Parametric optimization of process parameters for Electric discharge Machining of Tungsten carbide (93% WC and 7%Co)

Abstract Nowadays there is a huge demand of High Strength Temperature Resistance (HSTR) alloys such as titanium, carbide, nimonics and ceramics in aerospace, defence and electronics. Among these alloys machining of tungsten carbide alloy is of interest, because of its numerous applications. Complex shapes of tungsten carbide are not generally made by traditional manufacturing process. To machine tungsten carbide with high accuracy, non-traditional machining process like Laser beam machining, Electron beam machining and Electrical discharge machining are a proper choice. In the present paper, the authors have machined Tungsten carbide (93% WC and 7%Co) with copper electrode. The machining is performed on EDM MODEL 500 X 300 ENC with VELVEX EDMVEL-2 as dielectric oil. The 17 experiments are carried out based on RSM (Box-Behnken) method. Further, in order to find the optimum combination grey relational approach is used. The results showed that pulse-on-time of 40μs, pulse-off-time of 2μs and current of 8A are optimum combination for machining of Tungsten carbide (93% WC and 7%Co). Lastly, the confirmation experiment has been conducted.


Introduction
With the increase in demand for complex and intricate shapes made from alloys of high strength and hardness have a wide variety of applications in industry, aerospace, marine, surgeries equipment etc. (Kumar et al., 2010;Singh 2016, 2018). These hybrid materials include stainless steel, tungsten carbide alloys, ceramics, glass etc. Stainless steel alloys find its applications in various fields such as power generation, medical applications, automotive industry, architecture and building construction. Similarly, other hybrid materials are also used in various fields. Machining of such alloys which are difficult to machine is a very important problem in industries nowadays. The demand for these newly developed materials has increased for the last 40 years in different engineering applications because of its outstanding material properties such as high strength, toughness, hardness, temperature strength, oxidation resistance etc. Machining of these materials are faced with severe difficulties because of high coefficient of expansion, work hardening rate, low thermal conductivity and high ductility etc. Low thermal conductivity property of material causes high temperature at the tool surface which is in contact with the surface of work-piece which often leads to reduction in tool life, surface finish and tolerances. Making complex and intricate shapes is very difficult to achieve with traditional machining methods. Chip formation is yet another phenomenon which is mostly affected because of its low thermal conductivity and high work hardening capability. High ductility property of some material has caused formation of built-up edges on the cutting tool and finally leads to chattering of tool.
It is hard to find a suitable tool material for machining these new materials using conventional machining methods. Therefore, a need for new machining methods known as non-traditional machining method raised for those, who would machine hybrid materials like ceramics, tungsten and its alloys, high strength polymers, stainless steel and other alloys. These nontraditional machining methods are capable of machining wide range of difficult machine materials disregard of their hardness. There are dozens of non-conventional machining process; most of them are having same applications. Therefore, non-traditional machining processes are often classified according to the energy used during machining (Groover, 2002). Table 1 presents the classification of non-traditional machining processes (Pandey and Shan, 2001). Abrasive jet machining (AJM), Water Jet machining (WJM), Ultrasonic machining (USM), Abrasive water jet machining (AWJM) are the processes under mechanical processes. Electrochemical machining (ECM), Electrochemical grinding (ECG), Electrochemical drilling (ECD) are the processes under electro chemical. Under chemical machining two processes i.e. Chemical machining and photochemical machining (PCM). Electron beam machining (EBM), Electric discharge machining (EDM), Laser Beam machining (LBM), Ion Beam Machining (IBM) are thermo-chemical processes. In the present paper, the authors have performed electric discharge machining of Tungsten carbide with an aim to obtain optimal parametric combination. The results show that pulseon-time of 40µs, pulse-off-time of 2µs and current of 8A are optimum combination for machining of Tungsten carbide (93% WC and 7%Co).

