Failed Convergence Test

[davidli506]

Hello All,

I created a three story frame with concentrated plastic hinges at the end of each element. I followed the example "Pushover Concentrated.tcl" http://opensees.berkeley.edu/wiki/index ... trated.tcl. Instead of applying a pushover analysis I used a Time History one. When my model was ran with an earthquake scale of 0.5 I got the same displacements and natural building frequency as a published research paper that I am comparing my model to.

My problem is when I run the model with a scale of 1.0 Opensees gives an error message of "failed to converge," "After: 100 iterations," and "returned: -3 error flag." I tried to increase the number of iterations, but that did not fix it. Is this a problem with the model being subjected to a loading that is too large?

Thank you,
David

[vesna]

You may want to try changing solution algorithm when it does not converge. In the example manual you will find examples on how to do that.

[davidli506]

Vesna,

When you say change solution algorithm do you mean changing the type of algorithm I'm using to analyze? I am currently using "algorithm newton." I am assuming that I should choose one of the other choices in opensees.

Thank you,
David

[vesna]

Look at this example:
http://opensees.berkeley.edu/wiki/index ... typical.29

Look at files: Ex5.Frame2D.analyze.Dynamic.EQ.Uniform.tcl and LibAnalysisDynamicParameters.tcl to see how analysis was defined for this example and how convergence problem was handled.

[Eli]

Hi
Here is a part of the warning message about convergence problems that I have due to time history analysis:

OpenSees > analyze failed, returned: -3 error flag
WARNING: CTestEnergyIncr::test() - failed to converge
after: 1000 iterations
NewtonLineSearch::solveCurrentStep() -the ConvergenceTest object failed in test(
)
DirectIntegrationAnalysis::analyze() - the Algorithm failed at time 11.2796
OpenSees > analyze failed, returned: -3 error flag
WARNING: CTestEnergyIncr::test() - failed to converge
after: 1000 iterations
NewtonLineSearch::solveCurrentStep() -the ConvergenceTest object failed in test(
)
DirectIntegrationAnalysis::analyze() - the Algorithm failed at time 11.2818

I used the convergence loop as in this example: http://opensees.berkeley.edu/wiki/index ... e_Analysis and http://opensees.berkeley.edu/wiki/index ... typical.29
but I could not fix it.
Is there any way to solve it?

[hongqi]

i meet the same problem,and the code stop running at the time which is ont the ground motion end time.


> Hi
> Here is a part of the warning message about convergence problems that I have due to
> time history analysis:
>
> OpenSees > analyze failed, returned: -3 error flag
> WARNING: CTestEnergyIncr::test() - failed to converge
> after: 1000 iterations
> NewtonLineSearch::solveCurrentStep() -the ConvergenceTest object failed in test(
> )
> DirectIntegrationAnalysis::analyze() - the Algorithm failed at time 11.2796
> OpenSees > analyze failed, returned: -3 error flag
> WARNING: CTestEnergyIncr::test() - failed to converge
> after: 1000 iterations
> NewtonLineSearch::solveCurrentStep() -the ConvergenceTest object failed in test(
> )
> DirectIntegrationAnalysis::analyze() - the Algorithm failed at time 11.2818
>
> I used the convergence loop as in this example:
> http://opensees.berkeley.edu/wiki/index ... e_Analysis
> and
> http://opensees.berkeley.edu/wiki/index ... typical.29
> but I could not fix it.
> Is there any way to solve it?

[fmk]

for the case were you are using the -initial if it fails, change a 0 at the end of the test to a 1. this will print the tolerance at every step. if the tolerance is decreasing you might be able to achieve convergence simply by increasiing the number of steps from 1000 to something larger. how large depends on your model. there are other things to try, i suggest playing with this first as for most well behaved non-ductile structures this works best.

[baikunthasilwal]

When I analyzed Dynamic Analysis of 2-Story Moment Frame available in http://opensees.berkeley.edu/wiki/index ... ment_Frame. I got the following warning: Failed to converge. Then, although I modified the convergence tolerance limit to lower values and increase the number of iteration to 100,000, the dynamic result was also unsuccessful. Please advice me how to get the satisfactory result or better performance in analysis.

I am analyzing three story plane frames for Incremental dynamic analysis. But I am not sure of getting the reliable results. I need someone help to understand the fundamental concepts and practical application of nonlinear dynamic analysis with concentrated hinges.

The following is the warning messaged I got during analysis.

OpenSees > source MRF_2Story_Concentrated.tcl
Modified Ibarra-Medina-Krawinkler Model with Bilinear Hysteretic Response

T1 = 0.8258941430172435 s
T2 = 0.22336556868371923 s
Model Built
Running dynamic analysis...
WARNING: CTestNormDispIncr::test() - failed to converge
after: 10 iterations
NewtnRaphson::solveCurrentStep() -the ConvergenceTest object failed in test()
DirectIntegrationAnalysis::analyze() - the Algorithm failed at time 6.587
OpenSees > analyze failed, returned: -3 error flag
Dynamic analysis did not converge
The current time is: 6.586000
OpenSees >

Please suggest the required modification or any other helpful feedback either for modeling, nonlinear issue, hinge assignment, or any other analyze options to get the results in OpenSees.

Baikuntha Silwal
Graduate student