Need help with identifying the underlying issues that might be causing errors

[abira]

Hi,

I am running a dynamic analysis for a 10 storey 5 bay steel frame, both geometric and material non-linearity is considered. I am running a probabilistic analysis, using a sample size of 500. All the material properties and loads are modelled as random. Out of 500 simulations around 4/5 simulations produce following errors:

1. Negative eigenvalues which produces a domain error

2. CorotCrdTransf2d::compElemtLengthAndOrientWRTLocalSystem: 0 length

Does this error indicate a deeper issue with my model or is there a bug. Are there any tips to determine what is causing the issue?

Thanks

[mhscott]

Yes, it's a model issue

[abira]

Thanks for the reply @mhscott. Is there any documentation available that can help me understand these errors. I did find the following https://opensees.berkeley.edu/OpenSees/ ... ource.html while searching for second error but I am not able to clearly understand what it is indicative of. Is there any other reference material available that can help me understand better.

Thanks

[abira]

Anyone looking for similar errors, some of the reasons are mentioned below:

1. Negative Eigenvalues: One of the reasons that might be causing this error is not invoking wipe Analysis before starting an eigenvalue analysis. Not invoking wipeAnalysis after the application of gravity loads (i.e., static analysis) was one of the reasons I was getting negative eigenvalues. In addition to this, not assigning proper nodal constraints when using a combination of beam elements, spring elements, truss elements, etc. in your model might be causing this error.

2: CorotCrdTransf2d::compElemtLengthAndOrientWRTLocalSystem: 0 length. This error is caused when the length of the beam column element is zero. Unintentionally assigning same coordinates to two different nodes may cause this error. I think the same error would be encountered with any geometric transformation if the length of beam-column element is zero. It is not just specific to corotational transformation.

Thanks
Abira Rather