This study examines thrust force behavior in conventional drilling (CD) and ultrasonic-assisted drilling (UD) of multi-layer AA6061 aluminum sheets fabricated via accumulative roll bonding (ARB), integrating statistical design of experiments, machine learning prediction, and meta-heuristic optimization. Drilling tests were conducted on 3–12 layer configurations using three point angles (118°, 130°, 140°), three feed rates (0.08, 0.15, 0.25 mm/rev), and three spindle speeds (500, 1000, 1500 rpm) under both CD and UD conditions. Analysis of variance (ANOVA) identified feed rate ( p 0.98 in training and >0.92 in testing, with mean absolute percentage errors below 8%; feature importance ranked feed rate highest (∼42%), followed by point angle, number of layers, and spindle speed. Optimization via the Grey Wolf Optimizer (GWO) yielded minimum thrust forces of 252.62 N for CD and 211.6 N for UD at the combination of 140° point angle, 0.08 mm/rev feed, 1500 rpm, and 12 layers, achieving desirability = 1.000. Results underline the effectiveness of coupling DOE analysis with RF prediction and GWO optimization to capture process–geometry–material interactions, and demonstrate that UD combined with optimal tool geometry enables significant thrust force reduction in ARB-processed aluminum, offering actionable strategies for low-force machining.
Azar et al. (Mon,) studied this question.