Crack path determination for non-proportional mixed-mode fatigue

Turbine engine components such as fan and compressor blades experience complex combinations of steady and vibratory loads that lead to in-service cracking in directions that cannot be predicted by current fracture criteria. Accurate crack path predictions are required in order to characterize the risk and extent of damage resulting from liberation of a fractured ligament from rotating components. Under proportional in-phase mixed Mode I/Mode II loading conditions, crack growth direction has been observed in some materials to shift from tensile-dominated Mode I to shear-dominated Mode II or mixed-mode crack growth at higher proportions of initial Mode II loading, but non-proportional loads are not well-characterized. An extensive database of crack growth direction under non-proportional 2-D mixed-mode loading conditions is required to expand crack path prediction models, which are likely to vary between alloys. An approach based on linear elastic fracture mechanics (LEFM) is desired in order to implement the model in crack growth software such as the boundary element-based fracture analysis package FRANC3D.;A novel specimen configuration has been designed and analyzed for generation of wide ranges of mixed-mode loading conditions in a single test. This specimen and a more conventional thin-walled tubular specimen have been used to test polycrystalline nickel-base superalloy Inconel 718 under proportional in-phase and 3 kinds of non-proportional fatigue loading. Stress intensity factors for the various configurations have been analyzed with FRANC3D.;Modal transition from Mode I (tensile) to Mode II (shear) crack branching has been observed in several load cases. Qualitative microscopy of fracture surfaces was used to characterize the crack growth behavior. An LEFM approach based on an effective stress intensity factor range, which incorporates the maximum value and range of each appropriate stress intensity (Mode I or Mode II), has been used to successfully predict the crack deflection angles, and in most cases to quantify modal transition, within each load case considered. Variability between load cases and specimen configurations points to the limitations of LEFM, or at least the stress singularity-based approximation of crack tip stress fields, in providing a general predictor of crack path behavior across all types of non-proportional mixed mode loading.