Optimization and experimental verification of stamping process parameters for battery pack top cover based on GA-BP algorithm

To more accurately resolve the problem of process quality optimization in the top cover stamping process, an optimization strategy combining Auto Form platform and Genetic Algorithm optimized Back Propagation Neural Network intelligent algorithm is projected. Taking the maximum thinning rate of the top cover stamping forming as the optimization objective, a 3D numerical model is established, and AutoForm finite element software is used for numerical simulation. Orthogonal experiments are designed, and the experimental results are subjected to range analysis to determine the primary and secondary relationships of evaluation indicators affecting the top cover stamping forming, as well as the optimal process parameter combination (A1 stamping speed 1000mm/s; B4 blank holder force 80T; C3 friction coefficient 0.12; D2 concave convex die gap 1.26mm), a small batch trial stamping experiment was conducted, and it was found that a small number of samples had wrinkling and cracking. Design Latin hypercube sampling, construct GA-BP database, train the database, optimize to obtain the optimal process parameter combination, and conduct numerical simulation verification. Furthermore, conduct stamping experiments to compare and verify the quality of the samples. It was found that the sample quality was better than that of the top cover stamping experiment under the optimal process parameter combination obtained from orthogonal experiments. This verifies that the parameter optimization model of GA-BP algorithm has higher accuracy and effectiveness than the traditional orthogonal experiment range optimization analysis method.