Influence of generator parameters on cutting width during WEDM process

The primary objective of this paper is to rigorously analyse the influence of selected parameters on the cut width in the context of Wire Electric Discharge Machining (WEDM). The parameters under investigation, each varied at different levels, include pulse width, servo reference mean voltage, time interval between two pulses, discharge frequency, and wire feed speed. As the workpiece material, sintered carbide K10 was used. Data for the response variable, cut width, was meticulously collected using two high-precision instruments: the Zoller Genius 3s and the Alicona InfiniteFocusSL. Among the parameters studied, the mean reference voltage emerged as the most influential variable. It holds a commanding lead with a percentage contribution of as much as 82.4 %, making it nearly 15 times more impactful than the second most influential parameter, which is pulse width. At its lowest setting, or first level, the mean reference voltage results in a record low cut width. Conversely, at its highest setting, or third level, it leads to a record high cut width. This comprehensive analysis not only quantifies the relative importance of each parameter but also provides valuable insights into the optimal settings for achieving desired cut widths. The findings have significant implications for improving the efficiency and precision of WEDM processes.


Introduction
Tungsten carbide and its composite that is also named as hard metal is used for die and cutting industries and even, nowadays, it is playing a prime role in gas, oil, energy, mining, and health industry.However, having a significant issue, it is arduous to machine with traditional techniques as it is not only hard but brittle too.Hence, this unconventional way, Electrical Discharge Machining (EDM) can be a way out to shape it precisely [1].A limited number of traditional methods exist for machining tungsten carbide, for example turning [2], milling [3] and grinding [4].
Wire Electrical Discharge Machining (WEDM) is a spark erosion technique employed for crafting intricate 2D and 3D geometries in electrically conductive materials using a wire electrode (Figure 1).Sparks are generated in a dielectric fluid medium between the workpiece and the wire electrode, either flushed with or submerged in the fluid.The high level of dimensional accuracy and superior surface finishes achievable with WEDM make it especially useful in the production of stamping dies, extrusion dies, and prototype components.In the absence of WEDM, achieving such precision in workpieces would necessitate extensive manual grinding and polishing [6].Given that no mechanical interaction occurs between the workpiece and the electrode, materials of any hardness level can be processed, provided they possess adequate electrical conductivity [7].There are several ways how to measure cutting width (Kerf value).In [8] authors used Universal Profile Projector with an accuracy of 0.001 mm.It was measured at three different points of the kerf in each run.
In wire electrical discharge machining (WEDM), the gap between the wire and the workpiece typically varies from 0.025 to 0.075 mm and is precisely maintained by a computer-controlled positioning mechanism.The cut width, or 'kerf,' is determined by adding the wire diameter (which ranges between 0.05 and 0.4 mm) to twice the distance of the wire-workpiece gap.The efficiency and production rate of WEDM are largely dictated by the material removal rate (MRR).The primary objective when configuring machining settings is to optimize for the highest possible MRR while minimizing the kerf.The calibration of these machining parameters is heavily influenced by the expertise of the operators and the guidelines provided in parameter tables from machine tool manufacturers [6].In [9] authors used a neural network to determine key parameters in the cutting process.In [10] authors investigated the machining parameters on gap width, surface roughness and properties of white layer.

