Abstract:

In order to obtain the forming parameters of materials under the conditions of complex states of stress and strain accurately and improve the accuracy of finite element numerical simulation of sheet metal forming, a new method based on the advanced RBF surrogate model was  presented. By introducing a radial correction coefficient into the radial basis kernel function of the RBF surrogate model, and using PSO to optimize the radial correction coefficient, the approximation accuracy of the model could be improved. Applying the PSO-RBF model to a nonlinear function, it shows that this method can improve the accuracy of the model effectively. At the same time,the PSO-RBF model was applied into the parameter reverse of sheet forming. The results show that the material parameters based on the PSO-RBF model can reflect the sheet forming materials flow trend and strain distribution more accurately.

Key words: sheet forming;radial basis function(RBF);particle swarm optimization(PSO) , algorithm;parameter inverse

摘要：

为了准确获取材料在复杂应力应变状态下的板料成形本构参数,提高板料成形有限元数值模拟的精度,提出了基于改进径向基函数代理模型的板料成形参数反求优化方法。将径向修正系数引入径向基函数（RBF）核函数中,利用粒子群算法（PSO）对径向修正系数进行优化,提高模型的预测精度。将PSO-RBF模型应用到一个非线性测试函数中,结果表明,PSO-RBF模型比RBF模型的预测精度提高很多;同时将PSO-RBF模型应用到板料成形本构参数反求中,代替有限元模型进行正问题计算,可节省计算成本和提高效率。结果表明,基于PSO-RBF模型反求优化得到的材料参数,能够更加准确地反映材料的流动趋势以及应变分布。

关键词: 板料成形, 径向基函数, 粒子群算法, 参数反求