Weird crack behaviour?

[Low12]

Hey guys, i have a question regarding crack analysis in ABAQUS.

I am using the CZM to initiate and propagate a crack in a coating. Recently i was playing around with partitions (vertical white lines) so that multiple crack can emerge sequentially. After having a closer look at the output file i noticed, that within only 2 frames the crack orientation changed and the crack kind of "jumped" (withing the white circle area). What could be a possible explanation for such a behaviour? Is ABAQUS well suited for a complex crack analysis or is it just useful if you want to investigate the first crack that appears in the system?

Left picture represents the state 1 increment before the right picture.
Thank you very much in advance.

 

[FEA way]

Try playing around with mesh. Refine it but also use different technique. In this case less regular mesh (like free meshing with advancing front) may be better because cracks may not propagate properly along the element edges.