Prediction and Elimination of Galling in Forming Galvanized Advanced High Strength Steels (Ahss)

Home , Galling, Lubrication

PREDICTION AND ELIMINATION OF GALLING IN FORMING GALVANIZED ADVANCED HIGH STRENGTH STEELS (AHSS)

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree of Doctor of Philosophy in the Graduate

School of The Ohio State University

Hyunok Kim, M.S.

* * * * *

The Ohio State University 2008

Dissertation Committee: Approved by Professor Taylan Altan, Adviser Professor Bharat Bhushan ______Professor Gary L. Kinzel Adviser Associate Professor Mark E. Walter Mechanical Engineering Graduate Program

ABSTRACT

To improve crash worthiness and fuel economy, the automotive industry is using increasingly Advanced High Strength Steels (AHSS). In addition having limited formability, compared to mild steels, AHSS require high pressures at the tool-workpiece interface. This leads to frequent lubricant film break down and galling as well as reduction in tool life, resulting in a considerable challenge to stamping engineers. Galling is a form of adhesive wear and the economic impact of galling upon stamping production is significant due to the increase of scrap rate and die maintenance cost.

The major objective of this study is to predict and reduce galling in forming galvanized AHSS. For this purpose, a preliminary model was developed to predict the severity of galling for given lubricants, die materials and coatings in forming galvanized AHSS.

In this study, Finite Element Analysis (FEA) of selected stamping operations with AHSS was conducted to determine the critical pressure conditions that exist in practical stamping. Thus, an attempt was made to develop laboratory tribotests, e.g. Twist Compression Test (TCT), Deep Drawing (DDT) and Strip Drawing Test (SDT) that emulate practical stamping conditions. By using these tribotests, the performance of lubricants, die materials and coatings was evaluated for various grades of AHSS in terms of coefficient of friction (COF) and galling. The results of this study helped to develop a preliminary model for predicting galling by considering the empirical relationships between important tribological parameters (e.g. interface pressure, ii

sliding length, surface roughness, lubricant viscosity, hardness of zinc-coatings and tool coatings) and galling in forming galvanized AHSS. This model was applied to estimate the total number of parts that can be formed before severe galling occurs. Therefore, this model can offer a cost effective way to select practical and best combinations of lubricant, tool material and tool coating for reducing galling in forming galvanized AHSS. In addition, it can be further developed for use as a scheduling tool for the die maintenance before galling occurs, thus the unexpected downtime of stamping process can be eliminated.

iii

DEDICATION

To my beloved wife, Haw-Young, my lovely children, Benjamin, Samuel and Joseph.

ACKNOWLEDGMENTS

I would like to express my most sincere gratitude to my adviser, Professor Taylan Altan, for the research opportunity at the Engineering Research Center for Net Shape Manufacturing (ERC/NSM) and his exceptional guidance with thoughtful advice during my Ph.D. program at the Ohio state University. My great appreciation extends to Professor Gary Kinzel, Professor Bharat Bhushan, and Professor Mark Walter for serving on my dissertation committee and providing valuable comments.

Sincere thanks are extended to all the students and visiting scholars at the ERC/NSM, for their discussions and assistances in various aspects of this research work. My special thanks to Dr. Soo-Sik Han, Hyunjoong Cho, Yingyot Aue-u-lan, Ajay Yadav, Kyungbo Kim, Hariharasudhan Palaniswamy, Nimet Kardes, Partchapol Sartkulvanich, Ibrahim Al-Zkeri, Serhat Kaya, Johann Mai, Quigguang Yan and Dr. Gracious Ngaile.

Also, I would like to thank Mr. Frank Goodwin at the International Lead Zinc Research Organization (ILZRO) for the research grant (#ZCO-51 “Control of Galling During Forming Galvanized Advanced High Strength Steels”) that enabled this study to continue.

Finally, I want to thank my parents, pastor Keun Sang Lee and friends in church for their pray, support and encouragement during my Ph.D. program.

VITA

January 06, 1974 ...... Born - Seoul, South Korea

2000...... B.S. Mechanical Engineering, Sung Kyun Kwan University, Seoul, Korea

2002...... M.S. Mechanical Engineering, University of Michigan, Ann Arbor, US

2002 - 2008 ...... Graduate Research Associate, Engineering Research Center for Net Shape Manufacturing (ERC/NSM), The Ohio State University, Columbus, Ohio, US

PUBLICATIONS

Research Publications

1. Kim, H., Sung, J., Sivakumar, R., and Altan, T., “Evaluation of Stamping Lubricants using the Deep Drawing Test”, International Journal of Machine Tools and Manufacturer, Vol. 47/14, pp. 2120-2132, 2007.

2. Kim, H., Nargundkar, N., and Altan, T., “Prediction of Bend Allowance and Springback in Air-Bending”, ASME Transactions: Journal of Manufacturing Science and Engineering, Vol. 129, pp. 342-351, 2007.

3. Kim, H., Sung, J., and Altan, T., “Investigation of Galling in Forming Galvanized AHSS by Using the Twist Compression Test (TCT)”, accepted for the publication to the Journal of Materials Processing Technology, November 2007.

4. Kim, H., Han, S., Yan, Q., and Altan, T., ”Evaluation of Tool Materials, Coatings and Lubricants in Forming Galvanized Advanced High Strength Steels (AHSS)” accepted for the publication to the Annals of the CIRP, January 2008.

5. Kim, H., Shriniwas, P. and Altan, T., “Evaluation of New Lubricants for Cold Forging Without Zinc Phosphate Coating”, the 37th International Cold Forging Group (ICFG) Plenary Meeting, Istanbul, Turkey, Sep. 13-15, 2004.

6. Altan, T., and Kim, H., “Improvement of Tribological Conditions in Tube Hydro-Forming (THF) by using Environmental Friendly Lubricant Systems and Textured Tubes”, Proceedings of the 2005 NSF DMII Grantees Conference, Scottsdale, Arizona, Jan. 3-6, 2005.

7. Altan, T., and Kim, H., “Experimental and Numerical Studies on Friction & Lubrication in Metal Forming Processes”, Proceedings of the 2006 NSF DMII Grantees Conference, St. Louis, Missouri, July 24-27, 2006.

FIELDS OF STUDY

Major Field: Mechanical Engineering

Studies in: Metal Forming Technology, Design and Tribology

vii

TABLE OF CONTENTS Page ABSTRACT ...... ii

DEDICATION ...... iv

ACKNOWLEDGMENTS ...... v

VITA ………...... vi

LIST OF FIGURES ...... xiv

LIST OF TABLES ...... xx

LIST OF SYMBOLS ...... xxii

GLOSSARY ...... xxiii

CHAPTER 1 INTRODUCTION ...... 1

1.1. Application of AHSS in Forming Automotive Structural Parts ...... 1

1.2. Galvanized AHSS ...... 4

1.3. Galling in Forming Advanced High Strength Steels (AHSS) ...... 7

1.4. Research Motivation ...... 9

CHAPTER 2 RESEARCH OBJECTIVES ...... 10

CHAPTER 3 STATE-OF-THE-ART REVIEW ...... 11

3.1. Friction Models Used in Metal Forming Analysis ...... 11

3.2. Lubrication Mechanisms in Metal Forming ...... 14

viii

3.2.1. Primary Lubrication Mechanisms in Metal Forming ...... 14 3.2.2. Secondary Lubrication Mechanisms in Metal Forming ...... 16

3.3. Galling in Sheet Metal Forming Process ...... 16 3.3.1. Studies on Wear Models for Galling ...... 16 3.3.2. Tribotests to Evaluate Galling in Cold Forming Process ...... 17 3.3.3. Effect of Surface Texturing on Galling ...... 19

3.4. Lubrication, Tool Materials and Tool Coatings used in Forming AHSS ...... 20 3.4.1. Lubrication in Forming AHSS ...... 20 3.4.2. Tool Materials in Forming AHSS ...... 21 3.4.3. Tool Coatings in Forming AHSS ...... 25

CHAPTER 4 INVESTIGATION OF GALLING USING THE TWIST COMPRESSION TEST ...... 27

4.1. Principle of TCT ...... 27

4.2. Effects of Zinc Coatings on Galling ...... 28 4.2.1. Preparation of Sheet Specimens and Lubricants ...... 28 4.2.2. Test Conditions ...... 31 4.2.3. Experimental Results ...... 31 4.2.4. Effects of Galvanized Coating Characteristics on Galling Behavior...... 38 4.2.5. Change in the Surface Topography of Tool for increasing COF ...... 39

4.3. Effects of Lubricants on Galling ...... 48 4.3.1. Characterization of Sheet Coatings and Microstructures ...... 48 4.3.2. Experimental Conditions ...... 51 4.3.3. Experimental Results ...... 52 4.3.4. Discussions and Conclusions ...... 57

4.4. Effect of Tool Material and Surface Treatments on Galling ...... 57

4.4.1. Characterization of Tool Materials and PVD coatings ...... 57 4.4.2. Experimental Conditions ...... 59 4.4.3. Experimental Results ...... 60 4.4.4. SEM / EDS Analyses of Galling ...... 70 4.4.5. Discussions and Conclusions ...... 72

4.5. Effect of Interface Temperature on Galling ...... 73 4.5.1. Temperature Measurement During the Test ...... 73 4.5.2. Finite Element Analyses for TCT ...... 76

CHAPTER 5 EVALUATION OF LUBRICANTS USING THE DEEP DRAWING AND IRONING TESTS ...... 81

5.1. Principle of Deep Drawing and Ironing Tests ...... 82

5.2. Experimental Setup ...... 85 5.2.1. Description of the Tooling ...... 85 5.2.2. Test Procedures ...... 86 5.2.3. Lubricant Selection and Application Method ...... 87 5.2.4. Characterization of Sheet and Tool Surfaces ...... 88 5.2.5. Experimental Conditions ...... 89

5.3. Deep Drawing Tests Results ...... 90 5.3.1. Load-Stroke Curves ...... 90

5.3.2. Comparison of Maximum Punch Force, Fmax ...... 93 5.3.3. Comparison of Perimeter and Flange Draw-in Length ...... 94

5.4. Ironing Test Results ...... 95 5.4.1. Load-Stroke Curves ...... 95 5.4.2. Sidewall Thinning Distributions ...... 96

5.5. Finite Element Analyses of Deep Drawing and Ironing ...... 97 5.5.1. FE Models for Deep Drawing and Ironing ...... 98 5.5.2. Determination of Simulation Parameters and Thermal Properties ...... 99 x

5.5.3. FE Results of Load-Stroke Curve ...... 100 5.5.4. Predictions of Contact Pressure at the Tool-Workpiece Interface ...... 102 5.5.5. Temperature Distribution at the Tool-Workpiece Interface ...... 104

5.6. Summary and Conclusions ...... 106 5.6.1. Summary ...... 106 5.6.2. Conclusions ...... 107

CHAPTER 6 INVESTIGATION OF GALLING BY USING THE STRIP DRAWING TEST ...... 108

6.1. Experimental Setup ...... 108 6.1.1. Design of Experiment ...... 108 6.1.2. Description of SDT Tooling ...... 110

6.2. Evaluation of Die Coatings ...... 111 6.2.1. Load-Stroke Curves ...... 114 6.2.2. Surface Topography Change in Die Coatings Before and After the Test ...... 115 6.2.3. Ranking of Galling ...... 118

6.3. Evaluation of Lubricants ...... 119 6.3.1. Load-Stroke Curves ...... 121 6.3.2. Comparison of Maximum Punch Force ...... 122 6.3.3. Comparison of Strip Elongation ...... 124

6.4. Finite Element Analyses for Strip Drawing and Strip Ironing ...... 125 6.4.1. Preparation of FE Simulation Model...... 125 6.4.2. Load-Stroke Curve Predictions ...... 126 6.4.3. FE Predictions of Contact Pressures and Temperature Increase ...... 127

6.5. Summary and Conclusions ...... 129 6.5.1. Summary ...... 129

6.5.2. Conclusions ...... 130

CHAPTER 7 DEVELOPMENT OF A PRELIMINARY MODEL FOR PREDICTION OF GALLING ...... 131

7.1. Basic Modeling Concepts ...... 131 7.1.1. Perception of Galling in Stamping Galvanized AHSS ...... 132 7.1.2. Effects of Contact Pressure on Galling...... 133 7.1.3. Effects of Hardness of Zinc coatings and Tool Coatings on Galling ...... 133 7.1.4. Effects of Surface Roughness on Galling ...... 134 7.1.5. Effects of Lubricant on Galling ...... 135

7.2. Preliminary Model for the Prediction of Galling ...... 136 7.2.1. Regression Analysis for Determining the K value (Coefficient of Galling Rate) ...... 140

CHAPTER 8 A CASE STUDY OF GALLING PREDICTION MODEL IN FORMING B-PILLAR ...... 142

8.1. Finite Element Analyses of Forming B-Pillar Part ...... 142 8.1.1. Preparation of FEM Simulation Model ...... 143 8.1.2. FE Prediction of Thinning and FLD Diagram ...... 145 8.1.3. FE Prediction of Maximum Contact Pressure in Forming B- pillar Part ...... 145

8.2. Application of Galling Prediction Model ...... 147

CHAPTER 9 OVERALL SUMMARY AND CONCLUDING REMARKS ...... 150

REFERENCES ...... 152

APPENDIX A - TWIST COMPRESSION TEST MACHINE ...... 159

APPENDIX B - DEEP DRAWING TEST TOOLING ...... 161

APPENDIX C - IRONING TEST TOOLING ...... 163

xii

APPENDIX D - VISCOUS PRESSURE BULGE (VPB) TEST ...... 164

APPENDIX E - STRIP DRAWING TEST TOOLING ...... 170

APPENDIX F - STRIP DRAWING TEST RESULTS AND INPUT DATA FOR FEM SIMULATIONS ...... 174

APPENDIX G - LINEAR REGRESSION ANALYSIS RESULTS ...... 176

xiii

LIST OF FIGURES Figure Page Figure 1.1: The use of AHSS/UHSS in vehicles [Courtesy of SSAB] ...... 2 Figure 1.2: The change in usage of materials for Body In White parts [Pfestorf, 2005]...... 3 Figure 1.3: Application of galvanized metals on GM-10 [AISI 2004] ...... 4 Figure 1.4: Galvannealing Process for Hot-Dip Galvannealed Steel Sheet [Courtesy of POSCO] ...... 5 Figure 1.5: Cross section of GA Coating [Courtesy of ILZRO] ...... 6 Figure 1.6: Comparison of the structure of GA and GI coatings [Courtesy of Rooij et al. 2001B] ...... 6 Figure 1.7: Macroscopic / Microscopic / Nanoscopic scale views of galling (material transfer) in metal forming process ...... 7 Figure 1.8: Schematic of Galling in Sheet Metal Forming [Olsson 2006] ...... 8 Figure 1.9: Factors influencing on tribological failures in forming AHSS ...... 9 Figure 3.1: Relationship between contact pressure and frictional shear stress ...... 12 Figure 3.2: Stribeck curve showing onset of various lubrication mechanisms ...... 15 Figure 3.3: Ball scratching test [Courtesy of Carlsson et al. 2001A] ...... 17 Figure 3.4: Strip reduction test [Courtesy of Olsson et al. 2004] ...... 18 Figure 3.5: The U-bending test [Courtesy of Sato et al. 1998] ...... 18 Figure 3.6: Strip draw test (left) and Draw bead test for galling (right) [Courtesy of Vermeulen et al. 2001] ...... 19 Figure 3.7: Surface treatment effects on tool wear in U-channel drawing of DP steel EG, thickness 1 mm [IISI, 2006] ...... 21 Figure 3.8: Schematic of the tooling construction with tool steels as inserts in cast iron die in forming AHSS at VOLVO [Liljengren et al. 2006] ...... 24 Figure 3.9: Schematic of tooling with ceramic material inserts for stamping AHSS steel [Fuller et al. 2004] ...... 25 Figure 4.1: Schematic of TCT ...... 28 Figure 4.2: TCT tool insert (D2 steel) and sheet specimen (Dimensions are in mm) ...... 29

xiv

Figure 4.3: Micrographs and surface roughness of the sheet specimens before the test ...... 29 Figure 4.4: COF-Time curves obtained from TCT with different lubricants for DP 500 GI at 100 MPa interface pressure ...... 32 Figure 4.5: Ranks of galling on tool samples tested ...... 33 Figure 4.6: Ranks of powdering on sheet samples tested ...... 33 Figure 4.7: Galling observed in the tested tool at 170 MPa interface pressure (left) and the micrograph of tool surface (right) ...... 34 Figure 4.8: Qualitative comparison of the severity of powdering for various zinc coatings and lubricants tested at 100 and 170 MPa interface pressure ...... 36 Figure 4.9: Comparison of surface roughness of sheet samples tested with Lub A ...... 37 Figure 4.10: Comparison of surface roughness of sheet samples tested with Lub B ...... 37 Figure 4.11: SEM pictures of DP600 GI and DP500 GI...... 38 Figure 4.12: COF-Time curves obtained from the TCT up to COF=0.1 / 0.2 / 0.3 ...... 40 Figure 4.13: Images of a) D2 TCT tool sample tested and selected areas for surface analysis, b) micrograph of tool surface before the test ...... 40 Figure 4.14: Topographical change in tool surface at different COF with DP500 GI ...... 41 Figure 4.15: Topographical change in tool surface at different COF with DP600 Bare ...... 42 Figure 4.16: Topographical change in tool surface at different COF with DP600 GI ...... 43 Figure 4.17: Components of AFM machine (left) and the principle of AFM (right) ...... 45 Figure 4.18: Change in surface topography of tool samples as the COF increases (Measured area = 90 µm x 90 µm) ...... 46 Figure 4.19: Variation of Surface Topography of the tool sample tested for DP500GI with Lub B at 170 Mpa ...... 47

Figure 4.20: Variation of Ra-value of tool samples tested up to different COF ...... 47 Figure 4.21: SEM images of GA and GI coatings ...... 49 Figure 4.22: Microstructure of Dual Phase and TRIP steels ...... 51 Figure 4.23: Surface roughness of GI/GA coated sheets tested at 50 MPa ...... 56 Figure 4.24: Surface roughness of GI/GA coated sheets tested at 170 MPa ...... 56 Figure 4.25: Micrographs of tool surface before the test ...... 58 Figure 4.26: TCT results of PVD coated tools with DP600 Bare (uncoated) and Lub B at 50 and 170 MPa Contact Pressures ...... 68 Figure 4.27: Surface roughness of GA coated sheets tested with various tool materials and tool coatings ...... 69 xv

Figure 4.28: Surface roughness of GI coated sheets tested with various tool materials and tool coatings ...... 69 Figure 4.29: SEM photographs of the galled surface on D2 TCT tool (a) 2 mm resolution and (b) 50 µm in testing with GI coated sheet, (c) 2 mm resolution and (d) 50 µm in testing with GA coated sheet ...... 70 Figure 4.30: Temperature measurement using a thermocouple ...... 74 Figure 4.31: Variations of temperature near the tool-sheet interface measured during the test for DP500 GI with Lubes A and B at 170 MPa ...... 75 Figure 4.32: Variations of temperature near the tool-sheet interface measured during the test for DP600 Bare with Lubes A and B at 170 MPa ...... 75 Figure 4.33: Variations of temperature near the tool-sheet interface measured during the test for DP600 GI with Lubes A and B at 170 MPa ...... 76 Figure 4.34: FE model for TCT (left) and COF-time curve input data(right) ...... 77 Figure 4.35: Temperature distributions in the tool and sheet specimen at selected points for tracking temperature generation with FEA ...... 78 Figure 4.36: Comparison of FE results and experiment for DP600 Bare with Lubricant A at 170 MPa Interface Pressure ...... 79 Figure 4.37: FE prediction of pressure distribution at the tool-workpiece interface ...... 80 Figure 5.1: Schematic of deep drawing test and tool dimensions ...... 83 Figure 5.2: Schematic of ironing test ...... 84 Figure 5.3: Deep Drawing Tooling at ERC/NSM...... 85 Figure 5.4: The schematic of deep drawing and ironing test procedure ...... 87 Figure 5.5: Micrographs of DP590 GA specimen in different measurement scales ...... 88 Figure 5.6: Comparison of load-stroke curves obtained by testing three samples at a same testing condition ...... 90 Figure 5.7: Load-stroke curves obtained for various lubes tested at a low BHF ...... 91 Figure 5.8: Deep drawn vs. fractured samples ...... 92 Figure 5.9: Load-stroke curves obtained for various lubes tested at a high BHF ...... 92 Figure 5.10: Maximum punch force attained from deep drawing tests with various lubricants ...... 93 Figure 5.11: Comparison of perimeter at the flange of drawn cups coated with different lubricants ...... 94

Figure 5.12: Draw-in length, Ld, for different lubes tested at BHF 30 and 70 tons ...... 95 Figure 5.13: Load-stroke curves measured for ironing tests with various lubricants ...... 96

xvi

Figure 5.14: Comparison of sidewall thinning distributions obtained from the ironing tests with different lubricants ...... 97 Figure 5.15: a) FE model of the round cup drawing and b) Flow stress of sheet material (DP590GA) used for FE simulations ...... 98 Figure 5.16: FE model of the round cup ironing process ...... 99 Figure 5.17: Comparison of load-stroke curves predicted by FE simulations with deep drawing experiment with Lub B at BHF 30 tons ...... 101 Figure 5.18: Comparison of load-stroke curves predicted by FEM with ironing experiments ...... 102 Figure 5.19: FE simulation results at BHF 30 and 70 tons with COF =0.05, a) Tracking points for calculating pressure distribution at 76 mm stroke and b) Pressure distribution at the deformed sheet and the die corner radius ...... 103 Figure 5.20: FE prediction of pressure distribution between sheet and ironing die at 27 mm stroke ...... 104 Figure 5.21: FE prediction of temperature distribution in the final drawn cup (BHF=30 tons and COF=0.05) ...... 105 Figure 5.22: FE prediction of temperature distribution in the drawing die (BHF=30 tons and COF=0.05) ...... 105 Figure 5.23: FE prediction of temperature distribution at the workpiece and ironing die in FE simulation with m=0.1 ...... 106 Figure 6.1: Thinning distribution on the strip predicted by FEM ...... 109 Figure 6.2: Schematic of strip drawing test ...... 110 Figure 6.3: Strip drawing test tooling ...... 111 Figure 6.4: Micrographs of various tool coatings before the test ...... 112 Figure 6.5: Drawn and ironed specimens in different views ...... 113 Figure 6.6: Load-stroke curves obtained from strip drawing tests for various die coatings under dry and lubed conditions ...... 114 Figure 6.7: Microscope examination for the ironing zone of a die insert ...... 115 Figure 6.8: Comparison of surface topography change of various PVD-coated and uncoated die surfaces at 5 mm corner radius ...... 116 Figure 6.9: Change in surface topography of uncoated and coated dies ...... 117 Figure 6.10: Galled die surface after the test ...... 118 Figure 6.11: Ranking of galling for various die coatings ...... 118 Figure 6.12: Load-stroke curves obtained for various lubricants with an uncoated die ...... 121 xvii

Figure 6.13: Maximum punch force obtained from the SDT for GA coated sheet ...... 122 Figure 6.14: Maximum punch force obtained from the SDT for GI coated sheet ...... 123 Figure 6.15: Strip elongation of DP590 GA specimens after the test ...... 124 Figure 6.16: Strip elongation of DP600 GI specimens after the test ...... 125 Figure 6.17: FE model for the SDT ...... 126 Figure 6.18: Comparison of FE predictions of load-strokes with experiments ...... 126 Figure 6.19: Pressure distributions of strips for a) drawing and b) ironing and c) tested specimens ...... 128 Figure 6.20: FE prediction of temperature increase in SDT and SIT (COF=0.1) ...... 129 Figure 7.1: Development of galling in forming galvanized AHSS ...... 132 Figure 7.2: Classification of coating failure mode [Courtesy of Katagiri et al. 2007] ...... 134 Figure 7.3: Sliding length vs. GSI in SDT/SIT ...... 140 Figure 7.4: K-values determined for three different lubricant viscosities ...... 141 Figure 8.1: B-Pillar structure in a 4 door sedan model [ULSAB, 2002] ...... 143 Figure 8.2: B-Pillar simulation model ...... 144 Figure 8.3: Thinning distribution and FLD analysis of B-pillar simulation ...... 145 Figure 8.4: Contact pressure distribution in the B-pillar part ...... 146 Figure 8.5: Sliding length and maximum contact pressure on the B-pillar part ...... 148 Figure 8.6: Maximum number of B-pillar parts expected for various die surface coditions and lubricants ...... 148 Figure 8.7: Schematic of the tooling construction with PVD coated tool steel insert in cast iron die for forming AHSS B-pillar part ...... 149 Figure A.1: TCT machine at IRMCO ...... 159 Figure B.1: Deep Drawing Tooling at ERC/NSM …………...... 161 Figure B.2: Dimensions of Deep Drawing Tooling………...... 162 Figure B.3: A polished deep drawing die before the test…...... 162 Figure C.1: Ironing tooling at ERC/NSM…...... 163 Figure C.2: Ironing ring die dimensions...... 163 Figure D.1: Schematic of the tooling used for the VPB test; (a) before the test (b) after the test with a bulged sample………………...... 165 Figure D.2: Geometry of the sheet before and after the test...... 165 Figure D.3: Algorithm used to determine the flow stress curve from the VPB test...... 167 Figure D.4: Comparison of the dome height in tested samples (h=dome height)...... 168 xviii

