Effective parametric analysis of machining curvature channel using semicircular curved copper electrode and OHNS steel workpiece through a novel curved EDM process

Electrical discharge machining (EDM) is the spark erosion machining process in which the material is removed from the workpiece through the spark erosion principle. During machining, a series of discharge sparks are generated and it is possible to machine the hard material with the complex geometry [1]. The complex three dimensional geometry and hard material parts requires the machining characteristics, which can be implemented for the performance improvement in the EDM process [2]. Similarly, the inventions in the EDM process are increasing. A new EDM process in which the non linear channel is developed through a customized experimental setup and that channel can be used for cooling purpose in some industrial application [3]. Further, the machining of customized channel through EDM requires a specific application. It is achieved by providing the various mechanisms through which, it is possible to obtained the various desired channels. Although one mechanism is different from another for generating the customized channel, its behavioral study is analyzed for developing a new mechanisms [4]. EDM research can be carried out for improving the behavioral domain such as performance measures, optimization and EDM process parameters during machining [5].

Some research findings indicate that the customized geometry of electrode increases the performance of the EDM in many ways. Okada et al developed the customized setup for machining a channel through EDM on specific materials and its parametric effect are studied [6]. Also, the machining parameters study is analyzed using Taguchi's design of experiment and experiments are performed through the alternative mechanism used on EDM [7]. However, a new rotary EDM process in which the electrode is rotating at certain rpm on the particular workpiece material for generating a machining channel. The effect of output characteristics is analyzed by material removal rate (MRR), Tool wear rate (TWR), surface roughness (SR) and overcut [8]. Again, the analysis of machining parameters on workpiece material are carried out on the EDM [9]. With the unique manufacturing process used for analyzing the machining parameters, multi-response optimization is employed. It helps in improving the mean of the EDM process [10]. Marichamy et al discussed a new approach to validate the output mean on EDM by implementing a new manufacturing process adopted for developing the workpiece material. The machineablity on new workpiece material are analyzed, and its effect on machining parameters are discussed by parametric optimization using Taguchi method [11]. J Laxman and K Guru Raj reviewed the methodological procedure utilized in EDM process through the Taguchi method and regression model. The predicted results indicate the sustainability of the regression model and its implementation [12]. An integrated study is involved in EDM process to evaluate the mathematical models for the optimization of machining parameters [13]. In order to identify the relationship with multiple performance models, an TOPSIS methology is suggested for evaluating the optimum condition in EDM machining process [14]. Machining parameters and machining characteristics are the two important criterion for practical implementation of TOPSIS method in the EDM process [15]. To summarize, most of the multiple performance characteristics are optimized by adopting the Taguchi and TOPSIS in industrial application [16].

In this paper, a curved machining channel has been machined with four important input variables: sparking current, pulse on time, pulse off time and angular sensitivity through a novel curved EDM device. The angular sensitivity is the new curved machining parameter selected for the curvature channel machining through curved EDM process. The semicircular curved copper electrode and OHNS workpiece have been considered during experimentation, and its parametric analysis has been carried out. Hence, the output performance characteristic for the optimum level is formulated, in order to maximize the performance using the Taguchi method, regression analysis and TOPSIS Method.

2.1. Innovative device detail

The experiments were conducted on a 'Electronica 500 X 300' Z axis numerically controlled (ZNC) EDM machine manufactured in Pune, India. A curved attachment is developed and placed on the table of the ZNC EDM machine to execute the oscillating drive to the semicircular curved copper electrode. It involves drive motor, timing gears driven mechanism, bearings house assembly, nylon bolts, nuts, aluminum support for the bearings house assembly and the locking mechanism for the electrode as shown in figure 1. As illustrated in figure 1, the locking mechanism is used for locking the semi circular curved copper electrode with the rectangular plate through a grub screw. The top side of a rectangular plate consist of a specific hole of a diameter equal to the thickness of the curved electrode. A new hole is developed in perpendicular with the existing hole in order to align the grub screw. The threads are generated on a new hole through the tapping process for locking the grub screw. Allen key is used for tightening the grub screw with a certain torque. During tightening, it is necessary that the semicircular curved copper electrode is inserted in the top hole of the rectangular plate for easy locking. Features of curved machining mechanism are as discussed:

• (i)  
  Forward-reverse motion to the electrode.
• (ii)  
  Flexible in design and machining.
• (iii)  
  Electrode supporting holder for different radius.
• (iv)  
  Special flushing during machining.

