An overload retardation model based on constraint factors

Application of a single overload to a constant amplitude loading will significantly affect subsequent crack propagation rate. Therefore, it is necessary to evaluate the influence of overloads on crack growth rates for the fatigue life prediction of structures subject to variable amplitude loadings. The existing fatigue crack growth models for structures subject to variable amplitude loadings can be classified into three categories: the plastic zone based models, the empirical crack closure models and the strip yield models. Of these, the strip yield models can make precise predictions for the overloading effect, but they are too complicated for engineering applications. The plastic zone based models are simple, but they generally include empirical parameters which often lack clear physical significance and have to be determined by a lot of experiments. By combining the advantages of the plastic zone based models and the strip yield models, the three-dimensional constraints near the crack tip are introduced and a new overload retardation model is developed. In this model, the plastic zone induced by the overload is affected by the three-dimensional constraints, and it leads to significant influence on the crack closure and growth rate. On condition that the crack growth rate at constant amplitude loading is obtained, the overload retardation factor of this model can be determined by a retardation test under any overload ratio. The retardation effects of other overload ratios can be predicted by the obtained factor. With the aid of the factor and the sound basis of exact prediction of the crack propagation under constant amplitude loading by the constraint theory, the developed model is proved to be efficient on aluminium alloys such as 2024-T3, 2024-T351, and 6061-T6. In most cases, its life prediction errors are within 20% compared with the results of the crack propagation tests with overloads. The errors are still less than 34% even in large-scale yielding situations. Moreover, from the verification of 2024-T351, the new model is shown to be effective for predicting the thickness effect on overloading retardation.