the overall formability of the specimen, even
if they do reduce required deformation forces in the process.

11.4.4 Springback Reduction Using EAF

Springback can be a big concern when metal forming. Currently, springback is

reduced by costly trial-and-error processes or by slow and intensive modeling
strategies. In 2009, Green et al. examined the effectiveness of applied electricity
when reducing springback in formed sheet metal specimens. In this work, Al6111
sheet strips were used. Specifically, shape retention testing was performed, in
which the sheet strip specimens were bent around a round die, clamped in place,
and then a single pulse of electricity was applied. Immediately after the pulse of
electricity, the clamps were removed and the springback was measured.
Current density and duration of the single pulse were varied as part of this
research. Figure 11.47 shows the amount of springback reduction that was achieved
from this work 0. From the figure, for a single 1.5-s-long electrical pulse, the
302 11  Applications of Electrically Assisted Manufacturing

Fig. 11.47  Post-deformation
springback reduction using
EAM [9]. Springback in
Al6111 sheet metal strip
specimens was completely
eliminated with a single 1.5-s
electrical pulse of 120 A/mm2
or greater post-deformation,
but prior to removing the
material from the die

Fig. 11.48  Percent reduction
in springback due to shape
retention tests [9]

effects of springback can be completely eliminated with a current density of 120 A/

mm2. The percent reduction in springback from the same specimens is quantified
in Fig. 11.48 [9]. From this work, it can be concluded that applying electricity to
a part after deformation could be a viable method for reducing springback effects.

11.4.5 Electrically Assisted Micro-Forming

Siopis et al. [24–26] examined how different microstructure properties affect the
effectiveness of EAM in micro-extrusion experiments. Specifically, it was con-
cluded that a finer-grained material, with more grain boundary area, enhanced the
electroplastic effect, whereas a larger-grained material, with less grain boundary
area, lessened the effect [24]. Another work by Siopis and Kinsey [25] determined
that the effectiveness of EAF increased as the dislocation density within the metal
11.4  Experimental Findings for Alternative EAF Processes 303

also increased, as a result of cold-working prior to EAF experiments. Further,

research by Dzialo et al. [26] showed that, above the critical threshold current
density, the efficiency of the applied electricity increased with the percentage of
alloyed Zn in copper specimens during micro-forming.

11.5 Overhead Transmission Line Design Using EAF

Overhead transmission lines (OHTL’s) deliver electricity from the generation point
(i.e., the power plant) to the customer (i.e., homes or businesses). In the EAF tech-
nique, electricity is applied to a metal while it is mechanically deformed. Overhead
electric transmission lines are simply coated metal cables in static tension loading
with high-power electricity flowing through them. From the statements above, there
are some similarities between the EAF technique on a forging process and what goes
on within a high-power electric transmission line. This section will include an expla-
nation of the electrical system grid, different transmission line structures and set-ups,
commercially available conductors and sizing procedures, conductor sag and what it
means, the effect of high temperatures on transmission line longevity, and finally how
the EAF technique explained in this thesis could be applied to the analysis of OHTL’s.

11.5.1 The Electricity Transmission Grid

Electrical transmission lines are placed throughout the USA, thus forming a trans-
mission “grid” that can transport the electricity to where it is needed. There are
two types of transmission lines (high-voltage and low-voltage). The high-voltage
lines are used to transport the electricity over very large distances quickly with
little loss. The low-voltage lines are used to deliver the electricity closer to the cus-
tomers. High-voltage lines can be rated from 100 up to 1,000 kV, whereas low-
voltage lines can be rated from 4 to 46 kV [27]. Note that the voltage coming
directly into a residence is typically 240 V; the above values are for distribution.
In general, the size of the conductors is determined from estimating the poten-
tial power use of the customers at the end of the lines. Outages can occur if the
voltage in the lines is more than that for which the line is rated. If one link in the
grid is down, the electricity can reroute itself; thus, adding more voltage to the
amount which the other line was already carrying. Ultimately, the overloading of
several lines within a grid could lead to a blackout.

11.5.2 Transmission Line Structures and Setups

There are several different types of transmission line structures. For each of these
structures, there are clearance limitations and structural compliances. Most recently,
the H-frame design structure has been replaced with the various single-frame
304 11  Applications of Electrically Assisted Manufacturing

structures because they require less right-of-way restrictions. When setting up a new
transmission line system, there are several main procedures that need to be followed
(e.g., installing the foundations, erecting the towers, pulling the conductors), which
is further explained in [27]. The installation of the conductors is an important step in
the transmission structure finalization. After a regulated amount of time has passed
since setting the footings for the transmission towers, one can then begin to pull the
conductor lines. In most cases, the wires are mounted to vehicles and pulled through
the stringing hardware to position the lines and set the sag in the lines.

11.5.3 Commercial Conductors and Sizing

There are several types of conductors that can be used as transmission lines; how-
ever, the most-common is the steel-reinforced aluminum conductor (ACSR). This
conductor has a relative low-cost and high weight-to-strength ratio compared to its
competitor conductors. Several other conductor types are an all aluminum conduc-
tor (AAC) and an all aluminum alloy conductor (AAAC) [28]. The different con-
ductor types are named after different objects. Specifically, the AAAC conductors
are named after types of trees, while the ACSR conductors are named after differ-
ent animals (the ACSR conductor used in the example in the following subsec-
tions is named “DOG”). According to Castro [29], the ACSR-type conductor can
be sized two different ways. First, the most economical method is to use Kelvins
Law, which states, “The most economical area of a conductor is that for which the
annual cost of energy losses is equal to the interest on that portion of the capital
outlay which may be considered as proportional to the weight of the conductor.”
Second is the more efficient sizing method, in which the conductor is sized based
on balancing the I2R losses and the installation cost. Figure 11.49 shows how the
most efficient conductor is sized. To put the cost of an OHTL into perspective,

Fig. 11.49  Determination of
the most efficient conductor
size [29]. The most efficient
method of sizing a conductor
is to balance the power losses
and the installation costs,
since these are inversely
proportional
11.5  Overhead Transmission Line Design Using EAF 305

a 200-mile single-circuit 500-kV, 3-conductor line is approximately $700,000 per

mile [27]. The main factors impacting the price of a transmission line are the ter-
rain, the location, the overall distance, the configuration of the conductors, and
any relevant environmental regulations. Wires which shield the conductor bundles
from lightning and grounding are normally installed above the bundles.

11.5.4 Conductor Sag

Because of all the different weather and loading conditions, one can never eliminate
sag in transmission lines and, thus, it must be designed in. The sag is a measurement
of the vertical difference between the line connection point on the adjacent tower
and the lowest vertical point within a particular span of the conductor. There are reg-
ulations as to how low a transmission line can be to the ground. Equation (11.26)
shows how to calculate the sag of a parabolic-shaped line [30].

wL 2
Scond = (11.26)
8Tcond

where Scond is the sag in meters, w is the conductor weight in N/m, L is the hori-
zontal span length in meters, and Tcond is the conductor tension in N. In order to
use the above equation, the conductor weight must be provided in N/m. Equation
(11.27) shows how to convert the weight from kg/km to N/m.
wc · 9.81 N
w= [=] (11.27)
1, 000 m
where wc is the weight of the conductor in kg/km.
The sag in a transmission line is directly related to the tension in the line, and
this can be affected by the following variables [30]:
• An increase in temperature of the line can lead to thermal expansion of the con-
ductor, thus resulting in an increase in its length, as shown in Eq. (11.28).

�L = αexp · T · S (11.28)

where α is the coefficient of thermal expansion, T is the temperature increase in

°C, and S is the span length in meters.
• An increase in the wind speed can produce an extra force on the conductor, thus
increasing the tension in the conductor. This can elongate the conductor by way
of elastic stretching, as shown below in Eq. (11.29).
(T − To )
�L = (11.29)
E·A
where To is the initial tension in N, T is the final tension in N, E is the coeffi-
cient of elasticity, and A is the cross section of the conductor in meters.
306 11  Applications of Electrically Assisted Manufacturing

• If ice forms on the conductor, it will increase the overall diameter of the con-
ductor and will also add weight to the line. This will increase the tension and
can lead to elastic stretching as was calculated in the equation above.
• As the transmission lines age, they go through countless cycles of loading and
unloading and heating and cooling, which can weaken the overall strength of
the line.

