Ultrasonic Vibration Assisted Manufacturing of High-Performance Materials

Home , Lausanne, Ning (surname)

Ultrasonic vibration assisted manufacturing of high-performance materials

Fuda Ning, B.S., M.S.

A Dissertation

Industrial, Manufacturing, and Systems Engineering

Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of

DOCTOR OF PHILOSOPHY

Approved

Dr. Weilong Cong Chair of Committee

Dr. Hong-Chao Zhang

Dr. Golden Kumar

Dr. George Zhuo Tan

Dr. Wei Li (Dean’s Representative)

Mark Sheridan Dean of the Graduate School

May, 2018

To My Family and Fleeting Time.

Texas Tech University, Fuda Ning, May 2018

ACKNOWLEDGMENTS First and foremost, I would like to express the deepest appreciation to my advisor, Dr. Weilong Cong, for his supervision and tremendous help throughout the past few years. He continually conveyed a rigorous attitude toward research and scholarship, which inspired me all the time. This dissertation would not have been completed without his professional guidance.

I would like to thank my committee members, Dr. Hong-Chao Zhang, Dr. Golden Kumar, and Dr. George Zhuo Tan, for their valuable advice on my dissertation. I also want to thank Dr. Wei Li to serve as the Dean’s Representative for my defense.

My sincere appreciation goes to the U.S. National Science Foundation for the financial support through award CMMI-1538381. I would like to extend my thanks to the Gradual School at Texas Tech University for the Doctoral Dissertation Completion Fellowship to enable me to devote full-time effort to finalizing this dissertation in the very last year of my Ph.D. study.

Special thanks to the members of Dr. Cong’s group, Mr. Yingbin Hu, Mr. Hui Wang, and Mr. Yuanchen Li. Working with all of you at Texas Tech University has been becoming an unforgettable experience.

This journey would not have been possible without the support of my family. I am especially grateful to my parents and my wife who always support me emotionally and encourage me in all of my pursuits.

iii

Texas Tech University, Fuda Ning, May 2018

TABLE OF CONTENTS ACKNOWLEDGMENTS ...... III

ABSTRACT ...... XII

LIST OF TABLES ...... XIV

LIST OF FIGURES ...... XV

CHAPTER I ...... 1 INTRODUCTION ...... 1 1.1 Background and motivation ...... 1 1.2 Research questions ...... 3 1.3 Research goal and objective ...... 5 1.4 Research methodology ...... 6 1.5 Significance of the research...... 7 1.6 Structures of this dissertation ...... 7 References ...... 8

CHAPTER II ...... 11

A LITERATURE REVIEW ON ULTRASONIC VIBRATION-ASSISTED (UV-

A) MANUFACTURING PROCESSES ...... 11 Abstract ...... 11 2.1 Background ...... 12 2.2 UV-A mechanical manufacturing processes ...... 13 2.2.1 UV-A conventional machining ...... 13 2.2.2 UV-A densification ...... 16 2.2.3 UV-A forming ...... 19 2.2.4 Ultrasonic consolidation...... 21 2.2.5 Remarks on the effects of ultrasonic vibration ...... 24 2.3 UV-A thermal manufacturing processes ...... 25 2.3.1 Nonlinear effects of ultrasonic vibration on liquid melting materials ... 25 2.3.2 UV-A thermal non-traditional machining ...... 27 2.3.3 UV-A casting ...... 29

Texas Tech University, Fuda Ning, May 2018

2.3.4 UV-A fusion welding ...... 33 2.3.5 UV-A laser cladding ...... 35 2.4 Concluding remarks ...... 37 References ...... 38

CHAPTER III ...... 46

ULTRASONIC VIBRATION-ASSISTED (UV-A) HOLE-MAKING OF CFRP

COMPOSITES: A COMPARISON WITH CONVENTIONAL GRINDING ...... 46 Abstract ...... 47 3.1 Introduction ...... 47 3.1.1 Properties and applications of CFRP composites ...... 47 3.1.2 Drilling in CFRP composites ...... 48 3.2 Experimental conditions and measurement procedures ...... 50 3.2.1 Properties of workpiece material ...... 50 3.2.2 Experimental set-up and conditions ...... 51 3.2.3 Measurement procedures for output variables ...... 53 3.3 Experimental results and discussions ...... 56 3.3.1 Effects on cutting force ...... 56 3.3.2 Effects on torque ...... 59 3.3.3 Effects on surface roughness ...... 60 3.3.4 Effects on hole diameter...... 62 3.3.5 Effects on material removal rate ...... 63 3.4 Conclusions ...... 65 References ...... 66

CHAPTER IV ...... 69

ULTRASONIC VIBRATION-ASSISTED (UV-A) HOLE-MAKING OF CFRP

COMPOSITES: DESIGN OF EXPERIMENT WITH A CUTTING FORCE MODEL ...... 69 Abstract ...... 70 4.1 Introduction ...... 70

Texas Tech University, Fuda Ning, May 2018

4.1.1 Drilling of CFRP composites ...... 70 4.1.2 Cutting force in RUM of CFRP composites ...... 71 4.1.3 Purpose of the chapter ...... 72 4.2 Mechanistic predictive model and design of experiments ...... 72 4.2.1 Mechanistic predictive model for cutting force in RUM of CFRP composites ...... 72 4.2.2 Experimental verification for mechanistic predictive model ...... 76 4.2.3 Design of Experiments ...... 77 4.3 Experimental results and discussions ...... 79 4.3.1 Main effects on cutting force ...... 79 4.3.2 Two-factor interaction effects on cutting force ...... 81 4.3.3 Three-factor interaction effects on cutting force ...... 82 4.4 Conclusions ...... 83 References ...... 84

CHAPTER V ...... 87

ULTRASONIC VIBRATION-ASSISTED (UV-A) HOLE-MAKING OF CFRP

COMPOSITES: A MECHANISTIC ULTRASONIC VIBRATION AMPLITUDE MODEL ...... 87 Abstract ...... 88 5.1 Introduction ...... 88 5.2 Microscope observation method ...... 90 5.3 Development of mechanistic amplitude calculation model ...... 92 5.3.1 Approach to amplitude calculation model development and notations . 92 5.3.2 Major assumptions in model development ...... 94 5.3.3 The relationship between cutting force and indentation depth ...... 94 5.3.4 The relationship between indentation depth and material removal rate 95 5.3.5 Ultrasonic vibration amplitude calculation model ...... 96 5.4 Experimental set-up and conditions ...... 97 5.4.1 Workpiece properties ...... 97

Texas Tech University, Fuda Ning, May 2018

5.4.2 Experimental set-up and cutting force measurement ...... 98 5.4.3 Cutting tool variables and machining variables ...... 100 5.5 Pilot experimental verification ...... 101 5.6 Conclusions ...... 104 References ...... 105

CHAPTER VI ...... 109

ULTRASONIC VIBRATION-ASSISTED (UV-A) SURFACE GRINDING OF

CFRP COMPOSITES: A COMPARISON WITH CONVENTIONAL SURFACE GRINDING ...... 109 Abstract ...... 110 6.1 Introduction ...... 110 6.2 Kinematic motion analysis ...... 112 6.3 Experimental set-up and conditions ...... 115 6.3.1 Workpiece material properties ...... 115 6.3.2 Experimental set-up and conditions ...... 115 6.3.3 Measurement procedures for output variables ...... 117 6.4 Results and discussion ...... 118 6.4.1 Effects on infeed cutting force ...... 118 6.4.2 Effects on axial cutting force ...... 119 6.4.3 Effects on torque ...... 121 6.4.4 Effects on surface roughness ...... 122 6.5 Conclusions ...... 123 References ...... 123

CHAPTER VII ...... 126

ULTRASONIC VIBRATION-ASSISTED (UV-A) SURFACE GRINDING OF

CFRP COMPOSITES: A MECHANISTIC MODEL ON CUTTING FORCE IN

THE FEED DIRECTION ...... 126 Abstract ...... 127 7.1 Introduction ...... 127

vii

Texas Tech University, Fuda Ning, May 2018

7.2 Model development ...... 131 7.2.1 Approach to model development ...... 131 7.2.2 Assumptions in model development ...... 132 7.2.3 The kinematics analysis of single abrasive grain in RUM surface grinding ...... 132 7.2.4 Relationship between indentation depth of abrasive grain into workpiece and radial grain force ...... 134 7.2.5 CFRP micromechanical analysis for obtaining elastic modulus and Poisson’s ratio...... 135 7.2.6 Material removal volume by one abrasive particle ...... 136 7.2.7 Material removal rate and radial grain force model ...... 139 7.2.8 The calculation of feed-direction cutting force ...... 140 7.3 Experimental set-up and conditions ...... 142 7.3.1 Workpiece material properties ...... 142 7.3.2 Experimental set-up and cutting force measurement ...... 143 7.3.3 Experimental conditions...... 144

7.4 Fracture volume factor KV acquisition and model verification ...... 145

7.4.1 Obtaining the value of fracture volume factor KV ...... 145 7.4.2 Model validity verification and predicted influences...... 146

7.4.3 Experimental validation of KV value ...... 149 7.5 Conclusions ...... 150 References ...... 151

CHAPTER VIII ...... 155

A FUNDAMENTAL INVESTIGATION ON ULTRASONIC VIBRATION-

ASSISTED (UV-A) LASER ENGINEERED NET SHAPING OF STAINLESS STEEL ...... 155 Abstract ...... 156 8.1 Introduction ...... 156 8.2 Experimental procedures ...... 159

viii

Texas Tech University, Fuda Ning, May 2018

8.2.1 Materials ...... 159 8.2.2 Experimental set-up and parameters ...... 160 8.2.3 Characterizations of geometries, microstructures, and properties ...... 162 8.3 Results and discussion ...... 163 8.3.1 Powder utilization efficiency ...... 164 8.3.2 Deposited shape geometry ...... 165 8.3.3 Surface roughness ...... 166 8.3.4 Geometry of molten pool and dilution zone ...... 167 8.3.5 Cross-sectional morphologies and microstructures ...... 168 8.3.6 Tensile properties ...... 171 8.3.7 Microhardness ...... 171 8.4 Conclusions ...... 173 Reference...... 174

CHAPTER IX ...... 177

MICROSTRUCTURES AND MECHANICAL PROPERTIES OF INCONEL 718

SUPERALLOY BULK PARTS FABRICATED BY ULTRASONIC

VIBRATION-ASSISTED (UV-A) LASER ENGINEERED NET SHAPING ..... 177 Abstract ...... 178 9.1 Introduction ...... 178 9.2 Experimental procedures ...... 181 9.2.1 Materials ...... 181 9.2.2 Experimental set-up and parameters ...... 181 9.2.3 Characterizations of microstructure and mechanical properties ...... 183 9.3 Results and discussion ...... 187 9.3.1 Porosity ...... 187 9.3.2 Grain microstructure ...... 189 9.3.3 Phase composition ...... 191 9.3.4 Precipitated phase morphology ...... 193 9.3.5 Tensile properties ...... 196

Texas Tech University, Fuda Ning, May 2018

9.3.6 Microhardness ...... 198 9.3.7 Dry sliding wear resistance ...... 199 9.4 Conclusions ...... 200 References ...... 201

CHAPTER X ...... 204

ULTRASONIC VIBRATION-ASSISTED LASER ENGINEERED NET

SHAPING OF INCONEL 718 SUPERALLOY BULK PARTS: EFFECTS OF

ULTRASONIC FREQUENCY ...... 204 Abstract ...... 205 10.1 Introduction ...... 205 10.2 Experimental procedures ...... 207 10.2.1 Materials ...... 207 10.2.2 Experimental set-up and parameters ...... 207 10.2.3 Characterizations of molten pool, porosity, and microstructures ...... 209 10.3 Results and discussion ...... 212 10.3.1 Effects on molten pool geometry ...... 212 10.3.2 Effects on temperature profiles within molten pool ...... 213 10.3.3 Effects on temperature fluctuation of single layer and whole parts ..... 215 10.3.4 Effects on porosity ...... 217 10.3.5 Effects on grain microstructures ...... 219 10.4 Conclusions ...... 220 References ...... 222

CHAPTER XI ...... 225

ULTRASONIC VIBRATION-ASSISTED LASER ENGINEERED NET

SHAPING OF TIB REINFORCED TI MATRIX COMPOSITES: EFFECTS ON

MICROSTRUCTURE AND MECHANICAL PROPERTY ...... 225 Abstract ...... 226 11.1 Introduction ...... 226 11.2 Experimental methods ...... 228

Texas Tech University, Fuda Ning, May 2018

11.2.1 Materials ...... 228 11.2.2 Experimental set-up and conditions ...... 230 11.2.3 Measuring procedures ...... 231 11.3 Experimental results ...... 233 11.3.1 Porosity and pore size ...... 233 11.3.2 TiB growth and primary TiB size ...... 235 11.3.3 TiB whisker distribution ...... 236 11.3.4 Grain size ...... 238 11.3.5 Microhardness ...... 240 11.4 Discussions ...... 241 11.5 Conclusions ...... 245 References ...... 246

CHAPTER XII ...... 250

CONCLUSIONS AND SCIENTIFIC CONTRIBUTIONS ...... 250 12.1 Conclusions ...... 250 12.2 Scientific contributions ...... 252

APPENDIX A ...... 254

PUBLICATIONS SINCE PH.D. STUDY ...... 254

Texas Tech University, Fuda Ning, May 2018

ABSTRACT High-performance materials are broadly applied in many key industries to meet the high requirements of material strength, light weight, corrosion resistance, high-temperature capability, functionality, etc. Manufacturing of high-performance materials is necessary to produce the end-use parts or components that can be effectively applied in industries. However, traditional manufacturing methods could generate various manufacturing problems, which would limit the broad applications of end-use parts.

As a type of high-performance materials, carbon fiber reinforced plastic (CFRP) composites have found remarkably increasing applications in the aerospace and automotive industries due to their superior properties. In order to manufacture the CFRP composites to final end-use parts, machining processes including hole making and surface grinding are always planned after the molding processes to generate features on the final parts. However, CFRP composites exhibit a poor machinability in the traditional machining processes, leading to low part quality and low manufacturing efficiency. As another type of high-performance materials, metal alloys such as nickel and titanium possess outstanding properties even at extremely high temperatures. For this reason, they have been extensively used to manufacture engine components in aerospace, defense, and marine industries. Recently, additive manufacturing of these alloys has gained numerous attention with the advantages of wasted material reduction and manufacturing efficiency improvement. However, fabrication defects and uncertain microstructures are inevitably induced in the additively manufactured metallic parts, which are greatly detrimental to the part qualities and mechanical properties.

Based on the problems during the manufacturing of those two types of high- performance materials as mentioned above, it is thereby of great significance to develop a high-quality and high-efficiency manufacturing technique to effectively reduce the issues. In recent years, ultrasonic vibration has attracted great interests in assisting numerous manufacturing processes with significant improvements in the

xii

Texas Tech University, Fuda Ning, May 2018 manufacturing process performances. Compared with other mechanical vibrations, ultrasonic vibration possesses a frequency that is much higher than the natural frequency of a system. Due to this reason, ultrasonic vibration can maintain or even improve the stability of the manufacturing system without adding harmful low- frequency vibrations. In this dissertation, ultrasonic vibration-assisted (UV-A) manufacturing processes will be thus conducted to seek the potential solutions for the aforementioned manufacturing problems.

In this dissertation, a comprehensive literature review on UV-A manufacturing processes is firstly conducted to provide fundamental knowledge on the mechanism of ultrasonic vibration’s actions in all the manufacturing processes. Then, endeavors have been made to facilitate the effective fabrication of the high-performance materials by introducing ultrasonic vibration as an assisted technique in the manufacturing processes. In particular, UV-A machining (hole making and surface grinding) of CFRP composites as well as UV-A laser engineered net shaping of stainless steel, nickel alloys, and titanium matrix composites are experimentally and theoretically investigated. The results show improvements in both process performance and manufactured part quality, which can be attributed to the remarkable influences of ultrasonic vibration. The investigations in this dissertation will help to establish a high-quality and high-efficiency process to improve the CFRP machinability and performance of additively manufactured metal and metal matrix composite parts. In addition, the fundamental understanding and knowledge generated in this dissertation will benefit the area of UV-A manufacturing of high-performance materials.

xiii

Texas Tech University, Fuda Ning, May 2018

LIST OF TABLES Table 3.1 Properties of workpiece material ...... 50 Table 3.2 Tool parameters ...... 53 Table 3.3 Machining conditions ...... 53 Table 4.1 Input variables in the cutting force model for CFRP composites ...... 73 Table 4.2 Experimental conditions for model verification ...... 76 Table 4.3 Low and high levels of input variables ...... 77 Table 4.4 Design matrix and experimental results ...... 78 Table 4.5 Analysis of variance for full factorial model ...... 79 Table 4.6 Three-factor interaction effects analysis ...... 83 Table 5.1 Input variables in model development ...... 93 Table 5.2 Properties of CFRP workpiece material ...... 98 Table 5.3 Cutting tool parameters ...... 100 Table 5.4 Experimental conditions...... 101 Table 6.1 Properties of CFRP workpiece material ...... 116 Table 6.2 Cutting tool variables ...... 117 Table 6.3 Experimental conditions...... 117 Table 7.1 Properties of CFRP workpiece material ...... 142 Table 7.2 Identifications of cutting tools in the experiments ...... 144 Table 7.3 Experimental conditions...... 145 Table 8.1 The UV-A LENS manufacturing parameters for AISI 630 thin wall deposition ...... 161 Table 9.1 The major chemical compositions of Inconel 718 powders...... 181 Table 9.2 The LENS manufacturing parameters for IN718 bulk part fabrications ...... 183 Table 9.3 The LENS manufacturing parameters for IN718 tensile specimen fabrications ...... 187 Table 10.1 The LENS manufacturing parameters for IN718 bulk part fabrications ...... 209 Table 11.1 Input fabrication variables for TiB-TMC part fabrication ...... 231

xiv

Texas Tech University, Fuda Ning, May 2018

LIST OF FIGURES Figure 2.1 Illustrations on UV-A turning process (after [6]) ...... 15 Figure 2.2 Illustrations on UV-A grinding processes ...... 16 Figure 2.3 Illustrations on UV-A densification mechanism in three stages (after [21]) ...... 17 Figure 2.4 Illustrations on UV-A pelleting of cellulosic biomass [30] ...... 18 Figure 2.5 Pellets produced by (a) pelleting without ultrasonic vibration and (b) UV-A pelleting [28] ...... 19 Figure 2.6 Procedures of UV-A hot embossing process (after [36]) ...... 20 Figure 2.7 Procedures of UV-A extrusion process at one ultrasonic vibration cycle (after [38]) ...... 21 Figure 2.8 Illustrations on ultrasonic consolidation process (after [45]) ...... 22 Figure 2.9 Fine grains around the interface regions between the first and second layers in the aluminum alloy parts fabricated by ultrasonic consolidation (after [50]) ...... 23 Figure 2.10 Optical micrograph of aluminum alloy parts fabricated by ultrasonic consolidation under (a) the smaller ultrasonic amplitude and (b) the larger ultrasonic amplitude (after [54]) ...... 23 Figure 2.11 Actions and influences of ultrasonic vibration in UV-A mechanical manufacturing processes. (UC is ultrasonic consolidation.) ...... 25 Figure 2.12 Actions and influences of ultrasonic vibration in UV-A thermal manufacturing processes (after [4]) ...... 27 Figure 2.13 Illustrations on UV-A EDM system set-up (after [64]) ...... 29 Figure 2.14 Machined surfaces of parts fabricated by (a) EDM process and (b) UV-A EDM process (after [63]) ...... 29 Figure 2.15 Ultrasonic vibration imported for solidifying melts during casting process (after [74]) ...... 30 Figure 2.16 Microstructures of aluminum alloy A356 parts fabricated (a) without and (b) with ultrasonic vibration (after [75]) ...... 31 Figure 2.17 Interactions between cavitation bubbles and dendrite (a) before fragmentation and (b) after fragmentation (after [80]) ...... 32

Texas Tech University, Fuda Ning, May 2018

Figure 2.18 Tensile properties of magnesium alloy AZ91 parts made by the casting with and without ultrasonic vibration (after [70]) ...... 33 Figure 2.19 Illustrations on the ultrasonic vibration implementation in UV-A arc welding system (after [83-85]) ...... 34 Figure 2.20 Weld appearance on the cross-section of AISI 304 sample (after [83]) ...... 35 Figure 2.21 Schematic of UV-A laser cladding for coating fabrication (after [92]) ...... 36 Figure 2.22 Online temperature measurement in (a) laser cladding without ultrasonic vibration and (b) UV-A laser cladding (after [97]) ...... 37 Figure 3.1 Illustration of rotary ultrasonic machining ...... 49 Figure 3.2 Illustration of Fiber structures in CFRP ...... 50 Figure 3.3 Rotary ultrasonic machining (RUM) experimental set-up ...... 51 Figure 3.4 Illustration of the tool used in both RUM and grinding ...... 52 Figure 3.5 Typical relationship between cutting force and time (in RUM) ...... 54 Figure 3.6 Illustration of the machined hole and rod ...... 55 Figure 3.7 Cutting force comparison between RUM and grinding when tool rotation speed changed ...... 57 Figure 3.8 Cutting force comparison between RUM and grinding when feed rate changed ...... 58 Figure 3.9 Torque comparison between RUM and grinding when tool rotation speed changed ...... 59 Figure 3.10 Torque comparison between RUM and grinding when feed rate changed ...... 60 Figure 3.11 Surface roughness comparison between RUM and grinding when tool rotation speed changed ...... 61 Figure 3.12 Surface roughness comparison between RUM and grinding when feed rate changed ...... 62 Figure 3.13 Hole diameter comparison between RUM and grinding when tool rotation speed changed ...... 63 Figure 3.14 Hole diameter comparison between RUM and grinding when feed rate changed ...... 64

xvi

Texas Tech University, Fuda Ning, May 2018

Figure 3.15 Material removal rate comparison between RUM and grinding when tool rotation speed changed ...... 64 Figure 3.16 Material removal rate comparison between RUM and grinding when feed rate changed ...... 65 Figure 4.1 Illustration of rotary ultrasonic machining ...... 71 Figure 4.2 Model developing processes ...... 73 Figure 4.3 Material removal mechanism by one abrasive partial ...... 75 Figure 4.4 Comparisons of predicted and experimental results ...... 77 Figure 4.5 Main effects of variables on cutting force ...... 80 Figure 4.6 Two-factor interaction effects of variables on cutting force ...... 82 Figure 4.7 The best combination for making the lowest cutting force at a three-factor interaction ...... 83 Figure 5.1 Illustration of RUM ...... 89 Figure 5.2 Illustrations of machined rod and hole ...... 91 Figure 5.3 Measurement of ultrasonic vibration amplitude on a microscopic picture in RUM of Ti ...... 91 Figure 5.4 Machined rod surfaces of different materials ...... 92 Figure 5.5 Amplitude calculation model development procedures ...... 93 Figure 5.6 RUM system set-up ...... 99

Figure 5.7 Typical curve of cutting force Fz versus cutting time in RUM process (a. The tool was feeding into the workpiece; b. The tool fully fed into the workpiece; and c. The tool was leaving the workpiece.) ...... 100 Figure 5.8 Illustration of cutting tool ...... 101 Figure 5.9 Ultrasonic vibration amplitude measurement (a. Illustration of the designed CFRP/Al stack; b. The manufactured stack after RUM process; and c. The morphology of the machined Al surface) ...... 103 Figure 5.10 Comparisons between calculated model results and experimental results of ultrasonic vibration amplitude ...... 104 Figure 6.1 Illustration of rotary ultrasonic surface machining (RUSM) process ...... 111

xvii

Texas Tech University, Fuda Ning, May 2018

Figure 6.2 The trajectory of a single abrasive grain in (a) RUSM and (b) conventional surface grinding (CSG) in the certain period ...... 114 Figure 6.3 RUSM system set-up ...... 117 Figure 6.4 Measurements of surface roughness ...... 118 Figure 6.5 Comparisons of infeed cutting force between RUSM and CSG ...... 120 Figure 6.6 Comparisons of axial cutting force between RUSM and CSG ...... 120 Figure 6.7 Comparisons of torque between RUSM and CSG ...... 122 Figure 6.8 Comparisons of surface roughness between RUSM and CSG ...... 123 Figure 7.1 Illustration of rotary ultrasonic machining (RUM) surface grinding process ...... 129 Figure 7.2 Feed-direction cutting force modeling procedures ...... 131 Figure 7.3 The kinematic trajectory of a single abrasive grain in RUM surface grinding process ...... 133 Figure 7.4 Micromechanics analysis of CFRP workpiece ...... 135 Figure 7.5 Material removal volume by a single abrasive grain in one ultrasonic vibration cycle ...... 137

Figure 7.6 Relationship between grain force Fg and feed-direction force Fx in RUM surface grinding process ...... 141 Figure 7.7 Calculation methodology for feed-direction cutting force Fx ...... 142 Figure 7.8 RUM surface grinding set-up ...... 144 Figure 7.9 Comparisons of predicted and experimental feed- direction cutting forces under different machining variables ...... 147 Figure 7.10 Relationships between ultrasonic vibration amplitude A and indentation depth δ ...... 149 Figure 7.11 Comparisons of predicted and experimental feed- direction cutting forces under different tool variables...... 149

Figure 7.12 Fracture volume factor Kv values under different experimental conditions ...... 150 Figure 8.1 Laser engineered net shaping (LENS) process ...... 158

xviii

Texas Tech University, Fuda Ning, May 2018

Figure 8.2 AISI 630 powder morphology ...... 160 Figure 8.3 Schematic of ultrasonic vibration-assisted LENS system set-up ...... 161 Figure 8.4 Effects of ultrasonic vibration on powder utilization efficiency during 5-layer thin walls deposition ...... 164 Figure 8.5 Comparisons on (a) length, (b) width, (c) height, and (d) flatness of 5-layer thin walls between LENS with (W) and without (W/O) ultrasonic vibration ...... 165 Figure 8.6 Effects of ultrasonic vibration on surface roughness during single-layer and 5-layer thin walls deposition ...... 167 Figure 8.7 Effects of ultrasonic vibration on (a) molten pool width, (b) dilution zone depth, and (c) heat affected zone depth during 5-layer thin walls deposition ...... 168 Figure 8.8 Cross-sectional morphologies and microstructures at different locations of AISI 630 samples parts fabricated by LENS without and with ultrasonic vibration ...... 170 Figure 8.9 Tensile properties of the AISI 630 samples fabricated by LENS without and with ultrasonic vibration (UV) ...... 171 Figure 8.10 Microhardness of the AISI 630 parts fabricated by LENS without and with ultrasonic vibration (UV) ...... 172 Figure 9.1 Inconel 718 powder morphology ...... 181 Figure 9.2 Schematic of ultrasonic vibration-assisted LENS system set-up ...... 183 Figure 9.3 (a) SEM image of porosity of fabricated IN718 parts and (b) porosity analysis/calculation using image processing technique ...... 185 Figure 9.4 The intercept method for grain size measurement based on the ASTM E112-13 standard ...... 185 Figure 9.5 Designed tensile specimen based on ASTM E8 standard ((a) dimensions of tensile specimen and (b) tensile samples fabricated by LENS) ...... 186 Figure 9.6 (a) Processed SEM images for calculating porosity and (b) porosity values under different fabrication conditions ...... 188 Figure 9.7 (a) Microstructural characteristics and (b) grain size evolution under different processing conditions ...... 190

xix

Texas Tech University, Fuda Ning, May 2018

Figure 9.8 EDS analysis results of IN718 parts fabricated by LENS (a) without ultrasonic vibration and (b) with ultrasonic vibration at the laser power of 270 W ...... 192 Figure 9.9 EDS analysis results of IN718 parts fabricated by LENS (a) without ultrasonic vibration and (b) with ultrasonic vibration at the laser power of 350 W ...... 193 Figure 9.10 The phase morphology at different sections of IN 718 parts fabricated by LENS (a) without ultrasonic vibration and (b) with ultrasonic vibration at laser power of 270 W ...... 194 Figure 9.11 The phase morphology at different sections of IN718 parts fabricated by LENS (a) without ultrasonic vibration and (b) with ultrasonic vibration at laser power of 350 W ...... 195 Figure 9.12 Comparisons on (a) stress-strain curve, (b) yield strength, (c) ultimate tensile strength (UTS), and (d) ductility of LENS-fabricated IN718 parts under different processing conditions...... 197 Figure 9.13 Effects of grain size on yield strength and UTS ...... 197 Figure 9.14 Vickers microhardness measurement on the transverse surface of the fabricated parts ...... 198 Figure 9.15 A comparison on the width of IN718 parts after dry sliding under different fabrication conditions ...... 200 Figure 10.1 Schematic of ultrasonic vibration-assisted LENS system set-up ...... 208 Figure 10.2 Methodology to generate the temperature profile of the molten pool ...... 210 Figure 10.3 Measurement of molten pool length L and width W through the thermal image ...... 210 Figure 10.4 (a) OM image of porosity of fabricated IN718 parts and (b) porosity analysis/calculation using image processing technique ...... 211 Figure 10.5 Effects of ultrasonic vibration frequency on the (a) length and (b) width of the molten pool ...... 213 Figure 10.6 Effects of ultrasonic vibration frequency on temperature profiles at various positions along the molten pool surface ...... 215 Figure 10.7 Effects of ultrasonic vibration frequency ((a) 0 kHz, (b) 25 kHz, (c) 33 kHz, and (d) 41 kHz) on the profile of peak temperature values within a single layer ...... 216

Texas Tech University, Fuda Ning, May 2018

Figure 10.8 Effects of ultrasonic vibration frequency ((a) 0 kHz, (b) 25 kHz, (c) 33 kHz, and (d) 41 kHz) on the profile of peak temperature values within the four-layer bulk parts ...... 217 Figure 10.9 OM images for calculating porosity of LENS-fabricated IN718 parts under ultrasonic vibration frequencies of (a) 0 kHz, (b) 25 kHz, (c) 33 kHz, and (d) 41 kHz ...... 218 Figure 10.10 Effects of ultrasonic vibration frequency on the porosity of LENS-fabricated IN718 parts ...... 219 Figure 10.11 Grain microstructures of LENS-fabricated IN718 parts under ultrasonic vibration frequencies of (a) 0 kHz, (b) 25 kHz, (c) 33 kHz, and (d) 41 kHz ...... 221 Figure 10.12 Effects of ultrasonic vibration frequency on the grain size value of LENS-fabricated IN718 parts ...... 222 Figure 11.1 Morphologies of as-received (a) CP-Ti powder, (b) B powder, and four-hour ball milling mixed CP-Ti and B powders at (c) lower and (d) higher magnifications, respectively ...... 229 Figure 11.2 Schematic of ultrasonic vibration-assisted LENS system set-up ...... 230 Figure 11.3 (a) SEM image of porosity of fabricated TiB-TMC parts and (b) porosity analysis/calculation using image processing technique ...... 232 Figure 11.4 The measurement procedures for grain size of TiB-TMC parts ...... 233 Figure 11.5 Porosity values under different processing conditions (W/O: without ultrasonic vibration; W: with ultrasonic vibration) ...... 234 Figure 11.6 The variations of pore sizes of TiB-TMC parts fabricated at a laser power of (a) 200 W and (b) 300 W ...... 235 Figure 11.7 (a) Illustration of two types of TiB whiskers and (b) comparison of primary TiB whisker size among different processing conditions ...... 236 Figure 11.8 Distribution of TiB within Ti matrix under processing conditions of (a) 200 W/O, (b) 200 W, (c) 300 W/O, and (d) 300 W ...... 237 Figure 11.9 EDS mapping results of TiB-TMC parts fabricated by different processing conditions ...... 238

xxi

Texas Tech University, Fuda Ning, May 2018

Figure 11.10 QCN microstructural characteristics under different processing conditions ...... 239 Figure 11.11 Comparisons of grain size among different processing conditions...... 240 Figure 11.12 Comparison of microhardness on the transverse surface of TiB-TMC parts fabricated under different conditions ...... 241

xxii

Texas Tech University, Fuda Ning, May 2018

CHAPTER I

INTRODUCTION

1.1 Background and motivation In recent years, demand for high-performance materials is on a significant rise worldwide. The high-performance materials are the ones those are introduced into many key industries to meet the higher requirements in the areas of material strength, light weight, corrosion resistance, high-temperature capability, functionality, etc. [1, 2]. The typical high-performance materials include composites (such as fiber reinforced polymers, ceramic reinforced metals, etc.), metal alloys (such as nickel, titanium, etc.), multi-functional materials (such as fabrics combined with electronics, smart materials, etc.), and nanomaterials (nanotubes or nanoparticles) [1, 2].

Manufacturing of high-performance materials is necessary to produce the end- use parts or components that can be effectively applied in industries. However, traditional manufacturing methods could generate various manufacturing problems, which would limit the broad applications of end-use parts [3]. For example, as a type of high-performance materials, carbon fiber reinforced plastic (CFRP) composites have found remarkably increasing applications in the aerospace and automotive industries due to the superior properties, such as high stiffness-to-weight ratio, excellent fatigue and wear resistance, high dimensional stability, etc. [3, 4]. In order to fabricate the CFRP composites to near-net-shape components, molding processes are usually applied. Additional machining processes including hole making and surface grinding are further required to generate features on the final parts [5]. Specifically, hole making of CFRP components is commonly desired for the assembly purpose in the aircraft industry, while surface grinding of CFRP components is necessary to produce functional surfaces on the parts applied in the automotive industry. However, CFRP composites exhibit a poor machinability in the traditional machining processes due to the abrasiveness of carbon fiber and heterogeneity of the materials, which could remarkably reduce the load-bearing capability and service life of end-use CFRP parts

1 Texas Tech University, Fuda Ning, May 2018

[6]. Such problems result in low part quality and low manufacturing efficiency, providing great challenges to the industry. As another type of high-performance materials, high-value nickel and titanium alloys possess outstanding corrosion resistance, high fatigue strength, and excellent oxidation resistance even at extremely high temperatures [7]. With this reason, these metal alloys are the most attractive candidates to manufacture engine components such as turbine blades, rotors, exhaust systems, etc. for the applications in aerospace, defense, and marine industries [8, 9]. Most of these end-use components own complex structures, indicating a large amount of materials would be certainly removed in subtractive manufacturing processes. Since nickel and titanium alloys exhibit high cost and poor machinability, subtractive manufacturing of these alloys for the final part production is not a cost-effective method [10]. Recently, additive manufacturing of nickel and titanium alloys has gained numerous attention due to the reduction of wasted materials, the decrease of manufacturing costs, and the improvement of manufacturing efficiency [11-13]. However, fabrication defects and uncertain microstructures are inevitably induced in the additively manufactured metallic parts, which are greatly detrimental to the part qualities and mechanical properties. Based on the problems during the manufacturing of those two types of high-performance materials as mentioned above, it is thereby of great significance to develop a high-quality and high-efficiency manufacturing technique to effectively reduce the issues.

Ultrasonic vibration has attracted great interests in assisting numerous manufacturing processes at a relatively low equipment cost [14]. Such assisted technique has been strongly proven to exert significant improvements on the manufacturing process performances. Ultrasonic vibration is a physical wave that propagates at frequencies of more than 20 kHz, which is beyond the conventional upper limit of human hearing. Compared with other mechanical vibrations, ultrasonic vibration possesses a frequency that is much higher than the natural frequency of a system. Due to this reason, ultrasonic vibration can maintain or even improve the stability of the manufacturing system without adding harmful low-frequency

2 Texas Tech University, Fuda Ning, May 2018 vibrations. To realize the effective supply of ultrasonic vibration to the materials being manufactured, a reliable contact between an ultrasonic generator and the materials is always achieved by utilizing a medium such as a horn/tool holder assembly, an ultrasound probe, an ultrasonic vibration platform, etc. [14]. Ultrasonic vibration can be transferred through solid-phase or liquid-phase materials with its extraordinary properties. Based on the material types during the transmission of ultrasonic energy, ultrasonic vibration-assisted (UV-A) manufacturing processes can be divided into two main categories, including UV-A mechanical manufacturing (such as machining, densification, forming, consolidation, etc.) and UV-A thermal manufacturing (such as thermal non-traditional machining, casting, fusion welding, laser cladding, direct laser deposition, etc.). In this dissertation, UV-A machining and UV-A direct laser deposition in each category will be conducted to seek the potential solutions for the aforementioned manufacturing problems of CFRP and metal alloys, respectively.

1.2 Research questions This dissertation will consist of two main parts. In the first part, UV-A hole making and surface grinding of CFRP will be investigated. Such technique is proposed with the purpose of reducing the problems in traditional machining of CFRP, such as large cutting force, high tool wear, bad surface quality, etc. [15]. Thus, the first research question is what the process improvements will be as compared with the process performance in traditional hole making (using core grinding) and surface grinding of CFRP.

In UV-A hole making of CFRP, cutting force is one of the most important output variables, which is directly related to the tool wear and surface quality. Experimental investigations have been carried out to study the effects of input variables on the cutting force [16-18]. However, the two-factor and three-factor interaction effects of input variables on cutting force are still unclear. The second research question is what the statistical significance of input variables on cutting force will be in UV-A hole making of CFRP.

3 Texas Tech University, Fuda Ning, May 2018

In UV-A hole making of CFRP, ultrasonic vibration amplitude is an intermediate variable that has direct influences on almost all output variables [19, 20]. The measurement of ultrasonic vibration amplitude would be thus beneficial to predict the results of output responses. However, there is a lack of methods to measure ultrasonic vibration amplitude in UV-A hole making of CFRP. Developing a mechanistic model to investigate the ultrasonic vibration amplitude turns out to be crucial. Hence, the third research question is how to effectively and accurately establish such a mechanistic model to reveal the relationships between input variables and ultrasonic vibration amplitude in UV-A hole making of CFRP.

The mechanistic model of axial cutting force in UV-A hole making of CFRP has been developed by analyzing the material removal mechanism [21]. In UV-A surface grinding of CFRP, however, the dominated component of cutting force is the feed-direction cutting force and the material removal mechanism is different with that in UV-A hole making of CFRP due to the kinematic differences. Thus, the fourth research question is how feed-direction cutting force will be generated and what the material removal mechanism will be in UV-A surface grinding of CFRP.

On the other hand, UV-A direct laser deposition (using laser engineering net shaping (LENS) process) of metal alloys and their composites will be demonstrated in the second part. LENS can be considered as a thermal manufacturing process since it involves material melting and solidification behaviors. It has been reported that the direct implementation of ultrasonic vibration in thermal manufacturing processes (such as casting, fusion welding, etc.) can induce nonlinear effects including acoustic streaming and cavitation [14]. Owing to these effects, ultrasonic vibration enables to reduce porosity, refine the microstructure, and increase the material homogeneity in the fabricated parts [22]. Thus, the fifth research question is whether or not ultrasonic vibration can reduce fabrication defects and improve mechanical properties of the parts by affecting the material melting and solidification behaviors in LENS of metal alloys.

Ultrasonic frequency is a very important variable of the ultrasonic vibration,

4 Texas Tech University, Fuda Ning, May 2018 which will affect the ultrasonic vibration performance in assisting LENS processes. Currently, the available literature on the effects of ultrasonic frequency on the performance of thermal manufacturing are very scarce, which may be associated with the difficulties of ultrasonic frequency adjustment especially in the LENS process. The importance of investigations on ultrasonic frequency effects has been thereby overlooked. The sixth research question will be what the effects of ultrasonic vibration frequency on molten pool geometry, porosity, and grain microstructures are in UV-A LENS process.

In addition, UV-A LENS of titanium matrix composites will be presented. Investigations on LENS of titanium matrix composites have shown that internal weaknesses such as coarse geometry and heterogeneous distribution of the reinforcement have been inevitably induced, lowering the mechanical properties of the fabricated parts [23, 24]. These weaknesses are difficult to be simultaneously alleviated simply by conducting parametric studies on the LENS process or post heat treatment of the parts [23]. Therefore, the seventh research question is whether or not ultrasonic vibration can effectively improve the characteristics of the reinforcement in LENS of titanium matrix composites.

1.3 Research goal and objective The research goal of this dissertation is to generate fundamental knowledge and provide a significant advance in UV-A subtractive and additive manufacturing of high-performance metal and composite parts. In pursuit of this research goal, the objective of this dissertation is to identify the improvements of both process performance and part quality by integrating ultrasonic vibration in machining of CFRP and in LENS of metal alloys and composites. Specifically, nine research tasks driven by the seven research questions mentioned above will be performed in this dissertation, with the purpose of

(1) evaluating the improvement of process performance by applying ultrasonic vibration in hole making and surface grinding of CFRP (in two tasks).

5 Texas Tech University, Fuda Ning, May 2018

(2) identifying the significance of input variables on cutting force and determining the controllable input variables in UV-A hole making of CFRP.

(3) explaining the relationships between input variables and ultrasonic vibration amplitude and providing a guide for building models to predict output variables in UV-A hole making of CFRP.

(4) understanding material removal mechanism and predicting the feed- direction cutting force for effectively controlling the occurrence of surface damages in UV-A surface grinding of CFRP.

(5) exploring the mechanism of ultrasonic vibration’s actions during the melting and solidification process in UV-A LENS of metal alloys (in two tasks).

(6) investigating the effects of ultrasonic vibration frequency on molten pool geometry, porosity, and grain microstructures in UV-A LENS of metal alloys.

(7) seeking the influences of ultrasonic vibration on the reinforcement behaviors in UV-A LENS of composites.

1.4 Research methodology In this dissertation, theoretical and experimental investigations will be conducted to study the mechanism and phenomena occurred in the UV-A manufacturing of high-performance materials. In particular, mechanistic models will be developed to predict the intermediate and output variables in UV-A hole making and surface grinding of CFRP. Such theoretical investigation will be beneficial to provide the explanations for some experimentally observed phenomena. On the other hand, experimental studies will include comparative study, the design of experiment, process control, microstructural analysis through optical microscopy and scanning electron microscopy, and testing of physical and mechanical properties. For the experimental results, quantitative data analysis will be performed to answer these research questions.

6 Texas Tech University, Fuda Ning, May 2018

1.5 Significance of the research The obtained knowledge in this dissertation will fill the research gaps in the literature on UV-A hole making and surface grinding of CFRP as well as UV-A LENS of metal alloys and metal matrix composites. In UV-A hole making and surface grinding of CFRP, the obtained results will provide solutions to the existing problems. In UV-A LENS of metal alloys and metal matrix composites, the understanding of ultrasonic vibration effects on molten pool features, microstructural evolution, mechanical performance, and part quality will be generated.

The investigations in this dissertation will help to establish a high-quality and high-efficiency process to improve the CFRP machinability and performance of additively manufactured metal and metal matrix composite parts. The results will also benefit subtractive and additive manufacturing of high-performance materials utilized in the key industries.

1.6 Structures of this dissertation This dissertation consists of twelve chapters. After this introductory chapter, a literature review on ultrasonic vibration-assisted (UV-A) manufacturing processes will be presented in Chapter II which summarizes fundamental knowledge on the mechanism of ultrasonic vibration’s actions in both mechanical and thermal manufacturing processes. Then, the dissertation will be demonstrated in two main parts with nine research tasks. Each task has been published or submitted and these specific research tasks are elaborated as follows:

(1). UV-A hole making of CFRP composites will be illustrated in Chapters III to V to conduct a comparative study on process performance between UV-A hole making and conventional core grinding, to investigate effects of input variables on cutting force, and to develop a novel mechanistic modeling method for ultrasonic vibration amplitude measurement through cutting force, respectively. Furthermore, UV-A surface grinding of CFRP composites will be given in Chapters VI and VII to evaluate the improvement of process performance in comparison with conventional surface grinding and to establish a novel mechanistic model for the feed-direction

7 Texas Tech University, Fuda Ning, May 2018 cutting force prediction, respectively.

(2). UV-A LENS of metallic alloys and metal matrix composites will be described in Chapters VIII to XI. Effects of ultrasonic vibration on the geometrical features, porosity, grain size, phase morphology, microstructural characteristics, and mechanical properties (tensile properties, microhardness, and wear resistance) will be investigated with different types of feedstock materials (stainless steel, nickel-based superalloy, and titanium matrix composites). In addition, the effects of ultrasonic vibration frequency will be explored in UV-A LENS of nickel-based superalloy.

Finally, the conclusions and scientific contributions of this dissertation will be demonstrated in Chapter XII.

References [1] Flower, H.M., 1995, High performance materials in aerospace, 1st edition, Springer, London, UK. [2] Gu, D.D., 2015, Laser additive manufacturing of high-performance materials, 1st edition, Springer, Berlin, Germany. [3] Park, K.Y., Choi, J.H., and Lee, D.G., 1995, Delamination-free and high efficiency drilling of carbon fiber reinforced plastics, Journal of Composite Materials, 29(15), pp. 1988–2002. [4] Davim, J.P., and Reis, P., 2003, Drilling carbon fiber reinforced plastics manufactured by autoclave-experimental and statistical study, Materials and Design, 24(5), pp. 315–324. [5] Amir, A., Ye, L., and Chang, L., 2016, Drilling conditions on hole quality for CFRP laminates, in Proceedings of the American Society for Composites, September 19-22, Williamsburg, VA, USA, pp. 1–16. [6] Jia, Z.Y., Fu, R., Niu, B., Qian, B.W., Bai, Y., and Wang, F.J., 2016, Novel drill structure for damage reduction in drilling CFRP composites, International Journal of Machine Tools and Manufacture, 110, pp. 55–65. [7] Zhong, C.L., Gasser, A., Kittel, J., Wissenbach, K., and Poprawe. R., 2016, Improvement of material performance of Inconel 718 formed by high deposition- rate laser metal deposition. Materials & Design, 98, pp. 128–134. [8] Chen, Y., Lu, F., Zhang, K., Nie, P., Hosseini, S.R.E., Feng, K., and Li, Z., 2016, Dendritic microstructure and hot cracking of laser additive manufactured Inconel 718 under improved base cooling, Journal of Alloys and Compounds, 670, pp. 312–321. [9] Parimi, L.L., Ravi, G.A., Clark, D., and Attallah, M.M., 2014, Microstructural and texture development in direct laser fabricated IN718, Materials

8 Texas Tech University, Fuda Ning, May 2018

Characterization, 89, pp. 102–111. [10] Tabernero, I., Lamikiz, A., Martínez, S., Ukar, E., and Figueras, J., 2011, Evaluation of the mechanical properties of Inconel 718 components built by laser cladding, International Journal of Machine Tools and Manufacture, 51(6), pp. 465–470. [11] Jia, Q.B., and Gu, D.D., 2014, Selective laser melting additive manufacturing of Inconel 718 superalloy parts: densification, microstructure and properties, Journal of Alloys and Compounds, 585, pp. 713–721. [12] Irwin, J., Reutzel, E.W., Michaleris, P., Keist, J., and Nassar, A.R., 2016, Predicting microstructure from thermal history during additive manufacturing for Ti-6Al-4V, ASME Journal of Manufacturing Science and Engineering, 138(11), pp. 111007–111017. [13] Gu, D.D., Meiners, W., Wissenbach, K., and Poprawe, R., 2012, Laser additive manufacturing of metallic components: materials, processes and mechanisms, International Materials Reviews, 57(3), pp. 133–164. [14] Komarov, S.V., Kuwabara, M., and Abramov, O.V., 2005, High power ultrasonics in pyrometallurgy: current status and recent development, ISIJ International, 45(12), pp. 1765–1782. [15] Abrate, S., and D.A. Walton, 1992, Machining of composite materials. Part I: Traditional methods, Composites Manufacturing, 3(2), pp. 75–83. [16] Feng, Q., Cong, W.L., Pei, Z.J., and Ren, C.Z., 2012, Rotary ultrasonic machining of carbon fiber-reinforced polymer: feasibility study, Machining Science and Technology, 16(3), pp. 380–398. [17] Cong, W.L., Feng, Q., Pei, Z.J., Deines, T.W., and Treadwell, C., 2011, Dry machining of carbon fiber reinforced plastic composite by rotary ultrasonic machining: effects of machining variables, Proceedings of the ASME 2011 International Manufacturing Science and Engineering Conference, Corvallis, OR, pp. 363–371. [18] Liu, J., Zhang, D., Qin, L., and Yan, L., 2012, Feasibility study of the rotary ultrasonic elliptical machining of carbon fiber reinforced plastics (CFRP), International Journal of Machine Tools and Manufacture, 53(1), pp. 141–150. [19] Cong, W.L., Pei, Z.J., Feng, Q., Deines, T.W., and Treadwell, C., 2012, Rotary ultrasonic machining of carbon fiber reinforced plastic composites: using cutting fluid versus cold air as coolant, Journal of Composite Materials, 46(14), pp. 1745–1753. [20] Cong, W.L., Pei, Z.J., Feng, Q., Deines, Srivastava, A., Riley, L., and Treadwell, C., 2012, Rotary ultrasonic machining of CFRP composites: a study on power consumption, Ultrasonics, 52(8), pp. 1030–1037. [21] Cong, W.L., Pei, Z.J., Sun, X., and Zhang, C.L., 2014, Rotary ultrasonic machining of CFRP: a mechanistic predictive model for cutting force, Ultrasonics, 54(2), pp. 663–675.

9 Texas Tech University, Fuda Ning, May 2018

[22] Abramov, O.V., 1987, Action of high intensity ultrasound on solidifying metal, Ultrasonics, 25(2), pp. 73–82. [23] Banerjee, R., Collins, P.C., Genc, A., and Fraser, H.L., 2003, Direct laser deposition of in situ Ti-6Al-4V-TiB composites, Materials Science and Engineering: A, 358(1), pp. 343–349. [24] Banerjee, R., Genc, A., Hill, D., Collins, P.C., and Fraser, H.L., 2005, Nanoscale TiB precipitates in laser deposited Ti-matrix composites, Scripta Materialia, 53(12), pp. 1433–1437.

10 Texas Tech University, Fuda Ning, May 2018

CHAPTER II

A LITERATURE REVIEW ON ULTRASONIC VIBRATION- ASSISTED (UV-A) MANUFACTURING PROCESSES

Abstract In recent years, ultrasonic technology has been extensively applied in numerous manufacturing processes to improve the process performance and part quality. This review paper presents a broad overview of the recent progress in ultrasonic vibration-assisted (UV-A) manufacturing processes reported in the literature. Based on the ultrasonic energy propagation through solid or liquid phases, UV-A manufacturing processes can be divided into mechanical manufacturing processes (including conventional machining, densification, forming, and consolidation) and thermal manufacturing processes (including thermal non-traditional machining, casting, fusion welding, and laser cladding). The results from a great number of published investigations have strongly evidenced the significant influences of ultrasonic vibration during the material processing. In UV-A mechanical manufacturing processes, ultrasonic vibration can reduce the machining force through intermittent cutting between machine tools and the workpiece, decrease the load in both densification and forming processes due to the reduced friction, and increase the bond at the workpiece interface by breaking oxide layers in ultrasonic consolidation. In UV-A thermal manufacturing processes, ultrasonic vibration can exert nonlinear effects (acoustic streaming and cavitation) on the solidification behavior of liquid melting materials. The precision and quality of parts fabricated by thermal manufacturing will be thus improved through various phenomena such as element homogenization, material degassing, crack reduction, microstructure refinement, etc. This literature review aims to provide a comprehensive overview on ultrasonic vibration influences in different UV-A manufacturing processes and guide the future development of ultrasonic vibration assisted technologies.

Keywords: Ultrasonic vibration; UV-A manufacturing; Mechanical manufacturing; Thermal manufacturing; Acoustic streaming; Cavitation.

11 Texas Tech University, Fuda Ning, May 2018

2.1 Background Ultrasound corresponds to a physical wave that propagates at frequencies of more than 20 kHz, which is beyond the conventional upper limit of human hearing. Such wave can propagate in different mediums depending on their intrinsic elastic properties and densities [1, 2]. The mechanical vibration generated by ultrasound provokes tiny displacements of the molecules of the medium from their resting position. The formed ultrasonic vibration can propagate as a wave traveling to other parts of the medium. The ultrasound wave and its propagation can be influenced by many factors such as ultrasonic frequency, amplitude, etc. [2]. Ultrasonic frequency is related to the times of complete oscillations generated per second. Both ultrasonic frequency and amplitude determine the value of ultrasonic intensity, which is the amount of power passing through per unit cross-section area. To realize the effective supply of ultrasonic vibration to the materials being processed, it is essential to provide a reliable contact between an ultrasonic generator and the materials by utilizing a medium such as a horn/tool holder assembly, an ultrasound probe, an ultrasonic vibration platform, etc. [3].

The concept of applying ultrasonic vibration to improve the process performance was generated for a long time [3]. Since then, the use of ultrasonic vibration in many manufacturing processes have been extensively studied and few investigations have been reported on the adverse influences of ultrasonic vibration [2]. Compared with other low-frequency vibrations, ultrasonic vibration possesses a frequency that is much higher than the natural frequency of a manufacturing system. With this reason, ultrasonic vibration can maintain or even improve the stability of system without adding harmful low-frequency vibrations in the manufacturing system. According to the results from a large number of published investigations, ultrasonic vibration has been strongly proven to exert significant influences and play important roles in assisting different manufacturing processes.

Ultrasonic vibration can be transmitted through solid-phase or liquid-phase materials with extraordinary properties. Based on the transmission medium, UV-A

12 Texas Tech University, Fuda Ning, May 2018 manufacturing processes introduced in this literature review are divided into two main parts, including UV-A mechanical manufacturing processes (conventional machining, densification, forming, and consolidation) and UV-A thermal manufacturing processes (thermal non-traditional machining, casting, fusion welding, and laser cladding). In the UV-A mechanical manufacturing processes, ultrasonic vibration is applied to change the contact behaviors between the machine tools and processed solid parts, thereby increasing manufacturing efficiency and improving the part quality. In UV-A thermal manufacturing processes, however, ultrasonic vibration is introduced to affect the liquid melting material solidification behaviors. Acoustic streaming and cavitation, two nonlinear effects induced by ultrasonic vibration, are always caused within the molten materials through periodically positive-negative pressures and violent movements [4]. These effects can be used to explain phenomena occurred in UV-A thermal manufacturing processes.

This chapter provides a comprehensive overview of the recent achievements on the utilization of ultrasonic vibration in various manufacturing processes. The effects of ultrasonic vibration on the process performance and part quality are elaborately discussed for each manufacturing process. Considerable efforts have been devoted to a better understanding of the phenomena associated with the principles of ultrasonic vibration in UV-A manufacturing processes, which will be also beneficial for understanding the topics presented in the following chapters.

2.2 UV-A mechanical manufacturing processes

2.2.1 UV-A conventional machining The history of ultrasonic vibration being utilized in assisting the conventional machining could trace back to the 1950s when ultrasonic vibration was applied in turning of metal workpieces such as aluminum, brass, mild steel, and cast iron [5]. Since then, ultrasonic vibration has also been used in assisting various conventional machining processes, such as turning, milling, drilling, and grinding [6-9], to improve the machining efficiency and produce parts with a high precision.