Zoom In Zoom Out Reset image size

As compared to linear EDM machining principle,the curved machining principle involves the curved electrode rotating through a specific direction of the axis of rotation with the certain angular speed. The forward and reverse movement of the electrode provides the curvature direction and in-depth penetration during machining on the workpiece.

2.2. Workpiece and electrode for curved machining

The workpiece used for carrying out the experiments is oil hardening non-shrinking steel (OHNS) applied in cooling channel machining of the plastic injection mold. The workpiece is shaped in rectangular pieces of length 40 mm, width 12.5 mm and height 40 mm. The electrode is a semi circular ring with a radius of 16.5 mm and a cross-sectional thickness of 3.2 mm. Experiments conducted using nine pure semicircular copper electrodes with nine OHNS workpieces. These combinations are obtained from the design of experiment. Every experiment required the individual curved electrode and a workpiece for machining as shown in figure 2 respectively. The chemical composition of the semicircular curved copper electrode and OHNS workpiece are shown in table 1.

Zoom In Zoom Out Reset image size

Table 1.  Chemical composition of workpiece.

Workpiece	Oil hardened non shrinking steel
Material	C (Carbon)	Mn (Magnesium)	W(Tungsten)	Cr (Chromium)	V(Vanadium)	Fe (Iron)
Percentage contribution (%)	0.94	0.99	0.49	0.49	0.99	remaining
Electrode	Semicircular curved copper electrode
Material	Pure Copper
Percentage contribution (%)	99

Both the end face of the individual curved electrode were finished using a bench grinder, and all sides of the individual workpiece were cleaned. The measurement of Electrode Wear Rate (EWR) and Material Removal Rate (MRR) have been obtained by using the difference in weight measured before and weight after the machining of the workpieces and electrodes with respect to the machining time and density of material. The measurement of weight is performed using METTLER-Toledoc 0.19g SB-24001DR analytical balance.

The power supply was successfully connected to the closed circuit in which the workpiece and electrode were attached to the negative and positive terminals, respectively. After completing each experiment, the workpiece and electrode are separated and weighed. The machining time is kept constant at 120 min for every experiment, and it is measured by using a digital watch. The density of electrode and workpiece material are selected as 8.96 g cm⁻³ and 8.67 g cm⁻³ respectively. The EWR and MRR are calculated using the formula as described in equations (1) and (2):

$$\begin{eqnarray}&&{EWR}=\displaystyle \frac{({Electrode}\ {{weight}}_{{before}}-{Electrode}\ {{weight}}_{{after}})\ast 1000}{({Density}\ {of}\ {electrode})\ast ({Machining}\ {time})}\end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&{MRR}=\displaystyle \frac{({Workpiece}\ {{weight}}_{{before}}-{Workpiece}\ {{weight}}_{{after}})\ast 1000}{({Density}\ {of}\ {workpiece})\ast ({Machining}\ {time})}\end{eqnarray} \tag{ 2 }$$

Further, the curved machining angle (CMA) (θ) is the critical output performance characteristic in this experimentation. It is a challenging task to observe the curved channel machined inside the rectangular workpiece. The curved machined channel was measured by splitting the workpiece into two halves. The machine used for dividing the workpiece is wire cut EDM manufactured by Electronica, Pune. The machining angle was measured between the curved channel machined tip inside the workpieces to the top horizontal surface of the workpiece. The curved angle was measured by using the profile projector, and the average of all three reading has used.

The linear curvature depth (L) is another dependent performance characteristic in the curved EDM machining process. The vertical distance obtained after generating the curved channel machined results in the linear curvature depth in the workpiece. This characteristic elaborates the depth observed during machining of the curved channel by the forward-reverse motion of the curved electrode.

After completion of each experiment, the electrode has separated from the electrode holder assembly along with the workpiece. The tip of the curved electrode has appropriately aligned with the workpiece to generate the curved machining. Based on the machining characteristics, the performance is analyzed.

