Sensitivity analysis of history dependent material mechanical models for numerical simulation of welding process

Abstract

Yield stress of 6013-T6 aluminium alloy was tested on Gleeble 1500D thermal–mechanical system at predesigned temperatures during different typical thermal cycles, in order to accurately reflect the influence of weld thermal history on material properties. The typical thermal cycles were referred to the temperature field simulation results of real welding process. The changes of yield stress were obtained directly from the stress–strain curves generated by the tensile tests. The tests were more accurate than previous publications, where only the yield stresses at room temperature after thermal history were tested or calculated from microstructure evolution model. Experimental results showed that the changes of yield stress during the cooling stage of typical thermal cycles followed one set of curves. These yield stress–temperature curves were different from those during the heating stage. Temperature and temperature history dependent material model M2 and M3 were established based on the experimental results. M2 model was perfectly plastic model while work hardening effect was considered in M3 model. Compared with conventional temperature dependent material model M1, the distributions of longitudinal residual stress and strain obtained with temperature and temperature history dependent models fit better with published results. Yield stress of the material at the weld zone decreased a lot after having experienced weld thermal history and longitudinal compressive plastic strain at the weld zone recovered to some extent during the cooling stage in M2 and M3 models. These were the main causes for lower peak longitudinal residual tensile stress in M2 and M3 models.