11.5.5 Effect of Temperature on Transmission Line Longevity

The heating of the transmission lines is currently the most critical attribute respon-
sible for the line sag and the longevity of the lines. The heating effect is due to the
large power running through the lines. Initially, the lines are sized for an estimated
amount of electricity, but as more is run through, their temperature increases sig-
nificantly. Standard operating temperatures of OHTL’s may be around 30–70 °C;
however, lines have been known to reach up to 175 °C or more when overloaded
[31]. In the case of aluminum conductors, the degradation in the aluminum
strength due to heat is cumulative and irreversible.
In [31], the effect of high-temperature cyclic loadings on an ACSR conductor with
compression dead-ends and full tension compression splices was examined though
evaluation of the tensile strength, hardness, and metallurgical changes. Cyclic tests at
250, 500, 750, and 1,000 thermal cycles were run at temperatures of 100 and 175 °C.
The average hardness values of the conductors decreased at both temperatures;
however, they significantly decreased for the 175 °C tests, so much so that they
could not even be read using the Rockwell H scale for the 500- and 100-cycle
tests.
The same conductors were tested in tension in compliance with ANSI C119.4
class 1 full tension. In this case, the conductor is required to withstand 95 % of its
rated breaking strength (RBS), which was 4,066 lb in this case. The tension force
decreased as the number of cycles increased for each respective temperature. For
the 100 °C tests, the 1,000-cycle test produced a tensile strength below the required
95 % RSB. For the 175 °C tests, the 250-, 500-, and 1,000-cycle tests all produced
tensile strengths below the 95 % RSB. Hence, this indicates that the cycling of
high temperatures has a notable effect on the tensile strength of the conductor.
The metallurgical evaluation in this work involved lightly magnifying the con-
ductors to try to evaluate them. No differences could be seen at this level of mag-
nification; however, grain size should be evaluated. It is expected that there will
be differences in the pre-test and post-test samples, since there was a significant
effect on the hardness and tensile strength.
This specific work supports the statement that the temperature of the conductor
can dictate its integrity. Some ideas to improve the conductor life would be to mini-
mize the temperature and duration for which the conductor is at a high temperature.
To this end, new technology has produced wireless sensors that can be attached to
the conductor and can be used to monitor the temperature in real time [32].
11.5  Overhead Transmission Line Design Using EAF 307

11.5.6 Applying EAF Modeling to OHTL Sag Calculations

Of note, from all the research that has been collected on transmission line design
and the sag in transmission lines, the main reason for the increase in the sag is due
to the temperature increase by way of thermal expansion when excess electricity is
supplied through a transmission line [33]. However, what if there were actual effects
due to the electroplastic effect as well? For example, in the EAF modeling strategy
of this thesis, the applied electricity is separated into the portion that assists deforma-
tion and the portion that assists resistive heating. The resistive heating portion of the
applied electricity is all that is accounted for in the thermal expansion calculation.
For example purposes, a “DOG” ACSR conductor will be examined with the
following specifications [34]:
• Overall line diameter of 14.15 mm
• Line weight of 3.89 N/m
• Calculated breaking load (CBL) = 32.68 kN
• Initially set tension in the cable = 20 % CBL = 6.536 kN
• Tower span = 275 m
Using Eq. (11.26) above, the conductor sag can be calculated, as shown Eq. (11.30):
N
3.89 m · (275 m)2
Sag = = 5.6 m (11.30)
8 · (6,536 N)

The typical amount of sag in an OHTL can be from about 3 to 10 m, depending on
the electrical conditions. Aside from the sag on the line, it is also important to ana-
lyze the tension, and corresponding stress in the line. The stress can be calculated
from Eq. (11.31):
 
Tension [N]
Cable Stress = =  2  = MPa (11.31)
Area m

The stress on this particular OHTL is shown in Eq. (11.32):

6,536 N
Cable StressDog_Conductor = π 2
= 41.56 MPa (11.32)
4 (0.01415 m)

Although the stress in the ACSR is below the yield strength limit for high-strength
steel [35, 36], it may already be higher than the yield strength limit for 1xxx elec-
trical aluminums depending on the temperature of the conductor [37]. In addition,
the lines could potentially be rated for carrying several hundred or even thousands
of megawatts (MW) of power through them. Experimental works have shown that
EAF has the ability to significantly reduce the stress needed to plastic deforma-
tion in both aluminum and steel metals [8, 19, 20, 23, 38]. With this being said,
the direct electrical effects may lower the plastic deformation strength of the lines
even further than what the heating can attribute toward.
308 11  Applications of Electrically Assisted Manufacturing

The heat balance for overhead conductors can be shown in Eq. (11.33).

dTc
qc + qr + mCp = qs + I 2 R(Tc ) (11.33)
dt
where qc is the convection heat loss, qr is the radiation heat loss, qs is the solar
heat gain, mCp is the total conductor heat capacity, and R(Tc) is the resistance of
the electrical conductor as a function of temperature [39]. The term in this equa-
tion represents the watt losses in the conductor, where all of the power is assumed
to contribute toward resistive heating. However, if the stress in the OHTL is close
to the yield strength of the conductor and electricity is now flowing, there may be
a possibility that some plastic deformation could take place.
Overall, applied electrical power has been proven to have the capability to
greatly reduce a metal’s yield strength above a material-specific, electrical power
threshold. Therefore, a portion of the conductor sag, which was originally consid-
ered to be due to heating effects, may actually be due to direct electrical effects
and could be represented by an EEC, specific to the conductor material properties,
the loading conditions, and the electrical power flowing through the line.
To bring the electroplastic effect on OHTL’s into perspective, data from [8] will
be used to estimate currents which may be above the electrical threshold of materi-
als similar to the materials in the ACSR-DOG conductor. From this reference, the
electrical thresholds for Al6061 and A2 tool steel were found to be 60 and 45 A/
mm2, respectively. (Note that these threshold current densities were for compres-
sion and were for specific die speeds. These current densities are just intended to
provide an idea of how much current could be needed to potentially have an effect
on the conductor strength.) For the given diameter of the conductor (14.15 mm),
Eqs. (11.34) and (11.35) calculate the amount of electrical current where effects
on conductor strength would be notable, based on the electrical thresholds from
[8].
A π 2

AmpsAl = Cd · Area = 60 · (14.15 mm) ≈ 9,400 A (11.34)
mm2 4

A π 2

AmpsSteel = Cd · Area = 45 · (14.15 mm) ≈ 7,100 A (11.35)
mm2 4
The large OHTL’s coming from the electrical generation stations may be able to
carry this amount of electrical current. However, the EAF experiments explained
throughout this thesis were conducted where the electrical power was supplied as
constant current, with varying voltage. The power in OHTL’s is supplied in the
opposite manner, where the electricity in the lines is described in terms of voltage,
since the voltage running through the lines is constant, and the current fluctuates
depending on the power usage of the customers. In order to determine the magni-
tude of the reduction in the strength of the conductor due to direct electrical effects
(i.e., the EEC), specialized experiments would need to be run, which are explained
in the following subsection.
11.5  Overhead Transmission Line Design Using EAF 309

11.5.7 Future Work to Determine EEC Values for OHTL’s

Although a full relationship between transmission line sag calculations and

invented EAF modeling strategy will not be covered in this section, the following
is an envisioned procedure to follow to begin to do this. It is known that all the sag
calculations consider the heating effect due to electricity the sole contributor to the
sagging. However, the invented EAF modeling strategy indicates that there could
also be a direct electrical effect on the deformation of the transmission line.
In order to begin to determine whether there is a direct electrical effect and
attempt to quantify, it is suggested to run an EAF tension test at an electrical
power combination comparable to what is seen daily in the transmission lines.
Then, compare the power profile of this test to a conventional tension test with-
out EAF. By calculating the difference in the mechanical power profiles, one can
determine the EEC using the mechanical approach (as described in Chap. 4). This
coefficient will describe the percentage of applied electricity that actually con-
tributed to assisting the deformation and did not just account for heating. If the
EEC is very low, then the assumption that only heat contributes to sag is sufficient.
However, if the EEC is quite large, then there are unaccounted affects which could
amplify the magnitude of the transmission line sag beyond that of which can be
calculated using current transmission line sag calculations.

References

1. Salandro WA, Bunget C, Mears L (2011) Electroplastic modeling of bending stainless steel
sheet metal using energy methods. J Manuf Sci Eng 133(1):10
2. Salandro WA, Bunget C, Mears L (2010) Modeling and quantification of the electroplastic
effect when bending stainless steel sheet metal. In: International manufacturing science and
engineering conference (MSEC 2010-34043), p 10
3. Hill R (1950) The mathematical theory of plasticity. Oxford University Press, Oxford
4. Hosford WF, Caddell RM (1993) Metal forming: mechanics and metallurgy, 2nd edn.
Prentice Hall, Englewood Cliffs
5. Weinmann KJ, Shippell RJ (1978) Effect of tool and workpiece geometries upon bending
forces and springback in 90° V-die bending of HSLA steel plate. In: Proceedings of the 6th
North American metal working research conference, pp 220–227
6. Altan T, Ngaile G, Sheng G (2004) Cold and hot forging: fundamentals and applications.
ASM International, Materials Park
7. Allegheny Ludlum Corp., Product data sheet (304SS)
8. Perkins TA, Kronenberger TJ, Roth JT (2007) Metallic forging using electrical flow as an
alternative to warm/hot working. J Manuf Sci Eng 129(1):84–94
9. Green CR, McNeal TA, Roth JT (2009) Springback elimination for Al-6111 alloys
using electrically-assisted manufacturing (EAM). Trans North Am Manuf Res Inst SME
37:403–410
10. Jones E, Jones JJ, Mears L (2013) Empirical modeling of direct electric current effect on
machining cutting force. In: 2013 International manufacturing science and engineering con-
ference (MSEC 2013-1229)
11. Jones JJ, Mears L (2010) Empirical modeling of the stress-strain relationship for an upsetting
process under direct electrical current. Trans North Am Manuf Res Inst SME 38:451–458
310 11  Applications of Electrically Assisted Manufacturing

