On Testing Requirements for Viscoplastic Constitutive Parameter Estimation

Abstract

Parameter estimation in viscoplastic models using nonlinear least squares fitting shows potential for reducing testing requirements. An efficient test matrix must include information to estimate each parameter, but must impose load histories in which the material behavior is dominated by only a few of the constitutive parameters at any time. Sensitivity coefficients are derived for each model parameter in MATMOD and evaluated for several load histories. Results show the adequacy of the test matrix performed to provide information for parameter evaluation.