An iterative framework for crack nucleation and propagation in high cycle fatigue

The objective of numerical frameworks properly devoted to fatigue life estimation is becoming of great interest in the mechanical field. Indeed, these approaches allow for the investigation of both crack nucleation and propagation stages, reducing the amount of cost and time resources that are necessary in classical experimental approaches. Unfortunately, current solutions necessitate either the crack nucleation definition or multiple mesh refinements that require specific hardware equipment for the simulation process. This work presents a numerical framework for predicting the fatigue life and fracture behaviour of specimens under High Cycle Fatigue (HCF) conditions. An iterative Python-based algorithm was developed and integrated with the open-source Finite Element Method (FEM) solver Code_Aster for simulating both the crack nucleation and propagation stages. The nucleation stage was investigated using the multiaxial critical plane formulation proposed by Smith-Watson-Topper, while the propagation stage was modelled by means of the eXtended Finite Element Method (XFEM) combined with the Paris law. Then, the proposed numerical framework was validated through experimental activities conducted on two different materials, i.e., C45 and 20MnCr5 steels. A detailed material characterization in terms of hardness, roughness, static and fatigue properties was performed with the specific purpose of calibrating the developed numerical framework. A proposal of an innovative curve for nucleation prediction was also presented for a complete fracture behaviour analysis based directly on S-N curves. Thus, experimental and numerical results were compared in terms of crack analysis and number of fatigue cycles, allowing for the validation of the presented work.