large deformation divergence

nylanyla Member


Could you please give me your point of view about the following problem when running my simulation?

Basically there are two bodies (A and B) one (A) surrounds the upper part of another (B) and is bonded to this latter. In fact, I have assigned a frictionless contact between the anterior surface of body B and the inners surface of body A. A displacement is applied to body B in order to move anteriorly until touching the inner surface of body A and unfold it to allow the stretching due to the displacement. the top surface of body A is constrained in the space in the three degrees of freedom.

During the solution, the body B moves correctly against body A (picture 1), but then high distortion occurs at the border that if fixed and divergence occur (picture 2).

Do you have an idea on how to solve the problem?

The material of body A is hyperelastic, Mooney Rivlin to allow a great deformation.




  • peteroznewmanpeteroznewman Member
    edited June 17

    If you have SOLID187 quadratic tet elements, use Keyop(6)=1 for mixed u-P formulation and use many substeps.

  • nylanyla Member
    edited June 18

    Yes, there is solid element 187 10-nodes and I have already written the command keyopt (6)=1. My substeps are 10, 10, 100 since at the beginning no bisection occurs and simulation runs well. The first bisection occurs around iteration 50 that is about 50% os simulation. So do you think I should increase both the minimum and maximum number of substeps? I have also used the snippet command nequit, 100 since it happened days ago that simulation stopped to converge without errors, probably due to the standard number of iterations exceeded.

    A couple of things:

    -I have seen that in the solver output that some elements are associated to SOLID187 how can I know what bodies they belong to?

    -I had posted a topic about mesh refinement "mesh refinement study" since you gave me a response I had labelled it as "is solution", but then I had decided to include a picture of my plot to show you the result and if one of my assumption is possible. Could you please have a look on that even if it was solved? 

  • peteroznewmanpeteroznewman Member
    edited June 18

    If your simulation runs well for a while at substeps 10/10/100, then gets into trouble at about 50% of the applied load, it is a good idea to break the simulation into two load steps. The first load step is half of the total and can complete at 10/10/100 then the second load step to get to the full load can begin at 100/100/1000 substeps. That way, you don't slow down the first half of the application of load.

    Read this post to see how to identify which bodies have the SOLID187 elements. If you want to be sure, you can put a command snippet under each body in the geometry branch of the outline to assign the element type. There are some examples on this site.

    On your last request, I replied in your other discussion.


  • nylanyla Member
    edited June 18

    Hi Peter,

    thanks for the tips. I have run the two steps simulation by increasing the number of substeps in the second half, but it has just diverged without errors except "The solver engine was unable to converge on a solution for the nonlinear problem as constrained.  Please see the Troubleshooting section of the Help System for more information". I've got also this info "The time for solution exceeded 1 hour. A High Performance Computing (HPC) license will reduce your solution time. Consider using an HPC license or purchasing an HPC license from your ANSYS Representative".

    The Newton-Raphson residual plot is the following 

    elements violations are associated with the area represented in the plot.

    However, It seems that the problem is the hardware configuration probably.  I'm trying to run with two additional cores, but I don't think that it solves.

  • nylanyla Member
    edited June 23


    I can definitely say that the problem is not the PC, I have tried to increase the number of substeps as you suggested and to refine the mesh by creating a body sizing in body A and a face sizing in B, but the error is related to highly distortion of elements in the region of the previous image.  I'm thinking the problem is due to the fold of body A since the top surface slightly folds inwards and then stretches out.

    The material is a Mooney Rivlin hyperelastic, thus it is quite deformable.

    Do you have any idea?


Sign In or Register to comment.