Electric discharge machining
Electric discharge machining (EDM) came into existence in 1766 and was discovered by Joseph Priestley, it is a non-traditional machining process (Rao, 2001,Singh andSharma 2017). Electric discharge machining has other names such as die sinking, spark machining, wire erosion, spark eroding or wire burring. In this process work-piece and tool electrode are immersed in the dielectric fluid and DC power supply is supplied to them. Due to discharge current the dielectric fluid is ionized which causes spark formation between electrode and work-piece which leads to the removal of material from workpiece by fusion and vaporization mechanism (as shown in Fig. 1).

Fig.1. Electrical discharge machining (EDM) process
The process has become more popular nowadays because of its capability to machine hard electrically conductive materials which is difficult with the conventional process (Jahan et al., 2011), but it has certain limitations (Kumar et al., 2010).
• Alteration in the properties of outer machined surface.
• Requires post processing after machining.
• Formation of heat affected zone (HAZ) and micro cracks.
• Low material removal rate. To deal with these problems, researchers have used abrasives such as silica, aluminium oxide, graphite, silicon carbide, etc. to achieve improved performance.

Research Methodology
The authors developed a basic research framework as shown in Fig. 2 below which presents a step by step procedure adopted in this study. The steps involved are as follows. 1. Firstly, selection of material and methodology is performed. 2. After conducting literature survey varying and fixed input parameters are selected. 3. Numbers of pilot experiments were done to find the scale of varying input parameters. 4. RSM Box-Behnken approach with 5 centre point is being used. A total of 17 experiments were performed. 5. Best combination is being found out among the 17 experiment by using GRA method. 6. Graphs are being plotted for various input and output parameters

Basic concept of Grey Relation Analysis
GRA (Grey Relational Analysis) is based upon grey theory given by Deng, 1989 it is an effective method to handle the uncertainty in the multi-input data. In this literature the GRA is used in various fields such as engineering, as well as in industrial and forecasting areas. RSM, GRA and entropy analysis was been successfully used to find the best parametric combination.

Taguchi
In the present article, the author conducted experiment to find out optimal combination for machining Die steel H13 using Taguchi L9 method.
Current is the significant parameter to obtain high MRR and low TWR for machining Die steel H13.

GRG
This article showed a detailed comparison of different input parameters.
The author found that pulse-ontime is the most important input parameter among the rest input parameters for machining tungsten carbide alloy. 8.

Taguchi
In the present article the author used Taguchi L9 method is utilizedto compare the impact of pulse on time, pulse off time and current on the output parameters such as EWR and MRR.
The author found that pulse on time has noticeable impact on MRR. Pulse off time and current has impact on TWR.
In general current is the most significant factor.

RSM
In the present article author has lead down a comparison between copper electrode and La2O3 coated copper electrode with different amount of concentration of La2O3 .And found out best concentration of La2O3.
The comprehensive study showed that La2O3 concentration of 1.2g/l showed improved micro-hardness and reduction of 2.61% in electrode wear rate.
The steps in GRA are as follows: 1. Normalization of experimental results 2.
Calculate GRC (Grey relational coefficient) which represent the relationship between normalized and ideal results.

4.
Calculate GRG (Grey relational grade) which is arithmetic mean of GRC (grey relational coefficients).

5.
Providing ranks to grey relational grade which will give the optimal result The normalized value obtained as mentioned in step 1 are obtained using equation (1) "higher is better" for material removal rate and micro-hardness and equation (2) "lower is better" for surface roughness.

GRC (grey relational coefficient):
The grey relational coefficient 0, ( )represents the relationship between normalized and ideal results. The GRC coefficient is expressed in equation ( Where ∆ is minimum deviational sequence,∆ is maximum deviational sequence, ∆ 0 ( )is deviation sequence and is distinguishing coefficient which range from 0 to 1 and is set as 0.5

GRG (Grey relational grade)
The grey relational grade is the arithmetic mean of grey relational coefficient (GRC)as described in the equation (4).
Where βi is the grey relational grade for ith experiments and n stands for number of responses.

