关键词: 物性参数　人工神经网络　识别　应变控制

Abstract: It’s very essential to correctly determine the material parameter in constitutive model for appropriately analyzing and simulating the process of metal forming. Based on Hill’s new yield criterion, the approach of artificial neural networks is used to identify the material parameter m in variant stress states during the process of material deformation. By the experiment of tension and torsion test of thin walled tube on MTS , the experimental strain increment is obtained and compared with theoretical strain increment which obtained by simulation with m value 2 and the identified m value. The results of comparison validate the correctness of identified m and deformation control method using the identified m.

Key words: Artificial neural networks, Deformation control, Identification, Material parameter