仿真驱动的汽车轻量化模具多目标优化设计

打开文本图片集

中图分类号：TG76 文献标志码：A

Abstract：Automobile lightweight technology can reduce the quality and use cost of the whole vehicle without sacrificing safety and stability. In order to give full play to the advantages of various lightweight materials and meet higher lightweight requirements， this paper uses mixed materials to design automobile parts. Aiming at the problems of unstable casting quality and low production efficiency in injection molding process， a multi-objective optimization strategy combining genetic algorithm （GA） and radial basis function neural network （RBFNN） was proposed. This method takes advantage of GA's global search ability and does not depend on the gradient information of objective function，so it is suitable for dealing with complex problems such as discontinuity and non-differentiability. At the same time，with the help of RBFNN's good fitting and prediction performance of nonlinear relationship，the optimization accuracy and efciency are improved. The method of this study is not only suitable for plastic injection molding process，but also covers the optimization of metal casting process， thus achieving a wider range of applications. The results show that the optimization method achieves the goal of weight reduction to a certain extent，and improves the structural reliability.

Key words： Simulation driven； Automotive lightweighting； Mold manufacturing； Multiobjective optimization;Genetic algorithm;RBF

0 引言

随着计算机技术的不断进步，模具注射成型工艺得到了显著提升[1。（剩余6841字）