OBRABOTKAMETALLOV Vol. 27 No. 2 2025 technology 39.945 log 3.716 i SR x P ; (12) 4.98 log 0.262 i TWR x P ; (13) ( . ) ( , ) ( , . ) ( ) n R MRR MRR TWR TWR SR SR U P n R W P n R W P n R W , (14) where n is a trial number; R is a repetition number; повторение WMRR, WTWR и WSR are the assumed values of weighting coefficients. Specific values: n (trial number) = 1, 2, 3, ..., 18; R (repetition number) = 1, 2, 3… . The utility values are computed using Expression (14) for 18 experimental conditions and three repetitions of MRR, SR, and TWR, as shown in Table 7 for both raw data and S/N ratios. Given that utility is a quality characteristic favoring larger values, the S/N ratio is calculated using Expression 3. The data from Table 7 are presented as main effect plots for both S/N ratios and raw data. Ta b l e 7 Utility data Trail No. Raw data (Utility Values) S/N ratio (dB) R1 R2 R3 1 3.913 3.922 3.885 11.8358 2 3.356 3.333 3.406 10.5391 3 2.787 2.786 2.800 8.9149 4 4.226 4.095 4.115 12.3486 5 3.489 3.557 3.517 10.9327 6 3.040 3.068 3.050 9.6936 7 2.945 2.934 2.956 9.3822 8 2.559 2.603 2.558 8.2086 9 4.192 4.158 4.182 12.4184 10 2.742 2.853 2.807 8.9407 11 4.264 4.276 4.271 12.6095 12 3.825 3.823 3.830 11.6546 13 3.681 3.631 3.736 11.3215 14 2.540 2.541 2.527 8.0836 15 4.199 4.234 4.234 12.5103 16 2.137 2.218 2.227 6.8203 17 4.511 4.536 4.506 13.0980 18 3.798 3.792 3.803 11.5906 From the utility data (Table 7), it is clearly stated that the highest S/N ratio is obtained for Trial 17 (13.0980 dB), which has the best performance with the highest utility values. This suggests an optimal combination of EDM parameters, leading to improved overall machining quality. Conversely, in the case of the lowest S/N ratio, Trial 16 (6.8203 dB) shows the poorest performance, indicating suboptimal process parameters that result in low utility values. The utility values range from 2.137 to 4.536, showing significant variation in machining performance across different trials. Higher S/N ratios correlate with higher utility values, confirming the robustness of the optimal parameter settings. Some trials exhibit closely matched R1,
RkJQdWJsaXNoZXIy MTk0ODM1