The Application of Multi-layer Perception Networks in the Parameters Optimization of Stamping Forming Process

In the stamping process of high-strength steels, how to adjust each processing parameters to meet the needs of parts deformed is the key problem to be solved in the craft of stamping forming. The main work in this paper is to gain the optimum parameters by making use of neural network to fit experimental datum of stamping. In the experiment of dieing, three craft parameters can be adjusted, the measurement precisions of formed parts are measured. The measurement precisions are applied as the input of neutral network. The adjustable craft parameters are applied as the output of neutral network. The multi layers perception neutral network is constructed by the theory of generalization performance to the question of dieing forming of high-strength steels. Then the craft parameters are given by multi layers perception network, and the feasibility of above method is verified by the experiment of dieing.