13 Texas Tech University, Fuda Ning, May 2018

In UV-A conventional machining processes, a proper design of ultrasonic vibration system is important to the successful implementation of a machining process. Such system is usually composed of an ultrasonic power supply, an ultrasonic transducer, an electric motor, a feeding device, and a control panel. The ultrasonic power supply generates the high-frequency electrical energy and then transmit it to the ultrasonic transducer. Such energy is converted into the mechanical vibration via a piezoelectric/magnetostrictive plate located within the transducer. The vibration is further amplified and then transmitted to the cutting tool through a horn/tool holder assembly [10]. The fundamental principles of ultrasonic vibration in conventional machining processes are similar. As a typical UV-A conventional machining process, UV-A turning has been reported to be suitable for cutting various difficult-to-machine materials with qualitative improvements in the process [6]. In UV-A turning, ultrasonic vibration could be independently applied to the cutting tool in a tangential direction (direction of cutting velocity), a feed direction, or a radial direction, as shown in Figure 2.1. Based on the cutting tooltip trajectories, UV-A turning has two main groups including 1-D UV-A turning and 2-D UV-A turning. In the 1-D cutting system, cutting tool vibrates only in one direction (mostly in tangential or feed direction). The 2-D UV-A cutting system allows the tooltip to vibrate simultaneously in both tangential and feed directions forming an elliptic trajectory. The cutting tool in the UV-A turning can periodically contact with the workpiece, thus enabling the materials to be removed by the intermittent cutting mode rather than the continuous interaction between the cutting tool tip and workpiece in the conventional turning [6, 11]. Such a different cutting behavior can facilitate the chip separation, reduce frictional forces, and then decreases cutting forces [12]. It also can reduce the deformation zone in a workpiece, extending tool life and improving the surface integrity of the machined parts [5, 11, 13].

14 Texas Tech University, Fuda Ning, May 2018

Workpiece Z

Y X Rotation Feeding

Ultrasonic vibration on tangential direction

Ultrasonic vibration on radial Ultrasonic direction vibration on feed direction Figure 2.1 Illustrations on UV-A turning process (after [6])

UV-A grinding, also known as rotary ultrasonic machining (RUM), is another widely used UV-A conventional machining process. In this process, ultrasonic vibration is implemented on the abrasive cutting tool and/or workpiece. The rotating cutting tool is fed into the workpiece at a constant feedrate through the vertical direction with a vertical ultrasonic vibration for the hole-making purpose (Figure 2.2a) or through the horizontal direction with vertical and/or horizontal (along with the feeding direction) ultrasonic vibrations for the surface grinding purpose (Figure 2.2b). Specifically, in the UV-A hole-making process [14-17], the contact mode between the end face of cutting tool and workpiece is intermittent, generating the impacts between abrasives and workpiece. Thus, the materials are primarily removed by the abrasive indentation caused by ultrasonic vibration through microchipping. Compared with conventional core grinding [17], UV-A hole-making produces a smaller indentation depth of cut, resulting in lower cutting force, less tool wear, and better surface quality. On the other hand, in the UV-A surface grinding process [18-20], the abrasives on the peripheral face of the cutting tool are mainly involved in cutting the workpiece with continuous contacts between the abrasive grains and workpiece under the vertical ultrasonic vibration. In such a process, the abrasives on the end face of the tool are responsible for the surface formation on the machined tracks. Similar to UV-A hole

15 Texas Tech University, Fuda Ning, May 2018 making, UV-A surface grinding also generates a smaller indentation depth than the conventional surface grinding due to the larger trajectory length at the same material removal volume, leading to lower cutting forces in both axial and feeding directions [18]. However, the larger surface roughness is achieved due to the occurrence of surface damages induced by the vertical ultrasonic vibration impact [18]. Such grinding defect may be avoided with the introduction of horizontal ultrasonic vibration, in which abrasive grains on the tool end face are involved in the surface grinding.

Coolant Coolant Rotation Rotation Z flow in flow in

Y X Vertical Vertical ultrasonic ultrasonic Feeding vibration vibration Abrasive portion Coolant Coolant Coolant flow out flow out flow out

Fee ding

orkpiece Workpiece Abrasive W and/or portion horizontal ultrasonic vibration (a) For hole making (b) For surface grinding Figure 2.2 Illustrations on UV-A grinding processes

2.2.2 UV-A densification Ultrasonic vibration has been introduced to densify powders especially the metal powders to obtain an increased green body density [21-24]. To achieve the optimal UV-A densification behavior, ultrasonic vibration should be imported at an optimum combination of ultrasonic frequency, vibration amplitude, and applied time [22, 25].

Figure 2.3 illustrates the UV-A densification mechanism in three stages [21]. At the first stage, the powder compacts inside the die experience a vertically applied

16 Texas Tech University, Fuda Ning, May 2018 pressure. The pressure is applied to overcome the barrier friction between the die wall and particles as well as the inter-particle friction. When the high-frequency horizontal ultrasonic vibration is applied to the die at the second stage, the contact time between the particles and die wall decreases, leading to the reduction of barrier friction. In addition, the friction of inter-particles could be lowered by the particle resonance induced by ultrasonic vibration. Thus, at the last stage, it can be expected that the powder compact density would be enhanced due to the reduced frictions [24]. Such improvement of the compact characteristics is also found when a vertical ultrasonic vibration is implemented onto both an upper punch and a lower punch for pressing the metal powders [22, 26]. Compared to conventional densification, UV-A densification can produce green bodies with a 10% higher density and a 20% higher hardness [21]. The ultrasonic vibration in the densification process can also generate better-arranged and fewer-fractured structures with a uniform density distribution.

1st stage 2nd stage 3rd stage Applied pressure Applied pressure

Applied pressure

Inter-particle friction

Horizontal Horizontal ultrasonic ultrasonic Barrier vibration vibration friction

Applied pressure

Applied pressure Applied pressure Figure 2.3 Illustrations on UV-A densification mechanism in three stages (after [21])

Pelleting process for biomass powder densification in biofuel manufacturing is another densification process, enabling the improvement of the overall utilization efficiency of transportation infrastructure and storage systems during biofuel production [27]. UV-A pelleting, a newly developed pelleting method, is characterized by combining ultrasonic treatment and pelletizing as one process that produces biomass pellets with lower pelleting force, higher durability, higher cellulose

17 Texas Tech University, Fuda Ning, May 2018 accessibility, and larger density [28, 29]. Figure 2.4 shows the schematic of UV-A pelleting of cellulosic biomass [30]. The ultrasonic vibration is applied to the pelleting tool in UV-A pelleting. In UV-A pelleting, ultrasonic vibration enables the biomass powders to vibrate up and down, which can well arrange the particles in a pellet form. Such an arrangement decreases pores in pellets, leading to the reduction of pelleting force. Moreover, ultrasonic energy could produce an intense hydro-mechanical shear force [31]. This impact could break the lignin shell of the particle, provide an additional bond in pellets, and increase cellulose accessibility [30, 32]. A clear contrast between pelleting without ultrasonic vibration and UV-A pelleting for producing pellets is given in Figure 2.5. It is notable that compared with pelleting without ultrasonic vibration, UV-A pelleting results in a more compact cylindrical shape pellet with fewer loose particles [28].

Pelleting tool

Feeding

Ultrasonic vibration

Mold

Biomass

Figure 2.4 Illustrations on UV-A pelleting of cellulosic biomass [30]

18 Texas Tech University, Fuda Ning, May 2018

Figure 2.5 Pellets produced by (a) pelleting without ultrasonic vibration and (b) UV-A pelleting [28]

2.2.3 UV-A forming Ultrasonic vibration has been used as an effective assisted technique to reduce the forming load and to enhance the part quality in various forming processes including hot embossing, extrusion, wire drawing, etc.

Hot embossing has become a popular method to replicate precise micro features onto large plates. However, a long cycle time of the process could affect the success of this technology [33]. In order to solve the problem, investigations on UV-A hot embossing process have been conducted [33-35]. As shown in Figure 2.6, the UV- A hot embossing process begins by heating both steel molds and material (e.g. glass) to the molding temperature at the first stage. Then the mold gently embosses the glass with a pressure at the second stage. As the embossing continues at the third stage, ultrasonic vibration will be imposed on the upper mold once the displacement reaches a certain value (1.5 mm in the work of [36]) and the vibration will not stop until the maximum displacement (2 mm in the work of [36]) is achieved. At this moment, the glass is held for 30 s before it will be released from the molds at the final stage. The actions of ultrasonic vibration could cause an increase in local temperature between the molds and glass, enabling the softening glass to easily flow into the mold cavities with the formation of more accurate molded structures and better surface quality [37].

19 Texas Tech University, Fuda Ning, May 2018

Embossing force/pressure would be therefore decreased rapidly owing to the softened materials induced by the ultrasonic vibration in the hot embossing process [36, 37].

1. Heating 2. Start 3. Import vertical 4. End 5. Demolding embossing ultrasonic vibration embossing Ultrasonic horn Pressure Upper steel mold

Glass Lower steel mold Lower die Figure 2.6 Procedures of UV-A hot embossing process (after [36])

In UV-A extrusion process, the ultrasonic vibration can be applied either on the die or on the punch. It has been reported that ultrasonically vibrating the dies is more practical to reduce the extrusion force and material flow stress by decreasing the sliding friction between the workpiece and die [38-40]. The schematic diagram of a typical UV-A extrusion process is shown in Figure 2.7. In this case, the ultrasonic vibration is imposed on the die. Due to the relative movement between the workpiece and die, micro gaps can be formed at the die-workpiece interface within each ultrasonic vibration cycle. Such phenomenon could be considered as the intermittent contact behavior resulting from the ultrasonic vibration properties. This would modify the contact pressure features and reduce the friction at the interface, thus causing the reduction of extrusion force. It is also found that extrusion force decreases with the increase of both ultrasonic frequency and amplitude owing to the increased separation time of the die shoulder from the flowing material [39].

Similarly, ultrasonically vibrated dies are proposed to be applied in wire drawing to improve the drawing performance [41]. Comparisons between conventional and UV-A wire drawing processes have been performed. Similar to UV- A extrusion, the micro gaps can be created between the die and workpiece in the UV- A wire drawing, leading to the reduction of drawing force by nearly 30% [41]. In addition, the gaps could allow the lubricant to flow in, improving the die lubricating

20 Texas Tech University, Fuda Ning, May 2018 conditions. Compared with conventional wire drawing, UV-A wire drawing has been proven to effectively increase drawing speed and greatly improved the surface quality of wires [41, 42].

Extrusion (1) (2) (3) (4) (5)

force Punch

t i

Die l

A *

Ultrasonic 2 Vibration Pressure Shoulder

Workpiece Micro gaps (for illustration purpose) Figure 2.7 Procedures of UV-A extrusion process at one ultrasonic vibration cycle (after [38])

2.2.4 Ultrasonic consolidation Ultrasonic consolidation is a solid-state joining process that fabricates three- dimensional components through ultrasonic welding of layered metal tapes or foils [43]. Besides metal parts, multi-material engineering parts, such as embedded electronics, dissimilar metals, metal matrix composites, etc., can also be effectively fabricated in ultrasonic consolidation process [44]. In such process, ultrasonic vibration is locally applied to the metal tape or foil using a rolling ultrasonic sonotrode under the normal force to induce localized friction, as shown in Figure 2.8 [45]. The frictional heat is considered to be the primary bonding source that is more than twice that of heating from intense plastic deformation [46]. In addition, shear scrubbing induced by ultrasonic vibration can break up the oxides on the surfaces, allowing for the generation of nascent surface and formation of an atomic bonding between layers [47]. With ultrasonic energy, the foils would experience a significant acoustic softening in the ultrasonic field. Such softening effect occurs in the regions of the

21 Texas Tech University, Fuda Ning, May 2018 metal lattice which are known to conduct the mechanisms of plastic deformation [48, 49].

Ultrasonic Normal force transducer Rotation Sonotrode

Tape Ultrasonic Substrate Vibration

Figure 2.8 Illustrations on ultrasonic consolidation process (after [45])

Characteristics of weld formation during ultrasonic consolidation have been explored by investigating the microstructural evolution [50]. The results show that fine and equiaxed grains in the weld interface can be obtained due to the recrystallization that is occurred as a consequence of the heating and shear deformation along the ultrasonic vibration direction. The grain refinement at the joint regions between the first and second layers of the aluminum alloy parts could be observed from the Figure 2.9 [50]. Similarly, the microstructural evolution has been investigated elsewhere [51, 52] to study the bonding mechanism, evidencing the important actions of ultrasonic vibration during the solidification and recrystallization in the ultrasonic consolidation process.

In ultrasonic consolidation, the ultrasonic amplitude is considered to have a significant influence on the process performance since it greatly affects dynamic interfacial plastic shear strains at the contacting asperities [46]. Ultrasonic consolidation with a higher ultrasonic amplitude has been reported to fabricate the parts with larger tensile properties that almost approach the value of bulk materials [53]. In addition, fewer voids and better bonding are achieved in the fabricated aluminum parts processed under the larger ultrasonic amplitude, as shown in Figure

22 Texas Tech University, Fuda Ning, May 2018

2.10 [54]. The reason is that the larger amplitude could result in 100% linear weld density with a more compact foil layer. More importantly, the increase in ultrasonic amplitude enhances the plastic deformation and promotes plastic flow to provide favorable conditions for both fewer voids and stronger bonds [55].

2nd layer

Fine grain region

20 μm 1st layer

Figure 2.9 Fine grains around the interface regions between the first and second layers in the aluminum alloy parts fabricated by ultrasonic consolidation (after [50])

(a) (b)

Voids

200 μm Substrate 200 μm Substrate Figure 2.10 Optical micrograph of aluminum alloy parts fabricated by ultrasonic consolidation under (a) the smaller ultrasonic amplitude and (b) the larger ultrasonic amplitude (after [54])

Ultrasonic consolidation is also found to be applicable for creating metal matrix composites at room temperature. The strength of the reinforcement-matrix interface has been studied by analyzing the interface failure and the results suggest

23 Texas Tech University, Fuda Ning, May 2018 that an increased composite length would avoid the occurrence of composite damage [45]. In order to improve the understanding of the interfacial shear strength of metal matrix composites, Hehr et al. [56] investigated the interface strength using single fiber pullout tests and found that the bond mechanisms were linked with the function of fiber surface finish.

2.2.5 Remarks on the effects of ultrasonic vibration In UV-A mechanical manufacturing processes, the workpiece is not thermally damaged and no liquid melting materials are formed. The effects of ultrasonic vibration are only applied to the solid workpiece materials. Based on the investigations discussed above, the actions and influences of ultrasonic vibration on the performance of mechanical manufacturing processes are summarized in Figure 2.11.

There are four main actions of ultrasonic vibration in mechanical manufacturing processes. The intermittent contact is induced between the workpiece and manufacturing tools such as cutting tools, pelleting tools, dies, etc. to decrease the contact time. It will help to promote the chip separation in machining and reduce the friction or pressure at the interface in most mechanical manufacturing processes. Ultrasonic vibration also leads to the particle resonance in the densification process. Thus, a well arrangement of the particle can be achieved to reduce the inter-particle friction and pores. Moreover, material softening action can be caused by the ultrasonic vibration to increase the plastic flowability of the materials in the embossing process. The last action induced by ultrasonic vibration is shear scrubbing, resulting in the breakage of surface oxides to facilitate the bonding formation in UAM. All these actions and associated indirect actions can ultimately influence the process performance, including the reduction of various types of forces (such as cutting force, pelleting force, embossing force, extrusion force, drawing force, etc.), improvement of tool life, and enhancement of manufactured part quality (such as surface integrity, compact density, accuracy, etc.).

24 Texas Tech University, Fuda Ning, May 2018

Direct influences Ultimate influences (Indirect actions)

Cutting force

n i

Easy chip n Tool life i separation h

Actions of c a

ultrasonic M vibration Surface integrity Friction/pressure

Intermittent reduction n o

Compact density i

contact t

i s

Particle Pelleting force n e

Ultrasonic resonance Pore reduction D vibration properties Embossing force Material

softening g

Plastic flowability n

enhancement Part accuracy i

m r

Shear o scrubbing F Extrusion/drawing Surface oxides force breakage

Bonding formation C U

Figure 2.11 Actions and influences of ultrasonic vibration in UV-A mechanical manufacturing processes. (UC is ultrasonic consolidation.)

2.3 UV-A thermal manufacturing processes Ultrasonic vibration is also applicable in the thermal manufacturing processes since it has the capability of supplying the acoustic energy with an effective transmission from the ultrasonic generator to molten materials. In UV-A thermal manufacturing processes, providing a high-intensity ultrasound wave is crucial to achieving desired manufacturing outcomes. Thus, during the propagation of such wave, two main nonlinear effects are produced and their specific characteristics will be introduced in Section 2.3.1.

2.3.1 Nonlinear effects of ultrasonic vibration on liquid melting materials The nonlinear effects, as the basis for UV-A thermal manufacturing processes, are mainly caused due to the ultrasound energy attenuation [3]. The two nonlinear effects of ultrasonic vibration include acoustic streaming and cavitation, which can

25 Texas Tech University, Fuda Ning, May 2018 facilitate liquid material movements and further result in different direct and ultimate influences. Based on the reported investigations on UV-A thermal manufacturing processes [4], the specific actions and influences of ultrasonic vibration are summarized in Figure 2.12.

Acoustic streaming is the steady flow that is driven by momentum transfer due to the absorption of acoustic oscillations in the liquid materials. Such phenomenon was originally explained by J. Lighthill considering that the spatial variation of the Reynolds stress on the liquid materials resulted in the acoustic streaming [3]. The acoustic streaming is of great importance in controlling the physical and chemical phenomena in the liquid materials being processed with the ultrasonic vibration. Cavitation is a physical phenomenon that causes the generation of tiny bubbles or cavities followed by their growth, pulsation, and collapse in the liquid materials [3]. Such phenomenon appears when the acoustic pressure exceeds the critical value (namely cavitation threshold), which is the minimum acoustic energy required to initiate the cavitation. The cavitation threshold is affected by the content of nonmetallic solid and gaseous inclusions in the liquid [57]. A larger microbubble within the liquids will lead to a lower cavitation threshold. It is believed that the tensile stress induced by the ultrasound wave in its rarefaction phase is responsible for the formation of cavities [3]. The performance of a cavity is determined by both liquid material properties and acoustic field characteristics. Once the cavity collapses, a high- intensity shock is instantaneously emitted along the wave propagation.

Both acoustic streaming and cavitation affect the mass and heat transfer as well as the crystal growth during the liquid melting material solidification. Consequently, several indirect actions including stirring and mixing, crystal dispersion, and crystal nucleation will be generated. These indirect actions of ultrasonic vibration are also considered as the direct influences for the further induced phenomena, as shown in Figure 2.12. These phenomena serve as the ultimate influences on the liquid melting materials in thermal manufacturing processes, which will be presented in the following sections.

26 Texas Tech University, Fuda Ning, May 2018

Ultimate influences Direct influences Actions of Homogenizing ultrasonic (Indirect actions) material

vibration composition t

n Stirring and Degassing e mixing

Acoustic m and reducing

streaming e porosity

o m

Ultrasonic Improving s vibration l

a Dispersion of flatness i

properties r crystals

e t

a Reducing m cracks

Cavitation d

i u

q Nucleation of i

L crystals Refining grains

Figure 2.12 Actions and influences of ultrasonic vibration in UV-A thermal manufacturing processes (after [4])

2.3.2 UV-A thermal non-traditional machining Thermal energy can be involved in the material removal in thermal non- traditional machining processes, such as laser beam machining and electrical discharge machining (EDM), causing localized temperature high enough to melt or vaporize particles on the workpiece surface.

Ultrasonic vibration has been reported to be utilized in laser beam machining processes for hole making [58, 59] and surface processing [60-62]. A novel UV-A laser hole making system set-up has been designed, in which the impacts induced by ultrasonic vibration between the laser beam and workpiece could improve the hole surface quality by reducing the resolidified and redeposited particles on the hole surfaces [58]. Similar investigation [59] has been conducted by applying ultrasonic vibration to the workpiece to eject the laser-ablated particles for the improvement of hole quality and machining efficiency. With the assistance of ultrasonic vibration, micro holes with a significant increase of the hole depth can be generated in the UV-A laser hole making process. Compared with hole making, surface processing using UV- A laser machining method gains more attention due to its higher material removal rate.

27 Texas Tech University, Fuda Ning, May 2018

According to the reported results [60, 61], with the ultrasonic vibration, the area of heat affected zone within the laser-irradiated sample decreases and a smaller surface crater can be formed. Such phenomenon indicates that the ultrasonic vibration during laser surface processing could delay the interaction between the laser and processed part and accelerate the heat dissipation due to the enhanced surface convection effects. It is also observed that ultrasonic vibration forms a near-field surface cooling effect to improve the surface finish and to prevent the surface oxidation [62].

Besides the laser beam machining, EDM has been found to combine with ultrasonic vibration to achieve the UV-A EDM process that enables an improvement of material removal rate [63-65]. It can be seen from Figure 2.13 that the transducer and horn were attached between the spindle and tool electrode. The tool electrode vibrates ultrasonically with the adjustable ultrasonic amplitude [64]. Due to the cavitation of ultrasonic vibration [63, 65], the molten material of the workpiece could be easily removed, which improves the spark gap condition and further results in the increase of machining efficiency. On the other hand, the impact of the electron on the workpiece is produced during the vibration of tool electrode and the impact has a tangential component that could remove the peaks of the craters on the machined surface. Thus, the surface roughness of the workpiece can be reduced with the lower overlapping of the craters [64]. Another important improvement by UV-A EDM over conventional EDM is that micro-cracks on the machined surface could be remarkably reduced, as shown in Figure 2.14 [63]. This is attributed to the acoustic wave produced by ultrasonic vibration that greatly mitigates the generation of re-solidified materials, leading to the decreased thickness of the heat affected zone (HAZ) and lower thermal stresses for fewer micro-cracks [63, 64].

28 Texas Tech University, Fuda Ning, May 2018

Spindle Ultrasonic generator

Transducer

Horn _ Tool electrode Ultrasonic vibration Guider

+ Workpiece

Electrical Dielectric source liquid

Figure 2.13 Illustrations on UV-A EDM system set-up (after [64])

(a) (b)

Micro-cracks

40 μm Machined surface 40 μm Machined surface Figure 2.14 Machined surfaces of parts fabricated by (a) EDM process and (b) UV-A EDM process (after [63])

2.3.3 UV-A casting Since the middle of the last century, ultrasonic vibration has been applied as a dynamic nucleation method to the melts during casting [66, 67]. Ultrasonic vibration treatment of melts in casting is beneficial for controlling microstructure, mechanical properties, and part quality of light alloys (such as aluminum alloy [57, 68], magnesium alloy [69, 70], etc.), and light alloy based composites [71-74]. Most

29 Texas Tech University, Fuda Ning, May 2018 investigations [67, 69-71, 74-78] have implemented the ultrasonic vibration by dipping the ultrasonic probe into the melts to disperse the molten materials prior to the casting process. Then, the melts would be poured into the preheated crucible for the solidification process under the ultrasonic vibration again, as shown in Figure 2.15 [74].

Figure 2.15 Ultrasonic vibration imported for solidifying melts during casting process (after [74])

It has been reported that the cavitation effect induced by ultrasonic vibration could facilitate degassing and filtrating the melts prior to the casting process [57]. In this case, the primary crystals are refined and better distributed in the cast as a result of the ultrasonic treatment. A similar finding has been obtained in the investigation on the influence of ultrasonic vibration on the pure aluminum alloy melt prior to the solidification [75]. The ultrasonic vibration is applied to the melts for 10 min to remove the slag and degas the melts that would be rested before being poured into the mold for cooling to the room temperature. Figure 2.16 shows the microstructures of the aluminum alloy A356 parts fabricated without and with ultrasonic vibration [75]. It is notable that the silicon phases exhibit a long bar shape with an average size of 50 µm, while they are in a short rod shape with a smaller average size of about 5 µm and

30 Texas Tech University, Fuda Ning, May 2018 are shown in a nearly uniform distribution in the aluminum phase when ultrasonic vibration is utilized.

(a) (b)

Silicon phase

Aluminum phase

20 μm 20 μm

Figure 2.16 Microstructures of aluminum alloy A356 parts fabricated (a) without and (b) with ultrasonic vibration (after [75])

For the nanocomposite fabrication in UV-A casting [71, 73], ultrasonic vibration is also found to enable nanoparticles to be well dispersed and incorporated into the magnesium alloy matrix under the ultrasonic amplitude of around 20 µm. Owing to the ultrasonic cavitation effects, the transient impacts coupled with a high temperature of about 5000 °C and an extraordinarily rapid cooling rate of 1010 K/s enable the breakage of nanoparticle clusters and clean of particle surface [79]. A microstructure refinement is obtained with the average grain size reducing from 820 µm to 360 µm when ultrasonic vibration is propagated into the T10 steel melts [66]. The grain refinement is achieved during the nucleation stage, which is considered to be caused due to the acoustic streaming and cavitation effects of the high-intensity ultrasonic vibration [69, 74, 75]. With the continuous increase of the ultrasonic power, the impulse force produced during the collapse of the cavitation bubbles is getting larger so that the growing crystals vibrate more violently. Thus, more nuclei would be generated causing the formation of finer and better-distributed grains [66, 67]. The dendrite fragmentation is also found to facilitate the formation of refined grains and it is caused by the large instantaneous pressure and temperature fluctuations in the melts [75].

31 Texas Tech University, Fuda Ning, May 2018

The dynamic behavior of cavitation and effects on the dendrite fragmentation are investigated by a numerical analysis with an experimental high-speed digital imaging [80]. The interaction between ultrasonic cavitation bubbles and dendrites is shown in Figure 2.17. Due to the high attenuation of ultrasonic vibration in the liquid materials, cavitation bubbles are primarily created around the ultrasonic probe [81]. It can be clearly seen that the primary dendrite is initially bent and then fragmented into small pieces by the local shock waves in 0.44 ms, evidencing an important role of ultrasonic vibration effects on the microstructure refinement. It has been also found that acoustic streaming induced by ultrasonic vibration has no direct influences on the dendrite fragmentation, however, such dynamic behavior plays an important role in facilitating the continuous fragmentation of newly growing dendrites by flowing the cavitation bubbles to the dendrite arrays [80].

(a) Ultrasonic probe (b) Ultrasonic probe

Cavitation Cavitation bubble bubble

50 μm 50 μm

Figure 2.17 Interactions between cavitation bubbles and dendrite (a) before fragmentation and (b) after fragmentation (after [80])

The phenomena of melt degassing, elemental composition homogenization, and microstructural refinement caused by ultrasonic vibration would further enhance the mechanical properties of the parts processed by UV-A casting. It can be seen from the Figure 2.18 that yield strength (YS), ultimate tensile strength (UTS), and ductility of magnesium alloy AZ91 are increased after introducing ultrasonic vibration into casting process [70]. AZ91 alloy also exhibits an increased UTS with the increase of ultrasonic power due to more uniform and finer dendrites as well as smaller

32 Texas Tech University, Fuda Ning, May 2018 microstructural phases [67]. The improvement of YS, UTS, and elongation under the condition of ultrasonic vibration can also be found elsewhere for aluminum alloys [75], steels [66], and magnesium based composites [73].

200 PartCast iwithoutng witho uultrasonict ultrasoni cvibration vibration PartCast iwithng w iultrasonicth ultrasoni cvibration vibration 160

120

0 YS UT S Ductility (MPa) (MPa) (‰) Figure 2.18 Tensile properties of magnesium alloy AZ91 parts made by the casting with and without ultrasonic vibration (after [70])

2.3.4 UV-A fusion welding Ultrasonic vibration has been reported to be applied in fusion welding, including arc welding and laser beam welding. UV-A arc welding for improving the welding processes have obtained extensive attention during the last decade. Such technology has three main electrode methods including gas tungsten arc welding (GTAW), shielded metal arc welding (SMAW), and gas metal arc welding (GMAW) [82]. In the UV-A arc welding system, ultrasonic vibration can be applied either on the electrode or on the workpiece, as illustrated in Figure 2.19. Specifically, Figure 2.19(a) shows the UV-A GTAW system set-up that axially attaches an ultrasonic radiator to the tungsten electrode for the direct transmission of ultrasonic vibration to the molten pool of the weld within the workpiece [83]. On the other hand, Figure

33 Texas Tech University, Fuda Ning, May 2018

2.19(b) shows the UV-A SMAW system set-up that applies the vertical-direction ultrasonic vibration directly to the bottom of the to-be-welded workpiece [84, 85].

Figure 2.19 Illustrations on the ultrasonic vibration implementation in UV-A arc welding system (after [83-85])

A significantly increased weld penetration can be achieved in the UV-A GTAW of stainless steel 304, as shown in Figure 2.20. Such phenomenon appears due to the fact that the continuously decreased arc pressure in the conventional GTAW process can be solved by realizing the longitudinal oscillation of the electrode into the molten pool in UV-A GTAW [83]. Similar results have also been reported elsewhere in arc welding of aluminum alloys [86]. In addition, ultrasonic vibration can lead to remarkably refined microstructures in the weld, associated with the increase in the tensile strength of the welded joint [87, 88]. Apart from the weld microstructure changing from columnar dendrite to fine equiaxed dendrite, unmixed zones can be fully removed resulting in the increased corrosion resistance in the welds fabricated by UV-A SMAW process [84, 85]. These benefits are attributed to the nonlinear effects induced by ultrasonic vibration, strongly influencing the liquid material solidification process.

In contrast to the common ultrasonic vibration manners presented above, a different approach for importing ultrasonic vibration to neither electrode nor the

34 Texas Tech University, Fuda Ning, May 2018 workpiece has been proposed in UV-A GMAW process [89, 90]. The researcher integrated the welding torch with the ultrasonic transducer and the filler metal was axially fed through the inner concentric hole. Thus an ultrasonic radiation field was imposed nearby the arc region. According to the reported results, the use of ultrasonic vibration gave rise to more stable and higher quality welds.

In addition to arc welding, Kim et al. [91] added the ultrasonic vibration to the bottom of the plate with the specimen on the top during laser beam welding. It was found that welding defects were greatly suppressed by the ultrasonic vibration due to cavitation effects in the molten pool during UV-A laser welding.

(a) (b)

2 mmm 2 mm

Figure 2.20 Weld appearance on the cross-section of AISI 304 sample (after [83])

2.3.5 UV-A laser cladding Laser cladding involves material melting and solidification behaviors to fabricate high-performance coatings. In recent years, ultrasonic vibration has been widely used in assisting the laser cladding process. Figure 2.21 shows the schematic of UV-A laser cladding process for coating fabrication [92]. The powders are fed into the coaxial feeding nozzle through a shield gas and they are then formed into a solid coating layer on the substrate by the laser radiation. During the cladding process, the vertical ultrasonic vibration is generated and transmitted from substrate to the molten pool of the coatings. In order to provide a distinct effect, the ultrasonic power supply should output enough power for the ultrasonic transducer to supply a sufficient ultrasonic amplitude on the substrate with small ultrasonic transmission attenuation [93].

35 Texas Tech University, Fuda Ning, May 2018

Reflective mirror Laser Focusing lens

Powder & shield gas Coaxial feeding nozzle

Coating

Vibration Substrate direction

Ultrasonic Ultrasonic power supply transducer

Figure 2.21 Schematic of UV-A laser cladding for coating fabrication (after [92])

Many investigations have been conducted to study the effects of ultrasonic vibration on the microstructural and mechanical performance of the coating produced by laser cladding process. Compared with laser cladding without ultrasonic vibration, UV-A laser cladding can result in the formation of yttria-stabilized zirconia ceramic coatings with a flatter geometry and finer microstructures on the Ti-6Al-4V plate [92]. The flat coating outline is considered to be caused due to the deformation of gas-liquid interface induced by the radiation pressure in the acoustic field [94]. In addition, the microstructural refinement is achieved as a consequence of dendrite fragmentation and nucleation associated with the bubble collapsing of acoustic streaming and cavitation. Besides microstructural refinement, more uniformly distributed elements can also be generated during UV-A laser cladding of nickel-based alloy Ni60 coating on the steel substrate [95]. This is ascribed to the increase of molten material fluidity in a homogenized temperature field induced by ultrasonic vibration [95, 96]. Fan monitored and collected the online temperature of the molten pool in UV-A laser cladding and laser cladding without ultrasonic vibration using a thermal infrared imager [97]. The measuring results are shown in Figure 2.22 and a smaller temperature fluctuation can be revealed in UV-A laser cladding process. The

36 Texas Tech University, Fuda Ning, May 2018 homogenized distribution of temperature field in the molten pool has been also verified by conducting the finite element simulation during UV-A laser cladding of TiC/FeAl composite coating [98].

(a) 3000 (b) 3000

2500 2500

/ /

e 2000

a r

r 1500 1500

p m

m 1000 1000

T T 500 500

0 0 0 19 39 59 80 103 123 143 166 190 210 0 19 39 59 80 103 123 143 166 190 213 Time/s Time/s Figure 2.22 Online temperature measurement in (a) laser cladding without ultrasonic vibration and (b) UV-A laser cladding (after [97])

In laser-clad coatings, cracking was a common fabrication defect that would greatly limit the wide applications of coatings. According to the reported works [96, 99], the cavitation and mixing functions of ultrasonic vibration could improve the solidification state, reduce the residual heat stress, and thus decrease the sensitivities of cracking. The evidence in cracking reduction in UV-A laser cladding has been also provided by Chen et al. [93]. This investigation presents that the decrease of internal tension stress resulted from dendrite fragmentation could alleviate or even eliminate the cracking around the interface. Therefore, UV-A laser cladding could be a potential strategy to obtain a larger bond between coating and base materials by ultrasonically mixing and stirring molten powders and substrate [92].

2.4 Concluding remarks Great efforts have been made in the past to utilize the desirable effects of ultrasonic vibration, causing the emergence of a new field of ultrasonic vibration- assisted (UV-A) manufacturing. In this chapter, a comprehensive and systematic review on UV-A manufacturing processes were conducted. The present review summarized fundamental knowledge on the mechanism of ultrasonic vibration’s actions in the manufacturing processes. In UV-A conventional machining processes,

37 Texas Tech University, Fuda Ning, May 2018 ultrasonic vibration could help to reduce the machining force through intermittent cutting between machine tools and the workpiece. Thus, surface quality and dimensional accuracy could be improved compared with those in conventional machining processes without ultrasonic vibration. Similarly, in other UV-A mechanical manufacturing processes such as densification and forming, the introduction of ultrasonic vibration could also decrease the friction between the machine and particles/parts, leading to the reduction of force in the processes. Owing to these benefits, better-arranged powder compacts with a higher compact density and forming parts with a better surface quality were achieved in UV-A densification and UV-A forming, respectively. Furthermore, in UAM, ultrasonic vibration served as an energy source to soften the foil surface with the breakage of oxide layers, resulting in a strong bond during the formation of 3D parts.

In UV-A thermal manufacturing processes, importing ultrasonic vibration into molten materials would generate nonlinear effects including acoustic streaming and cavitation. These effects could change the mass and heat transfer as well as the crystal growth during liquid melting material solidification and crystallization processes. Several indirect actions including stirring and mixing, crystal dispersion, and crystal nucleation were caused, which induced other ultimate influences on the liquid melting materials. The reported results showed that ultrasonic vibration could help to homogenize the chemical contents by liquid agitation, reduce segregation and prevent grains from growing, refine the microstructures, decrease the temperature gradient, arrange particles distribution, and reduce porosity and cracks. Consequently, the precision and quality of thermally manufactured parts would be improved through these phenomena. This chapter generates a significant body of knowledge on ultrasonic vibration influences in different UV-A manufacturing processes, which will be of great benefit for understanding the tasks presented in the following chapters.

References [1] Laugier, P., and Haïat, G., 2011, Introduction to the physics of ultrasound, In Bone quantitative ultrasound, pp. 29–45.

38 Texas Tech University, Fuda Ning, May 2018

[2] Ziskin, M.C., 1993, Fundamental physics of ultrasound and its propagation in tissue, Radiographics, 13(3), pp. 705–709. [3] Komarov, S.V., Kuwabara, M., and Abramov, O.V., 2005, High power ultrasonics in pyrometallurgy: current status and recent development, ISIJ international, 45(12), pp. 1765–1782. [4] Abramov, O.V., 1987, Action of high intensity ultrasound on solidifying metal, Ultrasonics, 25(2), pp. 73–82. [5] Masuko, M., and Kumabe, J., 1958, Study on the ultrasonic turning, Japanese Society of Precision Engineering, 1, pp. 56–60. (in Japanese) [6] Babitsky, V.I., Kalashnikov, A.N., Meadows, A., and Wijesundara, A.A.H.P. 2003, Ultrasonically assisted turning of aviation materials, Journal of Materials Processing Technology, 132, (1), pp. 157–167. [7] Shen, X.H., Zhang, J., Xing, D., and Zhao, Y., 2012, A study of surface roughness variation in ultrasonic vibration-assisted milling, Journal of Advanced Manufacturing Technology, 58(5–8), pp. 533–561. [8] Liao, Y.S., Chen, Y.C., and Lin, H.M., 2007, Feasibility study of the ultrasonic vibration assisted drilling of Inconel superalloy, International Journal of Machine Tools and Manufacture, 47(12), pp. 1988–1996. [9] Qin, N., Pei, Z.J., Treadwell, C., and Guo, D.M., 2009, Physics-based predictive cutting force model in ultrasonic-vibration-assisted grinding for titanium drilling, Journal of manufacturing science and engineering, 131(4), pp. 111–119. [10] Cong, W.L., and Ning, F.D., 2015, Chapter 2 Rotary Ultrasonic Machining of CFRP Composites, in Machinability of Fibre-Reinforced Plastics, J.P. Davim (Ed.), Walter de Gruyter GmbH & Co KG, pp. 31–81. [11] Jamshidi, H., and Nategh, M.J., 2013, Theoretical and experimental investigation of the frictional behavior of the tool-chip interface in ultrasonic vibration assisted turning, International Journal of Machine Tools and Manufacture, 65, pp. 1–7. [12] Xu, W.X., and Zhang, L.C., 2015, Ultrasonic vibration-assisted machining: principle, design and application, Advances in Manufacturing, 3(3), pp. 173–192. [13] Zhong, Z.W., and Lin, G., 2006, Ultrasonic assisted turning of an aluminum- based metal matrix composite reinforced with SiC particles, The International Journal of Advanced Manufacturing Technology, 27(11–12), pp. 1077–1081. [14] Cong, W.L., Pei, Z.J., Feng, Q., Deines, T.W., and Treadwell, C., 2012, Rotary ultrasonic machining of CFRP: a comparison with twist drilling, Journal of Reinforced Plastics and Composites, 31(5), pp.313–321. [15] Ning, F.D., Wang, H., Cong, W.L., and Fernando, P.K.S.C., 2017, A mechanistic ultrasonic vibration amplitude model during rotary ultrasonic machining of CFRP composites, Ultrasonics, 76, pp. 44–51. [16] Ning, F.D., and Cong, W.L., 2015, Rotary ultrasonic machining of CFRP: design of experiment with a cutting force model, in Proceedings of the 2015 International Manufacturing Science and Engineering Conference, June 8–12, 2015, Charlotte, USA, pp. V001T02A040–V001T02A048. [17] Ning, F.D., Cong, W.L., Pei, Z.J., and Treadwell, C., 2016, Rotary ultrasonic

39 Texas Tech University, Fuda Ning, May 2018

machining of CFRP: a comparison with grinding, Ultrasonics, 66, pp.125–132. [18] Ning, F.D., Wang, H., Hu, Y.B., Cong, W.L., Zhang, M., and Li, Y.Z., 2017, Rotary ultrasonic surface machining of CFRP composites: a comparison with conventional surface grinding, 45th SME North American Manufacturing Research Conference, June 4–8, 2017, Los Angeles, CA, USA. [19] Ning, F.D., Cong, W.L., Wang, H., Hu, Y.B., Hu, Z.L., and Pei, Z.J., 2017, Surface grinding of CFRP composites with rotary ultrasonic machining: a mechanistic model on cutting force in the feed direction, International Journal of Advanced Manufacturing Technology, 92(1-4), pp. 1217–1229. [20] Wang, H., Ning, F.D., Hu, Y.B., Fernando, P.K.S.C., Pei, Z.J., and Cong, W.L., 2016, Surface grinding of carbon fiber–reinforced plastic composites using rotary ultrasonic machining: Effects of tool variables, Advances in Mechanical Engineering, 8(9), pp. 1–14. [21] Cha, H.R., 2011, Densification of the nanopowder by using ultrasonic vibration compaction, Reviews on Advanced Materials Science, 28(1), pp. 90–93. [22] Tsujino, J., Ueoka, T., Suzuki, H., Shinuchi, S., and Hashimoto, K., 1991, Ultrasonic vibration press of metal and ceramic powders using complex vibration and vacuum condition, in Proceedings of IEEE International Ultrasonics Symposium, pp. 973–978. [23] Emeruwa, E., Jarrige, J., Mexmain, J., Billy, M., and Bouzouita, K., 1990, Powder compaction with ultrasonic assistance, Journal of Materials Science, 25(2), pp. 1459–1462. [24] Cha, H.R., and Nakamura, K., 2010, Analysis of densification mechanism under ultrasonic compaction using force balance model, Proceedings of Symposium on Ultrasonic Electronics, 31, pp. 521–522. [25] Kikuchi, S., Koyama, D., and Nakamura, K., 2008, A design of ultrasonic compaction tools for metal powder magnetic core of motors, in Proceedings of IEEE International Ultrasonics Symposium, pp. 2142–2145. [26] Khasanov, O.L., and Dvilis, E.S., 2008, Net shaping nanopowders with powerful ultrasonic action and methods of density distribution control, Advances in Applied Ceramics, 107(3), pp. 135–141. [27] Hess, J.R., Wright, C.T., and Kenney, K.L., 2007, Cellulosic biomass feedstocks and logistics for ethanol production, Biofuels, Bioproducts and Biorefining, 1(3), pp. 181–190. [28] Zhang, M., Song, X.X., Pei, Z.J., and Wang, D.H., 2015, Chapter 9 Ultrasonic Vibration-Assisted Pelleting of Cellulosic Biomass for Biofuel Production, In Production of Biofuels and Chemicals with Ultrasound, Springer, Netherlands, pp. 243–267. [29] Zhang, M., Song, X.X., Pei, Z.J., Deines, T.W., and Treadwell, C., 2012, Ultrasonic vibration-assisted pelleting of wheat straw: an experimental investigation, International Journal of Manufacturing Research, 7(1), pp. 59–71. [30] Li, Y.Z., Ning, F.D., Cong, W.L., Zhang, M., and Tang, Y.J., 2016, Investigating pellet charring and temperature in ultrasonic vibration-assisted pelleting of wheat straw for cellulosic biofuel manufacturing, Renewable Energy, 92, pp. 312–320.

40 Texas Tech University, Fuda Ning, May 2018

[31] Kuttruff, H., 1991, Ultrasonics Fundamentals and Applications, Elsevier Science Publishers, New York. [32] Zhang, P.F., Pei, Z.J., Wang, D.H., Wu, X.R., Cong, W.L., Zhang, M., and Deines, T., 2011, Ultrasonic vibration–assisted pelleting of cellulosic biomass for biofuel manufacturing, Journal of Manufacturing Science and Engineering, 133(1), pp. 1–7. [33] Liu, S.J., and Dung, Y.T., 2005, Ultrasonic vibration hot embossing: a novel technique for molding plastic microstructure, International Polymer Processing, 20(4), pp. 449–452. [34] Khuntontong, P., Blaser, T., and Schomburg, W.K., 2009, Fabrication of molded interconnection devices by ultrasonic hot embossing on thin polymer films, IEEE Transactions on Electronics Packaging Manufacturing, 32(3), pp. 152–156. [35] Liu, S.J., and Dung, Y.T., 2005, Hot embossing precise structure onto plastic plates by ultrasonic vibration, Polymer Engineering & Science, 45(7), pp. 915– 925. [36] Hung, J.C., Tsai, Y.P., and Hung, C., 2013, Development of a new apparatus for ultrasonic vibration-assisted glass hot embossing process, Precision Engineering, 37(1), pp. 222–227. [37] Tsai, Y.P., Hung, J.C., Yin, L.C., and Hung, C., 2012, Ultrasonic vibration- assisted optical glass hot embossing process, The International Journal of Advanced Manufacturing Technology, 60(9–12), pp. 1207–1213. [38] Bunget, C., and Ngaile, G., 2011, Influence of ultrasonic vibration on micro- extrusion, Ultrasonics, 51(5), pp. 606–616. [39] Mousavi, S.A., Feizi, H., and Madoliat, R., 2007, Investigations on the effects of ultrasonic vibrations in the extrusion process, Journal of Materials Processing Technology, 187, pp. 657–661. [40] Hung, J.C., and Huang, C.C., 2012, Evaluation of friction in ultrasonic vibration- assisted press forging using double cup extrusion tests, International Journal of Precision Engineering and Manufacturing, 13(12), pp. 2103–2108. [41] Murakawa, M., and Jin, M., 2001, The utility of radially and ultrasonically vibrated dies in the wire drawing process, Journal of Materials Processing Technology, 113(1), pp. 81–86. [42] Qi, H.Q., Yuan, J.B., and Xie, T., 2008, Investigations on the effects of ultrasonic vibrations in the wire drawing, In 2008 IEEE International Ultrasonics Symposium (IUS), Beijing, China, pp. 2134–2137. [43] Black, J.T., and Kohser, R.A., 2012, Materials and Processes in Manufacturing, John Wiley and Sons Inc. [44] Obielodan, J., and Stucker, B., 2014, A fabrication methodology for dual– material engineering structures using ultrasonic additive manufacturing, The International Journal of Advanced Manufacturing Technology, 70(1–4), pp. 277–284. [45] Hahnlen, R., and Dapino, M.J., 2014, NiTi–Al interface strength in ultrasonic additive manufacturing composites, Composites Part B: Engineering, 59, pp. 101–108.

41 Texas Tech University, Fuda Ning, May 2018

[46] Sriraman, M.R., Gonser, M., Fujii, H.T., Babu, S.S., and Bloss, M., 2011, Thermal transients during processing of materials by very high power ultrasonic additive manufacturing, Journal of Materials Processing Technology, 211(10), pp. 1650–1657. [47] Chua, C.K., and Leong, K.F., 2015, 3D Printing and Additive Manufacturing Principles and Applications, World Scientific Publishing Co. Pte. Ltd., Singapore. [48] Langenecker, B., 1966, Effects of ultrasound on deformation characteristics of metals, IEEE Transactions on Sonic and Ultrasonic, 13, pp. 1–8. [49] Yang, Y., Ram, G.J., and Stucker, B.E., 2009, Bond formation and fiber embedment during ultrasonic consolidation, Journal of Materials Processing Technology, 209(10), pp. 4915–4924. [50] Shimizu, S., Fujii, H.T., Sato, Y.S., Kokawa, H., Sriraman, M.R., and Babu, S.S., 2014, Mechanism of weld formation during very-high-power ultrasonic additive manufacturing of Al alloy 6061, Acta Materialia, 74, pp. 234–243. [51] Dehoff, R.R., and Babu, S.S., 2010, Characterization of interfacial microstructures in 3003 aluminum alloy blocks fabricated by ultrasonic additive manufacturing, Acta Materialia, 58(13), pp. 4305–4315. [52] Mariani, E., and Ghassemieh, E., 2010, Microstructure evolution of 6061 O Al alloy during ultrasonic consolidation: An insight from electron backscatter diffraction, Acta Materialia, 58(7), pp. 2492–2503. [53] Foster, D.R., Dapino, M.J., and Babu, S.S., 2013, Elastic constants of ultrasonic additive manufactured Al 3003–H18, Ultrasonics, 53(1), pp. 211–218. [54] Sojiphan, K., Sriraman, M.R., and Babu, S.S., 2010, Stability of microstructure in Al3003 builds made by very high power ultrasonic additive manufacturing, In Proceedings of the Solid Freeform Symposium, Austin, Texas, USA, pp. 362– 371. [55] Ram, G.J., Yang, Y., and Stucker, B.E., 2006, Effect of process parameters on bond formation during ultrasonic consolidation of aluminum alloy 3003, Journal of Manufacturing Systems, 25(3), pp. 221–238. [56] Hehr, A., and Dapino, M.J., 2015, Interfacial shear strength estimates of NiTi–Al matrix composites fabricated via ultrasonic additive manufacturing, Composites Part B: Engineering, 77, pp. 199–208. [57] Eskin, G.I., and Eskin, D.G., 2003, Production of natural and synthesized aluminum-based composite materials with the aid of ultrasonic (cavitation) treatment of the melt, Ultrasonics Sonochemistry, 10(4), pp. 297–301. [58] Zheng, H.Y., and Huang, H., 2007, Ultrasonic vibration–assisted femtosecond laser machining of microholes, Journal of Micromechanics and Microengineering, 17(8), pp. 58–61. [59] Chiu, C.C., Chang, C.H., and Lee, Y.C., 2010, Ultrasound assisted laser machining and surface cleaning, In 2010 5th IEEE International Conference on Nano/Micro Engineered and Molecular Systems (NEMS), pp. 872–875. [60] Alavi, S.H., and Harimkar, S.P., 2015, Melt expulsion during ultrasonic vibration-assisted laser surface processing of austenitic stainless steel,

42 Texas Tech University, Fuda Ning, May 2018

Ultrasonics, 59, pp. 21–30. [61] Biswas, S., Alavi, S.H., and Harimkar, S.P., 2015, Laser surface melting of Ti– 6Al–4V under the influence of ultrasonic vibrations, Materials Letters, 159, pp. 470–473. [62] Kang, B., Kim, G.W., Yang, M., Cho, S.H., and Park, J.K., 2012, A study on the effect of ultrasonic vibration in nanosecond laser machining, Optics and lasers in engineering, 50(12), pp. 1817–1822. [63] Abdullah, A., Shabgard, M.R., Ivanov, A., and Shervanyi-Tabar, M.T., 2009, Effect of ultrasonic-assisted EDM on the surface integrity of cemented tungsten carbide (WC-Co), International Journal of Advanced Manufacturing Technology, 41(3–4), pp. 268–280. [64] Cao, M.R., Yang, S.C., Li, W.H., and Yang, S.Q., 2008, Experimental research on small holes by electrical discharge machining combined with ultrasonic vibration and assisted inwall polish with the ultrasonic vibrating, Key Engineering Materials, 359, pp. 374–378. [65] Jia, Z.X., Zhang, J.H., and Ai, X., 1997, Study on a new kind of combined machining technology of ultrasonic machining and electrical discharge machining, International Journal of Machine Tools and Manufacture, 37(2), pp. 193–199. [66] Liu, Q., Zhai, Q., Qi, F., and Zhang, Y., 2007, Effects of power ultrasonic treatment on microstructure and mechanical properties of T10 steel, Materials letters, 61(11), pp. 2422–2425. [67] Aghayani, M.K., and Niroumand, B., 2011, Effects of ultrasonic treatment on microstructure and tensile strength of AZ91 magnesium alloy, Journal of Alloys and Compounds, 509(1), pp. 114–122. [68] Jian, X., Xu, H., Meek, T.T., and Han, Q., 2005, Effect of power ultrasound on solidification of aluminum A356 alloy, Materials Letters, 59(2–3), pp. 190–193. [69] Liu, X., Osawa, Y., Takamori, S., and Mukai, T., 2008, Grain refinement of AZ91 alloy by introducing ultrasonic vibration during solidification, Materials Letters, 62(17), pp. 2872–2875. [70] Nie, K.B., Wang, X.J., Wu, K., Zheng, M.Y., and Hu, X.S., 2011, Effect of ultrasonic vibration and solution heat treatment on microstructures and tensile properties of AZ91 alloy, Materials Science and Engineering: A, 528(25), pp. 7484–7487. [71] Cao, G., Choi, H., Konishi, H., Kou, S., Lakes, R., and Li, X., 2008, Mg– 6Zn/1.5% SiC nanocomposites fabricated by ultrasonic cavitation–based solidification processing, Journal of Materials Science, 43(16), pp. 5521–5526. [72] Mula, S., Padhi, P., Panigrahi, S.C., Pabi, S.K., and Ghosh, S., 2009, On structure and mechanical properties of ultrasonically cast Al–2% Al2O3 nanocomposite, Materials Research Bulletin, 44(5), pp. 1154–1160. [73] Nie, K.B., Wang, X.J., Hu, X.S., Xu, L., Wu, K., and Zheng, M.Y., 2011, Microstructure and mechanical properties of SiC nanoparticles reinforced magnesium matrix composites fabricated by ultrasonic vibration, Materials Science and Engineering: A, 528(15), pp. 5278–5282.

