Recently I observed the following:

There were two rooms divided by a mesh screen that is commonly installed on house windows or screen doors. Both rooms had external doors closed and no ventilation and the air was not moving. In one room somebody left a burning cigarette in the ashtray and the bouyancy-driven flow was the only perceptible flow in the room. Smoke went to the cieling and then moved more or less in all directions at very low speed. Smoke reached the screen that partitioned two departments and basically turned as if it was a solid wall. Maybe there was some propagation trough the screen but it was insignificant when compared to the smoke that stayed inside. Smoke did not penetrate trough the screen although

1. There was some (small, say 0.01 m/s) velocity component perpendicular to the screen 2. Screen opening is enormous (say 0.001 m in diameter)

Question is - why the smoke does not go trough a screen?

Interesting problem, I don't have a good answer, but I have a question - How do you know that there was a velocity component perpendicular to the screen (passing through the screen I assume)? If the smoke didn't go this way then how could you "see" or "learn" about this perpendicular velocity?

There was no visible smoke in the next room. There had to be a component of the velocity perpendicular to the screen because smoke was slowly moving along the cieling towards the screen.

My theory is that the boundary layer from from the wires start to mix early on (close to the front of the screen). So, one has a fully-developed flow case at hand even though we are dealing with a wire-thin distance. This in itself is strange considering conventional rules of thumb when it comes to flow trough pipes.

Part of the incident flow hits the wires and then turns around. In the process, this counter-flow mixes with the incident flow and reduces the average speed of the incident air that makes it to the open part of the screen.

This speed reduction may amount to a significant reduction of the average speed (which is already very low) through the screen. It may well be that I did not wait long enough to see substantial amount of smoke making it trough the screen.

I would like to compute some of these effects, at least order-of-magnitude, but I am not sure how to do that. Specifically, I don't know

1. how to describe the thickness of a boundary layer around the wire.

2. How to estimate the average speed reduction of the air speed as it approaches a screen.

I would appreciate very much if you could provide some guidance in this respect.

2. How to estimate the average speed reduction of the air speed as it approaches a screen.

You can derive a boundary layer approximation of flow over a flat plate with transpiration (suction or bleeding). I think a simple form of that is available in Schlichting. That's assuming the flow has already turned parallel to the screen.

Adrin Gharakhani

(1). This is not a CFD question. (2). If there is a burning cigarette in a room, then, someone must have been in the room before. (3). If he was smoking, then he must be either thinking or working hard in the room. He could be relaxing in the room. So, he must have been in the room for a while. (4). The room air temperature is always lower than the body temperature. So, if a person has been sitting in the room smoking and relaxing (or working), then his body temperature will heat up the surrounding air and set up convection. He is also likely to heat up the chair, which will give up heat after the person is no longer in the room. (5). So, depending on the size of the room, and the time duration which the person was in the room, the convection of the air could stay in motion for a while, with the additional heat source from the seat and the floor. And the donut type of flow motion will stay in the room for a while. The air flow has to come down to the floor in order to satisfy the continuity equation. (fron the ceiling to the floor along the wall or the screen surface). (6). It is likely that the wall of the room is also colder than the air in the room. In this case, the air on the wall will have to flow downward toward the floor. Thus it will enhance the existing donut type air motion in the room. (7). So, the smoke is unlikely to go through the screen, except that, the screen wire will create lateral smoke flow motion and diffusion. The air in the room behind the screen is something like a dead-water region.

Hi,

I've got an idea (and I grant you that it might be silly):

1) I do not believe that the smoke does not actually go through the screen - it's just that we cannot see it any more.

2) If I want to get good mixing (read: high turbulence), the screen is just what I need - that (grid!) is how you make turbulence. Also, even if the velocity is low, the Re number based on the diameter of the wire will still be decent. Isn't it possible that the smoke actualy goes through the screen but the turbulence mixes it so well with the rest of the air that you cannot see it any more. I suppose the best way to check this is to go to the other side of the screen and have a smell - you should still be able to smell the smoke even if it is well mixed.

How about it,

Hrv

I am sure that all these factors that you mentioned have a part in this event, but I doubt that they play a major role when compared to the active source of smoke - burning cigarette. Temperature gradients (floor-cieling-walls) are probably not as high as across smoke-air boundary.

