Parapet Wind Pressure on Low Roof

[CURVEB]

What would one use for a parapet wind pressure on a low roof adjacent to a tall building? Calculating this load per ASCE7-10 would you take h as 250' or 50' (specifically for the parapet design in question)?

Would your answer change in any way if the low roof height was closer to the upper roof, say 230'?

Also, at what point would one consider a parapet to become more of a wall with openings along the top as opposed to a parapet assuming there is roof above, as in the below sketch? At some point, the wall would enclose the space such that you no longer have exterior pressures on the back side. I would think that there is a point where the space inside of the wall becomes partially enclosed, then fully enclosed as the top of the parapet gets closer to the roof above.

I guess the more specific question is at what point would it transition from a "partially enclosed" condition to a "parapet"? Is a 1' gap considered partially enclosed, but 2' or more you would treat it as a parapet?

Thanks in advance.

 

[phamENG]

I would take h as 50'. I would use the actual h value for the roof in question. So if it's at 248.9 feet, I'd use that.

For your second question, it depends a lot on what's around. How is there a roof there? Are you looking a glass guardrail setup on a balcony with another balcony above? If so, I wouldn't consider that a parapet as there's nowhere for the air to flow on the backside, and as such no suction pressure to develop. Wind design is a fluid flow problem. What is the fluid (the air) doing as it flows past the building? Parapets receive the loads they do because 1) there's a pressure acting against the surface and 2) they cause the flow to separate as the air goes by creating turbulent eddies that create a vacuum on the backside. If there's a wall there just a few feet away, you develop a back pressure that prevents that suction. If you don't, and you've somehow invented skyhooks to suspend this roof, then leave wind design to the peasants and go enjoy your billions of dollars.

Kidding aside, what scenario is creating this? I'd be inclined to say that I would do parapet as long as the parapet is half of the space between roofs or less. More than that and it starts getting a little gray. I'd try to stick with parapet loading as long as it doesn't make for an unreasonable connection or member size. If it does, I'd start looking for a way to rationally reduce the load.

 

[CURVEB]

Thanks for the input. To be more clear about the design issue, it is an enclosure of an existing roof-top deck. I say rooftop, but it's actually the level below the roof, so there is a slab above that covers the entire area and is supported on columns at the perimeter. Just typing it out has brought up another question: is the floor for this level technically even a roof surface, because the actual roof of the structure is one level above? While it might be exposed to some wind uplift, it's probably not nearly as much as the level above which is the actual roof, so in the case of leeward wall pressure + roof uplift pressure, I think there's a reduction there.

For the windward wall + leeward wall condition, I agree with you that if there is a sheltering wall or building just a few feet back (as in the case of a balcony) you couldn't necessarily see the windward pressure on the backside of the parapet.

While this all passes the test of reasonable engineering judgment, I wouldn't know how to justify that the code supports these reductions. The main crux of the issues seems to be how 'h' is defined as the building height, not the height of a portion of the building, and the fact that section 30.1.4 prohibits taking any reductions for apparent shielding. To further confuse the matter, figure 30.4-3 seems to imply that if the step between the upper and lower roof is significant, you would use the lower roof height (h2) to determine the lower roof pressures. Perhaps the newest edition of ASCE7 clarifies conditions like this.

 

[skeletron]

Replying using NBCC 2015 as a reference:

1. My tired brain suggests that I would bracket the solution. Design the parapet for a 50' height assuming no adjacent building as a starting point. Then, assume that there is complete shading on the far side. Assess the difference. Then start to mingle with the actual leeward pressure on 250'.

2. I suspect that the H/d ratio of the building also plays a roll. If you can use a triangular loading, you may get a more accurate result.

3. NBCC 2015 deals with stepped roofs (Fig. 4.1.7.6D) that may be of assistance.

4. Save for a scale wind tunnel design, I suspect the lack of shielding provisions in the code is a consequence of wind loads (fluid flow) being relatively difficult to generalize and predict.

5. For the second exercise, I believe there is a way to calculate the gust factor based on the area of your opening. So, use the gap as a continuous opening and relate it to the code provisions.

 

[phamENG]

I think it's important to really dig into the meaning of qz vs. qh and why each is used, when it's used. The windward pressures are applied using qz, which results in a variable distributed load, increasing with height (though conservatively simplified to a UDL below 15ft). That makes sense - winds are slower at the ground and faster above the ground, generally, due to frictional resistance to flow. qh is used for everything else because the suction effects aren't as clear cut as the windward pressures. As the flow separates at the edges of the building, the suction generated becomes harder to generalize along the building's height. Therefore, the codes present a conservative general case (use the worst value). The roof less so, since qh = qz at the roof. But that's another important point. The roof loading is calculated based on its height, not some arbitrary point in space. So using qh for the roof level in question (50ft) fits the intent of the code.

A tough revelation that I had to come to grips with: the ASCE wind provision are only directly applicable to about 20% of the structures I design. I don't do high rise, but the architects are always coming up with something that puts it out of the simple cases. 7-10 really only deals with rectangles thoroughly. There are provisions for domes, but they're a bit incomplete. There are stepped roofs, but clearly they don't cover them very thoroughly and they are dimensionally limited. 7-16 does introduce L-shaped buildings and canopies attached to the side of structures - I'm not sure how much else. The fact of the matter is that the code is limited in scope and, no matter how you justify it, if it doesn't fit into that very narrow window you're activating the accepted literature or wind tunnel clause.