Thickness Distribution Prediction of Superplastic Forming Parts Based on Machine Learning and Computer Aided Engineering

In recent years, the high-speed and high-performance needs of aircraft put forward higher requirements for superplastic forming process, in which uneven thickness distribution is one of the important factors affecting the performance of formed parts, and it increases the cost of design and manufacturing. Therefore, this study proposes using neural networks to predict the thickness distribution of superplastic formed parts. We established BP neural network and PSO-BP neural network to predict the thickness distribution of cross-section edges of box-shaped parts. The prediction performance of the two models is compared and the results show that the PSO-BP neural network has a higher prediction accuracy, especially in the area of the round corners. It is demonstrated that the method is feasible and significantly improves the computational efficiency.