Warning message for ForceBeam-Column element

[AbRiahi]

I have modeled a steel braced frame subjected to earthquake ground motion. In order to simulate buckling response, I used two fiber-based force BeamColumn elements for each bracing member with an imperfection value. But in some large scale factors I have got the following warning messages related to the bracing members (e.g. element 38):

Warning-ForceBeamColumn2d :: update-fail to get compatible element forces & deformation for element : 38 (dw<2.88303e-8)
Domain::update-domain failed in update
Newmark::update()-failed to update the domain
Warning-Modified Newton::Solve CurrentStep()-the integrator failed in update ()

I want to know:
Is these warning messages are necessarily related to unstability of the structure or it may be due to some numerical problems which can be avoided by some tricks? How can be sure? In the latter case how can it be tackled?

[vesna]

the message tells you that your element 38 can not reach compatibility between the element forces and deformations. In a way this means a failure of your element. However you can try few things before you announce the failure of the element:

1) check your section (make sure you have enough fibers)
2) try -iter option on the element (tolerance of 10e-12 should be good enough)
3) make sure NIPs is not greater than 5
4) replace one force-based element with several displacement-based elements

Let me know if any of this worked for you.

[wuhaoshrek]

vesna wrote:
> the message tells you that your element 38 can not reach compatibility
> between the element forces and deformations. In a way this means a failure
> of your element. However you can try few things before you announce the
> failure of the element:
>
> 1) check your section (make sure you have enough fibers)
> 2) try -iter option on the element (tolerance of 10e-12 should be good
> enough)
> 3) make sure NIPs is not greater than 5
> 4) replace one force-based element with several displacement-based elements
>
> Let me know if any of this worked for you.

Hi, vesna,

Why you suggest commander 2), i.e., NIPs is not greater than 5?

And what do you mean by failure of the element?

[vesna]

more than 5 IPs usually cause stress concentrations at the element ends. The best way to choose the number of IPs for nonlinear elements is by trying to match integration weight at the element end with the plastic hinge length.

What I meant by the failure of the element is that it could not reach convergence at the internal level. It does not necessarily indicate the the collapse of the element. You need to look at the element response to see if its behavior indicates a collapse.