OPTIMIZATION OF PROCESS PARAMETERS OF WIRE ELECTRICAL DISCHARGE MACHINING OF Ti6Al4V ALLOY USING TEACHING-LEARNING-BASED OPTIMIZATION

Transcription

1 OPTIMIZATION OF PROCESS PARAMETERS OF WIRE ELECTRICAL DISCHARGE MACHINING OF Ti6Al4V ALLOY USING TEACHING-LEARNING-BASED OPTIMIZATION Submitted by JAYDEEP R. ODEDARA Under guidance of JINESH B. SHAH M.Tech. (Thermal Engineering) Assistant Professor, Department of Mechanical Engineering, AITS, RAJKOT A thesis Submitted to Gujarat Technological University in Partial Fulfillment of the Requirements for The Post Graduate Degree of Master of Engineering In Production Engineering May, 2016 ATMIYA INSTITUTE OF TECHNOLOGY AND SCIENCE, YOGIDHAM GURUKUL, RAJKOT, GUJARAT

2 CERTIFICATE This is to certify that research work embodied in this thesis entitled OPTIMIZATION OF PROCESS PARAMETERS OF WIRE ELECTRICAL DISCHARGE MACHINING OF Ti6Al4V ALLOY USING TEACHING-LEARNING-BASED OPTIMIZATION was carried out by Mr. Jaydeep Odedara ( ) studying at Atmiya Institute Of Technology And Science, Rajkot (003) for partial fulfillment of Master of Engineering degree in Production Engineering to be awarded by Gujarat Technological University. This research work has been carried out under my guidance and supervision and is up to my satisfaction. Date: Place: Guide Asst. Prof. Jinesh B. Shah Mechanical Department AITS, Rajkot Head of Department Dr. P. S. Puranik Mechanical Department AITS, Rajkot Principal Dr. G. D. Acharya AITS, Rajkot Seal of Institute I

3 INDUSTRY CERTIFICATE II

4 COMPLIANCE CERTIFICATE This is to certify that research work embodied in this thesis entitled OPTIMIZATION OF PROCESS PARAMETERS OF WIRE ELECTRICAL DISCHARGE MACHINING OF Ti6Al4V ALLOY USING TEACHING-LEARNING-BASED OPTIMIZATION was carried out by Jaydeep R. Odedara at Atmiya Institute of Technology and Science - Rajkot for partial fulfillment of Master of Engineering degree to be awarded by Gujarat Technological University. He has complied with the comments given by the Dissertation phase I as well as Mid Semester Thesis Reviewer to my satisfaction. Date: Place: Jaydeep R. Odedara Guide Asst. Prof. Jinesh B. Shah Mechanical Department AITS, Rajkot III

5 PAPER PUBLICATION CERTIFICATE This is to certify that research work emboided in this thesis entitled OPTIMIZATION OF PROCESS PARAMETERS OF WIRE ELECTRICAL DISCHARGE MACHINING OF Ti6Al4V ALLOY USING TEACHING-LEARNING-BASED OPTIMIZATION was carried out by Mr. Jaydeep R. Odedara (Enrollment No: ) at Atmiya Institute of Technology and Science,Rajkot,Gujarat (Institute code-003) for partial fullfillments of Masters of Engineering degree to be awarded by Gujarat Technology University, has published article entitled Optimization of the process parameters of wire cut EDM A review for the publication by the International Journal for Scientific Research & Development, Volume 3, Issue 10, December -2015, ISSN(O): during December, Date: Place: Jaydeep R. Odedara Guide Asst. Prof. Jinesh B. Shah Mechanical Department AITS, Rajkot Head of Department Dr. P.S. Puranik Mechanical Department AITS, Rajkot Principal Dr. G.D. Acharya AITS, Rajkot Seal of Institute IV

6 THESIS APPROVAL CERTIFICATE This is to certify that research work embodied in this thesis entitled OPTIMIZATION OF PROCESS PARAMETERS OF WIRE ELECTRICAL DISCHARGE MACHINING OF Ti6Al4V ALLOY USING TEACHING-LEARNING-BASED OPTIMIZATION was carried out by Jaydeep R. Odedara at Atmiya Institute of Technology and Science (003) is approved for the degree of Master of Engineering with specialization of Production Engineering by Gujarat Technological University. Date: Place: Examiner s Sign and Name: ( ) ( ) V

7 DECLARATION OF ORIGINALITY We hereby certify that we are the sole authors of this thesis and that neither any part of this thesis nor the whole of the thesis has been submitted for a degree to any other University or Institution. We certify that, to the best of our knowledge, the current thesis does not infringe upon anyone s copyright nor violate any proprietary rights and that any ideas, techniques, quotations or any other material from the work of other people included in our thesis, published or otherwise, are fully acknowledged in accordance with the standard referencing practices. Furthermore, to the extent that we have included copyrighted material that surpasses the boundary of fair dealing within the meaning of the Indian Copyright (Amendment) Act 2012, we certify that we have obtained a written permission from the copyright owner(s) to include such material(s) in the current thesis and have included copies of such copyright clearances to our appendix. We declare that this is a true copy of thesis, including any final revisions, as approved by thesis review committee. We have checked write up of the present thesis using anti-plagiarism database and it is in allowable limit. Even though later on in case of any complaint pertaining of plagiarism, we are sole responsible for the same and we understand that as per UGC norms, University can even revoke Master of Engineering degree conferred to the student submitting this thesis. Date: Signature of Student: Signature of Guide : Name of Student: Name of Guide : Enrollment No: Institute Code : VI

8 ACKNOWLEDGEMENT I wish to express my sincere gratitude and regards to my project guide, Asst. prof. Jinesh B. Shah. His guidance and support throughout the project has been a major factor in the successful completion of the present work. This work would not have culminated into the present form without his invaluable suggestions and generous help. I am also very thankful to my co-guide Asst. prof. Pallav M. Radia, the person who makes me to follow the right steps during a research project. I am thankful to all faculties and friends at Atmiya Institute of Technology & Science, Rajkot who not only provided valuable suggestions and constant help during my work. I am sincerely thankful to Mr. Vinay Surelia (Managing director-surelia wire-cut PVT LTD) for providing me all the data and allowing me to visit the company whenever it is required to do so. Above all, I am forever thankful to my parents as well as my friends for their invaluable time and encouragement. JAYDEEP R. ODEDARA VII

9 TABLE OF CONTENTS TITLE PAGE CERTIFICATE INDUSTRY CERTIFICATE COMPLIANCE CERTIFICATE PAPER PUBLICATION CERTIFICATE THESIS APPROVAL CERTIFICATE DECLARATION OF ORIGINALITY ACKNOWLEDGEMENT TABLE OF CONTENTS LIST OF FIGURES LIST OF TABLES NOMENCLATURE ABSTRACT I II III IV V VI VII VIII XI XIII XIV XV Chapter 1 Introduction Introduction Basic principle of WEDM process Mechanism of material removal rate WEDM process parameters and responses WEDM process parameters WEDM process Responses Advantages of WEDM process Disadvantages of WEDM process Application of WEDM process Company profile 7 VIII

10 Chapter 2 Literature review Introduction Remarks from Literature review Problem identification Objectives 13 Chapter 3 Design of experiments Introduction to Design of Experiments Selection of process parameters Design of experiments by Taguchi method 15 Chapter 4 Experimental work and results WEDM machine Workpiece information Typical Properties of Titanium Grade 5 Alloy Applications of Titanium Grade 5 Alloy Wire Material Experimental results Main effect plots Main effect plots for brass wire Main effect plots for diffusion zinc coated wire 27 Chapter 5 Regression analysis Definition Linear regression Non-linear regression Regression analysis equation Simple linear regression analysis Data resulting from regression equation 35 IX

11 Chapter 6 Optimization using TLBO Introduction to TLBO Teaching Learning Bases optimization Teacher phase Learner phase Steps to implement TLBO Optimization Multi-objective function and limits of variables Results after optimization using TLBO algorithm 43 in MATLAB Chapter 7 Conclusion and Future scope Conclusion Future Scope 46 References 47 Appendix A List of TLBO code used 49 Appendix B Review Card 58 Appendix C Compliance report 64 Appendix D Paper publication certificate 65 Appendix E Plagiarism report 66 X

12 LIST OF FIGURES Figure 1.1 Schematic diagram of basic principle of WEDM 2 Figure 1.2 Detail of WEDM cutting gap 3 Figure 1.3 Ishikawa Cause and Effect Diagram for WEDM Process 6 Figure 4.1 ELEKTRA SPRINTCUT 734 WEDM Machine 18 Figure 4.2 Spool of wire material 21 Figure 4.3 Mitutoyo s SJ-201 surface tester machine 22 Figure 4.4 Figure 4.5 Workpiece material after conducting experiments by Brass wire Workpiece material after conducting experiments by diffusion zinc coated wire Figure 4.6 Main effect plots for Cutting rate (Brass wire) 26 Figure 4.7 Main effect plots for Surface roughness (Brass wire) 26 Figure 4.8 Main effect plots for Cutting rate (Diffusion zinc coated wire) 27 Figure 4.9 Figure 5.1 Figure 5.2 Figure 5.3 Figure 5.4 Figure 5.5 Figure 6.1 Figure 6.2 Main effect plots for Surface roughness (Diffusion zinc coated wire) Model order and their generated regression line with data points Residual Plots for Cutting rate for linear regression (Brass wire) Residual Plots for Surface roughness for linear regression (Brass wire) Residual Plots for Cutting rate for linear regression (Diffusion zinc coated wire) Residual Plots for Surface roughness for linear regression (Diffusion zinc coated wire) Distribution of marks obtained by learners taught by two different teachers Model for the distribution of marks obtained for a group of learners Figure 6.3 Flow-chart of the working of TLBO algorithm 39 XI

