Multi-response optimization of carbon fiber reinforced polymer (CFRP) drilling using back propagation neural network-particle swarm optimization (BPNN-PSO)

An integrated approach has been applied to predict and optimize multi-performance-characteristics, being optimum thrust force (F_Th), torque (M), hole entry delamination (FD_en) and hole exit delamination (FD_ex), in the drilling process of carbon fiber reinforced polymer (CFRP). The drilling operation was performed by using a full factorial design of experiments with two different drill geometry (DG), three diverse levels of spindle speed (n), and feeding speed (V_f). The quality characteristics of F_Th, M, FD_en, and FD_ex were smaller the better. Back propagation neural network (BPNN) was first performed to model the drilling process and to predict the optimum drilling responses. Particle swarm optimization (PSO) was executed to attain the best combination of drilling parameters levels that would give optimum performance. The influences of drill geometry, speeds of spindle, and feeding speed on the responses were examined by using the response graphs. In addition, the scanning electron microscope (SEM) photos of the drilled hole are also provided to show the difference of the hole quality before and after optimization. The outcome of the confirmation experiment disclosed that the integration of BPNN and PSO managed to substantially predicted and enhanced the multi-performance characteristics accurately.