ASCE 7-16 snow drift load on adjacent building with separation 2

[bnickeson]

thread507-443423

I have a very similar question to the above referenced thread where ASCE 7-10 and 7-16 seem to have nonsensical requirements for leeward snow drifting where there is a separation between buildings. If the separation is less than 20 feet, 7-10 and 7-16 are requiring the full snow drift (hd) be applied over a width of 6*hd or (6h-s), whichever is less. The 6*hd is even more width than the standard leeward snow drift width of 4*hd since the code doesn't specify that it can be truncated or that the 6*hd can be measured from the face of the taller building. So what 7-10 and 7-16 seems to be suggesting is that you have to apply more snow drift if there's a separation between buildings than you do if there is no separation which is completely nonsensical.

In the four years since that last thread, has anyone gotten any clarification from ASCE on if this is an error in the code?

 

[driftLimiter]

What exactly are you suggesting is an error? The code + commentary seem to lay it out pretty clearly to me.

The idea is that there is aerodynamic shade from the roof of the upper structure. And if you meet the criterion in Ch. 7 then you apply the drift loads. The entire concept of drift is about snow that is suspended in the air by wind flow. The code gives that wind profile approximately 1:6 rise/run from he upper roof down.

If you think about how drifting snow accumulates I don't think this is non-sensical at all. It flows with the wind. The wind doesn't stop and take a 90degree turn towards the ground after it sweeps off the upper roof.

 

[bnickeson]

The idea of snow drifting is not nonsensical. I get that, I get the mechanism. The application of it per the current code isn't very rational though. For instance, if you have zero separation you apply hd drift over 4*hd width. If you have a one foot separation you apply hd over 6*hd width. You are required to apply more snow because of that one foot separation. If you have a 12' separation, you again apply hd over 6*hd in an area of roof where you likely wouldn't have had to apply any snow drifting at all if your separation was zero. If you have a separation of 19', you again apply hd over 6*hd but if your separation is 21' you don't have to apply any drift at all (this assumes a somewhat tall vertical height difference between buildings).

Again, these provisions don't make any rational sense.

 

[RWW0002]

@ Drift Limiter - Take a look at the sketches in the attached thread and the follow-up discussion. While drift on structures within the aerodynamic shade makes sense, the application of the limits make for some odd conditions.

For instance - Look at Case I in the original sketch - building with 19.5' gap vs Case 3 same building, same gap, but with roofing between the structures, and Case 4 - Same building with a 20.5' gap.

It seems the general consensus was that new research gave more insight into drift within the aerodynamic shade, the actual code provisions and limits set make for some odd conditions that are a bit counterintuitive. - See previous thread for discussion.

To my knowledge the provisions have not changed for ASCE7-22, but I have not had to dig into it too deeply yet. Waiting for my next project with a 19' gap between buildings..

 

[phamENG]

Again, these provisions don't make any rational sense.

I disagree. They make rational sense - though there are some inconsistencies. Here is a Structures article about snow drifting in general with a good summary on the physics of leeward snow drifting. The 1/6 slope falls in between the two, which probably makes sense when looking at the longer distance. So I think that replacing the S-20/S with (6h-S)/6 is a good move. The overall shape makes more sense. It does have the unfortunate side effect of creating a discontinuity at 20ft of separation.

As to why that discontinuity exists...I'd wager that it has to do with probability of drift formation. As the separation grows, the likelihood of a drift forming probably falls off at some point maybe it's in the 18-20ft range? I don't know.

I would be curious to hear what ASCE's response is if anyone has bothered to inquire - but I think the nicest word I can use for calling the standard writers' decision irrational would be 'hasty.' There are a lot of things that go into the limits we use, and unless you're prepared to read every single referenced piece of research and meeting minutes/discussion notes from the committee meetings, it would probably make more sense (to say nothing of having a shred of respect) to dial back the language a bit. Read up on Chesterton's Fence. It's a good self calibration tool.

It'll all be moot shortly. ASCE is completely changing the drift calculation for 7-22.

 

[RWW0002]

I agree with you Pham that we do not always know all of the research and/or discussion that goes into the provision changes, and that this one in particular appears to have a sound research basis for a change. However, as many of us have noticed over the years, the simplification of the research into code provisions sometimes has unintended consequences.

