A neural network tool for identifying the material parameters of a finite deformation viscoplasticity model with static recovery

In the present paper, the inverse problem of parameter identification is so lved by using neural networks. In contrast to the commonly used optimizatio n methods, neural networks represent an explicit relation between the measu red strain, stress, time and the material parameters to be identified. The constitutive model under consideration describes finite deformation viscopl asticity and exhibits static recovery in both the isotropic and the kinemat ic hardening laws. To train the neural networks, a loading history is utili zed, which consists of a homogeneous uniaxial deformation including cyclic loading and relaxation phases. It is shown that the neural networks are abl e to identify physically meaningful sets of material parameters so that the constitutive model may predict experimentally observed material behavior i n a satisfactory manner. This is true even if complex loading histories are considered. (C) 2001 Elsevier Science B.V. All rights reserved.