12. Zhang W (2010) Intelligent energy field manufacturing: interdisciplinary process innovations
13. Long X, Khanna S (2005) Modeling of electrically enhanced friction stir welding process
using finite element method. Sci Technol Weld Joining 10(4):482–487
14. Ferrando WA (2008) The concept of electrically assisted friction stir welding (EAFSW) and
application to the processing of various metals. No. NSWCCD-61-TR-2008/1
15. Pitschman M, Dolecki J, Johns GW, Zhou J, Roth JT (2010) Application of electric current
in friction stir welding. In: 2010 International manufacturing science and engineering confer-
ence (MSEC 2010-34166)
16. Potluir H, Jones JJ, Mears L (2013) Comparison of electrically-assisted and conventional
friction stir welding processes by feed force and torque. In: 2013 International manufacturing
science and engineering conference (MSEC 2013-1192)
17. Jones JJ, Roth JT (2009) Effect on the forgeability of magnesium AZ31B-O when a continu-
ous DC electrical current is applied. In: ASME international manufacturing science and engi-
neering conference 2009, West Lafayette, IN, p 10 (2009)
18. Ross CD, Kronenberger TJ, Roth JT (2009) Effect of dc on the formability of Ti–6Al–4V. J
Eng Mater Technol 131:11
19. Roth JT, Loker I, Mauck D, Warner M, Golovashchenko SF, Krause A (2008) Enhanced
formability of 5754 aluminum sheet metal using electric pulsing. Trans North Am Manuf Res
Inst SME 36:405–412
20. Salandro WA, Jones JJ, McNeal TA, Roth JT, Hong ST, Smith MT (2008) Effect of electrical
pulsing on various heat treatments of 5xxx series aluminum alloys. In: ASME international
manufacturing science and engineering conference, Evanston, IL, 2008, p 10
21. Salandro WA, Jones JJ, McNeal TA, Roth JT, Hong ST, Smith MT (2010) Formability of Al
5xxx sheet metals using pulsed current for various heat treatments. J Manuf Sci Eng 132:11
22. Salandro WA, Khalifa A, Roth JT (2009) Tensile formability enhancement of magnesium
AZ31B-O alloy using electrical pulsing. Trans North Am Manuf Res Inst SME 37:387–394
23. Salandro WA, Roth JT (2009) Formation of 5052 aluminum channels using electrically-
assisted manufacturing (EAM). In: International manufacturing science and engineering con-
ference (MSEC 2009-84117), p 9
24. Siopis MS, Kinsey BL, Kota N, Ozdoganlar OB (2010) Effect of severe prior deformation on
electrical-assisted compression of copper specimens. In: International manufacturing science
and engineering conference (MSEC2010-34276), p 7
25. Siopis MS, Kinsey BL (2009) Experimental investigation of grain and specimen size effects
during electrical-assisted forming. In: International manufacturing science and engineering
conference (MSEC2009-84137), p 6
26. Dzialo CM, Siopis MS, Kinsey BL, Weinmann KJ (2010) Effect of current density and
zinc content during electrical-assisted forming of copper alloys. CIRP Ann Manuf Technol
59(1):299–302
27. Castro RD (1995) Overview of the transmission line construction process. Electr Power Syst
Res 35:119–125
28. Oncor Electric Delivery Company (2011) Index section 9 conductors, p 47
29. Castro RD (1995) Overview of the transmission line design process. Electr Power Syst Res
35:109–118
30. Ergon Energy Corp. (2012) Network lines standard guidelines for overhead line design, p 20
31. Di Troia G (2000) Effects of high temperature operation on overhead transmission full-ten-
sion joints and conductors, pp 1–7
32. Bernauer C, et al (2007) Temperature measurement on overhead transmission lines (OHTL)
using surface acoustic wave (SAW) sensors. In: 2011 International conference on electricity
distribution (CIRED2007_0788), p 4
33. Muhr M, Pack S, Schwarz R, Jaufer S (2006) Calculation of overhead line sags. In: 51st
Internationales Wissenschaftliches Kolloquium, p 10
34. Clydesdale Ltd. (1970) ABC, Copper, AAAC and ACSR conductor specifications. Conductor
specifications tables: technical specification for aluminum conductor steel reinforced (ACSR)
to BS 215 part 2: 1970. www.clydesdale.net
References 311

35. Rothman MF (ed) (1988) High-temperature property data: ferrous alloys. ASM International,
Metals Park, Ohio
36. Harvery PD (ed) (1982) Engineering properties of steel. ASM, Metals Park, Ohio
37. Aluminum alloys for cryogenic applications. The Aluminum Association, Washington, DC
(1999)
38. Heigel JC, Andrawes JS, Roth JT, Hoque ME, Ford RM (2005) Viability of electrically treat-
ing 6061-T6511 aluminum for use in manufacturing processes. Trans North Am Manuf Res
Inst SME 33:145–152
39. Zocholl SE, Guzman A (1999) Thermal models in power system production. Report from
Schweitzer Engineering Laboratories Inc., p 16
Appendix A

A.1 Experimental Findings for EAF Compression

There are two key EAF compression experimental works. Perkins et al. explored
the effects of electricity in compression on many different types of metal. In addi-
tion, Jones et al. compared the effects of different die speeds on the effectiveness
of compressing Mg AZ31B-O slugs. In this sub-section, these two works will be
highlighted, and an explanation of important EAF variables, derived from experi-
mentation, will be explained.

A.1.1 EAM Compression Results on Various Metals (Perkins et al.)

In the work by Perkins et al., EAM compression tests were run on various types of
metal alloys, including Aluminum Alloys (Al6061-T6511, Al7075-T6, Al2024-T4,
-T351), Copper-based Alloys (C11000, 360 Brass, 464 Brass), Ferrous-based
Alloys (304 Stainless Steel, A2 Steel), and Titanium Alloy [A.1]. In addition, the
work explored two different specimen sizes for several of the metals mentioned
above. The specimens were compressed 6.4 mm (a 60 % height reduction) with a
platen speed of 25.4 mm/min, while various starting current densities were applied
to the specimens constantly throughout each test. Figure A.1 shows stress-strain
curves for the Titanium Alloy at different starting current densities.
Conclusions from this work are as follows:
• There is an “electrical threshold” where electroplastic effects are produced,
which is specific to each material.
• The electrical threshold is a function of the starting current density, die speed,
and material.
• At the same die speed, some metals in this work expressed a low threshold
current density, while others showed a higher threshold current density (up to
4-times higher than others).

© Springer International Publishing Switzerland 2015 313

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2
314 Appendix A

Fig. A.1  Threshold effect

(Ti-G5) [A.1]. Once a
particular current density
level is reached, which is
specific to each metal alloy,
a significant force/stress
reduction due to EAM will
take place. This current
density level is denoted as the
“electrical threshold” level
at the particular electrical
settings

• The EAM technique produced similar results on two different size specimens,
leading to the conclusion that the effects are related to current density and not
current amplitude.

A.1.2 Effects of Different Die Speeds on EAM Compression

Effectiveness (Jones et al.)

Work by Jones et al. [A.2] determined that the electroplastic effect is strain rate
dependent. More specifically, in this work, Mg AZ31B-O specimens were com-
pressed while a constant current was applied. As the die speed was increased,
while the current remained constant, the magnitude of the electroplastic effect
decreased, as shown in Fig. A.2.
Conclusions from this work are as follows:
• The higher current densities produced greater formability improvements, how-
ever, the same benefits could be produced from lower current densities if the
platen speed was slowed down.
• A threshold current density was noticed where the formability was considerably
improved compared to the other lower current densities.
• The EAM technique in compression was found to be strain-rate dependent.

A.2 EEC Frequency Analysis

During analysis of conventional/EAF tests on 304SS material, it was noticed that

the EAF tests consistently produced load versus time profiles with greater noise
compared to the conventional tests (as seen in Fig. A.3). In the magnified, detailed
view in this figure, the EAF test profile had “spikes”. In this section, a frequency
analysis of the data is presented to investigate the frequency relationships to the
Appendix A 315

Fig. A.2  EAM die speed dependence (CD = 20 A/mm2) [A.2]. For the same electrical settings,
as the die speed is increased, the amount by which EAM has an effect on the metal’s formability
becomes reduced

Fig. A.3  Comparison of the forces for a conventional and EAF test and details on the “noise”
(Ti-G2) [A.3]. The EAF test load versus time profile shows much more “noise” in the profile than
the conventional compression test

EAF load versus time profile with greater noise [A.3]. A comparison of the load
recorded indicated the presence of a ‘noise’ in the data for the EAF. Figure A.3
shows the load for both tests for Ti-G2 at a die speed of 12.7 mm/min and an
initial current density of 25 A/mm2. A detailed view of the results is included.
Oscillation in the data is observed for the EAF test, while the conventional test
is clean. The Fast Fourier Transform (FFT) analysis on the data was performed to
316 Appendix A

Fig. A.4  Power spectrum x 10

4
5.5
(FFT) for an EAF test
5
(Ti-G2) [A.3]. The power

Power spectrum [N2/Hz]

spectrum shows the key 4.5
frequencies for this particular 4
set of electrical parameters 3.5
3
2.5
2
1.5
1
0.5
0
0 20 40 60 80 100 120 140 160 180 200
Frequency [Hz]

Fig. A.5  Position versus 3

time profile for conventional
and EAF compression of 2.5
Ti-G2. The position versus
time profile shows that the 2
Position (mm)

displacement is linear with

respect to time and that there
1.5
are no oscillatory portions of
the profile. This proves that Baseline
1
testing machine compliance EAF
was not the cause for the
oscillatory profile, as seen in 0.5
Fig. 5.23
0
0 2 4 6 8 10 12 14
Time (s)

rule out the influence from the electricity itself. The analysis for this particular test
found dominant frequencies in the spectrum of ~23 Hz and its multiples (Fig. A.4).
Figure  A.5 displays the position versus time profile for the conventional and
EAF compression tests in Fig. 5.23. This figure shows a linear displacement profile
with a constant die speed. This confirms that the compliance of the testing machine
was not responsible for the oscillatory portions of the load versus time plot above.
Figure A.6 plots the frequency as a function of initial current density and corre-
sponding current for the tests performed on stainless steel (SS304) at a die speed
of 12.7 mm/min. As the current density (i.e. current intensity/area) is increased, the
frequency is increased, which means a decrease in the period or time of one oscil-
lation. More specifically, as more electricity is applied, the oscillations occur at a
faster rate. Note that only the first peak frequency was considered here.
A first observation is that these values are lower than for titanium (G2,
Fig. A.3). While it is not exactly known why SS304 frequency values are less than
those of Ti-G2, some of the differences may be due to the crystalline structures or
Appendix A 317