Material and methods
Figure 2 shows the schematic illustration of the WEDM process.The machined material is sintered carbide K10.The diameter of the wire was 0.25 mm, and its tensile strength was up to 500 N/mm 2 .The core was made of pure copper wire (Cu), and the coating was multi-layered with a special GAMA γ Cu5Zn8 alloy.
In this research, a substantial number of variables were taken into account to improve the precision of the findings.The experimental design involved a total of 5 parameters, each with 3 levels.These parameters include pulse width (A), servo voltage (Aj), time between two pulses (B), frequency (FF), wire speed (Ws).
Table 1 displays the experimental parameters that underwent changes or variations during the study.In the experimental setup, the dielectric conductivity was rigorously maintained at a consistent level of 5 µS throughout the entire duration of the study.This stringent control of dielectric conductivity played a pivotal role in ensuring the stability and repeatability of the experimental conditions, thereby minimizing potential sources of variability and allowing for more accurate and reliable results.This level of consistency in dielectric conductivity was achieved through meticulous monitoring and adjustments as needed, reflecting our commitment to maintaining the integrity of the experimental environment.
Table 2 summarizes the constant parameters maintained throughout the experiment along with their corresponding values.Given that there were 5 control factors, each with three levels in the experiment, using a full factorial experimental design would have necessitated a significant number of runs, resulting in a study that would be both time-consuming and costly.To mitigate the need for an excessive number of runs, a Taguchi-based design of experiment was employed.Operating under the assumption that each input parameter in this experiment is independent, a Taguchi orthogonal array L27 was applied.
ROBOFIL 310 by Charmilles Technologies was used for conducting the experiments.As the workpiece material was used sintered carbide K10, with a composition of 93 % -96 % tungsten carbide (WC) and 4 % -7 % cobalt (Co), by Ceratizit.This specific material selection was made with careful consideration to the properties and characteristics required for the study, ensuring that it met the necessary criteria for the experimental objectives.
Table 3 shows the complete experimental design based on the Taguchi L27 matrix, it means that Table 3 contains a comprehensive layout of an experiment, meticulously designed according to a specialized approach or methodology known as the Taguchi L27 matrix.In experimental setup, special attention was devoted to the clamping of the workpiece, which was a rod made of sintered carbide with a diameter of 25 mm.Experimental setup can be seen in the Figure 3.To secure the sintered carbide rod, a custom-designed clamping mechanism was employed, capable of exerting uniform pressure across the surface of the workpiece.Prior to the machining operation, the workpiece was meticulously aligned to ensure that the WEDM wire would cut along the desired path with high precision.After completing all 27 experiments, the workpiece was cleaned and placed into the measuring device.The measurement took place on two measuring devices, Zoller Genius 3s and Alicona InfiniteFocusSL.
The measurement carried out using the Zoller Genius 3s device is visible in Figure 4. Acts is a universal measuring machine that includes 3 CCD cameras and LED lighting.It is equipped with a CNC control unit and a fully automatic measuring cycle.The usual use of this instrument is to measure the geometrical characteristics of cutting tools.Figure 5 shows the output from the measuring device.For better results of the experiment, another same measurement was also carried out, but using of the measuring device Alicona InfiniteFocusSL (Figure 6).Again, all 27 trials were measured three times, the resulting measured value was created by the arithmetic mean of these three measurements.Alicona InfiniteFocusSL is an efficient optical 3D measurement system that uses quick and easy measurement of shape and surface even on microstructural surfaces of objects.With this device and only one system, both the shape and the roughness of the components can be measured.Color images with a large depth of field and the possibility to set a suitable one is also feasible contrast and brightness.The device has a measuring field of 50 mm × 50 mm and a working area by up to 33 mm.By incorporating the Alicona InfiniteFocusSL into our experimental setup, we aim to achieve a higher level of precision and reliability in our measurements, thereby ensuring that the data collected is both accurate and representative of the phenomena being studied.Figure 7 shows the output from the Alicona InfiniteFocusSL measuring device.

