override negative pivots 1

[Hup]

Hi all,
I try to "override" a negative pivot. I am running a static linear analysis with a model containing a superelement. The stiffness matrix inside the SE is created outside Ansys and results from different calculations.

The matrix in SE is not positive definite, but the whole model (including SE) is "positive" (checked by combining all stiffness matrices in another program). Actually the model is really overconstrained but I still get a "large negative pivot" message.

Principally I suppose that even negative pivots in small regions of the model should result in great values of result DOF provided the rest of the model is "positive".

I tryed to switch the pivot chcking off by PIVCHECK,OFF but other errors stop the solution.

Has anybody encountered similar problem?

Thanks H-up

 

[Drej]

Switching the pivot check off will not solve the problem. Take a good look at your boundary conditions and make sure your model is adequately constrained - try a modal analysis on the structure and check for rigid body motion from the eigenvalue output.


------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[Hup]

Thanks for reply, Drej. The model basically consists of several solid-material brancheis fixed by its ends and the central part containing the superelement. So there can not be any kind of rigid body motion there.

Therefor I think the problem should have a physical solution.

 

[bagarbilla]

try a non linear solution .. by checking the non linear option in solution controls.. i suspect that there can be a large deflection occurring in your model...

 

[Hup]

Thanks, Bagarbilla, "NLGEOM,on" really helped to get results.I see that large negative pivot destroyed the original shape, but that's what I needed.

H-up