Fig. A.6  Frequency response Current intensity [Amps]

for EAF loads for SS304 0 100 200 300 400 500
as a function of current 7
intensity and initial current 6

Frequency [Hz]
density [A.3]. As the current 5
intensity or current density
4
is increased, the frequency is
also increased 3
2
1
0
0 5 10 15 20 25 30
2
Current density [Amps/mm ]

Fig. A.7  Grain boundaries,

dislocations, and applied
electricity [A.3]. This figure
depicts how the electrons will
interact with the dislocations
and aid them in moving past
the grain boundaries in the
metal

the different alloying components in each metal. Also, based on these results, one
may speculate that there is a cyclic softening/hardening phenomenon present dur-
ing EAF. The input of electric energy results in alternating of electroplastic soft-
ening of the material, when the electric energy helps the dislocation to surmount
obstacles and move, and hardening of the material due to deformation (schematic
of the electrons impacting the dislocations is shown in Fig. A.7). This is respon-
sible for the oscillations in the “detailed view” portion of Fig. A.3. The soften-
ing results in lowering the forming load, while the work hardening contributes
to increase in load, until there is again enough energy for dislocation movement.
The larger current density results from more electric power input, thus the oscil-
lation in load happens faster and the frequency is higher. More specifically, as
318 Appendix A

7 60000
Frequency
6 EEC*I 2
50000
5

EEC*I2 [Amps2]
Frequency [Hz]
40000
4
30000
3
20000
2

1 10000

0 0
0 100 200 300 400 500
Current intensity [Amps]

Fig. A.8  Variation of the frequency response and the electric energy component [A.3]. There is
a direct relationship between frequency/electric energy component and the current density

the average electrical input power is increased, the dislocations will be able to be
“pushed” across the grain boundaries quicker and more often, thus increasing the
frequency of the oscillations.
The amount of electric power contributing towards helping the deformation
(PEAF−def) was defined earlier as being a fraction of the total electric energy, where
the fraction is the Electroplastic Effect Coefficient. Equation (11.36) shows this
relation as a product of two terms:
 
PEAF−def = ξ VI = ξ RI 2 = ξ I 2 · R (A.1)

where R is the resistance. Figure A.8 plots the variation of the first term, ξ I 2
by considering the Electroplastic Effect Coefficient function determined ear-
lier for 304 SS from the EAF tests at a die speed of 12.7 mm/min. For current
intensities of up to about 333 A, the frequency profile and the first term profile
are nearly identical. However, at current intensities above 333 A, the frequency
profile becomes higher than the first term profile. The difference may be due to
the change in the resistivity, thus resistance, of the part with the increase in tem-
perature due to the electric field applied. The temperature of the workpiece dur-
ing EAF tests at a current density of 10 A/mm2 (167 A) was 75 °C, compared
to 250 °C at a current density of 25 A/mm2 (417 A). The resistivity of the metal
increased by 15 % between these two temperatures [A.4].
The cyclic aspect of the load recorded during the electrically assisted form-
ing test and the apparent proportionality of the frequency with the energy input
need to be further investigated. Some similarity is found with the oscillatory flow
observed at high-temperature tensile tests, mainly for steels and aluminum, named
the Portevin–Le Chatelier (PLC) effect [A.5]. The oscillation in the load is attrib-
uted to an unstable plastic flow, and depends on temperature and strain rate. The
commonly accepted explanation for the PLC effect is based on a dynamic strain
Appendix A 319

aging model, and is due to the interaction between the moving dislocations and
diffusing solute atoms. In addition to temperature and strain rate, the PLC depends
also on alloy composition, crystal lattice, dislocation density and grain size. Since
all these are parameters that also influence the electroplastic effect, further tensile
tests will be performed and the possible correlation between the PLC and electro-
plastic effects will be investigated.

A.3 Power Quality Effects

When analyzing the effect of applied electricity on the formability of metals, it is

also important to quantify the quality of the electricity, as it may have an effect on
the process. This section compares the power quality and resulting stress-elonga-
tion profiles of both pieces of equipment. Specifically, a Lincoln R3S DC welder
will be compared to a Darrah fully-controllable SCR power supply. Shown below
in Fig. A.9 are the voltage versus time waveforms of the Lincoln welder and the
Darrah power supply, respectively, generated by an oscilloscope (please note that
the waveforms are at different magnitudes). For the figure, the Darrah supply was
run at a current of 300 A. A multimeter was hooked to the positive and negative
leads, as well as an oscilloscope. For this amperage, the multimeter read 0.738 V,
and the waveform amplitude was about 428 mV.
Similar tests were run on both the welder and the Darrah power supply, with
3.81 mm-diameter Titanium-G2 compression slugs. Figure A.10 shows the engi-
neering stress-strain profiles for these EAF compression tests run at a platen speed
of 12.7 mm/min for 2.54 mm of deformation, while 300 A were applied by each
respective power source, as indicated on the figure.
From the figure a few key points can be made. First, the stress-strain profiles
of the specimen with electricity supplied by the Darrah supply was higher overall

Fig. A.9  Welder and power supply waveform profiles. Voltage versus time waveforms of the
Lincoln welder and the Darrah controllable power supply are shown, with the Darrah supply pro-
ducing more of a square-like wave as compared to the welder waveform
320 Appendix A

Fig. A.10  Ti-G2 engineering 1.80E+03

stress-strain profiles. The 1.60E+03
engineering stress-strain

Engineering Stress (Mpa)

profiles were generated from 1.40E+03
running EAF tests at 300 A, 1.20E+03
where the electricity was
1.00E+03
supplied with a Lincoln R3R
500 welder, as well as with a 8.00E+02
Darrah SCR power supply
6.00E+02

4.00E+02 Lincoln R3R 500 Welder

2.00E+02 Darrah SCR Power Supply

0.00E+00
0 0.1 0.2 0.3 0.4 0.5
Engineering Strain

compared to the Lincoln welder. Secondly, the “noise” in the stress-strain profiles
was much less for the tests run with the Darrah power supply as compared to the
Lincoln welder. One reason for this may be that the Darrah welder is more con-
trollable and that it consistently output current closer to 300 A as compared to the
Lincoln welder. In comparison, the Darrah unit is self-compensating, whereas the
Lincoln welder is manually controlled by potentiometer. Figure A.10 shows that
there is a small, but noticeable difference in the stress-strain profiles of identical
specimens/tests when a different power supply is used.

A.2.1 Power Supply Output Voltage Waveforms

The voltage waveforms produced from the Darrah power supply were further ana-
lyzed such to characterize the applied electricity. The purpose of this investigation
was to determine how much variation was present in the output voltage profiles,
and then to analyze if this magnitude of variation would have an effect on the elec-
troplastic effect. To determine this, output voltage profiles were recorded from the
power supply using an oscilloscope for two different current settings. The amount
of voltage variation was compared for each current setting, and was also normal-
ized with respect to the average voltage output profile. The work shows that the
voltage variation is a constant related to the specifications of the power supply and
the percentage of the variation as compared to the average voltage value decreases
as the power output increases.
In the experiments, constant currents of 400 and 800 A were run and the fol-
lowing attributes were discussed: (1) the overall shape of the voltage profile, (2)
the peak-to-peak difference, or the variation in the voltage profile, and (3) a com-
parison between the variation in the voltage and the average overall voltage profile
value. Then, a discussion into the effects of the variations in the voltage waveform
on the electroplastic effect will be covered. All of the voltage waveforms were
produced by connecting an oscilloscope to the positive and negative leads of the
Appendix A 321

Fig. A.11  Power supply voltage waveforms at 400 A (1.00 s horizontal increments, 100 mV ver-
tical increments). The variation in the voltage profile was about 50 mV, as compared to the aver-
age voltage value of about 100 mV

power supply, while the power supply produced the specified current magnitudes
in a closed loop.
Figures A.11 and A.12 show the voltage waveform profiles of the power sup-
ply as it is continuously producing 400 A, where Fig. A.11 is a magnification of
the same waveform. In each figure, there are three parts, in which the lower peak-
to-peak voltage, the higher peak-to-peak voltage, and the zero voltage designation
line (in red) are all described. The actual lower and higher voltage values between
the magnified and the un-magnified tests may vary slightly, since there was a
higher accuracy in defining the true peaks with the magnified waveform pictures.
Therefore, the voltage variation values reported in this section will be taken from
the magnified figures at each respective current. From Figure A.11, it can be seen
that there was a difference in the peak-to-peak voltage of 56 mV for the 400 A
tests. When compared to the average voltage value of about 88 mV for this respec-
tive current, this variation in the voltage profile spanned 64 % of the total average
voltage value.
322 Appendix A

Fig. A.12  Magnified power supply voltage waveforms at 400 A (2.00 ms horizontal increments,
100 mV vertical increments). The variation in the voltage profile was about 50 mV, as compared
to the average voltage value of about 100 mV

