snow drift

[bjb]

We have a condition where a new parapet will be tall enough to create a snow drift on an existing roof, which is framed with steel bar joists. REinforcing the joists is not desired because of piping running through the joists. We could design a rooftop platform to relieve the snow drift, but we are considering another approach.

Rather than build platform using wide flange beams, it has been suggested that we could install some light-gauge steel framing on the roof which would take the shape of the drift, and thererfore prevent the drift from forming. In other words, the drift is 5.7' high and 22.8' long, and we could build sloped framing that is 5.7' above the top of the roof against the parapet, and then dies into the roof 22.8' back into the building.
This would be like constructing a low slope gable that dies into the flat roof. In thread 507-126797, the poster indicates that a Professor from RPI is of the opinion that for a flat gable, probably no windward drift would form.

What we are considering is building a flat light-gauge steel gable that attaches to the new parapet about 8" below top of parapet. We then come back at an angle of about 10 degrees until we die into the existing flat roof, making sure that our new gable extends beyond the theoretical drift and that our angle is flatter than the theoretical angle. I think this eliminates a surface for a snow drift to form against. This will affect the way the roof drains, but that is being re-worked anyway.

What do you think?

 

[JAE]

Just check your new load pattern on the existing joists, including the extra dead load from the sloping roof framing. It may reduce the joist shear at the ends and may add to bending moment at midspan...or the other way around. Just need to model and check it against the existing joist shear/moment capacity envelope across the span.

 

[bjb]

Do you think the theory has merit in that the new flat gable would prevent the snow drift from forming? It seems to me that this concept doesn't provide a surface for the snow to pile up against, therefore eliminating a snow drift.

 

[EKizak]

As JAE stated check the existing joist for the extra dead load. Also if the new light gage shed is unheated, you will need to take that into account in the snow loading. Unfortunately this will increase the required snow load.

 

[bjb]

If we go this route we are going to put holes in the existing roofing for ventilation and heating.

 

[rrmiv]

I think your solution might work, but I think the key is attaching so close to the top of the existing parapet. I'm not so sure about the need to make sure your slope is "flatter than the theoretical angle". I don't know of a provision in ASCE7 that describes the maximum slope upon which a snow drift can occur. It seems to me if the snow has a vertical surface to drift against (assume you attached 3'-0" down from the existing top of parapet), then snow could drift up against it even if the new roof sloped 10 degrees. Can drift only accumulate on "flat" roofs?

 

[DaveAtkins]

I agree that this solution (assuming you address the issues mentioned above) does prevent snow from building up. Of course, at the bottom of the new sloped roof, you have somewhat of a "sawtooth roof" situation, but I typically don't get too concerned about snow buildup in roof valleys anyway.

DaveAtkins

 

[JAE]

The older codes used to have snow drift requirements for valleys, which is sort of what you have. But in the latest IBC codes and in the ASCE 7, I believe those provisions do not exist anymore.

 

[rrmiv]

Now I'm curious as to the opinion of others. Suppose the attachment is made 3'-0" down from the top of the existing parapet. Would you consider drift up against this barrier on the new sloping roof, say 10 degrees? I believe I would. Is it the slope of the new roof itself that prevents the drift or is it that no parapet is present to drift against? I think if a parapet is there, drift at a slight angle must be considered. At what angle is this not the case? I don't believe ASCE7 addresses this.

 

[bjb]

As we looked at this situation some more, we determined that builing this light gauge "gable" isn't the way to go for this particular building. We have intersecting drifts in corner conditions, so the geometry of the light gauge framing gets more complex. Also the magnitude of the drift is large-we have almost 900 feet of roof feeding the windward drift, and the parapet is 8'-6 above the top of the existing roof. I'm just not comfortable using light gage framing in these circumstances.

Thanks for the replies though. The light gauge framing option is an option we will keep in mind.

rrmiv:

In the condition you describe I would consider drift on the new roof. I think that the concept only works if you take the new framing to near the top of the parapet. You would need hc/hb to be <0.2 to not have to consider drift.

I'm not a fan of the continuing complexity of the codes, but I think it would be nice if they did do a little more research into snow drifting. Hopefully they would remember to keep it in a user friendly format for practising engineers, remembering that we have to perform to a budget and schedule which are often both inadequate.

 

[DaveAtkins]

I don't think slope of roof has anything to do with amount of snow drift (unless a roof is so steep and slippery that it won't even have balanced snow on it). If there is a step in the roof tall enough to cause drifting, there will be snow buildup on a lower, sloped roof.

DaveAtkins