Probabilistic fatigue life prediction model for alloys with defects: Applied to A206

► A model for the fatigue life prediction based on the statistical distribution of microstructure. ► The model predicted the fatigue life of A206 before and after friction stir processing well. ► The potential extent of improvement with microstructural refinement of A206 was predicted. A model for the prediction of fatigue life based on the statistical distribution of pores, intermetallic particles and grains is presented here. This has been applied to a cast Al alloy A206 before and after friction stir processing (FSP). The model computes the probability of initiating a small crack based on the probability of finding combinations of defects and grains on the surface. Crack initiation and the propagation life of small cracks due to these defect and grain combinations are computed and summed to obtain the total fatigue life. The defect and grain combinations are ranked according to total fatigue life and the failure probability computed. Bending fatigue experiments were carried out on A206 before and after FSP. FSP eliminated the porosity, broke down the particles and refined the microstructure. The model predicted the fatigue life of A206 before and after FSP well. The cumulative probability distribution vs. fatigue life was fitted to a three parameter Weibull distribution function. The scatter reduced after FSP and the fatigue threshold increased. The potential improvement in the fatigue life of A206 for a microstructure consisting of a finer distribution of particle sizes after FSP was predicted using the model.