Pareto-Based Multi-Objective Optimization of Drilling Kenaf/Epoxy Composites Using NSGA-II and MOPSO

The optimization of drilling parameters for natural fiber-reinforced polymer composites presents a significant challenge due to the simultaneous impact of machining conditions on mechanical loading and energy consumption. This study presents a surrogate-assisted, Pareto-based multi-objective optimization framework for drilling kenaf fiber-reinforced epoxy composites. A Taguchi L27 orthogonal array was used to investigate spindle speed, feed rate, and drill diameter, with thrust force, torque, and drilling power minimized simultaneously. Second-order surrogate models were developed and independently coupled with the Non-dominated Sorting Genetic Algorithm II (NSGA-II) and Multi-Objective Particle Swarm Optimization (MOPSO), each run for 500 generations/iterations. Experimentally, thrust force, torque, and drilling power ranged from 40.06–450.03 N, 0.020–1.00 N·m, and 3.65–251.33 W, respectively. Run 20 produced the lowest observed thrust force and torque, whereas Run 22 produced the minimum drilling power, demonstrating conflicting objectives. The surrogate models yielded R2 values of 90.19%, 82.21%, and 86.84% for thrust force, torque, and power, respectively. LOOCV R2 values were 70.91%, 50.20%, and 63.85%, indicating that the models are primarily suitable for interpolation within the investigated domain.