43 Texas Tech University, Fuda Ning, May 2018

[74] Su, H., Gao, W., Feng, Z., and Lu, Z., 2012, Processing, microstructure and tensile properties of nano–sized Al2O3 particle reinforced aluminum matrix composites, Materials & Design, 36, pp. 590–596. [75] Zhang, S., Zhao, Y., Cheng, X., Chen, G., and Dai, Q., 2009, High–energy ultrasonic field effects on the microstructure and mechanical behaviors of A356 alloy, Journal of Alloys and Compounds, 470(1), pp. 168–172. [76] Das, A., and Kotadia, H.R., 2011, Effect of high–intensity ultrasonic irradiation on the modification of solidification microstructure in a Si-rich hypoeutectic Al– Si alloy, Materials Chemistry and Physics, 125(3), pp. 853–859. [77] Shao, Z., Le, Q., Zhang, Z., and Cui, J., 2012, Effect of ultrasonic power on grain refinement and purification processing of AZ80 alloy by ultrasonic treatment, Metals and Materials International, 18(2), pp. 209–215. [78] Nagira, T., Nakatsuka, N., Yasuda, H., Uesugi, K., Takeuchi, A., and Suzuki, Y., 2015, Impact of melt convection induced by ultrasonic wave on dendrite growth in Sn–Bi alloys, Materials Letters, 150, pp. 135–138. [79] Suslick, K.S., Didenko, Y., Fang, M.M., Hyeon, T., Kolbeck, K.J., McNamara, W.B., Mdleleni, M.M., and Wong, M., 1999, Acoustic cavitation and its chemical consequences, Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences, 357(1751), pp. 335–353. [80] Shu, D., Sun, B., Mi, J., and Grant, P.S., 2012, A high-speed imaging and modeling study of dendrite fragmentation caused by ultrasonic cavitation, Metallurgical and Materials Transactions A, 43(10), pp. 3755–3766. [81] Xu, H., Han, Q., and Meek, T.T., 2008, Effects of ultrasonic vibration on degassing of aluminum alloys, Materials Science and Engineering: A, 473(1), pp. 96–104. [82] Cunha, T.V., and Bohórquez, C.E.N., 2015, Ultrasound in arc welding: A review, Ultrasonics, 56, pp. 201–209. [83] Sun, Q.J., Lin, S.B., Yang, C.L., and Zhao, G.Q., 2013, Penetration increase of AISI 304 using ultrasonic assisted tungsten inert gas welding, Science and Technology of Welding and Joining, 14(8), pp. 765–767. [84] Cui, Y., Xu, C.L., and Han, Q., 2006, Effect of ultrasonic vibration on unmixed zone formation, Scripta Materialia, 55(11), pp. 975–978. [85] Cui, Y., Xu, C., and Han, Q., 2007, Microstructure improvement in weld metal using ultrasonic vibrations, Advanced Engineering Materials, 9(3), pp. 161–163. [86] Dai, W.L., 2003, Effects of high-intensity ultrasonic-wave emission on the weldability of aluminum alloy 7075-T6, Materials Letters, 57(16), pp. 2447– 2454. [87] Watanabe, T., Ookawara, S., Seki, S., Yanagisawa, A., and Konuma, S., 2003, The effect of ultrasonic vibration on the mechanical properties of austenitic stainless steel weld, Quarterly Journal of the Japan Welding Society, 21(2), pp. 249–255. [88] Watanabe, T., Shiroki, M., Yanagisawa, A., and Sasaki, T., 2010, Improvement of mechanical properties of ferritic stainless steel weld metal by ultrasonic

44 Texas Tech University, Fuda Ning, May 2018

vibration, Journal of Materials Processing Technology, 210(12), pp. 1646–1651. [89] Fan, Y.Y., Fan, C.L., Yang, C.L., Liu, W.G., and Lin, S.B., 2013, Research on short circuiting transfer mode of ultrasonic assisted GMAW method, Science and Technology of Welding and Joining, 17(3), pp. 186–191. [90] Fan, C.L., Yang, C.L., Lin, S.B., and Fan, Y.Y., 2013, Arc characteristics of ultrasonic wave-assisted GMAW, Welding Journal, 92, pp. 375–380. [91] Kim, J.S., Watanabe, T., and Yoshida, Y., 1995, Ultrasonic vibration aided laser welding of Al alloys: improvement of laser welding-quality, Journal of Laser Applications, 7(1), pp. 38–46. [92] Wu, D.J., Guo, M.H., Ma, G.Y., and Niu, F.Y., 2015, Dilution characteristics of ultrasonic assisted laser clad yttria-stabilized zirconia coating, Materials Letters, 141, pp. 207–209. [93] Chen, C.Y., Deng, Q.L., and Song, J.L., 2005, Influence of Ni content and ultrasonic vibration to cracks in process of laser cladding, Journal of Nanjing University of Aeronautics and Astronautics, 37(11), pp. 44–48. (in Chinese) [94] Laborde, J.L., Hita, A., Caltagirone, J.P., and Gerard, A., 2000, Fluid dynamics phenomena induced by power ultrasounds, Ultrasonics, 38(1–8), pp. 297–300. [95] Shao, Y.L., Chen, X.P., Fu, D., and Zhang, H., 2014, Effects of synchronous ultrasonic vibration on microstructure and properties of laser cladding, Hot Working Technology, 43(10), pp. 160–162, 165. (in Chinese) [96] Deng, Q.L., Song, J.L., Chen, C.Y., and Hu, J.D., 2006, Cracking controlling methods for laser deposition formed metal parts, Chinese Patent, CN1737197 AB. (in Chinese) [97] Fan, P.X., Experimental study on laser cladding of titanium alloy with ultrasonic vibration. Shenyang Aerospace University, 2012. (in Chinese) [98] Li, D.Y., Zhao, L.Z., Zhang, J., and Jiang, X. 2015, Influence of ultrasonic vibration on temperature field of TiC/FeAl composite coating in laser cladding, Heat Treatment of Metals, 40(3), pp. 190–194. (in Chinese) [99] Song, J.L., Deng, Q.L., Ge, Z.J., Chen, C.Y., and Hu, D.J., 2006, The cracking control technology of laser rapid forming nickel-based alloys, Journal of Shanghai Jiaotong University, 40(3), pp. 548–551. (in Chinese)

45 Texas Tech University, Fuda Ning, May 2018

CHAPTER III

ULTRASONIC VIBRATION-ASSISTED (UV-A) HOLE-MAKING OF CFRP COMPOSITES: A COMPARISON WITH CONVENTIONAL GRINDING

Paper title:

Rotary ultrasonic machining of CFRP: a comparison with grinding

Published in:

Ultrasonics (2016), Vol. 66, pp. 125-132

Authors:

Fuda Ning1, Weilong Cong1, Zhijian Pei2, and Clyde Treadwell3

Authors’ affiliations:

1Department of Industrial, Manufacturing, and Systems Engineering, Texas Tech University, Lubbock, TX, USA.

2Department of Industrial and Systems Engineering, Texas A&M University, College Station, TX, USA.

3Sonic-Mill, 7500 Bluewater Road NW, Albuquerque, NM, USA.

46 Texas Tech University, Fuda Ning, May 2018

Abstract Carbon fiber reinforced plastic (CFRP) composites have been intensively used in various industries due to their superior properties. In aircraft and aerospace industry, a large number of holes are required to be drilled into CFRP components at the final stage for aircraft assembling. There are two major types of methods for hole making of CFRP composites in industry, twist drilling and its derived multi-points machining methods as well as grinding and its related methods. The first type of methods is commonly used in hole making of CFRP composites. However, in recent years, rotary ultrasonic machining (RUM), a hybrid machining process combining ultrasonic machining and grinding, has also been successfully used in drilling of CFRP composites. It has been shown that RUM is superior to twist drilling in many aspects. However, there are no reported investigations on comparisons between RUM and grinding in drilling of CFRP. In this chapter, these two drilling methods are compared in five aspects, including cutting force, torque, surface roughness, hole diameter, and material removal rate.

Keywords: Rotary ultrasonic machining (RUM); Carbon fiber reinforced plastic (CFRP) composite; Grinding; Drilling.

3.1 Introduction

3.1.1 Properties and applications of CFRP composites Carbon fiber reinforced plastic (CFRP) composites consist of two materials: carbon fibers and polymer. Within CFRP composites, carbon fibers are surrounded by the polymer matrix. The carbon fibers are used to support the load, while the polymer matrix is used to bind and protect the fibers and transfer the load to the reinforcing fibers [1-3].

CFRP composites have a variety of attractive properties, including low density (providing light-weight engineering solutions); high stiffness-to-weight ratio; excellent fatigue and wear resistance; high dimensional stability; and low friction coefficient, thermal expansion, electrical conductivity [1, 4, 5-8].

47 Texas Tech University, Fuda Ning, May 2018

Due to these superior properties, more and more CFRP composites have been widely used in many different applications, such as aerospace and commercial aircraft, automobile, sports goods, robot arms, bridges, chemical containers, and fishing rods [9-11]. Especially in aerospace and commercial aircraft industry, the usage of CFRP composites grows remarkably. For instance, in Boeing 787, the newest generation of commercial aircraft in the Boeing company, about 50% of the materials in weight are composites, leading to weight savings of 20% [12-13]. Besides, as the aircraft is engaged in corrosive environments, composites will not be easily subject to fatigue and corrosion damage due to their good fatigue and corrosion resistance, resulting in saving on maintenance costs [14].

3.1.2 Drilling in CFRP composites In CFRP composites applications, hole making is an important machining operation for assembly purposes. For example, a large number of holes need to be drilled for assembly of Boeing 787 aircraft [15]. Twist drilling, milling, and their derived methods were commonly used in the drilling of composites [16-18], but they had many drawbacks including high tool wear (short tool life), bad surface roughness, severe delamination, low hole accuracy, etc. To reduce or eliminate these problems, abrasive machining methods (grinding [19] and abrasive waterjet machining [20]), vibration assisted grinding [21], hybrid machining methods (rotary ultrasonic machining (RUM) [22-23]), and laser assisted drilling [24]) were also used in hole making of CFRP composite materials. However, abrasive waterjet machining had high consumable cost (slurry), and laser assisted twisting drilling consumed high energy. In addition, the hole accuracy of both methods was limited due to the differences between hole entrance and exit in abrasive waterjet machining and heat affected zone in laser-assisted drilling. It was reported that RUM of CFRP led to small cutting force, surface roughness, delamination, tool wear, etc. [25], and grinding in the drilling of CFRP had limited tool life and high abrasive wear of the tool [26]. However, there are no reported investigations on comparisons between RUM and grinding in the drilling of CFRP.

48 Texas Tech University, Fuda Ning, May 2018

RUM, a hybrid machining process, is a grinding process assisted with ultrasonic vibration, as illustrated in Figure 3.1. Compared with grinding, using RUM would not require significantly higher energy by analyzing energy consumption during RUM of CFRP composites [27]. The cutting tool is a core drill with metal-bonded abrasives. During RUM, the rotating tool vibrates ultrasonically (typically 20 kHz) in the tool axis direction and is fed towards the workpiece. Coolant is pumped through the center core of the cutting tool to wash out the swarf and prevent the cutting zone from overheating.

Figure 3.1 Illustration of rotary ultrasonic machining

Comparisons between two types of drilling methods (RUM and grinding) are conducted for the first time. This chapter makes comparisons in five aspects (cutting force, torque, surface roughness, hole diameter, and material removal rate). There are four sections in this chapter. Following this introduction section are experimental conditions and measurement procedures, experimental results and discussions, and conclusions.

49 Texas Tech University, Fuda Ning, May 2018

3.2 Experimental conditions and measurement procedures

3.2.1 Properties of workpiece material The CFRP workpiece used in this experiment, as illustrated in Figure 3.2, consisted of plain woven structured carbon fibers and the epoxy resin matrix. The carbon fiber yarn in the woven structure had an orientation of 0/90 degrees, a thickness of 0.2 mm, and a width of 2.5 mm. The workpiece contained 42 layers of carbon fibers and had the size of 200 mm × 150 mm × 16 mm. Specific properties of the workpiece material are listed in Table 3.1.

Carbon fiber

Epoxy resin

Figure 3.2 Illustration of Fiber structures in CFRP

Table 3.1 Properties of workpiece material Property Unit Value Density of CFRP kg/m3 1550 Hardness (Rockwell) HRB 70–75 Poisson’s Ratio (v12) - 0.34 Poisson’s Ratio (v13) - 0.34 Poisson’s Ratio (v23) - 0.42 Longitudinal Young’s modulus (E1) GPa 136 Transverse Young’s modulus (Et) GPa 10.5 In-plane shear modulus (G12) GPa 3.76 Density of epoxy matrix kg/m3 1200 Poisson’s ratio of epoxy matrix - 0.4 Young’s modulus of epoxy matrix GPa 4.5 2 Fracture toughness of epoxy matrix (Energy/Gc) J/m 500 Density of carbon fiber kg/m3 1800 Poisson’s ratio of carbon fiber - 0.3 Young’s modulus of carbon fiber GPa 230 2 Fracture toughness of carbon fiber (Energy/Gc) J/m 2

50 Texas Tech University, Fuda Ning, May 2018

3.2.2 Experimental set-up and conditions The experiments were performed on a rotary ultrasonic machine (Series 10, Sonic-Mill, Albuquerque, New Mexico, USA). The RUM experimental set-up included an ultrasonic spindle system, a data acquisition system, and a cooling system, as illustrated in Figure 3.3.

Figure 3.3 Rotary ultrasonic machining (RUM) experimental set-up

The ultrasonic spindle system was mainly comprised of an ultrasonic spindle, a power supply, and a motor speed controller. The power supply was used to convert conventional (60 Hz) electrical supply to high-frequency (20 kHz) electrical energy which was provided to a piezoelectric converter. The converter was located inside the ultrasonic spindle and could convert high-frequency electrical energy into high- frequency mechanical vibration. The ultrasonic vibration generated by the converter was amplified and transmitted to the cutting tool which was attached to the spindle, making the tool to vibrate vertically to the tool face at a high frequency. The

51 Texas Tech University, Fuda Ning, May 2018 amplitude of ultrasonic vibration was adjusted under different settings of output control of the power supply. The rotational motion of the tool was provided by the motor attached atop the ultrasonic spindle and various speeds were obtained by changing the motor speed controller. The data acquisition system mainly consisted of a dynamometer (Model 9272, Kistler Inc., Switzerland), a charge amplifier (Model 5070A, Kistler Inc., Switzerland), and an A/D converter. The data acquisition system in Figure 3.3 was involved in measuring the cutting force and torque during RUM of CFRP composites. The detailed information will be provided in Section 3.2.3. The cooling system was comprised of a pump, a coolant tank, a pressure regulator, flow rate and pressure gauges, and valves. Coolant was supplied to the spindle and the interface of machining.

The cutting tool used in both RUM and grinding was a metal-bonded diamond core drill (NBR Diamond tool corp., LaGrangeville, NY, USA), as shown in Figure 3.4. The detailed tool parameters are listed in Table 3.2.

O Abrasive D portion ID

Tool connection portion

th g n e l g in n u T

Figure 3.4 Illustration of the tool used in both RUM and grinding

The power and frequency of ultrasonic vibration in RUM were fixed at 40% (13.5 μm in amplitude for the tool used in this experiment) and 20 kHz, respectively, and there was no such vibration in grinding. Effects of tool rotation speed and feed rate on five output variables were investigated using both drilling methods, as shown

52 Texas Tech University, Fuda Ning, May 2018 in Table 3.3. The ranges of feed rate and tool rotation speed were chosen according to the experience from the authors’ preliminary experiments as well as the guidance from the RUM machine manufacturer, Sonic-Mill Inc. To make comparisons between RUM and grinding, the feed rate and tool rotation speed were kept the same. Four holes were drilled under each combination of input variables.

Table 3.2 Tool parameters Parameter Unit Value Outer diameter mm 9.6 Inner diameter mm 7.8 Tuning length mm 44.5 Abrasive material Diamond Grit size mesh # 60/80 Grain concentration 100 Number of slots 0 Bond B (metal)

Table 3.3 Machining conditions Feed rate (mm/s) Tool rotation speed (rpm) 0.1, 0.2, 0.3, 0.4 3000 0.5, 0.6, 0.7, 0.8 1000, 2000, 3000 0.5 4000, 5000

3.2.3 Measurement procedures for output variables The dynamometer was used for measuring cutting force in the axial direction and torque, and the electrical signals from it were amplified by the charge amplifier. Then the A/D converter transformed the electrical signals into digital signals, which would be collected by a data acquisition card (PC-CARD-DAS16/16, Measurement Computing Corporation, Norton, MA, USA) with the help of DynoWare software package (Type 2815A, Kistler Inc., Switzerland) on a computer. The sampling rate was 20 Hz during all experiments under the condition of 10 points being read per scan, where each data point would be an average value of 10 points, so that a large amount of data points could be avoided. As the maximum sampling rate of dynamometer system (6000 Hz) was much lower than the frequency of ultrasonic vibration, the detailed cutting force and torque affected by each cycle of ultrasonic vibration would not be explored [28-29].

53 Texas Tech University, Fuda Ning, May 2018

The measured cutting force fluctuated with time within a certain range. A typical curve of cutting force signal in time domain during RUM of CFRP is shown in Figure 3.5. Cutting force for drilling each hole was represented by the maximum cutting force value (Fz) [30-31]. The maximum value of cutting force was of the major concern since such instantaneous value would determine the maximum stress in the workpiece and influence the drilling operation performance, including the part accuracy, surface quality, tool life, etc.. The maximum value of cutting force was defined as the median of five values (two before, two after, and present data point) with the applied smooth function. Similarly, torque for drilling each hole was represented by the maximum torque value obtained by using the same principle.

Maximum cutting force

Figure 3.5 Typical relationship between cutting force and time (in RUM)

The machined hole and machined rod (material removed by RUM tool) in the two drilling methods were illustrated in Figure 3.6. A surface profilometer (Surftest- 402, Mitutoyo Corporation, Kanagawa, Japan) was used to measure surface roughness on the machined surface of each hole. The tested range and cut-off length were set as 4 mm and 0.8 mm, respectively. In this chapter, average surface roughness Ra was chosen to evaluate the machined hole surfaces. The measurement of Ra started at a location near the hole entrance and moved along the axial direction of the hole. Four measurements were performed with 90° between two adjacent measurements. Each

54 Texas Tech University, Fuda Ning, May 2018 measurement was repeated twice, leading to eight Ra values in total for each hole. The average of these eight values was used as the Ra value for each hole.

D r

Machined rod t

Machined hole

D Feed direction Entrance

Exit

Figure 3.6 Illustration of the machined hole and rod

The diameter of the machined hole was used to evaluate the hole quality. Hole diameter (D) was measured by a Vernier caliper (model IP-65, Mitutoyo Corp., Kanagawa, Japan). The measurements were conducted along two directions perpendicular to each other with each measurement repeated twice. Therefore, there were four values of hole diameter for each hole and the average of these four values was used to represent the hole diameter. The optical microscope (TS100, Nikon Corp., Tokyo, Japan) was also used to measure the hole diameter with the function of circular measurement to validate the measuring accuracy of the Vernier caliper.

Material removal rate (MRR) was calculated as the volume of removed material in form of cutting chips divided by machining time. It can be expressed by the following equation:

22  D/ 2  Dr / 2  h MRR   (1) T

where, D is the diameter of machined hole, h is the thickness of workpiece, T is the time for drilling the hole, and Dr is the diameter of machined rod, which was also measured by the Vernier caliper. Such machined rod could not be considered as

55 Texas Tech University, Fuda Ning, May 2018 the removed material in this equation since it was not removed in form of cutting chips.

3.3 Experimental results and discussions All the data points in Figures 3.7 – 3.16 are plotted by the average values from four drilled holes under the same condition. The maximum and minimum values of four holes were reflected in error bars.

3.3.1 Effects on cutting force A comparison of cutting force between RUM and grinding when tool rotation speed changed is shown in Figure 3.7. With the increase of tool rotation speed from 1000 to 5000 rpm, cutting force decreased in both RUM and grinding. In RUM, as feed rate was fixed, MRR (as well as the material removal volume by one abrasive particle V1) would not change. Increase in tool rotation speed resulted in an increase of effective cutting distance. To keep the material removed volume V1 unchanged, the indentation depth should decrease. In this case, effective cutting time Δt and max impact force F1 decreased accordingly [32]. In addition, cutting force can be calculated by [33]:

F  tfFi  ntfF1 (2) where,

n is the number of active abrasive grains on the end face of the cutting tool;

f is the frequency of ultrasonic vibration;

Fi is the interaction force.

Based on Equation 2, the number of active abrasive grains n and ultrasonic frequency f were constant, therefore, cutting force would decrease with the increase of tool rotation speed.

In grinding, with the increase of tool rotation speed, the penetration depth of the diamond grain into the workpiece material decreased. The penetration depth was

56 Texas Tech University, Fuda Ning, May 2018 associated with interaction force which determined the cutting force. The interaction force generated between diamond grains on the drill end surface and the workpiece material decreased with the decrease of penetration depth, leading to the decreased cutting force accordingly.

Figure 3.7 Cutting force comparison between RUM and grinding when tool rotation speed changed

For all different levels of tool rotation speed, cutting forces in RUM were always lower than those in grinding. The difference in cutting force between these two methods became smaller as tool rotation speed increased, and difference almost disappeared when tool rotation speed reached to 5000 rpm. During RUM process, high-frequency ultrasonic vibration was applied to the cutting tool making the abrasive grains penetrated into the workpiece. With the increase of tool rotation speed to 5000 rpm, the penetration depth was decreased to a quite small level and effective cutting distance was increased remarkably, which would significantly reduce or eliminate the influences of ultrasonic vibrations on the cutting force. The changes of cutting force in RUM and grinding within the range of tool rotation speed were 90 N and 125 N, respectively.

Figure 3.8 shows a comparison of cutting force between RUM and grinding

57 Texas Tech University, Fuda Ning, May 2018 under different settings of feed rate. As feed rate increased from 0.1 to 0.8 mm/s, cutting force for both RUM and grinding increased. According to Equation 1, material removal rate would increase with the increase of feed rate (h/T in the equation), which was caused by an increased penetration depth of the diamond grain into the workpiece material. With the similar reasons, the interaction force would increase and consequently led to the increase of the cutting force. Cutting force in grinding was higher than that in RUM and the difference between them was nearly the same (about 16 N) at almost all levels of feed rate. However, when feed rate was 0.1 mm/s, cutting forces in both RUM and grinding were nearly the same (about 115 N). When feed rate changed from 0.1 to 0.8 mm/s, the change of cutting force in RUM was 53 N which was lower than that in grinding (69 N).

Figure 3.8 Cutting force comparison between RUM and grinding when feed rate changed

Compared with grinding, RUM led to lower cutting force at all different levels of feed rate. In RUM, Fi in RUM was larger than that in grinding in a 1/ f period of time. However, t in RUM was much smaller than that in grinding. Therefore, Fti  in RUM was smaller than that in grinding, resulting in lower cutting force in RUM according to Equation 2.

58 Texas Tech University, Fuda Ning, May 2018

3.3.2 Effects on torque Figure 3.9 shows a torque comparison between RUM and grinding at different levels of tool rotation speed. For both RUM and grinding, torque decreased when tool rotation speed increased. Using grinding led to larger torque than using RUM at all levels of tool rotation speed. The difference in torque between RUM and grinding decreased remarkably (from 0.43 to 0.06 N∙m), when tool rotation speed increased from 1000 to 5000 rpm. With the increase of tool rotation speed, the change of torque in grinding was about 0.6 N∙m, which was larger than that in RUM (less than 0.2 N∙m).

Figure 3.9 Torque comparison between RUM and grinding when tool rotation speed changed

A comparison of torque between the two types of drilling methods when feed rate changed is shown in Figure 3.10. As feed rate increased from 0.1 to 0.8 mm/s, torque increased for both types of methods. At all levels of feed rate, torque in grinding was larger than that in RUM. The difference in torque between RUM and grinding was very small (about 0.03 N∙m) when feed rate was smaller than 0.3 mm/s. It became larger when feed rate increased from 0.3 to 0.6 mm/s, and remained almost constant (about 0.16 N∙m) when feed rate was larger than 0.6 mm/s. Within the range

59 Texas Tech University, Fuda Ning, May 2018 of feed rate, the change of torque in grinding was 0.3 N∙m, higher than that in RUM (0.18 N∙m).

Figure 3.10 Torque comparison between RUM and grinding when feed rate changed

The trends of effects of input variables on torque were similar to those on cutting force. As torque was correlative to cutting force during RUM and grinding, same explanations for the trends and comparisons of the cutting force between these two methods can be sued for those of torque.

3.3.3 Effects on surface roughness A comparison of surface roughness between RUM and grinding at different levels of tool rotation speed is shown in Figure 3.11. For both RUM and grinding, surface roughness decreased when tool rotation speed increased. With the increase of tool rotation speed, the machining interactions between cutting tool and workpiece within a certain period of time became more frequent, so that more convex materials could be removed by their further interaction with cutting tool. With this reason, the surface of the hole could be ground with lower surface roughness. Surface roughness of the hole generated by grinding was lower than that generated by RUM when tool rotation speed was 1000 rpm. When tool rotation speed increased from 2000 to 5000 rpm, using grinding led to higher surface roughness than using RUM. The difference

60 Texas Tech University, Fuda Ning, May 2018 in surface roughness between RUM and grinding was very small. Within the range of tool rotation speed, the change of surface roughness in RUM was 1 μm, larger than that in grinding (0.6 μm).

Figure 3.11 Surface roughness comparison between RUM and grinding when tool rotation speed changed

Figure 3.12 shows a comparison of surface roughness between the two types of drilling methods when feed rate changed. For both RUM and grinding, surface roughness increased with the increase of feed rate. As the cutting time decreased with the increase of feed rate in the same hole depth, the interaction time between hole surface and cutting tool decreased accordingly, leading to higher surface roughness. There was no obvious difference in surface roughness between RUM and grinding at different settings of feed rate. The changes in surface roughness for both RUM and grinding were about 0.65 μm.

61 Texas Tech University, Fuda Ning, May 2018

Figure 3.12 Surface roughness comparison between RUM and grinding when feed rate changed

3.3.4 Effects on hole diameter Hole diameter is one of the parameters to evaluate hole drilling consistency. Figure 3.13 compares the hole diameter between RUM and grinding at different levels of tool rotation speed. It can be seen that, when tool rotation speed increased, hole diameter decreased for both drilling methods. The possible reason is that with the increase of tool rotation speed, the roundness of tool rotation decreased, leading to a lower horizontal oscillation. As a result, hole diameter decreased. When tool rotation speed was low (1000 rpm), hole diameter in grinding was larger than that in RUM. However, at a higher tool rotation speed of 5000 rpm, hole diameter in RUM was larger than that in grinding. Hole diameters in both RUM and grinding were almost the same under other settings of tool rotation speed. The change in hole diameter for RUM (0.023 mm) was smaller than that for grinding (0.038 mm).

62 Texas Tech University, Fuda Ning, May 2018

Figure 3.13 Hole diameter comparison between RUM and grinding when tool rotation speed changed

Figure 3.14 shows a comparison of hole diameter between the two types of methods when feed rate changed. For both RUM and grinding, hole diameter decreased with the increase of feed rate. The cutting time decreased with the increase of feed rate so that materials removed from the cylindrical surface of the hole were less, resulting in decreased hole diameter. When feed rate was lower than 0.5 mm/s, hole diameter in grinding was larger than that in RUM. However, using grinding led to smaller hole diameter than using RUM when feed rate exceeded 0.5 mm/s. The change in hole diameter for RUM was 0.028 mm which is smaller than that for grinding (0.06 mm).

3.3.5 Effects on material removal rate A comparison of MRR between RUM and grinding at different settings of tool rotation speed is shown in Figure 3.15. It can be seen that with the increase of tool rotation speed, MRR in grinding decreased and MRR in RUM increased gently and then decreased. MRR in RUM was higher than that in grinding at high levels of tool rotation speed (4000 to 5000 rpm). When tool rotation speed was at low levels (1000 to 3000 rpm), MRRs in both RUM and grinding were about the same. The change of

63 Texas Tech University, Fuda Ning, May 2018

MRR in grinding (0.05 mm3/s) was larger than that in RUM (0.02 mm3/s) within the whole range of tool rotation speed.

Figure 3.14 Hole diameter comparison between RUM and grinding when feed rate changed

Figure 3.15 Material removal rate comparison between RUM and grinding when tool rotation speed changed

64 Texas Tech University, Fuda Ning, May 2018

Figure 3.16 shows a MRR comparison between RUM and grinding when feed rate changed. For both RUM and grinding, MRR increased linearly with the increase of feed rate. Both methods had no obvious difference in MRR at all levels of feed rate.

Figure 3.16 Material removal rate comparison between RUM and grinding when feed rate changed

3.4 Conclusions This chapter reported the comparisons between rotary ultrasonic machining (RUM) and grinding of CFRP for the first time. Five output variables were compared, including cutting force, torque, surface roughness, hole diameter, and material removal rate (MRR). The conclusions are drawn as follows:

(a) Cutting force, torque, surface roughness, and hole diameter decreased for both RUM and grinding with the increase of tool rotation speed. MRR in grinding decreased and MRR in RUM increased gently first, then decreased when tool rotation speed increased. With the increase of feed rate, cutting force, torque, surface roughness, and MRR increased, and hole diameter decreased, for both RUM and grinding. (b) Almost all the cutting force and torque in grinding were higher than those in RUM for both tool rotation speed and feed rate. In other words, equal MRR

65 Texas Tech University, Fuda Ning, May 2018

generated higher cutting force and torque in grinding than in RUM. (c) Surface roughness in grinding was higher than that in RUM under almost all conditions. This indicated that using grinding led to worse surface roughness. (d) The change of hole diameter in grinding was larger than that in RUM. This revealed that the hole size consistency was worse in grinding. (e) RUM had higher material removal rate than grinding at higher levels of tool rotation speed. MRR values in both RUM and grinding were nearly equal at all levels of feed rate and lower levels of tool rotation speed.

Compared with grinding, RUM performs better in the drilling of CFRP, including lower cutting force, lower torque, and better surface roughness. Compared with grinding machine, RUM machine required ultrasonic generation system and special tools (without welding joints). Such differences increased the investment of machine itself. However, the investment cost of RUM process mainly consisted of energy cost and tool cost. Tool cost of RUM was lower than that of grinding due to less tool wear resulting from lower cutting force and lower torque during RUM of CFRP. With the same energy cost, the total investment cost of RUM process would be decreased.

References [1] Chung, D.D.L., 2010, Composite Materials Science and Applications, 2nd edition, Springer London Ltd, London, UK. [2] Gay, D., Hoa, S.V., and Tsai, S.W., 2003, Composite Materials Design and Applications, 4th edition, CRC press, Boca Raton, FL, USA. [3] Kinet, D., Mégret, P., Goossen, K.W., et al., 2014, Fiber Bragg Grating Sensors toward Structural Health Monitoring in Composite Materials: Challenges and Solutions, Sensors, 14(4), pp. 7394–7419. [4] Tong, L., Mouritz, A.P., and Bannister, M.K., 2002, 3D Fibre Reinforced Polymer Composites, Elsevier Science Ltd, Oxford, UK. [5] Strong, A. B., 2008, Fundamentals of composites manufacturing: materials, methods and applications, 2nd edition, SME. [6] Park, K.Y., Choi, J.H., and Lee, D.G., 1995, Delamination-free and high efficiency drilling of carbon fiber reinforced plastics, Journal of Composite Materials, 29(15), pp. 1988–2002.

66 Texas Tech University, Fuda Ning, May 2018

[7] Davim, J.P., and Reis P., 2003, Drilling carbon fiber reinforced plastics manufactured by autoclave-experimental and statistical study, Materials and Design, 24(5), pp. 315–324. [8] Mallick, P.K., 1997, Composite Engineering Handbook, CRC Press, Marcel Dekker Inc., New York, NY, USA. [9] Sadat, A.B., 1995, Delamination and other types of damage of graphite/epoxy composite caused by machining, Applied Mechanics Division and the Materials Division: Machining of Advanced Materials, June 28–30, Los Angeles, CA, pp. 41–52. [10] Guu, Y.H., Hocheng, H., Tai, N.H., and Liu, S.Y., 2001, Effect of electrical discharge machining on the characteristics of carbon fiber reinforced carbon composites, Journal of Materials Science, 36(8), pp. 2037–2043. [11] Mazumdar, S. 2001, Composites manufacturing: materials, product, and process engineering, CRC Press, Marcel Dekker Inc., New York, NY, USA. [12] Quilter, A., 2001, Composites in aerospace applications. IHS White Paper, 444, pp. 1–3. [13] Boeing Co. website, 2014, 787 Dreamliner Program Fact Sheet, http://www.boeing.com/commercial/787family/programfacts.html. [14] IndustryWeek. website, 2012, How Composites are Strengthening the Aviation Industry, authored by Robert Yancey, http://www.industryweek.com/none/how- composites-are-strengthening-aviation-industry. [15] Boeing Co. website, 2006, Boeing 787 from the ground up, http://www.boeing.com/commercial/aeromagazine/articles/qtr_4_06/article_04_ 2.html. [16] Ramulu, M., Branson, T., and Kim, D., 2001, A study on the drilling of composite and titanium stacks, Composite Structure, 54(1), pp. 67–77. [17] Tsao, C.C. and Hocheng, H., 2004, Taguchi analysis of delamination associated with various drill bits in drilling of composite material, International Journal of Machine Tools and Manufacture, 44(10), pp. 1085–1090. [18] Abrate, S., and D.A. Walton, 1992, Machining of composite materials. Part I: Traditional methods, Composites Manufacturing, 3(2), pp. 75–83. [19] Soo, S.L., Shyha, I.S., Barnett, T., Aspinwall, D.K., and Sim, W.M., 2012, Grinding performance and workpiece integrity when superabrasive edge routing carbon fibre reinforced plastic (CFRP) composites, CIRP Annals-Manufacturing Technology, 61(1), pp. 295–298. [20] Kalla, D.K., Zhang, B., Asmatulu, R., and Dhanasekaran, P.S., 2012, Current research trends in abrasive waterjet machining of fiber reinforced composites, Materials Science Forum, 713, pp. 37–42. [21] Wang, X., Wang, L.J., and Tao, J.P., 2004, Investigation on thrust in vibration drilling of fiber-reinforced plastics, Journal of Materials Processing Technology, 148(2), pp. 239–244. [22] Li, Z.C., Pei, Z.J., Sisco, T., Micale, A.C., and Treadwell, C., 2007, Experimental study on rotary ultrasonic machining of graphite/epoxy panel,

67 Texas Tech University, Fuda Ning, May 2018

Proceedings of the ASPE 2007 Spring Topical Meeting on Vibration Assisted Machining Technology, Chapel Hill, NC, April 16–17, pp. 52–57. [23] Cong, W.L., Feng, Q., Pei, Z.J., Deines, T.W., and Treadwell, C., 2011, Dry machining of carbon fiber reinforced plastic composite by rotary ultrasonic machining: effects of machining variables, Proceedings of the ASME 2011 International Manufacturing Science and Engineering Conference (MSEC 2011), Corvallis, OR, USA, June 13–17. [24] Stock, J., Zaeh, M.F., and Conrad, M., 2012, Remote laser cutting of CFRP: improvements in the cut surface, Physics Procedia, 39, pp. 161–170. [25] Cong, W.L., Pei, Z.J., Feng, Q., Deines, T.W., and Treadwell, C., 2012, Rotary ultrasonic machining of CFRP: a comparison with twist drilling, Journal of Reinforced Plastics and Composite, 31(5), pp. 313–321. [26] Biermann, D., and Feldhoff, M., 2012, Abrasive points for drill grinding of carbon fibre reinforced thermoset, CIRP Annals-Manufacturing Technology, 61(1), pp. 299–302. [27] Cong, W.L., Pei, Z.J., Deines, T.W., Srivastava, A., Riley, L., and Treadwell, C., 2012, Rotary ultrasonic machining of CFRP composites: a study on power consumption, Ultrasonics, 52(8), pp. 1030–1037. [28] Feng, Q., Cong, W.L., Pei, Z.J., and Ren, C.Z., 2012, Rotary ultrasonic machining of carbon fiber-reinforced polymer: feasibility study, Machining Science and Technology, 16(3), pp. 380–398. [29] Liu, D., Cong, W.L., Pei, Z.J., and Tang, Y.J., 2012, A cutting force model for rotary ultrasonic machining of brittle materials, International Journal of Machine Tools and Manufacture, 52(1), pp. 77–84. [30] Cong, W.L., Pei, Z.J., and Treadwell, C., 2014, Preliminary study on rotary ultrasonic machining of CFRP/Ti stacks, Ultrasonics, 54(6), pp. 1594–1602. [31] Cong, W.L., Pei, Z.J., Deines, T.W., Liu, D.F., and Treadwell, C., 2013, Rotary ultrasonic machining of CFRP/Ti stacks using variable feedrate. Composites Part B: Engineering, 52, pp.303–310. [32] Cong, W.L., Pei, Z.J., Sun, X., and Zhang, C.L., 2014, Rotary ultrasonic machining of CFRP: A mechanistic predictive model for cutting force, Ultrasonics, 54(2), pp. 663–675. [33] Pei, Z.J., Prabhakar, D., Ferreira, P.M., and Haselkorn, M., 1995, A mechanistic approach to the prediction of material removal rates in rotary ultrasonic machining, Journal of Engineering for Industry, 117(2), pp. 142–151.

68 Texas Tech University, Fuda Ning, May 2018

CHAPTER IV

ULTRASONIC VIBRATION-ASSISTED (UV-A) HOLE-MAKING OF CFRP COMPOSITES: DESIGN OF EXPERIMENT WITH A CUTTING FORCE MODEL

Paper title:

Rotary ultrasonic machining of CFRP: design of experiment with a cutting force model

Published in:

Proceedings of the ASME 2015 International Manufacturing Science and Engineering Conference (MSEC2015), June 8-12, 2015, Charlotte, North Carolina, USA

Authors:

Fuda Ning and Weilong Cong

Authors’ affiliations:

Department of Industrial, Manufacturing, and Systems Engineering, Texas Tech University, Lubbock, TX, USA

69 Texas Tech University, Fuda Ning, May 2018

Abstract Drilling is one of very important machining processes in many applications of carbon fiber reinforced plastic (CFRP) composites. Rotary ultrasonic machining (RUM) has been successfully used in the drilling of CFRP composites to overcome poor machinability. Cutting force is one of the most important output variables for evaluating drilling process since it will greatly influence cutting temperature, tool wear, and surface conditions. Currently, there are no reported investigations on the effects of input variables on cutting force using design of experiment (DOE) method in RUM of CFRP composites. The five-variable two-level full factorial design has been conducted to study cutting force based on a mechanistic predictive model in RUM of CFRP composites. Main effects, as well as interaction effects of five process variables (vibration amplitude, tool rotation speed, feedrate, abrasive size, and abrasive concentration) on cutting force, are revealed.

Keywords: Rotary ultrasonic machining (RUM); Carbon fiber reinforced plastic (CFRP) composite; Design of experiment (DOE); Cutting force; Modeling.

4.1 Introduction

4.1.1 Drilling of CFRP composites Drilling is one of the most widely used machining processes, accounting for almost 40% of all the material removal operations in the aerospace industry [1]. The parts or components made out of CFRP composite after molding processes are usually near net shaped, however, additional drilling process is required at the final stage for future parts assembly integration [2, 3]. Due to CFRP composites’ superior properties of high strength-to-weight ratio, excellent fatigue and wear resistance, and good dimensional stability [3, 4], CFRP composites have a poor machinability during the drilling process, such as short tool life, fiber breakage, matrix cracking, surface delamination, and geometrical defects [5-8].

To solve some of these problems in drilling of CFRP composites, RUM was proposed to be used. Experimental investigations on comparisons between RUM and

70 Texas Tech University, Fuda Ning, May 2018 twist drilling [9] as well as between RUM and abrasive grinding [10] have been conducted. The results showed that compared with twist drilling and abrasive grinding methods, RUM is superior in most aspects of output variables.

RUM, a hybrid machining process, is an abrasive grinding process assisted with ultrasonic vibration, as illustrated in Figure 4.1. The cutting tool is a core drill with metal-bonded abrasives. During RUM, the rotating tool vibrates ultrasonically (typically 20 kHz) in the vertical axis and feeds towards the workpiece. Coolant is pumped through the center core of the cutting tool to wash out the swarf and prevent the cutting zone from overheating.

Rotation Coolant flow in

Ultrasonic vibration Feeding

Coolant Coolant flow out flow out

Workpiece Abrasive portion Figure 4.1 Illustration of rotary ultrasonic machining

4.1.2 Cutting force in RUM of CFRP composites Cutting force is one of the most important output variables for evaluating drilling process. Cutting force will directly affect output variables including cutting temperature, tool wear, and surface conditions [11], which brings about high parts quality and low machining cost [12].

71 Texas Tech University, Fuda Ning, May 2018

Experimental investigations have been carried out to study the effects of machining variables (ultrasonic power, spindle speed, feedrate, drill geometry, abrasive size, and coolant type) on output variables (cutting force, torque, surface roughness, tool wear, and material removal rate) in RUM of CFRP composites [13- 15]. Cong et al. [16] built a mechanistic predictive model that can explain the experimental phenomena and predict effects of various input variables on cutting force. Pilot experiments have been conducted to prove that such model can correctly reflect the relationship between input variables and cutting force. However, there are no reported investigations on the design of experiment (DOE) of cutting force in RUM of CFRP composites. To decrease the machining cost and energy, this chapter is going to study cutting force using DOE method based on the mechanistic predictive model.

4.1.3 Purpose of the chapter This chapter, for the first time, reports DOE according to the mechanistic predictive model to investigate the main effects as well as interaction effects on the cutting force in RUM of CFRP composites. A 25 (five-variable two-level) full factorial experiment is designed to investigate the main effects as well as interaction effects of five input variables (vibration amplitude, tool rotation speed, feedrate, abrasive size, and abrasive concentration) on cutting force in RUM of CFRP composites.

There are four sections in this chapter. Following with the introduction is descriptions for mechanistic predictive model and design of experiments. After that, experimental results will be analyzed in three aspects, respectively, including main effects, two-factor interaction effects, and three-factor interaction effects on cutting force. Finally, conclusions in this chapter will be presented.

4.2 Mechanistic predictive model and design of experiments

4.2.1 Mechanistic predictive model for cutting force in RUM of CFRP composites RUM can be treated as a combination of ultrasonic machining process and abrasive grinding process. A mechanistic predictive model considering ultrasonic

72 Texas Tech University, Fuda Ning, May 2018 machining as the predominant process was developed by Cong et al [16]. It was established by analyzing one particle first and then integrating all active abrasive particles those were engaged in machining to find out the relationship between input variables and cutting force. Machining of CFRP composites was based on brittle fracture and chip separation occurs by brittle fracture [17]. The main model developing steps are shown in Figure 4.2. All the input variables in this model are listed in Table 4.1.

Cutting force Input variables (F) Machining and Workpiece tool variables properties

Material removal Max impact force, rate (MRR) (Fi)

Material CFRP removal Number of micromechanics analysis active process abrasives (n)

E, ν, Kc Removed volume Indentation depth, for one abrasive (δ) (V1)

Fracture volume factor (Kv) Figure 4.2 Model developing processes

Table 4.1 Input variables in the cutting force model for CFRP composites Categories Input variables Unit Elastic modulus, E MPa Workpiece properties Poisson’s ratio, v Fracture toughness, Kc MPa  mm Outer diameter, Do mm Inner diameter, Di mm Tool variables Abrasive concentration, C Abrasive size, d mm Amplitude, A mm Frequency, f Hz Machining variables Feedrate, Fr mm/s Tool rotation speed, S rpm

In this model, the maximum impact force for one abrasive grain taking part in cutting was expressed as:

73 Texas Tech University, Fuda Ning, May 2018

F 81/ 2 Ed 1/ 2 3 / 2 F  i  (1) 1 n 3(1 2 )

where,

Fi is the maximum impact force between tool and workpiece.

δ is the indentation depth, and it could be expressed as:

2 1/ 3  9 (F / n)2 1 2      i    (2) 16 d / 2  E     

The number of active abrasive grains, n, on the end face of cutting tool can be obtained by:

2 2 0.88 103 CC3 3 aa4 (3) nAA33006.561 10  (π / 6)100dd

where,

Ca is the abrasive concentration;

 is the density of abrasive material, g/mm3,   3.52103 g/mm 3 for diamond;

2 2 2 A0 is the area of the cutting tool end face, mm , A0  π(Do  Di )/4 , ( Do and Di are the outer and inner diameters of cutting tool, respectively, mm).

Cutting force F could be expressed as:

1 1    1 1    81/ 2 nEd 1/ 2 3 / 2 F  n  arcsin 1 F   arcsin 1 (4)    1    2 2 π  A 2 π  A 3(1 ) Moreover, the material removal volume by one abrasive partial could be calculated by regarding material removal as the result of the brittle fracture. At the

74 Texas Tech University, Fuda Ning, May 2018 interface of abrasive grain and workpiece surface, the abrasive grain has a maximum indentation value δ in a specific period of time. This phenomenon results in the change of lateral crack length CL and lateral crack depth CH, as illustrated in Figure 4.3. Lateral cracks will form and propagate during the indentation of the material. The material will be removed from the workpiece if there are two adjacent indentations generating the lateral cracks.

Figure 4.3 Material removal mechanism by one abrasive partial

The material removal volume V1 by one abrasive could be determined as follows by simplifying the fracture zone.

3/ 4 1  F  1/ 2  DS π     1  2 (5) V1  KV  d      arcsin1  3  KC  60 f 2  A

where, KV is fracture volume factor that is a proportionality parameter.

Material removal rate (MRR) could be calculated either based on material removed volume and feedrate or through the summation of MRR of all abrasive particles on the end face of the cutting tool. The equations could be expressed as follows, respectively.

π(D 2  D 2 ) MRR  F A  o i F (6) r 0 4 r

75 Texas Tech University, Fuda Ning, May 2018

3/4 1/2 nππ S F1 2  MRR KV( D o  D i ) d    arcsin 1  (7) 180KAC  2  After equating Eq. (6) and Eq. (7), the obtained Eq. (8) can be associated with Eq. (4):

3/ 4  2 2 1/4 2 1/4 π(D  D ) n π S  F  2 1/ 2 1 1    o i    Fr  KV (Do  Di )  d     arcsin1   4 180 K 2 π  A   C    (8) 81/ 2 nEd 1/ 2 3/2 1 1     F   arcsin 1 2     3(1 ) 2 π  A

In this simultaneous equation, the indentation depth δ and cutting force F are unknowns. The further detailed information can be referred to the paper on this mechanistic predictive model [16].

4.2.2 Experimental verification for mechanistic predictive model A series of experiments were conducted to verify this mechanistic model by testing effects of input variables (vibration amplitude, tool rotation speed, and feedrate) on cutting force, as listed in Table 4.2. Parameters of tool variables were kept constant using only one tool. A total of ten experiments were conducted by varying each machining variable and keeping other variables the same. The comparisons of predicted and experimental results are shown in Figure 4.4. It can be seen that the trends of predicted results on effects of input variables (vibration amplitude, tool rotation speed, and feedrate) on cutting force agreed well with the trends of experimental results.

Table 4.2 Experimental conditions for model verification Input variable Unit Value Vibration amplitude, A mm 0.0035; 0.0135; 0.0225; 0.03 Tool rotation speed, S rpm 1000; 2000; 3000; 4000; 5000 Feedrate, Fr mm/s 0.1; 0.3; 0.5; 0.7; 0.9

76 Texas Tech University, Fuda Ning, May 2018

(a) Effect of vibration amplitude (b) Effect of tool rotation speed

(S = 3000 rpm; Fr = 0.5 mm/s) (A = 0.0135 mm; Fr = 0.5 mm/s)

4.2.3 Design of Experiments In this chapter, a 25 full factorial design was employed to test the effect of five input variables on cutting force at two levels (low and high), as illustrated in Table 4.3. These input variables include:

Table 4.3 Low and high levels of input variables Low level High level Input variables Unit (-1) (+1) Vibration amplitude, A mm 0.0035 0.03 Tool rotation speed, S rpm 2000 6000 Feedrate, Fr mm/s 0.1 0.8 Abrasive size, d mm 0.048 0.201 Abrasive concentration, C 50 150

77 Texas Tech University, Fuda Ning, May 2018

Vibration amplitude: Maximum absolute value of the ultrasonic vibration measured from the position of equilibrium; Tool rotation speed: Rotational speed of cutting tool; Feedrate: Feedrate of cutting tool; Abrasive size: Diameter of abrasive in the cutting tool; and Abrasive concentration: The number of abrasives in a unit volume.

Cutting force values have been calculated under the total 32 different combinations of the tests, as listed in Table 4.4.

Table 4.4 Design matrix and experimental results Tool Vibration Abrasive Abrasive Cutting rotation Feedrate No. amplitude size concentration force speed Fr (mm/s) A (mm) d (mm) C F (N) S (rpm) 1 -1 -1 -1 -1 -1 70.17 2 1 -1 -1 -1 -1 66.25 3 -1 1 -1 -1 -1 24.83 4 1 1 -1 -1 -1 23.39 5 -1 -1 1 -1 -1 506.68 6 1 -1 1 -1 -1 483.82 7 -1 1 1 -1 -1 177.85 8 1 1 1 -1 -1 168.56 9 -1 -1 -1 1 -1 37.18 10 1 -1 -1 1 -1 34.93 11 -1 1 -1 1 -1 13.18 12 1 1 -1 1 -1 12.37 13 -1 -1 1 1 -1 267.44 14 1 -1 1 1 -1 251.21 15 -1 1 1 1 -1 94.09 16 1 1 1 1 -1 88.42 17 -1 -1 -1 -1 1 73.00 18 1 -1 -1 -1 1 68.81 19 -1 1 -1 -1 1 25.87 20 1 1 -1 -1 1 24.35 21 -1 -1 1 -1 1 524.05 22 1 -1 1 -1 1 497.66 23 -1 1 1 -1 1 184.66 24 1 1 1 -1 1 174.50 25 -1 -1 -1 1 1 38.72 26 1 -1 -1 1 1 36.37 27 -1 1 -1 1 1 13.74 28 1 1 -1 1 1 12.90 29 -1 -1 1 1 1 277.03 30 1 -1 1 1 1 260.33 31 -1 1 1 1 1 97.82 32 1 1 1 1 1 91.91

78 Texas Tech University, Fuda Ning, May 2018

A commercial software, Design-Expert (version 9.0, Stat-Ease Corp., Minneapolis, MN, USA), was used for data processing. The significant effects of all the variables can be identified by functional module ANOVA (analysis of variance), as shown in Table 4.5. The significant level was set at 0.05.

Table 4.5 Analysis of variance for full factorial model

Source Mean Square F-value P-value A-Vibration amplitude 532.48 1020.83 < 0.0001* B-Tool rotation speed 160352.83 307415 < 0.0001* C-Feedrate 398273.67 763536 < 0.0001* D-Abrasive size 67237.09 128901 < 0.0001* E-Abrasive concentration 206.82 396.49 < 0.0001* AB 109.65 210.22 < 0.0001* AC 287.48 551.13 < 0.0001* AD 26.30 50.43 0.0004* AE 0.98 1.89 0.2187 BC 91977.73 176332 < 0.0001* BD 15648.45 29999.87 < 0.0001* BE 38.80 74.39 0.0001* CD 38660.58 74116.75 < 0.0001* CE 106.74 204.63 < 0.0001* DE 14.27 27.35 0.0020* ABC 57.87 110.95 < 0.0001* ABD 3.52 6.75 0.0408* ABE 0.31 0.59 0.4707 ACD 11.69 22.41 0.0032* ACE 0.66 1.27 0.3021 ADE 0.48 0.93 0.3731 BCD 9048.39 17346.81 < 0.0001* BCE 18.99 36.40 0.0009* BDE 2.25 4.31 0.0831 CDE 6.76 12.95 0.0114* *indicating significant model terms when p-values are less than 0.05

4.3 Experimental results and discussions

4.3.1 Main effects on cutting force It can be seen from Table 4.5 that main effects of all five input variables were significant (all P-values < 0.05). The most significant input variable was feedrate, followed by tool rotation speed, abrasive size, vibration amplitude, and abrasive

79 Texas Tech University, Fuda Ning, May 2018

concentration (according to F-values). These main effects of each input variable on cutting force are illustrated in Figure 4.5. It can be observed that cutting force decreased as vibration amplitude, tool rotation speed, and abrasive size increased, and as feedrate and abrasive concentration decreased.

The main effects of vibration amplitude, feedrate, and tool rotation speed were consistent with those reported in the investigations on RUM of different materials, including composites (CFRP composites [18] and ceramic matrix composites [19]), brittle materials (silicon [20], silicon carbide [21], and ceramics [22]), and ductile materials (aerospace stainless steel [23] and titanium alloy [24]). However, for the main effect of abrasive concentration, the trend in this model prediction was not consistent with that reported by Churi et al [21] on RUM of silicon carbide. The inconsistency mainly resulted from different material properties. Besides, it may be owed to the ideal case that the change of abrasive concentration will not affect any process condition. In the actual experiments conducted by Churi et al. [21], the two grinding tools with different abrasive concentrations hardly maintained the same condition, leading to an inconsistent effect on cutting force.

300 Vibration amplitude Tool rotation speed Feedrate 250

200

150

100

)

( 50

c r

o 0 f

Low High Low High Low High g 300

n i

t Abrasive size Concentration t

u 250 C

200

150

100

0 Low High Low High

Figure 4.5 Main effects of variables on cutting force

80 Texas Tech University, Fuda Ning, May 2018

4.3.2 Two-factor interaction effects on cutting force All the two-factor interaction effects on cutting force, as presented in Figure 4.6, were significant except the interaction effect between vibration amplitude and abrasive concentration. From F-values in Table 4.5, it can be seen that the most significant interaction effect was that between tool rotation speed and feedrate, followed by that between feedrate and abrasive size. The least significant term was the interaction between vibration amplitude and abrasive concentration.

At both levels of abrasive concentration, cutting force decreased as tool rotation speed and abrasive size increased, and as feedrate decreased. The cutting force changed at a similar rate for both levels of abrasive concentration with changing of vibration amplitude, tool rotation speed, feedrate, and abrasive size. It also can be seen that there were almost no differences in cutting force for both levels of abrasive concentration.

For both levels of abrasive size, cutting force decreased with increasing of vibration amplitude, increasing of tool rotation speed, and decreasing of feedrate. The cutting force decreased at a similar rate with changing of vibration amplitude for both levels of abrasive size. At low level of abrasive size, cutting force decreased at a faster rate with increasing of tool rotation speed and decreasing of feedrate. Low level of abrasive size led to high cutting force for both levels of vibration amplitude, tool rotation speed, and feedrate.

For both levels of feedrate, cutting force decreased with increasing of vibration amplitude and tool rotation speed. High level of feedrate resulted in high cutting force for both levels of vibration amplitude and tool rotation speed. The cutting force decreased at a similar rate with changing of vibration amplitude for both levels of feedrate. At high level of abrasive size, cutting force decreased at a faster rate with increasing of tool rotation speed.

At both levels of tool rotation speed, cutting force decreased with increasing of vibration amplitude. Low level of tool rotation speed resulted in high cutting force for

81 Texas Tech University, Fuda Ning, May 2018

the input variable of vibration amplitude. The cutting force decreased at a similar rate with increasing of vibration amplitude for both levels of tool rotation speed.

Low High Low High Low High Low High 400 320 # 240 Concentration

160 C 80

0 400 320 Abrasive

) 240 size

160 ( d

e 80

c r

o f

400 g 320

n Feedrate

i t

t 240 Fr

160 C 80

0 400

320 Tool

240 rotation

160 speed S 80

Vibration amplitude A Figure 4.6 Two-factor interaction effectsInput vofar variablesiables on cutting force

4.3.3 Three-factor interaction effects on cutting force Six of all the three-factor interaction effects on cutting force were significant according to the P-value in Table 4.5. The most significant term was the interaction effect among tool rotation speed, feedrate, and abrasive size, followed by the combination of vibration amplitude, tool rotation speed, and feedrate. The best combination (when the cutting force was the lowest) for different three-factor interactions was high level of tool rotation speed, low level of feedrate, and high level of abrasive size (leading to 13.0463 N cutting force), as illustrated in Figure 4.7. Other three-factor best combinations were listed in Table 4.6.

