modeling cross-sectional distorsion of beam

[Newuser2007]

Hi,
I wanted to model beam where cross-sectional distorsion is allowed. For example, top flange of an I-beam (W14x34)deforms.
I have tried Beam 189. But it is timoshenko beam with shear stress constant through the tickness. That means plane section remains plane. I don't want that.

For my application( bridge deck on stringers), there is local deformation of the stringer top flange.

How can i model that in ANSYS using beam element? I know I might try shell to model beams but that will use up resources.
Can anyone help please?
Thanks.
Newuser2007

 

[40818]

Not familiar with Ansys, but what your talking about is a on-linear analysis, as is obvious by "plane sections not remaining plane". You need to model with shell elements, and have at least 3 elements for the flange height. You will need to run geometric, material non-lineararity with follower forces.
It will chew up your run time as that seems critical to you.

Why is it so important to have the non-linearity in your model??

 

[EdR]

newuser:

I doubt you will find any beam element that will do what you are asking since that cannot be modeled by having nodes only at the beam centroid....that's not to say there may not be some element in some code that does this but it would quite rare and specialized.....Thus you will have to model the beam using shell elements....

As to why 40818 thinks geometric, material, and follower force nonlinearities "must" also be modeled I don't know...I would only consider them if something else in the model indicated such a requirement....and for the size section you are dealing with I doubt any of these items are important....

Ed.R.

 

[40818]

EdR;
For completion really, if your going to do non-linear you may as well make it good job and not a half measure!!

 

[Newuser2007]

Thanks for your replies.
I know I can do solid or shell model of the beams to get the local effects like the flange distortions. But that would eat up resources. Thats why i wanted to try beam elements first.

ANSYS beam 188/189 elements are only first order shear deformable and I will not get cross-section distortion effect.

From your responses it seems like I can't find an appropriate beam element.

The nonliniearities is not completely releveant here. I can do with or without it. The major point whther I can have an element to show that local deformation or not.

I would appreciate if anyone has any beam element in mind.

 

[cbrn]

Hi,
no, in ANSYS you don't have any. As you have already seen, the "best" beam formulation is 1st-order Timoshenko / parabolic formulation.
I'm with EdR in saying that the pure-beam formulation is stuck on the hypothesis that the sections remain plane. In Ansys, even if you define these sections with custom shapes and it seems that you "mesh" them, it's only a way to compute mass/inertia properties and the shear-center.
I don't think that a properly set-up shell model would crush your computer to inacceptable runtimes.
By the way, saying that the sections remain plane doesn't mean that they can't rotate one with respect to the other.

Regards

 

[Newuser2007]

Hi cnrn,
thanks for your reply. I will have to try to shell elements. Only thing is that to match up with the other part sitting on top of the I-beam, I will have to do a fine mech on the beam. This eats up memory. I am trying constraint equations/coupling to resolve this too. But still all contributes to complexity!