Modelling the fatigue life reduction of an AlSi9Cu3 alloy caused by macro-porosity

This study focuses on modelling the fatigue life reduction of an AlSi9Cu3 aluminium alloy caused by macro-porosity. The objective was to develop an engineering approach to fatigue life prediction for structural elements with macro-porosity defects. By following this approach a finite element analysis is first applied to assess the dependence of the strain-concentration factor, caused by a failure-dominant pore, on the nominal strain, the pore size and the proximity to the specimen surface. This dependence is then modelled by a nonlinear equation with three independent variables, i.e. the nominal strain, the pore size and the pore proximity to the specimen surface. The parameters of this equation are determined according to the results of the finite element analysis using a real-valued genetic algorithm. The proposed numerical approach was validated using experimental results. For this purpose, cylindrical specimens were manufactured by a high-pressure die casting and three levels of porosity were deliberately introduced into the specimens. The specimens were then tested at several strain levels and the corresponding fatigue life curves were estimated. The statistical significance of the fatigue life reduction due to the porosity level was statistically analysed and the experimental results were compared to the fatigue lives that were calculated with the presented strain-concentration model. The comparison showed a good agreement between the calculated and the experimentally obtained fatigue lives.