Optimal Process Design for Minimum Springback Based on RBF Network and Evolutionary Strategy

Springback is a very important factor to influence the quality of sheet metal forming. Since it is a multi-variable and high nonlinear problem, there hasn’t been an effective numerical or analytical approach proposed. In this paper, a method of generating optimal forming process for minimum springback is presented. The springback process was modeled and analyzed by using the finite-element method, so as to obtain the springback value for different process parameter as the sample signal of neural network. A Radial Basis Function Network (RBFN) was applied to simulate the complex springback process. An improved evolutionary strategy algorithm was used to optimize the identified model for minimum springback. It has been verified that the springback performance is better for sheet metal forming using the improved ES method than normal ES. The results indicate that optimal process parameters can be quickly and accurately accessed through the proposed approach.