Figure D.5: Comparison of the dome height in tested samples (h=dome height)...... 169 Figure D.6: Comparison of the flow stress curves obtained by VPB test for AKDQ, DP590, DP600, DP780, DP980 and TRIP780……………………………...... 169 Figure E.1: 3-D model of SDT tooling…………………………………….…………...... 170 Figure E.2: Engineering drawing of insert die………………………………………...... 171 Figure E.3: Engineering drawing of a die holder…….……………………………...... 172 Figure E.4: Engineering drawing of a fixture wing….………………………………...... 173 Figure F.1: Load-stroke curves obtained from SDT with TiCN coated dies……...... 174 Figure G.1: Relationship between K-values and lubricant viscosity with GI coating....178 Figure G.2: Relationship between K-values and lubricant viscosity with GA coating..179

xix

LIST OF TABLES Table Page Table 3.1: Selection of tool material for forming DP600, DP800 and DP1000 sheet materials [Liljengren et al. 2006] ...... 23 Table 4.1: Hardness values of metal substrates and zinc coatings ...... 30 Table 4.2: Properties of lubricants tested ...... 30 Table 4.3: Experimental conditions used for TCT ...... 31 Table 4.4: Qualitative Evaluation of Galling at 100 MPa Interface Pressure ...... 34 Table 4.5: Qualitative Evaluation of Galling at 170 MPa Interface Pressure ...... 35 Table 4.6: Comparison of the chemical compositions measured at both GI coatings of DP600 and DP500 sheet materials ...... 39 Table 4.7: EDS results of chemical components for DP590 GA and DP600 GI ...... 49 Table 4.8: Surface roughness and hardness of AHSS specimens ...... 50 Table 4.9: Test matrix to evaluate various stamping lubricants using the TCT ...... 52 Table 4.10: TCT results for various lubricants with a D2 tool material at 50 MPa interface pressure ...... 54 Table 4.11: TCT results for various lubricants with a D2 tool material at 170 MPa interface pressure ...... 55 Table 4.12: Test matrix to evaluate various tool materials and surface treatments using the TCT ...... 60 Table 4.13: TCT results for various lubricants with uncoated tool materials and GA coated DP steel at 170 MPa ...... 63 Table 4.14: TCT results for various lubricants with surface treated tool materials and GA coated DP steel at 170 MPa ...... 64 Table 4.15: TCT results for various lubricants with various tool materials and GA coated TRIP steel at 170 MPa ...... 65 Table 4.16: TCT results for various lubricants with various tool materials and GI coated DP steel at 170 MPa ...... 67 Table 4.17: EDS analysis results (a) spectrum results, (b) chemical composition for GI-induced galling, and (c) spectrum results, (d) chemical composition for GA-induced galling ...... 71 Table 4.18: Input data used in FE simulations ...... 77 xx

Table 5.1: Properties of lubricants tested ...... 88 Table 5.2: Conditions used in the deep drawing test ...... 89 Table 5.3: Conditions used in the ironing test ...... 90 Table 5.4: Input data used in FE simulations of deep drawing and iroing ...... 100 Table 6.1: Strip drawing simulation results at different die radius ...... 110 Table 6.2: Test conditions used in SDT for evaluating die coatings ...... 113 Table 6.3: Test conditions used in SDT for evaluating lubricants ...... 120 Table 7.1: Input data of TCT used for calculating GSI values ...... 138 Table 7.2: Relationship for GR and GSI in TCT ...... 138 Table 7.3: Input data of SDT/SIT used for calculating GSI values ...... 139 Table 8.1: Input data used for B-pillar simulation ...... 144 Table 8.2: Input data used for galling prediction model applied to forming B- pillar part ...... 147 Table D.1: Parameters used in the viscous bulge test………..………………………….... 168 Table F.1: Input data used for strip drawing simulations…..………………………….... 175 Table G.1: Results of linear regression analyses for various lubricant viscosities with GI coating…………………………………………………………………………….. 177 Table G.2: Results of linear regression analyses for various lubricant viscosities with GA coating……………...... 179

xxi

LIST OF SYMBOLS

Symbol Units Description

Heff (-) effective hardness of sheet and tool surfaces K (1/N) a coefficient of galling rate k (MPa) Shear strength of material

-2 -7 kw (-) Wear coefficient (10 ~ 10 )

Ls (mm) sliding length P (MPa) Normal pressure

Ra (mm or µm) Arithmetic average value of surface roughness

Ra,eff (-) effective surface roughness value of sheet and tool surfaces

Rz (mm or µm) Average peak-to valley height T (°C) Process temperature V (mm/sec) sliding velocity Y (MPa) Yield strength of the material

Zadh (mm) Wear depth α (-) Real contact area ratio σ (MPa) Effective stress or material flow stress ε (-) Total effective strain f’ (-) modified friction factor ∆t (sec) time interval

mr (-) modified shear factor

τf (MPa) Frictional shear stress µ (-) Coefficient of friction

xxii

GLOSSARY

AFM Acronym for Atomic Force Microscope, which is often used to measure the 3-D structure of engineering surfaces at high magnifications CVD Acronym for Chemical Vapor Deposition, which is a common process used for coating cutting tool inserts EDS Acronym for Energy Dispersive Spectroscopy, which is an advanced technique used to identify the elemental composition of a sample PVD Acronym for Physical Vapor Deposition, which is a common process used for coating cutting tool inserts. Scrap Rate The rate of cracked or fractured parts in production SEM Acronym for Scanning Electron Microscope, which is often used to observe the structure of wear surfaces at high magnifications Tool Wear Volume loss of the tool material at the contact interfaces with the chip and machined surface Viscosity A measure of the resistance of a fluid to being deformed by either shear stress or extensional stress SUS Acronym for Saybolt Universal Seconds, which is often used to indicate the viscosity of lubricant

xxiii

CHAPTER 1

INTRODUCTION

The increased use of Advanced High Strength Steel (AHSS) in the automotive industry, in order to improve crash worthiness and fuel economy, created several new challenges in forming these materials for manufacturing automotive structural components. One of these challenges is a form of adhesive wear, called galling, which leads to reduction in tool life. Forming of AHSS involves higher contact pressure and temperature at the tool-workpiece interface compared to forming mild steel. These unfavorable tribological interface conditions result in failure of commonly used lubricants, leading to galling which significantly reduces tool life.

1.1. Application of AHSS in Forming Automotive Structural Parts

Increased structural requirements for vehicle safety by National Highway Traffic Safety Administration (NHTSA), lower vehicle emissions and stringent Corporate Average Fuel Economy (CAFE) standards requires increased application of lighter materials in automobile construction to meet the standards. To satisfy these government regulations, automotive companies and their supplier industry have been focusing on the use of lightweight materials that can provide higher strength as well as better formability. Therefore, the use of

Advanced High Strength Steel (AHSS) and Ultra High Strength Steel (UHSS) has increased in automotive design, as shown in Figure 1.1.

Figure 1.1: The use of AHSS/UHSS in vehicles [Courtesy of SSAB]

AHSS is characterized as yield strengths from 210 to 550 MPa and tensile strengths from 270 to 700 MPa, while UHSS has yield strengths greater than 550 MPa and tensile strengths greater than 700 MPa. AHSS allows users to replace the parts made from conventional steels with thinner gages of AHSS. Thereby same or improved strength along with weight reduction can be achieved with AHSS steels.

Figure 1.2 provides a perspective on percentage usage of AHSS for Sport Utility Vehicle (SUV) Body In White (BIW) structural parts in the past and the current trend, reflecting the increased application of AHSS.

Figure 1.2: The change in usage of materials for Body In White parts [Pfestorf, 2005]

Despite all the efforts to commercialize several kinds of AHSS by the steel industry, forming of AHSS lead several challenges due to the higher strength and less formability of these materials compared to conventional steels and other problems such as a) inconsistency of material properties for the incoming steels, b) shear fracture during forming, c) higher contact pressure and temperature rise during bending, d) larger die stress and wear leading to reduced tool life, e) higher forming load and press capacity, and f) larger springback and sidewall curl leading to dimensional inaccuracy in the formed part.

1.2. Galvanized AHSS

Galvanized steel is considerably being used to reduce corrosion in automotive structural parts, as shown in Figure 1.3.

Figure 1.3: Application of galvanized metals on GM-10 [AISI 2004]

As the use of AHSS for automotive body panels increases, galvanizing technology for AHSS becomes also an important issue because the quality of galvanized or galvanneal coatings is significantly influenced by the characteristics of steel substrate, i.e. surface finish, microstructure and chemical reactivity of the steel surface, etc. Today, galvanizing technology is widely used in various AHSS, e.g. DP600/800/980, HSLA and TRIP 780 [Sreenivasan et al. 2006].

Galvanizing is a chemical process to coat the steel with layers of zinc. However, this process requires a rather complex facility. Figure 1.4 illustrates the galvannealing process to produce the Hot-Dip Galvanneal (GA) material. The galvanizing process begins with cleaning the coil to remove oily contaminations from the steel surface. This cleaning operation improves the coating affinity in the zinc pot. After cleaning, steel passes annealing and cooling operations to obtain the proper mechanical properties. In the next step, it passes the zinc pot to coat the zinc layers on the steel substrate. When steel is submerged in molten zinc, the chemical reaction permanently bonds the zinc to the steel through galvanizing. Therefore, the zinc permanently becomes a part of steel. The amount of coating can be controlled by the gas wiping, called Air Knife, Figure 1.4. After passing the zinc pot, the coated steel passes the GA furnace to obtain the uniform Fe-Zn alloy layers, which is important to improve the compatibilities of galvanized sheets with welding and painting.

Figure 1.4: Galvannealing Process for Hot-Dip Galvannealed Steel Sheet [Courtesy of POSCO]

A representative micrograph of the cross section in the GA coating is given in Figure 1.5. Eta (η) layer, the most outer layer, is just pure zinc (Zn) and the other layers, Zeta (ζ), Delta (δ) and Gamma (Γ) are composed of zinc and iron

in different ratios. The layer near the steel substrate has the higher iron content while the layer near the Eta (η) layer, pure zinc outer layer has the lowest iron content.

Figure 1.5: Cross section of GA Coating [Courtesy of ILZRO]

With the GA coating, hot-dip galvanized (GI) coating is also commonly used for automotive structural parts. Figure 1.6 compares the structure of GA and GI coatings.

Figure 1.6: Comparison of the structure of GA and GI coatings [Courtesy of Rooij et al. 2001B]

GA coating has multi layers of different phases and hardness levels, while GI coating has a single layer above the steel substrate. The hardness of GA is known to be much higher than the one of GI as shown in Figure 1.6. GA is known to have better formability and weldability compared to GI, meanwhile, GI has relatively better surface quality.

1.3. Galling in Forming Advanced High Strength Steels (AHSS)

The fundamental understanding of galling can be obtained in the microscopic and nanoscopic scales. At a microscopic scale, as shown in Figure 1.7, the tool-workpiece interface has numerous minute asperities and valleys. The magnitude, spacing and directionality of the surface topography in these mating surfaces play important roles not only in creating friction but also in sustaining or breaking a lubricant film designed to mitigate friction and wear.

Figure 1.7: Macroscopic / Microscopic / Nanoscopic scale views of galling (material transfer) in metal forming process

In the nanoscopic scale view as shown in Figure 1.7, the atomic layers (i.e. atomic layer scale of 0.2~100 nm) that are positioned within a regular metal

crystal lattice in the contacting bodies, are under the interface pressure (P) and

have a relative velocity (Vrel). Metallic surfaces are not ideally smooth and have a certain roughness even in the atomic scale [Klocke et al. 2002]. Therefore, friction takes place while atomic layers between tool surface and sheet surface slide in the relative motion leading to deflection and relaxation of atoms. During the relaxation of atomic layer, the deflected surface atoms oscillate and this oscillation transforms into thermal energy. This is why friction generates heat energy at the tool-workpiece interface.

In the nanoscopic scale view, galling begins with the transportation of atoms from the atomic layer in sheet surface to the atomic layer in the tool surface. In macroscopic scale view, galling can lead to the scoring of workpiece and reduction in too life as shown in Figure 1.8.

Figure 1.8: Schematic of Galling in Sheet Metal Forming [Olsson 2006]

Tribological failures in forming AHSS are the result of complex interactions among a large number of parameters, as shown in Figure 1.9. Galling begins with a lubricant failure and develops to the reduction of tool life.

Figure 1.9: Factors influencing on tribological failures in forming AHSS

1.4. Research Motivation

Galling, a form of adhesive wear, is one of the big challenges to the automotive industry in forming AHSS. Galling reduces tool life and also causes “scratches” or surface imperfections on the formed part, which increases the scrap rate in production. The major reason of galling in forming AHSS is the high contact pressure generated at the tool-workpiece interface. Therefore, a reliable prediction of interface pressure is significantly important to predict galling. Furthermore, it is clear that ability to control and reduce galling during sheet metal forming will have significant improvement in tool life and serious economic implications on production cost. Therefore, it is necessary to develop a reliable model to predict and reduce galling in forming AHSS. 9

CHAPTER 2

RESEARCH OBJECTIVES

The overall objective of the proposed research is to predict and reduce galling in forming galvanized AHSS. In particular, the specific objectives of the proposed research are to:

develop a methodology to evaluate lubricants, zinc-coatings, tool materials and tool coatings under near production conditions.

determine critical pressures and temperatures for initiation of galling at the tool-workpiece interface for selected lubricants, sheet and tool characteristics (e.g. material properties, surface roughness and hardness).

develop a model to predict galling in forming galvanized AHSS.

develop guidelines to select best and practical combinations of lubricant, die material and coating that reduce or eliminate galling in production.

CHAPTER 3

STATE-OF-THE-ART REVIEW

3.1. Friction Models Used in Metal Forming Analysis

In last several decades, two friction models have been commonly used to describe the frictional condition in metal forming processes. They are Amonton- Coulomb’s friction model (Eq. 3.1) and the shear friction model (Eq. 3.2). Both models quantify interface friction by lumping all of the interface phenomena, which are illustrated in Figure 1.9, into a non-dimensional coefficient or factor.

τ f = µ p (3.1) where µτ =coefficient of friction, p=normal pressure, f = frictional shear stress

However, in metal forming processes, the interface pressure, p, can reach a multiple of the yield strength of material. Thus, the linear relationship between

τf and p in Amonton-Coulomb’s model is not valid at high contact pressure levels because the shear stress, τf, cannot exceed the shear strength, k, of the deformed material that is normally workpiece. Therefore, the coefficient of friction becomes meaningless when µp exceeds τf. Thus, to avoid this limitation of Amonton-Coulomb’s model, the shear friction model was proposed by Orowan in 1943. In this model, as shown in Figure 3.1, the frictional shear stress, τf, at low

pressure is proportional to the normal pressure such as Amonton-Coulomb’s model, however it equals to the shear strength, k, at high pressure.

Figure 3.1: Relationship between contact pressure and frictional shear stress

In Eq. (3.2), m equals to zero for no friction and m equals to one for a sticking friction condition that is the case where sliding at the interface is preempted by shearing of the bulk material [Schey, 1983].

σ τσ==fm = mk f 3 01≤≤m (3.2) where fm =friction factor, = shear factor, k=shear strength and σ =flow stress

To consider the effect of real contact area ratio on friction, Wanheim and Bay [Wanheim et al., 1974 and Bay et al., 1975] proposed a modified shear friction model, namely the general friction model. As shown in Eq. (3.3), in this model, the friction shear stress, τf, is a function of the real contact area ratio, α. When two nominally flat surfaces are placed in contact, surface roughness causes

the discrete contact spots. The total area of all these discrete contact spots constitutes the real contact area and, in most cases of contact, this will be only a fraction of the apparent contact area. The ratio of the real contact area to the apparent contact area is known as the real contact area ratio, α.

σ τα==fkm' fr3 where (3.3) fm'= modifed friction factor , r = modified shear factor (as a function of real contact area) σα=flow stress, = the real contact area ratio = Real contact area/ Apparent contact area (Ar/Aa)

However, Wanheim and Bay’s model, does not consider the effects of lubricant behavior [Wilson 2004]. To take lubricant effects into account, a concept of complex model was proposed for boundary and mixed film lubrication regimes at the tool-workpiece interface by Bowden and Tabor (1967), as shown in

Eq. (3.4). In this model the frictional shear stress, τf , is given by

τ fa=+−ατ(1 α ) τ b where α = the real contact area ratio, (3.4) τa = average shear stress at contacting asperity peaks

and τb = average shear stress (lower stress) at the lubricant pockets

This model explicitly formulates important variables, the real contact area

ratio related to τa and the lubricant behavior related to τb that is influenced by viscosity, pressure, sliding speed and film thickness. If there is no lubricant at the

tool-workpiece interface then τb will be zero and the frictional shear stress, τf, will

be a function of real contact area ratio that equals to Eq. (3.3). To take into account of the lubricant behavior on friction, an artificial lubricant film was assumed to exist at the tool-workpiece interface and the variation of film thickness was calculated to characterize the variation of friction using the Reynolds equation in Fluid mechanics theory [Wilson et al. 1995 and Wilson, 2004]. However, although this approach has more detailed considerations to express the lubricant behavior, this model showed difficulties to apply to metal forming simulations.

3.2. Lubrication Mechanisms in Metal Forming

3.2.1. Primary Lubrication Mechanisms in Metal Forming

There are four primary lubrication mechanisms (i.e. dry / boundary / mixed film / hydrodynamic) observed in metal forming processes. As shown in Figure 3.2, the Stribeck curve illustrates the onset of these various types of lubrication as a function of lubricant viscosity, η, sliding velocity, v, and normal pressure, p [Schey, 1983].

Dry condition means no lubrication at the mating surfaces, thus friction is high. This condition is desirable in only a few selected forming operations (e.g. hot rolling of plates or slabs and non-lubricated extrusion of aluminum alloys).

Boundary lubrication is defined as a condition where the solid surfaces are so close together that surface interaction between mono or multi molecular films of lubricants and the solid asperities dominates the contact [Bhushan, 2002]. Boundary lubrication is the most widely encountered lubrication condition in stamping, forging and hydroforming.

Mixed-layer lubrication is also frequently encountered in sheet metal forming. In this case, the micro-peaks of the metal surface experience boundary 14

lubrication conditions and the micro-valleys of the metal surface become filled with the lubricant.

Hydrodynamic lubrication condition rarely exist in metal forming and it occurs at only under very specific conditions of velocity and temperature (e.g. sheet rolling operation).

Dry µ > 0.3 Dry Boundary Mixed ) µ

Hydrodynamic Boundary Lubrication 0.1< µ < 0.3 coefficient of friction (

ηv p Mixed-Layer Lubrication 0.03< µ < 0.1

Hydrodynamic or Full Film Lubrication

film thickness µ < 0.03

ηv p

Figure 3.2: Stribeck curve showing onset of various lubrication mechanisms

3.2.2. Secondary Lubrication Mechanisms in Metal Forming

In addition to the primary lubrication mechanisms, when the trapped lubricant permeates to the real contact surface, two different types of secondary lubrication mechanism, called Micro-Plasto-Hydro-Static Lubrication (MPHSL) and Micro-Plasto-Hydro-Dynamic Lubrication (MPHDL), may also develop. Beginning with the mixed-layer lubrication condition, the trapped and pressurized lubricant may escape from the valleys of the surface. When lubricant escapes from the valleys of the surface, it forms a thin film between the surface asperities and the die where boundary lubrication previously dominated. Thus, the frictional stress is further reduced. Experimental evidence of MPHSL and MPHDL mechanisms for sheet metal forming has been found by many researchers [Azushima et al, 1991, 1995, 1996, 2000A and 2000B, Bech et al. 1998 and 1999, Le et al. 2003, Lo et al. 1997, Mizuno et al. 1982, Shimizu et al. 2001, Sorensen et al. 1999, and Sutcliffe et al. 2001]. To reduce the surface friction, the optimized textured sheets can be used to enhance the MPHSL and MPHDL mechanisms [Pfestorf et al. 1998].

3.3. Galling in Sheet Metal Forming Process

Galling is a form of adhesive wear that occurs due to material transfer from the sheet to the forming tool during metal forming operations. In this section, analytical and experimental studies on galling in sheet metal forming are summarized.

3.3.1. Studies on Wear Models for Galling

Several wear models were developed to predict galling [Rooji et al. 2001A , 2001B and Heide et al. 2001, 2003]. All these models were developed by using 16

simple assumptions of the contact surfaces while sheet forming process involves much more complex contact phenomenon.

3.3.2. Tribotests to Evaluate Galling in Cold Forming Process

Several ASTM wear testing standards were developed to evaluate galling, mostly focused on mechanical rolling and sliding elements [Blau et al. 1999] but not with emphasis on conditions in metal forming. A ball scratching test was used to evaluate galling in sheet materials, as shown in Figure 3.3 [Carlsson et al. 2001A, 2001B and Heide et al. 2001, 2003]. However, in this test, the plastic deformation of the sheet in stamping process and its effect on galling were completely ignored. Therefore, the results cannot be applied directly to sheet metal forming operations.

Figure 3.3: Ball scratching test [Courtesy of Carlsson et al. 2001A]

Olsson et al. (2004) and Andreasen et al. (1998) used a strip reduction test to study on galling. In this test, the thickness of the metal strip is reduced intentionally between a non-rotating hardened tool pin and a supporting plate, as shown in Figure 3.4.

Figure 3.4: Strip reduction test [Courtesy of Olsson et al. 2004]

The reduction of strip thickness generates a high contact pressure and temperature close to production. The tendency towards galling is evaluated based on the length of the sheet metal that could be drawn over the tool before the first galling appears in the strip metal. The complex metal flow conditions that exist in real sheet forming processes were not reflected in this test.

The U-bending test was conducted for galling studies as shown in Figure 3.5 [Sato et al. 2000 and Skare et al. 2003]. Similar tests were conducted at Volvo in order to evaluate galling in forming galvanized and uncoated AHSS (DP600) with different tool materials (e.g. Calmax, Rigor, Sleipner, Sverker, Cast iron 0732, Cast iron 0741 and Vanadis) [Skare et al. 2003].

Figure 3.5: The U-bending test [Courtesy of Sato et al. 1998]

The strip draw test was extensively used to evaluate the galling, as shown in Figure 3.6 [Wichern et al 1999 and Vermeulen et al. 2001]. Draw bead test, Figure 3.6, was used to evaluate the effects of different zinc-coated cold rolled steel (galvannealed, electrogalvanized and galvanized) on the friction and galling behaviors [Dalton et al 1992]. Galvanized steel sheet material resulted in less interface friction compared to uncoated steels.

Figure 3.6: Strip draw test (left) and Draw bead test for galling (right) [Courtesy of Vermeulen et al. 2001]

3.3.3. Effect of Surface Texturing on Galling

Textured sheets are widely used in sheet metal forming to improve the performance of lubricants. The textured surface contains pockets which serve as lubricant retainers. The trapped lubricant can support the contact load partially by its contact pressure thereby reducing the contact areas and frictional stress [Kudo et al. 1965, Lo et al. 1997, Pfestorf et al. 1998 and Vermeulen et al. 2001]. During the plastic deformation, the pockets are squeezed. This results in the escape of lubricant from the pockets and reducing the friction between the sheet and the tool surface [Wichelborg et al. 2000]. The galling behaviours in textured sheet material with different texture dimensions were studied by Vermeulen et

al. (2001). Their experiments concluded that there exists an optimum size that gives best resistance to galling for the uncoated sheet material.

