Rate Data and Material Model Parameter Estimation

Abstract

Traditionally stress-strain data are used to estimate material parameters in viscoplastic material models. Since the models exist in rate form, they must be integrated in time for the duration of the laboratory tests to give response models for parameter estimation. In this paper the alternative is considered in which the rate form of the material models is fitted to time differentiated stress strain data. Material parameteres, their 95 percent confidence intervals, and their correlation coefficients are compared with their counterparts determined from the more traditional stress versus strain form of the data. It was found that the use of the stress rate form of the data gave well determined parameter estimates with equivalent or superior fitting statistics to the same data can be used to extract significant and distinct sensitivity coefficients to enchance the overall robustness of an optimization algorithm for a parameter estimation.