Characterization and simulation of the plastic behavior of steels subject to complex loading histories

Abstract

The present study aims to investigate material plastic behavior of advanced high strength steel (AHSS) under non-linear strain-path and evaluate constitutive model accuracy in finite element (FE) simulations, which includes some amount of complexity of the industrial application. For measuring material plastic behavior after strain-path changes, two-step loading experiment, which consists of two linear loading segment with a single abrupt strain-path change, has been widely adopted. For example, tension followed by tension or shear followed by tension along to different loading directions is the representative procedure for that. However, most of these methods are valid only for limited strain-path. In addition, the pre-strain stage of these methods cannot be conducted with a laboratory scale in the future, because the strength of AHSS sheets is increasing to meet the demands of industrial fields. An optimal experimental methodology that not only has the possibility of various non-linear strain paths but also can be adopted in very high strength of AHSS was proposed in this work. For the validation, a 0.72 mm thick EDDQ and a 1.6 mm thick TRIP1180 steel sheet samples were used. The former is a single-phase steel with an ultimate tensile strength (UTS) of 350 MPa, which is the reference material for low strength. Whereas the latter is multiphase steel with a UTS of 1180 MPa, which is the representative material for high strength steel. Getting uniformly deformed regions as much as possible within a reasonable load range and conducting in-plane biaxial tension tests after the pre-strain loading, that can provide a large variety of loading scenarios by changing the load ratio between the two orthogonal loading directions, were the most challenging tasks for the development of optimal experimental methodology. The dimension of specimens for pre-strain loading and in-plane biaxial tension were determined within experimental constraints through trial and error using experiments and FE simulations. Tension followed by tension with different loading directions, in-plane biaxial tension, and simple shear were successfully conducted with TRIP1180 and EDDQ with the proposed method. To investigate constitutive model accuracies of isotropic hardening (ISO), homogenous anisotropic hardening (HAH), and Yoshida-Uemori kinematic hardening (YU) model, FE simulations were conducted. After simulation, predicted springback profiles were mainly compared with the experimental profile. For this work, dual phase (DP) steel with 1.37 mm thickness and UTS of 780 MPa was selected. For the simple example, a U-draw bending simulation was conducted, because only a single change of load reversal occurred during forming. In addition, for the more complex example that is close to real part forming C-channel forming simulations were conducted. FE analysis was conducted to figure out the reason for the prediction accuracy depending on hardening models. HAH model shows good agreement in not only U-draw bending but also C-channel forming, while YU model can predict well only in U-draw bending. The result of this study indicates that material behavior under various loading scenarios not only for reverse-loading should be considered in the constitutive model to get accurate simulation results of industrial application. In addition, it was found that the most important factor for accurate prediction of springback profile is the final stress, according to FE analysis with the YU model.