13 Figure 6.4 Result values of the responses v/s iterations for brass wire 44 Figure 6.5 Result values of the responses v/s iterations for diffusion zinc coated wire 44 XII

14 LIST OF TABLES Table 2.1 Literature survey 8 Table 3.1 Range of the parameters 15 Table 3.2 Process parameters with their levels 15 Table 3.3 DOE using Taguchi s L27 orthogonal array 16 Table 4.1 Specification of WEDM machine tool 19 Table 4.2 Workpiece specification 19 Table 4.3 Material test report 20 Table 4.4 Properties of titanium grade 5 alloy 20 Table 4.5 Data obtained by conducting experiments by brass wire 24 Table 4.6 Data obtained by conducting experiments by diffusion zinc coated wire 25 Table 4.7 Result table 28 Table 5.1 Theoretical data from regression equation 35 XIII

15 NOMENCLATURE CR DOE GA IP OA SR SV T on T off TLBO W t WEDM Cutting Rate Design of Experiments Genetic algorithm Peak Current Orthogonal Array Surface Roughness Servo Voltage Pulse on time Pulse off time Teaching-Learning-Based optimization Wire tension Wire Electrical Discharge Machining XIV

16 OPTIMIZATION OF PROCESS PARAMETERS OF WIRE ELECTRICAL DISCHARGE MACHINING OF Ti6Al4V ALLOY USING TEACHING-LEARNING- BASED OPTIMIZATION Submitted by Jaydeep R. Odedara (Enrollment No: ) Supervised By Asst. Prof. JINESH B. SHAH M.Tech. (Thermal Engineering), Department of Mechanical Engineering, Atmiya Institute of Technology and Science, Rajkot ABSTRACT Wire-cut electrical discharge machining (WEDM) process is non-traditional type machining process. The material is removed by thermo-electric spark erosion process. Wire EDM machines can cut conductive metals of any hardness that are difficult or impossible to cut with the conventional methods. The process is used to produce intricate shapes and profile. Titanium grade5 (Ti6Al4V) material is selected as a workpiece material. Ti6Al4V material is used in many applications such as, aerospace, medical, military and other commercial application because of its high strength to weight ratio and its exceptional resistance to corrosion at elevated temperature. Brass wire and diffusion zinc coated wire is used as wire electrode material. The objective of this study is to find relation between process parameters such as pulse on time, pulse off time, peak current, wire tension and servo voltage for the cutting rate & surface roughness. Regression analysis is use to generate the equations for responses. Teaching-Learning-Based optimization algorithm is used to find the optimum values of process parameters. XV

17 CHAPTER 1 INTRODUCTION 1.1 INTRODUCTION Wire-electrical discharge machining (WEDM) is a non-traditional machining process in which the material is removed by the thermo-electric spark erosion process. In today s world, with the progress of mechanical industry, the use of materials having high hardness and toughness are increasing. Therefore, conventional machining methods are not work on this type of material. Hence, non-traditional machining techniques are used to machine such materials which are having high hardness and toughness. WEDM process can be used to cut any electrical conductive metal and alloys regardless of their hardness. WEDM process which is like a contour cutting with a band saw, a moving wire goes along a fixed path, cutting the workpiece with a number of electrical discharge sparks acting like cutting teeth. WEDM process can be used to create complex shapes and profile. In WEDM process, thin wire electrode is used, which transforms electrical energy into thermal energy to cut the material. The wire electrode does not touch the workpiece, but there is a small gap between the workpiece and the wire electrode. Therefore there is no mechanical stress produces during the process. The wire is kept in tension with the help of mechanical tensioning device, to decrease the tendency of producing incorrect parts. The wire material is made of copper, brass, molybdenum and also coated wire is used for some applications. 1.2 BASIC PRINCIPLE OF WEDM PROCESS The WEDM machine tool mainly contains four major components which are computer numerical control (CNC), power supply, mechanical section and dielectric system. The mechanical section mainly consists of main work table (X-Y), auxiliary table (U-V) and wire drive mechanism. The main work table (X-Y) is move in X and Y direction and it is driven by the D.C. servo motors. In WEDM machine two wire guides are located at the opposite sides of workpiece. The wire is continuously moving from wire feed spool and collected on take up spool. The wire is kept in Atmiya Institute of Technology and Science Page 1

18 INTRODUCTION tension between a pair of wire guides. The upper wire guide is supported by the U-V table, and can be move transversely along U and V axis with respect to lower wire guide which is stationary. [16] The electrical spark is generated by the pulse generator unit. The spark is generated in the small gap between the workpiece and the wire. The machining zone is constantly flushed by the dielectric fluid by the upper and lower nozzle on the both sides of workpiece. The deionized water is use as dielectric medium due to its low viscosity and fast cooling rate. The CNC controller gives path information for the operation. The schematic diagram of basic principle of WEDM process and details of WEDM cutting gap is shown in figure 1.1 and figure 1.2 respectively. Figure 1.1 Schematic diagram of basic principle of WEDM process [16] Atmiya Institute of Technology and Science Page 2

19 INTRODUCTION Figure 1.2 Detail of WEDM cutting gap [5] 1.3 MECHANISM OF MATERIAL REMOVAL RATE The electrical discharge spark is continuous generated between workpiece and wire electrode. The material is removed because of melting and vaporization caused by sparks. In WEDM, negative electrode is wire and the positive electrode is workpiece. The heat is produced due to continuous electrical sparks. The temperature generated is around 8000 C to C, therefore material is vaporize and melt. The eroded particles of material are flushed away be the top and bottom nozzle of deionized water. 1.4 WEDM PROCESS PARAMETERS AND RESPONSES The WEDM process is affected by the large number of process parameters. Therefore it is challenging to achieve optimal performance of the process. The main objective is to achieve higher productivity and better surface finish. The process parameters and output characteristics of the WEDM process is given below WEDM PROCESS PARAMETERS WEDM performance is affected by the many parameters which are given below. Atmiya Institute of Technology and Science Page 3

20 INTRODUCTION I. Pulse on time The pulse on time is the duration of time, for which current is flowing in each cycle. Pulse on time is denoted as T on and measured in micro second (µs). It is also known as pulse duration. With the larger value of pulse on time, material removal rate (MRR) is high and surface finish is rough because of the larger and deeper craters are formed. II. Pulse off time The pulse off time is the duration of time, between the two simultaneous sparks. It is also known as pulse interval. Pulse off time is denoted as T off and measured in micro second (µs). Pulse off time gives time to clear particles from gap between workpiece and wire. If the pulse off time is large then the machining time will be high. III. Peak current The peak current is the maximum value of current passing through the electrodes for that pulse. It is denoted by IP. As the value of peak current increases MRR and surface roughness increases. IV. Servo voltage The servo voltage is a reference voltage for the actual gap between the workpiece and wire. High voltage increases the gap and therefore increases in the flushing and machining of the process. It is denoted as SV. V. Wire feed Wire feed is the rate at which wire travels and continuous fed through wire guide. If the wire feed is low then there is a chance of wire breakage. Wire feed rate is denoted as W F. VI. Wire tension Wire tension is the amount of tension given to wire between upper and lower wire guide. If the thickness of the material is more, then more tension is required. It is denoted as W T. Atmiya Institute of Technology and Science Page 4

21 INTRODUCTION VII. Dielectric flushing pressure Dielectric flushing pressure is the input pressure of the dielectric. The range available for flushing pressure is either 1 (high) or 0 (low). For the thin workpiece, low input pressure is required. VIII. Servo feed Servo feed value decides the servo speed. It is denoted by SF. It can be given in normal feed mode and constant feed mode WEDM PROCESS RESPONSES The performance of the WEDM process is given by cutting rate, surface roughness, gap current and dimensional deviation. WEDM process responses are discussed below. I. Cutting rate In WEDM process, cutting rate is required characteristic and it should be high to increase production. It is the rate at which cutting process is done. In WEDM machine, cutting rate is given on display of machine and it is given in mm/min. II. Surface roughness Roughness can be defined as the texture of the surface. It is due to the irregularities in the surface. It is measured by vertical deviations of a real surface from its ideal form. If the deviations are large, the surface is rough. Surface roughness is measured with the help of Mitutoyo s surftest SJ-201 machine. Ra value is the average value of all peak and valley to the mean line. III. Dimensional deviation The dimensional deviation is measured with the help of micrometer. Dimensional deviation is measured from the cross section of specimen. Dimensional deviation is measured by the following equations. Atmiya Institute of Technology and Science Page 5