82 Design-Expert® Software Texas Tech University, Fuda Ning, May 2018 Factor Coding: Actual Cutting force (N) Cutting force (N) X1 = B: Tool rotation speed X2 = C: Feedrate X3 = D: Abrasive size 264.003 93.0593

Actual Factors A: Vibration amplitude = 0 E: Concentration = 0

High 503.055 176.394

36.8021 13.0463 High

Feedrate

Abrasive size

Low 69.5585 24.6101 Low Low High Tool rotation speed Figure 4.7 The best combination for making the lowest cutting force at a three-factor interaction

Table 4.6 Three-factor interaction effects analysis Lowest cutting Three-factor Best force interaction terms combinations value (N) ABC A(+1)B(+1)C(-1) 18.2525 ABD A(+1)B(+1)D(+1) 51.3996 ACD A(+1)C(-1)D(+1) 24.1445 BCD B(+1)C(-1)D(+1) 13.0463* BCE B(+1)C(-1)E(-1) 18.4432 CDE C(-1)D(+1)E(-1) 24.4165 A-Vibration amplitude; B-Tool rotation speed; C-Feedrate; D-Abrasive size; E- Abrasive concentration

4.4 Conclusions In this chapter, a 25 full factorial design is employed to test the effects of five input variables (vibration amplitude, tool rotation speed, feedrate, abrasive size, and abrasive concentration) on cutting force at two levels (low and high). The cutting force is calculated according to the mechanistic predictive model in RUM of CFRP composites. The main effects, as well as interaction effects among the variables, are investigated. The specific conclusions will be illustrated as follows:

(1) The most significant input variable is feedrate, followed by tool rotation speed, abrasive size, vibration amplitude, and abrasive concentration. The cutting

83 Texas Tech University, Fuda Ning, May 2018 force decreased as vibration amplitude, tool rotation speed, and abrasive size increased, and as feedrate and abrasive concentration decreased.

(2) All the two-factor interaction effects on cutting force are significant except for the combination of vibration amplitude and abrasive concentration with a p-value of 0.2187, larger than 0.05.

(3) Six of the three-factor interaction effects on cutting force are significant. The most significant interaction effect is the combination of tool rotation speed, feedrate, and abrasive size. The best combination for different three-factor interactions is high level of tool rotation speed, low level of feedrate, and high level of abrasive size (leading to 13.0463 N cutting force).

References [1] Subramanian, K., and Cook, N.H., 1977, Sensing of drill wear and prediction of drill life, Journal of Manufacturing Science and Engineering, 99(2), pp. 295–301. [2] Krishnaraj, V., Prabukarthi, A., Ramanathan, A., Elanghovan, N., Senthil Kumar, M., Zitoune, R., and Davim, J.P., 2012, Optimization of machining parameters at high speed drilling of carbon fiber reinforced plastic (CFRP) laminates, Composites Part B: Engineering, 43(4), pp. 1791–1799. [3] Gaitonde, V.N., Karnik, S.R., Rubio, J.C., Correia, A.E., Abrao, A.M., and Davim, J.P., 2008, Analysis of parametric influence on delamination in high-speed drilling of carbon fiber reinforced plastic composites, Journal of materials processing technology, 203(1), pp. 431–438. [4] Guu, Y.H., Hocheng, H., Tai, N.H., and Liu, S.Y., 2001, Effect of electrical discharge machining on the characteristics of carbon fiber reinforced carbon composites, Journal of materials science, 36(8), pp. 2037–2043. [5] Chen, W.C., 1997, Some experimental investigations in the drilling of carbon fiber-reinforced plastic (CFRP) composite laminates, International Journal of Machine Tools and Manufacture, 37(8), pp. 1097–1108. [6] Davim, J.P., and Reis, P., 2003, Study of delamination in drilling carbon fiber reinforced plastics (CFRP) using design experiments, Composite Structures, 59(4), pp. 481–487. [7] Arul, S., Vijayaraghavan, L., Malhotra, S.K., and Krishnamurthy, R., 2006, The effect of vibratory drilling on hole quality in polymeric composite, International Journal of Machine Tools and Manufacture, 46(3), pp. 252–259. [8] Krishnamoorthy, A., Rajendra Boopathy, S., Palanikumar, K., and Davim, J.P., 2012, Application of grey fuzzy logic for the optimization of drilling parameters for CFRP composites with multiple performance characteristics, Measurement, 45(5), pp. 1286–1296.

84 Texas Tech University, Fuda Ning, May 2018

[9] Cong, W.L., Pei, Z.J., Feng, Q., Deines, T.W., and Treadwell, C., 2012, Rotary ultrasonic machining of CFRP: a comparison with twist drilling, Journal of Reinforced Plastics and Composites, 31(5), pp. 313–321. [10] Ning, F.D., Cong, W.L., Pei, Z.J., Tang, Y.J., and Treadwell, C., 2014, Rotary ultrasonic machining of CFRP: a comparison with grinding, Ultrasonics, 66, pp. 125–132. [11] Ertunc, H.M., and Oysu, C., 2004, Drill wear monitoring using cutting force signals, Mechatronics, 14(5), pp. 533–548. [12] Mathew, J., Ramakrishnan, N., and Naik, N.K., 1999, Investigations into the effect of geometry of a trepanning tool on thrust and torque during drilling of GFRP composites, Journal of Materials Processing Technology, 91(1), pp. 1–11. [13] Feng, Q., Cong, W.L., Pei, Z.J., and Ren, C.Z., 2012, Rotary ultrasonic machining of carbon fiber-reinforced polymer: feasibility study, Machining Science and Technology, 16(3), pp. 380–398. [14] Cong, W.L., Feng, Q., Pei, Z.J., Deines, T.W., and Treadwell, C., 2011, Dry machining of carbon fiber reinforced plastic composite by rotary ultrasonic machining: effects of machining variables, Proceedings of the ASME 2011 International Manufacturing Science and Engineering Conference, Corvallis, OR, pp. 363–371. [15] Liu, J., Zhang, D., Qin, L., and Yan, L., 2012, Feasibility study of the rotary ultrasonic elliptical machining of carbon fiber reinforced plastics (CFRP), International Journal of Machine Tools and Manufacture, 53(1), pp. 141–150. [16] Cong, W.L., Pei, Z.J., Sun, X., and Zhang, C.L., 2014, Rotary ultrasonic machining of CFRP: a mechanistic predictive model for cutting force, Ultrasonics, 54(2), pp. 663–675. [17] Lazar, M.B., 2012, Cutting force modeling for drilling of fiber reinforced composites, Ph.D. Dissertation, Swiss Federal Institutes of Technology in Lausanne, Switzerland. [18] Cong, W.L., Feng, Q., Pei, Z.J., Deines, T.W., and Treadwell, C., 2012, Rotary ultrasonic machining of carbon fiber reinforced plastic composite: using cutting fluid versus cold air as coolant, Journal of Composite Materials, 46(14), pp. 1745–1753. [19] Li, Z.C., Jiao, Y., Deines, T.W., Pei, Z.J., and Treadwell, C., 2005, Rotary ultrasonic machining of ceramic matrix composites: feasibility study and designed experiments, International Journal of Machine Tools and Manufacture, 45(12), pp. 1402–1411. [20] Cong, W.L., Feng, Q., Pei, Z.J., Deines, T.W., and Treadwell, C., 2012, Edge chipping in rotary ultrasonic machining of silicon, International Journal of Manufacturing Research, 7(3), pp. 311–329. [21] Churi, N.J., Pei, Z.J., Shorter, D.C., and Treadwell, C., 2007, Rotary ultrasonic machining of silicon carbide: designed experiments, International Journal of Manufacturing Technology and Management, 12(1), pp. 284–298.

85 Texas Tech University, Fuda Ning, May 2018

[22] Jiao, Y., Hu, P., Pei, Z.J., and Treadwell, C., 2005, Rotary ultrasonic machining of ceramics: design of experiments, International Journal of Manufacturing Technology and Management, 7(2), pp. 192–206. [23] Cong, W.L., Pei, Z.J., Churi, N.J., and Wang, Q.G., 2009, Rotary ultrasonic machining of stainless steel: design of experiments, Transactions of the North American Manufacturing Research Institution of SME, 37(1), pp. 261–268. [24] Churi, N.J., Pei, Z.J., and Treadwell, C., 2006, Rotary ultrasonic machining of titanium alloy: effects of machining variables, Machining Science and Technology, 10(3), pp. 301–321.

86 Texas Tech University, Fuda Ning, May 2018

CHAPTER V

ULTRASONIC VIBRATION-ASSISTED (UV-A) HOLE-MAKING OF CFRP COMPOSITES: A MECHANISTIC ULTRASONIC VIBRATION AMPLITUDE MODEL

Paper title: A mechanistic ultrasonic vibration amplitude model during rotary ultrasonic machining of CFRP composites

Published in: Ultrasonics (2017), Vol. 76, pp. 44-51.

Authors: Fuda Ning1, Hui Wang1, Weilong Cong1, and P.K.S.C. Fernando2

Authors’ affiliations: 1Department of Industrial, Manufacturing, and Systems Engineering, Texas Tech University, Lubbock, TX, USA 2Department of Industrial and Manufacturing Systems Engineering, Kansas State University, Manhattan, KS, USA

87 Texas Tech University, Fuda Ning, May 2018

Abstract Rotary ultrasonic machining (RUM) has been investigated in machining of brittle, ductile, as well as composite materials. Ultrasonic vibration amplitude, as one of the most important input variables, affects almost all the output variables in RUM. Numerous investigations on measuring ultrasonic vibration amplitude without RUM machining have been reported. In recent years, ultrasonic vibration amplitude measurement with RUM of ductile materials has been investigated. It is found that the ultrasonic vibration amplitude with RUM was different from that without RUM under the same input variables. RUM is primarily used in machining of brittle materials through brittle fracture removal. With this reason, the method for measuring ultrasonic vibration amplitude in RUM of ductile materials is not feasible for measuring that in RUM of brittle materials. However, there are no reported methods for measuring ultrasonic vibration amplitude in RUM of brittle materials. In this study, ultrasonic vibration amplitude in RUM of brittle materials is investigated by establishing a mechanistic amplitude model through cutting force. Pilot experiments are conducted to validate the calculation model. The results show that there are no significant differences between amplitude values calculated by model and those obtained from experimental investigations. The model can provide a relationship between ultrasonic vibration amplitude and input variables, which is a foundation for building models to predict other output variables in RUM.

Keywords: Ultrasonic vibration amplitude; Rotary ultrasonic machining (RUM); Brittle material; Cutting force.

5.1 Introduction Rotary ultrasonic machining (RUM) has been successfully used in drilling brittle materials (such as alumina, zirconia, silicon, silicon carbide, etc.), ductile materials (titanium and stainless steel alloys), and composite materials (ceramic matrix composites and carbon fiber reinforced plastic composites), and has been proved to be an efficient and effective hole making process [1-21]. RUM is a hybrid nontraditional machining process that combines ultrasonic machining and abrasive grinding, as

88 Texas Tech University, Fuda Ning, May 2018 illustrated in Figure 5.1. In RUM, the metal-bonded diamond core drill performs as a cutting tool. The cutting tool rotates and axially feeds toward the workpiece at a constant feedrate (or a constant pressure) under an ultrasonic vibration frequency (typically 20 kHz). Coolant is pumped through the core of the drill, washing away the swarf and remaining the workpiece at a relatively low cutting temperature.

Rotation Coolant flow in

Feeding Ultrasonic vibration

Coolant Coolant flow out flow out

Workpiece Abrasive portion Figure 5.1 Illustration of RUM

Ultrasonic vibration amplitude, one of the most important input variables in RUM, is controlled by ultrasonic power supply. Numerous studies have found that ultrasonic vibration amplitude has direct influences on almost all output variables in RUM [7, 8, 13, 14, 22, 23]. Ultrasonic vibration amplitude measurement would be thus beneficial for exploring the explanations for some experimentally observed phenomena and predicting the results of output responses in RUM. Currently, there are three reported methods for measuring ultrasonic vibration amplitude in RUM: optical vibration sensor method, dial indicator method, and microscope observation method [22, 24]. However, the first two methods can only be used to measure ultrasonic vibration amplitude without RUM machining of materials. It was reported

89 Texas Tech University, Fuda Ning, May 2018 that with the increase of ultrasonic power from 20% to 40%, the ultrasonic vibration amplitude measured by the dial indicator method had similar trends but different values compared with that measured by the microscope observation method in RUM machining of stainless steel [22]. The major limitation of the microscope observation method is that the machining marks, indicating diamond grains trajectories, can only be observed on the RUM-machined surface of ductile materials (The details of the microscope observation method will be reviewed in Section 5.2.). In addition, some methods have been reported for measuring ultrasonic vibration amplitude in other applications [25-34], which can be potentially utilized in RUM without machining of materials. The investigations on all of these reported methods indicate that there is a lack of methods to measure ultrasonic vibration amplitude in RUM machining of brittle materials.

In this chapter, a mechanistic calculation model is developed to investigate ultrasonic vibration amplitude through cutting force in RUM of brittle materials. Carbon fiber reinforced plastic (CFRP) composites have been selected as the workpiece in this investigation based on two major reasons. (1) The material removal mechanism in RUM of CFRP has been identified as brittle fracture, and (2) the mechanistic calculation model can be simplified for RUM of other homogeneous brittle materials [35]. With this model, the relationships between ultrasonic vibration amplitude and different combinations of input variables in RUM of CFRP are established. The ultrasonic vibration amplitude in RUM of brittle materials can be calculated by the model, which will be verified by the microscope observation method using a specially designed aluminum-CFRP stack as the workpiece.

5.2 Microscope observation method The microscope observation method is similar to the quick-stop method [36- 38] for metal-cutting research where the cutting process is “frozen” for observations. The principle of this method is chasing and then measuring the trajectories of diamond grains relative to the machined surfaces. A machined hole and a machined rod are generated by the cutting tool in RUM, as illustrated in Figure 5.2. The diamond grain

90 Texas Tech University, Fuda Ning, May 2018 trajectories remain on the machined surfaces of both hole and rod, because the diamond grains exist on both outer and inner sides of the tool. Hence, the trajectories can be observed from the micrograph taken on the machined rod surfaces. The measurement methodology and amplitude reading mechanism are shown in Figure 5.3.

Machined rod

View surface Machined hole

Feeding Entrance

Exit

Workpiece

Figure 5.2 Illustrations of machined rod and hole

Slope )

A ne

( o li

Zer

m A

20 μm Figure 5.3 Measurement of ultrasonic vibration amplitude on a microscopic picture in RUM of Ti

The microscope observation method to measure the ultrasonic vibration amplitude is only applicable to the ductile materials since the machining marks (trajectories of diamond grains) on the machined surfaces are visible on the ductile materials (including stainless steel, titanium, aluminum), but not visible on the brittle materials as well as some composite materials (alumina ceramics, ceramic matrix

91 Texas Tech University, Fuda Ning, May 2018 composites (CMC), and CFRP), as shown in Figure 5.4. It is still unknown what the ultrasonic vibration amplitude is during RUM of brittle materials. Therefore, a method to measure the ultrasonic vibration amplitude in RUM of brittle materials is needed. Such method will be presented in Section 5.5.

Aluminum Titanium Ceramics CMC

Stainless steel CFRP Figure 5.4 Machined rod surfaces of different materials

5.3 Development of mechanistic amplitude calculation model

5.3.1 Approach to amplitude calculation model development and notations RUM can be treated as a combination of the ultrasonic machining process and abrasive grinding process. The approach of considering ultrasonic machining as the predominant process has been successfully used for RUM model development [5, 35, 39-42]. This approach can also be employed in this amplitude model development.

Figure 5.5 shows the major model development procedures which can be summarized as the following steps.

(1) Measure cutting force using a dynamometer; (2) Establish a relationship between cutting force and abrasive particle indentation depth; (3) Estimate the actual volume of material removed by one abrasive particle indentation in a single ultrasonic vibration cycle;

92 Texas Tech University, Fuda Ning, May 2018

(4) Calculate material removal rate (MRR) by aggregating the effects of all active abrasive particles; (5) Build a relationship between input variables (including feedrate and both inner and outer diameters of the tool) and MRR. (6) Finally, find out the relationship between ultrasonic amplitude and cutting force.

Relationship between A and F Input variables Cutting force (listed in Table 1) (F) Workpiece properties Machining and tool CFRP variables micromechanics process Max impact force Material removal (Fi) / rate (MRR) or Max impact force Mechanical for one abrasive properties of (F1) Number of homogeneous active brittle materials abrasives (n) E, ν, Kc Removed volume Indentation depth for one abrasive (δ) (V1) Fracture volume factor (Kv) Figure 5.5 Amplitude calculation model development procedures

RUM, a complex process, involves many input variables, including workpiece properties, tool variables, and machining variables. These input variables used in model development are listed in Table 5.1.

Table 5.1 Input variables in model development Category Input variable Unit Elastic modulus E MPa Workpiece Poisson’s ratio v properties Fracture toughness Kc MPa mm Outer diameter Do mm Tool Inner diameter Di mm variables Abrasive concentration Ca Abrasive size d mm Amplitude A mm Machining Frequency f Hz variables Feedrate Fr mm/s Tool rotation speed S rpm

93 Texas Tech University, Fuda Ning, May 2018

5.3.2 Major assumptions in model development The major assumptions for the RUM cutting tools are that all the diamond abrasive particles on the end face of a cutting tool are rigid spheres with the same size and embedded depth, all participating in cutting during each ultrasonic cycle. The major assumptions for workpiece material are that workpiece material is ideally brittle and it is removed in a brittle fracture mode. While using CFRP as the workpiece in model development, a heterogeneous material is allowed to be represented as an equivalent homogeneous material through micromechanics process [35, 43, 44]. Other assumptions for the CFRP composites include uniform and continuous fibers, a perfect bonding between matrix and fibers, and the void-free structures. Similar assumptions could be found in other investigations on the force model development for grinding (core drill) of CFRP [45-49]. Additional assumptions and simplifications will be discussed where necessary.

5.3.3 The relationship between cutting force and indentation depth In RUM, the cutting force can be calculated by [5, 35, 40-42]

F  tfFi  ntfF1 (1) where, Δt is the effective contact time when an abrasive particle penetrates into the workpiece, s; f is the ultrasonic vibration frequency, Hz; Fi is the maximum impact force between tool and workpiece, N; F1 is the maximum impact force between a single abrasive particle and workpiece, N; and n is the amount of active abrasive grains on the end face of cutting tool.

The effective contact time can be calculated by

1 π    t    arcsin1  (2) πf 2  A where, δ is the indentation depth of an abrasive grain into the workpiece, mm; and A is the ultrasonic vibration amplitude, mm.

94 Texas Tech University, Fuda Ning, May 2018

The relationship between the maximum impact force and indentation depth could be established as [40, 50]

81/ 2 nEd 1/ 2 3/ 2 F  (3) i 3(1 2 ) where, E is the elastic modulus of the workpiece material, MPa; d is the diameter of abrasive grains, mm; and ν is the Poisson’s ratio of the workpiece material.

After substituting Eq. (2) and Eq. (3) into Eq. (1), the following equation can be obtained

1 1    81/2 nEd 1/2 3/2 F  tfF   arcsin 1 i    2 (4) 2 π  A 3(1 ) Eq. (4) can also be expressed to obtain the amplitude

1  π 1.06πF(1 2 ) (5) A   1 sin  1/ 2 3/ 2   2 nEd  

5.3.4 The relationship between indentation depth and material removal rate In RUM, the material volume removed by one abrasive particle can be calculated by [35]

3/4 1  F  2 1/2  DS π     1  V1  KV   d      arcsin1  (6) 3  KC  60 f 2  A

where, KV is the fracture volume factor that is a proportionality parameter assumed to be constant regardless of input variables; KC is the fracture toughness represented by stress intensity factor, MPa mm ; S is the tool rotation speed, rpm; and

D is the cutting tool diameter, mm; The cutting tool has an outer diameter Do and an inner diameter Di due to the thickness of the core drill. Thus, D can be calculated by D  D D  o i . 2

95 Texas Tech University, Fuda Ning, May 2018

Then, the material removal rate can be theoretically calculated by summating material removal rates of all the abrasive particles on the end face of the cutting tool. The material removal rate can be expressed by

3/ 4 1/ 2 1/ 2 3/ 2 nπS  8 Ed   2 1/ 2 π    MRR  nfV  K (D  D )  d    arcsin 1 1 V o i  2       180  3KC (1 )  2  A (7)

In addition, it can be expressed in terms of the feedrate ( Fr ) and area of the cutting tool end face ( A0 ) based on the definition:

π(D2  D2 ) MRR  F A  o i F (8) r 0 4 r By equating Eq. (7) and Eq. (8), the relationship between A and δ can be obtained

1  π 47F K 3/ 4 (D  D )(1 2 )3/ 4  A   1 sin  r C o i  (9) 2 1/ 2 3/ 2 3/ 4 2 1/ 2   nSKV Ed   d    

5.3.5 Ultrasonic vibration amplitude calculation model By combining Eq. (5) and Eq. (9), the amplitude calculation model can be expressed as

1  3/ 4  π 47F K (D  D )(1 2 )3/ 4  A   1 sin  r C o i   2 1/ 2 3/ 2 3/ 4 2 1/ 2   nSKV Ed   d      1 (10)   π 1.06πF(1 2 )  A   1 sin  1/ 2 3/ 2    2 nEd  

The obtaining process of fracture volume factor ( KV ) for brittle materials was reported by Cong et al. [35] and Liu et al. [5]. Elastic modulus ( E f , Em ), Poisson’s ratio ( v f , vm ), and fracture toughness represented by energy ( Gcf ,Gcm ) of each CFRP component (carbon fiber and epoxy, respectively) are listed in Table 5.2. E, ν, and KC

96 Texas Tech University, Fuda Ning, May 2018 of CFRP can be calculated through micromechanics process [35, 43, 44, 51]. The number of active abrasive grains on the end face of cutting tool can be determined by

2   3 4 Ca [5], where, C is the abrasive concentration and ρ is the n  6.56110  3  A0 a  d   density of abrasive material, g/mm3. In the simultaneous Eq. (10), only indentation depth δ and ultrasonic vibration amplitude A are unknowns. Hence, A can be obtained from this mechanistic amplitude calculation model. δ also can be obtained by substituting A in either Eq. (5) or Eq. (9).

5.4 Experimental set-up and conditions

5.4.1 Workpiece properties The CFRP composite workpiece used in this investigation had a dimension of 200 mm × 150 mm × 16 mm and it was composed of carbon fibers and epoxy resin matrix. The plain woven fabric of carbon fibers had an orientation of 0/90 degrees, where the carbon fiber yarn had a thickness of 0.2 mm and a width of 2.5 mm. The CFRP contained 21 layers of fabric with 2 layers of carbon fiber for each. The workpiece properties are listed in Table 5.2.

In RUM of CFRP, the elastic modulus used in this model is

E E E  f m (11) V f Em VmE f The Poisson’s ratio in transverse direction can be calculated by

 fV f  mVm    E (12) E fV f  EmVm

The fracture toughness Kc can be calculated by

1/ 2 Kc  [2E (Gcf Vf  GcmVm )] (13)

In Eqs. 11–13, V f and Vm are the volume fractions of the fiber and the matrix, respectively.

97 Texas Tech University, Fuda Ning, May 2018

Table 5.2 Properties of CFRP workpiece material Property Unit Value Density of CFRP kg/m3 1550 Hardness (Rockwell) HRB 70–75 Poisson’s Ratio (v12) - 0.34 Poisson’s Ratio (v13) - 0.34 Poisson’s Ratio (v23) - 0.42 Longitudinal Young’s modulus (E1) GPa 136 Transverse Young’s modulus (Et) GPa 10.5 In-plane shear modulus (G12) GPa 3.76 Density of carbon fiber kg/m3 1800 Poisson’s ratio of carbon fiber (vf) - 0.3 Young’s modulus of carbon fiber (Ef) GPa 230 2 Fracture toughness of carbon (Energy/Gcf) J/m 2 Density of epoxy matrix kg/m3 1200 Poisson’s ratio of epoxy matrix (vm) - 0.4 Young’s modulus of epoxy matrix (Em) GPa 4.5 2 Fracture toughness of epoxy matrix (Energy/Gcm) J/m 500

5.4.2 Experimental set-up and cutting force measurement The experiments were carried out on a rotary ultrasonic machine (Series 10, Sonicmill, Albuquerque, NM, USA) that was mainly comprised of three subsystems, including an ultrasonic spindle system, a coolant system, and a data acquisition system, as illustrated in Figure 5.6.

The major components of the ultrasonic spindle system were an ultrasonic spindle integrated with an electric motor, an ultrasonic power supply, a control panel, and a hydraulic feeding device. The power supply output the high-frequency (20 kHz) electrical energy converted from the conventional line electricity. Such high-frequency electrical energy enabled a high-frequency mechanical vibration (namely ultrasonic vibration) with the help of a piezoelectric converter. Ultrasonic vibration was then amplified and transmitted to the cutting tool by an acoustic horn inside the ultrasonic spindle, leading to the cutting tool vibration at the frequency of 20 kHz. The ultrasonic vibration amplitude was adjustable due to the variations of the power supply output level. The motor atop the ultrasonic spindle provided the cutting tool rotations and the motor speed controller on the control panel could be adjusted for different tool rotation

98 Texas Tech University, Fuda Ning, May 2018 speeds. The coolant system consisted of a pressure regulator, flow rate gauges, valves, a pump, and a coolant tank, providing coolant to the spindle and the machining interface.

Ultrasonic spindle 3000 system Feeding device Data acquisition system Electric motor Control panel

Ultrasonic spindle Power supply

Transformer Computer and tool holder Tool Pressure Valve Pressure Flow rate Workpiece gauge regulator gauge A/D Fixture converter Valve Dynamometer Pump

Amplifier Coolant tank Coolant system Figure 5.6 RUM system set-up

The data acquisition system consisted of a dynamometer, a charge amplifier, an A/D converter, and a computer with software. The dynamometer (9272, Kistler Inc., Winterthur, Switzerland) was utilized to measure the axial cutting force Fz. Electrical signals (electric charges) from the dynamometer were amplified by the charge amplifier (5070A, Kistler Inc., Winterthur, Switzerland) and then transformed into digital signals by the A/D converter. A data acquisition card (PC-CARD- DAS16/16, Measurement Computing Corp., Norton, MA, USA) on a computer collected the digital signals with the help of a Dynoware software (2815A, Kistler Inc., Winterthur, Switzerland). Configured with the sensitivity values of the dynamometer, the cutting force values could be obtained. Figure 5.7 shows typical cutting force Fz fluctuating with time in RUM process. There are three major sections in this curve. Section (a) was the period when the cutting tool began to contact with

99 Texas Tech University, Fuda Ning, May 2018 the workpiece and the cutting force kept increasing until the abrasive portion of the tool end was fully involved in the workpiece. Section (b) was the stable cutting zone where all the abrasive grains on the tool participated in the machining process. Once the abrasive portion started leaving the workpiece, the cutting force was gradually decreased to zero, as shown in Section (c). The cutting force used to represent the cutting force F in Eq. (10) was the average value in Section (b).

1590

) 1060

(

z F Fx Fx (N) 350

a b c 0 0 10 20 30 40 50 60 TTimeime ( (S)S)

5.4.3 Cutting tool variables and machining variables The cutting tool, as illustrated in Figure 5.8, was a metal-bonded diamond core drill (N.B.R. Diamond Tool Corp., LaGrangeville, NY, USA). The details of cutting tool used in this study are listed in Table 5.3.

Table 5.3 Cutting tool parameters Parameter Unit Value Outer diameter (Do) mm 9.53 Inner diameter (Di) mm 7.82 Tuning length mm 44.5 Abrasive material Diamond Grit size mesh # 60/80 Abrasive concentration 100 Number of slots 0 Bond B (metal)

100 Texas Tech University, Fuda Ning, May 2018

Figure 5.8 Illustration of cutting tool

Considering the limitations of the experimental set-up (for example, ultrasonic vibration frequency was fixed at 20 kHz on the machine), only tool rotation speed, feedrate, and ultrasonic power were changed in the experiments. The ranges of these machining variables were selected based on the experience from the authors’ preliminary experiments [16, 17] as well as the guidance from the RUM machine manufacturer, Sonic-Mill Inc. The detailed experimental conditions are listed in Table 5.4.

Table 5.4 Experimental conditions Input variable Unit Value Tool rotation speed, S rpm 2000; 3000; 4000; 5000 Feedrate, Fr mm/s 0.2; 0.35; 0.5; 0.65; 0.8 Ultrasonic power % 30; 40; 50; 60

5.5 Pilot experimental verification As the machining trajectories on the RUM machined surfaces of CFRP composite were invisible, a special workpiece by stacking CFRP composite and ductile aluminum (Al) alloy was designed and manufactured for ultrasonic vibration amplitude measurement in RUM of CFRP, as shown in Figure 5.9a. The contacting surfaces of CFRP and Al were ground and polished to ensure them to be perfectly joined by bolts and nuts. During the RUM machining, the cutting tool primarily cut

101 Texas Tech University, Fuda Ning, May 2018 the CFRP composite at the joining interface, and the material removal mainly occurred within the CFRP composite. Only a small amount of Al was machined simultaneously to catch the abrasive trajectories without changing the dominated brittle fracture mode, as shown in Figure 5.9b. After the RUM machining, the CFRP and Al workpiece stack was disassembled. The grain trajectory could be observed from the morphology of the machined Al surface using an optical microscope, as shown in Figure 5.9c. Thus, the ultrasonic vibration amplitude could be experimentally measured.

Experiments were conducted by varying each machining variable and keeping other variables constant. Figure 5.10 show comparisons between amplitudes calculated from the mechanistic calculation model and those measured by the microscope observation measurement method. It can be seen that the trends of ultrasonic vibration amplitude with respect to the machining variables (ultrasonic power, tool rotation speed, and feedrate) obtained by model agreed well with those obtained by the experiments. At the highest level of ultrasonic power, the lowest level of tool rotation speed, or the highest level of feedrate, calculation model caused a slightly larger ultrasonic vibration amplitude than the experimental investigations. This was due to the fact that these levels of the input variables were the extreme machining conditions in RUM process, which resulted in a more fluctuated cutting force generating less accurate cutting force than that as predicted. On the other hand, RUM was a complex machining process that combined both ultrasonic assisted machining and abrasive grinding. In order to simplify the machining conditions, several major assumptions for both cutting tool and workpiece material were involved in the development of the mechanistic model, making the theoretical cutting conditions different with the actual ones. In addition, a small amount of Al materials were removed in the actual experimental conditions to facilitate the amplitude measurement. The existence of such non-CFRP materials would cause the errors between the predicted values and the measured results.

102 Texas Tech University, Fuda Ning, May 2018

a b Al Machined hole Observing area

CFRP

Aluminum alloy CFRP composite

Machined rod

Figure 5.9 Ultrasonic vibration amplitude measurement (a. Illustration of the designed CFRP/Al stack; b. The manufactured stack after RUM process; and c. The morphology of the machined Al surface)

It can be seen from Figure 5.10a that the ultrasonic vibration amplitude increased remarkably with the increase of ultrasonic power for both theoretical and experimental investigations in RUM of CFRP. The ultrasonic vibration amplitude was significantly affected by ultrasonic power. This trend was the same as the reported result in RUM of Ti [22]. Figures 5.10b and 5.10c show that ultrasonic vibration amplitude slightly decreased in RUM of CFRP with the increase of tool rotation speed and the decrease of feedrate for both model and experimental results. These trends were different from those in RUM of Ti. In RUM of Ti, ultrasonic vibration amplitude showed no significant variations with changes in tool rotation speed and feedrate. The major possible reason is that material removal mechanisms in RUM of CFRP (brittle fracture) and in RUM of Ti (ductile removal) are different.

103 Texas Tech University, Fuda Ning, May 2018

) 50

m Model μ 40 Experiments

0 Ultrasonic amplitude ( 30% 40% 50% 60% Ultrasonic power (%) a. Effects of ultrasonic power.

30 30

m) Model Model μ Experiments Experiments 20 20

10 10

0 Ultrasonic amplitude (μm) 0 Ultrasonic amplitude ( 2000 3000 4000 5000 0.2 0.4 0.6 0.8 Tool rotation speed (rpm) Feedrate (mm/s) b. Effects of tool rotation speed. c. Effects of feedrate. Figure 5.10 Comparisons between calculated model results and experimental results of ultrasonic vibration amplitude

5.6 Conclusions This chapter developed a novel modeling method for ultrasonic vibration amplitude measurement through cutting force in RUM of brittle materials. To the best knowledge of the authors, this is the only reported method with the capability of calculating theoretical ultrasonic vibration amplitude from the measured cutting force. In this investigation, ultrasonic vibration amplitudes calculated by the theoretical model had similar trends with those measured by experimental investigation. The ultrasonic vibration amplitude increased with ultrasonic power increasing, tool rotation speed decreasing, and feedrate increasing.

104 Texas Tech University, Fuda Ning, May 2018

Ultrasonic vibration amplitude, one of the most important input variables in RUM, has remarkable influences on cutting force, cutting temperature, tool wear, and edge chipping in RUM of brittle materials. Additionally, the mechanistic calculation model reported in this chapter can provide a relationship between ultrasonic vibration amplitude and different combinations of input variables, which is a foundation for building models to predict other output variables in RUM.

References [1] Li, Z.C., Jiao, Y., Deines, T.W., Pei, Z.J., and Treadwell, C., 2005, Development of an innovative coolant system for rotary ultrasonic machining, International Journal of Manufacturing Technology and Management, 7(2–4), pp. 318–328. [2] Li, Z.C., Pei, Z.J., Zeng, W.M., Kwon, P., and Treadwell. C., 2005, Preliminary experimental study of rotary ultrasonic machining on zirconia toughened alumina, Transactions of NAMRI/SME, 33, pp. 89–96. [3] Li, Z.C., Cai, L.W., Pei, Z.J., and Treadwell, C., 2006, Edge-chipping reduction in rotary ultrasonic machining of ceramics: finite element analysis and experimental verification, International Journal of Machine Tools and Manufacture, 46(12–13), pp. 1469–1477. [4] Li, Z.C., Pei, Z.J., Sisco, T., Micale, A.C., and Treadwell, C., 2007, Experimental study on rotary ultrasonic machining of graphite/epoxy panel, Proceedings of the ASPE 2007 Spring Topical Meeting on Vibration Assisted Machining Technology, pp. 52–57. [5] Liu, D.F., Cong, W.L., Pei, Z.J., and Tang, Y.J., 2012, A cutting force model for rotary ultrasonic machining of brittle materials, International Journal of Machine Tools and Manufacture, 52(1), pp. 77–84. [6] Cong, W.L., Pei, Z.J., Churi, N.J., and Wang, Q.G., 2009, Rotary ultrasonic machining of stainless steel: design of experiments, Transactions of the North American Manufacturing Research Institution of SME, 37, pp. 261–268. [7] Cong, W.L., Pei, Z.J., Van Vleet, E., and Wang, Q.G., 2009, Surface roughness in rotary ultrasonic machining of stainless steels, Proceedings of the IIE Annual Conference and Expo 2009 – Innovations Revealed. [8] Cong, W.L., Pei, Z.J., Deines, T.W., Wang, Q.G., and Treadwell, C., 2010, Rotary ultrasonic machining of stainless steels: empirical study of machining variables, International Journal of Manufacturing Research, 5(3), pp. 370–386. [9] Cong, W.L., Feng, Q., Pei, Z.J., and Treadwell, C., 2010, Comparison of superabrasive tools in rotary ultrasonic machining of stainless steel, Proceedings of the ASME 2010 International Manufacturing Science and Engineering Conference (MSEC), pp. 113–119. [10] Cong, W.L., Feng, Q., Deines, T.W., Pei, Z.J., and Treadwell, C., 2011, Dry machining of carbon fiber reinforced plastic composite by rotary ultrasonic

105 Texas Tech University, Fuda Ning, May 2018

machining: effects of machining variables, Proceedings of the ASME 2011 International Manufacturing Science and Engineering Conference, pp. 363–371. [11] Cong, W.L., Pei, Z.J., Deines, T.W., and Treadwell, C., 2011, Rotary ultrasonic machining of CFRP using cold air as coolant: feasible regions, Journal of Reinforced Plastics and Composites, 30(10), pp. 899–906. [12] Cong, W.L., Pei, Z.J., Feng, Q., Deines, T.W., and Treadwell, C., 2012, Rotary ultrasonic machining of CFRP: a comparison with twist drilling, Journal of Reinforced Plastics and Composite, 31(5), pp. 313–321. [13] Cong, W.L., Pei, Z.J., Feng, Q., Deines, T.W., and Treadwell, C., 2012, Rotary ultrasonic machining of carbon fiber reinforced plastic composites: using cutting fluid versus cold air as coolant, Journal of Composite Materials, 46(14), pp. 1745–1753. [14] Cong, W.L., Pei, Z.J., Feng, Q., Deines, Srivastava, A., Riley, L., and Treadwell, C., 2012, Rotary ultrasonic machining of CFRP composites: a study on power consumption, Ultrasonics, 52(8), pp. 1030–1037. [15] Cong, W.L., Pei, Z.J., Feng, Q., Deines, T.W., and Treadwell, C., 2012, Edge chipping in rotary ultrasonic machining of silicon, International Journal of Manufacturing Research, 7(3), pp. 311–329. [16] Ning, F.D., and Cong, W.L., 2015, Rotary ultrasonic machining of CFRP: design of experiment with a cutting force model, Proceedings of the ASME 2015 International Manufacturing Science and Engineering Conference, pp. V001T02A040– V001T02A048. [17] Ning, F.D., Cong, W.L., Pei, Z.J., and Treadwell, C., 2016, Rotary ultrasonic machining of CFRP: a comparison with grinding, Ultrasonics, 66, pp. 125–132. [18] Churi, N.J., Pei, Z.J., and Treadwell, C., 2006, Rotary ultrasonic machining of titanium alloy: effects of machining variables, Machining Science and Technology 10(3), pp. 301–321. [19] Churi, N.J., Pei, Z.J., and Treadwell, C., 2007, Rotary ultrasonic machining of titanium alloy (Ti-6Al-4V): effects of tool variables, International Journal of Precision Technology, 1(1), pp. 85–96. [20] Churi, N.J., Pei, Z.J., Treadwell, C., and Shorter, D., 2007, Rotary ultrasonic machining of silicon carbide: designed experiments, International Journal of Manufacturing Technology and Management, 12(1–3), pp. 284–298. [21] Feng, Q., Cong, W.L., Pei, Z.J., and Ren, C.Z., (2012) Rotary ultrasonic machining of carbon fiber reinforced polymer: feasibility study, Machining Science and Technology, 16(3), pp. 380–398. [22] Cong, W.L., Pei, Z.J., Mohanty, N., Van Vleet, E., and Treadwell, C., 2011, Ultrasonic vibration amplitude in rotary ultrasonic machining: a novel measurement method and effects of process variables, Journal of Manufacturing Science and Engineering, 133(3), pp. 1–6. [23] Cong, W.L., and Ning, F.D., 2015, Chapter 2 Rotary ultrasonic machining of CFRP composites, Machinability of fibre-reinforced plastics, 1, pp. 31–81.

106 Texas Tech University, Fuda Ning, May 2018

[24] Prabhakar, D., 1992, Machining advanced ceramic materials using rotary ultrasonic machining process, Master Thesis, University of Illinois at Urbana- Champaign [25] Hueners, B.W., 1983, Absolute ultrasonic amplitude measurement, calibration and troubleshooting of a wire bonder using a laser interferometer, International Journal of Hybrid Microelectron, 6(1), pp. 167–170. [26] Bindal, V.N., Jain, S.K., and Kumar, Y.A., 1986, A laser interferometer for vibration amplitude measurement of power ultrasonic sources, Indian Journal of Pure and Applied Physics, 24(12), pp. 584–587. [27] Boucaud, A., Felix, N., Pizarro, L., and Patat, F., 1999, High power low frequency ultrasonic transducer: Vibration amplitude measurements by an optical interferometric method, Proceedings of IEEE Ultrasonics Symposium, pp. 1095–1098. [28] Yarnitsky, Y., and Braun, S., 1967, Vibration-amplitude measurement on ultrasonic drill, Microtecnic, 21(3), pp. 297–298. [29] Yoneda, K., Tawata, M., and Hattori, S., 1979, Measurement of very small vibration amplitude in ultrasonic transducer by means of a laser probe, Proceedings of IEEE Ultrasonics Symposium, pp. 51–55. [30] Yost, W.T., and Cantrell, J.H., 1992, Absolute ultrasonic displacement amplitude measurements with a submersible electrostatic acoustic transducer, Review of Scientific Instruments, 63(9), pp. 4182–4188. [31] Lazara, K., Zayas, J.M., and Zajac, A., 1975, X-Ray Measurement of an ultrasonic wave amplitude in a crystal, Journal of Acoustical Society of America, 58(2), pp. 471–474. [32] Leonov, G.V., Khmelev, V.N., Savin, I.I., and Abramenko, D.S., 2005, Automation of the amplitude measurement process of ultrasonic oscillatory systems irradiating surface, Proceedings of the 6th Annual International Siberian Workshop and Tutorials on Electron Devices and Materials, pp. 64–67. [33] Golyamina, I.P., Polyakov, Z.I., and Khlopotunova, N.A., 1982, Meter for monitoring the vibration amplitude of an ultrasonic tool, Instruments and Experimental Techniques, 25(5), pp. 1304–1308. [34] Khmelev, V.N., Abramenko, D.S., Barsukov, R.V., and Lebedev, A.N., 2008, Usage features of contact and noncontact measuring methods of oscillation amplitude during adjustment process of ultrasonic devices, Proceedings of the 9th International Workshop and Tutorials on Electron Devices and Materials, pp. 223–226. [35] Cong, W.L., Pei, Z.J., Sun, X., and Zhang, C.L., 2014, Rotary ultrasonic machining of CFRP: A mechanistic predictive model for cutting force, Ultrasonics, 54(2), pp. 3663–3675. [36] Philip, P.K., 1971, Study of the performance characteristics of an explosive quick-stop device for freezing cutting action, International Journal of Machine Tool Design and Research, 11(2), pp. 133–144.

107 Texas Tech University, Fuda Ning, May 2018

[37] Griffiths, B.J., 1986, Development of a quick-stop device for use in metal cutting hole manufacturing processes, International Journal of Machine Tool Design and Research , 26(2), pp. 191–203. [38] Vorm, T., 1976, Development of a quick-stop device and an analysis of the frozen-chip technique, International Journal of Machine Tool Design and Research, 16(4), pp. 241–250. [39] Pei, Z.J., and Ferreira, P.M., 1998, Modeling of ductile mode material removal in rotary ultrasonic machining, International Journal of Machine Tools and Manufacture, 38(10–11), pp. 1399–1418. [40] Pei, Z.J., Prabhakar, D., Ferreira, P.M., and Haselkorn, M., 1995, A mechanistic approach to the prediction of material removal rates in rotary ultrasonic machining, Journal of Engineering for Industry, 117(2), pp. 142–151. [41] Qin, N., Pei, Z.J., Treadwell, C., and Guo, D.M., 2009, Physics-based predictive cutting force model in ultrasonic-vibration-assisted grinding for titanium drilling, Journal of Manufacturing Science and Engineering, 131(4), pp. 1–9. [42] Qin, N., Pei, Z.J., Treadwell, C., and Guo, D.M., 2011, Ultrasonic vibration- assisted grinding of brittle materials: a mechanistic model for cutting force, Proceedings of the ASME 2011 International Manufacturing Science and Engineering Conference (MSEC), pp. 127–136. [43] Barbero, E.J., 2010, Chapter 4 Micromechanics, Introduction to composite material design, CRC Press, 2, pp. 91–142. [44] Kaw, A.K., 2006, Chapter 3 Micromechanical analysis of a lamina, Mechanics of composite materials, CRC Press, 2, pp. 203–314. [45] Hocheng, H., and Tsao, C.C., 2005, The path towards delamination-free drilling of composite materials, Journal of Materials Processing Technology, 167(2–3), pp. 251–264. [46] Hocheng, H., and Tsao, C.C., 2006, Effects of special drill bits on drilling- induced delamination of composite materials, International Journal of Machine Tools and Manufacturing, 46(12–13), pp. 1403–1416. [47] Tsao, C.C., and Chiu, Y.C., 2011, Evaluation of drilling parameters on thrust force in drilling carbon fiber reinforced plastic (CFRP) composite laminates using compound core-special drills, International Journal of Machine Tools and Manufacture, 51(9), pp. 740–744. [48] Tsao, C.C., 2006, The effect of pilot hole on delamination when core drill drilling composites materials, International Journal of Machine Tools and Manufacturing, 46(12–13), pp. 1653–1661. [49] Dharan, C.K.N., 1978, Fracture mechanics of composite materials, Journal of Engineering Materials and Technology, 100(233), pp. 233–247. [50] Timoshenko, S.J., and Goodier, N., 1970, Theory of elasticity, McGraw-Hill, 3. [51] Matthews, F.L., Davies, G.A.O., Hitchings, D., and Soutis, C., 2003, Finite element modeling of composite materials and structures, CRC Press.

108 Texas Tech University, Fuda Ning, May 2018

CHAPTER VI

ULTRASONIC VIBRATION-ASSISTED (UV-A) SURFACE GRINDING OF CFRP COMPOSITES: A COMPARISON WITH CONVENTIONAL SURFACE GRINDING

Paper title:

Rotary ultrasonic surface machining of CFRP composites: a comparison with conventional surface grinding

Published in:

Procedia Manufacturing (2017), Vol. 10, pp. 557-567.

Authors:

Fuda Ning1, Hui Wang1, Yingbin Hu1, Weilong Cong1, Meng Zhang2, and Yuzhou Li1,3

Authors’ affiliations:

1Department of Industrial, Manufacturing, and Systems Engineering, Texas Tech University, Lubbock, Texas, 79409, USA

2Department of Industrial and Manufacturing Systems Engineering, Kansas State University, Manhattan, KS 66506, USA

3School of Electromechanical Engineering, Guangdong University of Technology, Guangzhou, Guangdong, 510006, China

109 Texas Tech University, Fuda Ning, May 2018

Abstract Rotary ultrasonic machining (RUM), a hybrid nontraditional process technology combining ultrasonic machining and grinding, has been proven to be a promising method for hole making of CFRP. Due to its advanced capabilities, RUM has been further extendedly applied in surface machining: rotary ultrasonic surface machining (RUSM). Carbon fiber reinforced plastic (CFRP) composites have found extensive applications in areas such as aerospace, automotive, and sports due to their superior material properties. CFRP components are usually near net shaped after molding processes, however, additional surface machining is still required to generate the final dimensions and functional surfaces of the advanced CFRP components especially with three-dimensional features. However, the investigations on RUSM of CFRP are very limited and there are no reported studies on comparisons between RUSM and conventional surface grinding (CSG) of CFRP. In this chapter, for the first time, a comparative study between these two processes of CFRP in the aspects of axial and infeed-directional cutting forces, torque, and surface roughness is conducted. In order to better understand the material removal differences between these two processes, the kinematic motions of the abrasive grains are also analyzed and compared.

Keywords: Rotary ultrasonic surface machining (RUSM); Conventional surface grinding (CSG); Carbon fiber reinforced plastic (CFRP) composites; Cutting force

6.1 Introduction Rotary ultrasonic machining (RUM), a hybrid non-traditional process, has been increasingly investigated owing to its advanced capability for machining difficult-to- cut materials [1]. Specifically, RUM has been shown to be a promising method for hole making with better machining performance compared to twist drilling [2] and core grinding [3]. Due to the superior advantages, RUM was further extended to rotary ultrasonic surface machining (RUSM), as illustrated in Figure 6.1. It can be seen that RUSM combines both ultrasonic machining and surface machining processes. The

110 Texas Tech University, Fuda Ning, May 2018 rotary cutting tool with diamond abrasives vibrates in the axial direction at a typically ultrasonic frequency of 20 kHz. Meanwhile, the workpiece horizontally feeds towards the tool at a constant feedrate for the material removal. The coolant is pumped through the core of the cutting tool to wash out the chips and prevent the grinding zone from overheating. A sufficient amount of cooling lubricant is thus guaranteed even at a large depth of cut in RUSM.

Coolant Rotation flow in Z

Y X

Ultrasonic vibration Abrasive portion Coolant flow out

F eed

Workpiece

Figure 6.1 Illustration of rotary ultrasonic surface machining (RUSM) process

Carbon fiber reinforced plastic (CFRP) composites have found remarkably increasing applications in the aerospace and automotive industries due to the superior properties [4-6]. The CFRP parts are usually fabricated using molding processes for near-to-net shapes and additional machining processes are further utilized to achieve the dimensional accuracy of the final parts [7-9]. However, CFRP composites exhibit a poor machinability due to the anisotropy and heterogeneity of the materials, considerably reducing the load-bearing capability and service life of CFRP components [8]. Among the machining processes, surface grinding is required to generate the final dimensions of the functional surfaces of the advanced CFRP components especially with three-dimensional features [10]. However, conventional surface grinding (CSG) reaches certain limits in further improving the surface quality

111 Texas Tech University, Fuda Ning, May 2018 and machining efficiency. Therefore, surface grinding of CFRP in a high-quality and high-efficiency approach turns out to be a crucial work.

RUSM has been experimentally and theoretically explored for surface grinding of CFRP [11,12] and other brittle materials, such as glass [13,14], ceramics [15], and SiC reinforced Si composites [10]. In RUSM of CFRP, effects of tool variables (including abrasive size, abrasive concentration, number of slots, and tool end geometry) and machining variables (including tool rotation speed, feedrate, depth of cut, and ultrasonic amplitude) on the process performance were studied. In addition, axial cutting force model in RUSM of CFRP was developed to predict the cutting force value [11]. All these investigations are focused on the understanding of relationships between input variables and output variables in RUSM. However, it is still unknown about the improvements of process performance by applying RUSM compared with using CSG.

In this chapter, for the first time, a comparison between RUSM and CSG of CFRP on axial and infeed-directional cutting forces, torque, and surface roughness will be conducted. In order to better understand the material removal differences between these two processes, the kinematic motions of the abrasive grains are also analyzed and compared.

6.2 Kinematic motion analysis In both RUSM and CSG, the abrasive grains located on the peripheral surface of the tool are primarily responsible for the material removal. Meanwhile, the abrasive grains on the end face of the tool are mainly responsible for the surface formation on the machined tracks [10]. It is assumed that (1) the abrasive grains are rigid spheres with the same diameter and their shapes do not change during the machining process; (2) all abrasive grains on the periphery of the cutting tool have the same height; and (3) the abrasive grains conform to the stochastic distribution [16].

The motion of abrasive grains in RUSM is a combination of spindle rotation, workpiece feed motion, and ultrasonic vibration in the direction of tool axis. Thus, the

112 Texas Tech University, Fuda Ning, May 2018 motion trajectory of a single abrasive grain in RUSM can be described by

 S x  R cos( t)  f t  30 r  S y  R sin( t) (1)  30 z  Asin(2ft)  where, R is the radius of the cutting tool, mm; S is the tool rotation speed, rpm; fr is the workpiece feedrate, mm/s; A is the ultrasonic vibration amplitude, mm; f is the ultrasonic vibration frequency, Hz; t is the processing time, s.

The velocity of the single abrasive grain can be expressed by

 S S v   Rsin( t)  f  x 30 30 r  S S  (2) v y  Rcos( t)  30 30  vz  2fA cos(2ft) 

The effective trajectory length LRUSM of the single abrasive grain during one ultrasonic vibration cycle can be calculated by

1 1 f 2 2 2 2 LRUSM  vx  vy  vz  dt (3) 0

Since the feedrate of workpiece is much smaller than the cutting velocity caused by the spindle rotation, the Eq. (3) can be simplified as

1 2 2 1    S  R  f   2  LRUSM     2  f  Acos2ft  dt (4) 0    30  

Similarly, the velocity and effective trajectory length LCSG in CSG can be expressed by

113 Texas Tech University, Fuda Ning, May 2018

 S S v   Rsin( t)  f  x 30 30 r  S S (5) v  Rcos( t)  y 30 30

1 1 f 2 2 2   S  R LCSG  vx  vy  dt  (6) 0 30 f

Based on the expressions above, the motion trajectories of a single abrasive grain in RUSM and CSG were illustrated in Figure 6.2(a) and Figure 6.2(b), respectively.

(a) (b) Workpiece Z Workpiece

X-Y plane

Motion trajectory of one grain

Figure 6.2 The trajectory of a single abrasive grain in (a) RUSM and (b) conventional

surface grinding (CSG) in the certain period

Thus, the material removal volume of the single abrasive grain in the certain period (1/f) can be calculated by [17]

3 L  L in RUSM 4  RUSM 1  F  1   1  2 2 V1  K V  d    L  or (7) 2  KC    L  LCSG in CSG

where, KV is a constant parameter; F1 is the maximum force for one abrasive grain, N; KC is the fracture toughness expressed by stress intensity factor, MPa mm ; δ is the indentation depth of one abrasive grain penetrating into the workpiece, mm; d

114 Texas Tech University, Fuda Ning, May 2018

is the average diameter of abrasive grains, mm; L is LRUSM in RUSM process and is

LCSG in CSG process.

Material removal rate (MRR) can be theoretically calculated by summing V1 of all the involved abrasive particles in the unit time. In addition, according to the definition of MRR in both RUSM and CSG, MRR also can be expressed in terms of the feedrate and effective cross section area of the cutting tool. Thus, the following equation can be established

MRR  n f V1  2 fr a p  R (8)

where, n is the number of active abrasive grains on the periphery of the cutting tool and ap is the depth of cut, mm.

6.3 Experimental set-up and conditions

6.3.1 Workpiece material properties The CFRP composite workpiece used in this investigation consisted of carbon fibers and epoxy resin matrix with the volume fractions of 67% and 33%, respectively. The workpiece was multi-directional CFRP composite laminated by 24 plies of epoxy resin and 23 layers of carbon fabric. The CFRP workpiece used in this experiment had the size of 150 mm × 18 mm × 18 mm. Properties of the CFRP workpiece material are listed in Table 6.1. Some of them were calculated by the micromechanical analysis.

6.3.2 Experimental set-up and conditions The experiments of this investigation were performed on a SONICMILL Series 10 rotary ultrasonic machine. The RUSM system set-up included an ultrasonic spindle system, a data acquisition system, a coolant system, and a horizontal feeding system, as shown in Figure 6.3. In the ultrasonic spindle system, the ultrasonic power supply was used to convert low-frequency line electricity to high-frequency electrical energy. Such energy was supplied to the piezoelectric converter located inside the ultrasonic spindle to generate the ultrasonic vibration at a high frequency (20 kHz). The ultrasonic vibration was amplified and then transmitted to the metal-bonded

115 Texas Tech University, Fuda Ning, May 2018 diamond tool, causing the tool to vibrate vertically. The motor assembled atop the ultrasonic spindle provided a rotation motion of the tool. Meanwhile, a NEWMARK linear stage with a maximum travel of 400 mm in the horizontal feeding system enabled an infeed motion of the workpiece to realize the RUSM process. The feedrate could be controlled by a motor controller and a software.

