existing steel bad jois

huangyhg

existing steel bad joist
i have a situating where mechanical engineers want to add an ac unit on existing roof. the roof construction is typical steel bar joist, metal deck, light weight concrete and water resisting membrane. of course there is a suspended ceiling system. nothing facny.
the building was designed in the late sixties. it appears that the structure was designed by an architect. the joists are spaced 6 foot on center with metal deck of unspecified height and three inches of concrete of unspecified density or type. this is of the plans sections and details.
it appears that the concrete is insulative type (i assume a weight of 2 psf per inch of thickness). the deck, i assume is one inch high based on 6 foot span. the steel joists are 16h7. using 20psf live load and computing 14-psf dead load i am able to determine that the joist load carrying capacity is almost maxed for its span. this is based on sji鈥檚 load tables for 16h7.
i am able to locate the ac unit so that two sides will be bearing directly on steel beams that are well below their allowable stress and deflection. no problem.
i followed procedure that was published by vulcarft in the 80 about designing joists for concentrated loads. here is what i found:
1.    the total moment capacity of 16h7 is above the new imposed uniform and concentrated load. from simple strength of material (shear and moment diagram)
2.    the end reaction capacity of the joist seat is slightly above the actual reactions from the imposed uniform and concentrated loads.
3.    when determining the uniform load on the joist using vulcraft鈥檚 procedure, i came up with a 鈥渃omputed equivalent uniform load鈥?that is about 10 pounds per foot larger that the joist allowable uniform load capacity.
i know that all the loads that i computed are based on full ll and dl combination acting simultaneously.  how does everyone feel about the excess 10 pounds per foot?
i plan on calling vulcraft and chatting with their engineers. meanwhile, if you were in my shoes, what would you do?
lutfi...  maybe i'm missing something but there appears to be a disconnect in item 3.  but given that, 10 lb/ft is about the weight of that joist.  also, if you notice the 16h table progression, each foot of span length accounts for about 22 lb/ft of capacity, so your 10 lb/ft is probably not significant in the overall scheme.  you might want to check the field dimensions carefully...you might not have a problem at all!.  
if you do decide you have an issue, i would consider strengthening the joists in place, since the joist seats appear to have capacity.  i'm assuming the location is in florida, and one problem i've noticed in florida failures is the sudden load reversal causing internal joist weld failures (usually occurs at the first diagonal as it tries to punch past the bottom chord), then progressive failure of the joists (one that i checked about 10 years ago looked like it had been unzipped).  when you have more dl+ll than was anticipated to begin with, you obviously have a tendency toward more deflection.  when high winds (even the strong afternoon thunderstorm variety)hit, the uplift will pick it up and drop it suddenly.  not a lot of upward motion, but sometimes enough to pop a few welds.
since this is 60's vintage, i doubt the joists were checked for uplift.  you might want to run a check on the first diagonal for uplift and see if more bottom chord lateral bracing will be required (on all, not the ones you're looking at for the ac unit).
one other thing to consider is the moisture condition of the lightweight insulating concrete.  in its equilibrium moisture content state (preferred), the moisture content will be around 10 percent by weight.  this will yield a unit weight of about 25-28 pcf (about what you are estimating).  actual in-place moisture contents can range from about 20-25 percent to over a 100 percent.  if you have a typical one, consider 30-50 percent moisture content (ask if they've ever had any roof leaks.  the answer will be "yes", of course).  with this moisture content, your unit weight will be about 33 to 38 pcf or more on the order of 3 or 4 lb/sf/inch.
lufti-
i would first check that the location of maximum moment computed with the concentrated load is within the varying "envelope" of moment resistance of the joists, not just the maximum.  the maximum allowable moment provided by the joist manufacturer at midspan is not necessarily true every where else in the joist (which is why "kcs" type joists were developed.)  ditto for the shear strength of the joist.  its probably not uniform across the entire   
ron,
did you notice i called it bad not bar joist!! that is how i feel about touching these old buildings. boy you never get the proper fee to doing them either.
let me elaborate on items number 3. i calculated the maximum moment due to uniform and concentrated load. then converted it to equivalent uniform load that would create the moment. i came up with 205 plf that is well below the capacity of the joist of 269 plf.
i calculated the new joist reaction due to the concentrated load. the new reaction is 4458 pounds. joist seat reaction capacity is 4900 pounds. which is good. however, i converted the maximum joist end reaction to equivalent uniform load that would generate this reaction and i cam up with 279 plf. the 279 plf exceeds the joist allowable uniform load capacity by 10 plf.
i agree with your points. i am in florida and uplift is a concern. i am not sure if the client wants to go and upgrade his roof structure to meet the uplift load reversal.
thanks for you input.
one question that occurs to me. you mentioned in your original post that:
"i am able to locate the ac unit so that two sides will be bearing directly on steel beams that are well below their allowable stress and deflection"
why not design load transfer steel to carry all the weight on the steel beams?
steve1,
1. the unit is rectangular.
2. there are two steel beams that are framing to a column and they are at 90 degrees of each other.
3. one beam is parallel with the joists and the other is supporting one end of the joists.
4. so envision that one of the short and one of the long sides are supported by the steel beam with the corner of the ac unit is directly at the column.
one other thing you can consider is to add diagonal truss-type bridging to spread the load out among more than just one or two joists.  this does work and can force the load path into multiple paths and reduce the added load on any single joist.  this is usually more appropriate out in the span of the joists, so with your condition, where you seem to be placing it near the ends of the joists, it may not be to helpful.
lutfi...ok that's clear.
using your same process and their maximum allowable end reactions, you get an equivalent uniform load of 306 lb/ft, which is greater than their stated 269 lb/ft.  this would lead one to believe that they are working from empirical data, as is often done with bar joists.  using their same ratios, your equivalent uniform load would be more on the order of 225 lbs/ft which is well below their stated maximums.
ron and others,
great forum. many thanks for everyone who contributed.
for everyone's benefit, vulcraft has a dos program that is free to download at this link
by the way, one key point is to make sure that concentrated loads are at a panel point. if not, make sure to add a strut that will transfer the load to the nearest panel point.
regards
how do you plan on connecting the the ac to the roof steel? often the best approach is to stub legs through the roof, and then seal with a tar box. prevention of roof leaks is usually the biggest concern to the client.
if you're going to use legs then it is still possible to support off of the two steel beams runninmg 90 degrees apart. three of the legs would be directly over steel and the fourth could be supported off a a 45 degree piece of steel that spans between the steel beams and picks up the remaining leg.
it's your call, but it's a lot easier to put new steel on top of the roof rather than than on the inside.