Predicting fatigue limits of defective A356-T6 and A357-T6 cast aluminum alloys using a hybrid empirical–machine learning approa

The present study develops a knowledge-based machine learning framework for estimating the fatigue life of Al alloy using a combination of data-driven models and an empirical equation. Three different ML models were developed and tested, including Support Vector Regression, Random Forest and Gaussian Process Regression, to predict the fatigue limit for the spherical defective cast A356-T aluminum alloy, considering different Second Dendrite Army spacing (SDAS) values and different defect sizes. The effectiveness of the models is assessed using standard analytical metrics, like Root Mean Squared Error (RMSE) and R-squared ( ). With R² value 0.99, the SVR model outperformed the others.