Question - Torsion Analysis of parallel HSS beam 2

[jsu0512]

I am currently being asked to design a double HSS that is subject to torsion along its length. The HSS at the back is supported by the concrete wall at its end (torsion resisted), and the front HSS, which I am designing, is planned to be attached to the front face of the back HSS with either bolts (at specified intervals) or welds, and both are subject to torsion.

My question is, how do I check the torsion in this situation where there are two HSS sections side by side? Should I consider these two HSS sections as a composite section? If so, what formulas should I use to check the welding (or bolt connections) to ensure these two HSS sections act as a composite in resisting torsion?

Additionally, the front HSS is not supported at its end and will be about 6 inches shorter than the back HSS. In this case, there is a 6-inch span where only the back HSS is present without being stiffened by the front HSS. I believe this span area also needs to be checked as a single HSS, as its torsion strength will be lower than that of the composite section. Am I correct?

 

[SWComposites]

The torsion of the single HSS section at the ends is probably going to control the design.

 

[JoshPlumSE]

If they weren't welded together continuously, then you could just add the J (torsional constant) of the two members and you'd be good.

This is actually a little complicated because you have a two celled enclosed section. I don't actually remember how to do this.

The conservative way is to assume that it's just a single hollow section and ignore the interior walls. Then J would be approximately equal to 4 times the enclosed area times the thickness divided by the perimeter.... I think.

 

[jsu0512]

The interior HSS is an existing member, and I am designing an outer HSS that will be attached to it. This is necessary because we are adding a bracket at mid-span. I need guidance on how to assess the connection between the two HSS members.

 

[KootK]

 

[SWComposites]

As stated above, the simplest method is to ignore the center webs and analyze it as one rectangular box; calculate the shear flow along the weld lines, and check the weld for that shear.

but you are going to get a stress concentration at the end of the new tube where it attaches to the longer tube. you probably should weld the web of the new tube to the web of the longer tube at the ends.

if you want a less conservative analysis, look up the analysis methods for the shear flow in a two cell aircraft wing box under torsion (its been way too long since I did that by hand to quickly outline the method).

 

[SWComposites]

KootK - the loading is via the bolted on bracket in the OP' figure.

 

[KootK]

KootK - the loading is via the bolted on bracket in the OP' figure.

Thanks. It doesn't appreciably change my recommendation though. Whatever gets delivered through the new HSS ought to just get transferred back to the original via plate bending. Given that, I'm not sure that there's much of a point to the new HSS. I'd just weld the bracket to the existing HSS directly.

 

[KootK]

Also, what's the overall span of this thing? I assume that it's not 6 mm. Maybe 6000 mm?

 

[jsu0512]

The existing HSS is only 6 inches wide and is not sufficient for the new bracket, which is why a new HSS is being considered. For simplicity, I have shown only one bracket in this discussion, but the project involves multiple brackets. We are considering adding a new HSS attached to the existing one. KootK, does your opinion still hold true, even if the bracket connects to both the new and existing HSS members?

 

[KootK]

Would also be useful to know what the connection to the existing CIP walls looks like.

 

[jsu0512]

Yea 6m (not mm). My mistake

 

[KootK]

KootK, does your opinion still hold true, even if the bracket connects to both the new and existing HSS members?

Mostly. I'd be inclined to handle it as shown below.

I'd only get into the weeds with the composite HSS business if it were truly the case that the existing HSS was deficient on it's own with respect to strength or stiffness. If the purpose of the HSS is just "reach" on the outer bracket bolt, there are simpler ways.

 

[jsu0512]

KootK

There are additional reasons for considering a continuous new HSS beam here. This beam not only support bracket but also supports the glazing channel at the front face of New HSS (see picture below). This glazing channel will run parallel to this new HSS beam. Do you see any better suggestions to address this requirement other than using a new HSS beam? I would appreciate your input.

 

[KootK]

Do you see any better suggestions to address this requirement other than using a new HSS beam? I would appreciate your input.

I need a bit more information:

1) Will the existing HSS, on it's own, be insufficient with respect to capacity or stiffness once the modification is made? It matters to me whether the new HSS is reinforcement, space filler, or both.

2) What does the connection of the existing HSS to the concrete walls look like? I'm most interested in knowing how many of the four tube walls are actually fastened to the concrete and which wall those are. Torsion cannot really be separated from the end support conditions.

3) Is the torsion in this setup really just a result of the eccentricity of a mullion post connection? I'm getting the sense that perhaps the new HSS is really there for lateral stiffness considering the HSS pair as a wind girt.

 

[KootK]

Speaking to welding vs bolting:

1) If you bolt the HSS together, I don't see how you would get any meaningful composite behavior.

2) If you weld the HSS as you've proposed, I worry about that overhead, flare bevel weld. It would be unpleasant for the guy with molten metal dripping onto his face and would be prone to pretty lousy weld quality.

 

[KootK]

Never mind the dimension. I see that must be about 12". Given that, you could potentially slip critical the bolts for composite action.

 

[jsu0512]

1.We have not yet confirmed if the existing beam can independently resist the torsion. Another engineer, who originally designed this beam, is currently reviewing its capacity. If the existing beam is deemed sufficient, would you then design the new beam such that its bottom flange can locally resist the bolt tension load at the fin bracket?
2.I am not certain about the exact configuration of the connection. My assumption is that all four sides of the HSS are restrained at the wall.
3.Yes, the torsion results from the cantilevered mullion at the bracket locations.

 

[KootK]

Nifty alternatives aside, I don't want to dodge your original questions.

My question is, how do I check the torsion in this situation where there are two HSS sections side by side? Should I consider these two HSS sections as a composite section?

Depends what you need. If you require the composite properties for a successful design and you're willing to connect the pieces accordingly, go composite. If the outer HSS is mostly just space filler, it might be simpler to just rely on the existing HSS for most purposes.

If so, what formulas should I use to check the welding (or bolt connections) to ensure these two HSS sections act as a composite in resisting torsion?

Treated rigorously, you've got three welding problems to solve.

1) Getting the applied torsion into the combined section. You'll want to do this locally using a model similar to what I posted initially.

2) Getting the applied torsion out of the new HSS and back into the existing where the new HSS terminates. This would also use a model similar to the sketch that I posted initially. It might also behoove you to add cap plates to the ends of the new HSS to keep the tube from warping at the ends.

3) Composite action welding. This turns out to be surprisingly easy:

a) Ignore the web.

b) Calculate the sectional St. Venant shear flow using the conventional theory: Link

c) Because transverse and longitudinal shears are complementary, your longitudinal shear flow will be the same as your transverse shear flow.

d) Provide intermittent welding to achieve the shear flow from [c].

Am I correct?

You are correct, the existing HSS needs to handle all of the load in the segment where the new HSS will be terminated.