Hollow section beams for composite decking

[TallEng]

Hey guys,

I'm looking at the design of a composite roof deck, using a profiled metal deck & lightweight concrete.

There's also a requirement for a parapet which will need supporting off the perimeter beams, so in addition to the out-of-balance loading from the roof deck/secondary beam reactions, I need to make consideration for the torsional resistance of the perimeter beams. My first thought was to use hollow section box beams (i.e. RHS), however the design software I'm using for the composite beam design doesn't allow for a hollow section to be designed and the only reference I can find in BS EN 1994 to hollow sections relates to concrete filled columns.

Is there something I'm missing, or can I use RHS sections with shear studs as the perimeter beams?

Thanks in advance!

 

[jayrod12]

Do you need it to be composite for the design of the HSS?

Alternately, you could use your standard wide flange beam and put angle braces for the bottom chord to an adjacent joist at each parapet post location. This is how we normally do it.

 

[TallEng]

Tomfh - looks like I've got some reading up to do! Can you recommend anything that goes into more detail regarding shear flow?

Jayrod12 - thanks for the suggestion, I'll look into how the detail will work in this instance.

 

[Hokie93]

How is the parapet framed? A more-preferable solution than using the spandrel beam for bracing is to frame the parapet such that it is continuous past the roof diaphragm and braced by the diaphragm, thus not imparting out-of-plane loading to the spandrel beam. This is readily accomplished with cold-formed steel framing, provided the architect is agreeable and not counting the square footage down to the square millimeter, carried to six decimal places. Failing structurally efficient parapet framing, I concur with the recommendation to laterally brace a wide-flange spandrel beam to eliminate the torsion or use a non-composite HSS. I don't think I have ever seen an HSS beam made composite with steel deck and concrete topping but there may not be a code prohibition against it.

 

[KootK]

...loading from the roof deck/secondary beam reactions

If you have secondary beams, you may well be able to just use those to brace the spandrel beam against torsion.

I can't think of any reason why you couldn't design an HSS as composite with a concrete slab. That said:

1) I'm not personally familiar with any software that does this and;

2) I'd rely on composite action only for gravity loads and not the torsion unless special detailing was provided to transfer the torsion into the slab.

3) Keep in mind that a beam resisting torsion often needs to deliver that torsion at it's support which can be a hassle.

4) For the sake design efficiency, you might just be able to trick your WF composite beam software into fudging a design for your HSS. Run the HSS as a wide flange beam with the following properties:

a) similar depth.
b) similar flange area.
c) similar or lesser web thickness (likely lesser).

I'd think that a conservative design approach. I don't really see how the design of a stud on an HSS flange would differ from the design of a stud on a wide flange. I think that the main difference would be that your HSS webs would tend to resist much more flexural tension than your typical WF web. So be it.

 

[Ingenuity]

The shearing forces can no longer flow directly from the studs into the web, they have to detour via the RHS top surface.

Which would also be the case where two rows of studs are required to the top flange in a WF section.

 

[Tomfh]

Which would also be the case where two rows of studs are required to the top flange in a WF section.

Yeah, maybe you’re right. It’s better in principle to have the components of web in the same vertical plane, but I’ve never actually calculated if it’s critical and worth worrying about.

 

[Agent666]

The stud loads are not going into the web, the horizontal stud force is transfered to the flange as an axial load. I've never seen anything before regarding studs located off the web being an issue, they can be right out on the edge of the flange at min edge distance and it'll still work as intended. The fact that codes have a minimum flange thickness (usually stud dia/2.5) is intended to make sure that the thing the stud is welded to is able to resist the loads and resulting stud deformation in fully mobilising the studs within the shear span.

Just design a composite RHS like any other composite beam, there is nothing special involved, go back to your standard and work things out from first principles if required if your software doesn't include the option. Its not overly hard.

 

[Tomfh]

The stud loads are not going into the web, the horizontal stud force is transfered to the flange as an axial load.

The stud loads do go into the web. The studs and the web form a composite web and so the horizontal and vertical shear loads have to get thru. In an I beam the shearing loads can go directly straight in, but if the studs are offset the loads have to take a more convoluted path. It may not be an issue, I've never done the numbers on an offset web. Maybe the RHS is strong enough that it doesn't matter...

 

[Agent666]

Longitudinal load goes into the web is proportional to the change in axial load due to change in moment between two points or axial loads introduced by studs, its no different to a ordinary beam in flexure essentially. Yes load gets into the web, but its not the full stud load at each and every stud that I'm interpreting from your posts.

Axial loads in the flanges change in proportion to the moment, as do the proportion making it to the web. You seem to be suggesting all of the load from a stud goes to the web unless I'm interpreting what you are saying incorrectly. But its only the longitudinal shear flow transferring across the flange/web interface. Which is usually quite small in comparison. I don't see anyone worrying about the shear lag effect or longitudinal shear in flanges due to either of these effects (bending or load introduction by studs or other means), because the shear flow is quite small by comparison.

Think about sizing a weld between a flange and web in a welded section for an analogous situation, weld often doesn't need to be that large even if the flanges are large or have a large force because one way to think about it is that only the change in load in the flange is transferred over a given length.

 

[Tomfh]

But its only the longitudinal shear flow transferring across the flange/web interface. Which is usually quite small in comparison. I don't see anyone worrying about the shear lag effect or longitudinal shear in flanges due to either of these effects

Yes that's what I'm referring to. The shear force into the flange. But reading more in the composite code AS2327.1 it doesn't seem a huge issue. The code explicitly covers RHS.

They seem to deal with the issue I'm fretting over about by saying the flange thickness must be at least 0.4x the stud diameter OR the studs must be directly over the web.

 

[Agent666]

Yeah, thats the stud dia/2.5 limit I noted in another form . The limits are there to ensure studs don't pull out/tearout of thinner flanges as they bend over under deformation when mobilising the shear strength, or precluding any burn through while welding as well.

In the latest AS/NZS2327:2017 theres also another tighter limit if you are dealing with fatigue or tension in the studs of stud dia/1.5. Theres also a limitation in this latest version for cold formed rectangular hollow sections, that the studs need to be manually welded for thicknesses not less than 4mm, I am not sure why but it seems like you'd maybe have to manually weld then at less than 4mm', not above 4mm thickness (could be an error), but it might be something the OP needs to review if relevant in his market.