关键词: Kriging模型, 模具寿命, 遗传算法, 阻力墙

Abstract: In view of the shortcoming of the traditional structure of flash cave in forging technology, a novel flash structure form for forging die：resistance wall which combines the advantages of conventional flash cave and wedge flash cave is proposed. The optimum parameters of resistance wall structure are obtained through optimization analysis with the objective of reducing the mold wear and the forming load, which provides the design basis for application of resistance wall. The latin hypercube method is used to carry out sampling of parameters, and the obtained samples are analyzed by finite elements simulation. The mold wear Kriging model and forming load Kriging model are established by taking the simulation result as response and the parameters of resistance wall structure as variables. Based on the above approximation models, the wear Kriging model and load Kriging model are converted into single objective function by linear weighting method. The optimum parameters of the resistance wall structure are obtained by using genetic algorithm for global optimization. This method has two characteristics：the Kriging model is suitable for computer simulation and the genetic algorithm is suitable for optimization with implicit functions. The crankshaft is taken as an example to validate the application of the optimization resistance wall structure.

Key words: Die life, Genetic algorithm, Kriging model, Resistance wall