3.4. Lubrication, Tool Materials and Tool Coatings used in Forming AHSS

3.4.1. Lubrication in Forming AHSS

AHSS forming requires enhanced performance of lubricant to form the part, because these steels generate larger interface pressures and temperatures. The interface pressure and temperature greatly affect lubricant viscosity and performance [Brown et al. 2006]. Lubricant must provide not only an interfacial barrier between die and sheet, but also a cooling effect to reduce the effect of heat generated from high deformation energy and friction. A recent study indicated that, while forming a UHSS automotive suspension part, the average temperature after die surface reached 116 ºC. in real production conditions [Jeffery 2004]. Under this elevated temperature condition, High Solids Polymer (HSP) lubricant was found to give consistent tool protection and friction control, while straight oil lost their performance [Jeffery 2004]. To endure this elevated tool temperature, several liquid lubricants such as some petroleum-based and water-based synthetic lubricants with Extreme Pressure (EP) additives are recommended by lubricant manufacturers [Brown et al. 2006]. Several organic dry film lubricants (e.g. mixed acrylic/polyurethane/polyester resin) were used to evaluate galling and tool wear for 55% Al-Zn coated steel sheet and hot dip galvanized steel using several laboratory tests (e.g. the scratching test and pin- on-disk test) [Carlsson et al. 2001A and 2001B]. Organic coatings were found to have lower friction and low galling tendency for both materials. A patented film that is made from urethane resin – containing silanol was used to withstand a

high temperature of 200 ºC involved in forming AHSS while other resin based films failed at elevated temperatures [Mori et al. 2000].

3.4.2. Tool Materials in Forming AHSS

Forming AHSS leads severe interface conditions that can accelerate adhesive and abrasive wears in stamping dies. Tool wear is often concentrated at the draw radii or draw beads that encounter higher contact pressure compared to other flat surfaces. Thus, die corner radius and draw bead radius, which are conventionally made from ductile cast iron, are used to be flame hardened to increase hardness for better wear resistance in automotive stamping of steel and aluminum alloy sheet materials. International Iron and Steel Institute (IISI) investigated the tool wear in forming of U-channel parts of AHSS with different tool materials, as shown in Figure 3.7.

Figure 3.7: Surface treatment effects on tool wear in U-channel drawing of DP steel EG, thickness 1 mm [IISI, 2006] 21

It could be observed that the conventional die material, Spheroidal graphite cast iron (GGG70L) with flame hardening resulted in significant amount of wear, while D2 steel hardened and plasma nitrated gave negligible wear.

Also, in the tool wear study conducted by Volvo, the stamping dies made from conventional spheroidal graphite cast iron gave galling in the tool after just 25 strokes [Skare et al. 2003]. This indicates that conventional cast die material does not have sufficient wear resistance and anti-galling properties against high contact pressure involved in forming AHSS sheet materials.

Bohler-Uddeholm developed several tool material manufactured by conventional ingot casting with or without Electro Slag Melting (ESR), Spray Forming (SF) or powder metallurgy (PM) for use in forming AHSS [Sandberg et al. 2004]. They are as follows: • AISI D2 – a high carbon/chromium tool steel with high amount of chromium carbides, 13% and a hardness range of 59-61 HRC after secondary hardening. • Carmo/Calmax – The first generation of car body die steels with a pure martensitic matrix structure. Optimized properties with regard to wear, ductility, weldability and hardness. Hardness of these materials is about 58 HRC after low temperature tempering. • Diemax – a new developed matrix steel with an optimal combination of ductility, hardenability and temper resistance at a hardness up to 57-58 HRC. The hardness is achieved by an high temperature tempering that facilitates surface coating. • Caldie – a new car body die steel developed from the old grade Carmo but with high temperature tempering gives a hardness of 60-62 HRC and while maintaining a very good ductility, hardenability and temper resistance. • Roltec – a spray formed grade in between PM and conventional cold work steels. Steel with a good combination of abrasive wear resistance and ductility with a hardness potential 64 HRC versus 61 HRC for D2 steel after secondary hardening. • Weartec – a spray formed grade with a very high abrasive wear and good ductility.

• Vanadis 4 Extra – an optimal combination of ductility and wear resistance, mixed abrasive and adhesive of all PM cold work steels. • Vanadis 6 – a PM steel with better abrasive wear resistance than Vanadis 4 Extra. • Vanadis 10 – a high vanadium alloy PM grade with the best combination of high abrasive wear resistance and ductility. • Vancron 40 – a nitrogen alloyed PM steel with very good low-friction properties for excellent galling and adhesive wear resistance.

Bohler-Uddeholm tested the tool materials with different sheet materials used in automotive industry and developed guidelines for selecting the tool materials in forming AHSS as shown in Table 3.1.

Sheet material : DP 600 Sheet thickness, t <1.2 mm Sheet thickness, t >1.2 mm Tool Coated Uncoated Coated Uncoated Comments Punch GGG 70 L GGG 70 L GGG 70 L GGG 70 L Steel inserts at expected wear areas Lower die GG25 GG25 GG25 GG25 Inserts Calmax+ Nitr Calmax+ Nitr+PVD Calmax+ Nitr Calmax+ Nitr+PVD All inserts are through hardened Sleipner Sleipner + PVD Sleipner + Nitr Sleipner + PVD All inserts are through hardened Blank holder GGG 70 L GGG 70 L GGG 70 L GGG 70 L Without draw beads Inserts Calmax+ Nitr+PVD Calmax+ Nitr Calmax+ Nitr+PVD With draw beads Sleipner + PVD Sleipner Sleipner + PVD With draw beads

Sheet material : DP 800 Sheet thickness, t <1.2 mm Sheet thickness, t >1.2 mm Tool Coated Uncoated Coated Uncoated Comments Punch Sleipner Sleipner Sleipner Sleipner Wrought and through hardened Calmax Calmax Calmax Calmax Through hardened Lower die GG25 GG25 GG25 GG25 Inserts Sleipner +PVD Sleipner +PVD Sleipner +PVD Sleipner +PVD Wrought and through hardened Sleipner 21 +PVD Sleipner 21 +PVD Sleipner 21 +PVD Sleipner 21 +PVD Through hardened Blank holder Sleipner +PVD Sleipner +PVD Sleipner +PVD Sleipner +PVD Wrought and through hardened Sleipner 21 +PVD Sleipner 21 +PVD Sleipner 21 +PVD Sleipner 21 +PVD Through hardened

Sheet material : DP 1000 Sheet thickness, t <1.2 mm Sheet thickness, t >1.2 mm Tool Coated Uncoated Coated Uncoated Comments Punch Sleipner Sleipner Sleipner Sleipner Wrought and through hardened Lower die GG25 GG25 GG25 GG25 Inserts Sleipner +PVD Sleipner +PVD Sleipner +PVD Sleipner +PVD Wrought and through hardened Sleipner 21 +PVD Sleipner 21 +PVD Sleipner 21 +PVD Sleipner 21 +PVD Through hardened Blank holder Sleipner +PVD Sleipner +PVD Sleipner +PVD Sleipner +PVD Wrought and through hardened Sleipner 21 +PVD Sleipner 21 +PVD Sleipner 21 +PVD Sleipner 21 +PVD Through hardened

Table 3.1: Selection of tool material for forming DP600, DP800 and DP1000 sheet materials [Liljengren et al. 2006]

Volvo used the new tool steels developed by Bohler-Uddeholm in the laboratory experiments for forming U channels as well as in production tools for structural parts. Based on the experiments and production results, preferred tool material for punch, blank holder and die in forming DP600, DP800 and DP 1000 are given in Table 3.1. Development of these special tool materials to withstand high temperature and pressure in forming AHSS has resulted in significant change in design and manufacturing of the tools in forming AHSS. Conventionally, the dies for automotive panels are cast specifically for the part to be manufactured as a single piece from spheroidal graphite cast iron. Similar procedure is followed for manufacturing tools for AHSS as well. However, at the high contact pressure and temperature locations such as corner radius in the punch, die and blank holder, the specially developed tool material are added as inserts to increase die life and reduce the die cost. An example tooling with tool steel inserts is shown in Figure 3.8 [Asnafi et al. 2004 and Liljengren et al. 2006].

Figure 3.8: Schematic of the tooling construction with tool steels as inserts in cast iron die in forming AHSS at VOLVO [Liljengren et al. 2006] 24

Fuller et al. (2004) investigated the use of ceramic tool material to form AHSS. Among the different ceramic materials, silicon nitride ceramic was found to be best suited to form AHSS materials. Figure 3.9 shows an example tool with ceramic insert for draw beads that have higher contact pressure and temperature during forming. Ceramics has negligible affinity to the sheet material and provides good thermal and wear resistances. However manufacturing of ceramic inserts with complex contour profiles required for stamping dies is normally not practical because of the manufacturing cost [Fuller et al. 2004].

Figure 3.9: Schematic of tooling with ceramic material inserts for stamping AHSS steel [Fuller et al. 2004]

3.4.3. Tool Coatings in Forming AHSS

Various die coating techniques are being used in automotive stamping industry to increase tool life. Examples are CVD (Chemical Vapour Deposition), PVD (Physical Vapour Deposition) and TRD (Thermal Reactive Diffusion), which is also known as TD (Toyota Diffusion) [Jeffery 2005 and Madorsky et al.

2005]. Forming AHSS requires good lubrications for increasing the performance of the conventional coatings, compared to forming mild steel. Production results at Volvo indicate that the additional lubricant is required to reduce wear and galling when forming AHSS using PVD or CVD coated tools [Liljengren et al. 2006]. Duplex treatments of tools that involve plasma nitriding followed by PVD coating could reduce the required lubrication by 75 % compared to conventional CVD and gave better anti-galling and wear resistance properties [Liljengren et al. 2006]. Recent studies through laboratory test (ball-on-disk test) and field test (deep drawing of example part) showed that several amorphous hard carbon coatings such as DLC (diamond-like carbon) and WC/C (hard amorphous hydrogenated carbon) with regular lubricant are promising die coating materials for galvanized AHSS, stainless steel, aluminium alloy sheet materials that are prone to galling [Murakawa et al. 1998, Sato et al. 1998 and 2000]. Also, amorphous hard carbon coatings performed better compared to commonly used PVD and CVD coatings [Podgornik et al. 2004B and Skare et al. 2003]. Production results at Volvo also indicates that WC/C offers good anti-galling properties and less wear in the dies used to make parts from AHSS without any lubrication [Liljengren et al. 2006].

Various tool materials and coatings were tested for use with zinc-coated deep drawing steel (DDS) by the Bending Under Tension (BUT) test [Shih et al. 2004]. Cast iron and cast iron with chrome plating gave the lowest friction coefficient from the BUT test in comparison to other tool materials (D2, Kirksite and Cast Steel), regardless of the type of zinc-coatings (electrogalvanized, hot dip galvanized and hot dip galvannealed).

CHAPTER 4

INVESTIGATION OF GALLING USING THE TWIST COMPRESSION TEST

In this study, the Twist Compression Test (TCT) and Finite Element (FE) analysis were used to i) evaluate the performance of galvanized coatings, lubricants and tool materials/coatings in terms of galling and powdering ii) understand the fundamental aspects of powdering and galling, and iii) determine the critical interface pressure and temperature that initiate galling or powdering in forming galvanized AHSS.

The following effects on the Coefficient of Friction (COF) and galling were considered in TCT: • contact pressure at the tool-workpiece interface • characteristics (surface roughness and hardness) of galvanized coating • lubricant viscosity • change in tool surface roughness • interface temperature between tool and workpiece

4.1. Principle of TCT

In the TCT, a rotating annular tool is pressed against a fixed sheet metal specimen while the pressure and torque are measured, Figure 4.1. The COF between the tool and the specimen is calculated by using Eq. (4.1).

Figure 4.1: Schematic of TCT

T µ = r ⋅ P ⋅ A (4.1)

Here, µ is the COF, T is the torque transmitted from the tool to the sheet metal specimen, r is the mean radius of the tool (11 mm), P is the contact pressure exerted by the tool on the sheet metal specimen, and A is the cross sectional area of the tool (220 mm2). Detailed descriptions of TCT machine are given in APPENDIX A.

4.2. Effects of Zinc Coatings on Galling

In this section, the TCT evaluation of galling for various zinc coated sheets is discussed.

4.2.1. Preparation of Sheet Specimens and Lubricants

The tool and specimen dimensions are given in Figure 4.2. Following each test, the used tool was replaced with a new tool that was polished to Ra values in the range from 0.15 to 0.2 µm.

Figure 4.2: TCT tool insert (D2 steel) and sheet specimen (Dimensions are in mm)

Figure 4.3 compares the micrographs of various zinc-coated sheet specimens and Ra-values measured by a mechanical stylus profiler before the test.

Figure 4.3: Micrographs and surface roughness of the sheet specimens before the test

To measure the hardness values of metal substrate and zinc-coatings, the Rockwell hardness test and the Vickers hardness test were conducted. In the Rockwell hardness test, three sheet samples were randomly selected and twenty 29

indentations were made on both sides of the sheet sample with the B-scale indenter at 100 kilogram force. The penetration depth for the indenter was about 130 µm. In the micro hardness test, three sheet samples were randomly selected for each sheet material and twenty indentations were made on both sides of sheet sample with the pyramid indenter at 10 gram force. While the coating thickness was about 5~ 15 µm per side, the penetration depth was obtained in the range of 1.04 ~ 4.68 µm.

Hardness DP500 GI DP600 GI AKDQ GA DDS GA

Rockwell hardness 80.06 (std.=0.60) 83.94 (std.=1.29) 44.26 (std.=0.62) 39.40 (std.=0.87) (HRB)

Vickers hardness 60.73 (std.=4.94) 52.87 (std.=5.84) 277.98 (std.=22.84) 290.29 (std.=38.45) (VHN) Note: std. = standard deviation

Table 4.1: Hardness values of metal substrates and zinc coatings

Two polymer-based wet lubricants were used in TCT. Detailed properties of lubricants are given in Table 4.2 and these lubricants contained pressure additives. The lubricants were applied on one side of the sheet specimen by using a pipette and a digital balance (1/1000 grams resolution). The applied lubricant quantity was in the range of 110 ~115 g/m2.

Properties Lub A (Product #146-492) Lub B (Product # 136-292)

Viscosity (cs) 126 562

Density (kg/m3) 1115.7 1114.5

pH 8.71 8.91

Solid (%) 57.51 13.97

Table 4.2: Properties of lubricants tested

4.2.2. Test Conditions

The test conditions are given in Table 4.3. In each experiment, the test was continued until the Coefficient Of Friction (COF) reached 0.3. Thus, COF=0.3 was used as an indicator for the start of metal-to-metal contact, based on experience in laboratory tests. The tool-workpiece contact pressures were selected based on the preliminary FE simulations of round cup drawing for DP 600.

Conditions Descriptions

Tool material D2 (with no coating)

Sheet materials DP 600 Bare (without coating) / thickness = 0.6 mm

DP 500 with GI / thickness = 0.8 mm

DP 600 with GI / thickness = 1.0 mm

DDS with GA / thickness = 0.8 mm

AKDQ with GA / thickness = 0.7 mm

Contact pressures 50, 100, 170 MPa

Testing speed 8.9 RPM (average surface speed = 10.35 mm/sec)

Testing lubes 2 wet lubricants (see Table 4.2)

Environment temp. 22 ºC (room temperature)

Table 4.3: Experimental conditions used for TCT

4.2.3. Experimental Results

Most tests were conducted up to COF 0.3 which was considered as the beginning of metal-to-metal contact. Typical Time versus COF curves are shown

in Figure 4.4. Similar data was obtained for different specimen materials and contact pressures, as shown in Table 4.3.

Figure 4.4: COF-Time curves obtained from TCT with different lubricants for DP 500 GI at 100 MPa interface pressure

After each test, the specimens and the tools were visually inspected for powdering (before cleaning with Acetone) and for galling (after cleaning with Acetone). Thus, the severity of galling and powdering was determined qualitatively (0- no galling / powdering to 3-most severe galling / powdering) as shown in Figure 4.5 and Figure 4.6.

Figure 4.5: Ranks of galling on tool samples tested

Figure 4.6: Ranks of powdering on sheet samples tested

Galling is a material transfer from the workpiece to the tool surface and usually forms a relatively strong adhesion on the tool surface. From the micrograph as shown in Figure 4.7, galled surface showed a heterogeneous material adhered on the tool surface. Powdering is the contamination of the tool- workpiece interface and it is the mixture of zinc powder, metal debris and lubricant residues. In general, the increase of powdering and galling causes a high coefficient of friction at the tool-workpiece interface.

Figure 4.7: Galling observed in the tested tool at 170 MPa interface pressure (left) and the micrograph of tool surface (right)

In our study, the performance of zinc coatings and lubricants were evaluated by ranking of galling and powdering. No galling was observed in TCT at 50 MPa, regardless of sheet materials, zinc-coatings and lubricants used in the tests. The qualitative severity of galling for two different interface pressures is compared in Table 4.4 and Table 4.5 for 100 and 170 MPa interface pressures.

Pressure Severity of Galling Sheet Material – Zinc Coating Lube (MPa) 0 0.5 1 1.5 2 2.5 3

DP600 – Bare A (No coating) B DP600 – GI A (Hot Dip Galvanized) B

100 DP500 – GI A (Hot Dip Galvanized) B DDS – GA A (Hot Dip Galvanneal) B AKDQ – GA A (Hot Dip Galvanneal) B

Table 4.4: Qualitative Evaluation of Galling at 100 MPa Interface Pressure

Pressure Severity of Galling Sheet Material – Zinc Coating Lube (MPa) 0 0.5 1 1.5 2 2.5 3

DP600 – Bare A (No coating) B DP600 – GI A (Hot Dip Galvanized) B 170 DP500 – GI A (Hot Dip Galvanized) B DDS – GA A (Hot Dip Galvanneal) B AKDQ – GA A (Hot Dip Galvanneal) B

Table 4.5: Qualitative Evaluation of Galling at 170 MPa Interface Pressure

Regarding lubricants, more galling was observed in the tests with Lub A, compared to Lub B, regardless of zinc coatings and the testing pressure. Regarding zinc coatings, DP 600 GI with Lub B showed minimum galling from the tests at both 100 and 170 Mpa, while DP 500 GI showed most severe galling with both lubricants at 170 Mpa. DP600 Bare and AKDQ GA showed minor galling at 100 Mpa, however, galling became more severe as the pressure increased to 170 MPa. DDS GA showed moderate galling at 170 Mpa and no galling at 100 MPa. The qualitative evaluation of powdering under various test conditions is given in Figure 4.8. In overall, GA coatings gave more powdering than GI coatings.

Figure 4.8: Qualitative comparison of the severity of powdering for various zinc coatings and lubricants tested at 100 and 170 MPa interface pressure

The surface of galvanized sheets was characterized by using the mechanical stylus profiler. The average surface roughness, Ra, of the sheet samples was measured before and after the test. As shown in Figure 4.9, the Ra- value decreased as the contact pressure increased to 50 Mpa, however, it increased with increasing contact pressure up to 170 Mpa. This implies that the asperities of sheet surface become flatten at 50 Mpa and as the pressure increase up to 170 MPa, the sheet surface becomes rougher because of the lateral shearing of asperities while the tool slides over the sheet surface. Larger data deviations were observed in the sheet samples tested at 170 Mpa, because the galled tool surface made severe scorings on the mating sheet surface. Taken overall in

Figure 4.9 and Figure 4.10, the sheet samples tested with Lub B gave smaller Ra values, which indicates less scoring on the sheet surface, than those tested with Lub A.

1.60

Before test 50 Mpa 1.40 100 Mpa 170 Mpa

1.20

1.00

0.80

0.60

0.40 Avg. Surface Roughness, Ra (µm) Roughness, Surface Avg.

0.20

0.00 DP600 GI DP600Bare DP500 GI DDS GA AKDQ GA

Figure 4.9: Comparison of surface roughness of sheet samples tested with Lub A

1.60 Before test 1.40 50 Mpa 100 Mpa 1.20 170 Mpa

1.00

0.80

0.60

0.40 Avg. Surface Roughness, Ra (µm) Roughness, Surface Avg. 0.20

0.00 DP600 GI DP600Bare DP500 GI DDS GA AKDQ GA

Figure 4.10: Comparison of surface roughness of sheet samples tested with Lub B

4.2.4. Effects of Galvanized Coating Characteristics on Galling Behavior

In the tests at 170 MPa, DP600 GI and DP500 GI showed different performances in reducing galling when they were tested with same Lub B. Depending on the galvanizing process conditions used in different steel companies, GI coating may have different characteristics in terms of chemical components (e.g. Zn and Al), surface finish and coating thickness. Therefore, both GI coating structures and their chemical compositions were examined by Scanning Electron Microscope (SEM) and Energy Dispersive Spectroscopy (EDS).

DP600 GI DP500 GI

Figure 4.11: SEM pictures of DP600 GI and DP500 GI

Based on the SEM analyses, DP600 GI coating shows a more uniform structure while DP500 GI coating showed high porosity as shown in Figure 4.11. From the spectral analyses with EDS, DP500 GI coating gave the relatively larger percentages of alloying elements compared to DP600 GI as shown in Table 4.6. Considering a fact that aluminized steels tend to cause more adhesion and galling than other steels, the aluminum content in the coating may influence the galling behavior.

DP600 GI DP500 GI Element Weight % Atomic % Weight % Atomic % C 0.51 2.64 0.59 2.96 O 0.54 2.12 1.53 5.73 Al 0.46 1.06 0.75 1.67 Fe 1.52 1.70 1.88 2.02 Ni 0.03 0.03 -0.01 -0.01 Zn 96.93 92.44 95.26 87.62 Total 100 100 100 100

Table 4.6: Comparison of the chemical compositions measured at both GI coatings of DP600 and DP500 sheet materials

4.2.5. Change in the Surface Topography of Tool for increasing COF

To investigate the change in tool surface topography as the COF increases, several tests were conducted up to COF=0.1, 0.2 and 0.3 individually at 170 Mpa. Figure 4.12 illustrates the COF-Time curves obtained from the tests for AKDQ GA at different COF values. Similar charts of other galvanized sheet samples were obtained from the tests. In each test, a new tool sample was used. For each tested sample, as shown in Figure 4.13-a, four different locations of tool surface were randomly selected for detailed surface analyses by using optical microscopy and Atomic Force Microscopy. The micrograph of initial tool surface is given in Figure 4.13-b.

Figure 4.12: COF-Time curves obtained from the TCT up to COF=0.1 / 0.2 / 0.3

Figure 4.13: Images of a) D2 TCT tool sample tested and selected areas for surface analysis, b) micrograph of tool surface before the test

Micrographs of Tool Surface Before/After the Test

Using the optical microscope, the surface of tool samples tested was qualitatively examined. Compared with the initial tool surface before the test Figure 4.13-b, the surface topography of tool significantly changed as the COF

increased up to 0.3, as shown in Figure 4.14. Material pick-up, white color dots as shown in Figure 4.14, started to appear on the tool surface at COF 0.1, Figure 4.14-a, and increased to be the severely galled surface at COF 0.3, Figure 4.14-c.

Testing conditions: DP500 GI and Lub B at 170 MPa

(a) Tool surface at COF 0.1

(b) Tool surface at COF 0.2

Figure 4.14: Topographical change in tool surface at different COF with DP500 GI

For the TCT with DP600 Bare at 170 Mpa, the micrographs of tool samples were compared in Figure 4.15. The material pick-up on the tool surface was not as severe as the case with DP500 GI. However, white colored tracks were generated at both inner and outer rims of tool surface, while the original tool surface was in relatively dark color.

Testing conditions: DP600 Bare and Lub B at 170 MPa

(a) Tool surface at COF 0.1

(b) Tool surface at COF 0.2

Figure 4.15: Topographical change in tool surface at different COF with DP600 Bare

Figure 4.16 shows the micrographs of tool surfaces tested against DP600 GI with Lub B at 170 MPa. Similar to the tool samples tested with DP600 Bare, Figure 4.15, the material pick-up on the tool surface was not as severe as the tool samples tested with DP500 GI.

