Plastic Torsion Capacity, CSA S16

[wadavis]

I'm working on a design following the letter of the code with the authority having jurisdiction (National Building Code of Canada, CAN/CSA S16), but I'm not finding references for using plastic torsional resistance. In this case an HSS square tubing is primarily loaded in torsion and I want to calculate the torsion capacity with a fully yielded cross-section, much like the plastic bending resistance of a class 1 beam.

To cut to the chase:
1. Are there any references to torsional resistance in S16 other than 14.10?
2. More importantly, are there any allowances to use the plastic capacity instead of the elastic capacity for torsion in Canadian or other design codes?

2.1. Any material explaining why it is a bad idea to calculate the torsion resistance of an HSS by the fully yielded cross section?​

Thanks for the help,
Wadavis
E.I.T.

 

[KootK]

I don't think that you'll find anything specifically addressing that in S16. As for things to watch out for, I'd give some thought to interaction with other stresses unless they are truly insignificant,

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[TLHS]

S16 pretty much expects you to work out torsion on your own. There's some guidance in the companion text put out by CISC "Limit States Design in Structural Steel" (Kulak and Grondin) but it's pretty much about W Sections as far as I recall.

For HSS, though, you should just be inducing a shear stress. So the appropriate thing to do would be check using the shear provisions of the code once you've determined that shear stress.

The common reference for torsion wouldn't be AISC Design Guide 9.

 

[gwynn]

As far as I remember (at the end of a long weekend), the only place not noted above where torsion is addressed for steel sections in the Canadian codes is in CSA S6 (the Canadian Highway Bridge design code). CSA S6 does explicitly treat torsion in HSS sections. Provided you used the strength reduction factor of 0.9 used in CSA S16 instead of the 0.95 used in CSA S6, I see no reason you could not apply the same methodology used there.

 

[BAretired]

I am by no means familiar with ultimate torsion but I do remember that for torsional stress in the elastic range, the Soap Film Analogy is used whereas for torsional stress at yield, the Sand Heap Analogy is used. The soap film is assumed to be a paraboloid surface whereas the sand heap is presumed to be conical. Googling sand heap analogy produces several articles which may be of interest.

BA

 

[KootK]

Two additional points:

1) I wouldn't hesitate to use AISC torsion provisions if they suit your application better. I've done so numerous occasions.

2) With a closed section, most engineers ignore warping stresses. However, warping stresses require some shear capacity for equilibrium. If you expend all of the section shear capacity on St. Venant torsion, I'm not sure where the reserve needed for warping resistance will come from.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[BAretired]

KootK,
I don't believe that warping is a consideration for HSS tubing. The warping constant, Cw is 0.0 for hollow structural sections.

BA

 

[wadavis]

Thanks for the input, a follow up for anyone interested:
Now that I've brushed up on my theory, the plastic capacity and elastic capacity will be the same for thin walled tubes with constant wall thickness (q=T/(2*A)). So the HSS will be designed accordingly with S16 reduction factors and with shear forces included according to 14.10.1.

I'm very interested in the shear flow interaction of shear plus torsion when past elastic, but that will have to be a project for another day.

 

[KootK]

@wadavis: thanks for the update. I think that it's uber-classy when the OP fills the group in on the outcome of an interesting technical query. I'm always amazed at how tricky things can get when we leave cookbook code design and fall back on mechanics of materials fundamentals.

@BA: I agree that warping torsion is generally not an issue with closed sections. However, warping torsion resistance requires the ability to transmit longitudinal shear stress. If 100% of the plastic shear/torsion capacity of an HSS is used for St. Venant torsion, it seems to me that you're back to having four discrete plates rather than a closed box section. Remember, Cw=0 because the section is able to transmit longitudinal shear to the individual plates comprising the section.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.