OBRABOTKAMETALLOV Vol. 27 No. 2 2025 technology Importantly, the true population mean (CIpop) lies completely within the confirmation experiment range (CICE). A narrow population confidence interval combined with consistent experimental data is an excellent indicator that the data are both reliable and consistent with expected trends. Precision increases as confidence interval widths decrease. The width of CIpop directly correlates with the accuracy and certainty of the estimates. Confirmation experiment intervals tend to be wider than those of regular tests because they incorporate variability inherent in experimental results, while the narrower population interval offers better reliability in estimating the true mean due to its stability. Wide confidence intervals indicate a high level of data variability, suggesting the need for improved optimization of experimental control parameters. The significant overlap between the two confidence intervals indicates that the experimental results statistically agree with the predicted values. Confirmation experiments The final step is to verify the achieved optimal levels of process variables by conducting experiments using these optimal values. The experiments were performed three times, and the average values were calculated. The average material removal rate (MRR) achieved was 8.852 mm³/min, the average surface roughness (SR) obtained was 2.818 µm, and the average tool wear rate (TWR) was 0.148 mm³/min. The following expression (equation 16) was used to compute the utility value: U MRR MRR SR SR TWR TWR P W P W P W ; (16) U 8.4434 0.33 4.7987 0.33 1.2352 0.3 3 4. 7775. The empirically derived utility value 4.7775 lies within the 95 % confidence interval of the utility range estimated for the utility function (UMRR, SR, TWR). The determined MRR results from three experimental runs exhibit strong consistency, which confirms the stability of the selected process parameters (Table 14). The increased MRR measurement shows that the machining method succeeds at material removal while maintaining stable control of additional performance characteristics. MRR measurements between experimental runs show slight variations due to sparks energy variations and changes in material intrinsic properties. The measured surface roughness falls within a restricted range that represents outstanding process stability. Ta b l e 1 4 Findings of confirmation experiments using optimal values of process variables Response parameters Optimal values of process variables Obtained experimental value Average value R1 R2 R3 MRR (mm3/min) A2 B3 C1 D3 E1 8.825 8.898 8.832 8.852 SR (µm) 2.812 2.829 2.813 2.818 TWR (mm3/min) 0.143 0.154 0.147 0.148 The surface quality improves when the assessment value of SR decreases towards 2.8 µm, as this benefits precise components that need minimal manufacturing post-treatment. The selected parameters yield effective optimization of surface quality because run-to-run variations remain low. The production benefits from low TWR when electrode life extends and the cost of machining declines. The process stability is confirmed by the average TWR value of 0.148 mm³/min, as this indicates that the machining tools remain functional for extended period before electrode replacement is required. Small differences in TWR result from irregularities in electrical discharge power and workpiece material composition distribution. Confirmation tests proved that the optimal manufacturing parameters generated from the multi-objective approach deliver expected results. The obtained experimental results matched well with the predicted values,
RkJQdWJsaXNoZXIy MTk0ODM1