Results and discussion
In Table 4, the measurement results obtained using the Zoller Genius 3s and Alicona Infinite Focus SL devices are presented.It is evident from the results that even though the absolute values of the measurements differ, the trend in the change of the cut width remains consistent.This consistency suggests that both devices are reliable for capturing the general trend, despite variations in absolute measurements.It is worth noting that the differences in absolute values could be attributed to the calibration methods or measurement sensitivities of the individual devices.Therefore, while the absolute values may vary, the overarching trend provides valuable insights into the behavior of the material being cut.
The experimental results are evaluated using statistical analysis.Initially, the Signal-to-Noise ratio (S/N ratio) is determined for all 27 conducted experiments focusing on the cut width, also known as 'Kerf '.This S/N ratio serves as a preliminary indicator of the quality and reliability of the experimental data.Following this, an Analysis of Variance (ANOVA) is employed to ascertain whether certain parameters have a statistically significant impact on the cut width.For this experiment, analyses are performed based on the data collected from measurements taken with the Zoller Genius 3s device.This device is particularly chosen for its precision and reliability in capturing the relevant metrics.The ANOVA will help in identifying which factors are most influential in affecting the cut width, thereby providing valuable insights for optimizing the cutting process.The statistical analysis aims not only to validate the experimental setup but also to offer a rigorous framework for interpreting the results.This comprehensive approach ensures that the conclusions drawn are both statistically and practically significant.
where n -number of repetitions, yi -i-th measuring.To obtain the signal-to-noise ratio, the formula (1) was applied.Table 5 presents the values of the S/N ratio.The S/N ratio stands for "Signal-to-Noise Ratio."It's a measure used to quantify the level of a desired signal (useful information) to the level of background noise (unwanted interference).In the context of experiments, a higher S/N ratio would generally indicate that the signal (in this case, the cut width or 'Kerf') is much clearer and more distinguishable from the noise, making the data more reliable.Conversely, a lower S/N ratio would suggest that the noise levels are relatively high, which could compromise the quality of your measurements.ANOVA is performed on the Signal-to-Noise ratios of the cut width (Kerf), obtained from measurements taken with the Zoller Genius 3s device.The results of the Analysis of Variance are displayed in Table 6, which were determined using the Minitab software.According to the Analysis of Variance (ANOVA) for the Signal-to-Noise ratios, as presented in Table 6, it can be concluded that the wire feed speed parameter (Ws) is statistically insignificant in affecting the cut width (Kerf).This is substantiated by the fact that its P-value exceeds the significance level of α = 0.05.Moreover, its percentage contribution to the variation in cut width is less than 1 %, indicating its negligible impact on the process.On the other hand, parameters such as pulse width (A), Servo voltage (Aj), time between two pulses (B), and discharge frequency (FF) are statistically significant in influencing the cut width.This is evidenced by their P-values being less than the significance level of α = 0.05.Among all statistically significant input parameters, the average reference voltage (Aj) has the highest sum of squares value, specifically 17.3649.This implies that it has the most substantial impact on the variation in cut width.
The percentage contribution of all input parameters can be visualized in Figure 8. Notably, the average reference voltage, with a value contributing more than 82 %, stands out as a highly influential factor in determining the observed cut width compared to other variables.This comprehensive analysis provides a robust framework for understanding the relative importance of various parameters in the cutting process.It also offers valuable insights for future optimization efforts, particularly focusing on those parameters that have shown a significant impact.

Conclusion
The input parameters proposed for the study included: pulse width, servo voltage, time between two pulses, discharge frequency, and wire feed speed.These five factors were varied at three different levels.According to the statistical analysis, the factor of pulse width ranked as the second most influential indicator on the cut width.Its percentage contribution was 5.6 %.The widest cut width for this parameter was observed at its third level.However, at the second level, the cut width was slightly below the value compared to the first level.This suggests that a higher value of pulse width does not necessarily correspond to a wider cut width.The most influential variable input parameter is the servo voltage.With a substantial lead in its percentage contribution of 82.4 %, it is almost 15 times more influential than the second most impactful factor.At its first level, the cut width reaches a record low value.On the other hand, at the third level, it achieves a record high value for cut width.In specific numerical terms, according to delta statistics, this difference amounts to as much as 82.8 µm.The third input parameter, time between two pulses, holds the same rank in the hierarchy of effects on the studied variable of cut width.Its percentage contribution is 5.2 %, which represents only a fraction of a percent compared to the more influential parameter.This parameter uniquely achieves the highest cut width at its second level.At the first level of its variable values, it provides the most favorable conditions for cut width.It has the second smallest impact on the cutting speed.Discharge frequency is the second least influential variable.Given this fact, it can be said that the parameters of discharge frequency, time between discharges, and pulse width have approximately the same impact on cut width.The least influential parameter is the wire feed speed.It has almost zero effect on both the cut width and the cutting speed.It was the only parameter categorized as statistically insignificant.

Figure 5 .
Figure 5. Output from the Zoller Genius 3s device

Figure 8 .
Figure 8.The proportion of the influence of input parameters on the width of the cut.

Table 2 .
Constant parameters applied in experiment.

Table 3 .
Design of experiment according to Taguchi L27 orthogonal array.

Table 4 .
Complete experiment design with measured values by Zoller Genius 3s and Alicona InfiniteFocusSL

Table 5 .
S/N ratio values.

Table 6 .
Analysis of variance for S/N ratios.