Table 6.1 Properties of CFRP workpiece material Property Unit Value Density of CFRP kg/m3 1600 Hardness (Rockwell) HRB 68–72

Poisson’s Ratio (v12) - 0.33

Poisson’s Ratio (v13) - 0.33

Poisson’s Ratio (v23) - 0.42

Longitudinal Young’s modulus (E1) GPa 155.6

Transverse Young’s modulus (E2) GPa 13.1

In-plane shear modulus (G12) GPa 4.7 Density of epoxy matrix kg/m3 1200 Poisson’s ratio of epoxy matrix - 0.4 Young’s modulus of epoxy matrix GPa 4.5 Fracture toughness of epoxy matrix J/m2 500 (Energy/Gc) Density of carbon fiber kg/m3 1800 Poisson’s ratio of carbon fiber - 0.3 Young’s modulus of carbon fiber GPa 230 Fracture toughness of carbon fiber J/m2 2 (Energy/Gc)

The cutting tool used in both RUSM and CSG was a metal-bonded diamond grinding tool provided by NBR Diamond Tool Corp. The detailed information of the cutting tool was listed in Table 6.2. In addition, effects of tool rotation speed and feedrate on the four output variables were studied in both RUSM and CSG. The ranges of tool rotation speed and feedrate were set based on the preliminary experimental results, as shown in Table 6.3. Other RUSM machining variables including depth of cut, ultrasonic frequency, and ultrasonic amplitude were fixed at 2 mm, 20 kHz, and 6 µm, respectively. The ultrasonic vibration was not provided in CSG. Three slots were machined under each combination of input variables.

116 Texas Tech University, Fuda Ning, May 2018

Data acquisition Ultrasonic system 3000 spindle system Feeding device Motor Electric controller motor Control panel Ultrasonic Computer spindle

Power supply A/D converter Pressure gauge Transformer and tool holder Pressure Flow rate Valve Channels Valve 1~4 regulator gauge Workpiece Pump Diamond tool Amplifier Fixture Dynamometer Motor Coolant Platform Abrasive tank portion Linear stage Fixture Feed direction Machine Horizontal feeding table Coolant system system Figure 6.3 RUSM system set-up

Table 6.2 Cutting tool variables Tool variables Value Outer diameter (mm) 9.6 Inner diameter (mm) 7.8 Tuning length (mm) 44.5 Mesh size 80/100 Abrasive grain size (mm) 0.165 Abrasive concentration 100

Table 6.3 Experimental conditions Group Tool rotation speed S (rpm) Feedrate Fr (mm/s) 1 3000; 4000; 5000; 6000 0.8 2 3000 0.4; 0.6; 0.8; 1

6.3.3 Measurement procedures for output variables The KISTLER 9272 dynamometer in the data acquisition system was used to measure infeed cutting force Fx, axial cutting force Fz, and torque M together with a KISTLER 5070 charge amplifier and a KISTLER 5697A A/D converter. The digital signals were collected by a KISTLER DynoWare software on a computer. The sampling rate of dynamometer system was set to 4000 Hz. Since the cutting force and torque changed with time and fluctuated in a certain range, the average value of the

117 Texas Tech University, Fuda Ning, May 2018 stable machining phase on the curve was selected to represent the cutting force and torque value.

A MITUTOYO surface profilometer was used to measure the surface roughness Ra. Test range and cut-off length were set at 4 mm and 0.8 mm, respectively. The average Ra value was used in this investigation. As shown in Figure 6.4, four positions were selected for the measurement of surface roughness on each slot. Two of them located in the cutting entrance area, and the other two located in the exit area.

Four measurements

Figure 6.4 Measurements of surface roughness

6.4 Results and discussion All the error bars in Figure 6.5 to Figure 6.8 represented the maximum, mean, and minimum values.

6.4.1 Effects on infeed cutting force

A comparison of infeed cutting force Fx between RUSM and CSG with the increase of tool rotation speed S or feedrate fr is illustrated in Figure 6.5. Under all different levels of S or fr, RUSM always generated lower Fx than CSG. By comparing Eq. (4) with Eq. (6), it can be found that effective trajectory length in RUSM was larger than that in CSG. As shown in Eq. (8), MRR was kept unchanged under any combination of machining variables, indicating the same material removal volume of the single abrasive grain V1 in both processes. According to Eq. (7), indentation depth

δ in RUSM should be smaller than that in CSG. In both RUSM and CSG processes, Fx

118 Texas Tech University, Fuda Ning, May 2018

is a function of one abrasive grain force F1 that can be calculated by [18]

81/ 2 Ed 1/ 2 3/ 2 F  (9) 1 3(1 2 )

where, E is the elastic modulus of CFRP and  is the Poisson’s ratio of CFRP.

Thus, it can be seen from Eq. (9) that at any level of S or fr, Fx in RUSM would be smaller than that in CSG due to the smaller δ.

In both processes, Fx decreased as S increased or fr decreased. It is notable in

Eq. (8) that MRR was not affected by the increase of S, resulting the fixed V1 in both processes. The increase of S led to an enhancement of effective trajectory length L in both processes, as shown in Eq. (4) with Eq. (6). Thus, a decreased δ was generated to maintain the same V1, causing the decrease of Fx in both processes. On the other hand, the decrease of fr reduced MRR and thus a decreased V1 would be caused based on Eq.

(8). Since the increase of fr did not change L in both processes, the decrease of δ was the reason that resulted in the decrease of V1, resulting in the decrease of Fx in both RUSM and CSG processes.

6.4.2 Effects on axial cutting force

Figure 6.6 shows the comparison of axial cutting force Fz between RUSM and

CSG with the increase of tool rotation speed S or feedrate fr. Similar to Fx, Fz in RUSM was always lower than that in CSG under any combination of machining variables. In RUSM, Fz can be expressed by [17,19]

1 1    81/ 2 nEd 1/ 2 3/ 2

Fz  n f t  F1    arcsin1  2 (10) 2 π  A 3(1 )

where, t is the time during which an abrasive particle is penetrating into the workpiece (effective contact time), s.

Since there was no ultrasonic vibration in CSG, the axial cutting force in CSG can be calculated by

119 Texas Tech University, Fuda Ning, May 2018

75 80 RUSM RUSM CSG CSG 60 65

45 (N) 50

Fx Fx Fx (N) 30 35

15 20 3000 4000 5000 6000 0.4 0.6 0.8 1 Tool rotation speed (rpm) Feedrate (mm/s) (a) Effects of tool rotation speed (b) Effects of feedrate Figure 6.5 Comparisons of infeed cutting force between RUSM and CSG

11 12 RUSM RUSM CSG CSG 9 10

7 8

Fz (N) Fz (N) 5 6

3 4 3000 4000 5000 6000 0.4 0.6 0.8 1 Tool rotation speed (rpm) Feedrate (mm/s) (a) Effects of tool rotation speed (b) Effects of feedrate Figure 6.6 Comparisons of axial cutting force between RUSM and CSG

81/ 2 nEd 1/ 2 3/ 2 F  n F  (11) z 1 3(1 2 )

The reason for the lower axial cutting force generation in RUSM was the same with that for infeed cutting force. In RUSM, the indentation depth δ was smaller than that in CSG, and Δt was smaller than 1/f due to the existence of ultrasonic vibration.

Thus, a lower Fz was achieved in RUSM than that in CSG at any combination of S and fr according to the Eqs. (10) and (11), respectively.

It can be seen that Fz showed the same decreasing trends with Fx when S increased or fr decreased. This was attributed to the relationships that Fz decreased as

120 Texas Tech University, Fuda Ning, May 2018

the results of a decreased δ generated by the increase of S or the decrease of fr in both

RUSM and CSG. In addition, at each combination of S and fr, Fz in RUSM or CSG was always smaller than Fx in the corresponding process. Such phenomena could be explained by the kinematic motion of the cutting tool in both processes. The volume of materials removed by the abrasives on the end face of tool was rather limited to only form the surface structure on the slot bottom. Whereas, the abrasives on the periphery face of tool removed the primary material volume, resulting in the larger Fx than Fz.

6.4.3 Effects on torque The torque could be calculated by multiplying cutting forces with their corresponding moment arms of forces. In this work, the total torque was equivalent to the sum of the torques generated from Fx and Fz by their coefficients kx and kz, respectively, as shown in Eq. (12).

M  kx Fxlx  kz Fzlz (12)

where, M is torque, N·m; lz and lx are the moment arms of related cutting force, m.

The comparison of torque between RUSM and CSG at each combination of tool rotation speed S and feedrate fr is illustrated in Figure 6.7. The results show that the torque in RUSM was always smaller than that in CSG at different levels of S or fr. Such phenomena were caused due to the positive relationships between torque and cutting forces in Eq. (12). RUSM expressed the lower values of both Fx and Fz than CSG, consequently causing the lower values of torque as well. In addition, the trends of torque were decreasing with the increase of S or with the decrease of fr in both processes. This was also attributed to the aforementioned relationships in Eq. (12).

121 Texas Tech University, Fuda Ning, May 2018

4 6.0 RUSM RUSM CSG CSG 3 4.5

2 3.0 Torque (N·m) Torque 1 (N·m) Torque 1.5

0 0.0 3000 4000 5000 6000 0.4 0.6 0.8 1 Tool rotation speed (rpm) Feedrate (mm/s) (a) Effects of tool rotation speed (b) Effects of feedrate Figure 6.7 Comparisons of torque between RUSM and CSG

6.4.4 Effects on surface roughness Figure 6.8 illustrates the comparison of surface roughness between RUSM and

CSG with the increase of tool rotation speed S or feedrate fr. Different with cutting forces and torque, surface roughness turned out to be larger in RUSM compared with that in CSG at each combination of S and fr. This phenomenon was ascribed to the different single abrasive grain trajectories on the tool end face in the two different processes. In RUSM, the vertical ultrasonic vibration induced the abrasive grains on the tool end face to produce more impacts on the machined surface. Whereas, only cutting and plowing effects of the abrasive grains occurred in CSG. Thus, more surface damages, such as micro-cracks, delamination, etc., were generated causing the larger surface roughness in RUSM [20]. During RUSM, using infeed-directional rather than vertical ultrasonic vibration during horizontal feeding process might be a potential solution to generate a smaller surface roughness than that in CSG.

Effects of S and fr on the surface roughness in both processes were the same with those on the cutting forces and torque. The increase of S or decrease of fr decreased the values of surface roughness in RUSM and CSG. With the increase of S or decrease of fr, the amount of active abrasive grains on the tool end surface participating in the machining increased in the unit time. As a result of it, the residual area on the machine surface was reduced, leading to the decrease of surface roughness

122 Texas Tech University, Fuda Ning, May 2018 in both processes.

1.6 1.6 RUSM RUSM CSG CSG 1.4 1.4

1.2 1.2

1.0 1.0 Surface roughness(µm) 0.8 Surface roughness(µm) 0.8 3000 4000 5000 6000 0.4 0.6 0.8 1 Tool rotation speed (rpm) Feedrate (mm/s) (a) Effects of tool rotation speed (b) Effects of feedrate Figure 6.8 Comparisons of surface roughness between RUSM and CSG

6.5 Conclusions A comparative study between RUSM and CSG of CFRP on axial and infeed- directional cutting forces, torque, and surface roughness was conducted in this chapter. The kinematic motions of the abrasive grains in these two processes were analyzed and compared to better understand the material removal differences. Development of axial and infeed-directional cutting force modeling was presented to explain the effects on the cutting forces in both RUSM and CSG. The results showed that RUSM always generated lower values of infeed cutting force, axial cutting force, and torque than CSG at each combination of tool rotation speed and feedrate. The main reason was due to the larger effective trajectory length in RUSM with a smaller indentation depth, leading to the smaller cutting forces and torque. However, RUSM exhibited the larger surface roughness compared with CSG resulting from the occurrence of surface damages induced by the vertical ultrasonic vibration. In the future investigations, horizontal ultrasonic vibration will be applied in RUSM process to eliminate the vertical impact of abrasives on the machined surface.

References [1] Zhang, J.H., Zhao, Y., Tian, F.Q., Zhang, S., and Guo, L.S., 2015, Kinematics and experimental study on ultrasonic vibration-assisted micro end grinding of

123 Texas Tech University, Fuda Ning, May 2018

silica glass, International Journal of Advanced Manufacturing Technology, 78(9–12), pp. 1893–1904. [2] Cong, W.L., Pei, Z.J., Feng, Q., Deines, T.W., and Treadwell, C., 2012, Rotary ultrasonic machining of CFRP: a comparison with twist drilling, Journal of Reinforced Plastics and Composites, 31(5), pp. 313–321. [3] Ning, F.D., Cong, W.L., Pei, Z.J., and Treadwell, C., 2016, Rotary ultrasonic machining of CFRP: a comparison with grinding, Ultrasonics, 66, pp. 125–132. [4] Davim, J.P., and Reis, P., 2003, Drilling carbon fiber reinforced plastics manufactured by autoclave-experimental and statistical study, Materials & design, 24(5), pp. 315–324. [5] Cong, W.L., and Ning, F.D., in: J.P. Davim (Ed.), Machinability of Fibre- Reinforced Plastics, Walter de Gruyter GmbH & Co KG, Berlin, 2015, pp. 31– 81. [6] Ning, F.D., Cong, W.L., Hu, Y.B., and Wang, H., 2017, Additive manufacturing of carbon fiber-reinforced plastic composites using fused deposition modeling: Effects of process parameters on tensile properties, Journal of Composite Materials, 51(4), pp. 451–462. [7] Jia, Z., Su, Y., Niu, B., Zhang, B., and Wang, F., 2016, The interaction between the cutting force and induced sub-surface damage in machining of carbon fiber- reinforced plastics, Journal of Reinforced Plastics and Composites, 35(9), pp. 712–726. [8] Jia, Z., Fu, R., Niu, B., Qian, B., Bai, Y., and Wang, F., 2016, Novel drill structure for damage reduction in drilling CFRP composites, International Journal of Machine Tools and Manufacture, 110, pp. 55–65. [9] Ning, F.D., and Cong W.L., 2015, Rotary ultrasonic machining of CFRP: design of experiment with a cutting force model, In: Proceedings of the 2015 International Manufacturing Science and Engineering Conference, Charlotte, North Carolina, USA, pp. V001T02A040–V001T02A048. [10] Bertsche, E., Ehmann, K., and Malukhin, K., 2013, An analytical model of rotary ultrasonic milling, International Journal of Advanced Manufacturing Technology, 65(9–12), pp. 1705–1720. [11] Liu, S., Chen, T., and Wu, C., 2017, Rotary ultrasonic face grinding of carbon fiber reinforced plastic (CFRP): a study on cutting force model, International Journal of Advanced Manufacturing Technology, 89(1–4), pp. 847–856. [12] Wang, H., Ning, F., Hu, Y., Fernando, P.K.S.C., Pei, Z.J., and Cong, W., 2016, Surface grinding of carbon fiber–reinforced plastic composites using rotary ultrasonic machining: Effects of tool variables, Advances in Mechanical Engineering, 8(9), pp. 1–14. [13] Zhang, C.L., Feng, P.F., Zhang, J.F., Wu, Z.J., and Yu, D.W., 2012, Theoretical and experimental research on the features of cutting force in rotary ultrasonic face milling of K9 glass, Applied Mechanics and Materials, 157, pp. 1674–1679. [14] Zhang, C., Zhang, J., and Feng, P., 2013, Mathematical model for cutting force in rotary ultrasonic face milling of brittle materials, International Journal of Advanced Manufacturing Technology, 69(1–4), pp. 161–170.

124 Texas Tech University, Fuda Ning, May 2018

[15] Pei, Z.J., and Ferreira, P.M., 1999, An experimental investigation of rotary ultrasonic face milling, International Journal of Machine Tools and Manufacture, 39(8), pp. 1327–1344. [16] Ning, F.D., Cong, W.L., Wang, H., Hu, Y.B., Hu, Z.L., and Pei, Z.J., 2017, Surface grinding of CFRP composites with rotary ultrasonic machining: a mechanistic model on cutting force in the feed direction, International Journal of Advanced Manufacturing Technology, 92(1–4), pp.1217–1229. [17] Cong, W.L., Pei, Z.J., Sun, X., and Zhang, C.L., 2014, Rotary ultrasonic machining of CFRP: a mechanistic predictive model for cutting force, Ultrasonics, 54(2), pp. 663–675. [18] Pei, Z.J., Prabhakar, D., Ferreira, P.M., and Haselkorn, M., 1995, A mechanistic approach to the prediction of material removal rates in rotary ultrasonic machining, Journal of Engineering for Industry, 117(2), pp. 142–151. [19] Ning, F., Wang, H., Cong, W., and Fernando, P.K.S.C., 2017, A mechanistic ultrasonic vibration amplitude model during rotary ultrasonic machining of CFRP composites, Ultrasonics, 76, pp. 44–51. [20] Sasahara, H., Kikuma, T., Koyasu, R., and Yao, Y., 2014, Surface grinding of carbon fiber reinforced plastic (CFRP) with an internal coolant supplied through grinding wheel, Precision Engineering, 38(4), pp. 775–782.

125 Texas Tech University, Fuda Ning, May 2018

CHAPTER VII

ULTRASONIC VIBRATION-ASSISTED (UV-A) SURFACE GRINDING OF CFRP COMPOSITES: A MECHANISTIC MODEL ON CUTTING FORCE IN THE FEED DIRECTION

Paper title:

Surface grinding of CFRP composites with rotary ultrasonic machining: a mechanistic model on cutting force in the feed direction

Published in:

International Journal of Advanced Manufacturing Technology (2017), Vol. 92, No. 1, pp. 1217-1229.

Authors:

Fuda Ning1, Weilong Cong1, Hui Wang1, Yingbin Hu1, Zhonglue Hu2, and Zhijian Pei3

Authors’ affiliations:

1Department of Industrial, Manufacturing, and Systems Engineering, Texas Tech University, Lubbock, Texas, 79409, USA

2Department of Mechanical Engineering, Texas Tech University, Lubbock, Texas, 79409, USA

3Department of Industrial and Systems Engineering, Texas A&M University, College Station, Texas 77843, USA

126 Texas Tech University, Fuda Ning, May 2018

Abstract For carbon fiber reinforced plastic (CFRP) composite components, especially advanced CFRP components with complex three-dimensional features, surface grinding is often needed to generate final dimensions and functional surfaces. Surface damages are frequently induced during surface grinding, reducing the load-bearing capability and service life of the components. Therefore, it is desirable to perform surface grinding of CFRP in a high-quality and high-efficiency way. Rotary ultrasonic machining (RUM) surface grinding has been investigated to machine CFRP for improved surface quality. Cutting force is one of the most important output variables for evaluating RUM surface grinding. The modeling of cutting force is essential to effectively control the occurrence of surface damages during RUM surface grinding of CFRP. In the RUM surface grinding process, the workpiece material is primarily removed by abrasives on the tool peripheral surface, thus it is essential to investigate the feed-direction cutting force model. However, such models are not available in the literature. In this study, for the first time, a mechanistic feed-direction cutting force model in RUM surface grinding of CFRP is established based on the assumption that the material is removed by brittle fracture. The mechanistic model has one parameter, fracture volume factor of the workpiece material, which needs to be determined by an experiment. There is a good consistency between theoretically predicted trends and experimentally observed results on the relationships between feed-direction cutting force and input variables.

Keywords: Rotary ultrasonic machining (RUM); Surface grinding; Carbon fiber reinforced plastic (CFRP) composites; Feed-direction cutting force; Mechanistic predictive model

7.1 Introduction Carbon fiber reinforced plastic (CFRP) composites have a variety of attractive properties, such as low density, high stiffness-to-weight ratio, excellent fatigue and wear resistance, high dimensional stability, and low thermal expansion coefficient [1- 4]. Due to these superior properties, the usage of CFRP composites has been

127 Texas Tech University, Fuda Ning, May 2018 remarkably increasing in the aerospace industry to produce lighter and more durable aircraft with higher fuel efficiency or larger payload [5, 6]. CFRP components often require additional machining processes (such as hole making, edge and surface machining, etc.) to improve the component integrity and dimensional accuracy for future assembly and other functional purposes in aerospace and automotive industries [5, 7-9]. During these machining processes, CFRP composites exhibit a poor machinability with considerable problems due to the anisotropic and heterogeneous material properties. For example, rapid tool wear and high cutting force can be primarily caused as a result of the highly abrasive nature of the carbon fibers and the low thermal conductivity of the resin matrix [5]. In this case, induced surface damages can lead to bad surface quality and low machining accuracy, reducing the load-bearing capability and service life of CFRP components [8].

Many studies have been conducted to improve the quality and efficiency in hole making of CFRP [8-13]. Apart from hole making, surface machining is also utilized to generate the final dimensions and functional surfaces of the advanced CFRP components especially with complex three-dimensional features [14]. There are two main conventional surface machining processes, surface milling and surface grinding. Compared with surface milling [15, 16], surface grinding using either cylindrical grinding tools [17] or flat grinding wheels [18] produces better surface quality of CFRP workpiece. However, short tool life and low machining efficiency still exist in the conventional surface grinding processes for CFRP. Therefore, it is desirable to develop new surface grinding processes for CFRP in a high-quality and high-efficiency way.

Rotary ultrasonic machining (RUM), a hybrid nontraditional process, has been increasingly investigated due to its relatively low cost and advanced capability for machining difficult-to-cut materials [19]. Specifically, RUM has been shown to be a promising method for hole making of CFRP with less tool wear, lower cutting force, and better surface quality compared to twist drilling [20] and core grinding [21]. A large number of studies on RUM of CFRP have been further conducted both

128 Texas Tech University, Fuda Ning, May 2018 theoretically and experimentally, such as modeling of cutting force and ultrasonic amplitude [22-24], enhancement of chip removal ability [25], utilization of cold air coolant [26], the study of power consumption [27], etc. RUM was further extended from hole making to surface grinding by Pei et al. for improving the machining performance of conventional surface grinding [28]. The RUM surface grinding process, as illustrated in Figure 7.1, combines both ultrasonic machining and surface grinding processes. It can be seen that the rotary cutting tool with diamond abrasives vibrates at an ultrasonic frequency (typically 20 kHz) in the tool axis direction. Meanwhile, the workpiece horizontally feeds towards the tool at a constant feedrate for the material removal. The coolant is pumped through the core of the cutting tool to wash out the chips and prevent the grinding zone from overheating.

Coolant Rotation flow in Z

Y X

Ultrasonic vibration Abrasive portion Coolant flow out

F eed

Workpiece Figure 7.1 Illustration of rotary ultrasonic machining (RUM) surface grinding process

Cutting force, one of the most important output variables for evaluating the RUM surface grinding process, is directly related to the tool wear and surface quality. Conducting the investigations on cutting force and its modeling are essential to better understand the mechanism of RUM surface grinding of CFRP for effectively controlling the occurrence of surface damages [5]. In RUM surface grinding, the abrasive grains located on the peripheral surface of the tool are primarily responsible for the material removal. Meanwhile, the abrasive grains on the end face of the tool are mainly responsible for the surface formation on the machined tracks [29]. In

129 Texas Tech University, Fuda Ning, May 2018 addition, unlike machining of metals, RUM surface grinding of CFRP should be conducted under a larger depth of cut [30], as the smaller depth of cut could easily lead to the surface damages due to ploughing and squeezing effects induced by the tool. With these reasons, compared with the axial cutting force, the feed-direction cutting force is the dominated component of the cutting force.

Cutting force features in RUM surface grinding of CFRP [31, 32] and other brittle materials, such as glass [33, 34], ceramics [35, 36], ceramic matrix composites [37, 38], and silicon matrix composites [14], have been experimentally and theoretically explored. In the experimental investigations, comparisons on axial cutting forces between RUM surface grinding and conventional surface grinding of CFRP [32] and glass [33] were conducted. In addition, effects of tool variables on the axial cutting forces in RUM surface grinding of CFRP [31] and ceramic [35] were studied. On the other hand, theoretical cutting force modeling was established based on the brittle fracture mode during RUM surface grinding of glass [34] and silicon matrix composites [14] for the axial cutting forces and feed-direction cutting forces, respectively. After that, theoretical models for predicting both axial cutting forces and feed-direction cutting forces during RUM surface grinding of dental ceramics [36] and ceramic matrix composites [37, 38] were developed. Recently, only Liu et al. presented a mathematic model to predict the axial cutting force in RUM surface grinding of CFRP [32], which was conducted on the basis of the modeling methodology for RUM hole making of CFRP [22]. However, the work of [32] considered the abrasive grains on the tool end face to be mainly responsible for the material removal in RUM surface grinding. The direct utilization of the RUM hole making methodology into the surface grinding process is not suitable due to the major kinematic differences between these two machining processes. To the best knowledge of the authors, there are no reported investigations on modeling of the feed-direction cutting force in RUM surface grinding of CFRP.

This chapter, for the first time, will present a novel mechanistic model to predict the feed-direction cutting force in RUM surface grinding of CFRP. The

130 Texas Tech University, Fuda Ning, May 2018

mechanistic model will be developed using kinematic and micromechanical analysis, and the brittle fracture removal mechanism. One experiment will be conducted to

obtain the value of fracture volume factor (KV) of the workpiece and comprehensive experiments will be performed to verify the validity of the established model by comparing the predicted values with the measured results. The model presented in this work will provide a guidance for other model developments in RUM surface grinding.

7.2 Model development In CFRP, both carbon fiber and thermoset matrix (for example, epoxy) are brittle materials and carbon fiber exhibits a low fracture toughness value [39]. The material is primarily removed by brittle fracture rather than ductile removal in machining of CFRP, remarkably different from machining of metal [5].

7.2.1 Approach to model development In this study, the main approach to develop the cutting force model starts from an analysis of a single abrasive grain and then sums up the effects of all active abrasive grains taking part in machining. The modeling of the feed-direction cutting force is organized in Figure 7.2. ? Input variables Cutting force Feed-directional (F) cutting force Input variables Workpiece Machining and (Fx) properWtioerkspiece tool variablesMachining & variables tool variables Material removal SummingM alla thxe impact force, Material removal Micromechanical feed-directional rate (MRR) (Fi) rate (MRR) analysis components of Fg

Material Ultrasonic vibration CFRP CwoFrkRpiePce Number of removal frequency Number of properties Grain force (Fg) active micromechanics analysis (f) (E, ν, KC) actiavbera sives (n) process abrasives (n) Volume removed Indentation by single abrasive E, ν, Kc Removed volume depth (δ) grain (V1) Indentation depth, for one abrasive (δ) (V1) Material Fracture volume removal factor (Kv) from experiment analysis Fracture volume factor (Kv) Figure 7.2 Feed-direction cutting force modeling procedures

131 Texas Tech University, Fuda Ning, May 2018

7.2.2 Assumptions in model development Several assumptions about conditions for CFRP micromechanical analysis in grinding processes are proposed for this study: (1). The bond between fibers and matrix is perfect; (2). The elastic moduli and diameters of fibers and space between fibers are uniform; (3). The fibers are continuous and parallel; (4). The fibers and matrix follow Hooke’s law (linearly elastic); (5). The composite is ideally brittle and void-free.

Other assumptions for RUM surface grinding of CFRP in the cutting force model development include: (1). The material is removed by brittle fracture; (2). The cutting tool is perfectly rigid; (3). The abrasive grains are rigid spheres with the same diameter and their shapes do not change during the machining process; (4). All abrasive grains on the periphery of the cutting tool have the same height; (5). The ultrasonic vibration during the machining process is stable (frequency and amplitude remain unchanged).

Additional assumptions and simplifications will be presented later once used.

7.2.3 The kinematics analysis of single abrasive grain in RUM surface grinding The motion of abrasive grains in RUM surface grinding is a combination of spindle rotation, workpiece feed motion, and ultrasonic vibration in the direction of tool axis. On the basis of RUM surface grinding kinematic motion, the trajectory model of a single abrasive grain can be established by

132 Texas Tech University, Fuda Ning, May 2018

 S x  Rcos( t)  f t  30 r  S y  Rsin( t) (1)  30 z  Asin(2ft)  where, R is the radius of the cutting tool, mm; S is the tool rotation speed, rpm; fr is the workpiece feedrate, mm/s; A is the ultrasonic vibration amplitude, mm; f is the ultrasonic vibration frequency, Hz; t is the processing time, s.

From Eq. (1), the trajectory of one abrasive grain in RUM surface grinding can be drawn by giving the initial values (R = 4.8 mm, S = 3000 rpm, f = 20 kHz, and A = 6 μm), as shown in Figure 7.3.

mm Ultrasonic -3 x 10 8 vibration 6

2 mm 0 mm 0.006-2

-4 -6 0 -8 -0.006 Rotation 4.84 3

42 5

1 4

3 3 0 2 ed 2 -1 1 Fe 0 -2 1 -1 -3 -2

-3 Z -4 0 -4 -1 X 4.8 -2 3 4 -3 1 2 -1 0 -4 Y -3 -2 -4.8 -4.8 -4 mm Figure 7.3 The kinematic trajectory of a single abrasive grain in RUM surface grinding process

The velocity of the single abrasive grain can be expressed by

133 Texas Tech University, Fuda Ning, May 2018

 S S v   Rsin( t)  f  x 30 30 r  S S vy  Rcos( t) (2)  30 30  vz  2fA cos(2ft)  The effective trajectory length L of the single abrasive grain during one ultrasonic vibration cycle can be calculated by

1 1 f 2 2 2 2 L  vx  vy  vz  dt (3) 0 Since the feedrate of workpiece is much smaller than the cutting velocity caused by the spindle rotation, the Eq. (3) can be simplified as

1 2 2 1   f   S  R  2 L     2  f  Acos2ft   dt (4) 0    30  

7.2.4 Relationship between indentation depth of abrasive grain into workpiece and radial grain force During RUM surface grinding, the cutting tool is always in continuous contact with workpiece in each ultrasonic vibration cycle, which differs from the intermittent contact mode in RUM process for hole-making.

The indentation depth δ of a single abrasive grain penetrating into the CFRP workpiece surface can be calculated by [40]

1 2 2 3  9 F 1 2      g    (5) 16 d / 2  E      where,

Fg is the radial grain force between one abrasive grain and workpiece, N; d is the average diameter of abrasive grains, mm; E is the elastic modulus of CFRP, MPa;

 is the Poisson’s ratio of CFRP. E and of both carbon fiber and epoxy in CFRP are known. The E and of CFRP will be estimated from those of carbon fiber and epoxy in the Section 7.2.5.

134 Texas Tech University, Fuda Ning, May 2018

By Eq. (5), the radial grain force Fg between one abrasive grain and CFRP workpiece can be expressed as

1 1 3 8 2  E d 2  2 F  (6) g 3(1 2 )

7.2.5 CFRP micromechanical analysis for obtaining elastic modulus and Poisson’s ratio “Micromechanics are the study of composite materials taking into account the interaction of the constituent materials in detail. Micromechanics allow the designer to represent a heterogeneous material (its properties vary from point to point) as an equivalent homogeneous material (it has the same properties everything), usually isotropic” [41], as illustrated in Figure 7.4(a) and Figure 7.4(b). Micromechanics process has been used to predict stiffness with great success. The stiffness of an isotropic material is completely described by two properties including the elastic modulus E and Poisson’s ratio ν. Using micromechanics process, the fracture toughness also can be obtained.

ics an ch me cro Mi

(b) Equivalent homogeneous material

(a) Heterogeneous material

Elastic modulus and Poisson’s ratio are different in longitudinal and transverse directions of the fiber. In RUM surface grinding of CFRP, the major machining load is applied on the longitudinal direction of the fiber. Thus, the elastic modulus and Poisson’s ratio of CFRP in the fiber longitudinal direction should be used in the model

135 Texas Tech University, Fuda Ning, May 2018 development. For simplicity, it is assumed that a cylindrical fiber is treated as a rectangular one, as illustrated in Figure 7.4(c). In reality, most micromechanics formulations do not represent the actual geometry of the fiber at all [41].

Rule of mixtures (ROM) formula [41] is used to predict the elastic modulus of

CFRP in the fiber longitudinal direction (E1)

E1  E f V f  EmVm (7) where

E f is the elastic modulus of fiber material in CFRP;

Em is the elastic modulus of matrix material in CFRP;

V f is the volume fraction of the fiber which can be calculated by volume of fiber V  ; f total volume

Vm is the volume fraction of the matrix which can be calculated by volumeof matrix V  . m total volume There is an assumption in this formulation which states that strains in the longitudinal direction of the fibers are the same as those in the matrix [41].

The Poisson’s ratio of CFRP composite materials in the longitudinal direction is

12  f V f  mVm (8) where vf is Poisson’s ratio of fiber material in CFRP and vm is Poisson’s ratio of matrix material in CFRP.

7.2.6 Material removal volume by one abrasive particle In order to understand the brittle fracture mechanism in RUM surface grinding of CFRP, it is necessary to analyze the interaction between an abrasive grain and the workpiece, as shown in Figure 7.5. The initial loading on the contact area of workpiece induces a permanent plastic deformation. The deformation zone is gradually enlarged until the load increases to the critical value, when median crack

136 Texas Tech University, Fuda Ning, May 2018 forms due to the stress concentration. Then lateral cracks are generated and continue to propagate as a result of the unloading caused by the motion of abrasive grain to the next position. The two adjacent lateral cracks can finally remove materials from the CFRP workpiece. Under this condition, the material removal volume of plastic deformation is contained in the volume removed by lateral cracks, indicating that materials are completely removed by brittle fracture [42].

Material removal volume by a grain Material indentation  removal volume by lateral cracks

Motion trajectory of a single grain in each ultrasonic cycle Workpiece Fg Abrasive grain

Plastic deformation d zone 

CH Lateral crack CL Median crack Workpiece Figure 7.5 Material removal volume by a single abrasive grain in one ultrasonic vibration cycle

The abrasive grain on the periphery face of tool moves along a sine wave. During each ultrasonic vibration period of time 1/f, the indentation δ of the abrasive grain keeps approximately the same. The abrasive grain slides a distance L on the workpiece surface due to the rotation of cutting tool. It can be seen from Figure 7.5 that the fracture cross section can be simplified as an area of a half ellipse with a semi- major axis of CL and a semi-minor axis of CH . So the material removal volume of the single abrasive grain in one ultrasonic vibration cycle V1 can be calculated by

137 Texas Tech University, Fuda Ning, May 2018

1 V   C C  L (9) 1 2 L H where,

CL is the lateral crack length, mm;

CH is the lateral crack depth, mm.

The lateral crack length (CL) and lateral crack depth (CH) are given by [43]

3  F  4  g  (10) CL  kL    KC  1 2 2 (11) CH  kH d    where,

kL is lateral crack length factor;

kH is lateral crack depth factor;

KC is the fracture toughness expressed by stress intensity factor, MPa mm .

can be calculated by [44]

2 KC  E1GC (12) 2 where GC is fracture toughness expressed in elastic energy release rate, J/m .

can be calculated by [44]

GC  2Gcf Vf  GcmVm  (13) G where cf and Gcm are fracture toughness in the elastic energy release rate for fiber and matrix, respectively.

By combining Eq. (12) with Eq. (13), the fracture toughness expressed in stress intensity factor for CFRP can be calculated by

138 Texas Tech University, Fuda Ning, May 2018

1 2 KC  [2E1(Gcf V f  GcmVm )] (14) After substituting Eqs. (4), (10), and (11) into Eq. (9), material removal volume of one abrasive particle can be calculated by

3 1 4 1 2 2 1  F  1    S  R  g 2 2 f   2  V1  KV  d       2  f  Acos2ft  dt (15)   0   2  KC   30  

where KV is a fracture volume factor ( KV  kL kH ) which is a proportionality parameter. It is assumed to be constant for CFRP and will be obtained from the experiment.

7.2.7 Material removal rate and radial grain force model The number of abrasive grains involving in the cutting on the periphery of the cutting tool can be determined by [45]

2  C  3 4 a (16) n  6.5610  3    R  ap  d   where,

Ca is the abrasive concentration;

 is the density of abrasive material, g/mm3,   3.52103 g/mm 3 for diamond;

a p is the cutting depth, mm.

Material removal rate (MRR) can be theoretically calculated by summing V1 of all the involved abrasive particles in the ultrasonic frequency. The MRR can be calculated by

2 3 3 4  C   F  1 MRR  n f V  3.28104  a   2  R a  f  K  g   d   2 2 1  3  p V      d    KC  1 (17) 1 2 2    S  R  2   f    2  f  Acos2ft   dt 0    30  

139 Texas Tech University, Fuda Ning, May 2018

According to its definition, MRR also can be expressed in terms of the feedrate and effective cross-section area of the cutting tool

MRR  2 fr  ap  R (18)

By equating Eq. (17) and Eq. (18), the relationship between Fg and δ can be obtained as

4 1  3  1 2 2   4 4  4  8 8  2  SR  2   3 3 3 9 3 2 3 f   (19) Fg  34.73 KC KV fr f Ca d d        2fA cos2ft  dt  0  30        1 1 3 8 2  E d 2  2 From Eq. (6), we know that F  . Thus, the simultaneous g 3(1 2 ) equations can be established as

 4  1 3  1 2 2    4 4  4  8 8  2   SR  2   3 3 3 9 3 2 3 f   Fg  34.73 KC K V f r f Ca d d        2fA cos2ft  dt  0  30          (20)  1 1 3 8 2  E d 2  2 F   g 3(1 2 )  In Eq. (20), only Fg and δ are two unknowns in the simultaneous two equations. Therefore, the radial grain force Fg can be obtained.

7.2.8 The calculation of feed-direction cutting force

The radial grain force Fg is considered as the basic component of feed- direction cutting force induced in abrasive machining and it is always in the radial direction of the rotation tool. The cutting force model in this work is derived by summing all the feed-direction components of the radial grain forces, as shown in Figure 7.6.

140 Top view Workpiece

Fx2 F Fg x3 F Fg x1 Grain 2 Fg Grain 3 Grain 1 Texas Tech University,R Fuda Ning, May 2018

Feed Tool Top view rotation Workpiece Front view Fx2 F Fg A,f x3 F Fg x1 Grain 2 Fg ap Grain 3 Grain 1 R Feed WorkpWieocrekpiece Feed Tool rotation Front view Figure 7.6 Relationship between grain force Fg and feed-direction force Fx in RUM A,f surfaceap grinding process

The coordinate of any abrasive grain within the contact area between cutting tool and workpieceFee dis WshownorkpWie oincrek Figurepiece 7.7(a) and can be described by

Pi  (y, z) i=1, 2, ···, n. (21) Many investigations assumed that abrasive grains are uniformly distributed on the surface of cutting tool. In this study, the abrasive grains conform to the stochastic distribution u, which is more suitable for the actual distribution of the abrasive grains on the tool. Therefore, y and z can be expressed as

y ~ u 0,  R  (22) z ~ u 0,ap  As shown in Figure 7.7(b), radian θi can be calculated by

y y   i   i (i=1, 2, ···, n.) (23) i   R R The relationship between grain force Fg and its feed-direction component force

Fxi is obtained by

Fxi  Fg sini (24) Therefore, the feed-direction cutting force Fx during RUM surface grinding of CFRP could be calculated by

n n Fx  Fxi  Fg sini (25) i1 i1

141 Texas Tech University, Fuda Ning, May 2018

z z Top viewTop view WorkpiecWeorkpiece

Pi (yi, zi) Pi (yi, zi) ap ap Fg Fxi Fg Fxi θi θi

Pi (yi, zi) Pi (yi, zi)

Feed FeTeodol Tool rotation rotation o o πR πR y y

(a) The coordinate of an abrasive grain (b) Relationships between Fg and Fxi

Figure 7.7 Calculation methodology for feed-direction cutting force Fx

7.3 Experimental set-up and conditions

In order to obtain the value of fracture volume factor KV and also to verify the validity of the established feed-direction cutting force model, a series of experiments were conducted to measure the cutting force results under different combinations of input variables.

7.3.1 Workpiece material properties The CFRP composite workpiece used in this investigation consisted of carbon fibers and epoxy resin matrix with the volume fractions of 67% and 33%, respectively. The workpiece was multi-directional CFRP composite laminated by 23 layers of carbon fabric and 24 plies of epoxy resin. The CFRP workpiece used in this experiment had the size of 150 mm × 18 mm × 18 mm. Properties of the CFRP workpiece material are listed in Table 7.1. Some of them were calculated by the micromechanical analysis.

Table 7.1 Properties of CFRP workpiece material Property Unit Value Density of CFRP kg/m3 1600 Hardness (Rockwell) HRB 68–72 Poisson’s Ratio (v12) - 0.33 Poisson’s Ratio (v13) - 0.33 Poisson’s Ratio (v23) - 0.42 Longitudinal Young’s modulus (E1) GPa 155.6 Transverse Young’s modulus (E2) GPa 13.1

142 Texas Tech University, Fuda Ning, May 2018

In-plane shear modulus (G12) GPa 4.7 Density of epoxy matrix kg/m3 1200 Poisson’s ratio of epoxy matrix - 0.4 Young’s modulus of epoxy matrix GPa 4.5 2 Fracture toughness of epoxy matrix (Energy/Gc) J/m 500 Density of carbon fiber kg/m3 1800 Poisson’s ratio of carbon fiber - 0.3 Young’s modulus of carbon fiber GPa 230 2 Fracture toughness of carbon fiber (Energy/Gc) J/m 2

7.3.2 Experimental set-up and cutting force measurement The experiments of this investigation were performed on a SONICMILL Series 10 rotary ultrasonic machine. The RUM surface grinding set-up was shown in Figure 7.8. In the ultrasonic spindle system, the ultrasonic power supply was used to convert low-frequency line electricity to high-frequency electrical energy. Such energy was supplied to the piezoelectric converter located inside the ultrasonic spindle to generate the ultrasonic vibration at a high frequency (20 kHz). The ultrasonic vibration was amplified and then transmitted to the metal-bonded diamond tool, causing the tool to vibrate vertically. The ultrasonic vibration amplitude could be adjusted by changing the output of the power supply. The motor assembled atop the ultrasonic spindle provided a rotation motion of the tool. Meanwhile, a NEWMARK linear stage with a maximum travel of 400 mm enabled an infeed motion of the workpiece to realize the RUM surface grinding process. The feedrate could be controlled by a motor controller and a software.

143 Texas Tech University, Fuda Ning, May 2018

Data acquisition Ultrasonic system 3000 spindle system Feeding device Motor Electric controller motor Control panel Ultrasonic Computer spindle

Power supply A/D converter Pressure Transformer gauge and tool holder Pressure Flow rate Valve Channels Valve 1~4 Diamond tool regulator gauge Workpiece Pump Amplifier Fixture Dynamometer Motor Coolant Platform Abrasive tank portion Linear stage Fixture Feed direction Machine Horizontal feeding table system Coolant system

Figure 7.8 RUM surface grinding set-up

The KISTLER 9272 dynamometer in the data acquisition system was used to measure the feed-direction cutting force Fx, together with a KISTLER 5070 charge amplifier and a KISTLER 5697A A/D converter. The digital signals were collected by a KISTLER DynoWare software. In the RUM surface grinding process, the cutting force changed with time and fluctuated in a certain range. Thus, the average value of the stable machining phase on the Fx curve was selected to represent the force value.

7.3.3 Experimental conditions RUM surface grinding involves many input variables include tool variables and machining variables. The major tool variables investigated in this work are abrasive grain size and abrasive concentration. The detailed information of the cutting tools was listed in Table 7.2.

Table 7.2 Identifications of cutting tools in the experiments Tool variables Tool #1 Tool #2 Tool #3 Tool #4 Tool #5 Tool diameter D (mm) 12.7 12.7 12.7 9.6 9.6 Mesh size 60/80 80/100 120/140 80/100 80/100

144 Texas Tech University, Fuda Ning, May 2018

Abrasive grain size d (mm) 0.215 0.165 0.115 0.165 0.165 Abrasive concentration Ca 100 100 100 75 100

In addition, four machining variables (tool rotation speed, feedrate, depth of cut, and ultrasonic amplitude) were studied and their ranges were set based on preliminary experimental results. The ultrasonic amplitude during RUM surface grinding was measured using the optical microscope method [46]. The specific experimental conditions could be found in Table 7.3. Each group of the experiment was designed to investigate the effects of each input variable.

Table 7.3 Experimental conditions

Depth of Ultrasonic Tool rotation speed Feedrate Group cut amplitude Tool # S (rpm) Fr (mm/s) ap (mm) A (µm) 1 2000; 3000; 4000; 5000 0.8 2 6 #5 2 3000 0.4; 0.6; 0.8; 1 2 6 #5 3 3000 0.8 1; 2; 3; 4 6 #5 4 3000 0.8 2 4; 6; 7; 8 #5 5 3000 0.4 1 5 #1; #2; #3 6 3000 0.4 1 6 #4; #5

7.4 Fracture volume factor KV acquisition and model verification

7.4.1 Obtaining the value of fracture volume factor KV

KV represents the fracture volume factor. It is assumed that KV is constant for the given CFRP material and is independent of input variables. Thus, it can be obtained by one experiment with one combination of input variables. In order to avoid using the upper and lower limit of each input variable range, a combination of tool rotation speed of 3000 rpm, feedrate of 0.8 mm/s, depth of cut of 2 mm, and ultrasonic amplitude of 6 µm was selected to calculate KV using the cutting tool of #5. By equating Eq. (17) and Eq. (18), KV can be calculated by

145 Texas Tech University, Fuda Ning, May 2018

MRR K  V MRR' 2 f  R  a  r p (26) 2 3 1 4 2 2 3 1 1  C   F  f  SR  4 a 2 g 2 2   2  3.2810     R  a p  f   d       2fA cos2ft  dt  3    0    d    KC   30   where, MRR’ is proportional to MRR of all the abrasive grains on the periphery of the cutting tool and MRR’ represents MRR when KV equals to 1. Fg can be calculated from the experimentally measured force Fx using Eq. (25) and δ can be thereafter obtained from Eq. (5). After calculation, the KV value was 0.508 and it would be input into the model to predict the feed-direction cutting force.

7.4.2 Model validity verification and predicted influences Based on the developed mechanistic model, the relationships between feed- direction cutting forces and input variables (including machining variables and tool variables) can be predicted. In order to verify the validity of the established model, all the six groups of experiments in Table 7.3 were conducted for the comparisons between the theoretically predicted values and experimentally measured results.

Comparisons of predicted and experimental feed-direction cutting forces regarding effects of machining variables are shown in Figure 7.9. The error bar with one standard deviation was used to represent the data distribution. It can be seen that the predicted trends with the increase of each machining variable agree well with the trends obtained experimentally. The predicted influences show that the increase of tool rotation speed S and ultrasonic vibration amplitude A led to the decrease of the feed- direction cutting force Fx, while feedrate fr and depth of cut ap had positive relationships with Fx. Similar findings regarding the predicted influences of S, fr, and A on axial cutting force using a mechanistic model were exhibited in the RUM hole making of CFRP [22].

146 Texas Tech University, Fuda Ning, May 2018

100 100 Predicted results Predicted results Experimental results Experimental results 80 80

60 60

40 40

Cutting force (N) Cutting force (N) 20 20

0 0 1000 2000 3000 4000 5000 6000 0.2 0.4 0.6 0.8 1 1.2 Tool rotation speed (rpm) Feedrate (mm/s) (a) Effects of tool rotation speed (b) Effects of feedrate

140 100 Predicted results Predicted results 120 Experimental results Experimental results 80 100 60 80

60 40 Cutting force (N)

Cutting force (N) 40 20 20

0 0 0 1 2 3 4 5 3 4 5 6 7 8 9 Depth of cut (mm) Ultrasonic vibration amplitude (um) (c) Effects of depth of cut (d) Effects of ultrasonic vibration amplitude Figure 7.9 Comparisons of predicted and experimental feed-direction cutting forces under different machining variables

The reasons for the trends of Fx under each machining variable were given as follows. During RUM surface grinding of CFRP, the predicted influences of S and fr on Fx were attributed to the relations in Eq. (19). It is notable that the increase of S or decrease of fr would result in a reduced radial grain force Fg, causing the decrease of

Fx. under the unchanged amount n of active abrasive grains during the machining, according to the Eq. (25). Different from the increase of S and fr, the increase of ap would not change Fg but led to an increased number of active abrasive grains involved

147 Texas Tech University, Fuda Ning, May 2018

according to Eq. (16). Thus, it can be found from Eq. (25) that Fx would increase with the increase of ap. To explain effects of ultrasonic vibration amplitude A on the cutting force, the material removal volume V1 of the single abrasive grain in one ultrasonic vibration cycle needs to be analyzed. Based on Eq. (17) and Eq. (18), V1 can be expressed by

2 f a  R 70.6 f d 2 r p r (27) V1   2 n f 3 Ca   f Eq. (27) shows that there are no causal relationships between V1 and A, which indicates that V1 would not be determined by A. Therefore, with the increase of A, the grain indentation depth δ decreased to keep the removed volume unchanged, as shown in Figure 7.10. It can be observed from Eq. (6) that the radial grain force Fg would be decreased, leading to the reduction of the feed-direction cutting force Fx.

Figure 7.11 shows the comparisons between predicted and experimental feed- direction cutting forces under different tool variables. A good consistency between the theoretically predicted trends and the experimentally observed trends was obtained. Fx decreased as tool abrasive size d increased or as abrasive concentration Ca decreased. The reasons for such phenomena could be explained by exploring the Eqs. (6), (16),

3/ 2 2 / 3 and (25). It can be concluded that Fx is proportional to d and Ca . Therefore, Fx decreased with the increase of d or with the decrease of Ca.

 1 2 A1 A2

148 Texas Tech University, Fuda Ning, May 2018

Figure 7.10 Relationships between ultrasonic vibration amplitude A and indentation depth δ

25 25 Predicted results Predicted results Experimental results Experimental results 20 20

15 15

10 10

Cutting force (N) Cutting force (N) 5 5

0 0 0.115 0.165 0.215 75 100 Abrasive size (mm) Abrasive concentration (a) Effects of abrasive size (b) Effects of abrasive concentration Figure 7.11 Comparisons of predicted and experimental feed-direction cutting forces under different tool variables.

7.4.3 Experimental validation of KV value

During the model development, it was assumed that KV value, independent to input variables, was constant for the given CFRP material and was obtained by an experiment under one combination of input variables. Apart from the combination of input variables for KV calculation in Section 7.4.1, all the other combinations of input variables in Table 3 were also used to obtain the tested KV values for the validation. 2 These KV values can be found in Figure 7.12. It can be seen that the R value between all the tested KV values and calculated KV value used in the model development is

0.954, indicating that KV value and its assumption used in the model development was reasonable.

149 Texas Tech University, Fuda Ning, May 2018

40 K =0.508 (Calculated) V 35 Tested points 30

/s) 3 25

MRR (mm 15 R =0.954

0 0 10 20 30 40 50 60 70 3 MRR' (mm /s)

Figure 7.12 Fracture volume factor Kv values under different experimental conditions

7.5 Conclusions A mechanistic model for the feed-direction cutting force in RUM surface grinding of CFRP was developed in this chapter. The kinematic motion of a single abrasive grain during RUM surface grinding was analyzed to calculate the material removed volume based on the brittle fracture mechanism. The radial grain force was obtained and then aggregated for feed-direction cutting force calculation by considering all the involved abrasive grains. One experiment is conducted to obtain the value of fracture volume factor (KV) of the workpiece material and comprehensive experiments are performed to verify the validity of the established model by comparing the predicted values with the measured results. The detailed conclusions were drawn as follows:

(1) The theoretically predicted trends of the feed-direction cutting force agreed well with the experimentally measured results with changes of different input variables.

(2) The influences of the input variables including machining variables and tool variables on the feed-direction cutting force were that the force decreased as tool rotation speed, ultrasonic vibration amplitude, and abrasive size increased and as

150 Texas Tech University, Fuda Ning, May 2018 feedrate, depth of cut, and abrasive concentration decreased. The reasons for these phenomena were discussed using the developed mechanistic model.

(3) KV value was successfully obtained by one combination of input variables. Experiments also show that KV is an approximately constant parameter for the given CFRP material.

Future investigations will focus on the development of feed-direction cutting force model in RUM of CFRP assisted with feed-direction ultrasonic vibration and study the effects of ultrasonic vibration directions on the cutting force.