To obtain the relationship between the input machining parameters and output characteristics, the adopted techniques used in the optimization is Taguchi method. Taguchi method is the best method in the design of experiment. The calculated experiments are obtained and their levels have defined according to the output requirements. The results obtained after experimentation with the quantity characteristics parameters have variation and contribute to the output characteristics. Some critical experiments have identified for the optimum EWR, MRR, CMA and L. According to the Taguchi method, the design of the experiments utilizes the following [17]:

• (i)  
  Identification of the output characteristics and the input machining parameters.In all the experiments, input characteristics with machining parameters are sparking current (IP), pulse on time (T_on), pulse off time (T_off) and angular sensitivity (SEN). Also, the output characteristics are EWR, MRR, CMA and L.
• (ii)  
  Identifying the various level number for the particular parameters.According to the capability of the ZNC EDM machine, the particular parameters and their levels are presented in table 2. The input and output performance characteristics with their levels are selected based on the initial trial experiment performed through the curved EDM process.
• (iii)  
  Selection of the proper orthogonal array and assignment of the input parameters to the orthogonal array as per the sequence of performance of the experiment.Orthogonal arrays are selected and constructed to facilitate experimental design using MINITAB 18 statistical software. It was found that the required orthogonal array as per the given experimental condition is L₉ and experimental outputs are displayed in table 3. Equations (1) and (2) are used to calculate the values of EWR and MRR respectively.

The experimental techniques used in the curved machining involves the complexity in optimizing the characteristics through a reduced number of experimentation activities. During the Taguchi method, the L₉ selection criterion depends on each row indicating the combination of the input characteristics with aligned level and the column indicates the formation of the orthogonal array. In this study, MINITAB 18 statistical software is used to identify and analyzed the various effect of the level selection (Level 1, Level 2 and Level 3) in curved machining during the observation of standard experimental results. Table 4 depicts the experimental results with S/N ratio for EWR, MRR, CMA and L.

The observed data characterizing the key effect for each level of the input characteristic are presented in tables 5–8. In curved EDM process, the main aim is to reduce the EWR and increase the MRR, CMA and L . Hence, the S/N ratio is distinguished for EWR (Smaller is better) and MRR, CMA and L (larger is better). Figures 3–6 show the key effect values has plotted for input process characteristics.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Further, analysis of variance (ANOVA) is used to determine the characteristics and relative importance of the variables. ANOVA results evaluate and identify the individual contribution of the selected input variables from the total model including the error. The S/N ratio is generally used to find the best set of controllable variables involved in the production process.

The ANOVA table was generated by MINITAB 18 statistical software. Tables 9–12 show the relevance of the different input characteristics towards the EWR, MRR, CMA and L. The entire observed result are confirmed at 95% confidence level and model interaction with nested parameters of pulse off time and pulse on time. It was found out that sparking current followed by the pulse off time and angular sensitivity had more significant impact over the EWR. However, the pulse off time has less influence over the EWR.

Again, the results of MRR explained that the contribution of angular sensitivity, sparking current and pulse on time is greatly affected by input parameters. Sparking current contribution is less during the MRR when the interaction are provided between the pulse on time and pulse off time.

Similarly, the CMA indicates that the contribution of the angular sensitivity is the primary input machining factor for machining the curved machining channel. Although, the sparking current has also contributed in curved EDM process when interacted with the pulse on time and pulse off time.

In L, the angular sensitivity indicates that significant depth is observed during the curved machining channel. The contribution of sparking current and pulse on time has significant impact on the output characteristics.

The main aim is to determine the input variables such as sparking current, pulse on time, pulse off time and angular sensitivity to optimize the EWR, MRR, CMA and L. It is difficult to determine the four input process characteristics relationship for optimizing four output characteristics with the electrode and workpiece material using Taguchi method, regression analysis and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method was employed.

The following sections frame a model for identifying the multi-objective optimization for the semi circular copper curved electrode and OHNS using regression analysis and TOPSIS method:

6.1. Formation of regression model for output characteristics

The regression equation is necessary to obtain the relationship from the input and output characteristics through a mathematical model. Hence, the linear relationship is required between the selected particular characteristics (Input and output). The derived regression equation for EWR equation (3), MRR equation (4), CMA equation (5) and L equation (6) are as follows:

$$\begin{eqnarray}\begin{array}{rcl}{EWR} & = & 0.121\,6-0.024\,0{{IP}}_{5}-0.054\,0{{IP}}_{20}+0.078\,0{{IP}}_{25}\\ & & +0.041\,4{{TOFF}}_{5}-0.059\,0{{TOFF}}_{93}+0.017\,6{{TOFF}}_{181}\\ & & -0.061\,5{{SEN}}_{1}+0.042\,7{{SEN}}_{51}+0.018\,8{{SEN}}_{100}\end{array}\end{eqnarray} \tag{ 3 }$$

$$\begin{eqnarray}\begin{array}{rcl}{MRR} & = & 0.300\,1+0.033\,1{{IP}}_{5}+0.007\,5{{IP}}_{20}-0.040\,6{{IP}}_{35}\\ & & -0.034\,2{{TON}}_{210}-0.024\,6{{TON}}_{577}+0.058\,7{{TON}}_{945}\\ & & -0.053\,4{{SEN}}_{1}+0.033\,1{{SEN}}_{51}+0.020\,3{{SEN}}_{100}\end{array}\end{eqnarray} \tag{ 4 }$$

$$\begin{eqnarray}\begin{array}{rcl}{CMA} & = & 77.556+5.111{{IP}}_{5}+0.778{{IP}}_{20}-5.889{{IP}}_{35}\\ & & -0.222{{TON}}_{210}-0.889{{TON}}_{577}+1.111{{TON}}_{945}\\ & & -7.556{{SEN}}_{1}+0.444{{SEN}}_{51}+7.111{{SEN}}_{100}\end{array}\end{eqnarray} \tag{ 5 }$$

$$\begin{eqnarray}\begin{array}{rcl}L & = & 32.086+2.828{{IP}}_{5}+0.591{{IP}}_{20}-3.419{{IP}}_{35}\\ & & -0.966{{TOFF}}_{5}+1.641{{TOFF}}_{93}-0.676{{TOFF}}_{181}\\ & & -3.989{{SEN}}_{1}+0.458{{SEN}}_{51}+3.531{{SEN}}_{100}\end{array}\end{eqnarray} \tag{ 6 }$$

The R² value indicates that the forecasts depicts 95.15%, 93.15%, 99.19% and 99.31% of the variance in EWR, MRR, CMA and L respectively. Tables 13–16 indicates that the selected regression models are the best one of the models that can be used with these parameters and parameter levels by R²_adj test.

Table 13.  Coefficients for regression analysis for EWR.

Term	Coef	SE Coef	T-Value	P-Value	VIF
Constant	0.121 6	0.013 4	9.1	0.012
IP
5	−0.024	0.018 9	−1.27	0.332	1.33
20	−0.054	0.018 9	−2.86	0.104	1.33
TOFF
5	0.041 4	0.018 9	2.19	0.16	1.33
93	−0.059	0.018 9	−3.12	0.089	1.33
SEN
1	−0.061 5	0.018 9	−3.25	0.083	1.33
51	0.042 7	0.018 9	2.26	0.153	1.33
S	R2	R2adj	PRESS	R2pred
0.040 098 2	95.15%	80.62%	0.065 118 4	1.87%

Table 14.  Coefficients for regression analysis for MRR.

Term	Coef	SE Coef	T-Value	P-Value	VIF
Constant	0.300 1	0.012 3	24.37	0.002
IP
5	0.033 1	0.017 4	1.9	0.198	1.33
20	0.007 5	0.017 4	0.43	0.71	1.33
TON
210	−0.034 2	0.017 4	−1.96	0.189	1.33
577	−0.024 6	0.017 4	−1.41	0.294	1.33
SEN
1	−0.053 4	0.017 4	−3.07	0.092	1.33
51	0.033 1	0.017 4	1.9	0.198	1.33
S	R2	R2adj	PRESS	R2pred
0.036 949 1	93.15%	72.61%	0.055 292 1	0.00%

Table 15.  Coefficients for regression analysis for CMA.

Term	Coef	SE Coef	T-Value	P-Value	VIF
Constant	77.556	0.484	160.13	0
IP
5	5.111	0.685	7.46	0.017	1.33
20	0.778	0.685	1.14	0.374	1.33
TON
210	−0.222	0.685	−0.32	0.776	1.33
577	−0.889	0.685	−1.3	0.324	1.33
SEN
1	−7.556	0.685	−11.03	0.008	1.33
51	0.444	0.685	0.65	0.583	1.33
S	R2	R2adj	PRESS	R2pred
1.452 97	99.19%	96.74%	85.5	83.50%

Table 16.  Coefficients for regression analysis for L.

