Snow Drift Simplification - Canada

[r-struct]

Engineers in Canada or any engineer in this forum who has to deal with snow loads:

Assuming I have a typical 120ft long by 60ft wide single story building. There are three main W-beam bearing lines supporting OWSJ (OWSJ length is 30ft). Beams supporting joists are all 20ft spans.

Say there is 6ft tall parapets all around and some 6ft tall roof top units (maybe 6, one every 20 ft bay) distributed uniformly at the middle beam line. Assuming snow drift is 80psf max, 12ft long, with base snow load of 35psf.

If you design all the beams and columns for around 35+2/3*(80-35) = 65psf uniform snow load based on engineering judgement and also to save time, also assuming water ponding will not be a problem, do you think it is too conservative or do you model the actual trapezoidal loads at parapet and roof top units? I just wanted to check how other engineers here usually deal with this.

Thanks in advance.

 

[dold]

I'd bet you that your RTUs are less than 15ft. long. If so, you do not have to consider drift at the units (according to ASCE 7-10 anyway, not sure about Canadian codes).

But...depending on how close RTUs are and how long they are, I can see the merit in considering those as drift 'accumulators'. But it doesnt take all that long to find the true reaction at your beam on gridline 2 (assuming the front wall runs along grid 1, back wall is grid 3. Side walls are grid A and G. Joists span from 1 to 2 to 3. beams run from A to G, correct?

As for the side walls where the joists are parallel to the wall and each joist away from the wall gets smaller uniform drift load...Yes we actually calculate the load on each joist up near the wall (I call it the drift zone). In fact, I calculated it so much that I made myself a handy spreadsheet. Nice and dynamic like...

edit to add pics

drift spreadsheet - DDZ

drift spreadsheet - no DDZ

 

[jayrod12]

Call me conservative, but I'd probably run the numbers using 80 psf everywhere. Let the joist guy decide what he wants to do, but then I don't give two shits if they move the units 14 times during design and tender, I know all beams are ok.

 

[skeletron]

I would take @jayrod12's approach because the optimization doesn't need to happen at the start.

 

[pmtrevisan1]

I design per Jarrod
never had an issue

 

[KootK]

Cool spreadsheets sold. I used to have a similar thing in MathCAD but nowhere near as well done.

 

[r-struct]

Thanks everyone for the replies.

Those are some nice spreadsheets dold. Yes, what you described is what I had in mind. If I remember correctly (not very sure), Canadian code recently allows up to 10ft and snow drift on roof top units can be ignored. I still think i makes sense to consider drift snow loads around rtu's.

Kootk, how do you simplify snow drift loads? Or do you analyze it in detail?

Thanks.

 

[KootK]

I do a lot in detail but should surely use Jayrod's approach more often as it is quite reasonable. Your original proposal actually strikes me as pretty reasonable too.

I use RISA Floor which allows one tapered area loads with ease. That tends to tempt me I to ill-advised detail.

 

[r-struct]

Good for you for being able to do a lot of things in detail . I've tried to use maximum snow drift value previously but was met with frown by some senior engineers.

I guess it kind of makes more sense to do it in a more detailed manner OR use a reasonable uniform value to avoid every conservative assumption spilling over up to foundation design (also to avoid being fired due to having too heavy designs).

 

[Yao1989]

65psf is good for beams. No way they are all 80psf same time. Different story for OWSJ and deck though. Either show it on loading plan or let OWSJ & deck supplier deal with it.