Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : September 16, 2019 - September 18, 2019
The contribution of this study is parameter identification for elastic-plastic model of gradient methods. The gradient methods provide us to identify the material parameters by minimizing objective function which represents difference between force-displacement curve of experiment and numerical simulation. In particular, the minimizing direction is determined by the gradients in which the tangential modulus is obtained from the analytical derivation of the objective function with respect to the material parameter in a similar manner to Mahnken and Stein. In this study, we employ Gauss-Newton method in which the gradient is approximated as the first derivative matrix. Throughout the numerical example, force-displacement curve numerically obtained from notched round bar with isotropic elastic-plastic model combined with Voce hardening rule as virtual experimental data is identified by the result with the material parameters optimized by our developed method. Throughout the numerical example, performance of our developed method is demonstrated by identifying force-displacement curve obtained from notched round bar with combination of isotropic elastic-plastic model and Voce hardening rule.
