Forecasting microhardness, surface roughness and metal removal rate in electrical discharge machining using GM(1,N|sin) power model

In the present study, a novel multivariate grey model called the GM(1,N|sin) power model was used to forecast metal removal rate, surface roughness and microhardness in nitrogen gas–liquid mixed electrical discharge machining. Tests were performed at various current levels, duty cycles and gas pressures and experimental data including metal removal rate, surface roughness and microhardness were collected. A novel periodic bidirectional tool rotation was added to the experimentation. The effect of process parameters on metal removal rate, surface roughness and microhardness was explored through analysis of variance and microstructural tests. The findings revealed that both current and gas pressure have a significant effect on the experimental results. When nitrogen is added to the base metal, hard intermetallic phases are formed within the alloy, resulting in strengthening. The nitrogen in the base metal caused fine and uniform grains and increased microhardness. Periodic bidirectional tool rotation resulted in sufficient dielectric fluid being transmitted to the machining zone to flush debris out of the machining zone and reduce surface defects. The grey model GM(1,N|sin) power model combined the sinusoidal function and power exponential to address nonlinear and periodic oscillation relationships between input and output variables. An accumulation generation operator and mean were introduced in this model and reduced uncertainty in the system. The performance of the GM(1,N|sin) power was measured by estimating the MAE, RMSE and R 2 values of 0.0765, 0.0853 and 0.9836 for MRR; 0.033, 0.0391 and 0.9874 for surface roughness; and 11.5806, 13.743 and 0.9739 for microhardness, respectively. These results indicated that the suggested model had a high degree of prediction accuracy. Consequently, the model proves to be an effective approach for solving nonlinear problems.