22 INTRODUCTION IV. Gap current In WEDM process, there is a small gap between workpiece and wire electrode. To start the cutting pulse of current is given by the pulse generator and the current passes through the material to be cut which is known as gap current. Gap current is displayed on the machine and measured in amperes. In figure 1.3, Ishikawa cause and effect diagram is shown. In this figure, all the affecting process parameters and output characteristics of the process is shown. Figure 1.3 Ishikawa Cause and Effect Diagram for WEDM Process [16] 1.5 ADVANTAGES OF WEDM PROCESS In WEDM process, there is small gap between the workpiece and the wire, therefore there is no mechanical stress during machining. All electrical conductive metals can be machined using WEDM process regardless of their hardness and toughness. There is no need of continuous operate the machine, operation can run on workpiece unattended. Fabrication of electrode is not required, because continuous travelling wire is use as electrode. Atmiya Institute of Technology and Science Page 6

23 INTRODUCTION 1.6 DISADVANTAGES OF WEDM PROCESS Capital cost is high. Process is not applied to very large workpiece. Formation of recast layer. 1.7 APPLICATIONS OF WEDM PROCESS Automotive industry Aerospace industry Tool and die making industry Medical and dental instrumentation Narrow slots and keyways 1.8 COMPANY OVERVIEW Name : Surelia Wire-cut PVT LTD Address : Mahendra industrial complex, Tagore road, Rajkot Industrial Guide : Mr. Niren Kadecha Phone No. : sureliawire@yahoo.com Website : The idea of the company is to become a reliable engineering partner and the most favored destination for machining operations by giving high quality using most recent innovation. The company provides various machining facilities which are given below CNC machining center CNC turning center CNC wire EDM Sinking EDM Other conventional machines Atmiya Institute of Technology and Science Page 7

24 CHAPTER 2 LITERATURE REVIEW 2.1 INTRODUCTION In this part, the literature review regarding the optimization of the process parameters of WEDM process is carried out. Optimum utilization of the capabilities of WEDM process requires the selection of an appropriate set of machining parameter. Remarks from the literature review are given in the section 2.2. Table 2.1 Literature survey Sr. No Title Author Journal Analysis and study the effects of various control factors of CNC-wire cut EDM for S7 steel Recent Developments in Wire EDM: A Review Parameter optimization of modern machining processes using teachinglearning-based optimization algorithm Spark gap optimization of WEDM process on Ti6Al4V Sonu Dhiman, Ravinder Chaudhary, V.K.Pandey Manpreet Singh, Harvinder Lal, Ramandeep Singh R. Venkata Rao, V. D. Kalyankar Kuriachen Basil, Dr. Josephkunju Paul, Dr. Jeoju M. Issac Mechanical Engineering : An International Journal ( MEIJ) International Journal of Research in Mechanical Engineering & Technology (IJRMET) Elsevier Engineering applications of artificial intelligence International Journal of Engineering Science and Innovative Technology (IJESIT) Year of Publication Optimization of WEDM process parameters on titanium alloy using Taguchi method Lokeswara Rao T. N. Selvaraj International Journal of Modern Engineering Research (IJMER) 2013 Atmiya Institute of Technology and Science Page 8

25 LITERATURE REVIEW Effects of Process Parameters on Material Removal Rate and Surface Roughness in WEDM of P20 Tool Steel Influence of machine feed rate in WEDM of titanium Ti-6Al-4V with constant current (6A) using brass wire Optimization of process parameters of wire EDM using zinc-coated brass wire Optimization material removal rate in WEDMing titanium alloy using the taguchi method Productivity and workpiece surface integrity when WEDM aerospace alloys using coated wires Evaluation of optimal parameters for machining brass with wire cut EDM A study on kerf and material removal rate in wire electrical discharge machining based on Taguchi method An investigation on wire wear in WEDM Jaganjeet Singh, Sanjeev Sharma Aniza Alias, Bulan Abdullah, Norliana Mohd Abbas Manoj Malik, Rakesh Kumar Yadav, Deepak Sharma, Manoj Danial Ghodsiyeh, Mohammadreza Askaripour Lahiji, Mahdi Ghanbari, Mostafa Rezazadeh Shirdar, Abolfazl Golshan M. T. Antar, S.L. Soo, D. k. Aspinwall, D. Jones, R. Perez C V S Parmeswara Rao, M M M Sarcar Nihat Tosun, Can Cogun, Gul Tosun Nihat Tosun, Can Cogun International Journal of Multidisciplinar y and Current Research Elsevier Procedia Engineering 41 International Journal of Advanced Technology & Engineering Research (IJATER) Research journal of applied science, engineering and technology Elsevier Procedia Engineering 19 Journal of scientific & industrial research Elsevier Journal of Materials Processing Technology Elsevier Journal of materials processing technology Atmiya Institute of Technology and Science Page 9

26 LITERATURE REVIEW 2.2 REMARKS FROM LITERATURE REVIEW Sonu Dhiman, Ravinder Chaudhary, V.K. Pandey, Analysis and study the effects of various control factors of CNC-wire cut EDM for S7 steel [1], In this paper, the effect of different process parameters like, pulse on time, pulse of time, servo voltage, peak current, wire feed, wire tension on cutting rate of S7 steel is studied. One factor at a time (OFAT) method is used. Cutting rate is increase with increase in pulse on time, peak current. Cutting rate is decrease with increase of pulse duration and servo gap voltage. Manpreet Singh, Harvinder Lal, Ramandeep Singh, Recent Developments in Wire EDM: A Review [2], this research paper gives brief idea of the WEDM process and researches done to increases performances. Also it gives development in new control systems, new guide to eliminate wire bending defects and new material for the wire electrode, new multi-layered electrodes developed for low wear rate and low breakage at increased currents. R. Venkata Rao, V. D. Kalyankar, Parameter optimization of modern machining processes using teaching-learning-based optimization algorithm [3], This paper gives brief idea about TLBO algorithm and it can be applied to modern machining processes like ultrasonic machining, abrasive water jet machining, wire electrical discharge machining etc. This paper also gives comparison between other optimization techniques and TLBO algorithm. TLBO algorithm is better than the other optimization algorithm in terms of results. Kuriachen Basil, Dr. Josephkunju Paul, Dr. Jeoju M. Issac, Spark gap optimization of WEDM process on Ti6Al4V [4], In this study the effect of voltage, dielectric process, pulse on time and pulse off time on spark gap of Ti6Al4V alloy is studied. Full factorial method is used. The pulse on time, pulse off time, the interaction of dielectric pressure & pulse off time and interaction of pulse on time & pulse off time are the most affecting parameters for spark gap. Lokeswara Rao T. and N. Selvaraj, Optimization of WEDM process parameters on titanium alloy using Taguchi method [5], In their paper they have selected Taguchi s orthogonal array (OA) under different condition of process Atmiya Institute of Technology and Science Page 10

27 LITERATURE REVIEW parameters. The regression equation is developed for the VMRR & Ra. Optimal combination of parameters is obtain for VMRR & Ra. Jaganjeet Singh, Sanjeev Sharma, Effects of Process Parameters on Material Removal Rate and Surface Roughness in WEDM of P20 Tool Steel [6], In this paper the effect of different process parameters on material removal rate and surface roughness was studied by Taguchi methodology. Also the results of the experimentation were analyzed by MINITAB software analytically as well as graphically. From the results it can be concluded that, MRR increases by increasing pulse on time and peak current. Surface roughness can be decreases by increases wire feed, pulse on time and peak current. Aniza Alias, Bulan Abdullah, Norliana Mohd Abbas, Influence of machine feed rate in WEDM of Titanium Ti-6Al-4V with constant current (6A) using brass wire [7], The objective of this paper is to find effect of different machine feed rates with constant current (6A) on Ti6Al4V. If machine feed rate is increase, the MRR and Kerf width increases. Smoother surface roughness is obtain with low machine feed rate. Surface topography of each experiment is also examined by using microscope. Manoj Malik, Rakesh Kumar Yadav, Deepak Sharma, Manoj, Optimization of process parameters of wire EDM using zinc-coated brass wire [8], In this paper, pulse on time, duty factor and pulse peak current have been taken for optimization of MRR, electrode wear rate and surface roughness using zinc coated brass wire. Grey based Taguchi method is used for the Design of Experiment (DoE). Danial Ghodsiyeh, Mohammadreza Askaripour Lahiji, Mahdi Ghanbari, Mostafa Rezazadeh Shirdar, Abolfazl Golshan, Optimization material removal rate (MRR) in WEDMing Titanium alloy (Ti6Al4V) using the Taguchi method [9], In this paper, Taguchi method is used. The behavior of control parameters pulse on time, pulse off time and peak current on the performance measures like material removal rate (MRR) and surface roughness is studied using ANOVA. Also, mathematical relationship between responses and variables is generated by response surface methodology (RSM). The most significant factor for MRR and surface roughness is peak current. The optimal condition for each parameter is found out. Atmiya Institute of Technology and Science Page 11

