基于数值模拟和遗传算法的粉末材料参数反求

Inverse method of powder material parameters based on numerical simulation and genetic algorithm

提出一种基于试验、有限元模拟以及优化算法的反向求解过程来得到2A12铝合金粉末材料本构参数的方法。利用热等静压机制备圆柱形试验件并获取测量参数,采用MSC.Marc软件中的Shima模型对2A12铝合金粉末热等静压成形过程进行数值模拟,再利用遗传算法(GA)对数值模拟中2A12铝合金粉末材料本构参数进行优化调整。结果表明,粉末体由于包套的屏蔽效应,各个方向上的收缩率有一定差异,成形零件的几何形状与最初并不相似;利用环形件进行模型验证,优化参数后进行数值模拟结果和实际测量结果的相对误差,与直接利用原始参数进行数值模拟结果和实际测量结果的相对误差相比,有明显降低。

An inverse method on powder material parameters was put forward based on experiment,simulation and optimization algorithm of the inverse solution process. Then,specimens were made by hot isostatic pressing equipment,and some geometric parameters were measured. Furthermore,the hot isostatic pressing process for powder material 2A12 was numerically simulated by Shima model in the MSC. Marc,and some of the parameters in the constitutive description were optimized and adjusted by Genetic Algorithm（ GA）. The results show that due to the shield effect of capsule,there are various shrinking ratios in different directions,so the final part is not similar with the original one. With the verification of circle specimen,the results show that the relative error between the simulated results and measured results decreases significantly comparing with the one only using original parameters.