Multiscale approach with RSM for stress-strain behaviour prediction of micro-void-considered metal alloy

This paper presents the concept of using a representative volume element (RVE) in a multiscale approach to predict the macroscopic stress-strain behaviour of a cast SS316L specimen under tension up to the point prior to necking. RVE models with various micro-void spatial configurations were built, and the effects of micro-voids and strain rate on the material properties (e.g., yield strength, ultimate tensile strength (UTS), ultimate tensile strain and strain hardening coefficient) were analysed. The spatial configuration of the micro-voids inside the cast SS316L specimen was acquired by the X-ray CT scanning system and each micro-void in the gauge length part was converted into a matching RVE model in the finite element (FE) analysis. Response surface methodology (RSM) was employed to investigate the effect of RVE configurations, i.e., the size of the RVE and the shape and spatial location of the micro-voids, on the material properties (yield strength and UTS) of the cast SS316L specimen at the macroscopic level, and then the optimal levels of the RVE configuration were determined. The stress-strain curve from the simulation did show a good agreement with the experimental results and hence the proposed concept was verified.