steel beam bowing

[Dereekoo]

I think you're being a bit pessimistic. Fig 3.15 of SCI P360 would suggest a series of joists can provide lateral restraint. And 2.5% of compression forces should be able to be held by joist hangers. (if the joists are parallel then red alert)

Bigger issue for me here is use of an open section for torsion resistance, wall returns instead of posts (not very impressed with the padstone/wall calc) and lack of dead load deflection. Oh and its not galvanised on.

Although they are indicative of floor supports with joists either side of the beam as opposed to a flat roof connection

 

[George_engineer]

Although they are indicative of floor supports with joists either side of the beam as opposed to a flat roof connection

Yeah - so it depends on the capacity to resist tension as well as compression. With fully fixed joist hangers, tied back to a ledger with a plywood structural deck/diaphragm overlaying the roof I think I'd be reasonably confident of full restraint.

 

[RonnyRaygun]

I think you're being a bit pessimistic. Fig 3.15 of SCI P360 would suggest a series of joists can provide lateral restraint. And 2.5% of compression forces should be able to be held by joist hangers. (if the joists are parallel then red alert)

Yes, but that assumes joists built in and from both sides of the beam. In this case the joists are only from one side and likely to be above the beam, not at the same level as it. The same paper says that reliance on friction isn't recommended, so it would be imperative to provide sufficient strapping or other method of restraint.

Always a risk to have to rely on the more subtle details on jobs such as this as some builders don't always follow the drawings to the letter...

 

[George_engineer]

Yes, but that assumes joists built in and from both sides of the beam. In this case the joists are only from one side and likely to be above the beam, not at the same level as it. The same paper says that reliance on friction isn't recommended, so it would be imperative to provide sufficient strapping or other method of restraint.

Always a risk to have to rely on the more subtle details on jobs such as this as some builders don't always follow the drawings to the letter...

Yeah - I only saw the image on my phone - now on my desktop it is clearly at door head level so no chance of lateral restraint.

I would not rely on friction - only when bolted + joist hangers. But I would be inclined to rely on one-sided so long as there was a diaphragm. Appreciate the freewheeling nature of some builders, but steel is expensive and it is a lot more economical to get the restraint in.

 

[tony1851]

We can argue the point on lateral restraint until the cows come home - it's one of those moot points, often where experience comes in. But where builders often 'do their own thing', its perhaps better to err on the safe side.
But this still doesn't address the issue of deflection. I've not done any figures - I think Ronny did some and IIRC they were quite excessive at 44mm - its just too much. The 8mm plate will bend downwards too far on that span, I suspect OP will have to consider a central post of some sort.

I wouldn't argue with the returns at 900mm - accepted they go outside the 'deemed to satisfy' rules in AD A, but still acceptable IMO - could easily be proved.

 

[George_engineer]

I'm not really arguing - I agree it hasn't got restraint. I was making the point that in principle, roof and floors can provide lateral restraint (to assure myself on some of my deigns...), but in this instance the beam is too low to benefit. If engineers can't try and make the best use of materials, then what are we for?

Open section here is a big mistake and there's no torsion or adequate deflection calculation in their docs.

Masonry returns could be ok if my friend the roof diaphragm helps. I'd be intrigued to see the calculation that would prove they are OK - genuine curiosity from me on how to arrange that calculation.

 

[tony1851]

Masonry returns could be ok if my friend the roof diaphragm helps. I'd be intrigued to see the calculation that would prove they are OK - genuine curiosity from me on how to arrange that calculation.

There are several ways: you can use the wall-panel permutations as shown in BS whatever-it-is, assuming no return, to see if the panel itself will be OK under wind load.
If that fails, treat the return as a beam under triangular load from the side panel - assuming a cracking pattern similar to those for concrete slabs. There will be considerable compression on the return due to the load from the beam - that will also help stability.