Thoughts on FEA models for "non-typical" steel connections

[rn14]

I'm curious on opinions about using FE analysis for unusual steel connections. More specifically, using plate/shell/solid elements to model a connection that doesn't fit into anything in AISC.
Do you ever do it?
Do you just look at stresses?
Do you accompany with some hand calcs to check plate buckling in compression and/or shear?
Is this technically "allowed".

I realize that people have pretty strong feelings about FEA in general, more specifically, those uncomfortable or unfamiliar with it tend to think it dangerous. For the sake of the conversation lets assume the person performing the modeling is familiar with the software being used and can correctly generate a model.

 

[JGard1985]

RN14, to focus the thread, do you have specific connection/example/sketch?

Jeff
Pipe Stress Analysis Engineer

http://www.xceed-eng.com

 

[rn14]

Jeff,
I don't have any sketches at the moment. Although a general question, the particular case that raised it this time is that we have been asked to design some "winch anchors" for snow grooming equipment. The want to try and use round tube that they already have. There would be 3 forces coming together at the round tube, the load from the grooming equipment and the reaction of 2 vertical braces.

 

[WARose]

I've never felt the need to model most steel connections as they (for the most part) are well covered by hand calculations/equations in the code. About the only time I have done any FEA modeling of steel connections is some studies I have done with regards to the behavior of typical connections /stiffeners under torsional loads. From that, I've gotten some rules of thumb I have used ever since.

 

[JGard1985]

Neat project! (I googled it winch anchors for snow grooming equipment) Amazing that such steep ski slopes are groomed this way!

We use FEA (ANSYS & LSDYNA) to evaluate complex structures and loading (impact, drops, fluid-structure interaction, etc.) for which there are limited code equations available. We recently used it to settle an argument for behavior of a bolted joint subject to external load.

In your case, I don't see why you couldn't use FEA. Not sure of the design requirements of winch anchors but FEA allows you to consider potential shock loading of the winch line, and energy methods if needed.

When we use FEA we always accompany it with hand calculations verifying results. Need to prevent garbage-in, garbage out. Checking stresses is always a needed, but really need to be cognizant of all FEA outputs:

-Deflection: Is structure really deflecting as you expect (common sense gut check), no additional boundary conditions
-Mesh Convergence: Is mesh adequately (accurately) capturing the response of structure
-Buckling: Are members in compression? If so, does the loading meet design code limits for member compresive strength. Bucking is one to watch closely in FEA!

Hope this helps



Jeff
Pipe Stress Analysis Engineer

http://www.xceed-eng.com

 

[WARose]

Jeff,
I don't have any sketches at the moment. Although a general question, the particular case that raised it this time is that we have been asked to design some "winch anchors" for snow grooming equipment. The want to try and use round tube that they already have. There would be 3 forces coming together at the round tube, the load from the grooming equipment and the reaction of 2 vertical braces.

Nothing wrong with using FEA for such a situation.....but if you can approximate it with a simple model/assumptions (on paper), it might be best to do so. (It's hard to say definitively without a picture.)

One pitfall of FEA for connections is all the stress concentrations you will get. You could model a connection (that works) right out of a AISC manual and get all sorts of excessive stresses. It takes a experienced hand to know how to handle those things.

My 2 cents.