Testing conditions: DP600 GI and Lub B at 170 MPa

(a) Tool surface at COF 0.1

(b) Tool surface at COF 0.2

Figure 4.16: Topographical change in tool surface at different COF with DP600 GI

AFM Analyses of Tool Surface Topology

The galling in sheet metal forming was studied by using various advanced surface analyzers such as Scanning Electron Microscopy (SEM), Transmission Electron Microscopy (TEM) and Auger Electron Spectroscopy (AES) [Carlsson et al. 2001A, Hogman et al. 2004, Sandberg et al. 2004 and Schedin et al. 1993]. However, all these methods can give only qualitative information on galling, similar to the micrograph from optical microscopy. In our study, the Atomic Force Microscopy (AFM) was used to examine the surface topography of tool samples quantitatively as well as qualitatively. From the AFM measurement, the surface roughness parameters such as Ra and RMS can be obtained. The mechanical stylus profiler was not able to give reliable data of surface roughness change in required resolution as COF increased. For the single measurement, the 90 µm x 90 µm area was scanned by the cantilever tip controlled by the x-y-z scanner. Figure 4.17 illustrates the major components of AFM and the principle of Dynamic Mode AFM that was used in our measurement. Basically, in AFM, a tinny tapping cantilever tip scans the topography of sample in nano-scale resolution. The 270 KHz operating frequency was used for the tapping motion of cantilever tip that has the stiffness of 42 N/m.

Figure 4.17: Components of AFM machine (left) and the principle of AFM (right)

The surface topography of tested tool samples was examined by using

AFM. As shown in Figure 4.18, the average surface roughness, Ra, tool, of tool samples increased as the COF increased. Under production conditions, galling is not easy to observe without running a large number of part. Therefore, the surface roughness parameters, Ra or RMS values, of dies were frequently used as pre-indicator for galling [Shih et al. 2004, Rooji et al. 2001A, 2001B and Schedin et al. 1993, Sato et al. 1998]. As the tool surface becomes rougher, the adhesion of workpiece material on the tool surface can be initiated, because asperities of the tool surface can plough into the relatively soft workpiece material.

Figure 4.18: Change in surface topography of tool samples as the COF increases (Measured area = 90 µm x 90 µm)

The galled surface of tool sample was examined by using AFM. Figure 4.19 illustrates the variation of surface topography of tool sample which was tested for DP500 GI with Lub B at 170 MPa. Three local areas, indicated as A, B and C in Figure 4.19, were selected for AFM measurements. As shown in Figure 4.19, the significant difference was observed in topography of the tool surface in terms of surface roughness, Ra,tool. The galled surface, indicated as window A in Figure 4.19, showed relative smooth surface, while the tool surface and the boundary between galling and tool surface showed much rougher surface (windows B and C in Figure 4.19).

Figure 4.19: Variation of Surface Topography of the tool sample tested for DP500GI with Lub B at 170 Mpa

The Ra, tool value was measured for the tool samples tested with DP600 GI and DP500 GI with Lub B at 170 Mpa.

600

Ra, initial tool

500 Ra, tool (COF=0.1) Ra, tool (COF=0.2) (nm) a 400 Ra,tool (COF=0.3)

300 Galling 3

200 Galling 0 Avg. Surface Roughness, R Roughness, Surface Avg. 100

0 Tool used for DP600GI Tool used for DP500GI

Figure 4.20: Variation of Ra-value of tool samples tested up to different COF

As shown in Figure 4.20, as the COF increases, the tool surface became rougher in the case of severe galling (Galling 3), while the tool surface roughness was slightly lower in the case of negligible galling (Galling 0).

4.3. Effects of Lubricants on Galling

TCT was used to evaluate the performance of lubricants in terms of galling and coefficient of friction. In this section, TCT results of various lubricants are discussed.

4.3.1. Characterization of Sheet Coatings and Microstructures

Two different zinc coatings, GA and GI, on the similar grade of AHSS (DP590 and DP600) were selected to investigate of the effect of zinc coating on galling behavior. In addition, two different microstructures of AHSS, Dual Phase (DP) and Transformation-Induced Plasticity (TRIP), were tested to investigate the effect of microstructure of sheet material on galling behavior.

The surface topography of both GA and GI coatings was taken by using the SEM. As shown in Figure 4.21, GA coating has more porous structure while GI coating shows relatively fine surface. From the EDS analyses, the chemical composition of these coatings was analyzed as shown in Table 4.7. GA coating was found to contain relatively higher alloying elementary compositions such as Fe, Al, Mn, etc., while GI coating showed 97 percentages of Zn element. This chemical difference is caused by a fact that GA coating is made through the annealing process that makes several transition layers from steel substrate to zinc layer.

Figure 4.21: SEM images of GA and GI coatings

DP590 GA DP600 GI Element Weight % Atomic % Weight % Atomic % C 1.02 4.81 0.51 2.96 O 2.42 8.56 0.54 5.73 Al 0.93 1.96 0.46 1.67 Ni 0.03 0.03 Mn 0.25 0.25 Fe 13.58 13.74 1.52 1.70 Zn 81.79 70.68 96.93 92.44 Total 100 100 100 100

Table 4.7: EDS results of chemical components for DP590 GA and DP600 GI

The surface roughness of GA and GI specimens was measured by using a

mechanical stylus profiler. Table 4.8 compares the average surface roughness, Ra, and the maximum peak-to-valley height, Rmax. GA specimens showed three to

four times larger Ra values compared to GI specimens

The micro-hardness of coating and a conventional hardness of specimen were measured by using the Vickers hardness tester and Rockwell hardness tester, respectively. As shown in Table 4.8, in comparison of micro-hardness of

coating structures, GA coating showed more than five times larger Vickers Hardness Number (VHN) compared to GI coating. In comparison of hardness of steel substrates, TRIP780 showed about 8 percentage higher HRB number compared to other materials.

Surface Roughness Parameters DP590 GA DP600 GI TRIP 780 GA

Ra (µm) 1.23 ± 0.12 0.33 ± 0.04 1.04 ± 0.15

Rmax (µm) 5.73 ± 1.07 1.28 ± 0.75 4.65 ± 1.06

Vickers Hardness (VHN) 316.14 ± 36.6 59.75 ± 7.3 334.32 ± 69.8

Rockwell Hardness (HRB) 90.17 ± 0.4 89.43 ± 0.5 97.03 ± 0.8

Table 4.8: Surface roughness and hardness of AHSS specimens

The microstructure of steel substrates was analyzed by using a standard metallurgical procedure (i.e. a series of different polishing steps and chemically etching with 2% nital etching). As shown in Figure 4.22, dual phase steel consists of a ferritic matrix containing a hard martensitic second phase in the form of islands. As the volume fraction of martensite increases, the strength of steel normally increases.

The microstructure of TRIP steel has retained austenite (a minimum of 5 volume percent) embedded in a primary matrix of ferrite. It contains hard phases such as martensite and bainite in varying amounts. During plastic deformation, the dispersion of hard second phases in soft ferrite causes a high work hardening rate, as observed in the DP steels. However, in TRIP steels the retained austenite also progressively transforms to martensite with increasing strain, thereby increasing the work hardening rate at higher strain levels. The TRIP steel has a lower initial work hardening rate than the DP steel, but the hardening rate persists at higher strains where work hardening of the DP begins to reduce. 50

Figure 4.22: Microstructure of Dual Phase and TRIP steels

4.3.2. Experimental Conditions

The test was focused on evaluating the performance of lubricants mating with GA and GI coated sheets. To isolate the effect of tool material on test results, only D2 tool steel was used as the tool material. As shown in Table 4.9, in our test, ten different wet and dry film lubricants were selected based on recommendations of automotive stamping companies, steel producers and lubricant suppliers. The two testing contact pressures, 50 and 170 MPa, were determined by preliminary FE simulation of cup drawing for DP600. Detailed test matrix is given in Table 4.9. All the test conditions were repeated in three times with new samples.

Testing parameters Descriptions

Sheet materials DP 590 GA, 1.24 mm (initial thickness)

DP 600 GI, 1.0 mm (initial thickness)

Tool materials D2 (uncoated and polished)

Testing Pressures 50 and 170 Mpa

Lubricants Lub A: Polymer based lube with EP additives (IRMCO, US)

(total 10 lubes) Lub B: Polymer based lube with EP additives (IRMCO, US)

Lub C: Water-soluble dry film lubricant (Castrol, UK)

Lub D: Water-soluble dry film lubricant (Quaker Chem., US)

Lub E: Water-free dry film lubricant (Zeller & Gmelin, Germany)

Lub F: Chlorinated soluble oil (Fuchs, US)

Lub L: Water-soluble dry film lubricant (Quaker Chem. US)

Lub M: Synthetic water soluble lubricant (Metalworking Lubricants, US)

Lub N: Straight oil (Quaker Chem., US)

Lub P: Straight oil (Zeller & Gmelin, Germany)

Table 4.9: Test matrix to evaluate various stamping lubricants using the TCT

4.3.3. Experimental Results

Galling behavior at low contact pressure

At 50 MPa contact pressure, most lubricants performs very well for both GA and GI coatings, as shown in Table 4.10. Especially, dry film lubricants, Lub

C, D, E and L, showed the steady-state response of COF after 120 sec, thus the test stopped in 180 sec. Lub F (Cholorinated soluble oil) gave a relatively severe galling with GA coating. Emulsion type lubricants, Lub F and Lub M showed difficulties to stay on GI coated sheet, thus audible unpleasant sounds occurred during the test and the test stopped. Since both Lub F and Lub M are water soluble lubricants, GI coating showed “hydrophobic” characteristic which is

repellant to water. The average roughness, Ra, of sheet surface was measured after the test. As the galling increases, the sheet surface is more severely scratched (i.e. scoring), thus the sheet surface roughness can be an indicator of galling. Figure 4.23 compares the surface roughness of GI and GA sheets tested with various lubricants at 50 MPa contact pressure. As shown in Figure 4.23, the

poor lubricants tend to increase the Ra-values significantly from the initial Ra values, while good lubricants reduce the Ra values.

Galling Behavior at High Contact Pressure

At 170 MPa contact pressure as shown in Table 4.11, several dry film lubricants (Lubes C, D and E), a synthetic lubricant (Lub M) and two straight oils (Lubes N and P) showed smaller galling than other lubricants for GA coating, while most lubricants, except two dry film lubricants (C and L), failed for GI coating due to galling and an unpleasant noise out. Similar to the test at 50 MPa, some lubricants showed the steady-state response of COF after 120 sec, thus the test duration cut off at 180 sec. Figure 4.24 compares the surface roughness of GI and GA sheets tested with various lubricants at 170 MPa contact pressure.

Similar to 50 MPa test results, good lubricants reduced Ra-values and poor

lubricants increased Ra-values. As shown in Figure 4.24, some of lubricants showed high deviations in the measurement error bar because of severe scoring occurred at high pressure test. 53

Pressure / Sheet Material / Lube Galling Powdering Avg. Time to COF 0.3 COF @ 180 sec Speed Tool Material A 1 2 105 B 0 1 108.5 C 0 0 0.12 - 0.13 D 0.16 0.16 0.15 DP590 GA / E 0 0 0.24 - 0.27 D2 Tool steel (uncoated) F 1.75 0.75 84.5 L 0.3 0.3 0.17 - 0.25 M 0.6 1.4 134.6 N 0 0 51.3 50 MPa / P 0 0 131 9 RPM A 0 0 63 54 B 0 0 70 C 0 0 0.15 D 0 0 40 DP600 GI / E 0 0 71.2 D2 Tool steel F 0.5 0 25 (Noise stop) (uncoated) L 0.3 0 0.2 M 1 0 33 (Noise stop)

N 0 0 65 P 0 0 110

Table 4.10: TCT results for various lubricants with a D2 tool material at 50 MPa interface pressure

Sheet Material / Interface Lube Galling Avg. Time to COF 0.3 COF @ 180 sec Pressure / Speed Tool Material Contamination A 2 1.78 55 B 1.3 1 37 C 1 1 0.15 D 0 0.5 0.08 DP590 GA / E 0 0 0.07 D2 Tool steel (Uncoated) F 2 1.16 92 L 2 2 92 M 0.75 0 0.15 N 0.25 0 67 170 MPa / P 0.16 0 170 9 RPM A 2 1.5 54 55 B 1.7 1.5 89 C 0.83 1 0.09 D 1 0.5 23 (Noise stop)

DP600 GI / E 1.75 1.5 70 (Noise stop) D2 Tool steel (Uncoated) F 2.66 1.66 12 L 1.16 0.5 0.13 M 2 1 40 (Noise stop)

N 2 2 11 (Noise stop)

P 2 0.66 37 (Noise stop)

Table 4.11: TCT results for various lubricants with a D2 tool material at 170 MPa interface pressure

Figure 4.23: Surface roughness of GI/GA coated sheets tested at 50 MPa

Figure 4.24: Surface roughness of GI/GA coated sheets tested at 170 MPa

4.3.4. Discussions and Conclusions

The following conclusions can be drawn from this test results: • At 50 MPa contact pressure, most lubricants showed negligible galling for both GI and GA coated AHSS. • At 50 MPa, DP590 GA showed more powdering than DP600 GI with the most of lubricants. • Lubricants F and M (water diluted lubes) did not stay on the surface of DP600 GI, thus a poor frictional response was observed with an audible unpleasant sound during the test. • Most dry film lubricants (C, D, E and L) gave a lower COF for GA coating than wet lubes at both 50 and 170 MPa contact pressures. • However, for GI coating, Lubes C and L gave a lower COF than Lubes D and E. • At 170 MPa contact pressure, the lubricants with GA coating showed less galling and interface contamination compared to the lubricants with GI coating. • In the test for GA coating at 170 MPa, most lubricants perform well, except lubricants A, F and L. • In the test for GI coating at 170 MPa, only lubricants B, C and L showed less galling in relatively longer sliding length, while other lubricants failed by generating galling with an audible unpleasant sound.

4.4. Effect of Tool Material and Surface Treatments on Galling

In this section, TCT results of various die materials with surface treatments (PVD coating or Plasma Ion Nitriding) are discussed.

4.4.1. Characterization of Tool Materials and PVD coatings

Tool materials were characterized by measuring surface roughness, Ra, and Rockwell hardness value (HRC). The hardness of PVD coatings were

measured by the micro hardness tester. Detailed results of these characterizations are given in Figure 4.25.

D2 Tool Steel Cast Iron

BUE

Vancron 40 (Powder Mat.) K340-PIN

Carmo – PIN / PVD K340-PVD

Figure 4.25: Micrographs of tool surface before the test

Cast iron material showed porous structures and the Ra value was relatively higher than other tool materials as shown in Figure 4.25. All the uncoated tool materials (e.g. D2, K340-PIN and Vancron 40) showed different surface topography while both PVD coated tool materials (e.g. Carmo-PIN/PVD and K340-PVD) showed similarly a white color embossed surface. Regarding the Rockwell hardness of tool substrate, Cast Iron showed the minimum hardness value, 41 HRC, while other tool materials gave the hardness value in a range of 56 ~63 HRC. As shown in Figure 4.25, the Carmo TCT tool was case hardened by using Plasma Ion Nitriding (PIN) and Physical Vapor Deposition (PVD) techniques. Depending on the case hardening technique, the hardness of surface layer varies as shown in Figure 4.25. This surface hardness is an important role to protect the tool surface from galling and chipping on the tool surface. Carmo- PIN/PVD gave the highest VHN compared to other tool materials, because it had an effect of superposition of PIN and PVD on the surface layer.

4.4.2. Experimental Conditions

The conditions for TCT were selected to evaluate the performance of tool materials and tool coatings with GA and GI sheets in terms of galling. In this test, several pre-screened lubricants from previous TCT were used to test various tool materials. Detailed test matrix is given in Table 4.12. All the test conditions were repeated in three times with new samples. The contact pressure and the testing speed were fixed as 170 MPa and 9 RPM, individually.

Testing parameters Descriptions

Sheet materials DP 590 GA, 1.24 mm (initial thickness)

DP 600 GI, 1.0 mm (initial thickness)

TRIP 780 GA, 1.0 mm (initial thickness)

Tool materials D2 (cold work tool steel) uncoated and polished

(total 6 matls.) K340 (cold work tool steel) with PVD (CrN)

K340 (cold work tool steel) with PIN

Carmo (cast steel) with PIN + PVD (CrN)

Vancron 40 (Powder Material) uncoated

A536 Grade 80-55-06 (Spheroidal Graphite Cast Iron)

Contact Pressure 170 MPa

Tested lubes Pre-screened lubes from the previous TCT

Table 4.12: Test matrix to evaluate various tool materials and surface treatments using the TCT

4.4.3. Experimental Results

Galling behavior for non-case hardened tool materials with GA sheets

With DP590 GA and TRIP780 GA sheets, various lubricants were tested with uncoated and non-case hardened tool materials (e.g. D2, Cast Iron and Vancron40), as shown in Table 4.13. The severity of galling and interface contamination, a mixture of powdering, metal debris and lubricant residue, were ranked in Table 4.13. In addition, Table 4.13 compares the average time recorded

until the COF reached to 0.3 and the steady state value of COF at 180 sec for several dry film lubricants. In general, for GA coated sheets, straight oil lubricants (Lubes N and P) and dry film lubricants (Lubes C, D and E) gave better effectiveness in reducing galling and interface contamination than other polymer based lubricants (Lub A and B) or water soluble lubricants (Lub F and M). Also, there was no distinguishable influence of hardness between DP590 GA and TRIP780 GA on galling and interface contamination. This implies that, in TCT, the effect of the strength of sheet material is difficult to be considered in the evaluation of galling, because there is negligible plastic deformation of sheet in TCT.

Galling behavior for case hardened tool materials with GA sheets

Three case hardened tool materials (e.g. K340-PVD, Carmo-PIN/PVD and K340-PIN) were tested with various lubricants and GA sheet (DP590 GA). As shown in Table 4.14, K340-PIN showed similar trend in the preference for the type of lubricant with other uncoated tool materials, e.g. D2, Vancron 40 and Cast Iron. However, PVD coated tool materials, K340-PVD and Carmo- PIN/PVD, gave very poor frictional response for the most of lubricants, except dry film lubricant, Lub C. As shown in Table 4.14, both PVD coated tools showed that the COF reached to 0.3 in a time range of 3 sec to 20 sec, while the other tool materials gave longer than 30 sec up to COF 0.3. This implies that hard coating of CrN may be adverse effects on lubrication. This phenomenon was also supported by other researchers [Klocke et al. 2002]. The hard coatings on tools can increase friction. From a micro-elasto-hydynamic point of view, the higher Young’s modulus of the coating material results in a lower amount of remaining microscopic lubricated area. This is caused by the fact that the liquid lubricant is more easily squeezed out of the contact, or that the remaining micro-lubricated

areas have a smaller film thickness under given load and sliding velocity. In the backward extrusion test, TiHfCrN coated tool gave about 20% higher maximum punch force than an uncoated tool [Klocke et al. 2002]. This hypothesis was also supported in the TCT result for Carmo-PIN with TRIP 780GA as shown in Table 4.15. This uncoated cast steel, Carmo-PIN, followed the tendency of other uncoated tool materials in terms of frictional response and lubricant preferences, as found in Table 4.14 and Table 4.15.

Sheet material / Interface Tool material Lube Galling Avg. Time to COF 0.3 COF @ 180 sec Pressure and Speed Contamination A 3 3 81.5 B 2 0 40 C 1 1 0.15 D2 Tool Steel D 1.5 1 113 (Uncoated) E 1 1 175 M 2 2 70 N 0 0 48 P 0.5 0.5 73 A 2.5 3 110 B 1.75 1.25 57.6 C 1.5 1.25 0.13 Dual Phase 590 GA D 1 1 162 / Cast Iron E 0.5 0.5 0.1

63 170 MPa and 9 RPM M 2 2 173 N 1.25 0 0.15 P 0.5 0.5 0.14 A 3 3 77 B 1.25 0.25 51 C 0.75 0.25 0.22 Vancron 40 D 1.5 1 37 (PM) E 1 1 133 M 1.75 1 117 N 0.25 0.5 71 P 0.25 0 86.4

Table 4.13: TCT results for various lubricants with uncoated tool materials and GA coated DP steel at 170 MPa

Sheet material / Interface Tool material Lube Galling Avg. Time to COF 0.3 COF @ 180 sec Pressure and Speed Contamination A 2 2 8 B 0.75 0 8.25 C 1 2 112 Carmo- D 0.5 0 4 PIN/ PVD E 0.5 1 18 M 0.75 0.5 3.3 N 0 0 10.5 P 0 0 4.2 A 2.5 2 16 B 0.5 0.5 11 C 2 1 117 Dual Phase 590 GA D 0 0 3 / K340-PVD E 0 0 19 170 MPa and 9 RPM

64 M 0.5 1 2.7 N 0 0 11 P 0.25 0 3.1 A 3 3 65 B 2 0.3 44.6 C 1 0.5 162.5 D 1 1 106 K340-PIN E 1 1 0.22 M 1.5 1 27 N 0.25 0 37.5 P 0.25 0 32

Table 4.14: TCT results for various lubricants with surface treated tool materials and GA coated DP steel at 170 MPa

Sheet material / Interface Tool material Lube Galling Avg. Time to COF 0.3 COF @ 180 sec Pressure and Speed Contamination B 2.5 1.5 48 D2 N 0 0 41.5 P 0 1 74.5 B 1.5 1.5 60 Cast Iron N 0.5 0 0.12

P 0.5 0 0.1

B 1.5 2 65 Vancron 40 N 0 2 57 TRIP 780 GA P 0 0.5 84

65 / 170 MPa and 9 RPM B 1.5 1.5 64 Carmo-PIN N 0.25 0 57.5 (No PVD coating) P 0 1 127.5 B 0.5 1 15 K340-PVD N 0 0 18 P 0 0 16 B 1.75 1 55 K340-PIN N 0 0 49 P 0.25 0 108.5

Table 4.15: TCT results for various lubricants with various tool materials and GA coated TRIP steel at 170 MPa

Galling behavior for various tool materials with GI sheets

Various tool materials were tested with DP600 GI. Three pre-screened lubricants, Lubes B, C and M were used for the test. A polymer based lubricant, Lub B, performed well with the most of uncoated tool materials, D2, Vancron 40 and K30-PIN, as shown in Table 4.16. However, PVD coated tool materials gave poor frictional response within 20 sec, which was similar in TCT results of PVD coated tools with GA coated sheets. To investigate the effects of zinc coating (GA and GI) and contact pressure on COF, DP600 bare (no zinc coating) was tested with PVD coated tools, Carmo-PIN/PVD and K340-PVD, at 50 and 170 MPa contact pressures. As shown in Figure 4.26, similar to previous test results of PVD coated tools with GA and GI sheets, a fast increase of COF was found also in the test with bare sheets at high contact pressure, 170 MPa. However, as the pressure decreased to 50 MPa, the frictional response (i.e. duration up to COF=0.3) improved significantly as shown in Figure 4.26. Therefore, it can be drawn that the frictional response of PVD coating is mainly affected by the contact pressure, not by zinc coating.

Sheet material / Interface Pressure and Speed Tool material Lube Galling Avg. Time to COF 0.3 COF @ 180 sec Contamination

D2 B 1.7 1.36 89

B 0.5 1.5 54.3

Cast Iron C 1 1 0.15

M 2 2 27

B 1 1 43

Vancron 40 C 2 1 0.15

Dual Phase 600 GI M 1 1 16 / 170 MPa and 9 RPM B 0.25 2.25 8.23 67 Carmo- M 0 1 4.4 PIN/ PVD C 1 2 87

K340-PVD C 0 1 56

B 0.83 1.67 48.5

K340-PIN C 1 1 151

M 1 1 11

Table 4.16: TCT results for various lubricants with various tool materials and GI coated DP steel at 170 MPa

Figure 4.26: TCT results of PVD coated tools with DP600 Bare (uncoated) and Lub B at 50 and 170 MPa Contact Pressures

Surface roughness of GA and GI sheets after the test

Figure 4.27 and Figure 4.28 compare the surface roughness, Ra, of GI and GA sheets tested with various tool materials and pre-screened lubricants at 170

MPa interface pressure. Ra-value was used as an indicator of galling, because severe galling increased scoring of sheet surface. As shown in Figure 4.27, the

poor lubricants tend to increase the Ra-values significantly from the initial Ra values, while good lubricants reduce the Ra values, which indicates relatively smooth surface.

