Dear Vicente, I advise you to check Morfeo/Crack for Abaqus, which allows for defining cracks with the XFEM feature of Abaqus and performing automated crack propagation under fatigue in the frame of LEFM. The software is distributed by GeonX, www.geonx.com. With this software you can perform 3D crack propagation without any manual intervention, and due to the theory being the XFEM and level set methods, you can readily handle crack topology changes. You could probably get an evaluation licence for a short period and then we would probably get in touch in the future as I am also working with GeonX from time to time . Regards. Frederic

so far i was using xfem, but i want to move to use normal fem, xfem has its advantages but allso its drawbacks. 

For 3D crack propagation, you will struggle with Abaqus only. That's the purpose of Morfeo/Crack - based on the XFEM, and competitors such as Zencrack with FEM or even Beasy with BEM.

Been working on / developping the XFEM since 2004 and I really see the power of the method.

If you are keen on using another software than abaqus,please check Morfeo/Crack standalone.

Have a look at my publications with colleagues from GeonX and Cenaero. The Morfeo/Crack solution and plug-in versions for commercial software is used by several industrials, mostly in aeronautics.

Good luck

yes, morfeo could be a solution (as xfem which is really my expertise field), but as i do not intend to do any crack propagation, i am mor einterested in using just normal finite elements; my doubt is what is the best porgram avalaible to build  a mesh for abaqus analysis.

all right then I get your point.

With Morfeo Crack you can load a crack as a step file, a mesh (for instance if you get it from µCT) or an analytical function.

For meshing, there are several options...  the thing is that XFEM does need mesh refinement in 3D just for having converged fields... TO refine the mesh locally I either

- define global mesh size + local mesh size with boxes.

- use the level sets and use the computed distance function to the crack tip as background field in gmsh in order to have a smooth transition between crack tip mesh size and global mesh size.

At GeonX we use either VirFac which is a commercial pre-processor including meshing or GMSH a free software that you can dowload on the web.

Otherwise, other software help you building meshes and models (e.g. products from Altair including hypermesh, or e.g. Ansys Workbench)

TO define a structured mesh for crack boxes with quarter-point elements, then the software you mentioned in your first post do the job.

Regards