Comparison of physically based constitutive models characterizing armor steel over wide temperature and strain rate ranges

Both descriptive and predictive capabilities of five physically based constitutive models (PB, NNL, ZA, VA, and RK) are investigated and compared systematically, in characterizing plastic behavior of the 603 steel at temperatures ranging from 288 to 873 K, and strain rates ranging from 0.001 to 4500 s ^-1. Determination of the constitutive parameters is introduced in detail for each model. Validities of the established models are checked by strain rate jump tests performed under different loading conditions. The results show that the RK and NNL models have better performance in the description of material behavior, especially the work-hardening effect, while the PB and VA models predict better. The inconsistency that is observed between the capabilities of description and prediction of the models indicates the existence of the minimum number of required fitting data, reflecting the degree of a model's requirement for basic data in parameter calibration. It is also found that the description capability of a model is dependent to a large extent on both its form and the number of its constitutive parameters, while the precision of prediction relies largely on the performance of description. In the selection of constitutive models, the experimental data and the constitutive models should be considered synthetically to obtain a better efficiency in material behavior characterization.