Operable partitions supported by composte bar joists

[txeng91]

I have been tasked with analyzing an existing composite floor system to support operable partitions Footprint is around 80x100 with composte bar joists spanning the 80’ dimensions between wf girders. They are wanting to add a significant amount of partitions, around 20 kips total, that will be supported by 6 joists, with the largest load a single joist is supporting around 10 kips or so. Big picture wise I see this as a relatively low risk modification since the joists are sized for 100 psf live load and the moment increase on the joists is typically in the 5%-10% raise. After many hours of calculations and research I think I can justify the additions with minimal stiffening, however I have a few concerns, see below:

1) The calculations I’ve obtained from the joist manufacturer (I’ve attached an example from one joist) show top chords and webs designed to capacity, with most the bottom chords and convections to the slab having quite a bit of extra capacity. One thing I don’t understand is that the axial load in the bottom chord is only a function of the total load and effective composite deck. My thought is that the axial load in the bottom chord should be calculated for pre composite load using the effective joist depth and the post composite load using the effective composite deck. The bottom chord would already be stressed prior to curing of the concrete and would maintain that stress based on the joist depth in addition to post composite stresses based on the composite depth. I have not been able to find any verbiage on this in SJI manuals.

2) The joists were originally designed in 1996 using ASD. The dead/live load split is about 50/50 so I am considering using LRFD in the analysis as this will give me around a 7.5% bump in capacity. The main reason for this is a lot of th truss webs were originally designed to 98%+ capacity and the partitions push many of them over 100% which would require stiffening. Any thoughts on this?

3) The max live load deflection will probably be in the 3”-4” range. On new construction I’ve done, manufacturers always have crazy deflection requirements like <1/2” andor L/1200. Has anyone had experience in retrofitting existing buildings witb partitions like this with high deflections and think this is doable?

Thanks for any responses

 

[BridgeSmith]

My only thought is that with 3"-4" of deflection, the partitions will have to have a significant gap at the bottom so they don't bind up. I would be asking if they could use a partition system that bears on the floor, with guides at the top, rather than one that hangs from the ceiling.

 

[txeng91]

The partitions actually have around a 3” gap at the base with a rubber stop that pulls out when it’s in place (attached) so I don’t think there’s any risk to the structure under high deflections. I ended up getting 2 1/2 to 3 1/2” total max deflection under full live loading, which includes secondary deflection imposed by the supporting girders. The partitions are only contributing 1/4” to 1/2” of that. And although reductions not allowed at 100 psf, the likelihood of that full load occurring over an 80’ span are zero to none.

There are some existing hanging partitions in this area that are going to be removed. The original structural plans called for 1” max live load deflection for joists supporting partitions. There must have been a discussion on the feasibility of this limit because the joist calculations show them checking the deflections using a 20 psf live load with a 3/4” limit. The manufacturer they are currently engaging of course has insane requirements (L/2160 w/ max 1/4” deflection). The bottom line is if they want to install them there’s going to have to be a compromise on what the acceptable deflection is.

I’ve never seen any bottom supported moving partitions, but I would still think they would have similar deflection requirements. Also there’s no way I’m scrapping all my work to this point to have to redo it on the floor below. The Architect already didn’t want to pay my fee to begin with, no way they’ll increase it for the redesign.

I’ve moved past item 3. I’ve let them know the deflections and if they want to sell they’re panels they’ll have to make it work. Would still love to hear if anyone has input on 1 & 2

 

[Hokie93]

Regarding item #1, the approach taken in the manufacturer's calculation agrees with the approach used in the SJI Standard Specifications for Composite Steel Joists, CJ-Series (SJI-CJ-1.0), dated May 10, 2006. The CJ-Series specification bases the design flexural strength (ΦM_n) on the lowest value of bottom chord tensile yielding, bottom chord tensile rupture, concrete crushing, and shear connector strength. For the two bottom chord-related limit states, both utilize the effective depth of the composite joist, which is defined as the vertical distance from the centroid of the steel joist bottom chord to the centroid of resistance of the concrete in compression. Put another way, an ultimate strength approach is used so superposition of pre- and post-composite stresses is not necessary.

Regarding item #2, using the LRFD methodology for strength purposes is appropriate. The SJI CJ-Series specification uses LRFD exclusively and that makes sense when determining the strength of a composite steel and concrete member. The existing composite joists do not know or care how they were originally designed.

What is the use of the floor supported by the composite joists? Assembly-type use? Classrooms? Offices? I am just curious.

 

[txeng91]

Thanks Hokie. I get what your saying, just seems counterintuitive to me, I’ll have to do some research and see if I can wrap my head around it.

Floor above is currently office space but will be assembly in the future. Also just realized the new system they are putting in place doesn’t appear to allow the panels to rotate about the track so they can’t actually stack them together outside of the stack room. Which if that’s the case it will make my analaysis so much easier. I was running about 40 different load cases and this will bring it down to maybe 5.