snow drif

huangyhg

snow drift
a client of ours is considering covering most of his (150,000 sq.ft.) flat warehouse roof with solar collector panels. the building is in toronto, canada and the roof live load in this area is for about 20" of snow. the base of the panels are 1'-0 off the roof and they are sloped at 45 degrees, total vertical height would be 6'-0. we originally designed the building and of course did not allow for this type of additional snow drift. my thought is that if i raise the base at least 3'-0 off the roof the wind would tend to scour the snow below and minimize (eliminate?) drift.
any thoughts?
thanks,
peter
you can't count on wind scour to relieve drifting. in my opinion, you have to consider the drift.
for the large roof area, you might want to rationalise it for full ground snow loading, but i would treat the areas for snow accumulation, at least locally.  the irregular surface will be a magnet for attracting snow.  the added weight of your panels will seriously impact the available live load capacity.
dik
if you spread out the solar array on the roof structure, the effect of any drifting would be minimized.
mike mccann
mmc engineering
no.  it appears roof loading will be increased.
you'd have the weight of the panels and their racks + the orignal weight of the (potential 20") of snow + the weight of the snow accumulating under (not melted and wind-blown and melted->refrozen back into ice) the panels.   at least wind load would not be increased too much: the panels will be somewhat sheltered by the raised edges around the existing flat roof.  wind load won't be reduced at all, but it won't be increased too much, that is.
what's the angle of the panels?  (aimed for best summer reception of the rays, i hope) at toronto's latitude of 43 degrees, that would be 43 degree nominal; for summer reception use 43 deg (up from horizontal) + 13 degrees = 56 degrees.   so snow will fall around and between the panels, plus accumulate on the total area panels themselves.  (of course, some will eventually fall of of the panels, sliding down tot he roof.  )
so potential snow accumulation surface has to account for more area: the original flat roof area plus the extra panel area angled at that 56 degrees.   
it/s not the additional area. but the localized drift caused by the panels.
mike mccann
mmc engineering
"at toronto's latitude of 43 degrees, that would be 43 degree nominal; for summer reception use 43 deg (up from horizontal) + 13 degrees = 56 degrees."   
think cook.   in summer, they need to use 43 deg (up from horizontal) minus 13 degrees (since the sun is higher in summer) = 30 degrees final angle.   
so final area covered by snow  = (original roof area covered by panels)/.866 = 1.15 original roof area.  

also...i'm not sure what kind of roof system (waterproofing) you have on the building, but installing anything on 1-foot high frames makes roofing, reroofing, and roof maintenance a nightmare.  in addition, for each frame you'll be making penetrations into the roof which are potentials for leaks.  check the national roofing contractor's assoc. requirements for the height of frames based on size of frame (canada follows many of nrca guidelines, even though us organization).  many of their provisions have also been adopted by building codes.

i would say you have two options:
1. spread them out enough that additional drift will be negligible (as others have suggested)
2. put them all in a single area and stiffen the supporting   
i agree with spreading them out.  the ibc and asce 7 allow you to ignore drifting if a roof projection is less than 15 feet long.
daveatkins