Metaheuristic Prediction Models for Kerf Deviation in Nd-YAG Laser Cutting of AlZnMgCu1.5 Alloy

In the present research, the AlZnMgCu1.5 alloy was machined via an industrial-type Nd-YAG laser cutting process. The Box–Behnken design of response surface methodology was used to plan the trials. The experiments were carried out by varying the nitrogen pressure (4–10 bar), pulse energy (2.5–5.5 J), cutting speed (10–18 mm/min), and pulse width (1.5–2 ms). ANOVA was conducted to assess the impact of process factors on response characteristics. The ANOVA results suggest that nitrogen pressure has the greatest influence on the input process parameters. A detailed investigation was conducted to examine the effects of various parameters on kerf deviation. The metaheuristic algorithms (i.e., Giant Trevally Optimizer—GTO; and Zebra Optimization Algorithm—ZOA) were implemented to determine the optimum process parameters for producing the best performance measures. A comparative analysis demonstrated that the parametric value provided by the GTO algorithm, which adheres to the ZOA method, yielded the lowest response. Optimization using GTO resulted in a 6.71% improvement in kerf deviation prediction accuracy compared to experimental values, while ZOA achieved a 2.37% improvement. Furthermore, GTO demonstrated superior computational efficiency, converging in 5.687 s, significantly faster than the 11.548 s required by ZOA. The optimal solution suggested by the GTO algorithm is further verified using a confirmation test on the random settings. In addition, the surface morphology of the laser-cut kerf surfaces was analyzed using SEM images. Through this, it is confirmed that the metaheuristic algorithm of GTO is more suitable for finding the optimum process parameters.