Abstract: In order to reduce the weight of the lower box of the battery box, the light material aluminum alloy and carbon fiber composite were used to replace the original Q235 steel material. After material replacement, in order to get the best thickness of new materials, a method of optimizing the thickness of aluminum alloy and carbon fiber parts is proposed by using neural network as the proxy model. The response of dimension variable was obtained by Latin hypercube test, and the optimum thickness of aluminum alloy support and reinforced beam and the thickness of carbon fiber angle were obtained based on neural network optimization. Finally, the optimal laying sequence of the box under carbon fiber is carried out in Optistruct. The optimization results show that the strength stiffness and the first mode frequency are improved while the weight of the battery box is reduced.

Key words: carbon fiber composite, battery box, neural network