Simulation-based S-shaped Track Pieces Stamping Resilience Research And Parameter Optimization

In recent years, in order to lightweight design, both the aluminum alloy sheet and high strength steel sheet have already been widely applied in the production of automobile cover parts. It makes springback problem more significant. The precision of the dimension and shape of stamping parts are directly affected which leads to hard assembly. In order to reduce production cost, improve the design efficiency and reduce the time of mold testing and fixing, optimize parameters and control springback based on results of stamping prediction have became the key of intelligent NC precise forming in current research.BenchMark2 "S-Rail-08" of NUMISHEET'2008 was taken as the typical study object of springback. Firstly, the S-Rail bending forming affecting factors, such as BHF (blank holder force), Drawbeads setting, Hardelling index, Mould clearance, Friction coefficient, blank thickness are determined by analyzing the machanic of S-Rail bending forming and single-factor experiments.Then, on the basis of finite element analysis samples and springback orthogonal experiments, Back propagation neural network prediction model and response surface method (RSM) prediction model are modeled and simulated. It is found that both the efficient and accuracy of RSM prediction model are better than a BP prediction model with the same samples.Finally, Non-Dominated Sorted Genetic Algorithm-Ⅱ(NSGA-Ⅱ) is proposed to optimize parameters and reduce springback based on the prediction simulation data. The Pareto results are confirmed by Isight and DYNAFORM simulation. As a result, springback of S-Rail is reduced and formability is also improved. This new approach has laid a foundation for the springback intelligent prediction and control of complicated parts stamping in the further study.