Joist Diagonals Cut Across Building

[T_Bat]

Hey Everyone,

I'm looking at at build-out of a 1950 era building with open web steel joists and concrete roof. Unfortunately, there has been a miscommunication and the contractor has cut out the one of diagonals in each joist across the entire building to allow for ducting. The diagonal is near the center of the joist (which is ideal) and my plan is to makie a vierendeel panel where the diagonal has been cut. I've already taken measurments of the joist and contacted SJI who pointed me towards an old manuf's literature.

The roof is concrete poured in metal lath (basically some kind of pereorated mesh decking). That being said I'm not sure I can count on the top chord of these joists to be braced - I can't verify any actual attachment between the deck/slab and the joists. It would be great for me to be able to count on more than just bridging and panel spacing for brace points (bridging for out of plane and panels for in plane). Does anyone know more about this construction type? Did they tie the roof slab and joists together in any way?

I'm also curious if just the presence of a slab on top can be counted on for any kind of bracing of the top chord. Not out of plane but for flexure. Seems like friction and some bond between the conrete and brae steel has occured. Just trying to think outside the box.

 

[T_Bat]

Pictures of the joist.

 

[T_Bat]

Forgot the close up of the roof...

 

[JAE]

I would not count on friction.
I would consider drilling small holes through each horizontal top chord leg along the length and installing 1/4" dia. screw anchors at a spacing that gets you the required unbraced length.
(Hilti Kwik Con II or Tapcon srews).

Any unbalanced loading on the roof will produce shear in your center "vierendeel" panel so you might think through how the roof is loaded.



Check out Eng-Tips Forum's Policies here:
faq731-376

 

[T_Bat]

Thanks JAE - I've seen this discussion about unbalanced load come up before but I guess I'm not sure what unblanced loads would apply. Typically, if this were new design wouldn you just apply uniform loads per ASCE 7? This is a low slope roof with minimal snow. Is it overly conservative to apply some kind if skip load with the roof live loading? would you normally do this for a joist if you were designing it?

 

[JAE]

T Bat - I guess I'd be careful not to just assume that I can use uniform Roof LL (not snow) and therefore my shear at the midspan (with the cut diagonal) is zero or close to zero.

My risk-averse sense would compel me to use engineering judgement here and visualize that a roof live load COULD possibly be unbalanced in some circumstances where shear would develop across that center area.
How you go about rationalizing some type of unbalanced loading is where your imagination kicks in to get you some level of safety beyond the uniform loading scenario.

Perhaps full LL on 1/2 span and half-live load on the other half?

Or typically, joists back in that time period had the web members designed for 50% of the joist end reaction. (Today SJI uses 25%).
You could perhaps design around that 25% or 50% level, which would be more than using partial unbalanced live load perhaps.


Check out Eng-Tips Forum's Policies here:
faq731-376

 

[T_Bat]

That makes sense. I was thinking full RLL on half and NO RLL on half. That seems overly conservative. Full on half and half on half sounds reasonable to me. Thanks!

 

[JAE]

Personally I'd go with 25% of the joist end reaction for shear in the center.

Check out Eng-Tips Forum's Policies here:
faq731-376

 

[JoshPlumSE]

JAE just beat me to the punch.... by default I'd start with 25% of the end shears for expected shear at the mid-span of the joist. That may end up being a bit conservative. But, it's a good place to start.

 

[T_Bat]

Got it - the unbalanced load definitely increases the demand on my joist. Normally I use plates or rods to reinforce joists but to get the addtional moment capacotu in the chords I'm looking at using a channel or plate with a stiffener on it. I'm leaning towards channel but per RISA section and Risa 3d either work. Still have to verify with some hand calcs but seems like the channel would be easiest....

 

[KootK]

Looking at 25% and partial loading possibilities would be my recommendation also. Philosophically, I think that one could make a pretty good case for partial loading on a single span truss. My suspicion is that those provisions were developed assuming continuous members with constant bending and shear capacities that would be unaffected by partial loading on a simple span. Obviously, with a truss, it's a different animal. Even with little snow, you don't want the roof collapsing when a modest dump does occur and the maintenance guys decides to do you a favor by shoveling it all to one side en route to removing it altogether.

It's hard to know without seeing a roof drainage plan but unbalanced, ponded water might also have the potential to create a shear demand in that panel. Which brings up another issue: when you have shear in that panel, you'll also have vierendeel shear deflection in that panel. That could start to add up.

I think that it's pretty reasonable to assume that the roof deck is bracing the top flange more or less continuously but it would be prudent to verify that somehow.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[JLNJ]

I'm guessing that 25% of the joist reaction will kill the joist via Vierendeel action. I don't think I'd want to go any less either due to the severe penalty for being wrong.

I'd be more inclined to devise a way to snake a diagonal back in, even if it ends up inside the duct.

 

[T_Bat]

I appreciate the responses everyone. I'm not sure about the diagonal inside of the duct - I'm no mechanical guy but that seems like trouble to me. I guess I'm not sure how exactly to design the panel 25% of the end reaction. Point load in the middle of the panel? Point load at the 1st joint adjacent to the panel?

Are joists, as a system, normally designed for 25% revesal on each diagonal? I know that each diagonal individually is designed for this load but is the whole system (the entire joist)?
Secondly I'm also not considering any load sharing between joists from the bridging. There is a line of x-bridging adjacent this panel for the entire legnth of the building.

 

[T_Bat]

I also meant to add that I ran an analysis with full ive load on half then half live load on the other as one case then the mirror image as another case.

 

[racookpe1978]

Snow loads came up, but are not wind/seismic loadings inherently unbalanced? Hurricane threat at all?

 

[T_Bat]

Inland - no hurricanes and low seismic. Negative wind loads are not an issue (no net uplifT) since the roof is concrete. I'm designing for the minimum positive CC wind combined with roof live load.

 

[JLNJ]

Maybe you make the diagonal round, put it in the duct, and tell the mechanical guy he has no choice.

Running duct is not an exact science. I think they can live with a small round pipe inside the duct.

 

[T_Bat]

JLNJ for the win! Turns out they want to just cut a hole in the duct, replace the diagonal, then seal the duct around it. I guess I'm so used to new construction where we typically try to avoid conflicts like this that I didn't think it was something the mechanical guys would go for.

Thanks everyone!