Research on laser processing technology of instrument panel implicit weakening line based on neural network and genetic algorithm

Abstract

Laser weakening technology has been found to be widely used in the processing of automobile instrument panel implicit weakening line. Pulse width, defocusing amount and processing speed were selected as the influencing factors of the experiment, and residual thickness was taken as the evaluation index. Orthogonal experiments were designed to explore the influence of process parameters on residual thickness. The Back propagation (BP) neural network residual thickness prediction model is constructed, and the results show that the maximum relative error is 10.96 % and the minimum error is 2.21 %. The weight and threshold of BP network are optimized by genetic algorithm to improve prediction accuracy, stability and convergence speed in training. The convergence speed of the optimized neural network is faster, and the maximum prediction error is less than 2.5 % and the minimum is 0.12 %. Finally, the optimized GA-BP neural network is used to predict the processing results under different process parameters. According to the requirement of energy consumption, processing efficiency and error, appropriate process parameters are formulated.