Page 1 of 1

Non-convergence during Analysis

Posted: Tue Jul 08, 2025 8:41 pm
by Tharuki
Dear STKO Team,
I am facing trouble in running my high rise building model due to non-convergence after some time steps. Can you help me with the identification of the problem.
This is the warning I get.
Increment: 1 | Iterations: 7 | Norm: 1.314e-04 | Progress: 0.017 %
Increment: 2 | Iterations: 5 | Norm: 6.839e-06 | Progress: 0.033 %
Increment: 3 | Iterations: 5 | Norm: 8.043e-10 | Progress: 0.050 %
WARNING: CTestEnergyIncr::test() - failed to converge
after: 400 iterations
current EnergyIncr: 66.0787 (max: 0.01) Norm deltaX: 36.231, Norm deltaR: 2.29797e+07
AcceleratedNewton::solveCurrentStep() -The ConvergenceTest object failed in test()
DirectIntegrationAnalysis::analyze() - the Algorithm failed at time 0.04
OpenSees > analyze failed, returned: -3 error flag
Reducing increment factor due to non convergence. Factor = 0.5
WARNING: CTestEnergyIncr::test() - failed to converge
after: 400 iterations
current EnergyIncr: 1076.49 (max: 0.01) Norm deltaX: 18.4436, Norm deltaR: 5.48282e+07
AcceleratedNewton::solveCurrentStep() -The ConvergenceTest object failed in test()
DirectIntegrationAnalysis::analyze() - the Algorithm failed at time 0.035
OpenSees > analyze failed, returned: -3 error flag
Reducing increment factor due to non convergence. Factor = 0.25

I have attached my model here.
Thank you in advance!

Re: Non-convergence during Analysis

Posted: Tue Jul 22, 2025 9:10 am
by STKO Team
Looking at the mode shapes, you can see that there is a core column that is not connected:
mode_shapes.png
mode_shapes.png (111.61 KiB) Viewed 1958 times
Debugging your model I noticed that:
  • The master nodes of your diaphragm are not the vertices of the first compound, but you incorrectly selected the nodes of that core wall
  • The you applied the forces to the vertices of the first geometries, but since they are not part of the diaphragm, they are singular in X-Y- and Rz DOFs.
Now I fixed it by changing the master-slave selection of the interactions, and now it seems to run:
CorrectWithcorecolumns.zip
(3.02 MiB) Downloaded 107 times
However I think there is still some mistakes. Looking at the first steps, from the displacements of the dynamic analysis you can see that also the other core walls have a strange displacement, and I noticed that you did not connected them to the beams of the frame (they are only attached with the diaphragm in the X-Y plane, but not in the Z DOFs), I don't know if this is meant to be like this or not. Double check your model
walls_disconnected.png
walls_disconnected.png (257.83 KiB) Viewed 1958 times