NFPA 92 and building with small atrium openings

[blackarrow90]

Hello

Currently, I am working on finding a solution for smoke extraction in an atrium building. The building has 5 floors with 10.000 sqft each. According to the NFPA 92, this is a type of spill plume, since the fire would start on one of the floors and spill to the open space in the middle. The problem is that the hole that represents the atrium is relatively small compared to the floor area (about 1000 sqft) and has escalators in it. So, there is no much room left for the smoke to go through, and once in the atrium, it gets easy to the upper floors where it poses danger to the people. NFPA 92 describes the atrium buildings where the atrium has a huge area and smoke is able to rise up more or less freely. Here, I have a hole in the ceiling where a lot of smoke comes through, from all four sides.
There is a smoke extraction point at the top of the atrium with a capacity of volumetric flow rate of 45 m3/s. The fire calculated spreads according to t2 curve (fast) and reaches around 1.5 MW before the sprinkler activates (each floor is equipped with sprinklers).
Using NFPA 92 calculation methods I get very high volumetric flow rates required to make proper smoke extraction. Is there any other way to address this specific case of the atrium?
Would be better to use a mechanical supply inlet instead of just natural inlet openings? Although the velocities of inlet air would be fine, I have concerns once the sprinkler activates. It will effectively lower the smoke temperature and take away some of its buoyancy.

 

[UFT12]

It seems NFPA 92 is not the applicable standard for your case. See Atrium definition in 3.3.1.

With just 1000 ft2 (93 m2) and escalators in it, it rather seems as a typical large communicating space which is usually (or should be) protected by draft curtains all around and dense sprinkler spacing. See for example figure A9.3.5 of NFPA 13. Some other codes like NFPA 101 require that these openings not to be used as means of egress so smoke management around these openings should not be the primary design intent.

 

[blackarrow90]

The draft curtains around the open space (or placed horizontally) would be the simplest solution, however, this is not desirable. There are draft curtains that are 1.3 feet high and located at the ceiling around the open space. This helps, but not a lot since these small curtains are quickly overrun by the smoke. The escalator stairs will not be used for evacuation, of course, and each floor has enough means of egress. However, each floor has to have a smoke-free layer under 2 m of height to allow for safe evacuation.

The smoke spilles from one of the floors into the open space. Controlling it is not easy, since the open space is small, there is not enough room for it to spread, and the smoke tends to spill further on the next floors. So the idea was to try to supply air in the floor above the fire origin, where the smoke is first to enter (blowing about 14 m3/s in addition to 3 m2 of windows and doors on the floor's facade). I know that the correct method would be to let the fresh air come in the floor under the floor where the fire originates, but since there is just not much extraction capacity at the top of the atrium, this only accelerates the smoke spread and fills the upper floors faster. This is verified in a CFD simulation as well.

So I am looking further for the solutions which would keep the floors more or less smoke-free for around 6 minutes, which are required for the evacuation. Letting more fresh air through bigger openings in the facade allows taking out the mechanical supply, but I am not sure if this would be the right solution. Especially because I can't simulate the effect of sprinklers, which will probably cool down the smoke and render the natural supply ineffective. So I lean towards mechanical supply because it should not be that much affected by smoke temperature, right?

 

[UFT12]

In general, mechanical supply is strongly discouraged (see argumentation in NFPA 92). You need to take advantage of the natural buoyancy of the smoke, regardless if the exhaust above is mechanical or not. Supplying air, especially from above the fire origin, will create turbulence and most likely will lead you to wrong pressurization and unpredictable patterns of smoke travel. These complex cases and for such openings that do not qualify as typical atriums are certainly out of NFPA 92 scope and the equations you find there most likely do not apply (see explanatory annexes).

If the escalators of these openings are not used as means of egress why not just letting the smoke travel to the topmost level and concentrate your efforts to meet the intent there. You do not need to meet a clear height at the escalators. Using draft curtains on each floor (1.3 ft do not qualify, see the standards for recommendations), will help to avoid spill to the floors above and concentrate the smoke only to the topmost level. I would strongly recommended them.

At the end, the best would be to simulate the case and run sensitivity analyses. See FDS software of NIST. Btw, if you are familiar with FDS, the smoke interface with sprinklers may not be able to be realistically simulated but using the right tricks to describe the HHR profile can lead you to worthy results since the sprinklers in reality are expected to reduce the smoke production.

 

[blackarrow90]

I understand that the natural supply would be the first option, however, until the smoke reached the extraction point at the top, its spreads and endangers the floors in between. I validated it using FDS and as you mentioned, "capped" the HRR curve once the sprinkler activates (HRR constant after about 3 minutes). As I mentioned earlier, my biggest concern is simulating fire by limiting the HRR only. In the case of sprinkler activation, the temperature will be lower, there will be turbulence, and relaying just on natural supply seems too risky for me. As soon as there is no sufficient buoyancy due to the cooling from sprinklers, the smoke will behave differently.
When you mentioned draft curtains on each floor, did you mean the ones that go from the ceiling to the bottom of the floor or fixed ones that are permanently attached to the ceiling (for example, glass panels or similar)?
I also simulated what happens when the supply air is blown mechanically into the floor above the floor where the fire origin is. It turned out that this supply of air acts as a "flushing" tool to keep the floor clear of smoke and allow people to evacuate. It is actually possible to meet the tenability criteria this way and limit the openings in the facade which has to be opened automatically after the fire's detection.

 

[UFT12]

Blowing air from corridors on floors (communicating space) towards the atrium/opening is actually supported by NFPA 92. The problem of this method is that it typically results in large amount of flows for large openings. If your simulation shows valid and realistic results (e.g. maintaining a clear height) that are supported by your data and is a realistic scenario to built, then this is the way to go.

The advice about draft curtains was based on two things, a) reduced spillage of plume under the floors above (plume typically widens due to entrainement) and b) reduced x-sectional area for the opposing blow of the supply air. The curtains are not meant as floor-to-ceiling separations rather, attached on ceiling and extending vertically downwards. At the end, whether they help on the given geometry or not, is something that can be shown only by running sensitivity analyses.