Numerical study of in-plane biaxial fatigue crack growth with different phase shift angle loadings on optimal specimen geometries

Nowadays for real world applications, mechanical components in the automotive, aerospace, aeronautical and other industries are subjected to multiaxial loading conditions. Although the fatigue behavior of materials like steel alloys, aluminum alloys or even magnesium alloys is fairly well established for uniaxial loading, one should not use this knowledge for biaxial loading. Developing new testing machines and new specimen geometries have been the previous goal of several authors. A new generation of smaller and more efficient biaxial fatigue testing machines has arrived on the market. Using electrical motors these machines are not able to achieve the higher loads as their hydraulic counterparts can, and therefore the cruciform specimen needs to be optimized. The authors have previously optimized the cruciform geometry for biaxial fatigue initiation, using a revolved spline to reduce the specimen center thickness. For crack initiation experimental results have proven that the obtained design detail is effective, but there are no studies about the behavior of these specimens for crack propagation. In this paper the authors firstly set out to determine the conditions for crack initiation using traditional criteria like Findley, Brown-Miller, Fatemi-Socie, Smith, Watson and Topper (SWT), Liu I and Chu, as a function of different biaxial loading with phase differences. On a second stage the authors compared the biaxial fatigue crack propagation on the optimal specimens, with the behavior of notched specimens, using the stress intensity factors for mode I and mode II. Several crack and loading parameters were studied, including the starting crack length and angle, and different loading paths. Different biaxial loadings were applied to the model, including 30°, 45°, 60°, 90° and 180° out-of-phase angles. Very similar small crack propagation parameters were obtained for both specimens, although as the crack growths the stress intensity factor for the optimal specimen do not behave as expected. Therefore limiting the use of this specimen for crack propagation.