D2 Caldie K340-PVD K340-PIN Cast Iron Vancron 40 3.0

2.5

2.0

1.5

Initial Ra of DP590 GA = 1.33 µm 1.0

0.5 Avg, Surface Roughness, Ra Surface Avg, (um)

0.0 Lub A Lub B Lub C Lub D Lub E Lub M Lub N Lub P Lubricant

Figure 4.27: Surface roughness of GA coated sheets tested with various tool materials and tool coatings

1.40 D2 Caldie K340-PIN 1.20 Cast Iron Vancron 40 1.00 Large Scoring

0.80

0.60

0.40 Small Scoring

0.20 Avg, Surface Roughness, Ra Surface (um) Avg,

0.00 Lub B Lub C Lub M

Lubricant

Figure 4.28: Surface roughness of GI coated sheets tested with various tool materials and tool coatings

4.4.4. SEM / EDS Analyses of Galling

Galling on the tool surface was qualitatively examined by using SEM. As shown in Figure 4.29, the galled surface of TCT tools tested with GI (a and b) and GA coatings (c and d) shows remarkable differences. Galling induced from GI sheet (a and b) shows a ductile structure, while galling induced from GA sheet shows a brittle structure. This difference can be explained by the characteristics of GI and GA coatings. GI coating has a more ductile structure while GA coating has more hard and brittle structures.

(a) (c)

(b) (d)

Figure 4.29: SEM photographs of the galled surface on D2 TCT tool (a) 2 mm resolution and (b) 50 µm in testing with GI coated sheet, (c) 2 mm resolution and (d) 50 µm in testing with GA coated sheet

In order to identify the source of galled material (either from zinc coating or steel substrate), spectrum analyses were conducted to analyze the chemical compositions of galling by using EDS, as shown in Table 4.17.

The chemical composition of GI-induced galling, corresponding to (a) and (b) in Table 4.17, showed 98% pure zinc and this was found to be very close to the original chemical composition, 97% Zn, of original GI coating as shown in Table 4.7. Also, for GA-induced galling as shown in (c) and (d) in Table 4.17, the chemical composition of galling (e.g. 80.32% Zn and 14.94% Fe) was found to be also close to the original composition of GA coating, e.g. 82% Zn, 13.58% Fe, etc.

(a) (c)

Element Weight % Atomic % Element Weight% Atomic%

C 0.46 2.34 C 0.60 2.77 O 4.13 14.31 (b) O 1.45 5.56 (d) Fe 14.94 14.83 Zn 98.09 92.10 Zn 80.32 68.09 Totals 100.00 Totals 100.00

Table 4.17: EDS analysis results (a) spectrum results, (b) chemical composition for GI-induced galling, and (c) spectrum results, (d) chemical composition for GA-induced galling

This EDS result leads an interesting finding that galling is a material transferred from the zinc coating, GI or GA, not from a steel substrate. This conclusion is also supported by different severities of galling between zinc-

coated sheet and bare sheet in TCT, as shown in Figure 4.14 and Figure 4.15, respectively. The tool surface tested with bare sheet (uncoated) shows a negligible galling while GI sheet caused severe galling. Therefore, the characteristic of zinc coating is one of important factors to determine the galling behavior.

4.4.5. Discussions and Conclusions

The following conclusions can be drawn from this test:

Effect of lubricant type • In TCT for GA coated sheet with uncoated tools (D2, Cast iron, Vancron 40 and K340-PIN) at 170 MPa, the straight oil lubricants (Lub N and Lub P) and dry film lubricants (Lub C, Lub D and Lub E) showed better effectiveness to reduce galling and interface contamination. • In TCT for GI coated sheet with uncoated tools (D2, Cast iron, Vancron 40 and K340-PIN) at 170 MPa, a polymer film lubricant (Lub B) and a dry film lubricant (Lub C) showed better effectiveness to reduce galling and interface contamination.

Effect of tool coating • PVD coated tools (Carmo-PIN/PVD and K340-PVD) showed poor frictional response regardless of sheet coatings and lubricants at 170 MPa contact pressure. • However, at lower pressure 50 MPa, the COF of these coated tools significantly improved in terms of the sliding distance up to COF = 0.3.

Effect of zinc coating • GA coated sheets slightly decreased galling and interface contamination compared to GI coated sheets.

• There was no significant differences in galling behaviors between DP590 GA and TRIP780 GA materials. • From EDS analyses, it was found that galling was mainly caused by zinc-coating layers (GI and GA), not from a steel substrate.

Effect of thermal hardening • Thermal through hardened tool (Vancron 40 and K340-PIN) showed slightly better performance for reducing galling compared to a ductile tool (Cast iron). • However, case hardened tool (K340-PIN) did not show distinguishable reduction of galling compared to non-case hardened tool (Vancron 40 and D2).

4.5. Effect of Interface Temperature on Galling

4.5.1. Temperature Measurement During the Test

To determine the effect of interface temperature upon lubricant effectiveness and galling, the temperature near the tool-workpiece interface was measured by using the thermocouple in the test at high contact pressure, 170 MPa. Details of measurement method are illustrated in Figure 4.30. While a thermocouple measured the temperature near the tool-sheet interface, the other thermocouple simultaneously measured the ambient temperature which was about 22 °C.

Figure 4.30: Temperature measurement using a thermocouple

A dummy sheet of 1 mm thickness was made to have a slot for the thermocouple. Thus, the temperature was measured at the bottom surface of the sheet specimen during the test, Figure 4.30. The specimen and the dummy sheet were held in position with two fixture wings, as shown in Figure 4.30. Temperatures and the COF were recorded during the test.

Figure 4.31 shows the temperature-time curves obtained from the TCT for DP500 GI with Lub A and B. As shown in Figure 4.31, the increase of temperature primarily depends on the COF-Time curve thus, Lub A gave the faster increase in temperature than Lub B. Since the plastic deformation is negligible in TCT, the temperature increase was caused by only frictional heating.

Figure 4.31: Variations of temperature near the tool-sheet interface measured during the test for DP500 GI with Lubes A and B at 170 MPa

Figure 4.32: Variations of temperature near the tool-sheet interface measured during the test for DP600 Bare with Lubes A and B at 170 MPa

In a similar manner, Figure 4.32 and Figure 4.33 compare the temperature- time curves obtained from the TCT for DP600 Bare and DP600 GI, respectively.

As shown in Figure 4.32 and Figure 4.33, DP600 GI gave slightly lower temperature increase than DP600 Bare. This resulted from a lower COF-time curve in TCT with DP600 GI.

Figure 4.33: Variations of temperature near the tool-sheet interface measured during the test for DP600 GI with Lubes A and B at 170 MPa

4.5.2. Finite Element Analyses for TCT

In the experiment, it is difficult to measure the temperature at the tool- sheet interface. Therefore, FE simulation of TCT was conducted to estimate the interface temperature using the commercial FEM software DEFORM 2-D. The FE model of TCT and the input data of COF-time curve are shown in Figure 4.34. The flow stress of the sheet materials (AKDQ and DP600) was obtained using the viscous pressure bulge test [Gutscher et al. 2004]. The COF-time input data was obtained from a TCT at 170 MPa interface pressure for lubricant A, as shown in Figure 4.34.

Figure 4.34: FE model for TCT (left) and COF-time curve input data(right)

Material input data and thermal properties used in FE simulation are summarized in Table 4.18.

Parameters /properties Workpiece Tool Workpiece holder

Material / Flow stress, σ, AKDQ / 800 ε 0.2 D2 Tool Steel P20 Tool Steel (MPa) & DP600 / 1028 ε 0.143

Object type Plastic Rigid Rigid

Angular speed of tool 8.9 RPM (=0.932 rad/sec) Contact pressure 170 MPa

Heat transfer coefficient 11 KW / m2-K

Coefficient of Friction Experimental data

Thermal conductivity 60.5 W / m-K 50.71 W / m-K 24.57 W / m-K Heat Capacity 3.41 J / m3-K 3.81 J / m3-K 2.78 J / m3-K Emissivity 0.95 0.7 0.7

Initial temperature (°C) 29 23 29

Table 4.18: Input data used in FE simulations

The temperatures were estimated at various locations in the tool and the sheet specimen, as shown in Figure 4.35.

Figure 4.35: Temperature distributions in the tool and sheet specimen at selected points for tracking temperature generation with FEA

In particular, the third point (Pt. 3) may correspond to the thermocouple location for temperature measurement. The predicted and measured temperatures are compared for a selected case in Figure 4.36.

80 FEA (Pt.1) FEA (Pt.2) 70 FEA (Pt.3)

Exp (Bottom Sheet) Exp (Bottom Sheet) 60 Exp (Ref Temp) FEA (Pt.1) 50 FEA (Pt.2) FEA (Pt.3) 40

30 Exp (Ref Temp) Temperature (Deg. C.) 20

0 0 20 40 60 80 100 120 140 Time (Sec)

Figure 4.36: Comparison of FE results and experiment for DP600 Bare with Lubricant A at 170 MPa Interface Pressure

The pressure distribution at the tool-specimen interface was also calculated by FE simulation of an example case, Figure 4.37. The COF-time input data for this example case was obtained from TCT of AKDQ GA with Lub B at 170 MPa. As expected the contact pressure shows a sharp increase at the edge of the annular tool. This pressure variation may cause a leaking of the lubricant that may result in very large local friction and lubricant break down.

Figure 4.37: FE prediction of pressure distribution at the tool-workpiece interface

CHAPTER 5

EVALUATION OF LUBRICANTS USING THE DEEP DRAWING AND IRONING

TESTS

In this part of the study, the deep drawing and ironing tests were used to evaluate the performance of various stamping lubricants in forming of DP590 GA round cup samples. All the lubricants tested are being used or considered for stamping productions.

In previous TCT, various stamping lubricants were evaluated. However, it is difficult to conclude the performance of lubricants with only the TCT results, because there is negligible plastic deformation in TCT. Therefore, to compensate TCT results, deep drawing and ironing tests were conducted to evaluate the performance of lubricants under near production conditions.

In deep drawing and ironing tests, the performance of lubricants was ranked via our evaluation criteria of force and geometry indicators. Deep drawing tests were conducted at two different blank holder forces, 30 and 70 tons, at a constant ram speed of 70 mm/sec. The ironing test was also conducted to further evaluate the performance of lubricants tested by deep drawing test, because the ironing test can provide higher pressure and temperature at the tool- workpiece interface than the deep drawing test. Most ironing tests were conducted at a constant ram speed without a blank holder force. There was no severe galling observed in both tests. From all the test results, polymer-based

thin film lubricants with pressure additives (e.g. Lubes A and B) showed better effectiveness than other lubricants in terms of force (e.g. maximum punch force and applicable BHF without cup fractures) and geometry indicators (e.g. draw-in length, flange perimeter and sidewall thinning). In deep drawing test, the perimeter or draw-in length of flange can be used as another friction indicator, instead of sidewall thinning distribution in the drawn cup. The smaller the flange perimeter, the better is the performance of the lubricant. In ironing test, the side wall thinning of ironed cup samples was compared. The smaller thinning distribution also indicates the better lubricant.

FE simulations were conducted to obtain better interpretations of the test results. The pressure distribution at the tool-workpiece interface was calculated and the temperature distributions in both workpiece and the die were predicted. From FE results, the maximum pressure and temperature were predicted to take place at the die corner radius and the corresponding contact area of deformed sheet. The pressure at this area was predicted to be larger than those in the straight die surface or straight flange area. This implies that the frictional condition at the die-sheet interface may change because of the large variation in pressure.

5.1. Principle of Deep Drawing and Ironing Tests

The deep drawing test was successfully used for evaluation of lubricants by various European automotive manufactures [Girschewski et al. 2000, Meiler et al. 2004, Tolazzi 2005 and Wagner 2002]. In deep drawing, the most severe friction usually takes place at the flange area as shown in Figure 5.1. The lubrication condition in the flange area influences (a) the thinning or possibly failure of the side wall in the drawn cup and (b) the draw-in length, Ld, in the

flange (please refer Figure 5.1). As the blank holder pressure, Pb, increases, the frictional stress, τ, also increases based on Coulomb’s law, as shown in Eq. (5.1). Therefore, lubricants can be evaluated in deep drawing by determining the maximum applicable blank holder force without fracture in the cup wall.

Figure 5.1: Schematic of deep drawing test and tool dimensions

τ =⋅µ Pb where τ = the frictional shear stress µ = the coefficient of friction P = the blank holder pressure b (5.1)

In using the deep drawing test, qualitative and quantitative analyses can be made to determine the effectiveness of lubricants, based on the following criteria: • The maximum punch force (the lower the force, the better the lubricant) • The maximum applicable blank holder force, BHF (the higher BHF that is applied without causing fracture in the drawn cup, the better is the lubricant) • Visual inspection of galling and zinc-powdering

• Measurement of draw-in length, Ld, in the flange (the larger the draw-in length, the better the lubrication)

• Measurement of the perimeter in flange area (the smaller the perimeter, the better the lubrication)

The deep drawing tests were conducted under process conditions that are present in practical stamping operations. Major emphasis is put on emulating: a) the sheet-die interface pressure levels that are similar to those occurring in production by adjusting BHF and b) punch speeds that are similar to those found in mechanical stamping presses.

In the ironing test, initially a round cup is deep drawn from a circular blank and later ironed. As shown in Figure 5.2, the sheet-die interface is subjected to higher contact pressure. Furthermore, the ironing die can be heated to the range of temperatures that exist in production.

Figure 5.2: Schematic of ironing test

The lubricants used in the ironing test are evaluated based on the following criteria:

• The maximum ironing load attained (the lower the load, the better the lubricant)

• Visual inspection of galling (good lubricant has less galling through die contact) • Surface topography (roughness and microscopic topology) of the ironed cup after test

5.2. Experimental Setup

5.2.1. Description of the Tooling

The deep drawing tooling is located in a 160-ton hydraulic press that has a maximum ram speed of 300 mm/sec, as shown in Figure 5.3. A draw die attached to the upper ram moves down and forms a cup sample over a stationary punch. The constant blank holder force is set to apply by the CNC-controlled hydraulic cushion pins. During the test, the punch force is measured by a load cell located at the bottom of punch and the displacement of die is recorded by a laser sensor.

Figure 5.3: Deep Drawing Tooling at ERC/NSM

In deep drawing test, the draw ratio (i.e. a ratio of blank diameter and punch diameter) was selected to be 2.0. The drawing depth was selected to be about 80 mm to have some flange area for measuring the draw-in length and the perimeter of the flange. Round blanks (304.8 mm diameter and 1.24 mm thick) were cut from Dual Phase (DP) 590 Galvannealed (GA) steel sheet. Detailed geometry of deep drawing tooling is available in APPENDIX B. The ironing tooling was a modified deep drawing tooling with an ironing ring insert above the draw die. Detailed geometry of ironing tooling is available in APPENDIX C.

5.2.2. Test Procedures

Detailed test procedures are given in Figure 5.4. After deep drawing, the cup geometry drawn to the height of 77 mm was trimmed to a height of 50 mm by machining. Analysis of the punch force-stroke curves and visual inspection of die and the tested cup surface were conducted after deep drawing and ironing stages.

Figure 5.4: The schematic of deep drawing and ironing test procedure

5.2.3. Lubricant Selection and Application Method

Six stamping lubricants were supplied by different lubricant manufacturers and were tested by deep drawing and ironing tests. To obtain the uniform coating weight of lubricant on sheet samples, the weight of sample was measured by a digital balance (1/1000 grams resolution) before and after

applying lubricants by using a pipette and a sponge roller. Detailed properties of lubricants are given in Table 5.1.

Properties Lub A Lub B Lub F Lub M Lub N Lub P

Viscosity (cs) 26 ~ 562

Density (kg/m3) 880 ~ 1152

Coating weight (g/m2) 11.5 ~ 13.7

Table 5.1: Properties of lubricants tested

5.2.4. Characterization of Sheet and Tool Surfaces

The surface roughness of the initial sheet blank and die surface was

measured by using a mechanical stylus profiler. The arithmetic average value, Ra, was mainly used for characterizing the surface roughness. Multiple measurements were conducted at different locations and directions. The span of a single measurement was set to be 4 mm. The Ra value of sheet surface was

found to be 1.04 µm and a drawing die was properly polished to obtain the Ra value of 0.1 µm before the test. In addition, optical micrographs of the sheet surfaces (DP590 GA) before the test were made as shown in Figure 5.5.

Figure 5.5: Micrographs of DP590 GA specimen in different measurement scales

5.2.5. Experimental Conditions

The conditions used for deep drawing and ironing tests are given in Table 5.2 and Table 5.3. In each experiment, three samples were tested at a constant ram speed of 70 mm/sec. Two levels of BHF were selected for deep drawing test based on the preliminary FE simulations of a round cup drawing for DP590 and experimental trials.

Conditions Descriptions

Tool material D2 (uncoated tool steel, 62 HRC and Ra=0.1 µm)

Sheet material DP 590 GA / thickness = 1.23 mm

Lub A: Polymer based lubricant with pressure additives

Lub B: Polymer based lubricant with pressure additives

Lub F: Chlorinated water emulsion lubricant Tested lubricants Lub M: Synthetic water emulsion lubricant

Lub N: Straight oil lubricant

Lub P: Straight oil lubricant

BHF 30 and 70 tons (metric)

Testing speed 70 mm/sec

Environment temp. 25 ºC

Table 5.2: Conditions used in the deep drawing test

Conditions Descriptions

Tool material A2 (uncoated tool steel, 58-60 HRC and Ra = 0.4 µm)

Sheet material DP 590 GA (Galvannealed) / thickness = 1.23 mm

Tested lubricants Same six stamping lubricants tested by deep drawing

Ironing ratios 4~9 % along the sidewall of drawn cup

Testing speed 70 mm/sec

Environment temp. 25 ºC

Table 5.3: Conditions used in the ironing test

5.3. Deep Drawing Tests Results

5.3.1. Load-Stroke Curves

In the deep drawing test, three sheet specimens were tested under a same testing condition. The test conditions were found to be reproducible, as indicated with load-stroke curves obtained from three sample runs as shown in Figure 5.6.

Figure 5.6: Comparison of load-stroke curves obtained by testing three samples at a same testing condition

The load-stroke curves measured in testing various lubricants were compared, as shown in Figure 5.7. The maximum punch force depends on the friction condition at the tool-sheet interface. Lub A and Lub B showed about 5 tons lower punch force than other lubricants as the stroke increased.

Figure 5.7: Load-stroke curves obtained for various lubes tested at a low BHF

As the BHF increases from 30 to 70 tons, sheet blanks coated by Lubes A and B were successfully deep drawn while sheet specimens coated by other lubes were fractured during deep drawing as shown in Figure 5.8.

Figure 5.8: Deep drawn vs. fractured samples

Figure 5.9 compares load-stroke curves of fully drawn cups and fractured cups at 70 tons BHF. This fracture is caused by the breakdown of lubricant film at high contact pressure.

Figure 5.9: Load-stroke curves obtained for various lubes tested at a high BHF

5.3.2. Comparison of Maximum Punch Force, Fmax

The maximum punch forces, Fmax, measured from the tests with various lubricants at both BHF 30 and 70 tons, are shown in Figure 5.10. The maximum punch force increased about 2~4 tons with increasing BHF because the higher contact pressure at the sheet-tool interface in the flange region causes higher friction forces. Most lubricants gave fully drawn cups at BHF 30 tons. However, the sheet samples coated with lubricants F, M, N and P were fractured at BHF 70 tons. When the Fmax exceeds 43 tons, it resulted in the sample fracture, Figure 5.10. Lub B showed a superior performance than other lubricants and Lub A gave the secondary lower Fmax, regardless of the BHF’s. Other lubricants F, M, N and P showed similar performance at the BHF of 30 tons.

Figure 5.10: Maximum punch force attained from deep drawing tests with various lubricants

5.3.3. Comparison of Perimeter and Flange Draw-in Length

The perimeter or draw-in length of flange can be used as another friction indicator, instead of sidewall thinning distribution in the drawn cup. The smaller perimeter indicates the better performance of lubricant. As shown in Figure 5.11, Lub A and Lub B gave the smaller perimeter within the measurement error bar compared to other lubes. The perimeter of fractured cup samples were not measured at BHF 70 tons, because the fracture point and shape were not consistent to calculate the average perimeter value from three samples.

Figure 5.11: Comparison of perimeter at the flange of drawn cups coated with different lubricants

In measuring the draw-in length, due to the non-circular shape of the flange area, four measurements were taken in different radial directions as

shown in Figure 5.12. Based on this criterion, as shown in Figure 5.12, Lub A and Lub B gave slightly longer draw-in lengths than other lubes. However, the difference of these two lubricants and other lubricants was not distinguishable enough to compare.

Figure 5.12: Draw-in length, Ld, for different lubes tested at BHF 30 and 70 tons

5.4. Ironing Test Results

5.4.1. Load-Stroke Curves

The load-stroke curves measured in ironing tests of various lubricants were compared for a fixed ram speed, 70 mm/sec, as shown in Figure 5.13. The maximum punch force depends on the friction condition at the tool-sheet

interface. Lub A and Lub B showed about 7-23 KN lower maximum punch force than other lubes as the stroke increased.

Figure 5.13: Load-stroke curves measured for ironing tests with various lubricants

5.4.2. Sidewall Thinning Distributions

In ironing test, the sidewall thinning changes depending on the interface friction between ironing die and workpiece, because the friction increases the tensile stress of sidewall during ironing. Therefore, the smaller thinning indicates a good lubrication. The thinning was calculated by measuring the sidewall thickness of ironed cup before and after the test. Three samples tested for each lubricant were measured and each sample was measured in four times with respect to circumferential direction. Therefore, average values of thinning were plotted along the side wall of ironed cup as shown in Figure 5.14.

Figure 5.14: Comparison of sidewall thinning distributions obtained from the ironing tests with different lubricants

As shown in Figure 5.14, Lub A and Lub B showed smaller thinning distribution than other lubricants. This result also corresponds to the deep drawing test results in terms of ranks for lubricants.

5.5. Finite Element Analyses of Deep Drawing and Ironing

In our experiments, it was not practical to measure the temperature and pressure at the tool-workpiece interface. Therefore, the thermal-mechanical coupled FE simulations were conducted by using the commercial FEM code, DEFORM-2D, to predict the interface temperature and pressure generated during the test.

5.5.1. FE Models for Deep Drawing and Ironing

By considering the actual geometry of tool and workpiece, 2D axisymmetric FE model was prepared to simplify the tool-workpiece configuration as shown in Figure 5.15-a. The material properties of DP590 were obtained by conducting the Viscous Pressure Bulge (VPB) Test (please refer Figure 5.15-b. Detailed procedure and results of VPB test are available in APPENDIX D. These data were used as the input data in the deep drawing and ironing simulations.

a) b)

Figure 5.15: a) FE model of the round cup drawing and b) Flow stress of sheet material (DP590GA) used for FE simulations

After completing deep drawing simulation, the cup geometry drawn to the height of 77 mm was trimmed to a height of 50 mm by deleting the elements beyond the cup height of 50 mm and interpolating the strains back into the trimmed geometry. Thus, the history of strain and stress obtained during deep

drawing stage was used as input for the ironing simulation. Figure 5.16 shows the FE model for the ironing process.

Figure 5.16: FE model of the round cup ironing process

5.5.2. Determination of Simulation Parameters and Thermal Properties

In thermal-mechanical coupled FE simulations, the sheet material properties, friction coefficient (µ or m) and heat transfer coefficient (HTC) are the critical input parameters to obtain the reliable simulation results. In deep drawing simulation, a constant value of COF (µ) was used. However, for ironing simulations, the constant shear friction factor, m, was used (please refer Eq. (1.2)). Thermal properties of tool and workpiece were obtained from the material database in DEFORM and the material handbook website [www.efunda.com/materials]. Heat transfer coefficient was selected from the previous sensitivity analyses of FEA and TCT results, since the same tool material (D2) and same lubricants were used in both tests. Details of thermal properties and simulation parameters used in FE simulations are given in Table 5.4.

The sheet was meshed with 3 elements along its thickness and with 1000 elements. The sheet was considered as a plastic object and the other objects (die, punch and blank holder) were considered as rigid.

