Creep–Fatigue Life Prediction of GH4169 Specimens Under Uniaxial and Multiaxial Conditions by Integrating Continuum Damage Mechanics and Machine Learning

Creep–fatigue interactions in aeroengine hot‐section components present significant challenges for life design and structural integrity assessment. This study proposes a framework integrating continuum damage mechanics (CDM) with machine learning (ML) to predict the creep–fatigue life of GH4169 specimens under both uniaxial and multiaxial conditions at 650°C. CDM simulations incorporating a multiaxial ductility factor generate a physics‐informed dataset covering different strain ranges, hold times, and stress states for smooth and hole specimens. Artificial neural network, random forest, and support vector machine (SVM) models are developed for crack initiation life prediction and validated against experimental data. The SVM model provides the best predictive performance. Based on the trained SVM model, the nonlinear degradation of creep–fatigue life and its strong sensitivity to hold time, particularly in hole specimens, are further analyzed. This approach provides an efficient and practical methodology for creep–fatigue life assessment of GH4169 components under complex stress states.