I am simulating soil pile interaction behavior. The other errors are " too many attempts made for this increment " or " time increment required is less than the minimum specified ". Is there any problem with the contact element? I am using a master-slave surface with penalty function.
You should provide details on the type of analysis and solver used for us get a grasp on your problem.
It is clear that convergence with contact is something that can generate a lot of issues, especially if you have large deformations involved in the contact area. Usually contact resolution drastically decrease the incremental steps that the solver can do, and causes the errors you mention.
If you specified a fixed time increment regarding the frequencies of interest, you should consider it regarding the stiffness of your contact area. Time steps between 10^-8s and 10^-5s are common for contact problems.
If you use the penalty formulation, make sure you calibrated your coefficient to be more that the relative stiffness of your interfaces, but not too high as it will induce numerical conditioning problems (in relation with your time step). I suggest you check this value and its correspondence in your unit system. Maybe decreasing it a few orders will not change the results but help the solver to converge.
Make also sure that you are using elements suited for contact. The 10 node tetrahedron has a variant C3D10M that is recommended for contact problems for example.
Other common strategies to ease convergence is to start with activated contact (using the adjust option in the contact pair), use conforming meshes at the interface, perform small load increments when starting from a zero deformation, ....
You also have numerical tweaks to allow Abaqus to perform more SDI iterations, release convergence tolerances on very difficult steps, etc.
I cannot be more precise with what you are reporting. I hope this helps.
- Badges
- Science topic