28 LITERATURE REVIEW M. T. Antar, S.L. Soo, D. k. Aspinwall, D. Jones, R. Perez, Productivity and workpiece surface integrity when WEDM aerospace alloys using coated wires [10], This study investigates the effect on productivity and surface integrity with the use of coated wires and uncoated wires. Comparison between Cu core coated wires and uncoated brass wire for two workpiece materials Udimet 720 nickel based super alloy and Ti-6Al-2Sn-4Zr-6Mo titanium alloy is given. The productivity of both workpiece material increase significantly with the use of coated wires. The variation in surface roughness is due to the imperfect sparks generated due to erosion of the wire, leads to wider machining gap. C V S Parmeswara Rao and M M M Sarcar, Evalution of optimal parameters for machining brass with wire cut EDM [11], In this paper, the effect of discharge current, voltage at rated wire speed and tension on MRR, surface roughness, cutting speed and spark gap is studied. Workpiece is cut with different wire composition of Cu and Zn. Mathematical relations are developed for cutting speed & workpiece thickness and for spark gap & workpiece thickness which are useful to estimate cutting time. Nihat Tosun, Can Cogun, Gul Tosun, A study on kerf and material removal rate in wire electrical discharge machining based on Taguchi method [12], This paper investigates the effect of machining parameters on the kerf width and MRR by Taguchi experimental design method. ANOVA method is used to find significant parameters and optimum machining parameter combination was obtained by S/N ratio. The objective is to find minimum kerf with maximum MRR. Based on the ANOVA highly effective parameter on kerf and MRR is open circuit voltage & pulse duration, while less effective factors are wire speed and dielectric flushing pressure. Nihat Tosun and Can Cogun, An investigation on wire wear in WEDM [13], This paper investigate the effect of cutting parameter on wire electrode wear. The process parameters selected are pulse duration, open circuit voltage, wire speed and dielectric fluid pressure. ANOVA is also used in this study. It is found that by increasing open circuit voltage and pulse duration, wire wear rate increases. WWR can be decrease by increasing the dielectric fluid pressure & wire speed. Atmiya Institute of Technology and Science Page 12

29 LITERATURE REVIEW 2.3 PROBLEM IDENTIFICATION In wire EDM process cutting conditions such as pulse on time, pulse off time, servo voltage, peak current and other machining parameters should be selected to optimize the economics of machining operations. The cost of numerically controlled machines is high compared to the conventional machines. Therefore it should be operate efficiently. So, it is required to improve responses such as cutting rate and to reduce surface roughness and dimensional deviation. Therefore optimization and parametric study of the cutting condition is required. I have visited Surelia wire-cut Pvt. Ltd. In search for the problem and from that I found that they are facing problems such as low cutting speed and high surface roughness. Therefore optimum set of parameters should be known. 2.4 OBJECTIVES To carry out experiments by using two different wire material To generate the mathematical relationship between input parameters and response To find out optimum values for process parameters by TLBO algorithm Atmiya Institute of Technology and Science Page 13

30 CHAPTER 3 DESIGN OF EXPERIMENTS 3.1 INTRODUCTION TO DESIGN OF EXPERIMENTS Design of experiments (DOE) can be defined as the technique for layout an experimental plan in the most consistent, efficient and statistical way. Design of experiments (DOE) can give the best mixture of the parameters for the process. In industry, DOE can be used to thoroughly examine the process parameters which influence the product quality. After recognizing the process conditions of the product which influence the product quality, it can be obvious to perform design of experiments. Design of experiments should be done in such a way that we can get the maximum information about the process from minimum possible number experiments because resource are limited, and thus it is essential to get the more information from each experiment that is being performed. Well-designed DOE can produce considerably more information and also required fewer numbers of experiments than unplanned or arbitrarily selected experiments. Design of experiments is done by firstly planning the process then selects the input process parameters. After that next step is identify the limiting values of the process parameters and selection of the required levels of the input parameters. 3.2 SELECTION OF PROCESS PARAMETERS Selection of the process parameters is based on the literature survey. For the selection of the levels of parameter different range of the parameter should be known. In this work, five process parameters are selected to study its effects on cutting rate and surface roughness. The ranges of the parameters are shown in the table 3.1 and selected process parameters and their three levels are shown in table 3.2. Atmiya Institute of Technology and Science Page 14

31 DESIGN OF EXPERIMENTS Table 3.1 Range of the parameters No. Parameter Range 1 Pulse-on time (T on ) Machine units 2 Pulse-off time (T off ) 0 63 Machine units 3 Peak current (IP) A 4 Wire tension (W t ) 1 15 Machine units 5 Servo voltage (SV) V Table 3.2 Process parameters with their levels [5] Parameter Level 1 Level 2 Level 3 Pulse-on time (T on ) Pulse-off time (T off ) Peak current (IP) Wire tension (W t ) Servo voltage (SV) Apart from the parameters that have been selected following parameters were kept constant during the experimental procedure. 1. Work material: Titanium grade 5, Ti6Al4V 2. Wire feed: 4 m/min 3. Servo feed: 2050 unit 4. Flushing pressure: 5 Kg/cm 2 5. Workpiece thickness: 4 mm 6. Peak voltage: 2 unit (110 Volt DC) 3.3 DESIGN OF EXPERIMENTS BY TAGUCHI METHOD I have use Minitab 16.0 software for design of experiments. In the Minitab 16.0 software there are several methods available to perform design of experiments like Taguchi design, Factorial design, Response surface method and Mixture design. Taguchi method is simple and the strength of the Taguchi method is that one can change many variables at a time and still have control over the experiments. Atmiya Institute of Technology and Science Page 15

32 DESIGN OF EXPERIMENTS Taguchi s DOE procedure 1. Identifying the process parameters 2. Selection of the levels of process parameters 3. Selection of the appropriate orthogonal array (OA) The Design of experiments are conducted using Taguchi s method in which the number of parameters are five with three levels of each parameter having first level as lower limit and the second level as intermediate value between upper limit and lower limit and the third level is the upper limit for that variable. Thus it will give Taguchi s L27 orthogonal array and this design is as shown in the below Table 3.3 No. Pulse-on time (T on ) Table 3.3 DOE using Taguchi s L27 orthogonal array Pulse-off time (T off ) Peak current (IP) Wire tension (W t ) Servo voltage (SV) Atmiya Institute of Technology and Science Page 16

33 DESIGN OF EXPERIMENTS No. Pulse-on time (T on ) Pulse-off time (T off ) Peak current (IP) Wire tension (W t ) Servo voltage (SV) Atmiya Institute of Technology and Science Page 17

34 CHAPTER 4 EXPERIMENTAL WORK AND RESULTS 4.1 WEDM MACHINE A wire-cut EDM machine (ELEKTRA SPRINTCUT 734) of Electronica Machine Tools Ltd. installed at Surelia wire-cut Pvt Ltd, Rajkot is use for the experimental work. Figure 4.1 shows the WEDM machine tool. The WEDM machine tool has the following specifications (Table 4.1). Figure 4.1 ELEKTRA SPRINTCUT 734 WEDM Machine Atmiya Institute of Technology and Science Page 18

35 EXPERIMENTAL WORK AND RESULTS Table 4.1 Specification of WEDM machine tool Design Table size Max. workpiece height Max. workpiece weight Main table traverse (X,Y) Wire electrode diameter Controlled axes Interpolation Dielectric fluid Dielectric fluid storage capacity Fixed column, moving table 400 x 650 mm 200 mm 500 kg 300, 400 mm 0.25 mm (standard) X, Y and U, V Linear & Circular Distilled water 150 litre 4.2 WORKPIECE INFORMATION Titanium grade 5 (Ti6Al4V) is taken as workpiece material. It has high strength to low weight ratio and outstanding corrosion resistance. Therefore it can be used in many applications. Also Ti6Al4V is most utilized titanium alloy. Table 4.2 Workpiece specification No. Workpiece Specification 1 Material Ti6Al4V Titanium grade 5 2 Grade 5 4 Thickness 4 mm 5 Width and Length 300 mm * 20 mm Al 5.5 to 6.75 % V 3.5 to 4.5 % 6 Chemical composition [17] Fe 0.1 to 0.3 % Mo 0.1 to 0.2 % Mn to % Ti Balance Atmiya Institute of Technology and Science Page 19

36 EXPERIMENTAL WORK AND RESULTS Table 4.3 Material test report Al % V % Mn % Mo % Fe % Cr % Ni % Si % Ti % TYPICAL PROPERTIES OF TITANIUM GRADE 5 ALLOY Table 4.4 shows some typical properties of titanium grade 5 (Ti6Al4V) alloy. Table 4.4 Properties of Titanium grade 5 alloy [18] PROPERTY VALUE Density 4.43 g/cc Tensile Strength, Ultimate 1170 MPa Tensile Strength, Yield 1100 MPa Modulus of Elasticity 114 GPa Shear Strength 760 MPa Thermal Conductivity 6.7 W/m-K Melting Point C Hardness (Brinell) 379 Hardness (Rockwell) APPLICATIONS OF TITANIUM GRADE 5 ALLOY Aerospace industry Biomedical applications such as implants Marine components Cryogenic parts Atmiya Institute of Technology and Science Page 20

37 EXPERIMENTAL WORK AND RESULTS Pressure vessels Steam turbine blades Chemical industry 4.3 WIRE MATERIAL In this work two types of wire materials are used as wire electrode for the WEDM process. 1. Brass wire Brass wire is the most commonly used wire. Brass is an alloy of copper and zinc. Brass wire has powerful combination of low cost, reasonable conductivity, high tensile strength, and improves flushability. The chemical composition of brass wire commonly used is Cu 65% Zn 35%. The diameter of the wire is 0.25 mm. 2. Diffusion zinc coated wire This type of wire is produced by the diffusion annealing process. This type of wire is produced by heat treating the zinc coated wire. Diffusion is the process whereby atoms diffuse from areas of high concentration to areas of lower concentration. The core of wire is made of brass material. The diameter of the wire is 0.25 mm. I. Brass wire II. Diffusion zinc coated wire Figure 4.2 Spools of wire material Atmiya Institute of Technology and Science Page 21

