Convergence Error in SFI-MVLEM Model

[Nayera Sakr]

Dear STKO Team,

I am writing to ask for you kind support in my problem. I built up my Model (C-shaped RC Wall) tested under cyclic loading, using SFI-MVLEM Element with Concrete CM and steel MPF.

I am facing a problem in the convergence. The analysis runs till a certain % and then fail to converge. So I don't know what could be the problem.

I attached the model. I would appreciate it if you could guide me solve the problem and let the analysis run completely.

Best Regards,
Nayera Sakr

[STKO Team]

You applied forces and you are doing a load-controlled analysis. But you have softening, so you will not be able to capture the descending branch.
Why don't you apply a displacement history instead?

[Nayera Sakr]

Dear STKO Team,

Thank you for your reply. I tried to apply the time series path as a displacement input instead of force input. and changed the analysis integrator to displacement control integrator. But the only limitations in this option is that it specifies the analysis to be in a specific DOF. So if I want to apply a bi-directional or diagonal loading, This couldn't be determined in the analysis.
I am running the analysis in both directions X and Y simultaneously (diagonal). so please advice, how could I consider the diagonal loading in the displacement control integrator.

Best Regards,

Nayera Sakr

[STKO Team]

If you impose a displacement (with the sp) you don't need to do a displacement-control analysis. Just do a load control.

When you have softening you have 2 options (They are equivalent only when the loading point is 1 (top of your specimen), but not in multi-storey buildings for example):

• Impose a force, and control its magnitude with a Displacement-Control integrator
• Impose directly a displacement and do a simple Load-Control integrator

With the second option of course you have the ability to apply 2 different sp load (imposed displacement) with 2 different time series in 2 different load patterns

[Nayera Sakr]

Dear STKO Team,

I tried to work on the displacement input and perform a load control analysis through imposing the displacement in a single point restraint, but unfortunately, the model is not converging as well.
kindly check the attached model.

I already tried it with the load input and it was running the analysis successfully. but after I modified the model to have a displacement input time series, it does not converge.

So please advice.
Best Regards,

Nayera Sakr

[Nayera Sakr]

I attached a model with the element MVLEM 3-D which is better in solving the convergence problems than the SFI-MVLEM 3-D.

[STKO Team]

Just change the constraint handler from Transformation to Penalty.
Now you have SP (imposed displacements) at the same place where you have MP constraints (EqualDOFs). This can give problems to the Transformation method.
In cases where the SP and/or MP constraints overlap, just go with the penalty.
In this case, you can use penalty value = 1.0e18. It's fine for your model

[Nayera Sakr]

Dear STKO team,

Thank you so much for your recommendation, it worked with the cyclic analysis of MVLEM 3-D and SFI-MVLEM 3-D. The only problem i am facing right now is that the pushover analysis of SFI-MVLEM 3-D is still giving a convergence error. So do you have any recommendations regarding this issue.

Really appreciating your support. Best Regards,

Nayera Sakr

[STKO Team]

You are modeling in N-mm:

• Try a higher reference horizontal load (for example 1.0e6)
  Now you have 1 N of reference load, so the load-multiplier (which is an unknown in the Displacement-Control) is very far from the solution. It's better to assign a reference load which is closer to the expected (so that the load-multiplier is close to 1)
• Use a less strict tolerance (1.0e-3 for example). 1.0e-6 millimeters is hard to achieve