Convergence Error

[darmawan]

Dear STKO Team

I am dealing with pushover, cyclic and nonlinear time history (NLTH) analysis of prestressed pile supported slab viaduct structure.
I use N, mm units in my STKO model.

I also model the structure using SAP2000.

I tried to adopt the STKO e-learning course to perform my STKO models, but i am still facing these problems:

• STKO gave natural period of 0.459 s, while SAP2000 gave 0.665 s. I used the assembled joint mass obtained from SAP to be assign as mass in OpenSees. So, I think the problem was on the stiffness not the mass. Is there any recommendation to "reduce" the stiffness of my OpenSees model?

• The pushover analysis failed in displacement about 70 mm. I tried to using different algorithm, but the analysis still did not converge.

• The cyclic analysis also failed to finish. The analysis stop in cyclic 31. is there any suggestion to make the analysis better?

• I assign the cyclic load condition at all top pile nodes, but in the cyclic analysis command, the selectionset for displacement control integrator was just for 1 vertex. Is it true?

• I tried to perform NLTH analysis using openseesMP. The analysis did not work well. and the monitor did not appear.

I uploaded some figures and also my STKO scd files. Any suggestions/recommendations will be very helpful and appreciated. Thank you in advance.

Best Regard,
Darmawan

[STKO Team]

Try to solve 1 issue at a time.
Start with the modal + pushover.

I cannot help you with the comparison to SAP2000. You have to make sure that everything match between them, to expect the same exact result.
Furthermore, you have nonlinear materials and pre-stress, try to understand how does SAP handles them in the Eigenvalue analysis.
For example, in your model, I made an Eigen also after the gravity and the results change due to cracking and application of the initial stress.

Regarding the convergence issues:
I changed something in your analyses (look at the attached file). However, you will still have convergence issues because your concrete model is extremely brittle. You should try to increase the fracture energy in both tension and compression.
I already changed it a bit, so that you can see how it can be done.
However, keep in mind that the issue of fracture energy is quite important. Keep in mind that in a force-based element it can depend on the size of the gauss-point. Do some google search about "Fracture energy regularization".

I also would change the model to a more robust one. The concrete07 and concreteCM can give very weird results under certain circumstances.

PSSV_Modal & Pushover Analysis.zip

(207.97 KiB) Downloaded 198 times

[darmawan]

Thank You!

[STKO Team]

You are welcome!