38 EXPERIMENTAL WORK AND RESULTS The experimental work is carried out at Surelia wire-cut Pvt. Ltd. Tagore road, Rajkot. The first 27 experiments was performed by the brass wire of 0.25 mm diameter and another 27 experiments was performed by diffusion zinc coated wire of 0.25 mm diameter, to determine the effects on response due to change of the wire material. The machining plan is to cut material in straight line of 20 mm in length and 10 mm in width. The thickness of the plate is 4 mm. 4.4 EXPERIMENTAL RESULTS After conducting the experiments two responses cutting rate and surface roughness are measured. Cutting rate is shown on the display of the machine. Cutting rate is measure from the machining time and surface roughness (Ra) is measure by Mitutoyo s SJ-201 surface tester machine. Result tables are given below Figure 4.3 Mitutoyo s SJ-201 surface tester machine Atmiya Institute of Technology and Science Page 22

39 EXPERIMENTAL WORK AND RESULTS Figure 4.4 Workpiece material after conducting experiments by Brass wire Figure 4.5 Workpiece material after conducting experiments by diffusion zinc coated wire Atmiya Institute of Technology and Science Page 23

40 EXPERIMENTAL WORK AND RESULTS Table 4.5 Data obtained by conducting experiments by brass wire No. T on T off IP Wt SV Cutting rate (mm/min) Surface roughness Ra (µm) Atmiya Institute of Technology and Science Page 24

41 EXPERIMENTAL WORK AND RESULTS Table 4.6 Data obtained by conducting experiments by diffusion zinc coated wire No. T on T off IP Wt SV Cutting rate (mm/min) Surface roughness Ra (µm) Atmiya Institute of Technology and Science Page 25

42 Mean Mean EXPERIMENTAL WORK AND RESULTS 4.5 MAIN EFFECT PLOTS The main effect plots are generated with the help of MINITAB 16 software. The main effect plot shows the relation of the parameters with responses MAIN EFFECT PLOTS FOR BRASS WIRE Main Effects Plot for CR Data Means Ton Toff IP Wt SV Figure 4.6 Main effect plots for Cutting rate (Brass wire) Main Effects Plot for SR Data Means 2.6 Ton Toff IP Wt SV Figure 4.7 Main effect plots for Surface roughness (Brass wire) Atmiya Institute of Technology and Science Page 26

43 Mean Mean EXPERIMENTAL WORK AND RESULTS Figure 4.6 and 4.7 shows the main effect plots for data means values for cutting rate and surface roughness respectively for the brass wire material. From figure 4.6, it can be seen that by increasing the value of pulse on time and peak current, cutting rate increases. The surface finish decreases with increasing the value of peak current and pulse on time MAIN EFFECT PLOTS FOR DIFFUSION ZINC COATED WIRE Main Effects Plot for CR Data Means Ton Toff IP Wt SV Figure 4.8 Main effect plots for Cutting rate (Diffusion zinc coated wire) Main Effects Plot for SR Data Means 2.4 Ton Toff IP Wt SV Figure 4.9 Main effect plots for Surface roughness (Diffusion zinc coated wire) Atmiya Institute of Technology and Science Page 27

44 EXPERIMENTAL WORK AND RESULTS Figure 4.8 and 4.9 shows the main effect plots for data means values for cutting rate and surface roughness respectively for the diffusion zinc coated wire material. Surface finish decreases with increase in pulse on time and peak current. Cutting rate is highly affected by pulse on time and peak current. 4.6 ANALYSIS OF RESULTS Table 4.7 Result table Response Unit Brass wire Diffusion zinc coated wire Min Max Min Max Cutting rate mm/min Surface Roughness µm From above table we can conclude the following things. Cutting rate: The value of the cutting rate is very important for any material to be cut rapidly. Brass wire gives highest cutting rate value of mm/min from the combination of input parameters. Diffusion zinc coated wire gives highest cutting rate value of mm/min from the combination of input parameters. Thus diffusion zinc coated wire material is the most desirable wire material for maximum cutting rate. Surface roughness: the surface roughness should minimum for any material to get better surface finish. Maximum surface roughness value for brass wire and diffusion zinc coated wire material is 2.82 µm and 2.65 µm respectively. Minimum surface roughness value for brass wire and diffusion zinc coated wire material is 1.93 µm and 1.33 µm. Thus diffusion zinc coated wire is desirable to use. Atmiya Institute of Technology and Science Page 28

45 CHAPTER 5 REGRESSION ANALYSIS 5.1 DEFINITION Regression analysis is statistical process which gives selection between the input parameters and output. Regression analysis gives output in such a way to fit the linear, quadratic, cubic, polynomial and non-linear curves. Regression analysis gives the correlation among the dependent variable and independent variable. Regression analysis is generally used for estimation and forecasting. It forms the equation from the given parameters LINEAR REGRESSION In linear regression type the output function is fitted with the equation of line. In this all inputs are given a coefficient to fit the equation. In this normal approach to fit the curves to the data by using the higher order polynomial term of square or cubic terms. We can choose the model order as per our requirements and which is best fit. By the use of squared and more exponent terms increase the bend in the curved fitted line. It is exceptionally uncommon to utilize more than a cubic term in general. For example linear regression equation can be written as given below. In which Y is dependent variable and B 0, B 1 B N are coefficient and X 1, X 2 X n are predictors. Y= B 0 + B 1 X 1 +B 2 X B n X n Figure 5.1 Model order and their generated regression line with data points Atmiya Institute of Technology and Science Page 29

46 REGRESSION ANALYSIS NON-LINEAR REGRESSION In regression analysis, non-linear regression analysis is method in which data are modeled by a function which is non-linear regression is done by the method of successive approximation. Non-linear regression produces an equation relationship among a response and prediction variables, and forecast new observation data and gives new equation. Non-linear regression can be used instead of linear regression because it gives the most flexible curve fitting functionality. Non-linear regression uses an algorithm to give best fit model step by step. 5.2 REGRESSION ANALYSIS EQUATIONS The regression analysis can be done using Microsoft office Excel and also by the Minitab software. In this work, Minitab 16.0 software is use for linear regression SIMPLE LINEAR REGRESSION ANALYSIS I have used linear regression analysis equation. On giving data table achieved from the experiments to the Minitab software it gives the following results. Atmiya Institute of Technology and Science Page 30

47 Frequency Residual Percent Residual REGRESSION ANALYSIS For Brass wire material, I. Regression Analysis: Cutting rate versus Pulse-on time (T on ), Pulse-off time (T off ), Peak current (IP), Wire tension (W t ) and Servo voltage (SV) The regression equation is: Cutting Rate = Ton Toff IP Wt SV Predictor Coef SE Coef T P Constant T on T off IP W t SV S = R-Sq = 92.0% R-Sq(adj) = 90.2% Residual Plots for CR Normal Probability Plot Versus Fits Residual Fitted Value 3.0 Histogram Versus Order Residual Observation Order Figure 5.2 Residual Plots for Cutting rate for linear regression (Brass wire) Atmiya Institute of Technology and Science Page 31

48 Frequency Residual Percent Residual REGRESSION ANALYSIS II. Regression Analysis: Surface roughness versus Pulse-on time (T on ), Pulse-off time (T off ), Peak current (IP), Wire tension (W t ) and Servo voltage (SV) The regression equation is: Surface Roughness = Ton Toff IP Wt SV Predictor Coef SE Coef T P Constant T on T off IP W t SV S = R-Sq = 90.9% R-Sq(adj) = 88.7% Residual Plots for SR Normal Probability Plot Versus Fits Residual Fitted Value Histogram Versus Order Residual Observation Order Figure 5.3 Residual Plots for Surface roughness for linear regression (Brass wire) Atmiya Institute of Technology and Science Page 32

49 Frequency Residual Percent Residual REGRESSION ANALYSIS For Diffusion zinc coated wire material, I. Regression Analysis: Cutting rate versus Pulse-on time (T on ), Pulse-off time (T off ), Peak current (IP), Wire tension (W t ) and Servo voltage (SV) The regression equation is: Cutting Rate = Ton Toff IP Wt SV Predictor Coef SE Coef T P Constant T on T off IP W t SV S = R-Sq = 93.4% R-Sq(adj) = 91.8% Residual Plots for CR Normal Probability Plot Versus Fits Residual Fitted Value 5 Histogram Versus Order Residual Observation Order Figure 5.4 Residual Plots for Cutting rate for linear regression (Diffusion zinc coated wire) Atmiya Institute of Technology and Science Page 33

50 Frequency Residual Percent Residual REGRESSION ANALYSIS II. Regression Analysis: Surface roughness versus Pulse-on time (T on ), Pulse-off time (T off ), Peak current (IP), Wire tension (W t ) and Servo voltage (SV) The regression equation is: Surface Roughness = Ton Toff IP Wt SV Predictor Coef SE Coef T P Constant T on T off IP W t SV S = R-Sq = 91.7% R-Sq(adj) = 89.7% Residual Plots for SR Normal Probability Plot Versus Fits Residual Fitted Value 2.50 Histogram Residual Versus Order Observation Order Figure 5.5 Residual Plots for Surface roughness for linear regression (Diffusion zinc coated wire) In this residual plot for regression analysis four graphs are shown. Residual versus fitted value graph shows the random pattern of the residuals on the both side of 0. Atmiya Institute of Technology and Science Page 34

