Big unrealistic spike using infill elements

04-Unexpected behaviour/errors
Post Reply
Giò
Posts: 1
Joined: 21 May 2022, 19:05

Big unrealistic spike using infill elements

Post by Giò »

I added the infill elements to a framed structure model; I point out that this bare structure model works flawlessly. By performing a dynamic analysis on the model, and using a white noise as input, with very low values, after many seconds of simulation without problems (more than 200 seconds), a disproportionate acceleration peak occurs suddenly and for no apparent reason, with unrealistic values! What could it be attributable to? Thank you
User avatar
z.gronti
Posts: 824
Joined: 16 Oct 2013, 08:14

Re: Big unrealistic spike using infill elements

Post by z.gronti »

Dear Giò,

Please use more strict convergence criteria, available in the project settings module.
Zoi Gronti
Seismosoft Srl.
Giò2
Posts: 3
Joined: 28 May 2022, 08:52

Re: Big unrealistic spike using infill elements

Post by Giò2 »

Dear Zoi,
I had to create a new account because the password recovery procedure doesn't work. Thank you for your suggestion. I made several attempts, trying to use more and more restrictive values for the convergence criteria, until the simulation became impossible to run. Unfortunately, while improvements have been made, the answer is still not reliable
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: Big unrealistic spike using infill elements

Post by seismosoft »

What you could do is the following:
1) Check other sources of divergence, if you carry out a search for 'convergence' or 'divergence' in the forum you will find plenty of resources.
2) check if you have step reduction, due to divergence. Check in the Convergence Details tab the locations/elements of problems
3) check the internal forces and deformations of the infills, to get insight of how the infills behave
Seismosoft Support
Giò2
Posts: 3
Joined: 28 May 2022, 08:52

Re: Big unrealistic spike using infill elements

Post by Giò2 »

Thanks for your support. I had no step reductions when using the default values for the convergence criteria; but I had huge anomalies. Instead, by using more stringent values for the convergence criteria, the step reductions could be continuous, even when the model response was regular. I can't check the internal forces and deformations of the infills, because I only allow the output of some nodes: being very long simulations and in very small steps, the output files are already huge! Nor can I do short tests, because the anomalies appear after at least 100 seconds. However, to avoid problems, I am working with a simpler 2D model, which does not seem to give those problems.
Thank you very much
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: Big unrealistic spike using infill elements

Post by seismosoft »

It could be a problem with very large loading. You could check (i) if you have specified possible strength reduction in the code-based checks and/or the performance criteria (ii) if you have reached the fracture/buckling steel strain, which is be default 20% (i.e. it is already very large). You can set the latter to 1 for instance, but if this is the case, it means that your loading is already extremely large.
Seismosoft Support
Giò2
Posts: 3
Joined: 28 May 2022, 08:52

Re: Big unrealistic spike using infill elements

Post by Giò2 »

Thank you for your reply. The load used is not large at all; indeed, it is very light. It is white noise with a standard deviation of 0.0016 m / (s * s) for both components
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: Big unrealistic spike using infill elements

Post by seismosoft »

You could carry out a search in this form with the 'convergence' or 'divergence' term. You will find plenty of resources. Also there is a section in Appendix A in SeismoStruct's User Manual, called 'Tips to Solve Convergence Problems'
Seismosoft Support.
Post Reply

Return to “04-Unexpected behaviour/errors”