You brought up an interesting point: room response as a whole. At these low speeds boundary layer grows laarge rather fast and the geometry of the room might be a significant part of the explanation.

My estimate is that the Reynolds number is on the order of 0.001!

If I assume that flow trough the screen is like a flow over the flat plate, boundary layer mixes almost instantaneously (close to the front of the screen) and one would think about that one deals with a fully developed laminar duct-type flow trough a very thin screen!

Smoke probably does make it through the screen, eventually. But I still think that the component of velocity perpendicular to the screen drops off sharply just before smoke hits the screen. Screen is like a 50% permeable wall. If my hypothesis is right, then the average speed perpendicular to the screen is perhaps a factor of 10 or even more lower than the speed 10 cm away from the screen. Like you, I beleive that smoke does make it through, but with a substantial time lag.

(1). At the point of the burning cigarette, it is hot. (2). But the hot gas and particles(the smoke) will quickly mix with the surrounding air, and the temperature of the mixture will also drop quickly. (3). So, it is really depends on the total amount of energy released from the object to heat up the air. The force produced will be a function of the volume of the mixture and the density difference. And that has to come from the total amount of energy released to the air. (4). Can a burning cigarette release enough energy to warm up a seat? I don't know. Because of the large wall area involved, a small temperature difference will definitely set up air motion in the room, regardless of the presence of the chair and the burning cigarette, I think.

I agree completely with your arguments: bouyancy flows exist even with small temperature differences. However, smoke visibly propagates propelled by the source energy and I think that it is the dominant component of the flow.

Hi, Marko,

Kghm, at these low speeds the smoke particles move with the air. The relative velocity of the smoke particles can be estimated if you wish but the answer is obvious. Therefore, if the smoke does not go through the screen hence the air does not go through the screen, too. This is certain. The calculations you want to perform are irrelevant.

Concentration of the smoke particles is usually too small to affect the flow. (Again, can be estimated). Therefore, the smoke just marks the fluid particles but does not affect the flow itself (This is not for sure, smoke particles can affect the effective density).

However, the smoke marks heated air. You may speculate on why warm air did not pass though the screen while the cold air did. This may be interesting speculation but probably irrelevant, too. There are many much simpler explanations why the air moved in the room in this way. For example, there was convective motion due to some other reasons (see John's message, or difference in wall temperatures, or ... many). The easiest way to get an answer is to repeat the experiment twice, once as it was to be sure that the effect is reproducible and then with the screen replaced with a closed door, for example. Quite probably, you will see the same.

The key thing here is that at very low speeds the permeability of a mesh screen can be much smaller than it seems from a simple observation of the area of the openings in it. The important quantity here is the drag coefficient of a wire and it tends to infinity as Re->0.

Rgds, Sergei.

Another reason could be that smoke is a solid aerosol. Aerosol particles track the fluid flow nicely for flow patterns in a free stream or through a big hole. But even a window screen can filter out a lot of of very tiny particles if the flow is slow enough. Small particles are attracted to the wires, and the aerosol particle concentration diffusion gradient drives more particles to the surface of the wire, even though the particles are much too small to be physically trapped in the holes. Look at at the wires on a window screen, and you'll see that they are very dirty, even though the particles look like they could have made it through the holes in the screen.

Hi, Phil,

So, you think the smoke was filtered by the screen. Kghm. No. It is not filtered even by a cigarette filter.

But this is an interesting exercise. The smoke was observed moving along the ceiling. You can hardly observe a smoke for more than several minutes before it diffuses (take 2 min). The distance it passed is approximately the size of the room (take 2 m). This gives a crude velocity estimate: V=2 m/ 120 sec \approx 1 sm/sec and the velocity through the screen can hardly be less which reproduces Marko's estimate (remember, I do not think there was an air flow through the screen, but if the movement of the air in the room were due to the suction through the screen our estimate would be reasonable).

Hence, to be captured, a smoke particle moving through the opening should develop a velocity with respect to the air which has the same order of magnitude, that is about 1 sm per second, due to the effects you mentioned. Well, I'll drop this now. Everything obvious. No way.

Rgds. Sergei.