Dear Code aster comunity,
I'm doing a 3D crack propagation study on a gear tooth root for my master thesis and the simulation stops after few cycles because the first propagation results in two cracks fonds. I'm using in DEFI_FISS_XFEM the ellipse pre-set shape inserted half way into the tooth root (as shown in the picture attached). On the other hand, the problem is less influent when using the shape demiplan (wouldn't be correct to use it for the validity of the study though)

How can i deal with this problem and let the simulation continue?
Should I reduce the number of NB_POINT_FOND?
I tired specifing which fond to propagate with NUME FOND in CALC_G_XFEM but the simulation stops anyway.
I hope you can help since it is the last step to finish my thesis.
Thank you in advance,
Andrea