Term	Coef	SE Coef	T-Value	P-Value	VIF
Constant	32.086	0.247	129.65	0
IP
5	2.828	0.35	8.08	0.015	1.33
20	0.591	0.35	1.69	0.233	1.33
TOFF
5	−0.966	0.35	−2.76	0.11	1.33
93	1.641	0.35	4.69	0.043	1.33
SEN
1	−3.989	0.35	−11.4	0.008	1.33
51	0.458	0.35	1.31	0.321	1.33
S	R2	R2adj	PRESS	R2pred
0.742 436	99.31%	97.23%	22.324	85.98%

6.2. Formation of TOPSIS model for output performance characteristics

TOPSIS is very well known and commonly used method. It is the best alternative method and indicates the shortest distance that is the Euclidean distance from the ideal solution. The various steps involved in TOPSIS are as follows:

• (i)  
  Determine the objective and identify the pertinent evaluation criteria.The objective of this paper is to optimize the output performance characteristics with the selected machining parameters are sparking current (IP), pulse on time (T_on), pulse off time (T_off) and angular sensitivity (SEN). Also, pertinent evaluation characteristics are EWR, MRR, CMA and L. Table 2 shows the detail related to the pertinent evaluation characteristics.
• (ii)  
  Preparation of a normalized decision matrix based on the inputs and outputs characteristics.Each row of the decision matrix is allocated to one alternative and each column to one criterion. Therefore, an element, x_ij of the decision matrix shows the performance of i^th alternative with respect to j^th criterion.

  $$\begin{eqnarray}&&{r}_{{ij}}=\displaystyle \frac{{x}_{{ij}}}{\sqrt{{\displaystyle \sum }_{i=1}^{m}{x}_{{ij}}^{2}}}\end{eqnarray} \tag{ 7 }$$

  Equation (7) is used to calculate the normalized values for the normalized matrix. Table 17 shows the linear normalized decision matrix obtained for the input and output characteristics.
• (iii)  
  Construction of a weighted normalized decision matrix.The criteria weights are obtained for all input and output characteristics which are set to 0.125. The ideal best value and the ideal worst value are denoted by V⁺ and V⁻ respectively from equation (8) and equation (9). Table 18 indicates that weighted normalized matrix with the beneficial and non beneficial values. The values having ideal best are sparking current, pulse on time, angular sensitivity, MRR, CMA and L and ideal worst are pulse off time and EWR.

  $$\begin{eqnarray}&&{V}^{+}=\left\{\left[\displaystyle \sum _{i}^{\mathrm{Max}}{V}_{{ij}}/j\epsilon J\right],\left[\displaystyle \sum _{i}^{\mathrm{Min}}{V}_{{ij}}/j\epsilon {J}^{{\prime} }\right]/1,2..N\right\}\end{eqnarray} \tag{ 8 }$$

  $$\begin{eqnarray}&&{V}^{-}=\left\{\left[\displaystyle \sum _{i}^{\mathrm{Min}}{V}_{{ij}}/j\epsilon J\right],\left[\displaystyle \sum _{i}^{\mathrm{Max}}{V}_{{ij}}/j\epsilon {J}^{{\prime} }\right]/1,2...N\right\}\end{eqnarray} \tag{ 9 }$$
• (iv)  
  Calculating the Euclidean distances and Performance Score for the values.Equations (10) and (11) are used for calculating the Euclidean distances for ideal best and ideal worst solution. The performance score (P_i) is obtained from the equation (12). Table 19 indicates the values of the ideal best and ideal worst with the performance score. The rank is also provided for obtaining the optimized experiment required for the curved machining EDM process in developing the curved machining channel.

  $$\begin{eqnarray}&&{S}_{i}^{+}={\left\{\displaystyle \sum _{j=1}^{N}{\left({V}_{{ij}}-{V}_{j}^{+}\right)}^{2}\right\}}^{0.5},\quad i=1,2,\,\ldots \,N\end{eqnarray} \tag{ 10 }$$

  $$\begin{eqnarray}&&{S}_{i}^{-}={\left\{\displaystyle \sum _{j=1}^{N}{\left({V}_{{ij}}-{V}_{j}^{-}\right)}^{2}\right\}}^{0.5},\quad i=1,2,\,\ldots \,N\end{eqnarray} \tag{ 11 }$$

  $$\begin{eqnarray}&&{P}_{i}={S}_{i}^{-}/\left({S}_{I}^{+}+{S}_{i}^{-}\right)\end{eqnarray} \tag{ 12 }$$

  In order to develop the machining channel through the curved EDM Process, fifth and third experiment as illustrated in table 19 are the best combination obtained through the TOPSIS Method as per the obtained rank.

