Machine learning–assisted evolutionary optimization of hole quality and surface integrity in abrasive waterjet drilling of polycarbonate

Abrasive Waterjet Machining (AWJM) is a reliable non-traditional technology for machining polymer-based engineered materials with low thermal degradation, dimensional inaccuracy, and surface damage. This study examines polycarbonate hole drilling performance using 125 full-factorial experimental trials using process parameters water pressure, standoff distance, and traverse rate. With a coordinate measuring machine and stylus-based profilometer, kerf angle, entry and exit circularity, and surface roughness were measured in a detailed metrological research. Four ML models were created to develop accurate predicting capabilities, with the Random Forest (RF) model performing best in all responses. RF predicted processes accurately with high coefficient-of-determination R² values (0.92, 0.82, 0.82, and 0.92) and low error coefficients (RMSE 0.196, 0.045, 0.042, and 0.452). The best drilling settings were found using evolutionary algorithms like Biogeography-Based Optimization (BBO), Particle Swarm Optimization (PSO), Salp Swarm Optimization (SSO), and Tug-of-War Optimization (TWO). Deng's similarity metric ranked SSO as the best optimizer based on many responses. The SSO produced the optimal settings (W_p = 250 MPa, S_d = 1.5, and T_r = 300 mm/min) for reduced kerf deviation, circularity and surface roughness. Under optimal circumstances, anticipated responses matched experimental results, proving the integrated ML-optimization framework's strength.