Hello Simulia community,
I am modeling a solid model (which is a part of a piping with a support) and it will be contacting an analytical rigid surface (which acts as a guide). See attached figure.
The two open ends of the solid model are kinematic coupled to the reference points on which displacement and angular velocity (to act as rotations) boundary conditions are applied.
The analysis is being conducted in two steps. in Step-1, a displacement controlled load is applied on the reference point of the analytical rigid surface to close the physical gap of 0.01" to establish initial contact. In Step-2, the displacement and angular velocity boundary conditions on the solid model are modified to actual values from initial values of zero. Linear elements are used to mesh the solid. I have used General Contact procedure.
The model runs fine. However, to simulate a different condition, I had to rotate the solid model instance about its Y axis by 180°, keeping everything else the same. The solution in this case does not converge from Step-2 onwards. It issues severe discontinuity iterations and the solution terminates after cutting back on the subsequent time step increments. I have tried several variations of the model as below but nothing worked.
- rotating only the solid instance keeping the analytical rigid surface the same.
- rotating both the solid and analytical rigid surface instances together.
- creating a whole new assembly by using the parts and re-defining all the surfaces, contacts, kinematic couplings, loads and boundary conditions etc.
- saving the model as new model and then trying the 1 to 3 variations as above.
- changing the coordinate system from global to local for couplings and boundary conditions.
However, the solution converges when I switch on the automatic stabilization in Step-2. I am unable to comprehend why the model works fine as created, but not when subsequent change, especially the rotation of the instances, is made. Please advice and prescribe a fix. Thanks in advance.
I have also noted in the case of un-converged solutions that the analytical rigid surface elongates in the visualization. What is the reasoning behind it ?