Figures A.13 and A.14 display the un-magnified and magnified voltage wave-
form profiles from the power supply when outputting 800 A, respectively. The dif-
ference in the peak-to-peak variation of the voltage profile was 72 mV, which was
16 mV larger than the voltage variation in the 400 A-voltage profile. Additionally,
the average voltage profile was 174 mV, which increased since the current was
also increased. The voltage variation was 41 % of the average voltage value.
The overall shape of the voltage output profiles for the two different currents
were consistent with each other. The average voltage value increased as the current
increased. Additionally, the variation in the voltage profile only increased slightly
(16 mV) from 400 to 800 A. The variation in the voltage is unrelated to the elec-
trical power output, but rather related to the capabilities of the power supply. The
voltage variation spanned a larger portion of the average voltage profile for the
lower current (64 %) and spanned a lower portion of the average voltage profile
for the higher current (41 %).
The variation in the output voltage profile was a constant related to the power
supply. Regardless of the current value that was chosen to be output, the varia-
tion within the voltage profile was a constant between 56 and 72 mV. At lower
Appendix A 323

Fig. A.13  Power supply waveforms at 800 A (1.00 s horizontal increments, 100 mV vertical
increments). The variation in the voltage profile was about 50 mV, as compared to the average
voltage value of about 180 mV. a lower voltage measurement (160 mV). b upper voltage mea-
surement (210 mV). c zero voltage level

currents, the voltage variation was a larger percentage of the average voltage
value, and this percentage decreased as the current that was output increased.
The theories supported by work from this thesis describe that the flowing elec-
trons from the applied electricity impact lattice obstacles and cause resistive heat-
ing, and additionally dislocation motion if the impacted obstacles are dislocations.
When electrical power is increased, the quantity of electrons flowing through the
circuit is also increased. The quantity of flowing electrons directly relates to the
magnitude of heating effects and the magnitude of dislocation motion because the
greater number of electrons will provide for more impacts with lattice obstacles.
The variation in the voltage profile will not have an effect on the EAF tech-
nique. With the DC voltage profiles, all of the voltage values are positive. As the
average voltage value is what is intended and there are no negative values, the
voltage variation will not matter. If the voltage profile is translated into the cur-
rent profile (i.e. same shape) and then into the power profile, there will be times
when extra power is supplied and times when less power is supplied. However,
overall, the average power supplied will be the value that the power supply was
programmed to output. In relating the electrical power to the quantity of flowing
324 Appendix A

Fig. A.14  Magnified power supply waveforms at 800 A (2.00 ms horizontal increments,

100 mV vertical increments). The variation in the voltage profile was about 50 mV, as compared
to the average voltage value of about 180 mV. a lower voltage measurement (138 mV). b upper
voltage measurement (210 mV). c zero voltage level

electrons, more or less electrons will be applied at high and low points of the volt-
age peaks, respectively. But as long as the average power is what was requested,
the total quantity of electrons that passed through the metal will be the same,
regardless of the magnitude of the voltage variation.
To continue this discussion on voltage variation, there may be particular man-
ufacturing processes that may be sensitive to the variation in the output voltage
profile. For example, tensile-based processes, where the cross-section of the work-
piece decreases throughout the process, are more sensitive to variations in heat-
ing. If the voltage and current profile variations produce higher-than-average and
lower-than-average electrical power fluctuations, this could affect heating in the
tensile-based processes because the quantity of electrons will be increasing and
the cross-section by which they are flowing will be increasing. In forging, where
the cross-section of the workpiece increases, the heating due to the electricity is
not as sensitive as with the tensile-based processes, because although the quan-
tity of electrons is changing, the cross-section that they are flowing through is
increasing.
Future experimentation to further analyze the effects of power supply output
variation could be performed. Specifically, the voltage profiles of 400 and 800 A
Appendix A 325

from the power supply could be compared to the output at the same currents
from different power supplies, and even from a battery. Additionally, electrically
assisted compression and tension tests could be run with the power produced from
a power supply and from a battery, and the workpiece temperature and mechanical
profiles could be compared.

References

[A.1] Perkins TA, Kronenberger TJ, Roth JT (2007). Metallic forging using electrical flow as an
alternative to warm/hot working. J Manuf Sci Eng 129(1):84–94
[A.2] Jones JJ, Roth JT (2009) Effect on the Forgeability of Magnesium AZ31B-O When a
Continuous DC Electrical Current is Applied. In: International manufacturing science &
engineering conference, MSEC 2009-84116, p 10
[A.3] Bunget C, Salandro W, Mears L (2012) Sensitivities when modeling electrically-assisted
forming. In: International manufacturing science & engineering conference, MSEC2012-
7334, p 9
[A.4] Allegheny Ludlum Corp., product data sheet and Clark AF, Childs GE, Wallace GH (1970)
Cryogenics, vol 10, p 295, August (1970) (304SS resistivity vs. temperature)
[A.5] Yilmaz A (2011) The Portevin–Le Chatelier effect: a review of experimental findings. Sci
Technol Adv Mater 12:063001, 16p
Appendix B

B.1 Mechanical Experimental

Findings for EAF Tension

The subsequent sections display experimental results for various EAF tests that
allow for a deeper understanding of the deformation mechanisms present dur-
ing EAF and the electroplastic theory. This appendix examines the mechanical
response of Mg AZ31 as a result of its industrial use, low density, and overall low
formability in common sheet forming processes. However, the testing presented is
still directly applicable to other materials.

B.1.1 EAF with Square Wave Current Application

For the testing performed under EAF conditions using a square wave current appli-
cation, several parameter sets were tested with several replications to examine the
influence of an applied electrical current. The present design space explored is listed
in Table B.1 and these do not necessarily represent optimal parameters for improved
formability with this material. For all of the EAF tests in this section, the electrical
current was applied using a square wave shape with a given duration and period (as
listed in Table B.1). Also, the magnitude of the applied square wave corresponds to
the current magnitude listed in the table. A schematic of the application scheme is
given in Fig. B.1. For Parameter Sets 1–8, EAF deformation tests were performed
and the average result is presented. Also, for Parameter Sets 1–4, stationary electrical
tests with the same square wave conditions were performed and the samples were
then tested under room temperature conditions to failure. The results of these tests
will be compared to the deformation behavior of the as-received material to see if the
stationary electrical testing had any influence on the mechanical response. It should
be noted that Parameter Set 5 was manually varied within the bounds listed in the

© Springer International Publishing Switzerland 2015 327

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2
328 Appendix B

Table B.1  EAF square wave testing conditions

Parameter Current Pulse Pulse Duty Wave Platen
set magnitude (A) duration (s) period (s) cycle shape velocity
(mm/min)
1 800 0.3 60 0.50 % Square 2.54
2 800 0.5 60 0.83 % Square 2.54
3 500 0.5 60 0.83 % Square 2.54
4 500 1.0 60 1.67 % Square 2.54
5 Varying 0.1 Varying 4–6 Varying Square 2.54
500–1,000
6 500 0.5 30 1.67 % Square 2.54
7 1,000 0.5 30 1.67 % Square 2.54
8 800 0.5 30 1.67 % Square 2.54

Fig. B.1  EAF square wave

electrical current application
scheme

table such that an approximate steady forming force was achieved. Also, for the EAF
test figures presented in this section it should be noted that the observed stress dis-
continuities are representative of the period the electrical current was applied.
For Parameter Sets 1–4, these testing conditions were for model construction
and to act as independent data sets for model validation for the tensile thermo-
mechanical modeling in Chap. 6. As a result, these tests were performed with vary-
ing electrical current magnitudes and pulse durations. The results for Parameter
Sets 1–4 are displayed in Figs. B.2, B.3, B.4 and B.5. As shown in all of the fig-
ures, the applied electrical current decreased the material’s flows stress as a result
of bulk heating, thermal expansion, and electroplasticity. After the application of
the electrical current was removed the temperature decreased and the material
strengthened until the next application of electrical current. Also, the overall frac-
ture strain was not significantly affected with these applied electrical conditions.
To summarize Parameter Sets 1–4, Fig. B.6 shows the flow stress reduction as a
function of strain and input electrical energy per unit resistance. As shown, higher
electrical energy input resulted in larger stress reductions and the constructed sur-
face of these testing conditions is almost planar. Also, it should be noted that with
increasing electrical energy input, this directly relates to higher peak temperatures
throughout the material. The maximum temperature reached spatially and over
time for Parameter Set 2, Parameter Set 4, Parameter Set 1, and Parameter Set 3
are 224, 148, 108, and 77, respectively.
Appendix B 329

Fig. B.2  Experimental EAF

flow stress results (Parameter
Set 1)

Fig. B.3  Experimental EAF

flow stress results (Parameter
Set 2)

In addition to the EAF testing of Parameter Sets 1–4, stationary tests were con-
ducted at the same electrical conditions, however, no deformation was imposed
during the application of current. After performing the stationary electrical testing,
the specimens were cooled to room temperature conditions and deformed to fail-
ure at room temperature. The results from this testing are presented in Fig. B.7 for
Parameter Sets 1–4 along with the average room temperature baseline with a ±3
standard deviation bound. Statistically, the average room temperature baseline and
its ±3 standard deviations represent 99.73 % of the normally distributed spread
for all baseline tests. It is assumed that the mechanical response follows a normal
distribution due to the variation in the sheets samples being random. As shown, all
of the room temperature tests fall within the ±3 standard deviation bounds and can
be considered to not be outliers from the data set. As a result, it can be concluded
that significant modifications did not occur from the application of current which
330 Appendix B

Fig. B.4  Experimental EAF flow stress results (Parameter Set 3)

Fig. B.5  Experimental EAF

flow stress results (Parameter
Set 4)

alter the strength response of the material. It should be noted that there is some
variation (<16 MPa) and this may be a result of some dislocation annihilation due
to temperature rise or possibly small amounts of grain growth from elevated tem-
peratures. Since this material was warm rolled it is basically in a recrystallized
state off of the rolling mill and would have a fairly low dislocation density. Also, it
is not expected that the material recrystallized since the recrystallization procedure
is to hold at 205 °C for one hour after 15 % cold work, while the maximum tem-
perature reached for a short time (a few seconds) was 224 °C from all parameter
sets [B.2]. Additionally, the temperatures reached were well below the annealing
temperature of 345 °C and the hot working range of 230–425 °C. Thus, it can be
Appendix B 331

