Abstract
Accurate prediction of creep and fatigue life in metal alloys is critical for ensuring the structural integrity and reliability of engineering components. This research presents a novel approach that leverages machine learning and physics-informed models to enhance the accuracy and robustness of these predictions for carbon and stainless-steel alloys. By incorporating domain-specific knowledge such as material microstructure, loading conditions and environmental factors into the machine learning framework. We capture the complex interactions that influence creep and fatigue behavior by employing physics-informed neural networks (PINNs) to incorporate governing equations for creep and fatigue enabling the model to learn from data while adhering to the underlying physical principles. We also employ Bayesian inference to quantify uncertainty in the predictions providing reliable estimates and confidence intervals. Our findings demonstrate the superior performance of the proposed approach compared to traditional methods offering a valuable tool for engineers in designing and maintaining durable and reliable metal alloy components.