Twin Beam Diaphragm

[JohnnySm]

Hi

I have a pair of steel beams (twin beams) running parallel to one another. supporting the same 440mm thick brick wall over.

One beam of the pair is however supporting a separate perpendicular steel beam creating a large point load.

I want both beams to work together and therefore rely on a load sharing diaphragm.

Beams are 2No. 203 UC 60's of 4m length with point load at 1m from support.

Standard diaphragm detail would be to use 152 UC sections at say 750mm c/c's. pre welded to one beam web and then bolted with end plate into the second beam once positioned into place.

Does anyone have any design theory on how to design the diaphragm to ensure the load is shared and to justify my detail.

kind regards

 

[struct_eeyore]

limit diaphragm deflection to absolute minimum - consider both effects of bending and shear. you could approximate the stress field with something like strut and tie model to check limit states. maybe a good problem for fem.

 

[KootK]

I'd think that you'd have to:

1) Put a 152 UC diaphragm right behind the supported beam and design that as a flexural continuation of the supported beam.

2) Have both the diaphragm member and the supported beam connect through common bolts at matched end plates such that you're creating a moment connection between the supported beam and the diaphragm.

I see #2 begin very difficult to make work with normal construction tolerances for this kind of thing. You might have other options if you could install something on top of the twin beams or below them but I don't imagine that either of those options is possible.

No matter what you do, it's hard to imagine getting good load share with the load delivered at the beam 1/4 point.

I think that you may well be stuck having to deal with the load in a single beam of the pair.

 

[BridgeSmith]

The only way I can envision getting substantial and predictable load sharing between the twin beams is if the perpendicular beam is above (or below) the twins and can be extended so that it bears on the midpoint of a diaphragm spanning between the twins.

 

[MIStructE_IRE]

I don’t believe you will achieve true load sharing in reality if one of the beams is taking the line load directly.

Editing this as I’ve just re-read that its a POINT load. In which case I think you have no chance of equally sharing this between beams with a diaphragm.

Is there any way the perpendicular beam can come in under the twin beams and hang from one of the diaphragm pieces? Then I could probably get on board. Attached crappy sketch!

 

[MIStructE_IRE]

Attached crappy sketch this time... Bloody phone!

 

[BridgeSmith]

I concur with the proposal in the "crappy sketch" (looks fine to me). I think over or under works the same. If that's not possible, you'd have more time and effort into FEA to try to figure out the load sharing than what it's worth. The cost difference to have beams of different sizes, or bump them both to the size required for the one to carry the perpendicular beam, can't be that large.

 

[JohnnySm]

Thank you all for your advice.
Am i being naive to think that the connection of the diapragm to the beam webs is simply shear only?

I think technically if i want one beam to adopt half the load from the other the diaphrgam must act like a cantiliver with moment connections at each end which would be near impossible to build . but given that the beams are not spaced that far apart a cantilver moment would be very small? welded end plate one end and pre welded bolts the other, for ease of construction.

My theory is that the whole system would't act much differently than if the point load was applied to the centre of the diaphragm itself.then the load would be shared between the twin beams equally no?

FEA model sounds like over kill here guys. I was hoping for a simple answer.

 

[BridgeSmith]

Whether the connection carries significant moment or very little depends on how it's detailed, so yes, if your perpendicular beam frames into one of the twin diaphragms with a connection that doesn't carry moment, then it can be a simple shear load to that diaphragm. However, even if you rigidly you connect the twin diaphragms, so that they become essentially one wide beam, you still have a load applied eccentrically to the center of that beam, and therefore a torsional moment in that built-up beam. The stresses in the twins will never be equal. The only way for the twins to share the load equally is for the load to be introduced along the centerline between the twin beams.

 

[KootK]

Am i being naive to think that the connection of the diapragm to the beam webs is simply shear only?

Yes, quite.

I think technically if i want one beam to adopt half the load from the other the diaphrgam must act like a cantiliver with moment connections at each end which would be near impossible to build .

Yeah, I'd say that's spot on.

but given that the beams are not spaced that far apart a cantilver moment would be very small?

It is small but, as you've clearly anticipated, you're heavily constrained by both space and constructability here. I've no doubt that, in practical terms, the right combination of member and connection stiffnesses could produce a near 50/50 split on the loads. But achieving those things will be no mean feat.

FEA model sounds like over kill here guys. I was hoping for a simple answer.

Agree completely. Keep it simple or don't do it.



HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[MIStructE_IRE]

Turn a ladder on its side and stand only on one of the side rails. See how much load goes into the other - practically none I would imagine.

I think the guys above are suggesting FE because it is unrealistic to expect this load share using static hand analysis.

I wouldnt bother with FE personally because I don’t believe in this as a load path.