Table 17.  Linear normalized decision matrix.

IP	TON	TOFF	SEN	EWR	MRR	CMA	L
0.071 1	0.107 6	0.014 2	0.005 1	0.233 2	0.291 8	0.320 8	0.305 5
0.071 1	0.295 6	0.263 8	0.262 3	0.158 3	0.364 8	0.355 0	0.379 3
0.071 1	0.484 1	0.513 4	0.514 3	0.264 4	0.427 3	0.385 0	0.394 1
0.284 3	0.107 6	0.263 8	0.514 3	0.120 8	0.312 7	0.367 9	0.386 8
0.284 3	0.295 6	0.513 4	0.005 1	0.029 2	0.229 3	0.295 2	0.286 6
0.284 3	0.484 1	0.014 2	0.262 3	0.304 0	0.458 6	0.342 2	0.336 4
0.497 5	0.107 6	0.513 4	0.262 3	0.641 3	0.260 6	0.303 7	0.290 0
0.497 5	0.295 6	0.014 2	0.514 3	0.558 0	0.302 3	0.333 7	0.319 7
0.497 5	0.484 1	0.263 8	0.005 1	0.141 6	0.281 4	0.282 3	0.276 2

Table 18.  Weighted linear normalized decision matrix.

Weight	0.125 0	0.125 0	0.125 0	0.125 0	0.125 0	0.125 0	0.125 0	0.125 0
	IP	TON	TOFF	SEN	EWR	MRR	CMA	L
	0.008 9	0.013 4	0.001 8	0.000 6	0.029 2	0.036 5	0.040 1	0.038 2
	0.008 9	0.036 9	0.033 0	0.032 8	0.019 8	0.045 6	0.044 4	0.047 4
	0.008 9	0.060 5	0.064 2	0.064 3	0.033 1	0.053 4	0.048 1	0.049 3
	0.035 5	0.013 4	0.033 0	0.064 3	0.015 1	0.039 1	0.046 0	0.048 3
	0.035 5	0.036 9	0.064 2	0.000 6	0.003 6	0.028 7	0.036 9	0.035 8
	0.035 5	0.060 5	0.001 8	0.032 8	0.038 0	0.057 3	0.042 8	0.042 1
	0.062 2	0.013 4	0.064 2	0.032 8	0.080 2	0.032 6	0.038 0	0.036 2
	0.062 2	0.036 9	0.001 8	0.064 3	0.069 8	0.037 8	0.041 7	0.040 0
	0.062 2	0.060 5	0.033 0	0.000 6	0.017 7	0.035 2	0.035 3	0.034 5
V+ (ideal best)	0.062 2	0.060 5	0.064 2	0.064 3	0.080 2	0.057 3	0.048 1	0.049 3
V− (ideal worst)	0.008 9	0.013 4	0.001 8	0.000 6	0.003 6	0.028 7	0.035 3	0.034 5

Table 19.  Euclidean distances and performance score.

Experiment No	${S}_{i}^{+}$	${S}_{i}^{-}$	Pi	Rank
1	0.101 9	0.081 2	0.183 1	3
2	0.076 0	0.082 1	0.158 1	8
3	0.087 3	0.097 4	0.184 6	2
4	0.066 1	0.101 9	0.168 0	7
5	0.101 7	0.084 4	0.186 1	1
6	0.054 4	0.102 8	0.157 2	9
7	0.117 7	0.062 4	0.180 1	5
8	0.073 7	0.107 7	0.181 4	4
9	0.078 1	0.099 9	0.177 9	6

6.3. Confirmation runs

The purpose of the confirmation run is to validate the conclusions drawn during the analysis phase. Optimum levels of the input machining parameters have been used for the prediction and confirmation of the improvement in output characteristics. In this study, a new experiment has formulated with a combination of parameters along with their levels A2B2C2D1, A3B1C3D1, A1B3C2D2 and A1B3C2D3 to obtain the characteristics respectively. The final step is to predict and verify the improvement in the output characteristics by using the response optimization in MINITAB 18 statistical software. Figure 7 shows the optimized response plot for the optimized machining condition.

