Support elasticity problem
At first I had modelled linear supports as pinned (fixed in Ux, Uy and Uz).
Was the instability reported for this model?
I tried to re-mesh last story (not helped).
What was the reason for doing this? Do you suspect that the instability is related to this story?
Artur Kosakowski
what do you mean reported?
Was the instability reported when you calculated model with pinned (not elastic) supports?
Nonetheless the value of displacements are rather acceptable, so because of any other options within Robot I have accepted results gained.
Try to create a new load case and define a small load in the node where the instability was reported for d.o.f it was reported (UX displacement -> FX force ; RX rotation -> MX moment etc.) and check the displacements and rotations in this node.
Regarding second question - re-mesh is not at all connected with instability. Model is well meshed. I got that appreciation that error no100 may be connected with meshing, therefore I have tried to re-mesh area, where any changes have been done.
Make a copy of the file and open it in the repair mode.
If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.
Artur Kosakowski
For the needs of your task, I have prepared a smaller model (only one floor - the instability above was also from that model). Then I deleted all of the loads and leave only DLL as load type. I added a small force to the exact node (72576) equals to Fx=10kN and done calculations one more time. This time instability was found in node 9744. Please see attached print screen of those two nodes.
What are the displacements and rotations when you apply FX=1kN in node 9744 for this 'reduced' model?
Artur Kosakowski
This indeed suggests that the instability may be just caused by large difference of values of elements of the stiffness matrix. Are you able to calculate this smaller mode with Skyline solver?
Artur Kosakowski
I have changed solver into Skyline, see attached print screen.
So we can't see if it also reports the instability.
I am using normal types of elements - slabs, walls etc. None of them with extraordinary dimensions...
The difference may be due to other reasons e.g. RLink elements user for linear releases, defining them for all edges of a panel or very small surface element generated during meshing and may be reported for correctly defined models.
If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.
Artur Kosakowski
@Anonymous wrote:
Artur, how can I be sure that results are correct? 🙂
By the end of day I will also try to run model with kz.
The support seems to have nothing to do with the instability as you made the corresponding test: 'The instability occurs even if supports are pinned'.
How shall I model linear releases?
If you have them defined for all edges of a slab you may add RZ support in any of its mesh nodes.
how can I be sure that results are correct? 🙂
This is what we were trying to verify by checking the results for any abnormal values:
'Nonetheless the value of displacements are rather acceptable, so because of any other options within Robot I have accepted results gained.'
and applying test loads in the locations of the reported instabilities which if anything there is wrong should give very large displacements or rotations.
If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.
Artur Kosakowski