Optimization of Stamping Process Parameters Based on Improved GA-BP Neural Network Model

Reasonable process parameters are the key measures to ensure the quality of stamping products. In order to reduce the risk of cracking and wrinkling of stamping products, an improved genetic algorithm is proposed and used to optimize the weights and thresholds of the BP neural network(BPNN). A surrogate model combining an improved genetic algorithm and BPNN(IGA-BPNN)is developed. Taking double C as the research object, the training samples and test samples are extracted through Latin hypercube. The training output of IGA-BPNN model is obtained by AutoForm simulation, and the mapping relationship between process parameters and forming quality is established. Then the mapping relationship is optimized by IGA to obtain the optimal process parameters. The results show that this method reduces the wrinkling of the flange edge of double C and obviously improves the forming quality.