Seismic Analysis not running fully for some values of PGA...

[nithinvl]

Hi

I have modeled a portal frame and tried doing non-linear time history analysis. But, at some PGA levels the analysis is not done for the specified duration. It stops in between. Kindly have a look at it. Also, this issue varies with different seismic files.

eg:
The analysis will be not be completely done for specified duration of analysis (say, 100s) for PGA 0.9g.
But when I make the PGA to say, 0.907g, the analysis is completely done for the specified duration.


I am attaching my model file. Kindly have a look at it and please resolve my issue. I am not able to progress forward because of this issue.

This is the link to the model file.
http://www.2shared.com/file/VfnpBFw2/PD_online.html

This is the link for seismic data file.
http://www.2shared.com/document/buO8v2XK/Borego.html

This is the link for units file.
http://www.2shared.com/file/op4NrQ8J/Li ... antR1.html

This is the link for log file.
http://www.2shared.com/document/K_SVDN-2/log.html

This is the file showing duration of analysis (Specified duration of analysis is 100s) done by OpenSees for different PGA.
http://www.2shared.com/document/_mvCFXzx/g_time.html

Thank You.
Nithin.

[vesna]

While some earthquakes may run smoothly the other ones may not. You need to play with the analysis object to try to resolve the convergence problem. Most of the time, the reduction of the time step of analysis can help. I suggest you to also look at your model and see if you can simplify it without compromising the accuracy of the system.

Also look at your results. It can be that some of the ground motions are causing failures of some of your elements.

[nithinvl]

Hello Vesna,

I thank you for your fast reply.

But, analysis is running completely for some specific values, when i do the analysis for higher PGAs.
For example, the analysis will not run fully for PGA 1.0g, but it will completely run for a higher PGA, say 1.15g.

I am doing analysis for about 5000 records. So, if the analysis stops in between, it will be a problem for me. Hence, kindly suggest me a remedy for this issue.

Also, can you please explain the message
"WARNING - ForceBeamColumn2d::update - failed to get compatible element forces & deformations for element: 2(dW: << 7.76905e-008)"

what does 'dW' mean ?.....Is it some sort of tolerance measurement ??

Regards
Nithin

[vesna]

There is no unique solution for this kind of problems. My impression is that you may be able to avoid those cases if you can find a way of simplifying your model without compromising its accuracy.

The error message indicates the convergence problem at the element level. When using force-based elements in your model there are internal iterations at the element level over which you do not have full control. The only parameters you can set up are the number of iterations and tolerance. The default values are 10 iterations and 1e-12 for tolerance. The message about dW tells you the value of the test after 10 iterations. You will get that message only if the test is greater than the tolerance.

[nithinvl]

Thank you....Vesna...for the immediate reply again.....