Fig. B.6  Stress reduction surface for PS1–PS4 at varying strain levels [B.1]

Fig. B.7  Experimental flow stress results of room temperature behavior for preceding stationary
electrical testing

concluded that stationary electrical treating did not significantly alter the mechani-
cal response, however, some dislocation annihilation or grain growth may have
occurred.
Parameter Set 5 was manually controlled and the magnitude and period were
varied throughout the testing. As can be seen in Fig. B.8 the material flow stress
332 Appendix B

Fig. B.8  Experimental EAF

flow stress results (Parameter
Set 5)

Fig. B.9  Experimental EAF

flow stress results (Parameter
Set 6)

can be significantly modified as a result of the applied current. With the observed
results seen for Parameter Set 5, it is suggested that the material flow stress can be
controlled with the application of electrical current. This type of process control
technique is discussed in Chap. 7 where constant forming force and constant stress
forming are designed and implemented during uniaxial forming.
In contrast to Parameter Sets 1–4, Parameter Sets 6–7 show experimental
results with a shorter pulse periods which translate to a greater amount of elec-
trical power being supplied to the sample with respect to time. The results for
Parameter Set 6 and 7 are shown in Figs. B.9 and B.10, respectively. Upon com-
paring the results from Parameter Set 3 to Parameter Set 6, it is seen that the stress
reductions at the same strain level are equivalent, however, Parameter Set 6 has
slightly greater elongation on average before failure (4.32 %).
Comparing Parameter Set 7 to Parameter Set 6 gives a significanly differ-
ent flow stress response where doubling the applied current did not double the
Appendix B 333

Fig. B.10  Experimental EAF flow stress results (Parameter Set 7)

Fig. B.11  Experimental EAF flow stress results (Parameter Set 8) and effect of specimen cool-
ing during EAF

observed stress reduction. This is expected as the electrical power is proprotion

to the square of the applied current. For Parameter Set 7, the current was discon-
tinued around a strain of 0.17 as a result of the stress approaching zero, so the
remainder of the test continued without current to material fracture. It is interest-
ing to note that the material was capable of hardening such that the stress reached
approximately 275 MPa while the prior stress state of the material was near zero.
The results in Fig. B.11 compare Parameter Set 8 without and with air cool-
ing applied uniformly to the specimen during EAF testing. The results show
334 Appendix B

significantly different flow stress responses where the test with cooling had rapid
increases in strength after the application of electrical current was discontinued
and lower flow stress reductions as compared to the non-cooled test. The rapid
increase in strength is a result of enhanced heat transfer (i.e. a reduced sample
temperature and thus a lower amount of thermal softening) resulting from forced
convection rather than only natural convection. This type of processing may be
advantageous to reduce thermal loads on the workpiece while also allowing for
the benefits of EAF. Additionally, an adaptation could be to track regions where
the material begins to neck and apply cooling to these regions to mitigate material
failure and allow for more uniform deformation. This could be accomplished with
in process Digital Image Correlation (DIC) data and a control system for the cool-
ing application.
To better understand the lower flow stress reductions of the cooled test as com-
pared to the non-cooled test, Fig. B.12 summarizes the stress reductions. As seen
in the figure, there is an average difference of 9 MPa in stress reduction comparing
the cooling and non-cooling tests. This lower stress reduction for the cooling test
is suggested to be a result of a reduction in the amount of temperature rise during
the application of current which leads to a reduced amount of thermal softening.
Although the exact temperatures were not recorded for these set of tests, it may
be stated that a difference of 9 MPa is quite small compared to the overall stress
reduction while an assumed temperature difference might be quite large by exam-
ining the quick increase in material strength after the discontinuation of electri-
cal current. Thus, this is an area that can further be explored as an ideal EAF test
would be performed without any bulk temperature rise (i.e. cooling application
technique that would suppress bulk joule heating) such that the observed reduction
in stress would be purely due to electroplasticity.

Fig. B.12  Summary of stress reductions for Parameter Set 8 with and without specimen cooling
Appendix B 335

B.2.2 EAF with Continuous Current Application

The influence of electrical current during deformation with a continuous wave

was also examined. These tests (summarized in Table B.2) were performed with
two testing velocities to examine the effect of strain rate and with varying current
magnitudes.
The results are presented in Fig. B.13 for the slow platen speed of 2.54 mm/min
at room temperature forming and EAF with a continuous current at 100, 150, and
200 A. The current application was applied at the beginning of the test while the
material was still being elastically deformed (i.e. prior to the material yield point).
As seen, with an increasing current magnitude the yield strength of the material
decreases. The observed decrease in the yield point can be a result of two possible
phenomenon. The first is that an increased bulk temperature due to Joule heating
adds thermal energy to the material lattice which allows for plastic deformation
to occur at a reduced level of stress. To quantify the temperatures during the test,
Fig.  B.14 shows the maximum temperature profile with respect to test time. As
shown, there is a significant difference in the initial temperatures when comparing

Table B.2  EAF continuous wave testing conditions

Test name Current magnitude (A) Wave shape Platen velocity
(mm/min)
100 A Continuous EAF 100 Continuous 2.54
150 A Continuous EAF 150 Continuous 2.54
200 A Continuous EAF 200 Continuous 2.54
100 A Continuous EAF 100 Continuous 25.4
150 A Continuous EAF 150 Continuous 25.4
200 A Continuous EAF 200 Continuous 25.4

Fig. B.13  Experimental
EAF flow stress results for
continuous application of
electrical current at
2.54 mm/min
336 Appendix B

Fig. B.14  Maximum
temperature response
for continuous current
application of electrical
current at 2.54 mm/min

Fig. B.15  Experimental
EAF flow stress results
for 200 A of continuous
electrical current at 2.54 mm/
min with discontinuation near
end of test

the 100 A to the 200 A test (100 °C at the yield point). The second effect is from
electroplasticity where the electrical current flow directly interacts with the mate-
rial lattice to reduce the stress as which material slip begins to occur. This concept
is not clearly visible in these tests, but the interaction with electrical current and
the yield point is clearly shown in the next section.
Figure  B.15 examines an individual continuous EAF test where a current of
200 A is applied to the material, and when an extremely low stress level is reached
(corresponds to 50 N forming load) the current application is discontinued. After
the applied current is stopped, the material then begins to cool and the material
begins to harden to approximately the initial yield strength of the as-received
material. This result from this test is important as this could be a processing strat-
egy that could be used in industrial applications. For example, the material can be
formed under very low stress conditions using EAF which is desirable and then
Appendix B 337

Fig. B.16  Experimental
EAF flow stress results for
continuous application of
electrical current at
25.4 mm/min

Fig. B.17  Temperature
response for continuous
current application of
electrical current at
25.4 mm/min

the material could be formed a small amount of the total strain to significantly
increase the strength of the material. The final strength of the material is critical
as this is the strength properties that the material will have in service unless addi-
tional post-forming thermal treatments are performed.
In addition, a faster platen velocity (10 times the initial velocity) was tested for
the continuous EAF tests, and the results are presented in Fig. B.16. The results
show similar trends, however, the yield strength of the material is not affected to
the extent of the slower platen velocity. This can be described again by examin-
ing the thermal response of the tests (Fig. B.17) where the maximum tempera-
tures are much lower at the point where yielding occurs. It should be noted that
the maximum temperatures observed for the faster platen speed reach higher peak
338 Appendix B

Fig. B.18  Experimental EAF flow stress results for continuous application of electrical current
at 2.54 and 25.4 mm/min

Fig. B.19  Experimental temperature at material yield stress for continuous application of elec-
trical current at 2.54 and 25.4 mm/min

temperatures (most notably for the 200 A test). This can be explained as a result of
a faster platen speed which creates an accelerated diffuse necking geometry. Due
to the diffuse neck, a smaller cross-sectional area is present and this creates an
area with higher heat generation which leads to a higher peak temperature. The
relation of area to heat generation is given in Eq. (6.16).
To compare the results from the two platen velocities, Fig. B.18 displays all of
the flow stress curves on one single figure. As seen the yield point is significantly
affected for the slower platen velocity and there is very little change for the faster
tests. Also, both platen velocities show reduced material flow stress with the slow
platen velocity having larger reductions in stress.
To specifically compare the maximum temperature at yield, the experimental
thermal results were analyzed and the observed temperatures are given in Fig. B.19.
Appendix B 339

It can be seen that with increasing continuous current magnitudes, the experi-
mental temperature at yielding increases. Also, the main difference is the slope for
each of the platen velocities. Thus, the significantly reduced yield point for the
slower platen speed can be partially due to the much higher temperatures observed
at the time of material yielding. It is interesting to note that the observed tem-
perature measurements at material yield are nearly linear for each platen velocity.
The significance or influence of this result is an area where more exploration is
warranted.

