When mixed mode loading conditions are applied, fatigue crack path deflection can occur under two different stable modes. The crack may either deviate under mode I (tensile mode) controlled conditions, or under mode II (shear mode), coplanar controlled propagation. Therefore, fatigue crack growth prediction methods are essential for component safety and durability design practices. This study explores the possibility of training artificial neural networks (ANN) to predict mixed mode fatigue crack growth. To achieve this, fatigue crack growth simulations on a compact tension shear (CTS) specimen, using Finite Element Analysis (FEA), were performed. These simulations reproduced various loading conditions, ranging from proportional loading to non-proportional mixed-mode conditions, by incorporating an additional superimposed static loading. Crack propagation direction was analysed using both the Maximum Tangential Stress (MTS) and Maximum Shear Stress (MSS) criteria and was validated against different author’s experimental testing. Then, an Artificial Neural Network was trained and tested using the obtained results. The network was capable of predicting the type of fatigue crack propagation mode that would occur, as well as the actual fatigue crack path.
No takes yet. Share an insight, caveat, or question.
Moita et al. (2025) studied this question.