St. Venant vs Warping torsion for I-beams 1

[CivilEngineerUK]

I am designing an I-beam for torsion.

The FE-program gives the St. Venant torsion. Is it allowed to design the I-beam using this St. Venant torsion only (and neglecting the warping torsion)? Is it a conservative assumption?

Eurocode states that I-beams shall be designed to warping torsion. But on the other hand the plasticity theory states that every equilibrium system that does not violate the yield strength is a safe approximation of the carrying capacity of the structure.

 

[JoshPlumSE]

Well, ignoring warping is not a conservative assumption as far as stress goes. It's conservative to ignore for deflection.

In some cases (support condition and loading), you could argue that warping isn't going to be very significant in the final design. But, it is best to do this from a position of knowledge.... after first estimating warping stresses.

 

[KootK]

1) If you ignore warping flexibility restraint you will always underestimate overestimate twisting deformation. In the case of an I section, your underestimate overestimate is likely to be a gross underestimate overestimate as warping flexibility restraint in such a section is usually a major contributor to rotational flexibility stiffness.

2) Section warping will introduce axial stresses in the flanges which will interact with the axial stresses imposed by the flexural and axial loads applied to the member causing first yields to occur sooner and asymmetrically.

3) If your software doesn't model warping flexibility stiffness, then it will be overestimating underestimating the torsional stiffness of the members and the amount of torsional force drawn to them. You'll want to keep an eye on that.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[Teguci]

If you are sizing a beam for torsion based on pure torsion only (St Venant) then you will be selecting a beam that is significantly larger than required. This is probably conservative, but the numbers should be checked. A quick check for warping stress can be done by considering just one flange under a lateral load* equal to torsion / (centroid distance between flanges) *point load or distributed depending on the torsional load distribution. This lateral bending/normal stress will act in conjunction with the beams other bending stresses. Check that these normal stresses do not exceed your allowable beam stress (does not violate the yield strength).

This approach is conservative for calculating the normal stresses because:
- you are ignoring the torsion taken by pure torsion action
- you are ignoring any contribution from the web

 

[JoshPlumSE]

KootK -

Just a quick correction on item #1. The warping effect adds stiffness to the section. Therefore, when you ignore the warping effect your twist is much, much higher for I beams.

That in turn makes item #3 incorrect as well. If your software ignores warping in the stiffness calculation, then the analysis will underestimate the amount of torsional force drawn to the member.

 

[canwesteng]

Fundamentally it should be okay for ultimate limit states - the flanges will yield under the added axial load and load should be taken by pure torsion. In saying this, you would need to check the section under service load to make sure you aren't yielding the flanges in that case. In which case you are already checking warping so you haven't saved any time.

 

[KootK]

Just a quick correction...

A pretty fundamental correction I'd say. Having given it some careful consideration, I agree on all counts. Thanks for taking the time to set me straight.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[CivilEngineerUK]

Thanks for all the answers!

I have another question:
I have an I-beam welded to a RHS-columns (no stiffener plates in the joint). The width of the beam flanges and width of the RHS-column is the same. Is this joint restraint for warping?

 

[KootK]

Nope. The column section can warp and the column itself rotate. If you want warping restraint at the ends, welding insome side plates is probably your best bet. Otherwise, your only warping restraint is that brought about near midspan as a result of symmetry.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[KootK]

Actually, to the extent that you can convincingly connect the beam flanges to the RHS walls parallel to be beam, you could use the side walls of the RHS similar to how side plates on the flanges are used. Do you want warping restraint here or are you trying to avoid it?

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[CivilEngineerUK]

I am trying to avoid it to be restraint.