Application of GA-BP artificial neural network to formability valuation for steel materials

Abstract

Hydroforming is the most widely used process in manufacturing steel products. Since the quality of the final parts are mostly influenced by formability of steel materials, How to determine the formability of steel materials becomes the key to improving the part quality. Back Propagation (BP) Neural Network has the ability of self-studying, self-adapting, fault tolerance and generalization was applied widely in the formability valuation of steel formability. But there are some defaults in its basic application, Such as low convergence speed, local extremes and so on. In this paper, a generic algorithm-back propagation neural network model (GA-BP) is developed to map the formability of steel materials. Firstly, a back-propagation artificial neural network is built, then 200 samples obtained from FE simulations to train and test the GA-BP ANN. Finally the verified experiment is carried out in a special hydroforming press. The results show that the proposed method is an effective tool for the evaluate formability of steel materials. This method proposed in this paper should provide a time-saving and easy way for post optimization.