Multi response optimization of thrust force and delamination in carbon fiber reinforced polymer (CFRP) drilling using backpropagation neural network-particle swarm optimization (BPNN-PSO)

Carbon fiber reinforced polymer (CFRP) composite materials used in the aircraft structural component, such as wings and rudder, have increased significantly. Drilling of these composite structure is essential to install fasteners for assembly. Thrust force (F_z) and hole exit delamination (FD_e) are the responses that used to evaluate the performance of drilling process. The quality characteristic of these responses is "smaller-is-better." This experiment aims to identify the combination of process parameters for achieving required multiple performance characteristics in drilling process of CFRP composite materials. The three important process parameters such as drill geometry (Pa), spindle speed (n) and feeding speed (V_f) were used as input parameters. Drill type was set at two different levels, while the other two were set at three different levels. Hence, 2×3×3 full factorials were used as design experiments, the experiments were replicated three times. The optimization was conducted by using the combination of backpropagation neural network method and particle swarm optimization method (BPNN-PSO). The architecture of developed BPNN network had three input layers, one hidden layer with nine neurons and two output layers. The activation functions of the hidden layer, an output layer, and network training were tansig, purelin and trainlm respectively. The minimum thrust force, torque, hole entry delamination and hole exit delamination could be obtained by using x type drill geometry, spindle speed and feeding speed of 2673 rpm and 153 mm/min respectively.