Input data for simulation Workpiece Die and Punch Blank holder

Material type DP590 D2 Tool steel P20 Tool steel

Object type Plastic Rigid Rigid

Time step size 0.1 mm/step (Total step No. = 800)

Blank holder force 30 and 70 tons

Heat transfer coefficient 11 KW / m2-K

Coefficient of Friction (COF) & Selected in a range of COF = 0.01~0.14 & m=0.05~0.3 Shear Friction Factor, m

Thermal conductivity 60.5 W / m-K 50.71 W / m-K 24.57 W / m-K

Heat Capacity 3.41 J / m3-K 3.81 J / m3-K 2.78 J / m3-K

Emissivity 0.95 0.7 0.7

Initial temperature (°C) 25

Table 5.4: Input data used in FE simulations of deep drawing and iroing

5.5.3. FE Results of Load-Stroke Curve

The load-stroke curves predicted by FE simulation with two different COF inputs (0.01 and 0.05) are compared with experiments with BHF of 30 tons and with a ram speed of 70 mm/sec, as shown in Figure 5.17. The small fluctuation in the load-stroke curve observed in the simulation was caused by the oscillation of contact nodes between the punch and a deforming workpiece. By selecting the

100

appropriate coefficient of friction (COF), FE simulation gave a good agreement with the maximum punch force and the overall trend with experiments.

Figure 5.17: Comparison of load-stroke curves predicted by FE simulations with deep drawing experiment with Lub B at BHF 30 tons

The load-stroke curves for ironing test were predicted by FE simulation with different friction coefficients. In ironing simulation, due to the high contact pressure at the die-workpiece interface, the shear friction factor, m, was used to define the frictional condition at the tool-workpiece interface. Figure 5.18 compares the load-stroke curves between selected FE predictions and experiments. The experimental results were within the FE predictions of m=0.05 and 0.12, as shown in Figure 5.18. However, the stroke corresponding to the maxim punch force was predicted differently by FEM compared to experiments. This may be caused by the fact that the rigid plastic model used in DEFORM does not consider the elastic deflection of cup specimen. Also, in our ironing

101

simulation, the deep drawn geometry was trimmed and the history of strain was extrapolated to the new trimmed geometry. Therefore, there is some discrepancy between FE model of ironing and the experimental model.

Figure 5.18: Comparison of load-stroke curves predicted by FEM with ironing experiments

5.5.4. Predictions of Contact Pressure at the Tool-Workpiece Interface

To better interpret the test results, FE simulation was used to predict the contact pressure at the die-sheet interface, since this variable is difficult to measure during the test. Figure 5.19 shows the die-sheet interface pressure predicted by FEM simulation at BHF 30 and 70 tons with COF=0.05. Using the point tracking function in DEFORM, about 70 nodes were selected to calculate the normal pressure values. At BHF 30 tons, the inlet and outlet of the die corner were predicted to generate a very high contact pressure, up to about 400 Mpa, while the pressure along the sheet curvature remained between 50 and 150 MPa.

102

Also, the contact pressure at the straight flange area was predicted to be in the range of 10 to 45 MPa. At BHF 70 tons, the contact pressures at the inlet and outlet of the die corner were predicted to be 5~10% higher than FE predictions at same points with BHF 30 tons. Therefore, when the sheet comes into the inlet of the die corner, a high contact pressure is expected and this pressure tends to increase as the BHF increases. This severe interface conditions can cause easily lubricant film break down and depending on the lubricant, the coefficient of friction at the die-sheet interface may increase significantly during the test.

Figure 5.19: FE simulation results at BHF 30 and 70 tons with COF =0.05, a) Tracking points for calculating pressure distribution at 76 mm stroke and b) Pressure distribution at the deformed sheet and the die corner radius

The pressure distribution was calculated for the ironing simulation with friction factor, m=0.1, that gave a good prediction of punch force compared to experiments as shown in Figure 5.20. The maximum pressure was predicted as 1 GPa at the interface.

103

Figure 5.20: FE prediction of pressure distribution between sheet and ironing die at 27 mm stroke

5.5.5. Temperature Distribution at the Tool-Workpiece Interface

By considering the heat transfer, the thermal-mechanical coupled simulation was conducted to predict the temperature increase and its distribution in the die and deformed sheet. BHF 30 tons and COF of 0.05 were selected for the simulation conditions.

Figure 5.21 presents the temperature distribution of the fully deformed cup at 80 mm stroke. The maximum temperature was predicted to be 86 ˚C along the die corner radius. This may be caused by the large plastic deformation induced by the high contact pressure and frictional shear stress at a given COF (i.e. τ = µp). After metal flows over this die corner, it was predicted to be cooled down. Although this maximum temperature, predicted for our given die/sheet geometry, is not high enough to affect the change in lubricant viscosity, this FEM result implies that the frictional condition along the deformed workpiec may change because of the large deviation of temperature and pressure. The temperature distribution in the drawing die was also predicted by FEM as shown

104

in Figure 5.22. As expected, the maximum temperature was predicted to occur at the die corner radius.

Figure 5.21: FE prediction of temperature distribution in the final drawn cup (BHF=30 tons and COF=0.05)

Figure 5.22: FE prediction of temperature distribution in the drawing die (BHF=30 tons and COF=0.05)

105

In ironing simulation, the interface temperature was predicted as shown in Figure 5.23. Friction factor, m=0.1, was used in simulation.

Figure 5.23: FE prediction of temperature distribution at the workpiece and ironing die in FE simulation with m=0.1

5.6. Summary and Conclusions

5.6.1. Summary • Various stamping lubricants were evaluated to form DP 590 GA material by using the deep drawing and iroing tests. • To predict the temperature distribution at the die-workpiece interface, FE simulations of these tests were conducted. • The performance of lubricants was evaluated by i) the maximum punch force, ii) the maximum applicable BHF, iii) visual inspection of zinc-powdering and galling and iv) draw-in length and perimeter in the flange of drawn cups.

106

5.6.2. Conclusions

The major conclusions drawn from this study are summarized as given below: • Deep drawing and ironing tests were able to distinguish the performance of different stamping lubricants with AHSS under near production conditions. • Based on performance evaluation criteria, Lubes A and B were most effective in regardless of blank holder forces in deep drawing test. • Lubes M, F, N and P gave good drawn parts at BHF 30 tons, however, they all showed fractures in BHF 70 tons. • Most lubricants were relatively effective for use in moderate deep drawing operations. • FE simulation results imply that the variation of pressure at the tool-workpiece interface can influence the local frictional condition. However, the temperature increase was not predicted so severe to change the lubricant behavior during the test. • FE prediction indicates that the contact pressure at the tool-sheet interface becomes more severe as the BHF increases. Therefore, in laboratory tribotests, it is necessary to evaluate the performance of lubricant at the relevant conditions that exist in production. • In both deep drawing and ironing tests, severe galling or powdering was not observed. • The major reason of this phenomena can be the relatively short sliding length between the die and sheets in both tests. The total sliding length for three specimens used at a test condition were only about 180 mm in deep drawing and 90 mm in iroing tests. Galling is usually difficult to observe without testing a large number of parts, which requires high material cost for laboratory tests. Therefore, it is desirable to develop a new test method to initiate galling by increasing the total sliding length as well as reducing the material cost, simultaneously.

107

CHAPTER 6

INVESTIGATION OF GALLING BY USING THE STRIP DRAWING TEST

In deep drawing and ironing tests, galling was not observed because the total sliding length between the die and three sheet specimens at a single test condition was not large enough to initiate galling on the die surface. Therefore, the strip drawing test (SDT) was developed to increase the total sliding length without wasting a lot of sheet material. This test is also capable to test the higher grades of AHSS (DP 780, TRIP 780 and DP 980), because the limited formability of these steels makes the cup drawing test difficult to conduct. The tooling for SDT was designed to offer many advantages to investigate the effect of various die radius and die materials with surface treatments on friction, lubrication and galling in forming AHSS. In this study, various die coatings and lubricants were evaluated by using SDT. The performance of die coatings was compared in terms of the severity of galling while the effectiveness of lubricants was evaluated by comparing the maximum drawing force and the geometry of specimen tested.

6.1. Experimental Setup

6.1.1. Design of Experiment

With preliminary FE simulations of strip drawing, four different die radii (5, 8, 10 and 12 mm) were determined to change the maximum contact pressure,

108

Pmax, in the range of 110 ~ 260 MPa without any necking of strip that was 356 mm long and 25.4 mm wide. A commercial FEM code, PAM-STAMP 2G, was used for the preliminary FE simulation. TRIP 780 of 1 mm thickness was used as the strip material and the sheet material properties were determined by VPB test. Detailed data of flow stress can be found in APPENDIX-D. The sidewall thinning was predicted by FE simulation as shown in Figure 6.1. The BHF was used as 75 KN.

Figure 6.1: Thinning distribution on the strip predicted by FEM

Details of FE results are summarized at Table 6.1. In addition, all these results were carefully analyzed to determine i) the dimensions of initial strip, ii) applicable BHF without any necking of strip, iii) the capacity of load cell to measure the punch force with a reliable resolution.

109

Die Radius (mm) Pmax (MPa) Max. Sidewall Thinning (%) Max. Punch Force (KN)

5 257.6 9.0 65

8 196.2 6.1 55

10 154.9 5.0 45

12 110.1 3.5 35

Table 6.1: Strip drawing simulation results at different die radius

6.1.2. Description of SDT Tooling

The die insert was made to have four different radii as shown in Figure 6.2. In addition, the insert was designed to freely adjust the die-punch clearance that can change the ironing ratio of strip up to 50%. FE simulation predicted the contact pressure at the ironing zone to be about 520 MPa for 20% ironing ratio. In SDT, two die inserts are used, thus the final deformed shape has a U-shape with 80 mm depth as shown in Figure 6.2. Detailed engineering drawings of the SDT dies are given in APPENDIX-E.

Figure 6.2: Schematic of strip drawing test

110

Figure 6.3 shows the tooling used for SDT. Two insert dies are attached to a draw die that moves with an upper ram in a 160 tons hydraulic press.

Figure 6.3: Strip drawing test tooling

6.2. Evaluation of Die Coatings

Strip drawing tests were conducted to evaluate the anti-galling property of three different PVD-coated dies and an uncoated die that were selected with the inputs from the tool manufacturers. All the die inserts were made of the same cold work tool steel, DC53. Die inserts were made in a sequence of i) heat treatment, ii) machining, iii) grinding/polishing and iv) PVD coating. Micrographs of initial die surfaces are compared in Figure 6.4. The Rockwell hardness of die material and the micro-hardness of die coating were measured. As shown in Figure 6.4, the hardness of TiCN showed 48% percent harder than other coatings, XNP and CrN, and about 300% harder than the uncoated die surface. It is also well known that, depending on the coating materials and methods (PVD, CVD and TD), the hardness of coating can be much stronger than the metal substrate.

111

Figure 6.4: Micrographs of various tool coatings before the test

Galling is difficult to observe by testing a small number of samples, thus in our study, 40 specimens were continuously tested for each die coating without cleaning the die surface under four different test conditions in terms of lubrication and contact pressure. After conducting twenty times drawing tests, additional twenty ironing tests were continued with the same die inserts at 14% ironing ratio. This ironing ratio was obtained by adjusting the clearance between the insert die and punch with 1.7 mm thickness metal shims. Tested samples are shown in Figure 6.5. The ironing area was localized in the middle of strip width as shown in Figure 6.5-b, because a cylindrical shape punch was used in the test. Through preliminary experiments and FE simulations, the BHF was determined

112

to be 50 KN and 10 KN for strip drawing and ironing, respectively, to avoid excessive thinning or fracture. Detailed test matrix is given in Table 6.2.

Figure 6.5: Drawn and ironed specimens in different views

Test conditions Descriptions

Sheet material DP 980 GI (initial thickness=1.4 mm)

DC53 uncoated (62 HRC, 830 VHN, Ra=0.19 µm)

Tool materials DC53 with PVD-CrN (62 HRC, 1300 VHN, Ra=0.16 µm)

(total 4 matls.) DC53 with PVD-XNP (62 HRC, 1332 VHN, Ra=0.25 µm)

DC53 with PVD-TiCN (62 HRC, 2500 VHN, Ra=0.34 µm)

Die radius 5 mm

Lubricant Straight oil (coated in 2.0~3.0 g/m2 by using a draw-down bar)

i) 10 drawings with lubricant

ii) 10 drawings without lubricant Test scheme iii) 10 ironings with lubricant

iv) 10 ironings without lubricant

Table 6.2: Test conditions used in SDT for evaluating die coatings

113

6.2.1. Load-Stroke Curves

The drawing force was measured as the stroke increased. Figure 6.6 compares the load-stroke curves obtained from the tests with different die inserts under dry and lubed conditions. As shown in Figure 6.6, PVD coated die inserts gave about 30 percentages (i.e. 6 KN) and 36 percentages (7 KN) higher punch force with respect to uncoated die insert (Bare) at lubed and dry conditions, respectively. The similar trend was found also in previous TCT results, which PVD coated tools gave relatively higher COF compared to uncoated tools. This consequence can be explained by a fact that a hard coating can cause a lubricant film break down relatively easily at high contact pressure compared to uncoated die, because the larger Young’s modulus of the hard coating [Klocke et al. 2002].

Figure 6.6: Load-stroke curves obtained from strip drawing tests for various die coatings under dry and lubed conditions

114

6.2.2. Surface Topography Change in Die Coatings Before and After the Test

Die surface was carefully examined to monitor the onset of galling while conducting 40 specimens. In addition, to examine the damage of coatings after the drawing and ironing tests, we took the micrograph of critical die surface that experienced high contact pressures during drawing and ironing tested as shown in Figure 6.7. Before the microscope examination, die inserts were cleaned with Acetone to remove any material pick up and oil contamination remained after the test.

Figure 6.7: Microscope examination for the ironing zone of a die insert

The micrographs of die surface were taken at a 5mm corner radius before and after the tests of 40 specimens. Since this area is expected to have the maximum contact pressure, the die coating can be damaged and it results in the initiation of galling. As shown in Figure 6.8, the uncoated die surface (Figure 6.8- A2) showed severe scratches along the die corner and all the coated dies showed their coating layers removed at the sharp die radius. The severity of the damage is slightly more in CrN (B2) and XNP (C2) compared to TiCN (D2).

115

Figure 6.8: Comparison of surface topography change of various PVD-coated and uncoated die surfaces at 5 mm corner radius

116

A sequential micrograph was taken for the ironing zone and was compared in Figure 6.9. An uncoated die surface showed severe scratches along the ironing zone, Figure 6.9. Among the coated dies, TiCN showed less damage of coating compared to CrN and XNP coatings. The micrographs of initial die coatings are given in Figure 6.4.

Figure 6.9: Change in surface topography of uncoated and coated dies

117

6.2.3. Ranking of Galling

The severity of galling was qualitatively ranked with a visual inspection. Figure 6.10 shows a severely galled die surface after testing 40 specimens.

Figure 6.10: Galled die surface after the test

Galling rank (GR) for various die coatings is compared in Figure 6.11. The first comparison of GR for various die coatings was made after drawing of 20 specimens and the second comparison was made after ironing of 20 specimens.

Figure 6.11: Ranking of galling for various die coatings

118

As shown in Figure 6.11, TiCN coating showed the best effectiveness in reducing galling under all the test conditions.

6.3. Evaluation of Lubricants

Various stamping lubricants were tested by SDT to evaluate their performances with uncoated and coated dies. All these lubricants were tested by previous TCT with various die materials and coatings. However, with TCT results only, it is difficult to conclude the effectiveness of lubricants at different die material conditions, because TCT is limited to generate a metal flow with plastic deformation that can significantly change the lubricant behavior and its effectiveness.

Detailed test conditions used in SDT are given in Table 6.3. Three specimens were tested for each condition. Two different zinc coatings, GA and GI, on the similar grades of AHSS (DP590/600) were used for the strip materials. An uncoated and TiCN coated die inserts were selected from the previous evaluation of die coating via SDT. After several experimental trials, blank holder force (BHF) was determined respectively as 40 KN for DP600 (thickness=1 mm) and 50 KN for DP590 (thickness=1.2 mm) to avoid any fracture. A ram speed was set to be 30 mm/sec.

119

Testing parameters Descriptions

Sheet materials DP 590 GA, 1.2 mm (initial thickness)

DP 600 GI, 1.0 mm (initial thickness)

Tool material and DC53 uncoated and PVD (TiCN) coated coatings

Die radius 8 mm

BHF 40 KN for DP600 GI and 50 KN for DP590 GA

Ram speed 30 mm/sec

Lubricants Lub B: Polymer-based lubricant with EP additives

(total 8 lubes) Lub C: Water-soluble dry film lubricant

Lub D: Water-soluble dry film lubricant

Lub E: Water-free dry film lubricant

Lub L: Water-free dry film lubricant

Lub M: Synthetic lubricant

Lub N: Straight oil

Lub S: Water-soluble lubricant

Table 6.3: Test conditions used in SDT for evaluating lubricants

Dry film lubricants (DFL) are divided into a water-soluble dry film lubricant and a water-free dry film lubricant. Both DFLs are applied in amount of 0.5~1.5 g/m2 at the rolling mill. They stick to the surface of sheet panels to provide better drawing performance compared to wet lubricants and to offer sufficient corrosion protection. Water-soluble DFL is not compatible with most

120

adhesives used in automotive body construction, while the water-free DFL has compatibility with almost all commonly used adhesives [Meiler et al. 2004].

6.3.1. Load-Stroke Curves

The punch force was measured during the strip drawing test. Figure 6.12 compares the load-stroke curves obtained from the tests for different lubricants with GA coated strips at an uncoated die (DC53 Bare).

Figure 6.12: Load-stroke curves obtained for various lubricants with an uncoated die

As shown in Figure 6.12, the maximum punch forces were observed at the end of stroke for the most of lubricants. Similar load-stroke curves were obtained from the tests with GA and GI coated strips at both uncoated dies and PVD (TiCN) coated dies. These results are available in APPENDIX-F.

121

6.3.2. Comparison of Maximum Punch Force

The maximum punch force was used to evaluate the performance of lubricants, because the total punch force is influenced by a frictional force at the tool-workpiece interface. Figure 6.13 compares the maximum punch forces obtained from the tests of different lubricants and die inserts with DP 590 GA.

Figure 6.13: Maximum punch force obtained from the SDT for GA coated sheet

A shown in Figure 6.13, Lub B and Lub C consistently showed a superior performance to other lubricants for both uncoated and TiCN coated die inserts. Especially, in the test with TiCN coated die, Lub E and Lub M showed also a reasonable frictional response, while other lubricants, (Lubes D, N and S) gave about 50% higher maximum punch force with respect to Lub B and Lub C. This implies that frictional force increased from uncoated die to the coated die, depending on the type of lubricant. Especially, a water-soluble DFL (Lub D) and

122

a straight oil (Lub N) showed about 18% increase of frictional force at TiCN coated die compared to an uncoated die. Comparing to SDT results with uncoated dies, Lub M gave a slightly lower punch force. This indicates that a synthetic lubricant (Lub M) can also be a reasonable selection with a PVD (TiCN) coated die.

Figure 6.14 compares the maximum punch forces obtained from the tests of different lubricants and two different die inserts with DP 600 GI. Similar to the results with GA coated sheets, Lubes B and C were most effective to reduce friction at both uncoated and coated die conditions. Especially, Lub L (water-free DFL) also gave an equivalent performance with Lubes B and C at uncoated die, although it gave the higher punch force at TiCN coated dies. A straight oil (Lub N) reduced friction in GI coated strips compared to GA coated strips. For the TiCN coated die, Lubes L, M and N showed also reasonable frictional responses.

Figure 6.14: Maximum punch force obtained from the SDT for GI coated sheet

123

6.3.3. Comparison of Strip Elongation

In SDT, frictional condition at the die and strip changes the sidewall thinning of drawn strip. As a result, the curvilinear length of strip is affected by the lubricant performance. The strip elongation was calculated by measuring the strip length before and after the test. The inner and outer side lengths of strips were measured and an average value was used for the comparison of strip elongations at different lubricants. Figure 6.15 and Figure 6.16 compare the strip elongation of GA and GI strips tested with various lubricants and two different die inserts. In overall, the ranks of lubricants based on this criterion showed a good agreement with ranks obtained from the maximum punch force comparison.

Figure 6.15: Strip elongation of DP590 GA specimens after the test

124

Figure 6.16: Strip elongation of DP600 GI specimens after the test

6.4. Finite Element Analyses for Strip Drawing and Strip Ironing

6.4.1. Preparation of FE Simulation Model

FE simulations were conducted for SDT of DP980 GI strip material to calculate the critical interface pressures and temperatures that may lead galling in the test. Two commercial FE codes, PAM-STAMP 2-G and DEFORM 3-D, were used and the simulation results were cross checked. 3-D FE model was prepared as shown in Figure 6.17. Detailed simulation conditions are available in APPENDIX-F.

125

Figure 6.17: FE model for the SDT

6.4.2. Load-Stroke Curve Predictions

FE predictions of load-stroke curve with different coefficients of friction were compared with two selected experimental results as shown in Figure 6.18.

Figure 6.18: Comparison of FE predictions of load-strokes with experiments

126

Experimental result of uncoated die condition showed a good agreement with FE result with COF=0.08, while other coated dies (XNP, CrN and TiCN) showed a good agreement with FE results with a range of COF=0.1 ~0.12.

6.4.3. FE Predictions of Contact Pressures and Temperature Increase

To investigate the effect of critical pressures and temperatures at the die- strip interface on galling, FE simulations were conducted to predict the interface pressure and temperature. Figure 6.19 showed the history of normal pressure on the sidewall of strip contacting with a die insert for drawing and ironing operations, individually. The maximum contact pressures were predicted to be 116 MPa in drawing and 410 MPa in ironing, respectively. As shown in Figure 6.19, the pressure distribution along the width of strip was found to be uniform in drawing, while the maximum pressure in ironing was predicted to be localized at the middle of strip width. This localization of pressure was caused by using a round punch in our test. As a result, the clearance between a punch and die varies along the strip width and the maximum ironing ratio was observed at the middle of strip width in tested samples, Figure 6.19-c.

127

Figure 6.19: Pressure distributions of strips for a) drawing and b) ironing and c) tested specimens

A thermal-mechanical coupled simulation was conducted to predict the temperature generations during strip drawing and ironing by using DEFORM 3- D. Detailed input data used for these simulations are available in APPENDIX-F. The temperature distribution was predicted along the deformed strip as shown in Figure 6.20. Considering the room temperature assumed to be 25°C, the maximum temperature was predicted to be about 33 °C for strip drawing and 37 °C for strip ironing at the end of stroke, Figure 6.20. Therefore, a temperature increase may not be a significant influence on the initiation of galling in our tests.

128

Figure 6.20: FE prediction of temperature increase in SDT and SIT (COF=0.1)

6.5. Summary and Conclusions

6.5.1. Summary • Various die coatings and stamping lubricants were evaluated to draw DP980 GI, DP 600 GI and DP 590 GA materials by using the SDT. • The severity of galling was compared for different die coatings with visual inspections after conducting 40 specimens of drawings and ironings. • Micrographs of die surfaces were taken to examine the damage in coating after the tests. • In lubricant evaluations, GI and GA coatings on the similar grades of AHSS were tested with various wet and dry film lubricants. • FE simulations were conducted to calculate the interface pressure and temperature in drawing and ironing tests.

129

6.5.2. Conclusions

The major conclusions drawn from this study are summarized as follows:

1. In the evaluation of die coatings, TiCN showed a superior anti-galling properties to other PVD coated dies and uncoated die.

2. This may be resulted from i) a higher hardness of TiCN (2500 VHN) compared to other coatings, CrN and XNP (1300~1330 VHN) and ii) chemical affinity between a tool coating and GI coating of the sheet.

3. As the contact pressure increased from 110 MPa at deep drawing to 410 MPa at ironing, galling increased considerably.

4. In our test, the effect of speed on galling was not considered, because our targeted stamping process of automotive body structural parts requires a relatively low stroking rate (10~14 SPM) [VOEST-ALPINE, 2006]. Although, the flash temperature at the micro-scale contact surfaces may be high enough to influence galling, the temperature calculated at the macro-scale tool- workpiece interface was found to be small enough to neglect its effect on galling.

5. In the evaluation of lubricants, a polymer based lubricant with EP additives (Lub B) and a water-soluble DFL (Lub C) was most effective in reducing friction at the tool-workpiece interface, regardless of sheet coatings and tool coatings.

6. A water-free DFL (Lub E) and a synthetic lubricant (Lub M) effectively reduced a friction in forming GA coated sheet with TiCN coated dies.

