SAP2000: NLTHA problem, Nonlinear vs Linear Time History Analysis

[handsome11184]

thread801-423249

Can anyone please explain that why SAP2000 results in higher deformation through linear time history analysis in comparison to its corresponding nonlinear time history analysis especially where plastic hinges form which is really weird.

Thank you very much in advance.

 

[JoshPlumSE]

Impossible for me to tell anything with any confidence. Though I can tell you what I would do:

1) Reduce my force input to the point where the non-linear results in no plastic hinging. Do the results match now?
a) If not, then you know there is something with the solution scheme or time steps or such that don't match up well.
b) If they do match then I would take a look a where the plastic hinges are forming in the analysis with the higher loads. You do get a lot of energy absorbed by plastic hinges. Could it be that this locally increases the deflection at these locations, but that it prevents the energy from getting into the portion of the structure you're currently using to compare the analyses.

 

[handsome11184]

Hi Josh,
Thank you very much for your reply.

First of all I should clarify that I am modeling a bridge structure.
1) I did reduce the structural load to linear level where the plastic rotations at the most critical hinge is negligible. Although both nonlinear and linear analyzes showed almost the same level of deformations, again, the displacement at the critical pier (where plastic hinge shows tiny plastic rotation) is a little less in nonlinear analysis.
2) I have tried 0.001 in time step. Still getting the same results.
3) Yes. You are right. Plastic hinges absorb energy but through plastic rotations which highly increases the displacement of the top if critical pier. However, I even obtain less deflection than linear analysis there. And, this is what driving me crazy.

This is what else I have done:
1) I set the damping to zero. linear displacement >>nonlinear displacement
2) I was originally using fiber type plastic hinges. I change the plastic hinge to idealized bi linear. Still, linear displacement >>nonlinear displacement.

I really need help with this please!

 

[bones206]

Take a close look at the hinge definitions to make sure the initial stiffness is correct. Perhaps the hinge elements are unintentionally adding stiffness to the system in the elastic range.

Are the deflected shapes the same in the linear-level-load for both models? Any differences?

Also, are there any differences in the fundamental frequencies in the two models?