Stress, strain, or energy? Which one is superior predictor of fatigue life in notched components? A novel machine learning-based framework

This paper introduces an efficient framework for accurately predicting the fatigue lifetime of notched components under uniaxial loading within the high-cycle fatigue regime. For this purpose, various machine learning algorithms are applied to a wide range of materials, loading conditions, notch geometries, and fatigue lives. Traditional approaches for this task have mostly relied on one of the mechanical response parameters, such as stress, strain, or energy. This study also concludes which of these parameters serves as a better measure. The key idea of the framework is to use the profile (field distribution represented by some points) of the mechanical response parameters (stress, strain, and energy release rate) to distinguish between different notch geometries. To demonstrate the accuracy and broad applicability of the framework, it is initially validated using metal materials, subsequently applied to specimens produced through additive manufacturing techniques, and ultimately tested on carbon fiber laminated composites. This research demonstrates the effective use of all three parameters in estimating fatigue lifetime, while stress-based predictions exhibit the highest accuracy. Among the machine learning algorithms investigated, Gradient Boosting and Random Forest yield the most successful results. A noteworthy finding is the significant improvement in prediction accuracy achieved by incorporating new data generated based on the Basquin equation.