Identifying the relevant parameters for a mesoscopic fatigue criterion using crystal plasticity finite element simulations of polycrystalline aggregates

International audienceMechanical components subjected to cyclic loadings suffer from fatigue phenomenon. In the case of metallic alloys, microstructurally short fatigue cracks might initiate at the surface. Initiation and micropropagation stages of these cracks generally represent more than half of the total fatigue life. Experimental studies showed that local microstructure parameters, such as crystalline texture and grain boundaries, play an important role in this process. Dislocations motions produce persistent slip bands emerging at the surface and create intrusions and extrusions from which cracks initiates. The local plastic strain mechanisms and their interaction with the microstructure are the object of numerous studies. The number of possible local configurations being infinite, a very large scatter is observed in high cycle fatigue test results, thus a statistical analysis becomes mandatory. To give better estimation of fatigue life, a mesoscopic scale fatigue criterion is needed. The study is based on finite element computations of polycrystalline aggregates On the one hand pure synthetic microstructures are used as input to investigate the problem with a statistical point of view. On the other hand, the input came from a three-dimensional microstructure characterization of a cracked zone extracted from a sample after fatigue loading, carried out by serial Electron BackScattered Diffraction measurements followed by mechanical polishing steps. On both approaches, local mechanical fields were analyzed to identify the relevant parameters to consider as the local fatigue criterion input. A fatigue indicator based on the intrusion-extrusion formation gave encouraging results, when compared to experimental results in the cracked grain. However, it was not able to discriminate this grain from the surrounding ones. A remarkable improvement was observed when taking into account the morphology of the grain in the indicator formulation