130

CHAPTER 7

DEVELOPMENT OF A PRELIMINARY MODEL FOR PREDICTION OF GALLING

In this chapter, a detailed procedure to develop a galling prediction model is presented. Through correlation analyses between experimental results and FE simulation results, a galling severity indicator (GSI) was conceived to predict the severity of galling that is Galling Rank (GR) in our study at given material properties (i.e. sheet material properties, surface roughness, lubricant viscosity and the hardness of tool/sheet surfaces) and process conditions (i.e. contact pressure and sliding length).

7.1. Basic Modeling Concepts

Based on the results obtained from various tribotests, a sequence of reasoning was made to draw empirical relationships between galling and tribological parameters (e.g. lubricant viscosity, hardness, surface roughness, pressure and sliding length). As a result, an empirical model was developed to predict the severity of galling (e.g. galling rank, GR values) at given tribological parameters.

131

7.1.1. Perception of Galling in Stamping Galvanized AHSS

When the tool and workpiece come into contact under normal pressure, numerous junctions can be made and the size of junctions can grow as the contact pressure increases. These junctions are sheared off when the relative sliding occurs between tool and workpiece. Th displaced materials can adhere to the tool surface or flow with lubricant as wear particles at the tool-workpiece interface. Galling is initiated when these particles damage the tool coating and strongly adhered to the tool substrate as shown in Figure 7.1. As a result, the surface roughness of tool increases and the incoming sheet surface is scratched more.

Figure 7.1: Development of galling in forming galvanized AHSS

In stamping galvanized AHSS, galling was found to be mainly caused by the zinc-coatings (GI or GA) on the sheet. Because the metal substrate is coated by zinc-layers, it is difficult for the steel substrate to directly adhere to the tool surface. This perception was supported by our TCT results between GI coated sheet and uncoated sheet (please see Figure 4.14 and Figure 4.15) and SEM/EDS results of galled materials (please refer Figure 4.29 and Table 4.17). Therefore, the characteristics of zinc-coating (i.e. hardness and surface roughness) are key material parameters to consider in galling prediction models. However, in punching (i.e. hole piercing) or trimming operations for AHSS, galling is known

132

to be involved with the metal substrate, because the trimming or piercing tool penetrates through the zinc-coating as well as the metal substrate [Hogman 2004].

7.1.2. Effects of Contact Pressure on Galling

In forming AHSS, the contact pressure was found to be considerably larger than that experienced in forming mild steels. As observed in our TCT and SDT results, high contact pressures degrade the performance of the lubricant, tool coating and tool surface finish, thus it accelerates the initiation of galling as the number of part increases.

From Archard’s adhesive wear model (1956) as given in Eq. (7.1), the wear depth (Zadh) is proportional to a wear coefficient (kw), the local contact pressure (P), local sliding velocity (V) and time interval (∆t), while it is inverse proportional to the hardness of workpiece (Hw).

P ⋅V ⋅ ∆t Z adh = kw (7.1) H w

Since galling is a type of adhesive wear, it is reasonable to assume that the severity of galling is also proportional to the contact pressure and sliding length that is the product of sliding velocity (V) and time interval (∆t).

7.1.3. Effects of Hardness of Zinc coatings and Tool Coatings on Galling

The hardness of GA coating is about five times higher than GI coating. As a result, GA coatings normally showed less galling compared to GI coatings.

133

However, since GA coating is more brittle than GI coating, GA can cause more contaminations (i.e. flaking and powdering) at the tool-workpiece interface.

Various tool materials and tool coatings were evaluated against galling in TCT and SDT. The CrN and XNP coated dies gave more severe galling in SDT. This may be caused by the failures of tool coating such as delamination, internal fracture and fragment as illustrated in Figure 7.2 [Katagiri et al. 2007].

Figure 7.2: Classification of coating failure mode [Courtesy of Katagiri et al. 2007]

To consider the effects of both zinc-coating and tool coating, the effective hardness was determined by summing the hardness of tool coating and zinc coating as Eq. (7.2), after conducting sensitivity analyses in applying various mathematical expression of the effective hardness to experimental data.

H eff = H tool−coating + H zinc−coating (7.2)

7.1.4. Effects of Surface Roughness on Galling

The surface roughness of workpiece and tool is known to significantly influence the lubricant and galling behaviors [Dalton et al. 1992]. The tool surface

134

roughness was used as an indicator for galling by [Heide et al. 2001, Podgornik et al. 2004B and Shih et al. 2004], because the rougher surface of tool increases scratches and adhesion from the workpiece material. As an example, for same

GI coatings in our previous TCT, DP 500 GI of Ra=0.72 µm showed much severe

galling while DP600 GI of Ra=0.34 µm gave a minimum galling with using a same lubricant (please refer to Table 4.5). Therefore, it is logical to consider the surface roughness values of sheet and tool surfaces in modeling. In our galling

prediction model, the equivalent surface roughness, Ra, eff, was used to present the effect of surface roughness of tool and workpiece on galling. The same

expression of equivalent surface roughness, Ra, eff, (Eq. 7.3) was used to analyze the pin-on-disk test results for adhesion [Fiorentino et al. 2005].

2 2 Ra, eff = Ra ,tool +Ra ,workpiece (7.3)

7.1.5. Effects of Lubricant on Galling

Lubricant is the first protection layer for the tool surface before the tool coating and tool substrate is exposed to the sheet surface. A high viscosity lubricant tends to sustain the large viscous shear stress along the film thickness compared to a low viscosity lubricant. Thus, it normally reduces the interface friction. However, it is not practical to use a high viscosity lubricant only for reducing friction. Automotive industry also considers the easy to clean the lubricant left on their parts after stamping, because more complex cleaning procedure results in high production cost.

Our experimental results showed that the performance of lubricants can eliminate or postpone the initiation of galling. Especially, the type of lubricant such as polymer-based lubricant, synthetic lubricant, dry film lubricant, straight

135

oil, etc, is an important factor to consider in forming GA or GI coated AHSS sheets. Therefore, with these findings, a galling prediction model should consider the viscosity and chemical affinity of lubricants mating with GI and GA sheets and tool coatings.

7.2. Preliminary Model for the Prediction of Galling

The experimental data obtained from this study were used to develop an empirical model to predict the severity of galling, i.e. galling rank. In our TCT and SDT, we evaluate the severity of galling qualitatively (0 - no galling to 3 - most severe galling) with visual inspections of tool surface after the test. The example pictures of galled tools are available in Figure 4.5 and Figure 6.10 for TCT and SDT, respectively. The ranks of galling were confirmed by examining the tested tool surfaces with microscopy and the surface roughness measurements of tool and sheet samples tested. Since these are qualitative observations or indirect consequences of galling, it is desirable to predict galling quantitatively. However, in our experiments, it was very difficult to reliably measure the weight of galled materials on the die, because the weight of galling was relatively very small, compared to the weight of die or tool inserts. Therefore, the Galling Severity Indicator (GSI) was conceived in form of Eq. (7.4) to consider the interface pressure (P), sliding length (Ls) and the effective values of surface roughness and hardness of tool and sheet coatings.

P× L × R GSI = s a,eff (7.4) H eff

136

It should be noted that the values of Ra,eff and Heff were calculated by using Equations (7.2) and (7.3) that were found in the literature [Fiorentino et al. 2005] and the sensitivity analyses of our experimental data.

To correlate the GSI number to our experimental data, Galling Rank (GR), an empirical relationship between GSI and GR was defined as Eq. (7.5).

P × L × R GR = K ⋅GSI = K ⋅ s a,eff (7.5) H eff

In our experiments, galling behavior was found to change with the viscosity of lubricant as well as chemical affinity between lubricant and zinc coatings. Therefore, to consider these effects on galling, the coefficient of galling rate, K, was conceived as shown in Eq. (7.5).

To determine the K-values for different lubricant viscosity levels, the linear regression analyses were conducted by using TCT and SDT data. Table 7.1 shows the input data of TCT used for calculating GSI values. Twelve cases of TCT data were used to calculate the GSI numbers. In these calculations, two different zinc-coatings (GI and GA), two lubricants with different viscosities and three different contact pressures were considered as shown in Table 7.1.

137

Parameters Descriptions

Pressure 50 / 100 / 170 MPa

Lub A = 126 centistrokes (585 SUS) Lubricant viscosity Lub B = 562 centistrokes (2606 SUS)

GI DP500 = 0.72 µm

Surface roughness (Ra) GA DP590 = 0.34 µm D2 Tool steel = 0.14 µm

GI coating = 60 VHN Hardness (H) GA coating = 316 VHN D2 Tool Surface = 700 VHN

Ls for GA-Lub A = 6825/5189/3552 mm for 50/100/170 MPa

Ls for GA-Lub B = 7040/3542/2373 mm for 50/100/170 MPa Sliding length (Ls) Ls for GI-Lub A = 5655/4940/12220 mm for 50/100/170 MPa

Ls for GI-Lub B = 4943/5135/15015 mm for 50/100/170 MPa

Table 7.1: Input data of TCT used for calculating GSI values

The GR and GSI values based on TCT results are compared in Table 7.2. For GI coating, when GSI number approach about 2000, the GR was predicted to be 3.0, most severe galling. However, for GA coating, GSI of 735.7 corresponded to GR 2.0. This discrepancy in GSI values between GI and GA coatings indicates different behaviors of a lubricant with GI and GA coatings. Therefore, K-value should be estimated to consider the effects of lubricant viscosities as well as lubricant behaviors with GI and GA coatings on GSI values.

Galling Rank (GR) Galling Severity Indicator (GSI)

P GI- GI- GA- GA- GI- GI- GA- GA- (MPa) Lub A Lub B Lub A Lub B Lub A Lub B Lub A Lub B

50 0 0 0 0 272.9 238.4 415.8 428.9

100 0.5 0.5 1.5 0 476.8 495.6 632.3 431.6

170 3 3 2 1.0 2004.9 2463.5 735.7 491.5

Table 7.2: Relationship for GR and GSI in TCT

138

The experimental data of strip drawing test (SDT) and strip ironing test (SIT), as shown in Table 7.3, were used for calculating GSI values. The maximum contact pressure was predicted at the die-strip interface by FE analyses of

SDT/SIT. The total sliding length, Ls, was calculated by multiplying the number of specimens, 40, with the measured sliding length, 60 mm, of tested strip with respect to the die corner radius, 5 mm.

Parameters Descriptions

Max. Contact Pressure 110 MPa for SDT and 410 MPa for SIT

Lubricant viscosity Lub N = 26 centistrokes (126 SUS)

DP980 GI (strip specimen) = 0.81 µm DC53 Bare (die insert) = 0.19 µm

Surface roughness (Ra) DC53 CrN (die insert) = 0.16 µm DC53 XNP (die insert) = 0.25 µm DC53 TiCN (die insert) = 0.34 µm

GI coating = 49.24 VHN DC53 Bare (die insert) = 830 VHN Hardness (H) DC53 CrN (die insert) = 1300 VHN DC53 XNP (die insert) = 1332 VHN DC53 TiCN (die insert) = 2500 VHN

Total Sliding length (Ls) Ls for SDT and SIT = 2400 mm

Table 7.3: Input data of SDT/SIT used for calculating GSI values

The relationship between the sliding length and GSI number was compared for different die surface conditions as shown in Figure 7.3. GSI increased rather monotonously up to 1200 mm that corresponded to twenty specimens tested by SDT and fast increased up to 2400 mm that corresponded to additional twenty specimens tested by SIT. An uncoated die insert (Bare) showed

139

the fastest increase of GSI compared to other coated die inserts. As a result, it reached GR 3.0 earlier than other coated dies in terms of sliding length that corresponds to the number of specimens. By comparing GSI number to GR (galling rank), the GSI of 300 was predicted to be critical for the most severe galling, GR=3, as shown in Figure 7.3. It should be noted that a lubricant at lower viscosity was used in SDT, while lubricants used in TCT had higher viscosities. Therefore, the GSI value corresponding to GR 3.0 were different in TCT and SDT.

Figure 7.3: Sliding length vs. GSI in SDT/SIT

7.2.1. Regression Analysis for Determining the K value (Coefficient of Galling Rate)

The K-values were determined for different viscosity values by conducting linear regression analyses of GR and GSI values, which were calculated based on TCT and SDT data. Our experimental data was fitted into the

140

linear regression model, Eq. (7.6), at each viscosity level (ν). Thus, our model will be in the following.

⎛ P × Ls × R ⎞ GR = K × ⎜ a,eff ⎟ + ε , where ε ~ N(0,σ 2 ) (7.6) ν V ⎜ H ⎟ ν ν ⎝ eff ⎠ν

Where εν = the error term, Kν= the predicator of K value, GRν=the predicator of GR value.

Detailed results of linear regression analyses are available in APPENDIX- G. From the linear regression analyses, K-values for three different viscosity levels and GI coating were determined as shown in Figure 7.4. These values were found to be statistically significant at the 95% confidence level.

Figure 7.4: K-values determined for three different lubricant viscosities

141

CHAPTER 8

A CASE STUDY OF GALLING PREDICTION MODEL IN FORMING B-PILLAR

In this chapter, the use of the proposed galling prediction model was demonstrated by applying it to forming the B-pillar part to estimate the maximum number of parts that can be formed before galling occurs.

8.1. Finite Element Analyses of Forming B-Pillar Part

FE simulations were conducted to form a B-pillar part that was investigated by ULSAB, Ultra-Light Steel Auto Body-Advanced Vehicle Concepts program (2002). B-pillar is one of good example structural parts being formed from higher grades of AHSS materials (i.e. UTS = 600~1000 MPa) to improve the crashworthiness for the side impact as shown in Figure 8.1. As a result, the requirements of tribological parameters (lubricant, die material and coating) are more demanding compared to forming other mild steels or Al alloy structural parts.

142

Figure 8.1: B-Pillar structure in a 4 door sedan model [ULSAB, 2002]

The major goals of B-pillar simulations are to: • predict the critical contact pressures during forming and correlate them to our tribotest conditions. • predict the maximum number of parts to produce without severe galling at a given lubricant, die material and zinc-coated sheet. • develop guidelines to select the best and practical selections of die material and lubricant to eliminate or reduce galling in forming galvanized AHSS.

8.1.1. Preparation of FEM Simulation Model

B-pillar simulations were conducted by using a commercial FEM software, PAM-STAMP 2-G. The FE model with geometries of tool and initial sheet blank are given in Figure 8.2.

143

Figure 8.2: B-Pillar simulation model

The material properties of sheet blank (TRIP600) was obtained by the tensile test and expressed with the Krupkowsky model. Detailed sheet material properties and other input data used in FE simulation are given in Table 8.1. The blank holder force (BHF) was determined by the FEM-based sensitivity analyses to avoid wrinkling and excessive thinning of the part.

Input data Descriptions

Sheet material TRIP 600 (initial thickness=1.5 mm)

n σ = K(ε 0 + ε ) where ε0=0.018, K=700 MPa and n=0.23, Sheet material properties Normal anisotropy ( r )=1.05, Elastic modulus (E)=210 GPa, Poisson ratio (υ)=0.3

BHF 23 tons

Coefficient of friction 0.14

Table 8.1: Input data used for B-pillar simulation

144

8.1.2. FE Prediction of Thinning and FLD Diagram

From the FE analyses, the maximum thinning distribution of the final part was predicted to be 24 % as shown in Figure 8.3. Based on FLD prediction of TRIP 600 by FEM, as shown in Figure 8.3, this part is expected to be successfully formed.

Figure 8.3: Thinning distribution and FLD analysis of B-pillar simulation

8.1.3. FE Prediction of Maximum Contact Pressure in Forming B-pillar Part

The contact pressure distribution at the tool-workpiece interface was predicted in B-pillar simulations. As shown in Figure 8.4, the contact pressure along the sidewall of the part was predicted to be about 200 MPa and the

145

maximum pressure was predicted to be 400 MPa at the sharp die corner as shown in Figure 8.4.

Figure 8.4: Contact pressure distribution in the B-pillar part

Galling may occur at die surfaces where the contact pressure is maximum. Therefore, it is important to emulate these critical pressure conditions in tribotests to evaluate the performance of lubricants and galling behaviour in forming AHSS. In addition, with this simulation prediction of contact pressure, our experimental conditions (i.e. the testing pressure range=50~410 MPa) in SDT and TCT were practical in emulating the conditions that exists in stamping AHSS parts.

146

8.2. Application of Galling Prediction Model

The developed galling prediction model was applied to stamping of the B- pillar part to predict the maximum number of parts that can be produced before galling rank reached 3.0, which requires die maintenance. Detailed input data used for galling prediction model are given in Table 8.2.

Input parameter Input data

Sheet materials DP 600 GI (57 VHN and Ra=0.33 µm)

D2 (700 VHN and Ra=0.14 µm)

Die materials Cast iron (510 VHN and Ra=0.35 µm)

(surface hardness and roughness) PVD coated (CrN) DC53 (1300 VHN and Ra=0.16 µm)

PVD coated (TiCN) DC53 (2500 VHN and Ra=0.34 µm)

Lubricant viscosity 26, 126 and 562 centistrokes

Maximum contact pressure in B- 400 MPa pillar simulation

Table 8.2: Input data used for galling prediction model applied to forming B- pillar part

Assuming the GR value as 3.0 (most severe galling), the maximum number of B-pillar parts were calculated by using a sliding distance of 24 mm at maximum contact pressure of 400 MPa, as shown in Figure 8.5.

147

Figure 8.5: Sliding length and maximum contact pressure on the B-pillar part

Twelve different combinations of die materials, PVD coatings and lubricants were taken into account for estimating the maximum number of parts.

Figure 8.6: Maximum number of B-pillar parts expected for various die surface coditions and lubricants

148

As shown in Figure 8.6, PVD coated tool steel (DC53) dies showed longer life than other uncoated tool steel (D2) or cast iron. However, it should be noted that with a low viscosity of lubricant (26 cs), PVD coated dies does not give remarkable longer life compared to uncoated tool steel and cast iron. Therefore, to extend the life of PVD coated die, a high viscosity lubricant with EP additives are recommended. With considering the relatively expensive cost of PVD coated tool steels, these specially coated die materials can be used as an insert dies for the critical locations where high contact pressures are expected in B-pillar die as shown in Figure 8.7. The rest of areas under lower pressures can be made of cast iron material with no coating.

Figure 8.7: Schematic of the tooling construction with PVD coated tool steel insert in cast iron die for forming AHSS B-pillar part

149

CHAPTER 9

OVERALL SUMMARY AND CONCLUDING REMARKS

In this study, a methodology that combines FE analyses and tribotests was developed to evaluate the performance of lubricants, tool materials and tool coatings in forming galvanized AHSS.

To identify the critical pressures or temperatures that may initiate galling in production, FE analyses of B-pillar part and cup drawing were conducted and the pressure at the tool-workpiece interface was found to be the dominant factor to cause lubricant film break down and galling. Therefore, these critical pressure conditions were emulated in the test conditions of various tribotests, e.g. Twist Compression Test (TCT), Deep Drawing/Ironing Tests (DDT/IT) and Strip Drawing/Strip Ironing Tests (SDT/SIT) to obtain the empirical relationship between important tribological parameters (lubricant viscosity, interface pressure, sliding length, surface roughness and the hardness of coating) and galling behaviors.

The results of this study helped to develop a preliminary model to predict galling in forming galvanized AHSS. The use of the developed galling prediction model was demonstrated by applying it to forming the B-pillar part to estimate the maximum number of parts that can be formed before galling occurs. In addition, a list of recommendations for selecting die materials, die coatings and lubricants were made for stamping the B-pillar part.

150

This preliminary model can provide a cost effective way to select practical and best combinations of lubricant, tool material and tool coating for reducing galling in forming galvanized AHSS. Also, it can be further developed for use as a scheduling tool for the die maintenance before galling occurs, thus galling and the unexpected downtime of stamping process can be eliminated.

The following research contributions were obtained from this dissertation work:

Development of a methodology for evaluating lubricants, zinc-coatings, tool materials and tool coatings to eliminate galling in forming galvanized AHSS.

Development of FEM simulation models (2-D and 3-D) for predicting the tool-workpiece interface pressures and temperatures in tribotests and forming AHSS automotive structural part.

Advancement in the understanding of the effect of interface pressure, sliding length, lubricant viscosity, hardness and surface roughness on the progress in galling behavior for uncoated and PVD-coated dies in stamping AHSS.

Development of a preliminary model to predict galling in forming galvanized AHSS.

Application of galling prediction model to stamping B-pillar part to schedule the die maintenance to eliminate unexpected press down time in production.

Contribution to the optimization of tribological conditions, cost reduction, and surface quality improvement aspects in stamping technologies.

Contribution to establish guidelines for automotive stamping industry to select the best and practical tribological system (chemical, mechanical surface treatment of die and sheet and lubricant) that reduce or eliminate galling in forming AHSS.