References [1]. Mallick, P.K., 1997, Composite Engineering Handbook, CRC Press, New York. [2]. Davim, J.P., and Reis, P., 2003, Drilling carbon fiber reinforced plastics manufactured by autoclave-experimental and statistical study, Materials & design, 24(5), pp. 315–324. [3]. Cong, W.L., and Ning, F.D., in: J.P. Davim (Ed.), Machinability of Fibre- Reinforced Plastics, Walter de Gruyter GmbH & Co KG, Berlin, 2015, pp. 31– 81. [4]. Ning, F.D., Cong, W.L., Hu, Y.B., and Wang, H., 2017, Additive manufacturing of carbon fiber-reinforced plastic composites using fused deposition modeling: Effects of process parameters on tensile properties, Journal of Composite Materials, 51(4), pp. 451–462. [5]. Karpat, Y., Bahtiyar, O., and Değer, B., 2012, Mechanistic force modeling for milling of unidirectional carbon fiber reinforced polymer laminates, International Journal of Machine Tools and Manufacture, 56, pp. 79–93. [6]. Jia, Z., Fu, R., Wang, F., Qian, B., and He, C., 2016, Temperature effects in end milling carbon fiber reinforced polymer composites, Polymer Composites, pp. 1–11. [7]. Jia, Z., Su, Y., Niu, B., Zhang, B., and Wang, F., 2016, The interaction between the cutting force and induced sub-surface damage in machining of carbon fiber- reinforced plastics, Journal of Reinforced Plastics and Composites, 35(9), pp. 712–726. [8]. Jia, Z., Fu, R., Niu, B., Qian, B., Bai, Y., and Wang, F., 2016, Novel drill structure for damage reduction in drilling CFRP composites, International Journal of Machine Tools and Manufacture, 110, pp. 55–65. [9]. Ning, F.D., and Cong W.L., 2015, Rotary ultrasonic machining of CFRP: design of experiment with a cutting force model, In: Proceedings of the 2015 International Manufacturing Science and Engineering Conference, Charlotte, North Carolina, USA, pp. V001T02A040–V001T02A048. [10]. Su, F., Wang, Z., Yuan, J., and Cheng, Y., 2015, Study of thrust forces and

151 Texas Tech University, Fuda Ning, May 2018

delamination in drilling carbon-reinforced plastics (CFRPs) using a tapered drill-reamer, International Journal of Advanced Manufacturing Technology, 80(5–8), pp. 1457–1469. [11]. Marques, A.T., Durão, L.M., Magalhães, A.G., Silva, J.F., and Tavares, J.M.R., 2009, Delamination analysis of carbon fibre reinforced laminates: evaluation of a special step drill, Composites Science and Technology, 69(14), pp. 2376– 2382. [12]. Turki, Y., Habak, M., Velasco, R., Aboura, Z., Khellil, K., and Vantomme, P., 2014, Experimental investigation of drilling damage and stitching effects on the mechanical behavior of carbon/epoxy composites, International Journal of Machine Tools and Manufacture, 87, pp. 61–72. [13]. Tsao, C.C., and Hocheng, H., 2005, Computerized tomography and C-Scan for measuring delamination in the drilling of composite materials using various drills, International Journal of Machine Tools and Manufacture, 45(11), pp. 1282–1287. [14]. Bertsche, E., Ehmann, K., and Malukhin, K., 2013, An analytical model of rotary ultrasonic milling, International Journal of Advanced Manufacturing Technology, 65(9–12), pp. 1705–1720. [15]. Pecat, O., Rentsch, R., and Brinksmeier, E., 2012, Influence of milling process parameters on the surface integrity of CFRP, Procedia CIRP, 1, pp. 466-470. [16]. Davim, J.P., and Reis, P., 2005, Damage and dimensional precision on milling carbon fiber-reinforced plastics using design experiments, Journal of Materials Processing Technology, 160(2), pp. 160–167. [17]. Soo, S.L., Shyha, I.S., Barnett, T., Aspinwall, D.K., and Sim, W.M., 2012, Grinding performance and workpiece integrity when superabrasive edge routing carbon fibre reinforced plastic (CFRP) composites, CIRP Annals- Manufacturing Technology, 61(1), pp. 295–298. [18]. Sasahara, H., Kikuma, T., Koyasu, R., and Yao, Y., 2014, Surface grinding of carbon fiber reinforced plastic (CFRP) with an internal coolant supplied through grinding wheel, Precision Engineering, 38(4), pp. 775–782. [19]. Zhang, J.H., Zhao, Y., Tian, F.Q., Zhang, S., and Guo, L.S., 2015, Kinematics and experimental study on ultrasonic vibration-assisted micro end grinding of silica glass, International Journal of Advanced Manufacturing Technology, 78(9–12), pp. 1893–1904. [20]. Cong, W.L., Pei, Z.J., Feng, Q., Deines, T.W., and Treadwell, C., 2012, Rotary ultrasonic machining of CFRP: a comparison with twist drilling, Journal of Reinforced Plastics and Composites, 31(5), pp. 313–321. [21]. Ning, F.D., Cong, W.L., Pei, Z.J., and Treadwell, C., 2016, Rotary ultrasonic machining of CFRP: a comparison with grinding, Ultrasonics, 66, pp. 125–132. [22]. Cong, W.L., Pei, Z.J., Sun, X., and Zhang, C.L., 2014, Rotary ultrasonic machining of CFRP: a mechanistic predictive model for cutting force, Ultrasonics, 54(2), pp. 663–675. [23]. Yuan, S., Zhang, C., Amin, M., Fan, H., and Liu, M., 2015, Development of a cutting force prediction model based on brittle fracture for carbon fiber

152 Texas Tech University, Fuda Ning, May 2018

reinforced polymers for rotary ultrasonic drilling, International Journal of Advanced Manufacturing Technology, 81(5–8), pp. 1223–1231. [24]. Ning, F., Wang, H., Cong, W., and Fernando, P.K.S.C., 2017, A mechanistic ultrasonic vibration amplitude model during rotary ultrasonic machining of CFRP composites, Ultrasonics, 76, pp. 44–51. [25]. Liu, J., Zhang, D., Qin, L., and Yan, L., 2012, Feasibility study of the rotary ultrasonic elliptical machining of carbon fiber reinforced plastics (CFRP), International Journal of Machine Tools and Manufacture, 53(1), pp. 141–150. [26]. Cong, W.L., Pei, Z.J., Deines, T.W., and Treadwell, C., 2011, Rotary ultrasonic machining of CFRP using cold air as coolant: feasible regions, Journal of Reinforced Plastics and Composites, 30(10), pp. 899–906. [27]. Cong, W.L., Pei, Z.J., Deines, T.W., Srivastava, A., Riley, L., and Treadwell, C., 2012, Rotary ultrasonic machining of CFRP composites: a study on power consumption, Ultrasonics, 52(8), pp. 1030–1037. [28]. Pei, Z.J., Ferreira, P.M., Kapoor, S.G., and Haselkorn, M.B.A.C., 1995, Rotary ultrasonic machining for face milling of ceramics, International Journal of Machine Tools and Manufacture, 35(7), pp. 1033–1046. [29]. Uhlmann, E., and Daus, N.A., 2001, Ceramics Materials and Components for Engines, Wiley-VCH Verlag GmbH, Weinheim, Germany, pp. 417–422. [30]. Gong, H., Fang, F.Z., and Hu, X.T., 2010, Kinematic view of tool life in rotary ultrasonic side milling of hard and brittle materials, International Journal of Machine Tools and Manufacture, 50(3), pp. 303–307. [31]. Wang, H., Ning, F., Hu, Y., Fernando, P.K.S.C., Pei, Z.J., and Cong, W., 2016, Surface grinding of carbon fiber-reinforced plastic composites using rotary ultrasonic machining: effects of tool variables, Advances in Mechanical Engineering, 8(9), pp. 1–14. [32]. Liu, S., Chen, T., and Wu, C., 2017, Rotary ultrasonic face grinding of carbon fiber reinforced plastic (CFRP): a study on cutting force model, International Journal of Advanced Manufacturing Technology, 89(1–4), pp. 847–856. [33]. Zhang, C.L., Feng, P.F., Zhang, J.F., Wu, Z.J., and Yu, D.W., 2012, Theoretical and experimental research on the features of cutting force in rotary ultrasonic face milling of K9 glass, Applied Mechanics and Materials, 157, pp. 1674–1679. [34]. Zhang, C., Zhang, J., and Feng, P., 2013, Mathematical model for cutting force in rotary ultrasonic face milling of brittle materials, International Journal of Advanced Manufacturing Technology, 69(1–4), pp. 161–170. [35]. Pei, Z.J., and Ferreira, P.M., 1999, An experimental investigation of rotary ultrasonic face milling, International Journal of Machine Tools and Manufacture, 39(8), pp. 1327–1344. [36]. Xiao, X., Zheng, K., and Liao, W., 2014, Theoretical model for cutting force in rotary ultrasonic milling of dental zirconia ceramics, International Journal of Advanced Manufacturing Technology, 75(9–12), pp. 1263–1277. [37]. Zhang, C., Yuan, S., Amin, M., Fan, H., and Liu, Q., 2016, Development of a cutting force prediction model based on brittle fracture for C/SiC in rotary ultrasonic facing milling, International Journal of Advanced Manufacturing

153 Texas Tech University, Fuda Ning, May 2018

Technology, 85(1–4), pp. 573–583. [38]. Yuan, S., Fan, H., Amin, M., Zhang, C., and Guo, M., 2016, A cutting force prediction dynamic model for side milling of ceramic matrix composites C/SiC based on rotary ultrasonic machining, International Journal of Advanced Manufacturing Technology, 86(1–4), pp. 37–48. [39]. Gay, D., Hoa, S.V., and Tsai, S.V., 2003, Composite Materials Design and Applications, CRC Press, New York. [40]. Pei, Z.J., Prabhakar, D., Ferreira, P.M., and Haselkorn, M., 1995, A mechanistic approach to the prediction of material removal rates in rotary ultrasonic machining, Journal of Engineering for Industry, 117(2), pp. 142–151. [41]. Kaw, A.K., 2006, Mechanics of Composite Materials, CRC Press, New York. [42]. Wang, Y., Lin, B., Wang, S., and Cao, X., 2014, Study on the system matching of ultrasonic vibration assisted grinding for hard and brittle materials processing, International Journal of Machine Tools and Manufacture, 77, pp. 66–73. [43]. Komaraiah, M., and Reddy, P.N., 1993, A study on the influence of workpiece properties in ultrasonic machining, International Journal of Machine Tools and Manufacture, 33(3), pp. 495–505. [44]. Matthews, F.L., Davies, G.A.O., Hitchings, D., and Soutis, C., 2003, Finite Element Modeling of Composite Materials and Structures, CRC Press, New York. [45]. Liu, D.F., Cong, W.L., Pei, Z.J., and Tang, Y.J., 2012, A cutting force model for rotary ultrasonic machining of brittle materials, International Journal of Machine Tools and Manufacture, 52(1), pp. 77–84. [46]. Cong, W.L., Pei, Z.J., Mohanty, N., Van Vleet, E., and Treadwell, C., 2011, Vibration amplitude in rotary ultrasonic machining: a novel measurement method and effects of process variables, Journal of Manufacturing Science and Engineering, 133(3), pp. 034501-1–034501-6.

154 Texas Tech University, Fuda Ning, May 2018

CHAPTER VIII

A FUNDAMENTAL INVESTIGATION ON ULTRASONIC VIBRATION-ASSISTED (UV-A) LASER ENGINEERED NET SHAPING OF STAINLESS STEEL

Paper titles: aMicrostructures and mechanical properties of Fe-Cr stainless steel parts fabricated by ultrasonic vibration-assisted laser engineered net shaping process bA fundamental investigation on ultrasonic vibration-assisted laser engineered net shaping process

Published in: aMaterials Letters (2016), Vol. 179, pp. 61-64. bInternational Journal of Machine Tools and Manufacture (2017), Vol. 121, pp. 61-69.

Authors: aFuda Ning and Weilong Cong bWeilong Cong and Fuda Ning

Authors’ affiliations:

Department of Industrial, Manufacturing, and Systems Engineering, Texas Tech University, Lubbock, TX, USA.

155 Texas Tech University, Fuda Ning, May 2018

Abstract Laser engineered net shaping (LENS), as a laser beam deposition additive manufacturing method, has been utilized as a key technology in the direct manufacturing or repairing of metal parts. However, deposition defects such as pores, cavity, micro-cracks, residual stress, and uncertain microstructures always exist in the LENS fabricated parts, which will greatly affect the qualities and mechanical properties. In this chapter, a novel ultrasonic vibration-assisted (UV-A) LENS process is proposed to reduce or eliminate the common defects due to the nonlinear actions and influences of ultrasonic vibration in molten materials. An experimental investigation is conducted on the effects of ultrasonic vibration on fabricated part geometry, powder utilization efficiency, surface roughness, geometry of molten pool and dilution zone, pores and micro-cracks, and grain size of the LENS-deposited stainless steel AISI 630 thin walls. The mechanical properties including tensile properties and hardness of the fabricated parts are evaluated and compared between UV-A LENS and LENS without ultrasonic vibration. The results show that ultrasonic vibration led to higher powder utilization efficiency, smaller flatness and surface roughness, and larger molten pool dimensions. Pores and micro-cracks were successfully reduced and crystal grains were significantly refined in UV-A LENS process. The improvement of these geometrical and microstructural characteristics induced by ultrasonic vibration further led to the increase in both tensile properties and hardness of LENS fabricated parts. The fundamental investigation in this work will help to establish an efficient and effective process for additive manufacturing and remanufacturing of metal parts with significantly improved qualities.

Keywords: Laser additive manufacturing; Laser engineered net shaping (LENS); Ultrasonic vibration; Microstructures; Mechanical properties.

8.1 Introduction The terminology of additive manufacturing (AM) technology was standardized and defined as “a process of joining materials to make objects from 3D model data, usually layer upon layer, as opposed to subtractive manufacturing methodologies” by

156 Texas Tech University, Fuda Ning, May 2018 the American Society for Testing and Materials (ASTM) in 2012 [1]. The AM technologies have been widely applied for metal materials manufacturing in numerous industries such as aerospace, marine, automotive, medical instrument manufacturing, tool manufacturing, etc. [2, 3]. Among all the major metal AM methods, laser additive manufacturing (LAM) has become the most popular method for direct deposition of metal materials, due to its advantages of high power density, excellent stability, and easy controllability [2-5]. LAM (including powder bed fusion AM and laser beam deposition AM) is a competitive method for fabrication of metal parts with complex structures which are highly expensive and difficult to be produced by conventional manufacturing processes. Compared with the powder bed fusion AM such as selective laser sintering/melting (SLS/M), laser engineered net shaping (LENS), as a laser beam deposition AM method, has the advantages of parts repairing and remanufacturing capability, high powder utilizing efficiency, high part building efficiency, smaller heat-affected zone, etc. [6, 7].

LENS process incorporates the features of laser cladding and is developed by Sandia National Laboratories, as illustrated in Figure 8.1. The substrate is melted by laser radiation to form a small molten pool that catches and melts metal powders. In the meantime, the powders are continuously delivered into the molten pool by a flowing inert gas (such as Argon) stream through the coaxial nozzle. The inert gas is also used to shield the LENS-fabricated parts from oxidation. With the melting of the powders, the volume of the molten pool is increased. After the leaving of laser beam radiation, the molten pool begins to solidify as a consequence of the heat dissipation. As the powder stream and laser beam (deposition head) moving according to the trajectory of designed structures, the first layer is deposited on the substrate. Afterwards, the laser deposition head ascends one layer thickness to a new position for the next layer deposition. The first layer can serve as the “substrate” which will be melted to form the initial molten pool of the next layer. Such process will be repeated many times until a designed three-dimensional (3D) near-net shape component is built layer by layer directly from the computer-aided design (CAD) model. The principles

157 Texas Tech University, Fuda Ning, May 2018 of LENS and laser cladding are mainly different in the details of processes. In LENS process, high precision and perfect control of the geometrical characteristics of the deposit can be obtained. In addition, unlike laser cladding that is mostly conducted on robots to modify or repair the surfaces by cladding thin-layer materials, LENS is performed on a three- or five-axis machine that interprets the programs translated directly from the sliced CAD model for building bulk parts layer by layer.

Z motion

Processing direction

Laser beam Powder stream Shield gas

Dilution Molten pool area 3rd layer 2nd layer Fabricated near-net 1st layer shape part Substrate Machine table X-Y

Figure 8.1 Laser engineered net shaping (LENS) process

Many investigations have been conducted to evaluate the microstructures and mechanical properties of LENS-fabricated metal parts. In these investigations, most fabricated parts exhibit various fabrication defects, including porosity [8-13, 16, 17], cavity and cracking [14], residual stress [15, 18], large heat-affected zone [3, 18-20], uncertain microstructures [10, 12, 21], etc., which will greatly affect the qualities and mechanical properties of the fabricated parts. Therefore, investigating a high-efficient, cost-effective, and high-quality LENS process to build metal materials is crucial.

In the present study, ultrasonic vibration-assisted (UV-A) LENS process is proposed to reduce or eliminate common defects in the fabricated metal materials. In order to present a fundamental understanding on UV-A LENS process, a comprehensive and deep study on the fabrication of AISI 630 stainless steel by LENS

158 Texas Tech University, Fuda Ning, May 2018 process with and without ultrasonic vibration is conducted for the first time. The main objective of this work is to identify the influence of ultrasonic vibration introduced on the geometrical, microstructural, and mechanical performance of the LENS fabricated parts. For this purpose, powder utilization efficiency, deposited shape geometry, surface roughness, molten pool and dilution zone geometry, pores and micro-cracks, and grain size of the as-deposited AISI 630 materials in LENS process without and with ultrasonic vibration are evaluated. The mechanical properties including tensile properties and hardness are also tested and compared. The obtained knowledge of this fundamental investigation will fill the research gaps in the literature on LENS of metal materials and provide a significant advance in additive manufacturing and remanufacturing of metal parts.

8.2 Experimental procedures

8.2.1 Materials AISI 630 is an important class of precipitation-hardening martensitic stainless steel that has been extensively employed in the industries of aerospace, marine, and chemical. AISI 630 is also referred to 17-4 alloy as it contains 17% chromium and 4% nickel, making it superior in strength and corrosion resistance at temperatures up to 316 °C. In this work, the as-received AISI 630 stainless steel powders (Carpenter Powder Products Inc., Bridgeville, PA, USA) were prepared as the depositing feedstock material with a particle size range of 45–105 µm, as shown in Figure 8.2. Low carbon steel plates (McMaster-Carr Co., Elmhurst, IL, USA) with dimensions of 100 mm × 50 mm × 6.4 mm were used as the substrate material. The plate surface was polished and then cleaned by acetone prior to the LENS fabrication.

159 Texas Tech University, Fuda Ning, May 2018

Figure 8.2 AISI 630 powder morphology

8.2.2 Experimental set-up and parameters The LENS process was conducted on a customized laser additive manufacturing system (450XL, Optomec Inc., Albuquerque, NM, USA) equipped with an ultrasonic vibration unit. The schematic of UV-A LENS system set-up is illustrated in Figure 8.3. The experimental set-up mainly included an IPG fiber laser source with a maximum output power of 400 W, a coaxial deposition head for powder and inert gas delivery, a three-axis motion numeric control system, an ultrasonic vibrator assembly, and an ultrasonic power supply. The ultrasonic power supply provided the high-frequency (41 kHz in this study) electrical energy to a piezoelectric ceramic vibrator. The ultrasonic vibration was generated and then transmitted onto the substrate via an acoustic transformer. Thus, a vertical ultrasonic vibration was generated on the surface of the substrate during UV-A LENS process.

160 Texas Tech University, Fuda Ning, May 2018

Beam delivery Laser system system

Laser beam Laser generator Power supply Fan

Hooper Z Chamber motion (system)

Delivery lines Powder Deposition head Feeding Nozzle Powder and motor inert gas flow Deposited Molten pool material Control Substrate Ar Inert gas Vibration system with Acoustic integrated transformer - Power computer Ceramic + supply vibrator Damper Powder & inert gas delivery system Y X X X-Y motion table Integrated UV-A unit

Figure 8.3 Schematic of ultrasonic vibration-assisted LENS system set-up

A single-bead thin wall structure with five layers was built by both UV-A LENS and LENS without ultrasonic vibration for the purpose of a fundamental investigation to provide the basic understanding on UV-A LENS method. The specific manufacturing parameters for thin wall fabrications were listed in Table 8.1.

Table 8.1 The UV-A LENS manufacturing parameters for AISI 630 thin wall deposition

Parameter Value Unit Laser power 350 W Powder flow rate 4 g/min Increment of Z axis 0.38 mm Laser spot size 0.1 mm Argon gas flow rate 6 L/min Axis feedrate 508 mm/min Number of layers 5 Ultrasonic frequency 41 kHz Ultrasonic power 60 W

161 Texas Tech University, Fuda Ning, May 2018

According to the values in Table 8.1, the theoretical energy density ɛ associated with the processes could be calculated by [22, 23] P   (1) v  h where, P is the laser power only in LENS process and it becomes the sum of laser power and ultrasonic power in UV-A LENS process, W; v is scanning speed, mm/s; h is the hatch spacing and it becomes equal to the spot size in the case of overlapping zero, mm. Thus the energy density for LENS process with and without ultrasonic vibration is 484 J/mm2 and 413 J/mm2, respectively.

8.2.3 Characterizations of geometries, microstructures, and properties The weight of the deposited five-layer thin wall could be obtained by measuring the mass difference between before and after LENS process using a precision balance (PN-6100 A, American Weigh Scales INC., Norcross, GA, USA). The measurement was repeated five times using five samples fabricated by UV-A LENS and LENS without ultrasonic vibration. Thus, the powder utilization efficiency could be calculated by the ratio of thin wall weight to the total mass of powder fed into the chamber over the LENS fabrication period.

The deposited shape geometries including length, width, height, and flatness were measured by a vernier caliper (IP-67, Mitutoyo Corp., Kanagawa, Japan). The surface flatness values were obtained using Global Backside Ideal focal plane Range (GBIR) measurement.

The surface roughness (Ra) on the top surface of the thin walls was measured by a surface profilometer (SJ210, Mitutoyo Corporation, Kanagawa, Japan). The tested range and cut-off length were set at 4 mm and 0.8 mm, respectively. The measurement of Ra was conducted on five samples with three times repeated for each sample. Thus, 15 Ra values in total were obtained for each fabrication condition. The

162 Texas Tech University, Fuda Ning, May 2018 fabricated surface morphology was observed by a field emission scanning electron microscopy (FE-SEM) (S4300, Hitachi Co., Tokyo, Japan).

To observe the geometry of molten pool and dilution zone, thin walls were deposited across two separate mating substrates and then fractured into two pieces by the tension force applied to the two mating parts. The geometries including molten pool width, molten pool depth, and heat affected zone (HAZ) depth were observed and measured using the FE-SEM. In addition, samples were prepared by standard metallographic techniques to observe and analyze the deposition defects and microstructures including porosity, micro-cracks, and grain size under the FE-SEM. The sectioned samples were ground and polished on a grinder-polisher machine (MetaServ 250, Buehler, Lake Bluff, IL, USA) and then etched in a 2% Nital solution for 2 mins to reveal the cross-sectional microstructures of the thin walls.

Tensile tests were conducted to obtain tensile properties using a universal testing machine (AGS-J, Shimadzu Co., Kyoto, Japan) at a crosshead speed of 2 mm/min. The relationships between force (N) and displacement (mm) were collected by a computer with the help of a data acquisition software (Trapezium, Shimadzu Co., Kyoto, Japan). In order to realize the tensile testing, the thin wall samples were deposited on two small separate substrates that were tightly closed to each other, so that the samples could be clamped by the machine grips. In addition, microhardness was measured using a Vickers microhardness tester (900-390A Phase II, Metal-Testers Inc., Nanuet, NY, USA). Both tensile test and microhardness test were performed five times to obtain the experimental results for evaluating effects of ultrasonic vibration on mechanical properties.

8.3 Results and discussion Boxplots were used to represent the data distribution of output variables and conduct the comparisons between two manufacturing conditions: UV-A LENS and LENS without ultrasonic vibration. All the boxplots had identical representations including mean, median, 25% and 75% percentile of confidence intervals, and outliers,

163 Texas Tech University, Fuda Ning, May 2018 as illustrated in Figure 8.4. The mean values were primarily utilized for a comparison on each output variable between LENS with and without ultrasonic vibration.

8.3.1 Powder utilization efficiency Powder utilization efficiency could directly determine the build height per layer and fabrication rate, which is far less than 100% in LENS process due to some objective reasons [24]. Effects of ultrasonic vibration on powder utilization efficiency during the five-layer thin walls fabrication under the two manufacturing conditions are shown in Figure 8.4. It is notable that the mean value of powder utilization efficiency increased from 20.7% to 23.2% when ultrasonic vibration was adopted in the LENS fabrication process. The results indicated that if keeping other LENS manufacturing parameters constant, processing with ultrasonic vibration led to a remarkably higher powder utilization efficiency than that without ultrasonic vibration. Such phenomenon may be ascribed to the extra input energy from ultrasonic vibration that improved the powder absorption rate during powder deposition. A larger powder utilization efficiency in UV-A LENS would promote the fabrication rate and economic benefit.

28 W W/O 26

24 Mean Median 22 99% 75%

20 25% 1% 18

Powder utilization efficiencyPowder (%)

16 Figure 8.4 Effects of ultrasonic vibration on powder utilization efficiency during 5- layer thin walls deposition

164 Texas Tech University, Fuda Ning, May 2018

8.3.2 Deposited shape geometry The capability of deposited shape formation was an important feature for LENS manufacturing and repairing of metallic parts. The effects of ultrasonic vibration on deposited shape geometry were evaluated. Comparisons of length, width, height, and flatness between UV-A LENS and LENS without ultrasonic vibration are shown in Figure 8.5. The length and width of the thin walls fabricated by both methods had no significant differences, but the thin walls fabricated with ultrasonic vibration had a higher mean value of build height and a remarkably smaller flatness. It can be seen that the mean height values of the samples fabricated by LENS with and without ultrasonic vibration are 1.35 mm and 1.26 mm, respectively. Thus, the related actual values of layer thickness are 0.27 mm and 0.25 mm. The actual layer thickness is dependent on how many powders are melted under each increment of Z axis by the input energy which is mainly determined by laser power, scanning speed, and powder feeding rate, leading to the difference between theoretical value and actual value of the layer thickness. In addition, during the actual LENS fabrication process, a portion of the previous layer will be re-melted under the laser radiation for the successive layer deposition, causing a smaller actual value of layer thickness than the theoretical one.

16.2 2.0 2.0 0.50 a b c d

15.9 1.6 1.6 0.45

15.6 1.2 1.2

0.40

15.3 0.8 0.8

Length (mm) Length

Width (mm)

Height (mm) Height Flatness (mm) 0.35 15.0 0.4 0.4

14.7 0.0 0.0 0.30 W W/O W W/O W W/O W W/O

Figure 8.5 Comparisons on (a) length, (b) width, (c) height, and (d) flatness of 5-layer thin walls between LENS with (W) and without (W/O) ultrasonic vibration

165 Texas Tech University, Fuda Ning, May 2018

The generation of the higher build height in UV-A LENS was caused due to the increased powder utilization efficiency as discussed above. The extra ultrasonic energy input enabled more feeding powders to be melted within each layer of the single wall structure. On the other hand, the ultrasonic vibration was imposed in the vertical Z direction rather than a horizontal X or Y direction, affecting little on the length and/or width. Additionally, the smaller flatness values in UV-A LENS were obtained due to the possible reason that the radiation pressure generated by ultrasonic in the acoustic field led to the deformation of gas-liquid interface that would generate a comparatively flat outline [25]. It can be concluded that ultrasonic vibration could increase the mean value of build height of the LENS-fabricated thin walls with significantly improved flatness.

8.3.3 Surface roughness Figure 8.6a shows a comparison of surface roughness value (Ra) between samples fabricated by UV-A LENS and LENS without ultrasonic vibration. Apart from five-layer thin wall samples, one-layer track samples were also deposited for top surface roughness measurement. This process could be considered as laser cladding, which is a surface modification and material repairing technology. It can be seen that after one layer deposition, there were no differences of surface roughness between both processes, indicating an unfulfilled influence of ultrasonic vibration on surface roughness after one-layer deposition. The mean Ra value of the five-layer thin walls fabricated by UV-A LENS was a little smaller than that without ultrasonic vibration.

Comparing the top surface morphologies in Figures 8.6b and 8.6c, it could be observed that fewer particles remained on the top surface of the five-layer walls fabricated by UV-A LENS. The balling effect associated with LENS process was mainly responsible for the undesired surface roughness, which was successfully alleviated by importing ultrasonic vibration into LENS process. During powder deposition, the molten pools generated by laser power would catch the powders those were melted and merged with the base materials. During the laser radiation, ultrasonic vibration energy in UV-A LENS could provide an additional heat source to the molten

166 Texas Tech University, Fuda Ning, May 2018 pool and enlarge the energy density. Marangoni convection would be thus increased leading to a higher surface disturbance in the molten pool. The higher surface disturbance could effectively decrease the mean value of surface roughness by preventing partially assimilated particles on the top surface of the sample [26]. Also, ultrasonic vibration provided intense stirring and mixing to the molten pool so that not fully melted particles might be shaken off the top surface.

20 WithW UV a b WithoutW/O UV

16 W 12 400 μm

8 c O

4 / W

Surface roughness Ra (µm) Ra roughness Surface

0 400 μm 1 layer 5 layers

Figure 8.6 Effects of ultrasonic vibration on surface roughness during single-layer and 5-layer thin walls deposition

8.3.4 Geometry of molten pool and dilution zone Effects of ultrasonic vibration on the geometry of molten pool and dilution zone during 5-layer thin walls deposition are illustrated in Figure 8.7. The measuring variables included molten pool width, dilution zone depth, and heat affected zone (HAZ) depth. It can be observed that compared with LENS without ultrasonic vibration, UV-A LENS generated larger molten pool width, dilution zone depth, and HAZ depth with a mean value increment of 0.12 mm, 0.11 mm, and 0.08 mm, respectively.

The reason is that UV-A LENS process generated a higher temperature in the molten pool and a higher cooling rate in the solidification process. The increase of

167 Texas Tech University, Fuda Ning, May 2018 cooling rate was due to the fact that cooling rate could be affected and determined by the total amount of the input energy [27]. As aforementioned, ultrasonic vibration enhanced Marangoni convection which was proportional to the square of the molten pool size and had a significant effect on the overall molten pool shape [28]. On the other hand, the energy density was varied between LENS (413 J/mm2) and UV-A LENS (484 J/mm2) processes. The acoustic streaming induced by ultrasonic vibration energy would result in a better mixing and agitation of elements within the molten pool, leading to the increase of molten pool size. The increase of geometrical dimensions in molten pool and dilution zone would improve the bonding strength between the deposited part and substrate.

1.5 0.5 1.0 a b c

1.2 0.4 0.8

0.9 0.3 0.6

a 0.6 0.2 0.4 Molten pool width (mm)

0.3 Dilution zone depth (mm) 0.1

b c 0.2 Heat affected zone depth (mm) 0.0 0.0 0.0 W W/O W W/O W W/O Figure 8.7 Effects of ultrasonic vibration on (a) molten pool width, (b) dilution zone depth, and (c) heat affected zone depth during 5-layer thin walls deposition

8.3.5 Cross-sectional morphologies and microstructures Figure 8.8 shows the cross-sectional morphologies and microstructures of the parts fabricated by LENS without and with ultrasonic vibration.

Figure 8.8a presents a typical arc outline of the single-bead wall samples fabricated by LENS without ultrasonic vibration. The width of the sample was not consistent and exhibited a bottom-up increase. The actual height of the sample was 2.56 mm and the cross-section area was 3.22 mm2. It can be observed from Figure 8.8b that the outline of the sample fabricated with ultrasonic vibration was

168 Texas Tech University, Fuda Ning, May 2018 comparatively flat and the width became uniform at different layers. In addition, the sample height increased to 2.64 mm that was almost identical with the designed height (2.66 mm). The cross-section area increased to 3.59 mm2, indicating higher powder utilization efficiency was obtained in UV-A LENS process. It can be concluded that ultrasonic vibration could positively affect sample morphologies with improved shape and dimensional accuracy.

As shown in Figure 8.8c, a large amount of pores were generated inside the samples fabricated by LENS without ultrasonic vibration due to the evolution of entrapped gas bubbles in the molten tracks. These pores were hardly expelled from the top surface prior to the metal solidification in LENS without ultrasonic vibration. It can be seen from Figure 8.8d that the pores were remarkably alleviated in the samples fabricated by LENS with ultrasonic vibration. The porosity value decreased from 0.68% to 0.35% when ultrasonic vibration was adopted during LENS process. The lower porosity was attributed to the acoustic streaming and cavitation actions of ultrasonic vibration in material solidification. The radiation pressure in the acoustic field changed the interface between the inner gas and metallic liquid and then break up the gas-evolved pores, leading to the decrease of porosity in the sample fabricated by UV-A LENS process. It can be concluded that utilizing ultrasonic vibration in LENS process could contribute to the fabrication of more dense parts.

As shown in Figure 8.8e, columnar grain structures with an average grain size of about 7–11 µm could be observed inside the parts fabricated by LENS without ultrasonic vibration. The columnar grains were found to grow along the bottom-up depositing direction, which was mainly caused by unidirectional heat conduction via the cooler substrate. Such heat transfer mode enabled grain structures to preferentially grow along the thermal gradient direction from the bottom of molten pool to the upper region. However, as shown in Figure 8.8f, parts fabricated with ultrasonic vibration generated equiaxed grain structures with smaller grain size (about 1.5–3 µm) and more uniform crystals. The refinement of the microstructures mainly resulted from the actions of acoustic streaming and cavitation induced by ultrasonic vibration. The

169 Texas Tech University, Fuda Ning, May 2018 pressure in the acoustic field made the solid-liquid interface morphologically unstable and fragmented the columnar grain tips. These dendrite tips were conveyed into the undercooled melt via convection. In the meantime, a large amount of grain nucleation was formed. Hence, the transition occurred from columnar structures to equiaxed structures, which would improve the homogeneity of the LENS-fabricated parts [29]. Therefore, the ultrasonic vibration was proved to be an effectively assisted approach to refine the grain size of stainless steel alloys.

A large number of micro-cracks could be observed at the bonding region of the samples fabricated without ultrasonic vibration (Figure 8.8g). The cracking was formed as a consequence of the residual stress generated by the drastic thermal gradient between the hot molten pool and the cold substrate. The formation of massive micro-cracks could lead to a weak bonding strength between the deposited part and the substrate. However, in UV-A LENS process, the fabricated parts exhibited remarkably fewer micro-cracks at the bonding zone, as illustrated in Figure 8.8h. The major reason was that the acoustic streaming and cavitation induced by ultrasonic vibration provided intense stirring and mixing in the molten pool. The molten powders and substrate material would be homogenized with less residual stress, thereby resulting in the fabrication of samples with fewer micro-cracks in UV-A LENS process.

a Without ultrasonic vibration b With ultrasonic vibration c d

Pores

2.56 mm 2.64 mm

5 μm 5 μm e Substrate Substrate f

g Micro-cracks h

2 μm 2 μm

2 μm 2 μm Figure 8.8 Cross-sectional morphologies and microstructures at different locations of AISI 630 samples parts fabricated by LENS without and with ultrasonic vibration

170 Texas Tech University, Fuda Ning, May 2018

8.3.6 Tensile properties The effects of ultrasonic vibration on tensile properties including tensile strength, yield strength, ductility, toughness, and Young’s modulus are shown in Figure 8.9. All the samples failed in the middle regions during tensile testing. It can be seen that the samples fabricated by UV-A LENS exhibited a remarkably higher tensile strength, yield strength, ductility, and toughness than those fabricated by LENS without ultrasonic vibration. The mean values of these properties increased by 144.2%, 117.7%, 108.5%, and 313.9%, respectively. The enhancement was obtained due to many possible reasons, including fewer pores and micro-cracks, finer grain size, better grains arrangement, and smaller residual stress induced by the ultrasonic vibration in LENS. The samples fabricated under both manufacturing conditions had similar Young’s modulus since this value was little affected by processing [30].

a 2500 Without UV With UV 2000

1500

1000

500

0 Tensile Yield Ductility Toughness Young’s strength strength (‱) (J*m-3*105) modulus (MPa) -1 (MPa) (10 GPa) 500 Figure 8.9 Tensileb properties of Without the AISI UV 630 samples fabricated by LENS without and 450 With UV

) with ultrasonic vibration (UV)

1 400

8.3.7 Microhardness350 65.02 μm 441 HV1 63.98 μm 456.8 HV1

The microhardness distribution along the depositing6 cross-section of the

. .

3 6 300 6 samples fabricated by LENS with and without ultrasonic20 μm vibration20 μm is shown in Figure

250 e c a

Microhardness (HV 8.10. It can be seen that using ultrasonicf vibration contributed to larger average r e 200 t n microhardness than without ultrasonic vibrationI at different positions of the deposited 150 Substrate Deposited part -1.2 -0.8 -0.4 0.01710.4 0.8 1.2 1.6 2.0 Distance from the interface (mm) Texas Tech University, Fuda Ning, May 2018 part from the interface. For instance, with ultrasonic vibration, the indentation area was smaller at 1.2 mm height of the deposited part from the interface, leading to a larger average microhardnessa 2500 of 456.8 HV1 than that of 441 HV1 in LENS without Without UV ultrasonic vibration. In addition, from the interface With to the UV substrate, microhardness 2000 sharply decreased to about 160 HV1 at -0.6 mm height underneath the substrate. 1500 The improvement of mechanical properties could be quantified by the classic

Hall-Petch equation 1000below:

1/ 2 500  i   0  Kd (2)   where, i is tensile0 or yield strength, 0 is a constant stress for steel material, Tensile Yield Ductility Toughness Young’s K is the Hall-Petch slope, andstre ndg tish sthetre nmeangth ( ‱grain) (sizeJ*m -3of*1 0steel5) mo materialdulus [9]. (MPa) -1 (MPa) (10 GPa) 500 b Without UV 450 With UV

)

1 400

350 65.02 μm 441 HV1 63.98 μm 456.8 HV1

. .

3 6 300 6 20 μm 20 μm

250 e

c a

Microhardness (HV

r e

200 t

n I

150 Substrate Deposited part -1.2 -0.8 -0.4 0.0 0.4 0.8 1.2 1.6 2.0 Distance from the interface (mm) Figure 8.10 Microhardness of the AISI 630 parts fabricated by LENS without and with ultrasonic vibration (UV)

The relation indicated that tensile or yield strength value positively correlated to the reciprocal root of the grain size and the strength value increased with the decrease of the grain size. In addition, the influence of grain size on strength was found to be identical to that on material hardness. The actions such as acoustic streaming, cavitation, stirring, and mixing induced by ultrasonic vibration facilitated

172 Texas Tech University, Fuda Ning, May 2018 the formation of crystals dispersion and nucleation, which contributed to the grain refinement. Consequently, the mechanical properties including tensile properties and microhardness of the parts were improved during UV-A LENS process.

8.4 Conclusions A fundamental investigation of the novel UV-A LENS process was conducted in this chapter. The objective is to evaluate the effects of ultrasonic vibration in the LENS manufacturing system on the geometrical and microstructural characteristics and further assess the material performance of LENS-fabricated parts. The conclusions have been drawn as follows:

(1) A significant increment of powder utilization efficiency would be obtained in UV-A LENS process. The extra energy imported by ultrasonic vibration was a possible reason for the improved powder absorption rate during material deposition. In addition, ultrasonic vibration could increase the build height of the LENS fabricated thin walls with smaller flatness. The higher powder utilization efficiency in UV-A LENS was responsible for the higher build height, and the radiation pressure generated by ultrasonic in the acoustic field facilitated the formation of the flat outline.

(2) Balling effect on the top surface was successfully alleviated by the intense stirring and mixing from ultrasonic vibration, leading to a smaller mean value of surface roughness in UV-A LENS process even though the data distribution was not significantly different from that in LENS process. Effect on surface roughness of bulk parts via a sufficient action of ultrasonic vibration on a larger area on the top will be investigated for the significance testing in the future work. As a consequence of the acoustic streaming induced by ultrasonic vibration in LENS process, geometrical dimensions in molten pool and dilution zone were increased, which could improve the bonding strength between the deposited part and substrate.

(3) Deposition defects including pores, cavities, and micro-cracks were eliminated or reduced in UV-A LENS process. The actions of acoustic streaming and

173 Texas Tech University, Fuda Ning, May 2018 cavitation induced by ultrasonic vibration could break up the gas-evolved pores and homogenize the molten pool with less residual stress. UV-A LENS process enabled the fabrication of samples with free of pores/cavities and fewer micro-cracks.

(4) Grain refinement was obtained in UV-A LENS due to the formation of a considerable amount of grain nucleation, thereby resulting in the improvement of tensile properties and microhardness of LENS-fabricated parts.

Reference [1] ASTM F2792-12a, 2012, Standard Terminology for Additive Manufacturing Technologies, ASTM International, West Conshohocken, PA, USA. [2] Chua, C.K., and Leong, K.F., 2015, 3D Printing and Additive Manufacturing Principles and Applications, World Scientific Publishing Co. Pte. Ltd., Singapore. [3] Thompson, S.M., Bian, L., Shamsaei, N., and Yadollahi, A., 2015, An overview of direct laser deposition for additive manufacturing; Part I: Transport phenomena, modeling and diagnostics, Additive Manufacturing, 8, pp. 36–62. [4] Das, S., Beama, J.J., Wohlert, M., and Bourell, D.L., 1998, Direct laser freeform fabrication of high performance metal components, Rapid Prototyping Journal, 4(3), pp. 112–117. [5] Frazier, W.E., 2014, Metal additive manufacturing: a review, Journal of Materials Engineering and Performance, 23(6), pp. 1917–1928. [6] Gu, D.D., Meiners, W., Wissenbach, K., and Poprawe, R., 2012, Laser additive manufacturing of metallic components: materials, processes and mechanisms, International materials reviews, 57(3), pp. 133–164. [7] Gasser, A., Backes, G., Kelbassa, I., Weisheit, A., and Wissenbach, K., 2010, Laser additive manufacturing, Laser Technik Journal, 7(2), pp. 58–63. [8] Tabernero, I., Lamikiz, A., Martínez, S., Ukar, E., and Figueras, J., 2011, Evaluation of the mechanical properties of Inconel 718 components built by laser cladding, International Journal of Machine Tools and Manufacture, 51(6), pp. 465–470. [9] Li, P., Yang, T.P., Li, S., Liu, D.S., Hu, Q.W., Xiong, W.H., and Zeng, X.Y., 2005, Direct laser fabrication of nickel alloy samples, International Journal of Machine Tools and Manufacture, 45(11), pp. 1288–1294. [10] Choi, J., and Chang, Y., 2005, Characteristics of laser aided direct metal/material deposition process for tool steel, International Journal of Machine Tools and Manufacture, 45(4), pp. 597–607. [11] España, F.A., Balla, V.K., Bose, S., and Bandyopadhyay, A., 2010, Design and fabrication of CoCrMo alloy based novel structures for load bearing implants using laser engineered net shaping, Materials Science and Engineering: C, 30(1), pp. 50–57.

174 Texas Tech University, Fuda Ning, May 2018

[12] Amano, R.S., and Rohatgi, P.K., 2011, Laser engineered net shaping process for SAE 4140 low alloy steel, Materials Science and Engineering: A, 528(22), pp. 6680–6693. [13] Durejko, T., Ziętala, M., Łazińska, M., Lipiński, S., Polkowski, W., Czujko, T., and Varin, R.A., 2016, Structure and properties of the Fe3Al-type intermetallic alloy fabricated by laser engineered net shaping (LENS), Materials Science and Engineering: A, 650, pp. 374–381. [14] Yu, J., Rombouts, M., and Maes, G., 2013, Cracking behavior and mechanical properties of austenitic stainless steel parts produced by laser metal deposition, Materials & Design, 45, pp. 228–235. [15] Liu, F., Lin, X., Yang, G., Song, M., Chen, J., and Huang, W., 2011, Microstructure and residual stress of laser rapid formed Inconel 718 nickel-base superalloy, Optics & Laser Technology, 43(1), pp. 208–213. [16] Lewis, G.K., and Schlienger, E., 2000, Practical considerations and capabilities for laser assisted direct metal deposition, Materials & Design, 21(4), pp. 417– 423. [17] Bi, G., Ng, G.K.L., Teh, K.M., and Jarfors, A.E., 2010, Feasibility study on the Laser Aided Additive Manufacturing of die inserts for liquid forging, Materials & Design, 31, pp. 112–116. [18] Krishna, B.V., Xue, W., Bose, S., and Bandyopadhyay, A., 2008, Engineered porous metals for implants, JOM, 60(5), pp. 45–48. [19] Bi, G., Sun, C.N., Chen, H.C., Ng, F.L., and Ma, C.C.K., 2014, Microstructure and tensile properties of superalloy IN100 fabricated by micro-laser aided additive manufacturing, Materials & Design, 60, pp. 401–408. [20] Su, X.B., Yang, Y.Q., Yu, P., and Sun, J.F., 2012, Development of porous medical implant scaffolds via laser additive manufacturing, Transactions of Nonferrous Metals Society of China, 22, pp. 181–187. [21] Shamsaei, N., Yadollahi, A., Bian, L., and Thompson, S.M., 2015, An overview of direct laser deposition for additive manufacturing; Part II: Mechanical behavior, process parameter optimization and control, Additive Manufacturing, 8, pp. 12–35. [22] Gu, D.D., and Shen, Y.F., 2009, Effects of processing parameters on consolidation and microstructure of W–Cu components by DMLS, Journal of Alloys and Compounds, 473, 1, pp. 107–115. [23] Liu, Z.C., Ning, F.D., Cong, W.L., Jiang, Q.H., Li, T., Zhang, H.C., and Zhou, Y.G., 2016, Energy consumption and saving analysis for laser engineered net shaping of metal powders, Energies, 9(10), pp. 763–774. [24] Zhang, K., Zhang, X.M., and Liu, W. J., 2012, Effects of processing parameters on powder utilization ratio during laser metal deposition shaping, Advanced Materials Research, 549, pp. 790–794. [25] Laborde, J.L., Hita, A., Caltagirone, J.P., and Gerard, A., 2000, Fluid dynamics phenomena induced by power ultrasounds, Ultrasonics, 38(1), pp. 297–300. [26] Shah, K., Pinkerton, A.J., Salman, A., and Li, L., 2010, Effects of melt pool

175 Texas Tech University, Fuda Ning, May 2018

variables and process parameters in laser direct metal deposition of aerospace alloys, Materials and Manufacturing Processes, 25(12), pp. 1372–1380. [27] Hofmeister, W., and Griffith, M., 2001, Solidification in direct metal deposition by LENS processing, JOM Journal of the Minerals, Metals and Materials Society, 53(9), pp. 30–34. [28] Drezet, J.M., Pellerin, S., Bezençon, C., and Mokadem, S., 2004, Modelling the Marangoni convection in laser heat treatment, Journal de Physique IV France, 120, pp. 299–306. [29] Kurz, W., Bezencon, C., and Gäumann, M., 2001, Columnar to equiaxed transition in solidification processing, Science and Technology of Advanced Materials, 2(1), pp. 185–191. [30] Black, J.T., and Kohser, R.A., 2012, Materials and Processes in Manufacturing, 11 ed., John Wiley and Sons Inc., Danvers, MA, USA.

176 Texas Tech University, Fuda Ning, May 2018

CHAPTER IX

MICROSTRUCTURES AND MECHANICAL PROPERTIES OF INCONEL 718 SUPERALLOY BULK PARTS FABRICATED BY ULTRASONIC VIBRATION-ASSISTED (UV-A) LASER ENGINEERED NET SHAPING

Paper title:

Ultrasonic vibration-assisted laser engineered net shaping of Inconel 718 parts: microstructural and mechanical characterization

Published in:

ASME Journal of Manufacturing Science and Engineering (2018), Vol. 140, No. 6, pp. 061012-1–061012-11.

Authors:

Fuda Ning1, Yingbin Hu1, Zhichao Liu1,2, Xinlin Wang1,2, Yuzhou Li3, and Weilong Cong1

Authors’ affiliations:

1Department of Industrial, Manufacturing, and Systems Engineering, Texas Tech University, Lubbock, TX 79409, USA.

2School of Mechanical Engineering, Dalian University of Technology, Dalian, Liaoning 116023, China.

3School of Electromechanical Engineering, Guangdong University of Technology, Guangzhou, Guangdong 510006, China.

177 Texas Tech University, Fuda Ning, May 2018

Abstract Laser engineered net shaping (LENS) has become a promising technology in direct manufacturing or repairing of high-performance metal parts. Investigations on LENS manufacturing of Inconel 718 (IN718) parts have been conducted for potential applications in the aircraft turbine component manufacturing or repairing. Fabrication defects, such as pores and heterogeneous microstructures, are inevitably induced in the parts, affecting part qualities and mechanical properties. Therefore, it is necessary to investigate a high-efficiency LENS process for the high-quality IN718 part fabrication. Ultrasonic vibration has been implemented into various melting material solidification processes for part performance improvements. However, there is a lack of studies on the utilization of ultrasonic vibration in LENS process for IN718 part manufacturing. In this chapter, ultrasonic vibration-assisted (UV-A) LENS process is thus proposed to fabricate IN718 parts for the potential reduction of fabrication defects. Experimental investigations are conducted to study the effects of ultrasonic vibration on microstructures and mechanical properties of LENS-fabricated parts under two levels of laser power. The results showed that ultrasonic vibration could reduce the mean porosity to 0.1%, refine the microstructure with an average grain size of 5 µm, and fragment the detrimental Laves precipitated phase into small particles in a uniform distribution, thus enhancing yield strength, ultimate tensile strength, microhardness, and wear resistance of the fabricated IN718 parts.

Keywords: ultrasonic vibration, laser engineered net shaping (LENS), Inconel 718 alloy, microstructures, mechanical properties

9.1 Introduction Inconel 718 (IN718) is a nickel-based superalloy with excellent combinations of high fatigue strength, outstanding corrosion resistance, and good oxidation resistance at elevated temperatures [1]. These superior properties make IN718 alloy an attractive candidate to fabricate turbine blades, rocket motors, and nuclear reactors those always serve in highly aggressive working environments [2]. Due to the excellent reliability in high-temperature applications, IN718 alloy has been in a

178 Texas Tech University, Fuda Ning, May 2018 remarkably increasing demand. However, conventional manufacturing of IN718 components especially with complex structures always results in a high cost and a long manufacturing cycle time. Therefore, it is desired to manufacture IN718 components in a shorter cycle time or re-manufacture worn parts to extend service life at a lower cost for meeting the high demand.

Laser engineered net shaping (LENS), one of the laser additive manufacturing techniques, has been applied as a competitive method for manufacturing and repairing functional and high added-value metal parts [3]. In LENS process, a designed three- dimensional (3D) near-net shape component can be directly built on the substrate track by track and layer by layer based on the computer-aided design (CAD) model. Due to advantages of feature adding and material saving, LENS has an excellent capability to produce complex structural components [4]. It has been reported by the U.S. National Institute of Standards and Technology (NIST) that LENS process exhibits significant cost savings and higher manufacturing efficiency than the traditional manufacturing of costly metal alloys used in the aerospace industry [5]. Such advantages are attributed to the fact that LENS process can fabricate or repair the product with fewer components either simultaneously or in the same location, whereas traditional manufacturing produces more intermediate components those will be assembled at different locations [5]. In recent years, LENS manufacturing or repairing of IN718 parts has attracted considerable interests in both industry and academia [2]. Many studies have been conducted to investigate the microstructures and mechanical properties of LENS-fabricated IN718 parts. Among the reported literature, effects of LENS processing parameters (including laser power [6], powder feeding rate [1], and laser scanning strategy [6-8]) on the workpiece performance were investigated. Several researchers also presented the comparisons on microstructures and mechanical properties between as-deposited IN718 parts and heat-treated IN718 parts fabricated by LENS [9-11]. However, fabrication defects such as porosity and heterogeneous microstructures have been inevitably induced in the LENS-fabricated IN718 parts, which are known to be detrimental to the mechanical properties [1]. It is crucial to

179 Texas Tech University, Fuda Ning, May 2018 investigate an improved LENS manufacturing technique to effectively reduce the porosity and homogenize the microstructures for obtaining enhanced mechanical properties of IN718 parts.

Ultrasonic vibration has been widely used as an assisted technique in melting material solidification processes such as casting, arc welding, etc. [12-15]. Since the vibration frequency is much higher than the natural frequency of a manufacturing system, ultrasonic vibration will not generate additional harmful low-frequency vibrations to the system. The direct implement of ultrasonic vibration in solidification and crystallization processes can induce nonlinear effects including cavitation, acoustic streaming, and acoustic radiation pressure [16]. Owing to these effects, ultrasonic vibration enables to reduce porosity, refine the microstructure, and increase the homogeneity of chemical contents in the fabricated parts. A fundamental investigation on ultrasonic vibration-assisted (UV-A) LENS manufacturing of 17-4 PH stainless steel parts was previously conducted by the authors [17, 18]. The results evidenced the porosity reduction and microstructure refinement induced by the ultrasonic vibration, leading to the enhanced tensile properties and microhardness. In addition, Wu et al. [19] applied ultrasonic vibration to refine microstructures of laser- cladded zirconia coatings and modify the dilution characteristics to enhance the coating bonding strength. However, there is a lack of studies on the utilization of ultrasonic vibration in LENS process for IN718 part manufacturing.

In the present investigation, IN718 parts were fabricated by LENS process without and with ultrasonic vibration under two levels of laser powers. The main objective of this work is to identify the sensitivity of microstructures to the introduction of ultrasonic vibration in LENS manufacturing system and further assess the variations of mechanical behaviors. For this purpose, microstructural features of the as-deposited IN718 parts, such as porosity, grain microstructure, and precipitated phase composition and morphology, were compared to evaluate the effects of ultrasonic vibration. The mechanical behaviors including tensile properties, microhardness, and wear resistance of the as-deposited IN718 were also studied.

180 Texas Tech University, Fuda Ning, May 2018

9.2 Experimental procedures

9.2.1 Materials Low carbon steel plates with dimensions of 100 mm × 50 mm × 6.4 mm were used as the substrate material. The plate surface was polished and then cleaned by acetone before the LENS fabrication was conducted. The spherical gas atomized Inconel 718 powders (Carpenter Powder Products Inc., Bridgeville, PA, USA) with a particle size range between 45 µm and 125 µm, as shown in Figure 9.1, were utilized for parts fabrication. The major chemical compositions of Inconel 718 powders were listed in Table 9.1.

Figure 9.1 Inconel 718 powder morphology

Table 9.1 The major chemical compositions of Inconel 718 powders

Element C Si Mo Nb Al Ti Co Cr Fe Ni Wt.% 0.031 0.05 3.05 5.04 0.47 0.9 0.09 19.28 18.576 52.3

9.2.2 Experimental set-up and parameters The LENS process was performed on a customized additive manufacturing machine (450XL, Optomec Inc., Albuquerque, NM, USA) combining with an ultrasonic vibration assisting system. Figure 9.2 is a schematic of UV-A LENS system set-up that mainly consisted of a 400W IPG fiber laser system, a four-jet coaxial

181 Texas Tech University, Fuda Ning, May 2018 powder and inert gas delivery system, a motion control system, an ultrasonic power supply, and an ultrasonic transducer assembly. The powders delivered by a flowing argon gas stream converged at the focal point of the laser beam. Both powders and laser beam were simultaneously ejected onto the substrate generating a small molten pool. The molten pool increased with the continuous powder deposition, and then rapidly solidified into a bump via heat dissipation as the axis moved to a new position. The movement trajectory complied with the computer model that directed the creation of one layer line by line. After one layer deposition, the jet elevated one layer thickness to maintain a constant focal point for subsequent layer deposition until the designed geometry was completed. In this study, the ultrasonic power supply provided the high frequency (41 kHz) electrical energy converting from the regular line electricity (110 V, 60 Hz). Thus, ultrasonic vibration was generated by a piezoelectric transducer and then transmitted onto the substrate by an acoustic transformer. The direction of ultrasonic vibration was perpendicular to the surface of the substrate, and the ultrasonic amplitude was around 5 µm under the ultrasonic power of 60 W in this work.

Block parts with dimensions of 8 mm × 8 mm × 4 layers were built by both UV-A LENS and LENS without ultrasonic vibration under low and high levels of laser powers. Such dimensions were selected for the purpose of providing the fundamental knowledge on UV-A LENS of IN718. A preliminary parameterization was studied to ensure the successful fabrication of IN718 block parts. The LENS manufacturing parameters for IN718 part fabrications are listed in Table 9.2. After fabrication, the actual average layer thickness was 0.65 mm or 0.78 mm at the laser power of 270 W or 350 W, respectively.

