unbalanced snow: ASCE 7-98 vs ASCE 7-02

[bjb]

In my state, we will be going to the 2003 IBC sometime later this year, and it's my understanding that 2003 IBC references ASCE 7-02.

In ASCE 7-98, which I am currently required to use, there is a case for the unbalanced snow load where it varies linearly from 1.2pf/Ce at the ridge to 1.2(1+beta)pf/Ce. This can give you huge unbalanced snow loads, far in excess of what we used to be designing for.

In ASCE 7-02, I don't see this case, and the loads don't appear to be a function of the roof slope anymore.

Does anyone know the reasons for this change? I don't find the commentary very helpful in either case. It seems that with every edition of ASCE-7 they change the way you calculate unbalanced snow. The provisions of ASCE 7-98 are especially onerous in my opinion.

 

[DaveAtkins]

Here in Wisconsin, we have the option of using the 1995 National Building Code of Canada for calculating unbalanced snow loads, because the ASCE 7-98 method is RIDICULOUS. With the ASCE 7-98 method, we literally had to use 2 X 6 purlins on long mini storage buildings because the unbalanced snow became so excessive.

Could it be that the new ASCE method follows this Canadian Code method?

DaveAtkins

 

[bjb]

The problem that I run into, is that often I calculate that my double top plates are overstressed in compression perpendicular to the grain, with similar problems under interior columns located in a wall.

The problem in NY is that we have our own unique snow map, which has grossly conservative ground snow loads for many areas of the state, especially in the area where I live. When we go to the 2003 IBC there is talk the ground snow loads will be adjusted downwards to reflect case studies that have been done.

Because ASCE keeps significantly changing the way unbalanced snow loads are calculated, I wonder what basis in reality the unbalanced provisions actually have?

I am designing a wood roof with a 4:12 pitch theta=18.4 degrees. My ground snow load is 65 psf, W is 24 feet, beta=1, snow density is 22.45 pcf, and pf is 46 psf. These parameters put me into the bottom case in Figure 7-5. The resulting unbalanced snow load is about 109 psf at the eave, and 55 psf at the ridge. We used to use 1.5ps/Ce in our old code, which would have resulted in 69 psf.

With ASCE 7-02, the unbalanced snow load is only 69 psf, and is constant on the lee side. With ASCE 7-02, beta=.5 because pg>40. Ce=1 for all cases. Either I am missing something basic, or this is plain ridiculous. Maybe a 3 year code cycle is just too often. Maybe if they had more time between issuing revisions some of the more ridiculous code provisions could have been weeded out.

 

[theclipper]

My problem with the unbalanced load varying down the roof is that it makes it more difficult to specify steel joists. I believe this loading requires the joists to become SP designated joists.

 

[sundale]

bjb,
Your numbers look right and I agree that the unbalanced loads are now getting far higher than before. I try and visualize a foot or two of standing water at 62.4 pcf and can never really buy that that is a realistic load.

The complexity of analysis is ridiculous given the accuracy to which the snow heights and water content is known. Here in Colorado, where you can get tremendous roof snow loads depending on the elevation, I think the density numbers are way too high for our (typically) "powder" snow and the 1.15 duration factor that assumes that this load is present for two months is also far comservative given our high altitude UV.

What do you do at a roof valley in regards to unbalanced snow loads? The UBC had the Cv boogie factor, which is now gone in ASCE 7-02. The only nebulous guidance given is "Winds from all directions shall be accounted for when establishing unbalanced loads." Does this really mean that you need to consider the 100 year meteorlogical event and then assume the wind deposited two leeward drifts into your roof valley by changing 90 degrees?

 

[bjb]

For a small valley, I usually don't worry about the valley drift, but if I think it is a significant condition, then I use the 97 UBC even though it isn't legally binding to me. The ground snow load map for NY is grossly conservative, so I know that any inaccuracy I have in ignoring small valleys isn't of practical significance. For example, NY is considering adopting a ground snow map that will result in a 20 psf ground snow load reduction in my area.

I agree that the codes are getting way to ridiculous. I am very dissatisfied with the IBC and ASCE-7, I wish they would simplify it. Most buildings built are one to two stories, and in my judgement there should be some simplified provisions regarding the loads for these buildings. The codes do have some simplified provisions, but I think they are too narrow in scope, or not very simple either.

 

[jheidt2543]

This raises are REALLY good question and maybe should be in a separate thread. Why do the major codes need to be changed every three years?

Is there really that much research going on that justifies these changes? Or, is it because the vast majority, not all, but the vast majority of the code committees are made up of engineering profs. and consultants that give seminars on all the updates? And, as noted, the complexity of the various codes is multipling. Just when we become familiar with a code, the next revision comes along!

The dollars generated by the sale of new codes and seminars needed to understand them is huge, not to mention the time consumed to go to them. Owners complain about the cost of construction, this is one place to look. (Maybe I've had too much coffee)

 

[bjb]

jheidt2543: I agree with you 100%. I suspect that many of the members on these committess are not subject to the same schedule and budget constraints that most practicing engineers are. They can spend a lot of time with complex load calculations, but it is a huge burden to the practicing engineer. And then when you consider all of the extra books you have to buy so that you can try to figure out what the code is saying, it does amount to a lot of money. They probably like the 3 year code cycle because it is profitable for them in terms of the extra books they end up selling. Sometimes I think that the people writing the code provisions are on a crusade to prevent practicing engineers from using their judgment in applying loads to a building.

 

[vmirat]

I have been battling the snow load issue here in Colorado for the past ten years. We are situated at the foot of the Rocky Mountains, so we are between case study and 15 psf ground snow load. Since I deal with DoD facilities, I have to resolve a conflict in ground snow load used between local A/E firms, local code authorities and DoD regulations. Local code requires the use of 30 psf "uniform" snow load, which most people interpret to mean flat roof snow load. DoD regulations used to list this particular location as 15 psf ground snow load, which meant the minimum ASCE 7 requirement of 20 psf would be used for flat roof snow load. I have advocated using a ground snow load of 30 but use the modification factors in ASCE 7 which usually brings that down to a flat roof load of 21 psf. After long discussions with the DoD office in charge of the regulation, I finally got the reg changed to the 30 psf Pg figure. However, I still get designs from private firms using the unfactored 30 psf value for flat roof because that is what they are used to using outside of the federal government.

A while ago, I called ASCE and asked them how they came up with the ground snow load maps. They said they were hand drawn. Not very exact in my mind. Of course, now we have web sites that tell you what your load criteria are, such as seismic. So, codes are only as good as the people that develop them.

I think codes have evolved over time as we learn more about the field of engineering and building construction, however I'm with jheidt2543 - three year cycles are too close.