151

REFERENCES

[AISI, 2004] AISI, “Automotive Steel Design Manual” (2004) [Andreasen et al. 1998] Andreasen, J.L., Bay, N., and Chiffre, L. De, “Quantification of Galling in Sheet metal forming by Surface Topography characterization,” International Journal of Machine Tools & Manufacture, Vol.38, pp.503-510, (1998) [Archard et al. 1956] Archard, J.F. and Hirst, W., “The Wear of Metals Under Unlubricated Conditions”, Proceeding of R. Soc. London A 236, pp.397-410 (1956) [Asnafi et al. 2004] Asnafi, N., Kjellsson, K., Johansson, T., “Blanking, stamping and trimming die experiences at Volvo cars”, International conference on recent developments in manufacture and use of tools and dies and stamping of steel sheets, pp263 – 274 (2004) [Azushima et al. 1991] Azushima, Uda, Yokohama and Kudo, “An Interpretation of the Speed Dependence of the Coefficient of Friction under the Micro-PHL Condition in Sheet Drawing”, Annals of CIRP(1991), Vol.40, pp.227-230 (1991). [Azushima et al. 1995] Azushima and Kudo, “Direct Observation of Contact Behaviour to Interpret the Pressure Dependence of the Coefficient of Friction in Sheet Metal Forming”, Annals of the CIRP, Vol.44, pp. 209-212 (1995). [Azushima et al. 1996] Azushima, Yoreyama, Yamaguchi and Kudo, “Direct Observation of Microcontact Behavior at the Interface between Tool and Workpiece in Lubricated Upsetting”, Annals of the CIRP, Vol. 45, pp. 205-210 (1996) [Azushima et al. 2000A] Azushima and Tanaka, “Lubricant Behavior Trapped within Pockets on Workpiece Surface in Lubricated Upsetting by Means of Direct Fluorescence Observation Technique”, Annals of the CIRP, pp. 165-168 (2000) [Azushima et al. 2000B] Azushima and Akira, “FEM Analysis of Hydrostatic Pressure Generated Within Lubricant Entrapped Into Pocket on Workpiece in Upsetting Process”, Transactions of the ASME, Vol.122, pp.822-827 (2000)

152

[Bay et al. 1975] Bay, N., Wanheim, T. and Petersen, A.S., “Ra and the average effective strain of surface asperities deformed in metal forming processes”, Wear, Vol.34, pp. 77-84 (1975) [Bech et al. 1998] Bech, Bay and Eriksen, “A Study of Mechanism of Liquid Lubrication in Metal Forming”, Annals of the CIRP, Vol.47, pp.221-226, (1998) [Bech et al. 1999] Bech, Bay and Eriksen, “Entrapment and escape of liquid lubricant in metal forming”, Wear, Vol. 233, pp.134-139 (1999) [Bhushan 2002] Bhushan, B., “Introduction to Tribology”, John Wiley & Sons, Inc. New York, US (2002) [Bhushan 2004] Bhushan, B., “Springer Handbook of Nanotechnology”, Springer-Verlag, Heidelberg (2004) [Blau et al. 1999] Blau, P.J. and Budinski, K.G., “Development and use of ASTM standards for wear testing,” Wear, Vol. 225-229, pp.1159-1170, (1999) [Bowden et al. 1967] Bowden, F.P. and Tabor, D., “Friction and Lubrication”, Methuen & Co. LTD, London, U.K. (1967) [Brown et al. 2006] Brown, M. and Bosler, P., “Selecting Stamping Lubricants for Advanced High-Strength Steels,” Metal Forming, Feb. (2006) [Carlsson et al. 2001A] Carlsson, P., Bexell, U., Olson, M., “Tribological performance of thin organic permanent coating deposited on 55% AL-ZN coated steel – Influence of coating composition and thickness on friction and wear”, Wear, Vol 251, pp. 1075-1084, (2001) [Carlsson et al. 2001B] Carlsson, P., Bexell, U., Olson, M., “Friction and wear mechanisms of thin organic permanent coatings on hot-dip coated steel”, Wear, Vol 247, pp. 88-99, (2001) [Dalton et al. 1992] Dalton, G.M., Schey, J.A., “Effect of bead finish orientation on friction and galling in Drawbead Test”, SAE Technical Paper, No. 920632, (1992) [Fiorentino et al. 2005] Fiorentino, A., Giardini, C. and Ceretti, E., “Study on surface contacts phenomena – Relation between Friction, Roughness and Material’s Properties” Internal Report of Technology and Manufacturing Systems Group, University of Brescia (2005) [Fuller et al. 2004] Fuller, K.H., Baumeister, J., Herold, R., “Ceramic forming tools to form advanced high strength steels”, New developments in sheet metal forming, Edited by K. Siegert, IFU, Stuttgart, pp 83-102 (2004)

153

[Girschewski et al. 2000] Girschewski, B., Pfestorf, M., “Influence of Dry Film Lubricants and Surface Structure on the Forming Behavior of Aluminum Sheets” Proceedings of International Conference of New Developments in Sheet Forming, University of Stuttgart, Germany, pp. 429-444, (2000) [Gutscher et al. 2004] Gutscher, G., Wu, H., Ngaile, G., and Altan, T., “Determination of flow stress for sheet metal forming using the viscous pressure bulge (VPB) test”, Journal of Materials Processing Technology Vol.146, pp.1-7, (2004) [Heide et al. 2001] Heide, E., Veld, A.J., Schipper. D.J., “The effect of lubricant selection on galling in a model wear test”, Wear, Vol.251, pp. 973-979, (2001) [Heide et al. 2003] Heide , E., Schipper. D.J., “Galling initiation due to frictional heating”, Wear, Vol.254, pp. 1127-1133, (2003) [Hogman 2004] Hogman, B., “Punching tests of EHS and UHS steel sheet”, Recent Advances in Manufacturing & Use of Tools & Dies and Stamping of Steel Sheets, Olofstrom, Sweden, Oct. 5-6, (2004) [IISI 2006] International Iron & Steel Institute (IISI), “Advanced High Strength Steel (AHSS) Application Guidelines” www.worldautosteel.org (2006) [Jeffery 2004] Jeffery, B., “Advanced lubricant technology for high-strength steel”, Stamping Journal, June 8th (2004) [Jeffery, 2005] Jeffery, B., “AHSS Tooling Requires Greater Protection,” Stamping Journal, September 13th (2005) [Katagiri et al. 2007] Katagiri, T., Yamasaki, Y., Yoshitake, A.,”Effect of Tool Coatings on Galling Prevention in Forming of High Strength Steel”, Proceedings of the International Conference on Tribology in Manufacturing Process, pp.41-46, Yokohama, Japan, Sep. 24-26 (2007) [Klocke et al. 2002] Klocke, F., Raedt, H.W., Maβmann, T., and Breuer, D., “Friction and Wear in Cold Forming“, the International Conference on Recent Developments in Metal Forming Technology, Columbus, Ohio, US, Oct. (2002) [Kudo, 1965] Kudo, H., “A note on the role of microscopically trapped lubricant at the tool-workpiece interface,” International Journal of Mechanical Science, Vol.7, pp.383-388, (1965). [Le et al. 2003] Le and Sutcliffe, “Evolution of Surface Pits on Stainless Steel Strip in Cold Rolling and Strip Drawing”, Transactions of the ASME, Vol. 125, pp.384-390 (2003)

154

[Liljengren et al. 2006] Liljengren, M., Kjellsson, K., Asnafi, N. and Johansson, T, “Die materials, hardening methods and surface coatings for forming high, extra high & ultra high strength steel sheets (HSS/EHSS/UHSS)”, Proceedings of IDDRG 2006, pp.597-604 (2006) [Lo et al. 1997] Lo, S. and Wilson, R., “A theoretical model of micro-pool lubrication in metal forming,” Proceedings of the 1st International conference on Tribology in Manufacturing, Gifu, Japan, pp.83-90, (1997) [Madorsky et al. 2005] Madorsky, Y., and Thompson, M., “Coating for stamping and forming tools-Three options for improved wear,” Stamping Journal, March 8, (2005) [Meiler et al. 2004] Meiler, M., Pfestorf, M., Merklein, M., and Geiger, M., “Tribological Properties of Dry Film Lubricants in Aluminum Sheet Metal Forming”, Proceedings of the 2nd ICTMP, June 15-18, 2004, Nyborg, Denmark, pp.489-500 (2004) [Mizuno et al. 1982] Mizuno, T., and Okamoto, M, “Effects of Lubricant Viscosity at Pressure and Sliding Velocity on Lubrication Condition in the Compression Friction Test on Sheet Metals” Journal of Lubrication Technology, Vol.104, pp.53-59 (1982) [Mori et al. 2000] Mori, Y., Yamazaki, M., and Shindo, H., “Film designed for lubricative coated steel sheets”, Progress in organic coating, Vol. 40, pp. 119-120, (2000) [Murakawa et al. 1998] Murakawa, M., Koga, N., Watanabe, S., Takeuchi, S., “Tribological behavior of amorphous hard carbon films against zinc- plated steel sheets”, Surface Coating Technologies, Vol. 184, pp. 425~ 430, (1998) [Olsson et al. 2004] Olsson, D.D., Bay, N., and Andreasen, J.L., “Prediction of Limits of Lubrication in Strip Reduction Testing,”, Annals of CIRP, Vol. 53/1, pp. 231-234, (2004) [Olsson 2006] Olsson, M., “Tribology in sheet metal forming and the tole of surface engineering”, The Altan Seminar on Metal Forming, Dalarna University, Borlänge, Dec. 4-5 (2006) [Orowan 1943] Orowan, E., “The calculation of toll pressure in hot and cold flat rolling”, Proceeding of Institutional Engineers Symposium, Vol.67, pp.140-150 (1943) [Pfestorf et al. 1998] Pfestorf, M., Engel, U. and Geiger, M., ”3D-Surface Parameters and Their Application on Deterministic Textured Metal Sheets” International Journal of Machine Tools & Manufacture, Vol.38, pp.607-614, (1998)

155

[Pfestorf 2005] Pfestdorf, M., “The Need for Multi-Phase Steel for the Body- in- White of a Passenger Car”, Great Designs in Steel Seminar, American Iron and Steel Institute (2005) [Podgornik 2004A] Podgornik, B., Hogmark, S., and Pezdirnik, J., “Comparison between different test methods for evaluation of galling properties of surface engineered tool surfaces,” Wear, Vol. 257, pp.843-851, (2004) [Podgornik 2004B] Podgornik, B., Hogmark, S., Sandberg, O., “Influence of surface roughness and coating type on the galling properties of coated forming tool steel”, Surface Coating Technologies, Vol.184, pp. 338- 348, (2004) [Rooji et al. 2001A] Rooij, M.B. and Schipper, D.J., “Analysis of Material transfer from a soft workpiece to a hard tool: Part I – Lump Growth Model,” Journal of Tribology, Vol. 123, pp.469-473, (2001). [Rooji et al. 2001B] Rooji, M.B., Schipper, D.J., “Analysis of material transfer from soft workpiece to a hard tool: Part – II – Experimental verification of the proposed lump growth model”, Transactions of ASME, Vol.123, pp.474-478, (2001) [Sandberg et al. 2004] Sanderberg, O., Johansson, B., “Tool steels for blankng and forming – new developments”, Proceedings of International conference on recent advances in manufacture & use of tools & dies and stamping of steel sheets, pp.53–70 (2004) [Sato et al. 1998] Sato, T. and Besshi, T., “Anti-galling evaluation in aluminum sheet forming,” Journal of Materials Processing Technology, Vol. 83, pp.185-191, (1998) [Sato et al. 2000] Sato, T., Besshi, T., Tsutsui, I., Morimoto, T., “Anti-galling property of a diamond like carbon coated tool in aluminum sheet forming”, Journal of Materials Processing Technology, Vol. 104, pp.21-24, (2000) [Schey 1983] Schey, J., “Tribolofy in Metalworking: Lubrication, Friction and Wear”, American Society for Metal, Metal Park, Ohio (1983) [Schedin et al. 1993] Schedin, E., Lehtinen, B., “Galling mechanisms in lubricated systems: a study of sheet metal forming”, Wear, Vol.170, pp.119- 130, (1993) [Shih et al. 2004] Shih, H. C., Belanger, P. J., Militisky, M., “Effects of Coating and Forming Condition on Friction”, Galvatech Conference, (2004) [Shimizu et al. 2001] Shimizu, Andreasen, Bech and Bay, “Influence of Workpiece Surface Topography on the Mechanisms of Liquid Lubrication in Strip Drawing. Transactions of the ASME, Vol.123, pp.290-294 (2001)

156

[Skare et al. 2003] Skare, T., and Krantz, F., “Wear and frictional behavior of high strength steel in stamping monitored by acoustic emission technique,” Wear, Vol.255, pp.1471-1479, (2003) [Sorensen et al. 1999] Sorenson, Bech, Andreason, Bay, Engel and Neudecker, “A Basic Study of the Influence of Surface Topography on Mechanisms of Liquid Lubrication in Metal Forming”, Annals of the CIRP, Vol.48, pp.203-208, (1999) [Sreenivasan et al.2006] Sreenivasan, N., Xia, M., Li, X., Lawson, S. and Zhou, N., “Improving Weldability of High Strength Galvanized Steel Sheet” US-GAP Program Review Meeting, Columbus, Ohio, US. Oct. (2006) [Sutcliffe et al. 2001] Sutcliffe, Le and Ahmed, “Modeling of Micro-Pit Evolution in Rolling or Strip-Drawing”, Journal of Tribology, Vol.123, pp. 791- 798, (2001) [Tolazzi et al. 2005] Tolazzi, M., Meiler, M., Merklein, M., “Tribological investigations on coated steel sheets using the dry film lubricant Drylube E1”, Advanced Materials Research, Vol. 6-8, pp. 565-572 (2005) [Vermeulen et al. 2001] Vermeulen, M., Scheers, J., “Micro-hydrodynamic effects in EBT textured steel sheet”, International Journal for Machine Tools and Manufacture, Vol.41, pp.1941-1951 (2001) [VOEST-ALPINE 2006] A private communication with Mr. Josef Faderl, VOEST- ALPINE, Austria (2006) [ULSAB, 2002] Ultra-Light Steel Auto Body (ULSAB), “Overview Report: ULSAB-AVC (Advanced Vehicle Concepts)” ULSAB_AVC Consortium (2002) [Wagner et al. 2002] Wagner, S., Kleinert, H., Zimmermann, R., “Dry Film Lubricants for Sheet Metal Forming,” Proceedings of International Conference of New Developments in Sheet Forming, University of Stuttgart, Germany, 2002, pp. 451-472 (2002) [Wanheim et al. 1974] Wanheim, Bay and Peterson, “A theoretically determined model for friction in metal working processes”, Wear, Vol.28, pp. 251- 258 (1974) [Wichern et al. 1999] Wichern, C.M., Van Tyne, C.J., “Frictional behavior of the sliding interface between an A2 steel die and zinc coated steel sheet”, Journal of Material Engineering and Performance, Vol.8, pp.571-576 (1999) [Wichelborg et al. 2000] Wichelborg, A., and Gunnarsson, L., “A frictional study of uncoated EBT steel sheets in bending under tension friction test”, Wear, Vol.237, pp.129-136, (2000)

157

[Wilson et al. 1995] Wilson, W., Hsu, T.C., and Huang, X.B., ”A realistic frictionmodel for computer simulation of sheet metal forming process”, ASME Transactions – Journal of Engineering for Industry, Vol. 117, pp. 202-209 (1995) [Wilson 2004] Wilson, W., “Modeling Tribology in Computer Simulations of Forming Processes”, the ICMTP Conference, Denmark (2004)

158

APPENDIX A -

TWIST COMPRESSION TEST MACHINE

The main components of the TCT machine at IRMCO in Evanston, IL, are shown in Figure A.1.

Figure A.1: TCT machine at IRMCO

The hydraulic motor is used to rotate the annual tool, Figure A.1. The hydraulic cylinder under the bottom plate is used to raise the workpiece against the tool. The pressure transducer is used to measure the pressure in the hydraulic cylinder. The tool-workpiece interface pressure can be easily calculated by using the measured pressure in the hydraulic cylinder and the ratio of cylinder and tool areas. The torque transducer is used to measure the torque

159

transmitted from the annual tool to the workpiece, Figure A.1. The torque transmitted from the tool to the workpiece causes the entire workpiece holder to try to rotate. However, because the locking device (which is not visible from outside view) is linked to the bottom of workpiece holder, the workpiece holder cannot actually rotate. Instead, a torque is transmitted from the workpiece holder to the locking device and it is measured by the torque transducer. The data acquisition system is used to acquire and compute the data from the pressure transducer and the torque transducer, Figure A.1.

160

APPENDIX B -

DEEP DRAWING TEST TOOLING

Details of schematic and tooling geometry are given in Figure B.1 and Figure B.2. The blank holder force (BHF) can be controlled up to 100 tons by the CNC hydraulic cushion pins.

Figure B.1: Deep Drawing Tooling at CPF-OSU

161

Figure B.2: Dimensions of Deep Drawing Tooling

Drawing die was properly polished to obtain the Ra value of 0.1 µm before the test as shown in Figure B.3.

Figure B.3: A polished deep drawing die before the test

162

APPENDIX C -

IRONING TEST TOOLING

Details of schematic and tooling geometry are given in Figure C.1 and Figure C. 2.

Figure C.1: Ironing tooling at ERC/NSM

Figure C. 2: Ironing ring die dimensions (all units are inch)

163

APPENDIX D -

VISCOUS PRESSURE BULGE (VPB) TEST

In order to conduct FE simulation of the deep drawing test, it is necessary to have the true stress-strain curve (i.e. the flow stress) of the sheet materials tested. Therefore, the viscous pressure bulge (VPB) test was used for the determination of flow stress for sheet material under biaxial state of stress. With the VPB test, larger strains, which are relevant for stamping operations, can be achieved compared to the standard tensile test.

Description of tooling and the test procedure

The ERC/NSM has developed a tooling for flow stress determination as shown in Figure D.1. This tooling provides the real time measurements of dome height and pressure. The VPB test uses a viscous medium to minimize the frictional effect of pressurizing medium and to easily handle instead of a liquid type medium. The tooling is designed for a double action hydraulic press. The upper die is connected to the slide of the press and the lower die is connected to the cushion of the press. The punch in the lower die is fixed with the press table. In the beginning of the test, the tooling is open and the blank sheet is placed between the upper and the lower dies. Then dies are closed to clamp the blank material and the slide moves down together with the entire die set. Consequently, the viscous medium is pressurized by the stationary punch and the sheet is bulged by the viscous medium flowing into the upper die. The sheet

164

is deformed until its forming limit is reached. The tooling has a lock-bead to ensure pure stretching without the flange material drawn into the die cavity. The press used for these experiments was a 160 ton hydraulic press with the CNC- controlled hydraulic cushion pins. The maximum blank holder force of the press is 100 ton. Details of VPB test are also available in [Gutscher et al. 2000].

Figure D.1: Schematic of the tooling used for the VPB test; (a) before the test (b) after the test with a bulged sample

Computation Procedure of Flow Stress

The membrane theory is used to determine the flow stress curve with the hydraulic bulge test, Figure D.2.

Figure D.2: Geometry of the sheet before and after the test

165

The effective stress, , and the effective strain, , can be calculated for the hydraulic bulge test using Equations (D.1) and (D.2).

(D.1)

(D.2)

Where εt = Strain at thickness, P = pressure, Rd = instantaneous radius of

curvature, td = instantaneous wall thickness at the apex of the dome, hd = instantaneous dome height, t0 = instantaneous sheet thickness, dc = die diameter,

Rc = Die corner radius, dsheet = Diameter of the sheet

From equations above it can be seen that four variables are needed to

determine the flow stress of a material: (a) instantaneous radius of curvature, Rd

(b) instantaneous wall thickness at the apex of the dome, td (c) instantaneous

dome height, hd (d) pressure, P.

The radius (Rd) and the thickness (td) are calculated as a function of the

dome height (hd) and the strain hardening exponent (n) by using a database, which was created by FEM. The procedure to determine the flow stress is shown in Figure D.3.

166

Figure D.3: Algorithm used to determine the flow stress curve from the VPB test

The database has been created using different n-values. It should be noted that a database for different n-values is only created once. Once created, the database can be used for any material. The strength coefficient (K) does not much influence the geometric variables. As shown in the flow chart, Figure D.3, the determination of flow stress starts by assigning an initial n-value. Then the flow stress curve and the K and n-values ( = K ) are calculated. This n-value is then used to calculate the thickness and the radius at the top of the dome and flow stress curve again. Iteration is continued until the difference between the new and the previous n-value is less than 0.001. Details of VPB test are also available in [Gutscher et al. 2000].

Experimental Results - Max. Dome Height

Various sheet materials were tested to compare the formability that is indicated by the maximum dome height without fracture. Four samples for each

167

material were tested. The first sample was tested until the sheet fractured at the bulge. Consequently the other samples were tested up to 5~10 % below the pressure measured at fracture and the maximum bulge height was measured. The test parameters used are shown in Table D.1.

Parameters

Test Duration Approx. 40 sec

Ram speed 1.0 mm/s

Clamping force 800 kN

Radius of fillet of cavity 6.4 mm

Size of the test sample 254 mm x 254 mm

Sheet Materials AKDQ (0.72), DP600 (1.0), DP590(1.24), DP980 (1.0), DP780 (1.0), TRIP780 (1.0) (Thickness in mm)

Table D.1: Parameters used in the viscous bulge test

The maximum dome height is compared for various sheet materials as shown in Figure D.4 and Figure D.5. With this comparison, the maximum value was obtained in AKDQ steel (34.1 mm) and the minimum value was obtained in DP980 steel (15.6 mm).

Figure D.4: Comparison of the dome height in tested samples of DP600, DP780 and DP980 (h = dome height)

168

Figure D.5: Comparison of the height in tested samples of AKDQ, DP590 and TRIP 780 (h = dome height)

The chart shown in Figure D.6 compares the flow stress curves for AKDQ steel, DP600, DP590, DP780, DP980 and TRIP 780 steels obtained by VPB tests. As shown in Figure D.6, this test is capable to reliably calculate the flow stress for two similar material grades, DP590 and DP600, provided by different suppliers. Both flow stress curves were found to be very similar from the VPB test.

Figure D.6: Comparison of the flow stress curves obtained by VPB test for AKDQ, DP590, DP600, DP780, DP980 and TRIP 780

169

APPENDIX E -

STRIP DRAWING TEST TOOLING

Detailed configuration of SDT tooling is given Figure E.1. Two inserts are used in the test with a strip of 356 mm long and 25.4 mm wide.

Figure E.1: 3-D model of SDT tooling

Insert die is assembled with two fixture wings and attached to a die holder. Fixture wing was designed for the assembly convenience. Detailed engineering drawings of these parts are given in Figure E.2, Figure E.3 and Figure E.4, individually.

170

171

Figure E.2: Engineering drawing of insert die

171

172

Figure E.3: Engineering drawing of a die holder

172

173

Figure E.4: Engineering drawing of a fixture wing

173

APPENDIX F -

STRIP DRAWING TEST RESULTS AND INPUT DATA FOR FEM

SIMULATIONS

Figure F.1: Load-stroke curves obtained from SDT for DP590 GA material with TiCN coated dies

174

Input data for simulation Workpiece Die and Punch Blank holder

Material type DP980 DC53 Tool steel P20 Tool steel

Object type Plastic Rigid Rigid

Time step size 0.2 mm/step (Total step No. = 400)

Blank holder force 25 KN

Heat transfer coefficient 11 KW / m2-K

Coefficient of Friction (COF) Selected in a range of COF = 0.05~0.14

Thermal conductivity 60.5 W / m-K 50.71 W / m-K 24.57 W / m-K

Heat Capacity 3.41 J / m3-K 3.81 J / m3-K 2.78 J / m3-K

Emissivity 0.95 0.7 0.7

Initial temperature (°C) 25

Table F.1: Input data used for strip drawing simulations

175

APPENDIX G -

LINEAR REGRESSION ANALYSIS RESULTS

In order to investigate the effect of viscosity level on the relationship between Galling Rank (GR) and Galling Severity Indicator (GSI), Eq. (G.1) was

formed to have the product of the Pressure (P), the Sliding length (Ls), the

Surface roughness (Ra) and the inverse of Hardness (H) as follows:

P × Ls × Ra GR = GSI = K × eff (G.1) H eff

For this analysis, our experimental data was fitted into the linear regression model given each viscosity level, V . Thus, our model will be in form of Eq. (G.2) as follows:

⎛ P × Ls × Ra ⎞ GR = K × ⎜ eff ⎟ + ε , where ε ~ N(0,σ 2 ) (G.2) V V ⎜ H ⎟ V V ⎝ eff ⎠V

For this statistical analysis, R program, a public statistical software was used. The outputs of regression analysis are summarized as follows:

176

• Linear regression results for the GI coated sheet

Viscosity level = 26

Coefficients: Estimate of K = 0.0085953 Std. Error = 0.0008388 t value = 10.25 Pr(>|t|) = 0.00198 Residual standard error = 0.5091 on 3 degrees of freedom Multiple R-Squared = 0.9722, Adjusted R-squared = 0.963 F-statistic = 105 on 1 and 3 DF, p-value = 0.001981

Viscosity level = 126

Coefficients: Estimate of K = 0.0014469 Std. Error = 0.0001529 t value = 9.465 Pr(>|t|) = 0.0110 Residual standard error = 0.3178 on 2 degrees of freedom Multiple R-Squared = 0.9782 Adjusted R-squared = 0.9672 F-statistic = 89.59 on 1 and 2 DF p-value = 0.01098

Viscosity level = 582

Coefficients: Estimate of K = 1.199e-03 Std. Error = 8.518e-05 t value = 14.07 Pr(>|t|) = 0.00501 Residual standard error = 0.215 on 2 degrees of freedom Multiple R-Squared = 0.99 Adjusted R-squared = 0.985 F-statistic = 198 on 1 and 2 DF p-value = 0.005012

From the above analysis, the regression coefficient is statistically significant on the all viscosity level = 26, 126, 582 at the 95% confidence level.

177

Viscosity level 26 126 562

Coefficient 0.008595345 0.001446872 0.001198628

Table G.1: Results of linear regression analyses for various lubricant viscosities with GI coating

Figure G.1: Relationship between K-values and lubricant viscosity with the GI coating

• Linear regression results for the GA coated sheet

Viscosity level = 58

Coefficients: Estimate of K = 6.649e-05 Std. Error = 1.507e-05 t value = 4.412 Pr(>|t|) = 0.142

178

Residual standard error = 0.03537 on 1 degrees of freedom Multiple R-Squared = 0.9511 Adjusted R-squared = 0.9023 F-statistic = 19.47 on 1 and 1 DF p-value = 0.1419

Viscosity level = 126

Coefficients: Estimate of K = 0.0013209 Std. Error = 0.0009608 t value = 1.375 Pr(>|t|) = 0.303 Residual standard error = 1.014 on 2 degrees of freedom Multiple R-Squared = 0.4859 Adjusted R-squared = 0.2288 F-statistic = 1.89 on 1 and 2 DF p-value = 0.3029

Viscosity level = 582

Coefficients: Estimate of K = 0.0008034 Std. Error = 0.0007032 t value = 1.142 Pr(>|t|) = 0.372 Residual standard error = 0.5501 on 2 degrees of freedom Multiple R-Squared = 0.3949 Adjusted R-squared = 0.09228 F-statistic = 1.305 on 1 and 2 DF p-value = 0.3716

From the above analysis, the regression coefficients are not statistically significant on the all viscosity level = 58, 126, 582 at the 90% confidence level.

Viscosty level 58 126 562

Coefficient 6.648869e-05 1.320888e-03 8.033592e-04

Table G.2: Results of linear regression analyses for various lubricant viscosities with GA coating

179

Figure G.2: Relationship between K-values and lubricant viscosity with the GA coating

180