182 Texas Tech University, Fuda Ning, May 2018

Beam delivery Laser system system

Laser generator Power supply Laser beam Fan

Hooper Z motion

Delivery lines Powder

Cladding head

Feeding Jet Powder and motor inert gas flow Deposited material Molten pool Inert gas Control Vibration Ar direction system with Substrate integrated - Transformer Power computer + supply Ceramic vibrator Damper Powder & inert gas delivery system Y X-Y motion table X X Chamber (system)

Figure 9.2 Schematic of ultrasonic vibration-assisted LENS system set-up

Table 9.2 The LENS manufacturing parameters for IN718 bulk part fabrications

Parameter Value Unit Laser power 270/350 W Powder flow rate 2.63 g/min Increment of Z axis 0.43 mm (theoretical layer thickness) Argon gas flow rate 6 L/min Axis feedrate (contour) 635 mm/min Axis feedrate (infill) 508 mm/min Hatch space 0.3 mm Hatch angle (first layer) 45 ° Hatch angle interval 90 °

9.2.3 Characterizations of microstructure and mechanical properties After LENS fabrication, samples were firstly ground and polished on a grinder-polisher machine (MetaServ 250, Buehler, Lake Bluff, IL, USA). Then, they were cleaned with acetone in an ultrasonic cleaner to remove contaminations on the

183 Texas Tech University, Fuda Ning, May 2018 sample surface. To reveal the microstructure, the samples were further etched with Kalling’s reagent (ES Laboratory, Glendora, CA, USA) for 5 min. The etchant components were hydrochloric acid (50%), cupric chloride (2%), water (2%), and methanol (balance). A field emission scanning electron microscope (FE-SEM) (S4300, Hitachi Co., Tokyo, Japan), equipped with a backscattered electron (BSE) detector, was used to observe the cross-sectional morphologies and microstructures of the fabricated parts. Additionally, elemental characterizations of featured phases in the selected micro-regions were carried out by an energy dispersive X-ray spectroscope (EDS) detector in the SEM. The spectra were collected by the point analysis under the conditions of a 51.2 µs process time (amp time), a 15 kV accelerating voltage, and a 30% dead time.

The porosity was defined as the ratio of measured micropore areas to the whole selected area. To quantify the porosity of IN718 parts fabricated under different conditions, the SEM images of the cross-sectional surface were processed using an image processing software ImageJ, as shown in Figure 9.3. The “circularity” function in ImageJ could help to determine the types of pore by identifying if the pore shape was spherical or non-spherical. Under each fabrication condition, three block parts were sectioned for the porosity measurement and then the calculation was conducted on five different areas of the same cross section. Thus, there were a total amount of 15 areas for the porosity calculation. The mean value with one standard deviation was used to describe and compare porosity among different fabrication conditions.

The circular intercept method in ASTM E112-13 standard [20] was used to measure the grain size, which began with drawing circular lines randomly in the SEM images and measuring the length (L) of these lines. Then, the total times (N) of lines crossing grain boundaries were counted. The grain size could be thus calculated by L/N. The circular lines used in this method could automatically compensate for deviations from equiaxed grain shape. Figure 9.4 shows the measuring procedures of grain size. The SEM image was processed to a black-white mode for a clear recognition of the grain boundary. The red dots were the positions where the lines

184 Texas Tech University, Fuda Ning, May 2018 intersected grain boundaries. The average grain size was calculated by dividing the total length of three circular lines over the total number of red dots. In order to better quantify the variation of grain sizes, the measurements were conducted on SEM images of five different IN718 parts. The mean and one standard deviation were used to represent the obtained results.

(a) (b) 15

i t

c 6

d 3

u B 0 0 5 10 15 20 25 (μm) Figure 9.3 (a) SEM image of porosity of fabricated IN718 parts and (b) porosity analysis/calculation using image processing technique

10 μm

Figure 9.4 The intercept method for grain size measurement based on the ASTM E112-13 standard

ASTM E8 standard [21] was followed to fabricate the IN718 tensile specimens in LENS without and with ultrasonic vibration. The detailed dimensional sketch and fabricated tensile samples are illustrated in Figure 9.5. Three tensile specimens were prepared for each processing condition. The LENS manufacturing parameters for

185 Texas Tech University, Fuda Ning, May 2018 tensile specimen fabrications are listed in Table 9.3. The samples were then cut out of the substrate and were ground to obtain the final shapes. The tensile test was performed on a universal testing machine (AGS-X, Shimadzu Co., Kyoto, Japan) with a 50 kN capacity load cell to evaluate the tensile properties. The gauge length was 12.5 mm and testing speed was 1 mm/min. A data acquisition software (Trapezium, Shimadzu Co., Kyoto, Japan) was used to collect the values of both force (N) and displacement (mm).

Microhardness on transverse surfaces of the polished parts was measured using a Vickers microhardness tester (900–390A Phase II, Metal-Testers Inc., Nanuet, NY, USA). A normal force of 9.8 N and a duration of 15 s were applied in the test. Ten random positions were tested for each sample.

The dry sliding wear test was conducted at the room temperature to evaluate the wear resistance of IN718 parts. ZrO2 ceramic bearing balls with a diameter of 3 mm served as the counterface material. The test conditions included a sliding speed of 3 mm/s, a normal force of 3 N, a sliding distance of 5 mm, and a duration of 1 hour. The width of sliding track, measured by the SEM, was used to characterize the wear resistance performance.

(a) 15.88 15.88 15.88 2.54

4.76 3.18

R=6.35 Unit: mm 51.82 (b)

Figure 9.5 Designed tensile specimen based on ASTM E8 standard ((a) dimensions of tensile specimen and (b) tensile samples fabricated by LENS)

186 Texas Tech University, Fuda Ning, May 2018

Table 9.3 The LENS manufacturing parameters for IN718 tensile specimen fabrications

Parameter Value Unit Laser power 270/350 W Powder flow rate 1.58 g/min Increment of Z axis 0.43 mm Argon gas flow rate 6 L/min Axis feedrate (contour) 762 mm/min Axis feedrate (infill) 508 mm/min Number of layers 3 Hatch space 0.3 mm Hatch angle (first layer) 60 ° Hatch angle interval 60 °

9.3 Results and discussion

9.3.1 Porosity Figure 9.6 shows the processed SEM images of LENS-fabricated IN718 alloy and porosity value under different fabrication conditions. The micropores generated in LENS process could be identified from two primary sources: gas entrapping and lack of fusion. The micropores caused by gas entrapping normally appeared as circular shapes, while those produced due to lack of fusion usually exhibited irregular shapes at the interfaces of adjacent particles. Micropores generated by both gas entrapping and lack of fusion were observed in the parts fabricated by LENS without ultrasonic vibration at the laser power of 270 W, as shown in Figure 9.6(a). It can be seen from Figure 9.6(b) that the parts possessed a porosity value of 0.88%. With the introduction of ultrasonic vibration at 270 W, gas-entrapped micropores were remarkably alleviated, and porosity value was reduced to 0.31%. The decrease of porosity was attributed to the effects of acoustic streaming and cavitation induced by ultrasonic vibration. These two major direct actions could affect the mass and heat transfer during the melting material solidification. Thus, indirect actions such as mixing and stirring were caused increasing the fluidity of melting materials in the molten pool. The entrapped gas was apt to aggregate and float upward under the drastic melting

187 Texas Tech University, Fuda Ning, May 2018 material movement, facilitating the escape of micropores from the molten pool prior to the melting material solidification. The similar phenomenon of porosity reduction influenced by ultrasonic vibration could also be found in UV-A casting of aluminum alloy [22].

Without UV With UV (a) A0 A1 Lack of fusion

W Lack of fusion

0 7 2 Gas entrapped

B0 B1

Gas W

entrapped

5 3

2 μm

(b) 1.2

1.0 0.88

0.8

0.6

Porosity (%) 0.4 0.31

0.2 0.2 0.09

0.0 A0 A1 B0 B1 Processing conditions Figure 9.6 (a) Processed SEM images for calculating porosity and (b) porosity values under different fabrication conditions

With the increase of laser power to 350 W, lack-of-fusion micropores were hardly observed due to the higher heat input. However, gas-entrapped micropores remained inside the parts after merely increasing the laser power to 350 W. This

188 Texas Tech University, Fuda Ning, May 2018 phenomenon was attributed to the fact that the argon gas entrapped in the molten pool did not have sufficient time to escape from the top of the molten pool during the highly transient solidification process. Besides, the increased laser energy could hardly dissolve the entrapped gas due to the low solubility of argon in metals [23]. It can be found that the fabricated part exhibited significantly reduced lack-of-fusion pores and gas-entrapped pores when the molten pool was imposed with ultrasonic vibration at 350 W, which was evidenced by the porosity value of only 0.09%.

9.3.2 Grain microstructure Due to the rapid solidification characteristic associated with LENS process, the obtained microstructures would be different from those in the cast or wrought parts. The microstructural grains of IN718 alloy produced by LENS without and with ultrasonic vibration at laser powers of both 270 W and 350 W are shown in Figure 9.7. The microstructural grain size was represented by mean with one-time standard deviation. It is notable from Figure 9.7(a-A0) that the coarse elongated grains were produced at the laser power of 270 W without ultrasonic vibration. The grain size achieved under this condition was 14.31 ± 1.96 µm, as illustrated in Figure 9.7(b). The elongated grain growth occurred along with the build direction due to the locally unidirectional solidification behavior. In the presence of ultrasonic vibration in LENS, the grains experienced the columnar-to-equiaxed transition (CET), as shown in Figure 9.7(a-A1), and the grain size was significantly decreased to 5.96 ± 1.52 µm. At the higher laser power level of 350 W, with the increase of laser power to 350 W, grain size (6.03 ± 2.57 µm) was smaller than that produced at the lower laser power level of 270 W. However, microstructural features in Figure 9.7(a-B0) indicated that the grains were not uniformly distributed, which could be verified by the relatively larger standard deviation of 2.57 µm. In contrast, the grains were further refined to 4.53 ± 1.04 µm with a more uniform distribution in UV-A LENS process, as shown in Figure 9.7(a-B1).

189 Texas Tech University, Fuda Ning, May 2018

(a) Without UV With UV A0 A1 W

0 7 2

B0 B1 n o W i

t 0 c 5 e 3 r i d

d l i 10 μm u B (b) 20

14.31±1.96 16 ) m µ

( 12

e z i s

6.03±2.57 n i 5.96±1.52

a 8 r

G 4.53±1.04

0 A0 A1 B0 B1 Processing conditions Figure 9.7 (a) Microstructural characteristics and (b) grain size evolution under different processing conditions

The grain refinement phenomenon induced by ultrasonic vibration could be explained by two prevailing ultrasonic refinement mechanism: cavitation-induced dendrite fragmentation and cavitation-enhanced nucleation [24, 25]. When ultrasonic vibration propagated in the melting materials, cavitation bubbles could be instantaneously generated. Once the bubbles collapsed, high-intensity shocks were emitted along the wave propagation, and the powerful pressure in the acoustic field could fragment the elongated dendrite tips. Meanwhile, these dendrite tips were conveyed into the undercooled melt via convection to form a large number of grain

190 Texas Tech University, Fuda Ning, May 2018 nuclei. With these ultrasonic cavitation effects, the CET occurred in the melting materials adjacent to the elongated dendritic front [26], leading to the remarkable reduction of microstructural grain size. With the increase of laser power, the higher applied heat could lead to a larger molten pool size, resulting in a lower temperature gradient in LENS process. It was reported that an increase in the laser power lowered the average ratio of temperature gradient to solidification velocity, thereby promoting the growth of equiaxed grains [26]. The transition from columnar to equiaxed grains would be beneficial for the grain size reduction. In addition to grain refinement, uniform distribution of grains was achieved due to the homogenized temperature field induced by ultrasonic vibration in the localized regions.

9.3.3 Phase composition Figures 9.8 and 9.9 show the EDS analysis results of IN718 parts fabricated by LENS without and with ultrasonic vibration at laser powers of 270 W and 350 W, respectively. At each fabrication condition, it can be clearly observed that point 1 (P1) in the dark matrix area was rich in Ni, Cr, and Fe, while point 2 (P2) on the white precipitated phase was rich in Ni, Cr, Nb, and Mo those were the major compositional elements of the Laves phase [10]. Similar identifications of Laves phase in the LENS- fabricated IN718 parts have also been discussed in other reported investigations [6, 11]. The Laves phase was composed of a Nb-rich brittle intermetallic compound with a typical composition of (Ni, Cr, Fe)2(Nb, Mo, Ti) [6]. The Nb segregation in IN718 alloy resulted in the formation of irregularly shaped Laves phase that was considered to be detrimental to mechanical properties. Such phase could facilitate the crack growth, decreasing the strength and plasticity of the IN718 alloy.

It was reported that Nb element played a crucial role in IN718 alloy with its impact on the solidification sequence and the strengthening phase precipitation [27]. The solidification of IN718 in LENS started with a primary liquid γ reaction and proceeded to cause an enrichment of interdendritic liquid in Nb, Mo, Ti, etc. until a eutectic type reaction occurred to terminate the solidification process [28]. Due to the poor diffusivity of large Nb atoms, Laves phase was difficult to be dissolved in the

191 Texas Tech University, Fuda Ning, May 2018 matrix of LENS-fabricated parts. However, with the actions of ultrasonic vibration, the long-bar Laves phase morphology could be fragmented into small-particle shape via high-intensity shock wave. The phenomena are shown in SEM images of Figures 9.8 and 9.9. The molten pool temperature could be increased due to the input of ultrasonic energy, thereby leading to a higher solubility of Nb in the matrix phase. The results indicated that ultrasonic vibration was favorable in partially dissolving Laves phase in the matrix. A homogeneous elemental composition could also be achieved due to the small-particle shaped Laves phase that was more uniformly distributed in the matrix. The morphological changes of Laves phase under ultrasonic vibration will be detailed in Section 9.3.4.

Figure 9.8 EDS analysis results of IN718 parts fabricated by LENS (a) without ultrasonic vibration and (b) with ultrasonic vibration at the laser power of 270 W

192 Texas Tech University, Fuda Ning, May 2018

Figure 9.9 EDS analysis results of IN718 parts fabricated by LENS (a) without ultrasonic vibration and (b) with ultrasonic vibration at the laser power of 350 W

9.3.4 Precipitated phase morphology Figures 9.10 and 9.11 show the microstructural interdendritic eutectics of the matrix and Laves phases at the top, middle, and bottom sections of IN718 parts fabricated by LENS without and with ultrasonic vibration at different levels of laser power.

It can be observed from Figure 9.10(a) that the Laves phase appeared in a long bar shape at different sections of IN718 parts fabricated by LENS without ultrasonic vibration, and it exhibited different growing directions at middle and bottom sections due to different directions of temperature gradient during deposition. With ultrasonic

193 Texas Tech University, Fuda Ning, May 2018 vibration, the long bar-shaped Laves phase was fragmented into small sphere-shaped and short bar-shaped particles at the bottom and middle sections of the sample, respectively, as shown in Figure 9.10(b). The phase fragmentation was mainly caused by the high-intensity shock in the acoustic field induced by ultrasonic vibration. Also, the Laves phase morphology exhibited a more uniform distribution and a relatively unidirectional growing direction compared with that in LENS without ultrasonic vibration. It can be inferred that ultrasonic vibration promoted the fabrication of IN718 parts with a homogeneous phase distribution. Due to the attenuation of ultrasonic energy with the increase of build height, it was hard to see considerable changes of Laves phase morphology at the top section of the sample.

Figure 9.10 The phase morphology at different sections of IN 718 parts fabricated by LENS (a) without ultrasonic vibration and (b) with ultrasonic vibration at laser power of 270 W

194 Texas Tech University, Fuda Ning, May 2018

With the further increase of laser power to 350 W, it can be observed from Figure 9.11(a) that long bar-shaped Laves phases still existed at all sections of the sample. By contrast, the implementation of ultrasonic vibration in LENS enabled to alleviate the precipitated Laves phase at all sections of the sample, as shown in Figure 9.11(b). Small particle-shaped Laves phase was formed with a uniform distribution in the matrix, which would benefit the fabrication of IN718 parts with reduced stress concentration and homogeneous elemental composition.

Figure 9.11 The phase morphology at different sections of IN718 parts fabricated by LENS (a) without ultrasonic vibration and (b) with ultrasonic vibration at laser power of 350 W

195 Texas Tech University, Fuda Ning, May 2018

9.3.5 Tensile properties Figure 9.12 illustrates the stress-strain curve, yield strength, ultimate tensile strength (UTS), and ductility of IN718 parts fabricated under different processing conditions. The values of these tensile properties were represented by mean with one standard deviation. At each level of laser power, UV-A LENS process caused a considerable improvement in yield strength and UTS but had no significant effects on ductility compared with LENS process without ultrasonic vibration. When laser power was increased from 270 W to 350 W, a remarkable increase of yield strength and UTS was also achieved in either UV-A LENS or LENS without ultrasonic vibration. It also can be found that there were no significant differences in ductility between the parts fabricated at low and high levels of laser power.

The improvement of yield strength and UTS can be explained by the grain refinement induced by ultrasonic vibration or higher laser energy input. Effects of grain size on yield strength and UTS are illustrated in Figure 9.13. It can be seen that both yield strength and UTS increased with the grain size decreased. Such trend was consistent with the positive correlation between yield strength/UTS and the reciprocal root of grain size in the classic Hall-Petch equation. Furthermore, ultrasonic vibration alleviated the porosity, fragmented the Laves phase, and facilitated the uniform distribution of elements, which also contributed to the enhancement of yield strength and UTS.

196 Texas Tech University, Fuda Ning, May 2018

(a) 1200 520 (A0) 270 W without UV (b) 270 W 350 W 482 (A1) 270 W with UV 1000 (B0) 350 W without UV (B1) 350 W with UV 480

434 800 440 425

600 400

A0 361

Stress (MPa) B1 B0

400 A1 360 Yield strength (MPa) 200 320

0 280 0 10 20 30 40 50 60 W/O W W/O W Processing conditions Strain (%) (c) 960 70 270 W 350 W (d) 270 W 350 W 884 880 60 787 800 768 735 50 40.5 41.8 36.8 720 40.1 40 640 30 560

1000 Ductility (%) 20 480

400 900 10

Ultimate tensile strength (MPa)

320 800 0 W/O W W/O W W/O W W/O W Processing conditions Processing conditions

700

Figure 9.12 Comparisons on (a) stress-strain curve, (b) yield strength, (c) ultimate 600 tensile strength (UTS), and (d) ductility of LENS-fabricated IN718 parts under 500 different processing conditions 400

1000 300 Yield strength 900 UTS

800

700

600 UTS (MPa)

Strength (MPa) Strength 500

400

300 8 9 10 11 12 13 14 15 UTS -1/2 Grain size (mm ) Figure 9.13 Effects of grain size on yield strength and UTS

197 Texas Tech University, Fuda Ning, May 2018

9.3.6 Microhardness Effects of ultrasonic vibration and laser energy on microhardness are illustrated in Figure 9.14. The boxplot is represented by mean, median, and one standard deviation to illustrate the values and statistical distribution of ten microhardness measurements. It can be seen that ultrasonic vibration resulted in a significant increase of microhardness from the mean value of 233 HV1 to that of 326

HV1 with an increment by 41% at laser power of 270 W. Increasing laser power to a higher level of 350 W could also lead to a significant increase of mean value to 306

HV1. Moreover, the utilization of ultrasonic vibration at the laser power of 350 W further increased the microhardness to a mean value of 352 HV1. It is known that the microhardness would be improved with the increase of UTS. Such relationship between microhardness and UTS indicated that the increase of microhardness was also attributed to the grain refinement induced by ultrasonic vibration and high laser energy input. The decrease of porosity aforementioned would be beneficial for a better fabrication of more uniformly dense parts, improving the part quality with the enhanced microhardness.

400 270 W 350 W

)

1 360

1 SD 320 Mean Median

280 -1 SD

Microhardness (HV 240

200 W/O W W/O W

Figure 9.14 Vickers microhardness measurement on the transverse surface of the fabricated parts

198 Texas Tech University, Fuda Ning, May 2018

9.3.7 Dry sliding wear resistance Figure 9.15 shows a comparison on the width of IN718 parts after dry sliding test under different fabrication conditions. The red dash line noted in the SEM images was employed to represent the width of each sliding track generated by the ZrO2 ceramic bearing ball. It can be seen that the width of the sample showed a significant decrease from 0.769 mm without ultrasonic vibration (UV) to 0.654 mm with UV at laser power of 270 W. At laser power of 350 W, ultrasonic vibration also led to the reduction but a smaller difference of the width from 0.596 mm to 0.548 mm. The decrease of width indicated that the wear resistance performance of IN718 parts could be enhanced in UV-A LENS process. As discussed above, compared with LENS without ultrasonic vibration, UV-A LENS produced finer grains at laser power of 270 W and led to a more uniform distribution of grains at laser power of 350 W. Thus, it can be inferred that the improvement of wear resistance of IN718 parts in UV-A LENS process would be more associated with the grain refinement. In addition, when laser power increased from 270 W to 350 W, wear resistance was improved in the parts fabricated either without UV or with UV. However, LENS process without UV led to a larger decrease in width from 0.769 mm to 0.596 mm in contrast to UV-A LENS, which could also be ascribed to the significant reduction of grain size that occurred at the laser power of 350 W without UV.

199 Texas Tech University, Fuda Ning, May 2018

Figure 9.15 A comparison on the width of IN718 parts after dry sliding under different fabrication conditions

9.4 Conclusions In this work, IN718 parts were fabricated by LENS process without and with ultrasonic vibration at two different laser powers. Porosity, microstructural grain size, phase composition, and interdendritic eutectic phase morphology, were compared to evaluate the effects of ultrasonic vibration. The mechanical behaviors including tensile properties (yield strength, UTS, and ductility), microhardness, and wear resistance of the as-deposited IN718 were also studied. The specific conclusions have been drawn as follows.

(1). Ultrasonic vibration led to a remarkable reduction of gas-entrapped pores in LENS-fabricated IN718 parts. Acoustic streaming and cavitation effects caused mixing and stirring actions that increased the fluidity, facilitating the escape of micropores from the molten pool. With the increase of laser power to a high level of 350 W, lack-of-fusion micropores were reduced as a result of higher heat input.

200 Texas Tech University, Fuda Ning, May 2018

(2). Finer grains with a uniform distribution were achieved in UV-A LENS of IN718 parts. Such phenomenon induced by ultrasonic vibration could be explained by cavitation-induced dendrite fragmentation and cavitation-enhanced nucleation.

(3). Ultrasonic vibration was favorable for dissolving Laves phase with an increased Nb content in the matrix phase. Long bar-shaped Laves phase morphology was fragmented into small sphere-shaped particles in a homogeneous distribution after the introduction of ultrasonic vibration especially at the higher level of laser power. This was mainly caused by the high-intensity pressure in the acoustic field induced by ultrasonic vibration.

(4). Due to the porosity reduction, grain refinement, Laves phase fragmentation, and uniform elemental distribution, the yield strength, UTS, microhardness, and wear resistance of IN718 parts fabricated by UV-A LENS were considerably increased.

References [1] Zhong, C.L., Gasser, A., Kittel, J., Wissenbach, K., and Poprawe, R., 2016, Improvement of material performance of Inconel 718 formed by high deposition- rate laser metal deposition, Materials & Design, 98, pp. 128–134. [2] Jia, Q.B., and Gu, D.D., 2014, Selective laser melting additive manufacturing of Inconel 718 superalloy parts: densification, microstructure and properties, Journal of Alloys and Compounds, 585, pp. 713–721. [3] Irwin, J., Reutzel, E.W., Michaleris, P., Keist, J., and Nassar, A.R., 2016, Predicting microstructure from thermal history during additive manufacturing for Ti-6Al-4V, ASME Journal of Manufacturing Science and Engineering, 138(11), pp. 111007–111017. [4] Gu, D.D., Meiners, W., Wissenbach, K., and Poprawe, R., 2012, Laser additive manufacturing of metallic components: materials, processes and mechanisms, International Materials Reviews, 57(3), pp. 133–164. [5] Thomas, D. S., and Gilbert, S. W., 2014, Costs and cost effectiveness of additive manufacturing: a literature review and discussion, National Institute of Standards and Technology (NIST) Special Publication, 1176, pp. 1–77. [6] Parimi, L.L., Ravi, G.A., Clark, D., and Attallah, M.M., 2014, Microstructural and texture development in direct laser fabricated IN718, Materials Characterization, 89, pp. 102–111. [7] Tabernero, I., Lamikiz, A., Martínez, S., Ukar, E., and Figueras, J., 2011, Evaluation of the mechanical properties of Inconel 718 components built by laser

201 Texas Tech University, Fuda Ning, May 2018

cladding, International Journal of Machine Tools and Manufacture, 51(6), pp. 465–470. [8] Liu, F.C., Lin, X., Yang, G.L., Song, M.H., Chen, J., and Huang, W.D., 2011, Microstructure and residual stress of laser rapid formed Inconel 718 nickel-base superalloy, Optics & Laser Technology, 43(1), pp. 208–213. [9] Zhao, X., Chen, J., Lin, X., and Huang, W., 2008, Study on microstructure and mechanical properties of laser rapid forming Inconel 718, Material Science and Engineering A, 478(1), pp. 119–124. [10] Qi, H., Azer, M., and Ritter, A., 2009, Studies of standard heat treatment effects on microstructure and mechanical properties of laser net shape manufactured Inconel 718, Metallurgical and Materials Transactions A, 40(10), pp. 2410–2422. [11] Lambarri, J., Leunda, J., Navas, V.G., Soriano, C., and Sanz, C., 2013, Microstructural and tensile characterization of Inconel 718 laser coatings for aeronautic components, Optics and Lasers in Engineering, 51(7), pp. 813–821. [12] Yang, Y., and Li, X.C., 2007, Ultrasonic cavitation-based nanomanufacturing of bulk aluminum matrix nanocomposites, ASME Journal of Manufacturing Science and Engineering, 129(2), pp. 252–255. [13] Cao, G.P., Konishi, H., and Li, X.C., 2008, Mechanical properties and microstructure of Mg/SiC nanocomposites fabricated by ultrasonic cavitation based nanomanufacturing, ASME Journal of Manufacturing Science and Engineering, 130(3), pp. 031105–031110. [14] Sun, Q.J., Lin, S.B., Yang, C.L., and Zhao, G.Q., 2013, Penetration increase of AISI 304 using ultrasonic assisted tungsten inert gas welding, Science and Technology of Welding and Joining, 14(8), pp. 765–767. [15] Watanabe, T., Shiroki, M., Yanagisawa, A. and Sasaki, T., 2010, Improvement of mechanical properties of ferritic stainless steel weld metal by ultrasonic vibration, Journal of Materials Processing Technology, 210(12), pp. 1646–1651. [16] Komarov, S.V., Kuwabara, M., and Abramov, O.V., 2005, High power ultrasonics in pyrometallurgy: current status and recent development, ISIJ International, 45(12), pp. 1765–1782. [17] Ning, F.D., and Cong, W.L., 2016, Microstructures and mechanical properties of Fe-Cr stainless steel parts fabricated by ultrasonic vibration-assisted laser engineered net shaping process, Materials Letters, 179, pp. 61–64. [18] Cong, W.L., and Ning, F.D., 2017, A fundamental investigation on ultrasonic vibration-assisted laser engineered net shaping of stainless steel, International Journal of Machine Tools and Manufacture, 121, pp. 61–69. [19] Wu, D.J., Guo, M.H., Ma, G.Y., and Niu, F.Y., 2015, Dilution characteristics of ultrasonic assisted laser clad yttria-stabilized zirconia coating, Materials Letters, 141, pp. 207–209. [20] ASTM E112-13, Standard Test Methods for Determining Average Grain Size, ASTM International, West Conshohocken, PA, 2013. [21] ASTM E8/E8M-09, Standard Test Methods for Tension Testing of Metallic Materials, ASTM International, West Conshohocken, PA, 2009.

202 Texas Tech University, Fuda Ning, May 2018

[22] Xu, H., Jian, X., Meek, T.T., and Han, Q., 2004, Degassing of molten aluminum A356 alloy using ultrasonic vibration, Materials Letters, 58(29), pp. 3669–3673. [23] Shao, S., Mahtabi, M. J., Shamsaei, N., and Thompson, S. M., 2017, Solubility of argon in laser additive manufactured α-titanium under hot isostatic pressing condition, Computational Materials Science, 131, pp. 209–219. [24] Wang, F., Eskin, D., Mi, J., Connolley, T., Lindsay, J., and Mounib, M., 2016, A refining mechanism of primary Al3Ti intermetallic particles by ultrasonic treatment in the liquid state, Acta Materialia, 116, pp. 354–363. [25] Chen, R.R., Zheng, D.S., Ma, T.F., Ding, H.S., Su, Y.Q., Guo, J.J., and Fu, H.Z., 2017, Effects of ultrasonic vibration on the microstructure and mechanical properties of high alloying TiAl, Scientific Reports, 7, pp. 41463–41477. [26] Gäumann, M., Bezencon, C., Canalis, P., and Kurz, W., 2001, Single-crystal laser deposition of superalloys: processing-microstructure maps, Acta Materialia, 49(6), pp. 1051–1062. [27] Liu, F., Lin, X., Leng, H., Cao, J., Liu, Q., Huang, C., and Huang, W., 2013, Microstructural changes in a laser solid forming Inconel 718 superalloy thin wall in the deposition direction, Optics & Laser Technology, 45, pp. 330–335. [28] Wang, Z., Guan, K., Gao, M., Li, X., Chen, X., and Zeng, X., 2012, The microstructure and mechanical properties of deposited-IN718 by selective laser melting, Journal of Alloys and Compounds, 513, pp. 518–523.

203 Texas Tech University, Fuda Ning, May 2018

CHAPTER X

ULTRASONIC VIBRATION-ASSISTED LASER ENGINEERED NET SHAPING OF INCONEL 718 SUPERALLOY BULK PARTS: EFFECTS OF ULTRASONIC FREQUENCY

Paper title:

Ultrasonic vibration-assisted laser engineered net shaping of Inconel 718 superalloy bulk parts: effects of ultrasonic frequency

To be submitted to:

ASME Journal of Manufacturing Science and Engineering

Authors:

Fuda Ning, Zhichao Liu, and Weilong Cong

Authors’ affiliations:

Department of Industrial, Manufacturing, and Systems Engineering, Texas Tech University, Lubbock, TX 79409, USA.

204 Texas Tech University, Fuda Ning, May 2018

Abstract Ultrasonic vibration-assisted (UV-A) laser engineered net shaping (LENS) of Inconel 718 parts have been proven to be a promising technology to improve the part quality and process performance. Ultrasonic vibration frequency, a very important variable of the ultrasonic vibration, can affect the ultrasonic vibration performance in assisting LENS processes. The applications of various ultrasonic frequencies attract less attention and the effects of ultrasonic frequency have been overlooked. In this chapter, Inconel 718 parts will be built using LENS process without and with ultrasonic vibration under three levels of ultrasonic frequency including 25 kHz, 33 kHz, and 41kHz. The effects of ultrasonic frequency on molten pool geometry, temperature profiles within the molten pool, peak temperature value fluctuation, porosity, and grain microstructures will be explored. This investigation will generate a better understanding of the link among ultrasonic frequency, temperature, molten pool, and microstructure in the UV-A LENS of Inconel 718 parts.

Keywords: ultrasonic vibration, laser engineered net shaping (LENS), Inconel 718 alloy, ultrasonic frequency, molten pool, temperature, porosity, microstructures

10.1 Introduction Laser engineered net shaping (LENS) is a laser additive manufacturing process that has been dramatically growing as a competitive method to manufacture and repair functional and high added-value metal parts [1]. The LENS process involves a track- by-track and layer-by-layer fabrication mode to build a three-dimensional (3D) part on a substrate based on the computer-aided design (CAD) model. Such process enables the production of complex structural components with benefits of feature adding, material and cost saving, and higher manufacturing efficiency [2, 3]. Recently, LENS of nickel-based alloys has attracted great attention for the fabrication of high- performance components utilized in the aerospace industry [4, 5]. Among the nickel- based alloys, Inconel 718 (IN718) is a widely used type due to the superior material properties that make it an attractive candidate to produce aerospace key parts such as turbine blades, rocket motors, nuclear reactors, etc. [4].

205 Texas Tech University, Fuda Ning, May 2018

Great efforts have been made to investigate the microstructures and mechanical properties of LENS-fabricated IN718 parts. Influences of process parameters including laser power, powder feeding rate, laser scanning speed, and laser scanning strategy on the part geometry/shape as well as part microstructural and mechanical performance were studied [6-10]. Characterizations were also conducted between as-deposited IN718 parts and heat-treated IN718 parts fabricated by LENS [11-13]. However, LENS-fabricated IN718 parts always exhibit inevitable fabrication defects, such as heterogeneous thermal distribution, large porosity, anisotropic microstructures, etc., which would generate adverse effects on the mechanical properties of the part [6]. It has been found that these defects are hard to be significantly alleviated simply by performing parametric studies and post heat treatment on the LENS-fabricated IN718 parts. Therefore, an effective improvement of LENS process to reduce the fabrication defects and enhance the performance of IN718 parts is thus of great necessity.

In order to diminish the fabrication defects, assisted techniques including ultrasonic vibration, electromagnetic field, and mechanical stirring have been employed in the thermal manufacturing processes such as casting, arc welding, cladding, etc., in which melting material solidification and recrystallization will always occur [14-19]. Among these assisted techniques, ultrasonic vibration has gained more attention due to the benefits including effective transmission of acoustic energy and relatively low equipment cost [20]. The direct introduction of ultrasonic vibration can induce two main nonlinear effects including cavitation and acoustic streaming [20]. These effects will cause several positive phenomena in the fabricated parts, such as porosity reduction, grain refinement, uniform distribution of the temperature field and chemical contents, etc. [21]. The authors have successfully implemented ultrasonic vibration in LENS process to fabricate IN718 parts [22]. It has been found that ultrasonic vibration could reduce the porosity, refine the microstructure, and fragment the detrimental Laves precipitated phase, thus enhancing yield strength, ultimate tensile strength, microhardness, and wear resistance of the

206 Texas Tech University, Fuda Ning, May 2018 fabricated IN718 parts [22]. Such an investigation can reveal the actions and influences of ultrasonic vibration in improving the LENS process, however, it is still unknown about the effects of ultrasonic vibration variables on the performance of LENS-fabricated IN718 parts.

Ultrasonic frequency is a very important variable of the ultrasonic vibration, which will affect the ultrasonic vibration performance in assisting thermal manufacturing processes. Currently, the available literature on the effects of ultrasonic frequency on the performance of thermal manufacturing are very scarce, which may be associated with the difficulties of ultrasonic frequency adjustment especially in the LENS process. The conditions of applying different ultrasonic frequencies attract less attention and the importance of ultrasonic frequency effects have been thereby overlooked [23]. In this study, IN718 parts will be built using LENS process without and with ultrasonic vibration under three levels of ultrasonic frequency. The effects of ultrasonic frequency on molten pool geometry, temperature profiles within the molten pool, peak temperature value fluctuation, porosity, and grain microstructures will be explored. This investigation will generate a better understanding of the link among ultrasonic frequency, temperature, molten pool, and microstructure in the ultrasonic vibration-assisted (UV-A) LENS of IN718 parts.

10.2 Experimental procedures

10.2.1 Materials The substrate utilized in this work was a low carbon steel plate with dimensions of 100 mm × 50 mm × 6.4 mm. The surface of the plate was polished and cleaned using acetone prior to the LENS fabrication. The spherical IN718 powders (Carpenter Powder Products Inc., Bridgeville, PA, USA) applied exhibited a particle diameter range from 45 µm to 125 µm.

10.2.2 Experimental set-up and parameters A customized laser additive manufacturing machine (LENS 450XL, Optomec Inc., Albuquerque, NM, USA) was employed to conduct the UV-A LENS fabrication

207 Texas Tech University, Fuda Ning, May 2018 process. As shown in Figure 10.1, UV-A LENS system set-up was mainly composed of a 400W IPG fiber laser system, a four-jet powder and inert gas delivery system, a motion control system, an ultrasonic vibration assisting unit (including an ultrasonic power supply and an ultrasonic transducer assembly), and an infrared (IR) camera unit integrated within the chamber. The flowing argon gas stream could convey the powders to converge at the focal point of the laser beam that could initially produce a molten pool on the substrate. Such a molten pool would be getting larger and larger with the absorption of continually ejected powders. When the laser beam moved to another position, the molten pool began to rapidly solidify into a solid bump because of the heat dissipation. The trajectory generated based on the digital 3D model could direct the formation of a single layer track by track. After this layer deposition, the cladding head increased with a layer thickness for the subsequent layer deposition until the final part was built.

Beam delivery Laser system system

Laser beam Laser generator Power supply Fan

Hooper Z motion

Pyroview software on PC Delivery lines Powder

Cladding head Feeding Nozzle Powder and motor inert gas flow Molten pool IR camera unit Control Vibration Ar Inert gas system with direction Substrate integrated - Transformer Power computer + supply Ceramic Damper Powder & inert gas vibrator delivery system Y X-Y motion table X X Chamber (system)

Figure 10.1 Schematic of ultrasonic vibration-assisted LENS system set-up

In this study, the ultrasonic power supply could convert the regular line electricity into the high-frequency electrical energy. The piezoelectric transducer

208 Texas Tech University, Fuda Ning, May 2018 could generate ultrasonic vibration that would be transmitted onto the substrate by an acoustic transformer. The ultrasonic vibration was in a vertical direction with an ultrasonic amplitude of around 5 µm. Different ultrasonic transducers were applied to allow for the ultrasonic vibrations under different ultrasonic frequencies.

Block parts with dimensions of 8 mm × 8 mm × 4 layers were deposited by LENS without and with ultrasonic vibration under three levels of ultrasonic frequencies, including 25 kHz, 33 kHz, and 41 kHz. The optimal process parameter window has been studied to ensure the successful fabrication of IN718 block parts. The specific LENS manufacturing parameters for IN718 part fabrications are listed in Table 10.1.

Table 10.1 The LENS manufacturing parameters for IN718 bulk part fabrications

Parameter Value Unit Ultrasonic frequency 0, 25 33, 41 kHz Ultrasonic amplitude 5 µm Laser power 350 W Powder flow rate 3.16 g/min Increment of Z axis 0.43 mm (theoretical layer thickness) Argon gas flow rate 6 L/min Axis feedrate 540 mm/min Hatch space 0.3 mm

10.2.3 Characterizations of molten pool, porosity, and microstructures An IR camera (Pyroview 768N, DIAS Infrared Systems, Dresden, Germany) with a measurement range of 1000 °C to 3000 °C was used to measure the temperature in the molten pool during the LENS process. The software (Pyrosoft Compact, DIAS Infrared Systems, Dresden, Germany) installed on the computer can collect the temperature data and thermal images during the measurement. The data acquisition frequency was 25 Hz and the measurement emissivity for IN718 material was set as 0.58. Figure 10.2 shows the methodology to generate the temperature profile within the molten pool using the thermal image. Hence, the temperature value and distribution along the profile line within the molten pool could be evaluated.

209 Texas Tech University, Fuda Ning, May 2018

Profile line

0.3 mm

Laser scanning direction Figure 10.2 Methodology to generate the temperature profile of the molten pool

Molten pool geometry (length L and width W) could be measured using the thermal image of the molten pool, as shown in Figure 10.3. It is known that the melting point of IN718 material was above 1260 °C, based on which the melting boundary of the molten pool could be thus achieved. Therefore, the length L and width W of the molten pool could be calculated by measuring the melting boundary. The molten pool geometry was measured five times and the mean value together with standard deviation would be used for the comparisons at different levels of ultrasonic frequency.

Melting boundary

1260 °C L 0.3 mm

Laser scanning direction Figure 10.3 Measurement of molten pool length L and width W through the thermal image

210 Texas Tech University, Fuda Ning, May 2018

After LENS fabrication, samples were sectioned, ground, and polished on a grinder-polisher machine (MetaServ 250, Buehler, Lake Bluff, IL, USA). Then, an ultrasonic cleaner was utilized to clean the sample surface with acetone to make it free of contaminations. The porosity of fabricated parts was observed using an optical microscope (DX-50, Olympus Corp., Tokyo, Japan), which was defined as the ratio of measured pore areas to the whole selected image area. In order to better calculate the porosity value, the observation was conducted seven times in both horizontal and vertical directions. Then, a total amount of 49 optical microscope (OM) images of the cross-sectional surface were combined without overlaps to identify the porosity value within the whole area of the fabricated part. The OM images were further processed via ImageJ software to “black & white” mode to better quantify the porosity of IN718 parts built at various ultrasonic frequencies, as shown in Figure 10.4.

(a) 0.5 mm

(b) n

1.5 o

t c

1.2 e

i d

0.9

0.6

l i

0.3 u 0.0 B 0 1 2 3 4 5 (mm) Figure 10.4 (a) OM image of porosity of fabricated IN718 parts and (b) porosity analysis/calculation using image processing technique

In order to reveal the grain microstructures, the samples were further etched with Kalling’s reagent (ES Laboratory, Glendora, CA, USA) for 5 min. The etchant was composed of 50% hydrochloric acid, 46% methanol, 2% cupric chloride, and 2% water. The grain microstructures of the fabricated parts were observed using the OM as well and the grain size value was measured using the circular intercept method

211 Texas Tech University, Fuda Ning, May 2018 based on the ASTM E112-13 standard [24]. The methodology for the grain size measurement could also refer to the previous work [22]. In order to better quantify the variation of grain sizes, the measurements were performed five times on different IN718 parts. The mean value and standard deviation were used to represent the obtained results.

10.3 Results and discussion

10.3.1 Effects on molten pool geometry Molten pool is the basic component of the part fabricated by LENS process. The geometry of the molten pool would affect the temperature distribution and further determine the part quality. Figures 10.5(a) and 10.5(b) show the effects of ultrasonic vibration frequency on the molten pool length and molten pool width, respectively. It is notable that both molten pool length and molten pool width increased along with the ultrasonic vibration frequency. Specifically, as the ultrasonic frequency increased from 0 kHz to 41 kHz, the mean value of molten pool length increased from 1.5 mm to 2.4 mm and that of molten pool width increased from 1.2 mm to 2 mm. The increment was 60% and 67%, respectively. The implementation of ultrasonic vibration energy would generate more heat input into the molten pool, which would enlarge the size as compared to the LENS without ultrasonic vibration (ultrasonic frequency of 0 kHz). In addition, an increase of ultrasonic vibration frequency could enhance the acoustic pressure, accelerating the heat convection of the molten pool. Under such a condition, the length and width of the molten pool would be increased.

212 Texas Tech University, Fuda Ning, May 2018

(a)

2.5

2.0

1.5 Molten pool length (mm) length pool Molten

1.0

0 25 33 41 Ultrasonic vibration frequency (kHz) (b) 2.5

2.0

1.5

Molten pool width (mm) width pool Molten 1.0

0 25 33 41 Ultrasonic vibration frequency (kHz) Figure 10.5 Effects of ultrasonic vibration frequency on the (a) length and (b) width of the molten pool

10.3.2 Effects on temperature profiles within molten pool Molten pool temperature is an essential attribute in LENS process, which determines the structural compositions of reactants in the molten pool and further affects the final microstructures and properties of the fabricated parts. The heat transfers by both conduction and convection as well as the flow pattern in the molten pool could be affected by the ultrasonic frequency. The temperature profiles along the molten pool surface at various layers under different ultrasonic vibration frequencies were compared, as shown in Figure 10.6. It can be seen that the increase of ultrasonic frequency resulted in the enhancement of peak temperature value within the molten

213 Texas Tech University, Fuda Ning, May 2018 pool at each layer. This phenomenon could be explained by the cavitation effect, which was a nonlinear action induced by the ultrasonic vibration in the melting materials. Cavitation was a physical phenomenon that produced the generation of tiny bubbles followed by their growth, pulsation, and collapse [20]. It has been reported that the highest temperature within the molten pool showed a linear relationship with the pressure emitted during the collapse of cavitation bubbles [25]. Hence, associated with the results shown in Figure 10.6, the peak temperature value could increase with the increase of acoustic pressure induced by the enhanced ultrasonic frequency. It is also worth noting that the temperature profile in LENS without ultrasonic vibration exhibited a drastic change at the peak temperature region, whereas the temperature profile showed a section of gentle changes at the peak temperature region when ultrasonic vibration with a frequency of 25 kHz was utilized regardless of the layer height. This could be attributed to the mixing and stirring caused by ultrasonic vibration that facilitated the mass and heat transfer within the molten pool. Thus, the temperature distribution tended to be relatively uniform so that the variation around the peak temperature was smaller. With the continuous increase of ultrasonic frequency to 41 kHz, a vigorous increase of the temperature occurred after the section of gentle changes. The enlarged impact force at the increased ultrasonic frequency could be a reason for this result, which indicated the optimal ultrasonic frequency in terms of the temperature distribution during LENS of IN718 parts. In addition, the molten pool length increased with the amount of built layer at each ultrasonic frequency. Such a phenomenon was a result of the heat accumulation in the build direction in LENS process.

214 Texas Tech University, Fuda Ning, May 2018

) 2400 2400 2400 2400

( C)

2200 2200 2200 2200

C) 



 e

r 2000 2000 2000 2000 u

th t a

4 r 1800 1800 1800 1800 e

layer p 1600 1600 1600 1600 m

e 1400 1400 1400 1400 Temperature ( Temperature

Temperature ( Temperature

1200 1200 1200 1200

Position of the profile line Position of the profile line Position of the profile line Position of the profile line

) 2400 2400 2400 2400

(

C) C)

2200 2200 2200 C) 2200

 



 e

r 2000 2000 2000 2000

rd u t

3 a 1800 1800 1800 1800

r e

layer p 1600 1600 1600 1600

m r

e 1400 1400 1400 1400

Temperature ( Temperature Temperature ( Temperature

Temperature ( Temperature

1200 1200 1200 1200

Position of the profile line Position of the profile line Position of the profile line Position of the profile line

) 2400 2400 2400 2400

d C)

l 



( C)

2200 C) 2200 2200 2200



e u r 2000 2000 2000 2000

nd u

B t

2 a 1800 1800 1800 1800

r e

layer p 1600 1600 1600 1600

m Temperature ( Temperature

Temperature ( Temperature

e 1400 1400 1400 1400

Temperature ( Temperature

1200 1200 1200 1200

Position of the profile line Position of the profile line Position of the profile line Position of the profile line

) 2400 2400 2400 2400



( C)

2200 2200 2200 C) 2200



 e

r 2000 2000 2000 2000

u t

st a 1800 1800 1800 1800 r

1 e

p 1600 1600 1600 1600

layer m Temperature ( Temperature

e 1400 1400 1400 1400 Temperature ( Temperature

Temperature ( Temperature Temperature ( Temperature T 1200 1200 1200 1200 PosPositionition ofo thef t profilehe p rlineofile PoPositionsition of o thef tprofilehe p linerofile PoPositionsition of o thef profilethe p linerofile PosPositionition of o thef t profilehe p linerofile

0 25 33 41 Ultrasonic frequency (kHz) Figure 10.6 Effects of ultrasonic vibration frequency on temperature profiles at various positions along the molten pool surface

10.3.3 Effects on temperature fluctuation of single layer and whole parts The peak temperature value was the highest temperature in the molten pool and would be used to represent the molten pool temperature fluctuation along with the deposition time for a single layer and whole parts. Figure 10.7 illustrates the effects of ultrasonic vibration frequency on the profile of peak temperature values within a single layer. The temperature larger than 1600 °C was plotted to better indicate the temperature fluctuation. It can be seen that LENS without ultrasonic vibration caused the temperature to fluctuate dramatically with an overall temperature change of 1067 °C. However, with the ultrasonic frequency increased to 25 kHz, a remarkably smaller temperature change of 868 °C was obtained. The peak temperature values showed a relatively uniform distribution, evidencing a stable thermal history in this condition. The continuous increase of ultrasonic frequency to 33 kHz and later on 41 kHz led to

215 Texas Tech University, Fuda Ning, May 2018 an increasing fluctuation of the temperature with the changes of 991 °C and 1053 °C, respectively. Such a phenomenon may be achieved due to the larger impact force generated at the larger ultrasonic frequency. Thus, the temperature distribution was not as uniform as that produced in the lower ultrasonic frequency of 25 kHz.

(a) (b)

2800 2800

C) C)

 2400  2400

2000 2000

Temperature ( Temperature 1600 ( Temperature 1600

1200 ΔT=1067 °C 1200 ΔT=868 °C

0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 (c) (d) Time (s) Time (s) 2800 2800

C) C)

 2400  2400

2000 2000

Temperature ( Temperature 1600 ( Temperature 1600

1200 ΔT=991 °C 1200 ΔT=1053 °C

0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 Time (s) Time (s) Figure 10.7 Effects of ultrasonic vibration frequency ((a) 0 kHz, (b) 25 kHz, (c) 33 kHz, and (d) 41 kHz) on the profile of peak temperature values within a single layer

As for the peak temperature trends within the whole part fabrication, as shown in Figure 10.8, it can be seen that the thermal history at the second layer was slightly different with that at the first layer in LENS without ultrasonic vibration. When ultrasonic vibration was implemented regardless of the ultrasonic frequencies, all the three conditions resulted in almost the same thermal history and no significant difference of the temperature distribution could be found between the first layer and the second layer. The results indicated that ultrasonic vibration could overall generate a similar temperature field especially at the beginning process of the fabrication.

216 Texas Tech University, Fuda Ning, May 2018

(a) (b)

2800 2800

C) 

 2400 2400

2000 2000

1600 ( Temperature 1600

Temperature ( Temperature

1200 1200

 C) 2400  2400

2000 2000

Temperature ( Temperature 1600 1600 Temperature ( Temperature

1200 1200

0 5 10 15 20 25 30 35 40 45 50 55 60 0 5 10 15 20 25 30 35 40 45 50 55 60 Time (s) Time (s) Figure 10.8 Effects of ultrasonic vibration frequency ((a) 0 kHz, (b) 25 kHz, (c) 33 kHz, and (d) 41 kHz) on the profile of peak temperature values within the four-layer bulk parts

10.3.4 Effects on porosity Porosity, a very important output variable, is a temperature-dependent property that needs to be adequately investigated because it determines the density of the IN718 part fabricated by LENS process. Porosity in LENS process was mainly generated from two sources, including lack of fusion and gas entrapment. Figure 10.9 shows the OM images for calculating porosity of LENS-fabricated IN718 parts under different ultrasonic vibration frequencies. It is notable that a large number of pores existed within the parts fabricated by LENS without ultrasonic vibration. Based on the level of pore size, it can be inferred that most pores were achieved due to the lack of fusion. Such a type of pores was resulted from the discrepancy in the temperature distribution in the molten pool, including the undesired occurrence of partially melted powders those flowed to the underlying layers [26].

217 Texas Tech University, Fuda Ning, May 2018

(a)

Substrate 0.5 mm (b)

Substrate 0.5 mm (c)

Substrate 0.5 mm

(d)

l i

0.5 mm u Substrate B Figure 10.9 OM images for calculating porosity of LENS-fabricated IN718 parts under ultrasonic vibration frequencies of (a) 0 kHz, (b) 25 kHz, (c) 33 kHz, and (d) 41 kHz

When ultrasonic vibration with a frequency of 25 kHz was implemented into the molten pool, the large lack-of-fusion pores were significantly reduced owing to the more uniform distribution of the temperature. The mixing and stirring actions also promoted the escape of those gas-entrapped pores from the molten pool. However, a larger ultrasonic frequency at 33 kHz and 41 kHz gradually weakened the phenomenon of the porosity reduction. As aforementioned, the uniform temperature distribution would be hardly achieved at a larger ultrasonic frequency resulting in the

218 Texas Tech University, Fuda Ning, May 2018 porosity increment again. The calculated porosity values at different conditions of the ultrasonic frequency are illustrated in Figure 10.10. The largest porosity value (1.239%) was obtained at no ultrasonic vibration condition. The porosity value was then decreased to 0.673% at the ultrasonic frequency of 25 kHz before it would increase to 0.937% and finally up to 1.08% at ultrasonic frequencies of 33 kHz and 41 kHz, respectively.

1.5

1.239 1.2 1.08

0.937 0.9

0.673

Porosity (%) 0.6

0.3

0.0 0 25 33 41 Ultrasonic vibration frequency (kHz) Figure 10.10 Effects of ultrasonic vibration frequency on the porosity of LENS- fabricated IN718 parts

10.3.5 Effects on grain microstructures Figure 10.11 shows the grain microstructures of LENS-fabricated IN718 parts under different ultrasonic vibration frequencies. It can be seen that the elongated coarse grains were generated in LENS without ultrasonic vibration. The vertical direction of grain growth was the result of the locally unidirectional solidification behavior, which was along with the build direction of the fabricated parts. The average grain size value was 13.92 m, as illustrated in Figure 10.12. With the introduction of ultrasonic vibration with a frequency of 25 kHz, the grains were transited to the relatively equiaxed morphology with a significantly reduced grain size of 6.69 m. It

219 Texas Tech University, Fuda Ning, May 2018 has been reported that the two simultaneously existed mechanisms induced by the ultrasonic vibration, including cavitation-enhanced nucleation and acoustic streaming- induced dendrite fragmentation, caused a high density of nuclei in the molten pool resulting in the grain refinement [27]. In addition, with the increase of ultrasonic frequency from 25 kHz to 41 kHz, the grain size was gradually reduced to 3.02 m and the distribution of grain microstructures tended to be more uniform, as evidenced by the OM images in Figure 10.11 and the smaller standard deviation especially at the ultrasonic frequency of 41 kHz in Figure 10.12. The reason for such a phenomenon might be that different ultrasonic frequencies produced a various of ultrasonic waves in the molten pool, affecting the acoustic pressure distribution. It was known that a larger ultrasonic frequency could lead to a higher acoustic radiation pressure and a larger number of cavitation bubbles, which could strengthen the nonlinear effects induced by the ultrasonic vibration. As a consequence, the cavitation-enhanced nucleation and acoustic streaming-induced dendrite fragmentation would facilitate the grain refinement at a higher level.

10.4 Conclusions IN718 parts were fabricated using LENS process without and with ultrasonic vibration under three levels of ultrasonic frequency including 25 kHz, 33 kHz, and 41kHz. The effects of ultrasonic frequency on molten pool geometry, temperature profiles within the molten pool, peak temperature value fluctuation, porosity, and grain microstructures were explored. It was found that the changes in ultrasonic frequency can modify the values of output variables in LENS-fabricated IN718 parts. Specifically, the increase of ultrasonic vibration frequency led to the increase of both molten pool length and molten pool width. In addition, the peak temperature value increased with the increase of acoustic pressure induced by the enhanced ultrasonic frequency. However, the temperature distribution around the peak temperature region was more uniform at a lower ultrasonic frequency (25 kHz) compared with that generated at a higher level of ultrasonic frequency (33 kHz or 41 kHz). Furthermore, a stable thermal history within the single layer was obtained at the ultrasonic frequency

220 Texas Tech University, Fuda Ning, May 2018 of 25 kHz by evaluating the peak temperature value fluctuations. Similarly, the continuous increase of ultrasonic frequency gradually led to the occurrence of the large thermal gradient. It was also found that the ultrasonic frequency of 25 kHz was the optimal selection to achieve the lowest value of porosity in the LENS-fabricated IN718 parts. Finally, the grain refinement phenomenon was strengthened with the increase of ultrasonic frequency due to the enhanced actions of both cavitation- enhanced nucleation and acoustic streaming-induced dendrite fragmentation induced by the ultrasonic vibration.

