Initializing type-2 residual stresses in crystal plasticity finite element simulations utilizing high-energy diffraction microscopy data

Abstract

Crystal plasticity finite element models have advanced in recent years by increasing their fidelity via incorporating physically based deformation mechanisms and more realistic microstructures. In order to better understand their effectiveness, researchers have looked at comparing their results with that of experiments at the grain level. In most of these efforts, the importance of incorporating initial residual stresses, representing the true grain morphologies, and imposing correct boundary conditions, which may have a significant effect on the results, is often overlooked. This work utilizes an available dataset of high-energy X-ray diffraction microscopy experiments conducted on a titanium alloy, Ti-7Al, providing complete grain averaged elastic strain tensors to investigate these three issues in crystal plasticity models. In this study, a method to initialize grain-level residual stresses is formulated and its effect on the correlation of results between simulations and experiments, is evaluated. The effect of grain morphology on the simulation results is evaluated by comparing results from simulations using exact 3D grain morphology with that of simulations using tessellated grain structures. In addition, the importance of applying physically realistic boundary conditions, obtained directly from experiments is also investigated. The findings of this work indicate that initial residual stresses and physically realistic boundary conditions play a key role in the results obtained from CPFE simulations and it is imperative that more importance is given to them. On the other hand, using exact grain morphology is not of much benefit while comparing data on the grain averaged scale as long as the microstructure contains information about the position of the grains, as is the case with tessellations, although grain morphology plays an important role in predicting intragranular strain localization.