Whether crack initiation and propagation in the finite element model could be obtained from the results of Fe-safe? Is it possible to know how the crack propagates?
we have done this in research projects with our fatigue code (now in Siemens Simcenter 3D) for crack initiation and at that time a university code for crack propagation. That worked out fine once the geometry (size, orientation, sharpness, etc.) was selected correctly.
The latter was done in that project by an expert, and (at least at that time) nobody really knows how to automate this step.
Today the second part typically is typically done with XFEM but the problem of the geometry of the initial crack is AFAIK still open.