(a) (b)

10 µm Figure 10.11 Grain microstructures of LENS-fabricated IN718 parts under ultrasonic vibration frequencies of (a) 0 kHz, (b) 25 kHz, (c) 33 kHz, and (d) 41 kHz

221 Texas Tech University, Fuda Ning, May 2018

13.92 16

6.69 8 Grain size (um) 4.89

4 3.02

0 0 25 33 41 Ultrasonic vibration frequency (kHz) Figure 10.12 Effects of ultrasonic vibration frequency on the grain size value of LENS-fabricated IN718 parts

References [1] Irwin, J., Reutzel, E.W., Michaleris, P., Keist, J., and Nassar, A.R., 2016, Predicting microstructure from thermal history during additive manufacturing for Ti-6Al-4V, ASME Journal of Manufacturing Science and Engineering, 138(11), pp. 111007–111017. [2] Thomas, D. S., and Gilbert, S. W., 2014, Costs and cost effectiveness of additive manufacturing: a literature review and discussion, National Institute of Standards and Technology (NIST) Special Publication, 1176, pp. 1–77. [3] Gu, D.D., Meiners, W., Wissenbach, K., and Poprawe, R., 2012, Laser additive manufacturing of metallic components: materials, processes and mechanisms, International Materials Reviews, 57(3), pp. 133–164. [4] Jia, Q.B., and Gu, D.D., 2014, Selective laser melting additive manufacturing of Inconel 718 superalloy parts: densification, microstructure and properties, Journal of Alloys and Compounds, 585, pp. 713–721. [5] Bennett, J.L., Wolff, S.J., Hyatt, G., Ehmann, K., and Cao, J., 2017, Thermal effect on clad dimension for laser deposited Inconel 718, Journal of Manufacturing Processes, 28, pp. 550–557. [6] Zhong, C.L., Gasser, A., Kittel, J., Wissenbach, K., and Poprawe, R., 2016, Improvement of material performance of Inconel 718 formed by high deposition- rate laser metal deposition, Materials & Design, 98, pp. 128–134. [7] Liu, F.C., Lin, X., Yang, G.L., Song, M.H., Chen, J., and Huang, W.D., 2011,

222 Texas Tech University, Fuda Ning, May 2018

Microstructure and residual stress of laser rapid formed Inconel 718 nickel-base superalloy, Optics & Laser Technology, 43(1), pp. 208–213. [8] Parimi, L.L., Ravi, G.A., Clark, D., and Attallah, M.M., 2014, Microstructural and texture development in direct laser fabricated IN718, Materials Characterization, 89, pp. 102–111. [9] Tabernero, I., Lamikiz, A., Martínez, S., Ukar, E., and Figueras, J., 2011, Evaluation of the mechanical properties of Inconel 718 components built by laser cladding, International Journal of Machine Tools and Manufacture, 51(6), pp. 465–470. [10] Zhang, Q.L., Yao, J.H., and Mazumder, J., 2011, Laser direct metal deposition technology and microstructure and composition segregation of Inconel 718 superalloy, Journal of Iron and Steel Research, International, 18(4), pp. 73–78. [11] Zhao, X., Chen, J., Lin, X., and Huang, W., 2008, Study on microstructure and mechanical properties of laser rapid forming Inconel 718, Material Science and Engineering A, 478(1), pp. 119–124. [12] Qi, H., Azer, M., and Ritter, A., 2009, Studies of standard heat treatment effects on microstructure and mechanical properties of laser net shape manufactured Inconel 718, Metallurgical and Materials Transactions A, 40(10), pp. 2410–2422. [13] Lambarri, J., Leunda, J., Navas, V.G., Soriano, C., and Sanz, C., 2013, Microstructural and tensile characterization of Inconel 718 laser coatings for aeronautic components, Optics and Lasers in Engineering, 51(7), pp. 813–821. [14] Cao, G.P., Konishi, H., and Li, X.C., 2008, Mechanical properties and microstructure of Mg/SiC nanocomposites fabricated by ultrasonic cavitation based nanomanufacturing, Journal of Manufacturing Science and Engineering, 130(3), pp. 031105–031110. [15] Patarić, A., Mihailović, M., and Gulišija, Z., 2012, Quantitative metallographic assessment of the electromagnetic casting influence on the microstructure of 7075 Al alloy, Journal of Materials Science, 47(2), pp. 793–796. [16] Kore, S.D., Date, P.P., Kulkarni, S.V., Kumar, S., Rani, D., Kulkarni, M.R., Desai, S.V., Rajawat, R.K., Nagesh, K.V., and Chakravarty, D.P., 2011, Application of electromagnetic impact technique for welding copper-to-stainless steel sheets, International Journal of Advanced Manufacturing Technology, 54(9), pp. 949–955. [17] Sun, Q.J., Lin, S.B., Yang, C.L., and Zhao, G.Q., 2013, Penetration increase of AISI 304 using ultrasonic assisted tungsten inert gas welding, Science and Technology of Welding and Joining, 14(8), pp. 765–767. [18] Ezatpour, H.R., Sajjadi, S.A., Sabzevar, M.H., and Huang, Y., 2014, Investigation of microstructure and mechanical properties of Al6061-nanocomposite fabricated by stir casting, Materials and Design, 55, pp. 921–928. [19] Xue, P., Ni, D.R., Wang, D., Xiao, B.L., and Ma, Z.Y., 2011, Effect of friction stir welding parameters on the microstructure and mechanical properties of the

223 Texas Tech University, Fuda Ning, May 2018

dissimilar Al–Cu joints, Materials Science and Engineering: A, 528(13), pp. 4683–4689. [20] Komarov, S.V., Kuwabara, M., and Abramov, O.V., 2005, High power ultrasonics in pyrometallurgy: current status and recent development, ISIJ International, 45(12), pp. 1765–1782. [21] Abramov, O.V., 1987, Action of high intensity ultrasound on solidifying metal, Ultrasonics, 25(2), pp. 73–82. [22] Ning, F.D., Hu, Y.B., Liu, Z.C., Wang, X.L., Li, Y.Z., and Cong, W.L., 2018, Ultrasonic vibration-assisted laser engineered net shaping of Inconel 718 parts: microstructural and mechanical characterization, ASME Transaction Journal of Manufacturing Science and Engineering, 140(6), pp. 061012-1–061012-11. [23] Wu, W., 2000, Influence of vibration frequency on solidification of weldments, Scripta Materialia, 42(7), pp.661-665. [24] ASTM E112-13, Standard Test Methods for Determining Average Grain Size, ASTM International, West Conshohocken, PA, 2013. [25] Rae, J., Ashokkumar, M., Eulaerts, O., von Sonntag, C., Reisse, J., and Grieser, F., 2005, Estimation of ultrasound induced cavitation bubble temperatures in aqueous solutions, Ultrasonics Sonochemistry, 12(5), pp. 325–329. [26] Susan, D.F., Puskar, J.D., Brooks, J.A., and Robino, C.V., 2006, Quantitative characterization of porosity in stainless steel LENS powders and deposits, Materials Characterization, 57(1), pp. 36–43. [27] Chen, X., Le, Q., Wang, X., Liao, Q., and Chu, C., 2017, Variable-frequency ultrasonic treatment on microstructure and mechanical properties of ZK60 alloy during large diameter semi-continuous casting, Metals, 7(5), pp. 173–185.

224 Texas Tech University, Fuda Ning, May 2018

CHAPTER XI

ULTRASONIC VIBRATION-ASSISTED LASER ENGINEERED NET SHAPING OF TIB REINFORCED TI MATRIX COMPOSITES: EFFECTS ON MICROSTRUCTURE AND MECHANICAL PROPERTY

Paper title:

Ultrasonic vibration-assisted laser engineered net shaping of TiB reinforced Ti matrix composites: effects on microstructure and mechanical property

Submitted to:

Additive Manufacturing

Authors:

Fuda Ning, Yingbin Hu, and Weilong Cong

Authors’ affiliations:

Department of Industrial, Manufacturing, and Systems Engineering, Texas Tech University, Lubbock, TX 79409, USA.

225 Texas Tech University, Fuda Ning, May 2018

Abstract Laser engineered net shaping (LENS), one of the laser-based additive manufacturing processes, has successfully produced TiB reinforced Ti matrix composites (TiB-TMC) with a quasi-continuous network (QCN) microstructure. However, internal weaknesses such as porosity, coarse primary TiB whisker, and heterogeneous distribution of TiB reinforcement have been inevitably induced in the fabricated TiB-TMC parts. Ultrasonic vibration has been broadly applied to affect the solidification behavior of liquid melting materials in thermal manufacturing processes such as casting, fusion welding, etc. Given the benefits of ultrasonic vibration in these processes, in this work, ultrasonic vibration-assisted (UV-A) LENS of TiB-TMC has been proposed and conducted. The results of this study show that due to the nonlinear effects of acoustic streaming and cavitation induced by ultrasonic vibration, porosity is significantly reduced and a relatively small variation of pore sizes is achieved. UV-A LENS also results in the formation of smaller TiB whiskers that distribute along the grain boundaries with a homogeneous dispersion. In addition, QCN microstructure is considerably finer than that produced by LENS process without ultrasonic vibration. The refinements of both reinforcing TiB whiskers and QCN microstructural grains further improve the mechanical property of TiB-TMC parts.

Keywords: Laser engineered net shaping (LENS); Ultrasonic vibration; Titanium matrix composites; Porosity; Grain refinement; Microhardness

11.1 Introduction Ceramic reinforced titanium (Ti) matrix composites have attracted great interests in aerospace and defense industries due to their excellent performance under severe friction and heavy load-bearing conditions. In particular, TiB reinforced Ti matrix composite (TiB-TMC) has seen a remarkable increase in demand. Compared with other ceramic reinforcements such as TiC [1], SiC [2], or TiN [3], TiB is the most suitable reinforcement for TMCs with many advantages, including chemically stable composition, high metallurgical bond with Ti matrix, and excellent stiffness and

226 Texas Tech University, Fuda Ning, May 2018 hardness [4, 5]. In addition, the residual stress at the interface between TiB and Ti would be suppressed due to their similar thermal expansion coefficient [6].

Reactive hot pressing (RHP) and casting processes have been reported to fabricate TiB-TMC parts to near net shapes. The effects of tailored microstructures on the mechanical properties have been investigated in both processes [7, 8]. These two traditional processes fabricate structure-restricted TiB-TMC parts with coarse microstructural grains at a high energy consumption [9]. Laser-based additive manufacturing (LAM) technology provides the capability of manufacturing complex structural parts with fine grains at a relatively lower energy consumption [10]. Laser engineered net shaping (LENS), one of the LAM techniques, is a competitive method for manufacturing composite parts due to the benefits of variable powder mixture ratio and controllable cooling rate [11]. LENS implements a focused laser energy source to directly build a designed three-dimensional (3D) component on the substrate in a track-by-track and layer-by-layer mode.

Several investigations have been conducted to study the microstructures and mechanical properties of in situ LENS-fabricated TiB-TMC parts. Banerjee et al. found that the LENS-fabricated TiB-TMC parts exhibited a microstructure consisting of microscale and nanoscale TiB precipitates [12-15]. Hu et al. summarized the mechanism of microstructure formation in LENS and compared it with that in RHP and casting [11]. It was found that finer grains were formed due to the rapid solidification rate and subsequent reheating in LENS. Follow-up studies were performed by Hu et al. to investigate the influences of process parameters (including laser power and Z-axis increment) [16] as well as strengthening and toughening effects [17] on the part performance. It was found that the increased ratio of eutectic TiB to primary TiB whiskers led to the enhancement of microhardness and ultimate compressive strength of TiB-TMC parts. However, internal weaknesses such as porosity, coarse primary TiB whisker, and heterogeneous distribution of TiB reinforcement have been inevitably induced in the LENS-fabricated TiB-TMC parts, which are detrimental to the mechanical properties. These weaknesses are difficult to

227 Texas Tech University, Fuda Ning, May 2018 be simultaneously alleviated simply by conducting parametric studies on the LENS process. In addition, post heat treatment of LENS-fabricated TiB-TMC parts would not have any significant influences on the size and distribution of TiB phase due to the rather limited solubility of boron in Ti matrix [13]. Therefore, it is crucial to investigate an improved LENS manufacturing method to effectively reduce the fabrication defects for obtaining enhanced mechanical properties of TiB-TMC parts.

Recently, ultrasonic vibration, electromagnetic field, and mechanical stirring have been applied in melting material solidification processes such as casting, arc welding, etc. in order to diminish the fabrication defects of metallic or composite parts [18-23]. Among these assistive techniques, ultrasonic vibration has gained more attention due to the effective transmission of acoustic energy at a relatively low equipment cost [24]. Authors’ previous works have shown a good feasibility of ultrasonic vibration implemented in LENS process for metallic part fabrication using stainless steel alloy [25, 26] or nickel alloy [27] powders. The results have indicated that ultrasonic vibration enables to reduce porosity, refine the microstructures, and increase the homogeneity of chemical contents in the LENS-fabricated metallic parts. However, there is a lack of studies on the utilization of ultrasonic vibration in TiB- TMC part manufacturing using LENS process.

In this study, TiB-TMC parts are built by LENS with and without ultrasonic vibration at two different levels of laser power. Porosity, TiB whisker size and distribution, and microstructural grain are evaluated to identify the influence of ultrasonic vibration on these features. Mechanical indentation test has been conducted to assess and compare the part performance of fabricated TiB-TMC. The obtained knowledge of this work will fill the research gaps in the literature on LENS of TiB- TMC materials and provide a significant advance in uncovering the processing- structure-property connection in UV-A LENS process.

11.2 Experimental methods

11.2.1 Materials

228 Texas Tech University, Fuda Ning, May 2018

The commercially pure Ti (CP-Ti) powder (Atlantic Equipment Engineers Inc., Upper Saddle River, NJ) and boron (B) powder (Chemsavers, Inc., Bluefield, WV) used in this work had an average particle size of 150 μm and 2 μm, respectively. A weight ratio of 98.4 to 1.6 between Ti and B powders was employed for part fabrication. As reported, such ratio facilitated the formation of fine-eutectic TiB reinforcement causing more strengthening and toughening effects on TMCs [11]. The as-received CP-Ti (Figure 11.1a) and B (Figure 11.1b) powders were premixed using a ball milling machine at 200 rpm rotation speed for four hours. After ball milling process, CP-Ti powders (Figure 11.1c) exhibited a more spherical shape with a uniform size and B powders were well attached to their surface, as shown in Figure 11.1d.

(a) (b)

200 μm 10 μm Ball milling

(c)

(d)

Boron powder

200 μm 10 μm Figure 11.1 Morphologies of as-received (a) CP-Ti powder, (b) B powder, and four- hour ball milling mixed CP-Ti and B powders at (c) lower and (d) higher magnifications, respectively

229 Texas Tech University, Fuda Ning, May 2018

11.2.2 Experimental set-up and conditions A customized LENS machine (450XL, Optomec Inc., Albuquerque, NM) integrating with an ultrasonic vibration assisting unit was used to build TiB-TMC components. Figure 11.2 shows the schematic of UV-A LENS system that integrated a 400W IPG fiber laser coupled with delivering both powder and inert shield gas into an enclosed chamber. An ultrasonic power supply provided the high-frequency electrical energy converting from the regular line electricity. Thus, ultrasonic vibration was generated by a piezoelectric transducer and then transmitted to the substrate via an acoustic transformer. The direction of ultrasonic vibration was perpendicular to the surface of the substrate. Block parts with dimensions of 8 mm × 8 mm × 10 layers were built by both UV-A LENS and LENS without ultrasonic vibration. The input fabrication variables were listed in Table 1, which were selected based on authors’ previous work [11, 16, 17].

Beam delivery Laser system system

Laser generator Power supply Laser beam Fan

Hooper Z motion

Delivery lines Powder

Cladding head

Figure 11.2 Schematic of ultrasonic vibration-assisted LENS system set-up

230 Texas Tech University, Fuda Ning, May 2018

Table 11.1 Input fabrication variables for TiB-TMC part fabrication Input fabrication variables Values Unit Laser power 200 and 300 W Scanning speed 11 mm/s Powder feeding rate 1.65 g/min Hatch distance 0.38 mm Z axis increment 0.42 mm Number of layers 10 First layer hatch angle 45 ° Interval angle 90 ° Oxygen level <200 ppm Ultrasonic frequency 40 kHz Ultrasonic amplitude 3 µm

11.2.3 Measuring procedures The transverse surface of fabricated parts was sectioned, and then ground and polished on a grinder-polisher machine. After that, an ultrasonic cleaner with acetone was used to remove the contamination on the machined surface. A field emission scanning electron microscopy (FE-SEM) (S4300, Hitachi Co., Tokyo, Japan) was used to observe the pores of the polished parts. To better quantify the porosity values, the SEM images were processed to the ones in a black and white mode using an image processing software ImageJ, as shown in Figure 11.3. The porosity value was defined as the ratio of measured pore areas to the whole selected area. The calculation of porosity was conducted on five different SEM images, and the mean value was used to describe and compare porosity among different fabrication conditions. In addition, the top 20 largest pores were selected from all the five different images to evaluate the variation of pore size under different fabrication conditions.

231 Texas Tech University, Fuda Ning, May 2018

(a) (b) 1 2

4 5

6 n

o 7 i

t 8

r i

d 9

g 10

n n

a 11 0.2 mm c 0.2 mm S Figure 11.3 (a) SEM image of porosity of fabricated TiB-TMC parts and (b) porosity analysis/calculation using image processing technique

In order to reveal the microstructure of the transverse surface, the polished surfaces were further etched by Kroll’s reagent (HF: 3%; HNO3: 6%; and water: balance) (Etchant Store, Suite N Glendora, CA, USA) for 20 seconds. SEM observation was performed to obtain the morphologies and microstructures of the fabricated parts. Additionally, elemental mapping results in the selected micro-regions were characterized by an energy dispersive X-ray spectroscopy (EDS) detector equipped with the SEM.

ASTM E112-13 linear intercept method was used to characterize the microstructural grain size [28]. The measurement began with drawing lines randomly in the images and measuring the length (L) of these lines. Then, the total times (N) of lines crossing grain boundaries was counted. Thus the grain size could be calculated by L/N. Figure 11.4 shows the measuring procedures of microstructural grain size of TiB-TMC parts. The red dots represented the positions where the lines intersected Ti grain boundaries. The average grain size was calculated by dividing the total length of five lines over the total number of red dots. In order to better quantify the variation of QCN microstructural grain size, the measurements were conducted on SEM images of five different TiB-TMC parts.

232 Texas Tech University, Fuda Ning, May 2018

Figure 11.4 The measurement procedures for grain size of TiB-TMC parts

Microhardness on transverse surfaces of the polished parts was measured using a Vickers microhardness tester (900–390A Phase II, Metal-Testers Inc., Nanuet, NY, USA) with a normal force of 9.8 N and a duration of 15 s. Ten random positions were tested for each sample.

11.3 Experimental results

11.3.1 Porosity and pore size Figure 11.5 illustrates the porosity value of TiB-TMC parts fabricated by LENS under different conditions. LENS without ultrasonic vibration at lower level of laser power (200 W/O, for short) led to the highest porosity of 0.79%. With the same laser power, UV-A LENS could fabricate the parts with a porosity value of 0.2%. When laser power was increased to 300 W without ultrasonic vibration (300 W/O), porosity value was reduced from 0.79% to 0.33%. A further reduction of porosity to 0.05% was achieved with the presence of ultrasonic vibration in LENS process at the laser power of 300 W.

233 Texas Tech University, Fuda Ning, May 2018

1.2

1.0

0.8

0.6

Porosity (%) 0.4

0.2

0.0 200 W/O 200 W 300 W/O 300 W Processing conditions Figure 11.5 Porosity values under different processing conditions (W/O: without ultrasonic vibration; W: with ultrasonic vibration)

In order to investigate the pore size variations in TiB-TMC parts, the top 20 largest pores of each sample were measured, as illustrated in Figure 11.6. Generally, the size of each pore with ultrasonic vibration was smaller than that without ultrasonic vibration. As shown in Figure 11.6a, the pore sizes were varied enormously among all the measured 20 pores in the parts fabricated at lower laser power without ultrasonic vibration, in which the largest pore area could be as large as 2000 µm2. However, the pore size and its distribution were remarkably decreased when ultrasonic vibration was implemented. It could be found that significantly large pores were eliminated and the large pore sizes were reduced into the range of 200 – 400 µm2. The variation of pore sizes with higher laser power of 300 W was shown in Figure 11.6b. The increase of laser power led to the reduction of the largest pore size to 1000 µm2, whereas the variation of these pore sizes was still relatively large. In UV-A LENS, a smaller range of pore sizes was obtained with the size lower than 250 µm2.

234 Texas Tech University, Fuda Ning, May 2018

(a) 2500 (b) 2500 200 W/O 300 W/O 200 W 300 W

2000 2000

)

2 2

1500 1500

1000 1000

Pore size (um Pore size (um

500 500

0 0 1 2 3 4 5 6 7 8 9 1011121314151617181920 1 2 3 4 5 6 7 8 9 1011121314151617181920 The number of pores The number of pores Figure 11.6 The variations of pore sizes of TiB-TMC parts fabricated at a laser power of (a) 200 W and (b) 300 W

11.3.2 TiB growth and primary TiB size Ti and B powders reacted during LENS fabrication, forming TiB reinforcement following Ti + B = TiB + G. TiB phases were identified by EDS analysis in authors’ previous investigations [16, 17]. In addition, TiB reinforcements were found to aggregate at the boundary of Ti matrix grains. There were two types of TiB reinforcement: coarser prismatic primary TiB and finer needle-like eutectic TiB. It can be seen from Figure 11.7a that TiB phases originated from a B particle in the core area and scattered to the surrounding region, forming a flower-like structure. Although the global content of B was 1.6 wt% (hyper-eutectic composition) in the mixed composite powders, the local ratio of B would be higher than 1.6 wt% due to the elemental agglomeration. Such B element distribution caused the formation of coarser primary TiB after the laser radiation. Since eutectic TiB was already formed at the nanoscale, the size reduction of microscale primary TiB whisker was more desired for TiB-TMC part fabrication.

In this study, since the length of primary TiB whisker was of similar size under different fabrication conditions, the width of primary TiB whisker size was used to represent the whisker size. Comparison of primary TiB whisker size under different processing conditions was shown in Figure 11.7b. The measured data was represented

235 Texas Tech University, Fuda Ning, May 2018 in the boxplot with values of mean, median, and one-time standard deviation. The mean value of primary TiB whisker size was 0.51 µm in the parts fabricated at lower level of laser power without ultrasonic vibration. In UV-A LENS, the mean size of TiB whisker was reduced to 0.33 µm with a smaller deviation. When laser power increased to 300 W, the mean whisker size was 0.59 µm with a data distribution similar to that achieved at 200 W laser power without ultrasonic vibration (200 W/O, for short). Such phenomena indicated that the increase of laser power exerted few effects on the reduction of primary TiB whisker size. In UV-A LENS at the laser power of 300 W, a significant decrease of mean size (0.2 µm) was achieved with a more uniform size distribution.

( a ) ( b ) 1.0 P r i m a r y T i B

1 S D 0.8

M e d i a n 0.6 M e a n

0.4 - 1 S D

0.2 Primary TiB whisker size (um)

0.0 200 W/O 200 W 300 W/O 300 W E u t e c t i c T i B 2 μ m Processing conditions Figure 11.7 (a) Illustration of two types of TiB whiskers and (b) comparison of primary TiB whisker size among different processing conditions

11.3.3 TiB whisker distribution Figure 11.8 shows the distribution of TiB whisker within Ti matrix with and without ultrasonic vibration. It is notable that TiB whiskers accumulated in an aggregate region and they scattered in a deficient region in LENS without ultrasonic vibration, as indicated by the yellow dash lines in Figures 11.8a and 11.8c. By contrast, a relatively uniform distribution of TiB whiskers was obtained in UV-A

236 Texas Tech University, Fuda Ning, May 2018

LENS, as shown in Figures 11.8b and 11.8d. To evidence the changes of TiB distribution within the matrix, EDS mapping results of TiB-TMC parts fabricated by different processing conditions were conducted and demonstrated in Figure 11.9. The element highlighted in green was B, which could suggest the distribution of TiB whiskers. At either level of laser power, the utilization of ultrasonic vibration resulted in a relatively uniform distribution of TiB reinforcement in the matrix.

Figure 11.8 Distribution of TiB within Ti matrix under processing conditions of (a) 200 W/O, (b) 200 W, (c) 300 W/O, and (d) 300 W

237 Texas Tech University, Fuda Ning, May 2018

Figure 11.9 EDS mapping results of TiB-TMC parts fabricated by different processing conditions

11.3.4 Grain size The microstructural grains of TiB-TMC formed by LENS without and with ultrasonic vibration at laser powers of both 200 W and 300 W were shown in Figure 11.10. It can be seen that the quasi-continuous network (QCN) microstructures were formed due to the connection and accumulation of eutectic TiB whiskers at the boundaries of newly crystallized Ti grains. However, the QCN microstructures were disordered and unremarkable in TiB-TMC parts fabricated by LENS without ultrasonic vibration regardless of the laser power level. With this case, large Ti grains were formed and TiB whiskers unevenly distributed around the Ti grain boundaries, leading to the implicit QCN. With ultrasonic vibration, QCN microstructural grains were rearranged and refined in a better-organized pattern, which caused the formation

238 Texas Tech University, Fuda Ning, May 2018 of clearer QCN microstructures. The QCN microstructural grain size was measured and compared among different processing conditions, as shown in Figure 11.11. At lower laser power, ultrasonic vibration led to a significant reduction of the average grain size from 8.5 µm to 5.1 µm. When laser power increased to 300 W, UV-A LENS process produced a smaller average grain size of 4.3 µm compared with that of 5.6 µm achieved in LENS without ultrasonic vibration. It can be seen that the standard deviation of grain size under each condition was similar.

Figure 11.10 QCN microstructural characteristics under different processing conditions

239 Texas Tech University, Fuda Ning, May 2018

Grain size (um) 4

0 200 W/O 200 W 300 W/O 300 W Processing conditions Figure 11.11 Comparisons of grain size among different processing conditions

11.3.5 Microhardness Comparisons of microhardness on the transverse surface of TiB-TMC parts fabricated under different conditions were illustrated in Figure 11.12. It can be seen that implementing ultrasonic vibration always led to larger mean microhardness with a smaller variation regardless of the laser power level. Specifically, with ultrasonic vibration, microhardness increased from 405 HV1 to 429 HV1 at lower laser power and increased from 428 HV1 to 488 HV1 at higher laser power, resulting in an increment rate of 6% and 14%, respectively. In addition, compared to the condition of 200 W/O, either utilizing ultrasonic vibration in LENS or increasing laser power to the higher level could produce a similar mean value of microhardness. The former action caused a smaller variation of microhardness, however, the latter one led to a quite scattered microhardness data distribution.

240 Texas Tech University, Fuda Ning, May 2018

520 1 SD

Mean 480 Median

) 1 -1 SD

440

400

Microhardness (HV Microhardness

360 200 W/O 200 W 300 W/O 300 W

Figure 11.12 Comparison of microhardness on the transverse surface of TiB-TMC parts fabricated under different conditions

11.4 Discussions Two nonlinear actions of ultrasonic vibration, including acoustic streaming and cavitation, would be caused within the melting materials, facilitating liquid material movements and further resulting in various direct and ultimate influences, as summarized in Figure 2.10 [29]. Acoustic streaming was the steady flow that was driven due to the absorption of acoustic oscillations and spatial variation of the stress in the liquid materials. Such action would remarkably enhance the mass flow, heat transfer, and solute exchange, resulting in a homogeneous temperature field. Cavitation, the other action, was a dynamic phenomenon that produced tiny bubbles or cavities followed by their growth, pulsation, and collapse in the liquid materials [24]. For a certain type of melting material, ultrasonic cavitation occurred only when the acoustic pressure exceeded the critical value (also known as cavitation threshold). The threshold represented the minimum ultrasonic intensity required to initiate the cavitation [30]. The ultrasonic intensity was defined as

241 Texas Tech University, Fuda Ning, May 2018

1 2 J  cv (1) 2 0 where

 is the density of the TiB-TMC medium, 4550 kg/m3; c is ultrasound speed in the melting TiB-TMC, 6100 m/s [31];

v0 is the velocity amplitude of oscillations, which could be expressed by

v0  2fA (2) where f is ultrasonic frequency, 40000 Hz;

A is ultrasonic vibration amplitude, 3 µm.

Thus, the nominal ultrasonic intensity in this work was calculated and the value was 7.88 × 106 W/m2. In order to produce high-intense wave to generate cavitation effects, the critical velocity amplitude for melting materials should be at least 10–4c [24]. Therefore, the threshold value of ultrasonic intensity can be calculated by

1 4 2 J  c(10 c) (3) t 2

It can reckon that the ultrasonic intensity achieved in this work was higher than the threshold value of ultrasonic intensity (5.16 × 106 W/m2), indicating that ultrasonic vibration could produce two nonlinear actions that effectively influenced the molten pool in LENS process. Ultrasonic vibration would be thus verified to be associated with the aforementioned experimental phenomena including porosity reduction, TiB whisker distribution uniformity, refinement of both primary TiB whisker and QCN microstructural grain, and microhardness enhancement.

242 Texas Tech University, Fuda Ning, May 2018

The high-intense acoustic radiation pressure induced by ultrasonic vibration could produce a static head Ps, which could be calculated by

Ps  2cAf (4)

In LENS-fabricated parts, pores were mainly generated from two main sources: gas entrapping and lack of fusion. Based on Eq. (4), the static head in this work was 2.1 × 107 Pa as high as 200 times of the atmospheric pressure. As a consequence, mass flow and heat transfer within the molten pool were effectively affected during the melting material movement, causing indirect actions such as mixing and stirring. The fluidity of melting materials would be thus increased, which allowed the entrapped gas to aggregate and float in a more drastic movement. Pores were prone to escape from the molten pool prior to the molten pool solidification, which could remarkably alleviate the formation of gas-entrapped pores. On the other hand, the higher level of laser energy facilitated a significant decrease of lack-of- fusion pores. The pores would be filled with more melting materials, resulting in the porosity reduction.

A relatively smaller variation of pore sizes could be achieved when ultrasonic vibration was applied regardless of laser power level. The radiation pressure could drive the pores to move in the direction of the induced acoustic field. When the pores were located on the path of an acoustic wave, the force F acting on these pores can be derived by [24]

11k 4R6J F  (5) 9c where, k is the wave number (the length of an acoustic wave divided by the wavelength), which is proportional to ultrasonic frequency;

R is the pore radius, mm.

243 Texas Tech University, Fuda Ning, May 2018

From Eq. (5), it can be seen that pores with a larger radius could flow in a drastic mode under a much higher driving force. Hence, these pores would more easily escape from the molten pool, achieving a smaller variation of pore sizes.

Effective actions of mixing and stirring induced by ultrasonic vibration could better disperse B particles, which would be hardly achieved during powder pre-mixing using ball milling. Such influence would promote the formation of fine eutectic TiB and suppress the growth of coarse primary TiB. Besides the decreased amount of coarse primary TiB, the reduction of primary TiB whisker size was obtained, which was pertaining to a type of structural refinement. This phenomenon was mainly attributed to the ultrasonically induced cavitation effect. The collapse of generated bubbles during cavitation occurrence produced a momentary shock along the wave propagation. The powerful pressure could impact the interface of coarse primary TiB whiskers, refining the whiskers and carrying the fragments away from the interface.

In the LENS-fabricated TiB-TMC parts, TiB reinforcement aggregated together by Van der Waals force. In UV-A LENS, ultrasonic vibration could well disperse and deagglomerate TiB reinforcements in the matrix. Extremely fast micro- jets and enormous shear forces generated by ultrasonic cavitation could effectively break agglomerated TiB reinforcements. On the other hand, acoustic streaming would produce a powerful mixing and stirring to well disperse the TiB whiskers. Both cavitation and acoustic streaming could homogenize temperature field and chemical compositions in the local regions, thereby improving the TiB reinforcement homogeneity, as indicated by the EDS mapping results on B elemental distribution. Such phenomenon was beneficial for the high-quality fabrication of titanium matrix composites reinforced by the non-metallic nano-sized TiB particles.

The formation of the QCN microstructure in LENS-fabricated TiB-TMC parts could be summarized as follows. Due to the laser radiation, TiB was in-situ formed once the reaction between Ti and B occurred and then it was fully melted into the molten pool. The melting material began to rapidly solidify once the laser beam left. A large nucleation rate could be generated as a result of the high degree of undercooling

244 Texas Tech University, Fuda Ning, May 2018

[32]. The liquid Ti was thus homogeneously nucleated, producing Ti nuclei throughout the matrix. Meanwhile, a heterogeneous nucleation of TiB phase occurred and the nucleated TiB grew into a long whisker shape. During the growth of Ti nuclei, TiB whiskers were driven to aggregate at the boundaries of Ti grains forming the QCN microstructure. These TiB whiskers would not be engulfed by Ti nuclei due to the low solubility of TiB in Ti [33].

When the ultrasonic vibration was imported, the grain refinement of QCN microstructure was achieved. The main reason for such refining phenomenon could be attributed to the cavitation-induced nucleation effects. Particularly, the extremely high acoustic pressure generated via the implosion of cavitation bubbles would change the local equilibrium conditions and crystallization kinetics [34]. Consequently, the undercooling of liquid phase became larger in UV-A LENS, improving both homogeneous Ti nucleation and heterogeneous TiB nucleation during the formation of QCN microstructure. The enhanced grain nucleation would further lead to the remarkable reduction of QCN microstructural grain size. On the other hand, with the increase of laser power, the higher applied heat resulted in a larger thermal gradient, which could accelerate the solidification rate that was a driving force causing grain refinement [25]. The finer grains would provide a larger area of grain boundaries to impede the dislocation motion, which was beneficial to the strengthening and toughening effects on the part performance. According to the classic Hall-Petch equation, the increase of microhardness could be attributed to the grain refinement. Additionally, the decrease of porosity and dispersion of TiB reinforcement aforementioned facilitated a better fabrication of more uniformly dense parts, improving the part quality with the enhanced microhardness.

11.5 Conclusions For the first time, this study implemented a UV-A LENS process to fabricate TiB-TMC parts starting with a blend of pure titanium and boron powders. The presence of ultrasonic vibration in LENS exhibited remarkable influences on porosity,

245 Texas Tech University, Fuda Ning, May 2018 pore size, TiB whisker size and distribution, QCN microstructural grain size, and microhardness. The specific conclusions have been drawn as follows.

(1). Ultrasonic vibration alleviated the formation of porosity and caused a small variation of pore sizes. The main reason was that acoustic streaming could remarkably affect the mass flow and heat transfer within the molten pool, producing powerful mixing and stirring actions to reduce the porosity and pore sizes.

(2). UV-A LENS also produced well-dispersed and deagglomerated TiB whiskers in the matrix. Such phenomenon was achieved due to the cavitation-induced shear force that could effectively break the TiB agglomerates.

(3). Microstructural refinements of both TiB whiskers and QCN microstructural grains were obtained in UV-A LENS, which was mainly attributed to the enhanced grain nucleation induced by ultrasonic vibration. The finer microstructures would provide a larger area of grain boundaries to impede the dislocation motion, thereby increasing the microhardness of fabricated TiB-TMC parts.

The obtained results of this work filled the research gaps in the literature on LENS of TiB-TMC and would provide a significant advance in exploring the processing-structure-property relationship in UV-A LENS process. Future work would be conducted to investigate the effects of ultrasonic vibration variables (such as frequency and amplitude) on microstructures and part performance of TiB-TMC parts fabricated by UV-A LENS process.

References [1] Gu, D.D., Meng, G.B., Li, C., Meiners, W., and Poprawe, R., 2012, Selective laser melting of TiC/Ti bulk nanocomposites: Influence of nanoscale reinforcement, Scripta Materialia, 67(2), pp. 185–188. [2] Das, M., Balla, V.K., Basu, D., Bose, S., and Bandyopadhyay, A., 2010, Laser processing of SiC-particle-reinforced coating on titanium, Scripta Materialia, 63(4), pp. 438–441.

246 Texas Tech University, Fuda Ning, May 2018

[3] Balla, V.K., Bhat, A., Bose, S., and Bandyopadhyay, A., 2012, Laser processed TiN reinforced Ti6Al4V composite coatings, Journal of the Mechanical Behavior of Biomedical Materials, 6, pp. 9–20. [4] Chandran, K.R., Panda, K.B., and Sahay, S.S., 2004, TiBw-reinforced Ti composites: processing, properties, application prospects, and research needs, JOM, 56(5), pp. 42–48. [5] Morsi, K., and Patel, V.V., 2007, Processing and properties of titanium-titanium boride (TiBw) matrix composites – a review, Journal of Materials Science, 42(6), pp. 2037–2047. [6] Gorsse, S., Le Petitcorps, Y., Matar, S., and Rebillat, F., 2003, Investigation of the Young’s modulus of TiB needles in situ produced in titanium matrix composite, Materials Science and Engineering: A, 340(1), pp. 80–87. [7] Huang, L.J., Wang, S., Dong, Y.S., Zhang, Y.Z., Pan, F., Geng, L., and Peng, H.X., 2012, Tailoring a novel network reinforcement architecture exploiting superior tensile properties of in situ TiBw/Ti composites, Materials Science and Engineering: A, 545, pp. 187–193. [8] Morikawa, D., Ramkumar, J., Mabuchi, H., Tsuda, H., Matsui, T., and Morii, K., 2002, Microstructure and mechanical properties of Ti-BN cast alloys prepared by reactive arc-melting, Materials Transactions, 43(9), pp. 2193–2196. [9] Zhang, Y.Z., Sun, J.C., and Vilar, R., 2011, Characterization of (TiB + TiC)/TC4 in situ titanium matrix composites prepared by laser direct deposition, Journal of Materials Processing Technology, 211(4), pp. 597–601. [10] Thompson, S.M., Bian, L., Shamsaei, N., and Yadollahi, A., 2015, An overview of direct laser deposition for additive manufacturing; Part I: Transport phenomena, modeling and diagnostics, Additive Manufacturing, 8, pp. 36–62. [11] Hu, Y.B., Zhao, B., Ning, F.D., Wang, H., and Cong, W.L., 2017, In-situ ultrafine three-dimensional quasi-continuous network microstructural TiB reinforced titanium matrix composites fabrication using laser engineered net shaping, Materials Letters, 195, pp. 116–119. [12] Banerjee, R., Collins, P.C., and Fraser, H.L., 2002, Laser deposition of in situ Ti-TiB composites, Advanced Engineering Materials, 4(11), pp. 847–851. [13] Banerjee, R., Collins, P.C., Genc, A., and Fraser, H.L., 2003, Direct laser deposition of in situ Ti-6Al-4V-TiB composites, Materials Science and Engineering: A, 358(1), pp. 343–349. [14] Banerjee, R., Genc, A., Collins, P.C., and Fraser, H.L., 2004, Comparison of microstructural evolution in laser-deposited and arc-melted in-situ Ti-TiB composites, Metallurgical and Materials Transactions A, 35(7), pp. 2143–2152. [15] Banerjee, R., Genc, A., Hill, D., Collins, P.C., and Fraser, H.L., 2005, Nanoscale TiB precipitates in laser deposited Ti-matrix composites, Scripta Materialia, 53(12), pp. 1433–1437. [16] Hu, Y.B., Ning, F.D., Wang, X.L., Wang, H., Zhao, B., Cong, W.L., and Li, Y.Z., 2017, Laser deposition-additive manufacturing of in-situ TiB reinforced titanium matrix composites: TiB growth and part performance, International Journal of Advanced Manufacturing Technology. doi:10.1007/s00170-017-

247 Texas Tech University, Fuda Ning, May 2018

0769-0. [17] Hu, Y.B., Cong, W.L., Wang, X.L., Li, Y.C., Ning, F.D., and Wang, H., 2018, Laser deposition-additive manufacturing of TiB-Ti composites with novel three- dimensional quasi-continuous network microstructure: Effects on strengthening and toughening, Composites Part B: Engineering, 133, pp. 91–100. [18] Cao, G.P., Konishi, H., and Li, X.C., 2008, Mechanical properties and microstructure of Mg/SiC nanocomposites fabricated by ultrasonic cavitation based nanomanufacturing, Journal of Manufacturing Science and Engineering, 130(3), pp. 031105–031110. [19] Sun, Q.J., Lin, S.B., Yang, C.L., and Zhao, G.Q., 2013, Penetration increase of AISI 304 using ultrasonic assisted tungsten inert gas welding, Science and Technology of Welding and Joining, 14(8), pp. 765–767. [20] Patarić, A., Mihailović, M., and Gulišija, Z., 2012, Quantitative metallographic assessment of the electromagnetic casting influence on the microstructure of 7075 Al alloy, Journal of Materials Science, 47(2), pp. 793–796. [21] Kore, S.D., Date, P.P., Kulkarni, S.V., Kumar, S., Rani, D., Kulkarni, M.R., Desai, S.V., Rajawat, R.K., Nagesh, K.V., and Chakravarty, D.P., 2011, Application of electromagnetic impact technique for welding copper-to-stainless steel sheets, International Journal of Advanced Manufacturing Technology, 54(9), pp. 949–955. [22] Ezatpour, H.R., Sajjadi, S.A., Sabzevar, M.H., and Huang, Y., 2014, Investigation of microstructure and mechanical properties of Al6061- nanocomposite fabricated by stir casting, Materials and Design, 55, pp. 921–928. [23] Xue, P., Ni, D.R., Wang, D., Xiao, B.L., and Ma, Z.Y., 2011, Effect of friction stir welding parameters on the microstructure and mechanical properties of the dissimilar Al–Cu joints, Materials Science and Engineering: A, 528(13), pp. 4683–4689. [24] Komarov, S.V., Kuwabara, M., and Abramov, O.V., 2005, High power ultrasonics in pyrometallurgy: current status and recent development, ISIJ International, 45(12), pp. 1765–1782. [25] Ning, F.D., and Cong, W.L., 2016, Microstructures and mechanical properties of Fe-Cr stainless steel parts fabricated by ultrasonic vibration-assisted laser engineered net shaping process, Materials Letters, 179, pp. 61–64. [26] Cong, W.L., and Ning, F.D., 2017, A fundamental investigation on ultrasonic vibration-assisted laser engineered net shaping process, International Journal of Machine Tools and Manufacture, 121, pp. 61–69. [27] Ning, F.D., Hu, Y.B., Liu, Z.C., Cong, W.L., Li, Y.Z., and Wang, X.L., 2017, Ultrasonic vibration-assisted laser engineered net shaping of Inconel 718 parts: a feasibility study, Procedia Manufacturing, 10, pp. 771–778. [28] ASTM E112–13., 2013, Standard test methods for determining average grain size, ASTM International, West Conshohocken, PA, USA. [29] Abramov, O.V., 1987, Action of high intensity ultrasound on solidifying metal, Ultrasonics, 25(2), pp. 73–82.

248 Texas Tech University, Fuda Ning, May 2018

[30] Eskin, G.I., and Eskin, D.G., 2003, Production of natural and synthesized aluminum-based composite materials with the aid of ultrasonic (cavitation) treatment of the melt, Ultrasonics Sonochemistry, 10(4), pp. 297–301. [31] Material Sound Velocities, Retrieved at https://www.olympus-ims.com/en/ndt- tutorials/thickness-gage/appendices-velocities. [32] McCartney, D.G., 1989, Grain refining of aluminum and its alloys using inoculants, International Materials Reviews, 34(1), pp. 247–260. [33] Tamirisakandala, S., Bhat, R.B., Tiley, J.S., and Miracle, D.B., 2005, Grain refinement of cast titanium alloys via trace boron addition, Scripta Materialia, 53(12), pp. 1421–1426. [34] Eskin, D.G., 2017, Ultrasonic processing of molten and solidifying aluminium alloys: overview and outlook, Materials Science and Technology, 33(6), pp. 636–645.

249 Texas Tech University, Fuda Ning, May 2018

CHAPTER XII

CONCLUSIONS AND SCIENTIFIC CONTRIBUTIONS

12.1 Conclusions In this dissertation, a literature review on ultrasonic vibration-assisted (UV-A) manufacturing processes was provided. Investigations on UV-A hole making and surface grinding of carbon fiber reinforced plastic (CFRP) composites and UV-A laser engineered net shaping (LENS) of stainless steel, nickel alloys, and TiB reinforced Ti composites were conducted to improve the process performance and manufactured part quality. The conclusions drawn from this dissertation include:

(1). Comparisons of cutting force, torque, surface roughness, hole diameter, and material removal rate between UV-A hole making and conventional core grinding of CFRP were made. It was found that UV-A hole making performed better than conventional core grinding during drilling of CFRP, leading to lower cutting force, lower torque, and better surface roughness, which answered the first research question.

(2). In UV-A hole making of CFRP, the most significant input variable is feedrate, followed by tool rotation speed, abrasive size, vibration amplitude, and abrasive concentration. All the two-factor interaction effects on cutting force are significant except for the combination of vibration amplitude and abrasive concentration. Moreover, six of the three-factor interaction effects on cutting force are significant. The most significant three-factor interaction effect is the combination of tool rotation speed, feedrate, and abrasive size. The results answered the second research question.

(3). A mechanistic calculation model was developed to investigate ultrasonic vibration amplitude through cutting force in UV-A hole making of CFRP. The ultrasonic vibration amplitude model established was beneficial for exploring the explanations for some experimentally observed phenomena and predicting the results of output variables in UV-A hole making of CFRP. The results show that ultrasonic vibration amplitudes calculated by the theoretical model had similar trends with those

250 Texas Tech University, Fuda Ning, May 2018 experimentally measured results in UV-A hole making of CFRP. The ultrasonic vibration amplitude increased with ultrasonic power increasing, tool rotation speed decreasing, and feedrate increasing. The results answered the third research question.

(4). The comparative study between UV-A surface grinding and conventional surface grinding of CFRP on the feed-direction and axial cutting forces, torque, and surface roughness has been conducted. The results showed that UV-A surface grinding always generated lower values of feed-direction cutting force, axial cutting force, and torque than conventional surface grinding at each combination of tool rotation speed and feedrate, due to the larger effective trajectory length with a smaller indentation depth in UV-A surface grinding. The results answered the first research question.

(5). The mechanistic model for the feed-direction cutting force prediction in UV-A surface grinding of CFRP has been developed. The kinematic motion analysis of one abrasive grain during UV-A surface grinding helped to calculate the material volume removed by the single abrasive grain based on the brittle fracture mechanism. The radial force for one abrasive grain was obtained and then aggregated for feed- direction cutting force calculation by considering all the involved abrasive grains. The theoretically predicted trends of the feed-direction cutting force have been found to be consistent with the experimentally measured results with different input variables. It was also shown that the feed-direction cutting force decreased as tool rotation speed, ultrasonic vibration amplitude, and abrasive size increased and as feedrate, depth of cut, and abrasive concentration decreased. The results answered the fourth research question.

(6). Ultrasonic vibration in LENS process resulted in higher powder utilization efficiency, flatter shape profile, weaker balling effect, larger molten pool size, fewer deposition defects (including pores, cavities, and micro-cracks), and finer grain structures in the fabricated stainless steel 17-4 parts. Mechanical properties including tensile properties and microhardness were thereby improved. These phenomena were mainly attributed to the actions of acoustic streaming and cavitation induced by ultrasonic vibration. The results answered the fifth research question.

251 Texas Tech University, Fuda Ning, May 2018

(7). A remarkable reduction of gas-entrapped pores and finer grains with a uniform distribution were achieved in UV-A LENS of IN718 parts. In addition, ultrasonic vibration was favorable for dissolving Laves phase in the matrix phase. Long bar-shaped Laves phase morphology was fragmented into small sphere-shaped particles in a homogeneous distribution. As a result, the yield strength, UTS, microhardness, and wear resistance of IN718 parts fabricated by UV-A LENS were considerably increased. The results answered the fifth research question.

(8). The increase of ultrasonic vibration frequency led to the increase of molten pool geometry and the peak temperature value. However, the temperature distribution around the peak temperature region was more uniform at a lower ultrasonic frequency (25 kHz) compared with that generated at a higher level of ultrasonic frequency (33 kHz or 41 kHz). It was also found that the ultrasonic frequency of 25 kHz was the optimal selection to achieve the lowest value of porosity in the LENS-fabricated IN718 parts. Finally, the grain refinement phenomenon was strengthened with the increase of ultrasonic frequency due to the enhanced actions of both cavitation- enhanced nucleation and acoustic streaming-induced dendrite fragmentation induced by the ultrasonic vibration. The results answered the sixth research question.

(9). UV-A LENS produced well-dispersed and deagglomerated TiB whiskers in the titanium matrix. Such phenomenon was achieved due to the cavitation-induced shear force that could effectively break the TiB agglomerates. Besides, microstructural refinements of both TiB whiskers and microstructural grains were obtained in UV-A LENS, which was mainly attributed to the enhanced grain nucleation induced by ultrasonic vibration. The results answered the seventh research question.

12.2 Scientific contributions (1). This dissertation discovered the advantages of UV-A hole making and surface grinding over the conventional machining processes of CFRP through experimental comparative studies, in order to determine the feasibility of UV-A machining of CFRP.

252 Texas Tech University, Fuda Ning, May 2018

(2). This dissertation provided fundamental knowledge on material removal mechanism and cutting force generation mechanism in UV-A surface grinding of CFRP through experimentation and mechanistic modeling for analyzing contact mode between abrasives and workpiece, in order to contribute to building an effective, efficient, and high quality CFRP machining process.

(3). This dissertation explored material melting and solidification mechanism by integrating ultrasonic vibration in LENS through experimentation and microstructural analysis, in order to advance the fabrication quality during LENS of metals and metal matrix composites.

253 Texas Tech University, Fuda Ning, May 2018

APPENDIX A

PUBLICATIONS SINCE PH.D. STUDY

Peer-reviewed Journal Articles: [1]. Ning, F.D., Hu, Y.B., Liu, Z.C., Wang, X.L., Li, Y.Z., and Cong, W.L., 2018, “Ultrasonic vibration-assisted laser engineered net shaping of Inconel 718 parts: microstructural and mechanical characterization,” ASME Transaction Journal of Manufacturing Science and Engineering, Vol. 140, No. 6, pp. 061012-1– 061012-11. (IF: 3.48) [2]. Ning, F.D., Cong, W.L., Wang, H., Hu, Y.B., Hu, Z.L., and Pei, Z.J., 2017, “Surface grinding of CFRP composites with rotary ultrasonic machining: a mechanistic model on infeed-directional cutting force,” International Journal of Advanced Manufacturing Technology, Vol. 92, No. 1, pp. 1217-1229. (IF: 2.298) [3]. Ning, F.D., Wang, H., Cong, W.L., and Fernando, P.K.S.C., 2017, “A mechanistic ultrasonic vibration amplitude model during rotary ultrasonic machining of CFRP composites,” Ultrasonics, Vol. 76, pp. 44-51. (IF: 2.281) [4]. Ning, F.D., Cong, W.L., Hu, Z.L., and Huang, K., 2017, “Additive manufacturing of thermoplastic matrix composites using fused deposition modeling: A comparison of two reinforcements,” Journal of Composite Materials. Vol. 51, No. 27, pp. 3733-3742. (IF: 1.595) [5]. Ning, F.D., Hu, Y.B., Liu, Z.C., Cong, W.L., Li, Y.Z., and Wang, X.L., 2017, “Ultrasonic vibration-assisted laser engineered net shaping of Inconel 718 parts: a feasibility study,” Procedia Manufacturing, Vol. 10, pp. 771-778. [6]. Ning, F.D., Wang, H., Hu, Y.B., Cong, W.L., Zhang, M., and Li, Y.Z., 2017, “Rotary ultrasonic surface machining of CFRP composites: a comparison with conventional surface grinding,” Procedia Manufacturing, Vol. 10, pp. 557-567. [7]. Ning, F.D., and Cong, W.L., 2016, “Microstructures and mechanical properties of Fe-Cr stainless steel parts fabricated by ultrasonic vibration-assisted laser engineered net shaping process,” Materials Letters, Vol. 179, pp. 61-64. (IF: 2.426) [8]. Ning, F.D., Cong, W.L., Pei, Z.J., and Treadwell, C., 2016, “Rotary ultrasonic machining of CFRP: A comparison with grinding,” Ultrasonics, Vol. 66, pp. 125-132. (IF: 2.281) [9]. Ning, F.D., Cong, W.L., Hu, Y.B., and Wang, H., 2016, “Additive manufacturing of CFRP composites using fused deposition modeling: Effects of process parameters on tensile properties,” Journal of Composite Materials. Vol. 51, No. 4, pp. 451-462. (IF: 1.595) (One of the most read articles in this journal) [10]. Ning, F.D., Cong, W.L., Qiu, J.J., Wei, J.J., and Wang, S.R., 2015, “Additive

254 Texas Tech University, Fuda Ning, May 2018

manufacturing of carbon fiber reinforced thermoplastic composites using fused deposition modeling,” Composites Part B: Engineering, Vol. 80, pp. 369-378. (IF: 4.644) (One of the most cited & downloaded articles in this journal) [11]. Cong, W.L., and Ning, F.D., 2017, “A fundamental investigation on ultrasonic vibration-assisted laser engineered net shaping process,” International Journal of Machine Tools and Manufacture, Vol. 121, pp. 61-69. (IF: 5.076)

Book Chapter: [12]. Cong, W.L., and Ning, F.D., 2015, Chapter 2 Rotary Ultrasonic Machining of CFRP Composites, in Machinability of Fibre-Reinforced Plastics, pp. 31-81, J. P. Davim (Ed.). Walter de Gruyter GmbH & Co KG.

Conference Proceedings: [13]. Ning, F.D., Hu, Y.B., Liu, Z.C., and Cong, W.L., 2016, “Microstructural and mechanical performance of Al2O3 nanoparticle reinforced 17-4 PH stainless steel bulk composite parts fabricated by laser engineered net shaping process,” 2016 Annual International Solid Freeform Fabrication Symposium, August 8-10, Austin, Texas, USA. [14]. Ning, F.D., Cong, W.L., Jia, Z.Y., Wang, F.J., and Zhang, M., 2016, “Additive manufacturing of CFRP composites using fused deposition modeling: effects of process parameters,” Proceedings of the ASME 2016 International Manufacturing Science and Engineering Conference (MSEC2016-8561), June 27 - July 01, 2016, Blacksburg, Virginia, USA. [15]. Ning, F.D., Cong, W.L., Wei, J.H., Wang, S.R., and Zhang, M., 2015, “Additive manufacturing of CFRP composites using fused deposition modeling: effects of carbon fiber content and length,” Proceedings of the ASME 2015 International Manufacturing Science and Engineering Conference (MSEC2015-9436), June 8- 12, 2015, Charlotte, North Carolina, USA. [16]. Ning, F.D., and Cong, W.L., 2015, “Rotary ultrasonic machining of CFRP: design of experiment with a cutting force model,” Proceedings of the ASME 2015 International Manufacturing Science and Engineering Conference (MSEC2015-9227), June 8-12, 2015, Charlotte, North Carolina, USA.

255