I asked a similar question to the author of the article you linked and his response indicated that the primary intent of the change in the code provision was to "eliminate from consideration small vertical gap cases". The code committee "only wanted engineers to worry about cases where the lower roof was close enough and low enough to be in the drift shadow" I have a feeling that in the process they unintentionally "created" the anomaly that has been pointed out. Additionally, there are not a ton of observations of "drifts across a gap" to show that the current approach is incorrect

In this case, based solely on the code provisions, I still wonder if the drift hd was intended to be applied at the face of the higher roof wall face and truncated across the gap, as opposed to applying the full drift at the edge of the lower roof. The commentary clearly indicates otherwise though, and there may be a clear reason for this (considering that wind and snow are a bit fickle that way..). in the words of Mr. O'Rourke "For these odd-ball geometries with few or no actual observations, it is better to be safe than sorry. "

 

[bnickeson]

Again, if they want to err on the side of conservatism and have us apply the full hd snow drift for smaller gaps, fine. I'm OK with that. However, it seems like there needs to be some sort of reduction or interpolation term once you get into the 15' to 20' gap or so. Apply a full drift if there's an 18' separation but zero drift if there's a 21' separation makes no sense at all.

Additionally, it's not that the code is requiring you to put the same drift load on a structure with a separation. It's requiring that you place MORE drift on the structure with a separation due to the 6*hd width. If that is based on wind tunnel testing I would love to see the data.

 

[RWW0002]

@ bnickeson - You are preaching to the choir a bit. I wrote the original thread and chased the reasons for the change as far as I was able/willing. On the project I was involved in at the time I either encouraged the design team to increase the distance between buildings or I reinforced the building for the full drift (I can't recall at the moment). My gut tells me that the code change was intended for 1 thing and happen to result in a higher drift condition for the case you (and I) pointed out, however there was not much data basis to "dial back" the provisions for the odd case of a gap approaching the 20' threshold. While I agree that having large jumps in loading at a somewhat arbitrary limit is not very practical in its application, it is what we are stuck with for now.

At the end of the day even with the best of intents we still end up with some odd code provisions from time to time. I don't think you have much choice here but to design to the current provisions, even if they are a bit counterintuitive at times. As PhamENG points out above, the provisions are being reworked currently. If this issue persists it might be worth chasing it a but further down the line.

 

[ProgrammingPE]

The 2010 code change to use a 1:6 wind shadow for leeward drift on adjacent structures makes perfect sense when your drift happens to be located above the wind shadow (it can be ignored) or when it is located right at the edge of the wind shadow (snow that would be above the 1:6 line gets cut off). It does, though, result in two clear anomalies when your drift falls entirely below the edge of the wind shadow.

The first anomaly is that adding a tiny separation distance between structures changes the drift from using the normal 4*hd to using 6*hd, which results in additional snow drift. While 6*hd makes sense when the snow is being trimmed by the 1:6 wind shadow line, it makes less sense when the drift is below this line. Using 4*hd that transitions to 6*hd may end up matching actual cases better, but is the additional complexity worth it here?

The second anomaly is the hard cut off for drift loads at 20 feet of separation. For drift locations in the wind shadow, the separation distance has no impact at all on the leeward drift load until it reaches 20 feet, where the drift load immediately disappears. Any time there is a sudden change in values, intuition should say that it is not correct. Some sort of transition region would match actual conditions much better, but it ends up being a question of what is the correct way to transition, and is it worth the added complexity.

ASCE 7-05 had a factor that went from 1 to 0 for separation distances of 0 feet to 20 feet. It appears that that did not actually match observations which is why the change was made. I imagine that that is partly why the hard limit of 20 feet exists, since the code committee didn't want to suddenly start requiring snow drift evaluations for buildings that are further than 20 feet away. It does seem reasonable, though, that a very tall building could blow its snow onto any structure that falls in its 1:6 wind shadow, but checking every structure for this could be overkill, especially without any reduction for separation distance.

Structural Engineering Software:

http://www.structuralcentral.com/tools/snow-load

Structural Engineering Videos:

http://www.youtube.com/structuralcentral