Leveraging Machine Learning for Enhanced Fatigue Life Prediction in Aluminum Alloys

This study introduces a data-driven approach, using machine learning (ML), to understand and predict fatigue failure in aluminum alloys. Traditional methods for deriving the S–N curve, which represents the relationship between stress and fatigue life, are costly and time-intensive. Fatigue mechanisms are multifaceted being influenced by numerous factors. The research proposes a novel method for estimating fatigue life at various stress amplitudes, forming the S–N curves. By integrating the key factors that govern fatigue life, ML techniques can effectively predict the S–N curves for aluminum alloys. The dataset used for this study was compiled from industry-accepted sources, and the model was trained using GBR algorithm. The model’s prediction of fatigue life yielded a mean squared error (MSE) of 0.46, indicating a high level of accuracy. Notably, the model was able to identify the features that most significantly impacted fatigue life and accurately predict the S–N curve, aligning closely with experimental data. The aim of this research is to aid material scientists and design engineers in examining the effects of different alloying element compositions on fatigue life. The model offers a preliminary estimate of fatigue life for various alloy combinations, providing a valuable tool for future investigations.