Rotating bending fatigue tests were executed on A356‐T6 wheels. The tests were performed both on components containing only the casting defects and on wheels in which artificial holes were machined. Results show the detrimental effect of sub‐superficial porosities on fatigue endurance. A predictive model was proposed. It resulted in a useful tool for the designer because knowing the applied stress, it is possible to estimate the number of cycles for crack nucleation considering the size of the defect, whereas knowing the endurance target requested it is possible to find the maximum defect size acceptable considering the applied stress.

The automotive wheel is a critical safety component in the vehicle and, for such a reason, it has also to meet strict requirements about technological properties. This component is produced by low pressure die casting technique and the casting defects related to the process have to be properly considered having a high effect in decreasing both static and dynamic resistance of the component. Effectively, casting defects like porosities influence the fatigue crack initiation and strongly affect the fatigue life too. One of the most common problem in the real component is the mismatch between the experimental data and literature. In fact, many scientific researches were carried out on small samples produced in a controlled condition and therefore it is difficult to direct transfer the laboratory results to a real cast component with a well-defined shape and different thicknesses.\ud In the present study, an aluminum alloy A356-T6 wheel was analyzed in order to correlate the fatigue performance taking in to account the casting defects.\ud The fatigue limit of the component was studied by rotating bending fatigue tests executed on the whole wheels. Microfractographic analyses on the broken wheels were carried out on the fracture surfaces using a Scanning Electron Microscope in order to identify the crack initiation zone: it was recognized that the crack always started from shrinkage porosities. The statistical population of these defects was therefore investigated on samples taken from the wheel in crack nucleation positions of the spoke and the maximum expected defect size on the component was estimated by the statistics of extreme values. The experimental fatigue limit was finally compared with the theoretical value predicted with the Murakami’s method