Zoom In Zoom Out Reset image size

The confidence range at 95% level Is estimated for the confirmation of the experiment. From the ANOVA table, calculated values are evaluated and presented for the validation of the results. Table 20 shows the results of confirmation obtained for the output performance characteristics. The corresponding optimum values displayed within the range of the confidence interval. It Is found that there is a significant effect on the output of performance characteristics with the input independent machining variables.

Table 20.  Results of confirmation runs for output performance characteristics.

Output performance characteristics	Optimal machining parameter combination	Corresponding optimum value	Confidence intervals (95%) range	Final optimized machining parameter	Average confirmation experimental value at optimum machining conditions
Electrode wear rate	A2B2C2D1	0.057 5	(−0.094–0.209 6)	A1B3C2D3	0.062 5
Material removal rate	A3B1C3D1	0.412 2	(0.272 0–0.552 4)		0.462 9
Curved Machining Angle	A1B3C2D2	90.86	(85.38–96.40)		90.00
Linear Curvature Depth	A1B3C2D3	40.086	(37.268–42.903)		37.325

The optimized machining parameters combination are experimentally evaluated, and the average confirmed experimental value are obtained. It indicates that the values obtained are in the range of the individual selected output characteristics.

Figure 11 shows the curved channel machined through curved EDM Process. The geometrical analysis is performed using the Mitutoyo Coordinate Measuring Machine (CMM) Model Crystal-Apex C544. As per the obtained data, the CMA, L and depth of channel are found to be 90.00^o, 37.325 and 3.225 mm respectively. The actual dimensions are verified with the actual design for validation. as shown in figure 9. Although, the surface finish is required to identify the surface texture obtained after machining the OHNS workpiece trough a semi circular copper curved electrode. The measurement is carried out on the Mitutoyo High precision surface roughness tester Model (SV 514) in which the average value of three consecutive readings before and after are found to be 3.154 μm and 3.641 μm as shown in figure 8.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The machined surface analysis has been done by scanning electron microscopy [18] (SEM) (JEOL, JCM-6000 PLUS) to investigate the microstructure of the workpiece surface before and after machining as shown in figures 10 and 11 respectively. SEM on the surface of OHNS die steel machined by curved EDM shows discrete craters with some volcanic features, and many spherical droplets left behind on the EDMed surface, which shows that the material removal mechanism is melting and evaporation. The top material in extremely high temperature region will vaporize, while the bottom surface material will melt. Two separate regions are clearly visible—one is the original material and another is the recast layer which has a brighter surface having white colour.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Small thermal gradient occurs at the lower pulse current, which resulted in formation of a thinner recast layer. As the pulse current increases, the recast layer appears thicker because at a higher pulse current, a steeper thermal gradient builds up resulting in a thermal effect beneath the melting zone. This behavior leads to a greater removal of molten layer that has not flushed out by the dielectric fluid as a result it re-solidifies and remains attached to the machined surface.

To establish a new cooling channel machined through a Curved EDM process, we proposed a simple innovative device installed on a ZNC EDM. It consist of the curved semicircular copper electrode with OHNS workpiece for electrical discharge machining including mechanical, electrical and electronic systems. The conclusions are summarised as follows.

• The innovative device can drive a semicircular curved copper electrode to move along a smooth curved trajectory during machining on EDM.
• A smooth curved cooling channels with various shapes and cross sections can be created by changing the electrode diameter and cross section area.
• This device can easily machined the internal and external trajectory.
• Effective cost of manufacturing and installation of this device is less compared to the existing mechanisms for curved machining.
• The combination of the three methods used in this novel curved EDM process effectively validates through the selection of material and machining parameters for the development of the curved machining channel.

The authors would like to express their sincere appreciation to Dr PM Padole, Director, VNIT Nagpur for their cooperation and support. The patent is filled with the invention title name 'LINEAR-CURVED ELECTRICAL DISCHARGE MACHINE (LCM)' with application no.201621020529 A regarding this innovative device.