Robust optimization design for stamping based on GA-BP proxy model and sparse polynomial chaotic expansion
For the purpose of carrying out a robust optimization design of stamping,an optimiza-tion design method based on GA-BP proxy model and sparse polynomial chaos expansions(sPCE)was proposed.Firstly,according to the actual situation of the production site,the stamping pro-cess design variables and forming quality indicators were selected,and then the sample points of stamping process were obtained by Latin hypercube sampling(LHS)and CAE numerical simula-tion.On this basis,the GA-BP proxy model was constructed by combining genetic algorithm and backpropagation neural network.Afterwards,based on sPCE,the uncertainty response of stamp-ing forming quality to design variables was analyzed.Furthermore,the robust optimization model of stamping was established by coupling the GA-BP proxy model and sPCE uncertain response model.Finally,non-dominated sorting genetic algorithm-Ⅱ(NSGA-Ⅱ)was used to solve the ro-bust optimization model in the full stamping design space.Taking a reinforcing plate under pillar A as an example,the robust optimization design of stamping was carried out and then the results showed that this method could accurately obtain the feasible optimal robust process parameter combination,and could be well applied to the robust optimization design of stamping.
NETL
NSTL
万方数据
