SIF XFEM

[Naman Verma]

I was looking up some fracture mechanics techniques using XFEM.. My problem is a plate with edge crack pure mode 1 loading. In the outputs of XFEM contour integrals. I requested for 5 contours.. Each contour gives different values of K.. (SIF) there are two things.. here.. XFEM_1 results are one thing and XFEM_2 results are totally another.. What are those XFEM_1 and XFEM_2 in output database.. ? Anybody ? ABaqus manual ddoesnt mention it at all. Can someone give a detailed practical tips to use XFEM for contour integral calculation and especially how to look for and how to make sense out of results n know we have got correct ones.. Thanks in advance

 

[FEA way]

Similar rules as the ones used in non-XFEM contour integral evaluation apply here as well. A common practice is to ignore the results for the first two contours and then average the remaining data (usually omitting output from nodes at each end since it tends to stand out from the rest unrealistically).

 

[Naman Verma]

Thank you for the reply. In my case, these are results that I am getting the Fifth contour results to seem to be far away from the rest of the contours. Should I take an average of the first four? or just ignore the first and last ie 1 and 5 contours and take an average of the rest. And could you please clarify more why we do that average?

 

[FEA way]

You may have to refine the mesh around the crack to reduce the differences. Then, once you are sure that the overall results are good, you can remove those that stand out strongly from the rest. Averaging is needed to get a single value of each K factor for a crack since that’s what we usually want to get from the analysis.

 

[Naman Verma]

WHILE CALCULATING SIF IN XFEM, Why it give results for two cracks ie XFEM_1 and XFEM_2? Which on is correct?

 

[FEA way]

According to the header of this table, there is a single crack and two crack front node sets. Check the documentation chapter "Contour integral evaluation" and make sure that your XFEM crack definition is correct.