ANSYS Beam 189 Section Properties Does Not Match AISC Section Properties

[strainstress]

Hi All:

When using ANSYS Beam 189 element the section properties obtained for a square tube (6x6x0.5) is different from that obtained using AISC Section Properties table. Obvisouly, this is due to the corner radius, which is not accounted for in ANSYS Section.

The results are therefore unconservative. Is it still acceptable to use BEAM 189 with section properties defined using SECTION or should I used BEAM4 element and use the AISC section properties ?

Thanks,

StrainStress

 

[rb1957]

? well what do you think ? clearly you think it's unacceptable (else you wouldn't've asked the question) and think using the BEAM4 element is the better way to go. i suspect someone in your office is telling you to use the BEAM189 element and you don't think it's right ?

you say "Obvisouly, this is due to the corner radius"; so it sounds like you've looked into the difference and figured out how the difference came about. if it is due to "corner radius" issues, is it that significant a difference ?

why not use both elements to see the difference in the models ?

Quando Omni Flunkus Moritati

 

[Drej]

Agree with rb1957. It's up to you to decide whether this is unacceptable or not using (obviously) the acceptance criteria you've generated for your design. In general, the differences in properties are small, and any difference should be bounded by other conservatisms you should have built into your robust design.


------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[Kherszal9]

I actually would say that you should use the BEAM189. The AISC codes are general, and depending on the material you are using, may actually be wrong. Beam4 is based off of Euler-Bernouli only, which does not account for Shear. While the AISC codes often are useful, from my understanding, sometimes they are outdated and simply not updated to newer methods or left more generalized because many people do not want to learn to use more advanced theories, so they are shot down.

Beam189 is a Timoshenko based beam element, which does incorporate shear stress and strain. Especially if you have a material that is even slightly anisotropic, such as many beams will be, this will at a minimum show a 3~5% difference for smaller deflections, which will increase and show a more obvious difference at large loads.

If you want to get to this level of considering end shapes, make sure to do material properties based off of directionality - depending on the material and manufacturing method you're looking at, a difference in strength there may show itself. For instance, if an I-beam was to be turned and loaded upon it's side, it would likely illustrate differences in response due to the manufacturing and crystal orientation internally, and not just its geometry.

As with all codes, FE or specified by groups like AISC, you can't take the values as exact in either case. One needs to investigate them and understand the assumptions being used.

That being said, if you are still that concerned with the geometry issue, then I recommend using one of the solid elements and creating the volume you need manually. Beam4 is a legacy element only for comparison to simple linear analytical results. It is not appropriate to use for actual design outside of demonstrating the differences in results.

I hope this helps,

Kherszal