A Computational Inverse Technique to Determine the Dynamic Constitutive Model Parameters of Concrete

摘要

In this paper, a computational inverse technique is presented to determine the constitutive parameters of concrete based on the penetration experiments. In this method, the parameter identification problem is formulated as an inverse problem, in which the parameters of the constitutive model can be characterized through minimizing error functions of the penetration depth measured in experiments and that computed by forward solver LS-DYNA. To reduce the time for forward calculation during the inverse procedure, radial basis function approximate model is used to replace the actual computational model. In order to improve the accuracy of approximation model, a local-densifying method combined with RBF approximation model is adopted. The intergeneration projection genetic algorithm is employed as the inverse solver. Through the application of this method, the parameters of HJC constitutive model are determined. Results show that the identified constitutive parameters' computational penetration depth and projectile deceleration-time curves are closely in accordance with experimental data. The proposed inverse approach is a potentially useful tool to effectively help identify material parameters.