Using "roof snow load" as a design load

[PBengineer]

I've been engineering small pole-type buildings for about 5 years now. 3 years ago, we switched over to the IBC from the 1997 UBC. Our company requires that our customers obtain snow loads from the counties for us to use in design. Many counties provide them with "roof snow loads", and I know that the "norm" of many civil/structural engineers has been in the past to take the ground snow load, reduce it, and use this as the "design" load for roof components. In fact, most counties accept this from other engineering companies.

The problem is that I use unbalanced snow loads in my design, which is usually 1 to 5 psf above ground snow, and considerably more than that for roof snow load. I can find no justification in either the old UBC or 2003 IBC to "stop" at the roof snow load calculation and use this. But I'm developing a reputation for "over-engineering" these "low-cost" structures (trust me-adding a few more purlins is a big deal to these contractors!). I've asked some older (that is "more experienced") engineers how they justify doing this, and no one has an answer except "I don't know-that's just the way it's always been done".

I'd really like some input from some senior civil/structural engineers on this.

Thanks

 

[Taro]

You need to get clarification from the counties as to what the "roof snow load" means. You seem to be assuming that they are giving you the flat roof snow load (p-sub-f) which you then must increase to account for unbalanced drifting, etc. However, some jurisdictions allow designers to design for a conservative uniform snow loading and ignore all drifting effects.

 

[aggman]

Assuming that you have a situation where IBC/ ASCE 7 requires you to design for unbalanced snow loads and you are required to meet the building code then I think you are doing the right thing. I would be sure that I was allowing for the lower importance factor as you see fit. You would be suprised at the number of failures due to snow drifting. If one fails and for whatever reason you are taken to court then you won't have much of a leg to stand on if you don't account for it. Having worked in an industry where I see structures under-designed regularly its not to hard for people to second guess you or to second guess yourself when something looks heavier than the other guys. Then one day that lighter structure will fail and yours does ok and people start to change their minds. Just be sure you are doing everything right and that you are accounting for you reductions correctly. I know that the wood building industry is really competitive (like metal buildings) so it's pretty easy for people to start taking short-cuts because it's never been a problem.

 

[dik]

as Taro suggests, clarify with building official... in a pinch you might consider the GSL=RSL/0.8 as is often the case... and then treat it as you normally would.

Dik

 

[ctcray]

In cases where the clarification from the B.O. is less than informative, I usually will just consider the required "roof snow load" as a separate roof load case. When checking drifting, sliding, and unbalanced, I base it off of the calculations in ASCE 7 based on "ground" snow load, from the ASCE maps.

I get the impression that many jurisidictions get confused by the term "ground snow load" so they don't properly specify it. In their infinite wisdom, they think "but we need to specify ROOF snow load!"

 

[DaveAtkins]

I agree with the comments above. But for determining unbalanced snow load, check if your State allows an "alternate" method. Here is Wisconsin, the State allows us to use a Canadian Code which yields much more reasonable unbalanced snow loads than ASCE 7.

DaveAtkins

 

[PBengineer]

(boy,I hope this isn't a dumb question...)

dik-where does GSL=RSL/0.8 come from? I've never seen this in the UBC or IBC. When would you use just this, and not account for drifting snow over the ridge on a gabled roof?

 

[dik]

PBEngineer:
Typically in Canadian Codes, the normal design roof load is 0.8 X GSL or Ss. We also include Sr which is a rain load in conjunction with the snow loading. Formulae for unbalanced conditions usually refer to Ss and Sr. We have a Ca for accumulation that uses Ss and Sr to determine the unbalanced conditions. It was based on this that I suggested using the 1/0.8 as a means of replicating the GSL condition.

Dave... thanks for the plug.

Dik

 

[J1D]

Right, 0.8 is the basic roof snow load factor. But in accordance with NBCC (National Building Code of Canada), the Specified Snow Load for roof includes rain load as well. As given in Sentence 4.1.7.1

S = Ss(Cb*Cw*Cs*Ca)+Sr

Where, Ss is the ground snow load; Sr is the associated rain load.

PBengineer, I think the roof snow load may certainly exceed ground snow load. But it is better to consider the load case by case.
In my cases, according to NBCC, S is smaller than Ss since

Cb=0.8; Cw=wind exposure factor, typically use 1.0, but can be 0.5 to 0.75; Cs=roof slope factor, typically 1.0, to be reduced when slope larger than 30 degrees; Ca=the snow accumulation factor, typically 1.0 to 1.25 (larger when the slope is steeper, but it is compensated by Cs at the same time), for large flat roof Ca may exceed 1.0 as well; Sr is usually smaller than 10% of Ss.

For flat and shed roofs or shallow gable (less than 15 degrees), arched and curved roofs you don’t need to consider unbalanced loading, just apply “full and partial loading” with “the specified uniform snow load in Sentence 4.1.7.1.(1), computed using C a = 1.0,distributed on any one portion of the loaded area, and half of this load on the remainder of the loaded area”. For steeper gable, arched and curved roofs you’ll consider the unbalanced loading (e.g., snow on half of a gable)as well, Ca can be up to 1.25

More likely, the roof snow load can be higher when you have snow drifting and sliding situations.

Hope this helps.

 

[VTPE]

ASCE assumes that their rof snow load includes rain on snow surcharge if the ground snow load is greater than 20 PSF:
The ground snow-load measurements on which this
Standard is based contain the load effects of light rainon-
snow. However, since heavy rains percolate down
through snowpacks and may drain away, they might not
be included in measured values. Where pg is greater
than 20 lb/ft2 (0.96 kN/m2), it is assumed that the full
rain-on-snow effect has been measured and a separate
rain-on-snow surcharge is not needed.

In addition, the sloped roof snow load is based on Pf and Ps, not Pg (although there is the beta factor that has Pg in it, you could be conservative with this factor)

 

[dozer]

PBEngineer,

All of the roof collapses that I'm familiar with were as a result of unbalanced snow loads. Don't let the money folks talk you out of checking unbalanced loads to save a few dollars.

Also, I would caution you, about getting snow loads from the counties. It's very likely that whoever is feeding you these numbers don't know their butt from a hole in the ground. Look at the snow maps and be sure the values they give you pass the sanity check. If it's in one of the blacked out areas . . . good luck.

 

[twinnell]

Use ASCE 7 to calculate design loads. Also, if you have a pop-up structure, such as a tower, above your roof...make sure you consider sliding snow. A lot of roofs fail because the snow from a high roof slides off and lands on a roof below. There is a section in ASCE 7 that covers those loads.

 

[IFRs]

Create a portfolio consisting of copies of newspaper or journal articles documenting failures from unbalancd loads to send to clients who don't know or care.