Numerical Investigation of Mixed-Mode Crack Growth

Abstract

Fatigue crack growth poses a critical challenge to service life of components, especially in safety-sensitive industries such as aviation and railway transportation, where structural components are subjected to millions of load cycles during service. Reliable prediction of crack growth direction and rate is essential for damage-tolerant design and maintenance planning. Stress intensity factor-based criteria are widely used when the crack is loaded in one single mode. However, their applicability for mixed-mode crack growth under non-proportional loading conditions remains limited. The Vector Crack Tip Displacement (VCTD) criterion has been proposed in the literature as an alternative, offering advantages in generality by relying on displacement measures rather than assuming linear elastic material with small-scale yielding. However, its performance under general mixed-mode fatigue loading conditions requires further investigation. In this study, several mixed-mode fatigue crack growth experiments from the literature are modeled within a two-dimensional Finite Element (FE) framework. The experiments include Four Point Bending, Compact Tension Shear, two variations of biaxial stress experiments, and a Twin-disc experiment. The performance of three formulations of the VCTD criterion are investigated in terms of their ability to predict crack growth direction. In addition, crack growth rate using Paris law and the resulting fatigue life are assessed independently along the predicted crack paths, and sequentially compared against experimental data from the literature. The VCTD criterion successfully captures crack growth direction in cases where the crack propagates along a straight path, with good agreement between predicted fatigue life and experimental results from the literature. However, all three formulations consistently fail to predict abrupt changes in crack direction, such as kinking and branching behavior. The origin of this discrepancy is discussed, considering the following possible contributing factors: microstructural features that may drive kinking in the physical experiments in ways not captured by the FE model, simplifications inherent to the FE framework that may affect the displacement fields at the crack tip, and a potential fundamental limitation of the VCTD criterion itself, which may lack a mechanism to detect abrupt directional changes.