Modelling microstructure evolution, creep deformation and damage in Type 316H stainless steel

The current Advanced Gas-cooled Reactors (AGR) in the UK are approaching their design life. The reactor operator, EDF Energy, is seeking to extend their life due to ecological and financial reasons. One challenge to plant life-extension is the accurate prediction of high-temperature deformation and failure of Type 316H stainless steel, used in components of the reactor boiler section. Under the operating conditions of these components (470-650°C; 10-300 MPa), creep and cyclic plasticity alter the material’s microstructure and failure response. Empirical models are used in industry to predict the deformation and damage response of Type 316H. Predictions can be in poor agreement with plant-recorded/experimental deformation response under complex loading histories. Furthermore, the models are deficient in capturing the sensitivity of the creep damage process to the microstructural state of the material. The overall aim of this research is to improve the understanding of high-temperature deformation and failure response of Type 316H under common plant conditions in order to make realistic life predictions. This is achieved through development of micromechanical models for deformation and damage, employed within polycrystalline modelling frameworks. An existing physically-based self-consistent model (SCM) for inelastic deformation, developed at Oxford University, captures accurately the global and grain-scale deformation response of Type 316H during short-term plasticity and creep. The model was found deficient in predicting cyclic deformation and stress relaxation. This study enhances the model by incorporating the fundamental physics of these processes. Predictions by the enhanced SCM provide insights into the evolution of deformation, microstructure and residual stress state of Type 316H under plant-relevant loading histories. The insights could inform assessment procedures in industry to more accurately account for the accumulation of inelastic strains and damage. Although the enhanced SCM predicts the global response of the material, localization effects, which are of importance to creep damage processes, are not captured. To address this, a crystal plasticity finite element (CPFE) model was developed. It employs the same micromechanical model as the SCM and predictions of global material response by the two frameworks are in agreement. The CPFE scheme captures stress and strain localization near grain boundaries and it was further extended to describe the grain interface response. A combined CPFE-interface element framework was developed and initially used to study the effects of grain-boundary sliding on creep deformation. Results suggest that the effects of grain-boundary sliding on the macroscopic deformation of Type 316H are limited. Features of the local stress and strain fields at grain boundaries, which could affect intergranular damage response, are captured by the CPFE-interface element scheme. A common creep damage mode in Type 316H under the operating range of interest is intergranular cavitation. Review of the literature confirmed that cavitation in Type 316H is controlled by cavity nucleation, which is not fully understood. In order to provide further insights into the physics of this process, existing strain-based empirical and stress-based (classical nucleation theory) nucleation models were modified in this study by considering experimentally-observed features of cavity nucleation in Type 316H. The models were employed locally within the developed CPFE-interface element framework. Modelling results suggest that the strain-based model as a function of local inelastic strain rate does not explain the physical nature of the nucleation process. By contrast, the modified classical nucleation theory captures both the macroscopic failure response and trends in distribution of cavities and failure in the microstructure. These findings outline key aspects of the nucleation process, which need to be examined experimentally. A number of missing features are identified in the mechanistic model, which need to be incorporated in future unified cavity nucleation theories.