Crystal plasticity simulations of materials to assess the fatigue response on the microscale are becoming increasingly popular in industrial application. However, in the case of parts made from cast Ni-base superalloys, the amount of pores in a part combined with the fine spatial discretization around the pores needed due to strong mechanical gradients quickly leads to a computationally impractical number of mesh elements. In this paper we show an alternative to explicit crystal plasticity modeling of part-scale porosity by introducing a Monte-Carlo submodel that recombines the fatigue response of single pores predicted by crystal plasticity into the fatigue response of the pore agglomerate. The model can be applied in realtime due to the use of precomputed crystal plasticity results. We demonstrate that fatigue indicator parameters predicted by the Monte-Carlo submodel agree well with those predicted by explicit crystal plasticity simulations. Lastly, we apply the proposed model to study the influence of different porosity volume percentages, pore sizes and pore morphologies on the fatigue indicator parameters of a cast MAR-M247 superalloy.
Read full abstract