Experiment details
The experiments are carried out on EDM MODEL 500 X 300 ENC as shown in Fig 3. Tungsten carbide alloy with 7% cobalt constituent is taken as work-piece and electrolytic copper is taken as electrode material, (as shown in figure 4(a) and (b) specifications in detail are mentioned in the table below, while EDM oil VELVEX EDMVEL-2 is used for experimentation. Properties and specifications of tool and work piece are presented in Table 3. The varying input parameters and levels are shown in Table 4 and the fixed input parameters are shown in Table 5 respectively. 7.

Material Removal rate
To calculate material removal rate initial and final weight of the samples are calculated after each experiment at same conditions as shown in equation (5): Where wi is initial weight (in g), wf is final weight (in g), t is machining time (in min) and d is density (in g/mm 3 ). MRR is calculated on weighing machine having a least count 0.0001g.

Micro-hardness
Micro-hardness depends upon the load applied (P) and the average of the diagonals of imprint (d). The indenter is square pyramidal in shape with apex angle of 136 ο is used. The hardness number is HV = 1.854 2 Where P the load is applied and in Newton and d the average value of two diagonals in millimetre. Micro-hardness (MH) is measured using micro-hardness tester (Model MVH -S AUTO) of ominitech industries Pune, India. Imprint diagonal size is measured using Quientiment software. The load of 1kg was applied for a dwell time of 15 second on each sample once.

Surface roughness
Surface roughness is measured on Perthometer (Model SJ-301 of Mitutoyo, Japan). Arithmetic mean method (Ra) is used to obtain surface roughness as shown in equation 7.
The equipment has a range of 0.01µm to 100µm. Each surface is measured thrice at three places and their average is taken as surface roughness value. The tracing length of 2.5mm is taken every time.

Parametric optimization of process parameters
In this part grey relational approach is utilized to optimize the EDM parameters.

GRA (grey relational analysis)
The RSM (Box-Behnken with 5 centre point) gives 17 experiments. Values of material removal rate, micro-hardness, surface roughness are normalized in 0 to 1 range using eq. (1) and eq. (2) as shown in table 6. Grey relational grade values which depict the relationship between ideal and normalized values are calculated using eq. (3) and along with that grey relation grade values which are the average of grey relational coefficients are calculated using eq. (4). The ranks are allotted to all the grey relational grade values in the decreasing order from maximum to minimum as shown in table 7. The average value of grey response for all the input parameters is shown in table 8.

Results and Discussions
The confirmatory experiment is carried out to find out predicted GRG using equation 8 Where αm = Mean of all GRG values αk=Average of GRG value at minimum rank optimize machining parameters for tungsten carbide alloy The confirmatory experiment values are presented in Table  9, Combinations A1B1C2 and A3B1C2are the optimal settings, it is observed that from initial setting of A1B1C2there is increase in material removal rate from 1.0477 to 2.3046, increase in micro-hardness of 1239 to 1361 and decrease in surface roughness from 1.86 to 1.26 and GRG value has also improved from 0.4529 to 0.8632. Hence from the above result it can be said that grey relation is successfully implemented to optimization.

Conclusions and future scope
The main objective in this work is exploring spark machining process for hard to machine material like tungsten carbide. (93% WC and 7%Co). Firstly experiments were performed in order to determine whether electrodes are acceptable for machining. Then input parameters and their working ranges by referring the literature survey were select-ed to investigate their influence on output characteristics i.e. material removal rate,surface roughness and micro-hardness. For this purpose, seventeen experiments were con-ducted. Further, to optimize the results, grey relation analysis method is used to handle the multi objective dataset. The best combination for machining Tungsten carbide over the given range is successfully obtained. The results show that experiment 14 has maximum GRG value with the optimal combination i.e. pulse-on-time 40 µs, pulse-off-time 2 µs and current 12A. In the future work other composition of tungsten carbide could be used to increase machining capa-bility of the tungsten carbide. The possibility of using suitable solvent to lower the emission of hazardous fumes can also be explored. The results can be further optimized by using evolutionary optimization techniques like, genetic algo-rithms, particle swam optimization or ant bee colony.