End-Milling of GFRP Composites with A Hybrid Method for Multi-Performance Optimization

The end-milling procedure has been widely used for machining glass-fiber-reinforced polymer composite (GFRP) materials. A complex interaction of reinforcing glass fibers with each other as well as the matrix element during the end-milling process can result in high cutting force (CF), surface roughness (SR), and delamination factor (DF) because of the anisotropic nature of GFRP. To reduce the three responses (CF, SR, and DF) at the same time, the end-milling cutting parameters, i.e., rotating speed (n), feed speed (Vf), and axial depth of cut (d), must carefully be determined. In this study, the end-milling of GFRP composites was investigated by utilizing a full factorial design of trials with three distinct values of n, Vf, and d. Also, a mix of genetic algorithms (GA) and backpropagation neural networks (BPNN) was administered to forecast the responses and obtain the optimized end-milling parameters. The firefly algorithm (FA), GA, and the integration of GA and the simulated annealing algorithm (SAA) were used to discover the best combination of end-milling parameter levels to reduce the responses' total variance. Later, the combination of BPNN and GA-SAA capable of accurately predicting multi-response characteristics and significantly improving multi-response characteristics was obtained through analyzing the confirmation experiment. © (2025), (Faculty of Engineering, Universitas Indonesia). All Rights Reserved.