51 REGRESSION ANALYSIS Histogram shows the general characteristics of the residual including spread and shape. A normal probability plot shows the residual values with regression line. Residual versus observation order shows all residuals in the order that the data was collected DATA OBTAINED FROM REGRESSION EQUATION The regression equation generated is used to form theoretical data and the data table generated using the regression analysis is shown as below. No. Table 5.1 Theoretical data from regression equation For brass wire For diffusion zinc coated wire Cutting rate (mm/min) Surface roughness (µm) Cutting rate (mm/min) Surface roughness (µm) Atmiya Institute of Technology and Science Page 35

52 CHAPTER 6 OPTIMIZATION USING TLBO 6.1 INTRODUCTION TO TLBO TLBO is a teaching-learning motivated algorithm based on effect of influence of a teacher on the output of learners in a class. Teaching-learning Based optimization (TLBO) is nature based optimization algorithm and solve various optimization problems efficiently. In TLBO, there are mainly two phases, one is teacher phase and another is learner phase. Output is taken in terms of results or grades. The teacher is taken as highly learned person. Teacher gives his or her knowledge to learners. The learner s outcome is depending on the quality of the teacher. Therefore good teacher trains learners for the good results in form of marks or grades. [3] TLBO is population based method. The global solution is obtained from a population of solutions. In TLBO, population is taken as group of learners. 6.2 TEACHING LEARNING BASED OPTIMIZATION Consider two different teachers as T 1 and T 2. They teach a subject of same content, in two different classes of same level learners. Figure 6.1 shows the distribution of marks obtained by the learners. In figure 6.1, curve 1 shows the marks obtained by the learners taught by T 1 and curve 2 shows marks obtained by learners taught by T 2. The normal distribution [14] is assumed for the marks gained. The normal distribution, ( ) ( ) Here, σ 2 = variance, σ = mean, X = value for which normal distribution required Atmiya Institute of Technology and Science Page 36

53 OPTIMIZATION USING TLBO Figure 6.1 Distribution of marks obtained by learners taught by two different teachers [14] As we can see in figure 6.1, the mean produces by T 2 is better than teacher T 1. The main difference among both results is their mean. Therefore it can be said that teacher T 2 is superior to teacher T 1. Now, from the above teaching process, a mathematical model is developed and applied for the optimization problem. Figure 6.2 shows, two curve A and curve B. M A is the mean of marks gained by learners in a class. The teacher is the most knowledgeable person in the society. Therefore best learner is turned as a teacher, as shown in T A in figure 6.2. The teacher tries to spread knowledge between learners. It will increase the knowledge level of whole class. Therefore teacher T A will try to move mean M A near their own level as per the capabilities. Therefore level of learners is increasing to a new mean M B. The quality of the students also affects the results. The quality of the students is evaluated from the mean value. Because of the efforts from the teacher T A, the mean value changes from M A to M B. Now students require a new teacher of better quality than themselves, as teacher T B, in curve B. The flowchart of the TLBO algorithm is shown in figure 6.3. Atmiya Institute of Technology and Science Page 37

54 OPTIMIZATION USING TLBO Figure 6.2 Model for the distribution of marks obtained for a group of learners [14] TEACHER PHASE A good teacher increases the level of learners up to his or her level in terms of knowledge. But it is depending on the level of learners. This follows a random process depending on many factors. Now, M i is the mean and T i is the teacher at any iteration i. T i will try to move M i near its own level. Therefore new mean will be T i labeled as M new. The solution is updated according to the difference among the existing and the new mean is Difference_Mean i = r i (M new T F M i ) Here, T F = teaching factor that decides the value of mean to be changed r i = random number in the range [0,1] The value of T F can be either 1 or 2, which is again a heuristic step and decided randomly with same probability as, T F = round [1+ rand (0, 1) {2-1}] Now, this difference modifies the existing result according to the below expression. X new,i = X old,i + Difference_Mean i Atmiya Institute of Technology and Science Page 38

55 OPTIMIZATION USING TLBO ` Initialize number of students (population), termination Calculate the mean design variables Identify the best solution (teacher) Modify solution based on best solution X new =X old +r(x teacher -(T f )Mean) Teacher Phase Reject No Is new solution better than the Yes Accept Select any two solutions randomly X i and X j, Yes Is X i better than X j No X new =X old +r(x i -X j ) Χ new =X old +r(x j -X i ) Student Phase Reject No Is new solution better than Yes Accept Is termination criteria satisfied? Yes No Final value of solution Figure 6.3 Flow chart of the working of TLBO algorithm [14] Atmiya Institute of Technology and Science Page 39

56 OPTIMIZATION USING TLBO LEARNER PHASE In this phase, the knowledge of learners is increases by teacher and also by the interaction among themselves. A learner will gain knowledge if the other learner has more knowledge than him or her. The learners phase is expressed as below. For i = 1:Pn Randomly select two learners X i and X j, where, i j If f (X i ) < f (X j ) X new,i = X old,i + r i (X i X j ) Else X new,i = X old,i + r i (X j X i ) End If End For Accept X new if it gives a better function value STEPS TO IMPLEMENT TLBO ALGORITHM The step by step procedure for the implementation of TLBO algorithm is given as below.[ ] Step 1: State the optimization problem and initialize the optimization parameters. Population size (Pn), number of generations (G n ), number of design variables (Dn), and limits of design variables (U L, L L ) are given. Define the optimization problem as: Minimize f(x) or Maximize f(x) Subject to X i x i = 1, 2.., Dn Where f (X) is the objective function, X is a vector for design variables such that LL,i x,i UL,i. Step 2: Initialize the population As per the population size and the number of design variable, a random population is produce. Population size specifies the number of learners. Design variables shows the subject offered. Atmiya Institute of Technology and Science Page 40

57 OPTIMIZATION USING TLBO [ ] Step 3: Teacher phase Calculate the mean of the population column wise. It gives the mean for the subject as M,D = [ m 1, m 2,,m D ] The best solution will act as a teacher for that iteration X teacher = X f (X) = min. The teacher will try to shift the mean from M,D towards X,teacher, which will act as a new mean for the iteration. So, M _new,d = Xteacher,D The difference between two means is expressed as Difference,D = r (M_new,D T F M,D ). The value of T F is selected as 1 or 2. The obtained difference is added to the current solution to update its values by X new,d = X old,d + Difference. Accept X new if it gives better function value. Step 4: Learner phase Learners increase their knowledge by interactions. The explanation for the learner phase is given in the section Step 5: Termination phase Stop if the maximum generation number is achieved; otherwise repeat from Step OPTIMIZATION Optimization can be defined as process of selecting optimum values of variables which gives the best suitable values of objective function. Optimization can be of single objective or multi-objective type. The optimization process is can be of minimization type or maximization type. Atmiya Institute of Technology and Science Page 41

58 OPTIMIZATION USING TLBO In this project work there are five variables namely pulse on time, pulse off time, peak current, wire tension and servo voltage. The outputs are cutting rate and surface roughness. Weight method is used for the multi-objective optimization. Cutting rate and surface roughness are the two different sub-objectives. The function is normalized to solve multi-objective problem MULTI-OBJECTIVE FUNCTION AND LIMIT OF VARIABLES I. For brass wire material, The first objective is to maximize the cutting rate. The equation for the cutting rate is CR 1 = Ton Toff IP Wt SV The second objective is to minimize the surface roughness. The equation for the surface roughness is SR 1 = Ton Toff IP Wt SV The above given single objective functions are mentioned together for multiobjective optimization. The normalized multi-objective function (Z) is formulated by giving weight factors in equation. [15] Maximize Z 1 = w (CR 1 / CR 1,max ) (1-w) (SR 1 / SR 1,max ) Here w = weight factor for the equation. CR 1,max and SR 1,max are the maximum and minimum values of the objective functions CR 1 and SR 1 respectively. II. For diffusion zinc coated wire material, The first objective is to maximize the cutting rate. The equation for the cutting rate is CR 2 = Ton Toff IP Wt SV Atmiya Institute of Technology and Science Page 42

59 OPTIMIZATION USING TLBO The second objective is to minimize the surface roughness. The equation for the surface roughness is SR 2 = Ton Toff IP Wt SV The normalized multi-objective function (Z) is, Maximize Z 2 = w (CR 2 / CR 2,max ) (1-w) (SR 2 / SR 2,max ) Here w = weight factor for the equation. CR 2,max and SR 2,max are the maximum and minimum values of the objective functions CR 2 and SR 2 respectively. The limits of the variable parameters are given as below, 105 T on T off IP W t SV 80 The multi-objective function is not depending on any constraints. Therefore it is of unconstrained type problem and for that the codes are given to the MATLAB for the optimization using TLBO algorithm. w= 0.6 Number of iterations= 500 Population= 35 The code is applied in MATLAB and modifications and some changes are done as per the problem and objective function RESULTS AFTER OPTIMIZATION USING TLBO ALGORITHM IN MATLAB After completing 500 iterations the optimum values of the multi-objective for the variables are given below. The population is taken as 35 in the program. The weight Atmiya Institute of Technology and Science Page 43