B.2.3 EAF with Incremental Current Application

In addition to examining the material flow stress response under square and con-
tinuous wave applications, this section examines the use of EAF in an incremental
manner. The term incremental means the electrically is applied to the sample, but
not during deformation. Additionally, the sample is allowed to cool to room tem-
perature before any deformation is imposed to the sample. The first experimen-
tal result is given in Fig. B.20 where room temperature deformation occurred to a
strain level of 0.07 mm/mm (corresponds to 5.08 mm of deformation) and then an
applied electrical pulse of 1,000 A for 1 s was applied while the sample was still
loaded, but no deformation was occurring. After the applied electrical pulse, the
sample was unloaded and allowed to cool to the room temperature. The sample
continued to be deformed at room temperature in steps of 5.08 mm until failure
(17 % increase over room temperature).
The main observation from this test is the modified yield point of the material
after the applied electrical current application. Since the material was stationary

Fig. B.20  Experimental EAF flow stress results for single pulse application of electrical current
during incremental forming
340 Appendix B

Fig. B.21  Temperature
response for single pulse
application of electrical
current during incremental
forming

during this time and only a load was applied, the bulk temperature effects are
expected to be small as a result of the maximum temperature only reaching 338 °C
for a very short period of time (Fig. B.21).
Again, for this material, the recrystallization temperature is 205 °C for 1 h with
approximately 15 % cold work and the annealing temperature for this material is
345 °C. Thus, this suggests that the applied electrical current had a direct influ-
ence on the material’s yield strength and the material was not annealed or recrys-
tallized purely from thermal effects. This observed effect could be a result of the
applied electrical current directly modifying the dislocation density of the material
and thus allowing for an instantaneous stress relief with the use of a single pulse
of electrical current. This is analogous to traditional incremental forming where
a conventional purely thermal process anneal treatment would be used to reduce
the material strength or remove some effects of cold work during processing. In
addition, as the stress of the material is reduced this then suggests a decrease in the
dislocation density of the material. As a result, the amount of dislocation annihi-
lated during this application allowed for the material to deform to a greater extent
over the room temperature baseline.
From a processing or economic standpoint this technique could be incorporated
directly in the forming equipment instead of having to remove the workpiece to
place in external annealing equipment which also has additional energy and capital
costs. After the incremental EAF pulse, the material was deformed under ambient
temperature conditions until failure. It is shown that without the application of cur-
rent the material returns to its prior flow stress before entering the plastic deforma-
tion regime.
The second experiment also applied EAF in an incremental format where the
material is only deformed under room temperature conditions and the current appli-
cation is applied at intervals where the specimen is loaded, but deformation is not
occurring. The results for this method and incremental room temperature forming
without any annealing are displayed in Fig. B.22. As seen, the EAF incremental
Appendix B 341

Fig. B.22  Experimental flow

stress results comparing RT
and EAF incremental forming

test reduced the material flow stress after each electrical pulse and also modified
the material yield strength. It should be noted that for the room temperature incre-
mental test the material yielded at the prior flow stress which was reached before
the unloading occurred. This is a result of the material being strengthened from cold
work imposed on the material and the material having this strength upon unloading.
Also, using this technique allowed for the specimen to have uniform elongation as
a result of the deformation only occurring at room temperature and without a tem-
perature gradient in the specimen (EAF tests with deformation display large thermal
gradients along the length axis of the specimen). This is beneficial as uniform strain
is present and the application of electrical current during the periods where defor-
mation is halted helps to reduce the material flow stress and reduce the dislocation
density of the material. Last, using this technique allowed for greater uniform strain
before fracture as compared to the room temperature and square wave EAF tests.
The increased amount of elongation can be attributed to the reduced strength of the
material (i.e. dislocation density reduction) which in turn reduces the local stresses
at pinned dislocation. The reduced local stresses in the lattice reduce the occurrence
of microvoid formation and cracking.

B.2.4 Mechanical Experimental Findings for EAF Tension

Conclusions

The main conclusions drawn from the presented experimental testing are:
• The EAF square wave tests decreased the material flow stress during the appli-
cation of current. The decrease is a result of thermal softening from bulk heat-
ing, thermal expansion, and electroplasticity. Also, the fracture strain was not
significantly affected by electrical parameters examined in this section.
• The electrical pretreating samples which were deformed at room temperature
conditions displayed only a small variation in mechanical response (<16 MPa).
342 Appendix B

Thus, it is suggested that the variation is a result of the dislocation density being
altered from electrical pretreating.
• The EAF square wave test with in-process manual control of the flow stress
level suggested that the material deformation force or flow stress could be con-
trolled, especially with a formal control strategy. This is investigated in Chap. 7.
• The EAF square wave tests that applied cooling with air showed that the stress
rapidly increased after the current was discontinued as compared to the non-
cooled test. Also, it was suggest that this technique could be used to reduce ther-
mal loads and possibly track and cool the local neck region during EAF. When
comparing the EAF cooled versus non-cooled test, there is an average differ-
ence of 9 MPa in stress reduction. This lower stress reduction for the cooling
test is suggested to be a result of a reduction in the amount of temperature rise
during the application of current which leads to a reduced amount of thermal
softening.
• The continuous current tests were performed at two platen velocities where
both showed reduced material flow stress, however, the slow velocity displayed
larger reductions. Also, the slow platen velocity showed a significant decrease in
the material yield point as compared to the faster velocity. This was a result of
the material temperature being much greater for the slower platen speed at the
point where yielding occurred.
• The EAF testing using an incremental method showed that the yield point of the
material was reduced after the applied electrical pulse. It should be noted that
the observed temperature increase from the electrical current was very unlikely
to cause the material to recrystallize or anneal from purely thermal effects
because of the measured temperature magnitude and the duration at an elevated
temperature. Last, this method allowed for uniform elongation during deforma-
tion with additional flow stress reductions.

References

[B.1] Jones JJ, Mears L (2013) Thermal response modeling of sheet metals in uniaxial tension
during electrically-assisted forming. J Manuf Sci. Eng 132:021011-1-11.
[B.2] United States Department of Defense (1998) Department of defense handbook: metallic
materials and elements for aerospace vehicle structures, MIL-HDBK-5H
Appendix C

C.1 Prior Cold Work Effect on EAF

Figure  C.1 displays the thermal profile of a stationary-electrical test performed

on 20 %CW/Annealed specimens at a starting current density of 10 A/mm2. The
maximum temperature of the annealed specimen was 200 °C, while the maximum
temperature of the worked specimen was 230 °C. Again, the cold-worked speci-
men temperature profile was significantly higher than the non-worked specimen
profile, and the temperature profiles did not level-off during the tests, but kept
increasing toward a steady-state value that would have been reached over a given
time period.
Figure  C.2 displays the thermal profile of a stationary-electrical test per-
formed on 30 %CW/Annealed specimens at a starting current density of 10 A/
mm2. Specifically, the maximum temperature of the anneal specimens was 200 °C
(which was the same as the annealed specimens at 20 %CW) and the maximum
temperature of the worked specimens was 300 °C. There was a very large differ-
ence in the thermal profiles of the worked and non-worked specimens, with the
worked specimen still producing the greater temperature profile by about 100 °C
compared to the annealed specimens.
Figure C.3 displays the thermal profile of a stationary-electrical test performed
on 40 %CW/Annealed specimens at a starting current density of 10 A/mm2. Again,
this was a large difference in the thermal profiles of the worked and non-worked
specimens, where the worked specimens reached a maximum temperature of about
365 °C and the annealed specimens peaked at about 200 °C.
Figure C.4 displays the thermal profile of a stationary-electrical test performed
on 50 %CW/Annealed specimens at a starting current density of 10 A/mm2.
Additionally, two contact load trials are shown in this figure. The annealed speci-
men temperature profile with the contact load of 2,225 N was significantly higher
than the other two profiles, which had a contact load of 62,300 N. This helps to
prove that the lower contact load has an effect on the specimen temperature, espe-
cially when it is still in the elastic region.

© Springer International Publishing Switzerland 2015 343

W.A. Salandro et al., Electrically Assisted Forming, Springer Series
in Advanced Manufacturing, DOI 10.1007/978-3-319-08879-2
344 Appendix C

Fig. C.1  20 %CW/Annealed 300

thermal profiles (stationary-
electrical tests at CD = 10 A/ 250
mm2). The difference in the 20% CW – 2225N

Temperature (°C)
thermal profiles (40 °C) stays 200
fairly consistent throughout
the duration of the stationary- 20% Anneal – 2225N
150
electrical test, as was the case
with the 10 %CW/Anneal 100
tests at the same current
density
50

0
0 2 4 6 8 10
Time (s)

Fig. C.2  30 %CW/ 350

Annealed thermal profiles 30% CW – 2225N
(stationary-electrical tests at 300
CD = 10 A/mm2). There is a
250
Temperature (°C)

consistent difference in the

thermal profiles (100 °C), 30% Anneal – 2225N
200
however, the magnitude of
the difference is much greater 150
than the 10 and 20 %CW/
Anneal tests 100

50

0
0 2 4 6 8 10
Time (s)

Fig. C.3  40 %CW/Annealed 400

thermal profiles (stationary-
350
electrical tests at CD = 10 A/ 40% CW – 2225N

mm2). The difference in the 300

Temperature (°C)

thermal profile marginally

increases over the duration 250
40% Anneal – 2225N
of the test to a difference of 200
about 160 °C
150

100

50

0
0 2 4 6 8 10
Time (s)
Appendix C 345

250
50% Anneal – 2225N

200

Temperature (°C) 150

100
50% CW – 62300N

50
50% Anneal – 62300N

0
0 2 4 6 8 10
Time (s)

Fig. C.4  50 %CW/Annealed thermal profiles (stationary-electrical tests at CD = 10 A/mm2 with

different static loads). This figure shows the difference that the static load can have on the tem-
perature of the part. The load of 2,225 N and the annealed part, which is still in the elastic region,
produced a much higher thermal profiles compared to another annealed specimen with a static
load of 62,300 N, which was well into the plastic region

