Comparison of self-consistent and crystal plasticity FE approaches for modelling the high-temperature deformation of 316H austenitic stainless steel

The present article examines the predictive capabilities of a crystal plasticity model for inelastic deformation which captures the evolution of dislocation structure, precipitates and solute atom distributions at the microscale, recently developed by Hu and Cocks (2015) and Hu et al. (2013). The model is implemented within a self-consistent framework and a crystal plasticity finite element (CPFE) scheme. Through direct comparison between the two CP schemes and with an extensive material database for Type 316H stainless steel, the different types of information and the degree to which the models are consistent with experimental observations are assessed. The study demonstrates an agreement between the SCM and the CPFE schemes, providing confidence in the micromechanical deformation model employed. The multi-scale approach also allows the effects of micro-scale deformation processes, related to dislocation-obstacle interactions, on the global deformation response to be captured. Modelling results from this study and their comparison to experimental observations show that deformation of polycrystalline materials, such as 316H stainless steel, is controlled by the evolution of microstructural state of the material and the redistribution of stress between individual grains. The study suggests that the SCM is a feasible tool to simulate and explain the deformation behaviour of complex alloys under industrially-relevant thermo-mechanical operating histories. The CPFE framework captures the effects of the variation in grain geometry and provides more detailed information about the variation of stress and strain within the individual grains, particularly their distribution near grain boundaries and triple points – which are important to understand in the context of damage development and failure. The SCM predicts a “stiffer”, more creep-resistant response than a CPFE model for a given set of material parameters due to the more highly-constrained deformation modes allowed in the model. As a result, material parameters calibrated using one modelling approach are not necessarily suitable for use in another approach – although parameters obtained when fitting the different models should not vary significantly