60 OPTIMIZATION USING TLBO factor is considered as per the requirement. Here the value of w is considered as 0.6. Figure 6.4 and 6.5 shows the values for the responses at 500 runs for brass wire and diffusion zinc coated wire respectively. Figure 6.4 Result values of the responses v/s iterations for brass wire Figure 6.5 Result values of the responses v/s iterations for diffusion zinc coated wire Atmiya Institute of Technology and Science Page 44

61 OPTIMIZATION USING TLBO Optimum values after computation using program are as below. For brass wire T on = T off = IP = W t = 5 SV = = 50 Cutting rate, CR 1 = 3.37 mm/min Surface roughness, SR 1 = 2.69 µm For diffusion zinc coated wire material, T on = T off = IP = W t = 15 SV = 80 Cutting rate, CR 2 = 5.35 mm/min Surface roughness, SR 2 = 2.62 µm The above given values are the final values for the multi-objective optimization of cutting rate and surface roughness for the given set of the input variables. Atmiya Institute of Technology and Science Page 45

62 CHAPTER 7 CONCLUSION AND FUTURE SCOPE 7.1 Conclusion The optimum value for the process parameters is obtained from the TLBO algorithm in MATLAB for the both wire material. The values are confirmed by experiments carried out at industry. The experimental values and theoretical values vary slightly for the objective function. From the experimental and research work, the best suited values of process parameters are T on = machine unit, T off = machine unit, IP = A, W t = 5 machine unit, SV = 50 V for brass wire material. For diffusion zinc coated wire material the values of process parameters are T on = machine unit, T off = machine unit, IP = A, W t = 15 machine unit, SV = 80 V. The diffusion zinc coated wire material gives the best result for the cutting rate and surface roughness. Thus it should use for the WEDM process of the Ti6Al4V material. 5.2 Future Scope In this work, parametric study is carried out for Ti6Al4V material and also optimization is done by TLBO algorithm. Pulse on time, pulse off time, peak current, wire tension and servo voltage are taken as input process parameters for the study on cutting rate and surface roughness and also two types of wire materials are used. The future scope in this process includes other parameters such as wire diameter, different type of wire material, flushing pressure, wire feed rate, conductivity of dielectric are to be investigated. The weight assigned to the objective function in optimization depends on the requirements of the industry. Other optimization algorithm also can be used. Atmiya Institute of Technology and Science Page 46

63 REFERENCES Research papers: [1] Sonu Dhiman, Ravinder Chaudhary, V.K. Pandey, Analysis and study the effects of various control factors of CNC-wire cut EDM for S7 steel MEIJ, Vol. 1, No. 1, May 2014, page no [2] Manpreet Singh, Harvinder Lal, Ramandeep Singh, Recent Developments in Wire EDM: A Review IJRMET, Vol. 3, Issue 2, May-October 2013, page no [3] R. Venkata Rao, V. D. Kalyankar, Parameter optimization of modern machining processes using teaching-learning-based optimization algorithm Elsevier, Engineering applications of artificial intelligence 26, 2013, page no [4] Kuriachen Basil, Dr. Josephkunju Paul, Dr. Jeoju M. Issac, Spark gap optimization of WEDM process on Ti6Al4V IJESIT, Vol. 2, Issue 1, January 2013, page no [5] Lokeswara Rao T. and N. Selvaraj, Optimization of WEDM process parameters on titanium alloy using Taguchi method IJMER, Vol. 3, Issue 4, July-August 2013, page no [6] Jaganjeet Singh, Sanjeev Sharma, Effects of Process Parameters on Material Removal Rate and Surface Roughness in WEDM of P20 Tool Steel International Journal of Multidisciplinary and Current Research, December 2013, ISSN: [7] Aniza Alias, Bulan Abdullah, Norliana Mohd Abbas, Influence of machine feed rate in WEDM of Titanium Ti-6Al-4V with constant current (6A) using brass wire Procedia Engineering 41 (2012) [8] Manoj Malik, Rakesh Kumar Yadav, Deepak Sharma, Manoj, Optimization of process parameters of wire EDM using zinc-coated brass wire IJATER, Vol. 2, Issue 4, July 2012, ISSN No: [9] Danial Ghodsiyeh, Mohammadreza Askaripour Lahiji, Mahdi Ghanbari, Mostafa Rezazadeh Shirdar, Abolfazl Golshan, Optimization material removal rate (MRR) in Atmiya Institute of Technology and Science Page 47

64 REFERENCES WEDMing Titanium alloy (Ti6Al4V) using the Taguchi method Research journal of applied science, engineering and technology, September 2012, ISSN: [10] M. T. Antar, S.L. Soo, D. k. Aspinwall, D. Jones, R. Perez, Productivity and workpiece surface integrity when WEDM aerospace alloys using coated wires 1 st CIRP conference on surface integrity (CSI), Procedia Engineering 19 (2011) 3-8 [11] C V S Parmeswara Rao and M M M Sarcar, Evalution of optimal parameters for machining brass with wire cut EDM Journal of scientific & industrial research, Vol. 68, January 2009, page no [12] Nihat Tosun, Can Cogun, Gul Tosun, A study on kerf and material removal rate in wire electrical discharge machining based on Taguchi method Journal of Materials Processing Technology 152, April 2004, page no [13] Nihat Tosun and Can Cogun, An investigation on wire wear in WEDM Journal of materials processing technology 134, October 2003, page no [14] R.V.Rao, V.J. Savsani, D.P. Vakharia, Teaching Learning-Based Optimization: An optimization method for continuous non-linear large scale problems Information Sciences 183, 2012, page no [15] R. Venkata Rao, Vivek Patel, Multi-objective optimization of heat exchangers using a modified teaching-learning-based optimization algorithm Applied Mathematical Modelling 37, 2013, page no Dissertation: [16] Rohit Garg, PhD thesis, Effect of process parameters on performance measures of wire electrical discharge machining National Institute of Technology, Kurukshetra, MAY-2010 Websites: [17] Arcam Ti6Al4V titanium alloy, [18] ASM material data sheet, Atmiya Institute of Technology and Science Page 48

65 TLBO Codes for brass wire: APPENDIX A LIST OF TLBO CODE USED function [CROut,SROut,XOut]= multi_brass_tlbo(w) %General Variables restrictions=zeros(5,1); iteration=500; design_variables=5; class_size = 35; %Dat Ton = ( )*rand(class_size,1); Toff = 14 + (52-14)*rand(class_size,1); IP = 70 + (230-70)*rand(class_size,1); Wt = 5 + (15-5)*rand(class_size,1); SV = 20 + (80-20)*rand(class_size,1); UL= [ ]; LL= [ ]; CROut= []; SROut= []; %Let's start the TLBO lastx=0; Xold=[]; Xnew=zeros(class_size,design_variables); X = [Ton, Toff, IP, Wt, SV]; firstx=x; i=1; j=1; while i <= iteration, Mx= mean(x);%mean Mx=Mx-std(X); CR=f1(X); SR=f2(X); Z= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); [Z,ind]=sort(Z,1,'ascend'); X=X(ind,:); [~,index]=max(z); % with initial values we get Objective function %making teacher Xteacher= X(index,:); Tf = round (1+rand()); r = rand(); Atmiya Institute of Technology and Science Page 49

66 LIST OF TLBO CODE USED difference=r*(xteacher - Tf*(Mx)); %Obtaining Xnew % Xnew=X + difference; Xnew(:,1)=X(:,1)+difference(1); Xnew(:,2)=X(:,2)+difference(2); Xnew(:,3)=X(:,3)+difference(3); Xnew(:,4)=X(:,4)+difference(4); Xnew(:,5)=X(:,5)+difference(5); %Teacher's proving Xnew vs Xold CrNew=f1(Xnew); SrNew=f2(Xnew); ZNew= (W*(CrNew/max(CrNew)))-((1-W)*(SrNew/min(SrNew))); %X = [Ton, Toff, IP, Wt, SV]; j=1; for j=1:class_size if (Z(j) <= ZNew(j)) %Here we compare the new versus the max X(j,:)=Xnew(j,:); Z(j)= ZNew(j); %If new is better end X(j,:)=max(X(j,:),LL); X(j,:)=min(X(j,:),UL); end %End Teacher's part changed %student part % Obtaining a new objective function, because it CR=f1(X); SR=f2(X); Z= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); % We will compare two elements from C in order to increase knoledge % between 'students' for j = 1:class_size, for jj=1:class_size, if j ~= jj % j = class_size Atmiya Institute of Technology and Science Page 50

