Surrogate Model to Predict Microstructure and Mechanical Properties in Stainless Steel Cladding under Reactor Operating Conditions

A machine-learning surrogate model was developed to provide rapid predictions of microstructural evolution and service lifetime for reactor materials under conditions of varying temperature and irradiation dose rate. To acquire high-fidelity training data, a Kinetic Monte Carlo (KMC) model was developed with the capability to simulate M₂₃C₆ and G phase precipitation kinetics in a 316 series stainless steel cladding. Experimentally reported behaviors of 316 SS in literature were linked to the kinetic parameters of the simulated precipitation in our model. The temperature and irradiation dose rate histories were generated synthetically for periods of up to 10,000 hours for the simulations. Precipitation progress parameters, such as volume fraction, number density, and particle size were then correlated using statistical methods to develop the surrogate model of these behaviors. Simultaneously, the mechanical properties of the simulated microstructures were evaluated using the microstructure-based Finite Element Method (FEM) analysis. The fidelity of this surrogate modeling to the predictions of our KMC model and FEM analysis is discussed.