Fig. C.5  30 %CW/Annealed 120

thermal profiles (stationary- 30% CW – 42275N
electrical tests at CD = 10 A/ 100
mm2 at 42,275 N)
Temperature (°C)

80
30% Anneal – 42275N

60

40

20

0
0 2 4 6 8 10
Time (s)

Figure  C.5 shows the thermal profiles of stationary-electrical compression

tests at a starting current density of 30 A/mm2, and a static load of 42,275 N. The
42,275 N represents the final load put on the specimens in order to work them to
30 % of their original height. The cold worked specimen temperature profile is
still higher than the non-worked profile. The maximum worked specimen tempera-
ture is 110 °C and the maximum annealed specimen temperature is 85 °C. Also
of note, is that the temperature difference between the %CW and Annealed ther-
mal profiles increase throughout the test duration. When comparing this figure to
Fig. C.2 (where there was only a load of 2,225 N on the parts), the temperatures
are much lower in this figure with the much higher load.
The thermal profiles in Fig. C.6 are for worked and non-worked specimens at
40 % of the original height, and under a static load of 50 kN (which was the final
346 Appendix C

Fig. C.6  40 %CW/Annealed 140

thermal profiles (stationary-
electrical tests at CD = 10 A/ 120
40% CW – 50063N
mm2 at 50,063 N)
100

Temperature (°C)
40% Anneal – 50063N
80

60

40

20

0
0 2 4 6 8 10
Time (s)

Fig. C.7  50 %CW/Annealed 140

thermal profiles (stationary-
electrical tests at CD = 10 A/ 120
50% CW – 62300N
mm2 at 62,300 N load)
100
Temperature (°C)

50% Anneal – 62300N

80

60

40

20

0
0 2 4 6 8 10
Time (s)

load required to deform the worked slugs to 40 %). The same trend follows, where
the worked specimen is much hotter than the non-worked specimen. The worked
specimen reached a maximum temperature of 117 °C and the annealed specimen
reached 82 °C. The same linear increasing trend as is shown in the previous figure
is also apparent in this figure. Compared to Fig. C.3, where the static load was
only 2,225 N, the difference between the worked and non-worked temperature
profiles increased only very slightly as the test progressed, whereas the profiles in
Fig. C.3 increased significantly as the tests progressed.
Figure C.7 displays the thermal profiles for stationary-electrical tests at a starting
current density of 50 A/mm2 for worked and non-worked specimens under a static
load of 62,300 N. The thermal profiles follow the same trend as the several figures
before it, in that the worked specimen temperature profile was hotter than the non-
worked profile, and the overall temperatures were much lower than the same worked
and non-worked specimens tested under a static load of 2,225 N. Specifically, the
worked specimen reached 120 °C, while the annealed specimen reached 80 °C.
Appendix C 347

Fig. C.8  40 %CW/Annealed 80
thermal profiles (conventional
70
compression). There was a 40% CW
greater difference between 60

Temperature (°C)
the annealed and worked
50
thermal profiles at 40 % 40% Anneal
(about 25 °C) compared to 40
30 %
30

20

10

0
0 1 2 3 4 5 6
Time (s)

Fig. C.9  40 %CW/Annealed 250

thermal profiles (EAF tests at
800 A). The difference in the 200
temperatures of the profiles is 40% CW
Temperature (°C)

about 40 °C, which is slightly

lower than the differences 150
for the 20 %CW/Anneal
40% Anneal
specimens 100

50

0
0 2 4 6 8
Time (s)

Figures C.8 and C.9 display the thermal profiles of 40 % worked/non-worked

specimens undergoing conventional compression and EAF compression, respec-
tively. The conventional compression tests in Fig. C.8 show a steadily increasing
difference in the temperature profiles between the worked and non-worked speci-
mens over the duration of the test. The worked specimen reached about 73 °C
and the annealed specimen reached about 60 °C. The EAF tests in Fig. C.9 show
a fairly consistent difference in the thermal profiles throughout the tests, as was
the case with the 20 % worked/non-worked EAF tests. In this case, the maximum
temperature is reached at about the halfway through the testing, where the worked
specimen is 195 °C and the annealed specimen is 155 °C.
The 40 % worked/non-worked tests are shown in Figs. C.10 and C.11 below.
The conventional compression tests in Fig. C.10 show a much larger difference in
the yield stress, due to the larger amount of cold-working. This stress difference
decreases throughout the tests, but there is still a small difference in the ending
stresses of the worked/annealed specimens (note that there was not a difference
between the 20 % worked/non-worked conventional compression specimens).
348 Appendix C

Fig. C.10  Stress-strain 3500

profiles (40 %CW/Annealed,
conventional compression). 3000

Engineering Stress (Mpa)

The stress-strain profiles of Conv – 40% CW
2500
the worked specimen were
much greater than the profiles 2000
of the annealed specimens at Conv – 40% Anneal
the beginning of the tests, and 1500
by the end of the tests, they
were only marginally higher 1000

500

0
0 0.1 0.2 0.3 0.4 0.5
Engineering Strain

Fig. C.11  Stress-strain 3000

profiles (40 %CW/Annealed,
EAF at 800 A). As was the 2500
Engineering Stress (Mpa)

case with the 20 %CW/

EAF – 40% CW
Anneal specimens, the 2000
stress-strain profile difference
at the end of the tests was 1500
greater than the respective EAF – 40% Anneal
conventional compression 1000
tests
500

0
0 0.1 0.2 0.3 0.4 0.5
Engineering Strain

Even though there is a higher dislocation density in the worked specimens, the
magnitude of the applied electrical power is able to overcome this difference in
dislocation density. With the 20 % cold-worked specimens, the applied electri-
cal power was able to overcome this dislocation density difference completely,
whereas the applied electrical power was not able to fully overcome the disloca-
tion density difference for the 40 % cold-worked specimens, and that is why there
is still a small difference between the stress-strain profiles of the conventional and
EAF compression tests. The EAF compression tests are depicted in Fig. C.11,
where there is an ending stress difference of about 350 MPa between the worked/
annealed specimens. There is still about the same stress difference of approxi-
mately 700 MPa at the elastic-plastic transition region. Additionally, the reduction
in flow stress at the end of the tests is different for conventional/EAF compres-
sion of each respective specimen condition. Specifically, the reduction in stress
between the conventional 40 %CW and the EAF 40 %CW tests is about 500 MPa
(3,000–2,500 MPa). Further, the difference between the conventional 40 %
annealed and the EAF 40 % annealed tests is about 750 MPa (3,000–2250 MPa).
This signifies that the high dislocation density in the %CW specimens affected
Appendix C 349

EAF, such that the applied electrical power did not have as great of an effect at
reducing stress as it had with the annealed specimens. To this end, the larger dif-
ference in dislocation density (due to the level of cold work) affected EAF’s effec-
tiveness to an even greater extent than the 20 % worked/non-worked specimens.

C.2 Microstructure Analysis under Tensile Loading

See Figs. C.12, C.13, C.14, C.15, C.16, C.17, C.18, C.19, C.20, C.21, C.22, C.23,
C.24, C.25, C.26, C.27, C.28, C.29, C.30 and C.31.

Fig. C.12  Microstructure
of stationary electrical test
(Sample 4) in Orientation
1 at L4

Fig. C.13  Microstructure
of stationary electrical test
(Sample 4) in Orientation
2 at L4
350 Appendix C

Fig. C.14  Microstructure
of electrically treated
deformation test (Sample 5)
in Orientation 1 at L1

Fig. C.15  Microstructure
of electrically treated
deformation test (Sample 5)
in Orientation 2 at L1

Fig. C.16  Microstructure of
EAF (800 A—0.5 s–60 s) test
(Sample 16) in Orientation
2 at L3
Appendix C 351

Fig. C.17  Microstructure of
EAF (800 A—0.5 s–60 s) test
(Sample 15) in Orientation
2 at L2

Fig. C.18  Microstructure of
EAF (800 A—0.5 s–60 s) test
(Sample 14) in Orientation
2 at L1

Fig. C.19  Microstructure of
continuous EAF (150 A) test
(Sample 13) in Orientation
2 at L3
352 Appendix C

Fig. C.20  Microstructure of
continuous EAF (150 A) test
(Sample 12) in Orientation
2 at L2

Fig. C.21  Microstructure of
continuous EAF (150 A) test
(Sample 11) in Orientation
2 at L1

Fig. C.22  Microstructure of
continuous EAF (200 A) test
(Sample 8) in Orientation
1 at L3
Appendix C 353

Fig. C.23  Microstructure of
continuous EAF (200 A) test
(Sample 7) in Orientation
1 at L2

Fig. C.24  Microstructure of
continuous EAF (200 A) test
(Sample 6) in Orientation
1 at L1

Fig. C.25  Microstructure of
continuous EAF (200 A) Test
(Sample 8) in Orientation
2 at L3
354 Appendix C

Fig. C.26  Microstructure of
continuous EAF (200 A) Test
(Sample 7) in Orientation
2 at L2

Fig. C.27  Microstructure of
continuous EAF (200 A) Test
(Sample 6) in Orientation
2 at L1

Fig. C.28  Microstructure
of room temperature
deformation to 5.08 mm
(Sample 9) in Orientation
1 at L1
Appendix C 355

Fig. C.29  Microstructure
of room temperature
deformation to 5.08 mm
(Sample 9) in Orientation
2 at L1

Fig. C.30  Microstructure
of room temperature
deformation to 5.08 mm
(Sample 10) with single
electrical pulse in Orientation
1 at L1

Fig. C.31  Microstructure
of room temperature
deformation to 5.08 mm
(Sample 10) with single
electrical pulse in Orientation
2 at L1