67 LIST OF TLBO CODE USED Xold = X(j,:); if Z(j)>Z(jj), X(j,1)=X(j,1)+rand()*(X(j,1)-X(jj,1)); X(j,2)=X(j,2)+rand()*(X(j,2)-X(jj,2)); X(j,3)=X(j,3)+rand()*(X(j,3)-X(jj,3)); X(j,4)=X(j,4)+rand()*(X(j,4)-X(jj,4)); X(j,5)=X(j,5)+rand()*(X(j,5)-X(jj,5)); Else X(j,1)=X(j,1)+rand()*(X(jj,1)-X(j,1)); X(j,2)=X(j,2)+rand()*(X(jj,2)-X(j,2)); X(j,3)=X(j,3)+rand()*(X(jj,3)-X(j,3)); X(j,4)=X(j,4)+rand()*(X(jj,4)-X(j,4)); X(j,5)=X(j,5)+rand()*(X(jj,5)-X(j,5)); end CR=f1(Xold); SR=f2(Xold); Zold= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); CR=f1(X(j,:)); SR=f2(X(j,:)); Znew= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); if Znew >= Zold X(j,:)=Xold; end % X(j,:)=max(X(j,:),LL); % X(j,:)=min(X(j,:),UL); end end %we will remove repeated values in order to avoid locals maximums CR=f1(X); SR=f2(X); Z= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); for j=1:class_size, for jj=j+1:class_size, if isequal(z(j),z(jj)), X(j,1) = ( )*rand(); X(j,2) = 14 + (52-14)*rand(); X(j,3) = 70 + (230-70)*rand(); Atmiya Institute of Technology and Science Page 51

68 LIST OF TLBO CODE USED X(j,4) = 5 + (15-5)*rand(); X(j,5) = 20 + (80-20)*rand(); end end end % Obtaining a new objective function, because it changed i=i+1; end CR=f1(X); SR=f2(X); Z= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); [~,index]=max(z); %CROut= [CROut max(f1(x))]; %SROut= [SROut min(f2(x))]; end CR=f1(X); SR=f2(X); Z= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); [~,index]=max(z); CROut=CR(index); SROut=SR(index); save('vars.mat') XOut=X(index,:); return %X = %For Brass wire material, function [CR]=f1(X) CR =-5.89+(0.0647*X(:,1)) + ( *X(:,2)) + ( *X(:,3)) - (0.0320*X(:,4)) - ( *X(:,5)); return function [SR]=f2(X) SR = (0.0252*X(:,1)) - ( *X(:,2)) + ( *X(:,3))-( *X(:,4))+ ( *X(:,5)); return Atmiya Institute of Technology and Science Page 52

69 LIST OF TLBO CODE USED TLBO Codes for diffusion zinc coated wire: function [CROut,SROut,XOut]= multi_zinc_tlbo(w) %General Variables restrictions=zeros(5,1); iteration=500; design_variables=5; class_size = 35; %Dat Ton = ( )*rand(class_size,1); Toff = 14 + (52-14)*rand(class_size,1); IP = 70 + (230-70)*rand(class_size,1); Wt = 5 + (15-5)*rand(class_size,1); SV = 20 + (80-20)*rand(class_size,1); UL= [ ]; LL= [ ]; CROut= []; SROut= []; %Let's start the TLBO lastx=0; Xold=[]; Xnew=zeros(class_size,design_variables); X = [Ton, Toff, IP, Wt, SV]; firstx=x; i=1; j=1; while i <= iteration, Mx= mean(x);%mean Mx=Mx-std(X); CR=f1(X); SR=f2(X); Z= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); [Z,ind]=sort(Z,1,'ascend'); X=X(ind,:); [~,index]=max(z); % with initial values we get Objective function %making teacher Xteacher= X(index,:); Tf = round (1+rand()); r = rand(); difference=r*(xteacher - Tf*(Mx)); %Obtaining Xnew % Xnew=X + difference; Atmiya Institute of Technology and Science Page 53

70 LIST OF TLBO CODE USED Xnew(:,1)=X(:,1)+difference(1); Xnew(:,2)=X(:,2)+difference(2); Xnew(:,3)=X(:,3)+difference(3); Xnew(:,4)=X(:,4)+difference(4); Xnew(:,5)=X(:,5)+difference(5); %Teacher's proving Xnew vs Xold CrNew=f1(Xnew); SrNew=f2(Xnew); ZNew= (W*(CrNew/max(CrNew)))-((1-W)*(SrNew/min(SrNew))); %X = [Ton, Toff, IP, Wt, SV]; j=1; for j=1:class_size if (Z(j) <= ZNew(j)) %Here we compare the new versus the max X(j,:)=Xnew(j,:); Z(j)= ZNew(j); %If new is better end end X(j,:)=max(X(j,:),LL); X(j,:)=min(X(j,:),UL); % End Teacher's part %student part % Obtaining a new objective function, because it changed CR=f1(X); SR=f2(X); Z= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); % We will compare two elements from Z in order to increase knowledge % between 'students' for j = 1:class_size, rr=rand(); for jj=1:class_size, if j ~= jj % j = class_size Xold = X(j,:); if Z(j)>Z(jj), X(j,1)=X(j,1)+rr*(X(j,1)-X(jj,1)); X(j,2)=X(j,2)+rr*(X(j,2)-X(jj,2)); X(j,3)=X(j,3)+rr*(X(j,3)-X(jj,3)); X(j,4)=X(j,4)+rr*(X(j,4)-X(jj,4)); Atmiya Institute of Technology and Science Page 54

71 LIST OF TLBO CODE USED X(j,5)=X(j,5)+rr*(X(j,5)-X(jj,5)); else X(j,1)=X(j,1)+rr*(X(jj,1)-X(j,1)); X(j,2)=X(j,2)+rr*(X(jj,2)-X(j,2)); X(j,3)=X(j,3)+rr*(X(jj,3)-X(j,3)); X(j,4)=X(j,4)+rr*(X(jj,4)-X(j,4)); X(j,5)=X(j,5)+rr*(X(jj,5)-X(j,5)); end CR=f1(Xold); SR=f2(Xold); Zold= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); CR=f1(X(j,:)); SR=f2(X(j,:)); Znew= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); if Znew >= Zold X(j,:)=Xold; end % X(j,:)=max(X(j,:),LL); % X(j,:)=min(X(j,:),UL); end end %we will remove repeated values in order to avoid locals maximums CR=f1(X); SR=f2(X); Z= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); for j=1:class_size, rr=rand(); for jj=j+1:class_size, if isequal(z(j),z(jj)), X(j,1) = ( )*rr; X(j,2) = 14 + (52-14)*rr; X(j,3) = 70 + (230-70)*rr; X(j,4) = 5 + (15-5)*rr; X(j,5) = 20 + (80-20)*rr; end end end Atmiya Institute of Technology and Science Page 55

72 LIST OF TLBO CODE USED % Obtaining a new objective function, because it changed i=i+1; end CR=f1(X); SR=f2(X); Z= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); [~,index]=max(z); %CROut= [CROut max(f1(x))]; %SROut= [SROut min(f2(x))]; end CR=f1(X); SR=f2(X); Z= (W*(CR/max(CR)))-((1-W)*(SR/min(SR))); [~,index]=max(z); CROut=CR(index); SROut=SR(index); save('vars.mat') XOut=X(index,:); return %X = %For zinc wire material, function [CR]=f1(X) CR = (0.100*X(:,1)) + ( *X(:,2)) + ( *X(:,3)) + (0.0541*X(:,4)) - ( *X(:,5)); return function [SR]=f2(X) SR = (0.0338*X(:,1)) + ( *X(:,2)) + ( *X(:,3))-( *X(:,4))-( *X(:,5)); return Problem file: %Now multi has the ability of change W factor (0<= W <=1) clear all close all [CR1Out,SR1Out,X1Out]= multi_brass_tlbo(0.6); [CR2Out,SR2Out,X2Out]= multi_zinc_tlbo(0.6); disp('this vector') X1Out % Vector Atmiya Institute of Technology and Science Page 56

73 LIST OF TLBO CODE USED disp('optimizes the objective functions CR and SR for brass') format bank disp('cr: ') CR1Out disp('sr: ') SR1Out disp('this vector') X2Out % Vector disp('cr: ') CR2Out disp('sr: ') SR2Out Atmiya Institute of Technology and Science Page 57

74 APPENDIX B REVIEW CARD Atmiya Institute of Technology and Science Page 58

75 REVIEW CARD Atmiya Institute of Technology and Science Page 59

76 REVIEW CARD Atmiya Institute of Technology and Science Page 60

77 REVIEW CARD Atmiya Institute of Technology and Science Page 61

78 REVIEW CARD Atmiya Institute of Technology and Science Page 62

79 REVIEW CARD Atmiya Institute of Technology and Science Page 63

80 APPENDIX C COMPLIANCE REPORT Comments during Dissertation Phase-1 No. Comments given by External Examiners 1. Topic and its treatment are very common. There is nothing new in the project. 2. Modify the topic. Change the process parameter to e.g. wire material, wire diameter, wire coating, dielectric velocity, wire tension etc. Modification done based on comments For more research, two wire materials are selected and also optimization technique will be applied for same. Topic has been modified. Process parameters have been changed to wire material, wire tension, servo voltage etc. Comments during Mid semester review No. Comments given by External Examiners 1. Understanding of the problem is very poor. 2. Basic fundamentals are not upto the level. 3. Candidate should perform the work by themselves. Modification done based on comments Reviewed various research papers regarding this problem. Basic fundamental is studied and understood. The experiments are performed by me with help of industrial guide. Atmiya Institute of Technology and science Page 64

81 APPENDIX D PAPER PUBLICATION CERTIFICATE Atmiya Institute of Technology and Science Page 65