Like others have joked about, it's really about where you want the air and temperature to go.
Let's say you have three fans in a computer case, one at the back, one at the front, and one on top of the CPU. If they all blow out of the case, then you've basically created a negative air pressure zone, and you will be sucking air through the vent holes elsewhere, and through any gaps in the case. This may have effects for where the dust is taken in.
If all of them blow into the case from the outside, you've created basically a high pressure zone, and air will be taken from the room and pushed out through the same cracks and vent holes elsewhere.
A fan if it's pushing against high pressure will inherently push less air because it's pushing against air that isn't moving as well. A fan that is trying to pull air out can also become inefficient if it is working hard to suck air in to eventually push out. Therefore it's best to have matching airflow - what you push out, you should try to also pull in somewhere else in more or less an equal CFM (cubic feet / minute).
For a room, it's best to have one area that is your intake and the other side of the room be the exhaust. This way you aren't just exhausting air you just pulled in. If you just have one fan, I suggest blowing out, which will draw air in from other spaces, such as under the door, because of the negative air pressure by pushing air out.
(I've worked these problems with indoor gardens and HVAC systems, where there's a lot higher CFM fans and spaces, but the theory is basically the same)
I strongly disagree on the setup with only one fan. Having only an exhaust fan in the room will draw nasty air in through cracks in the house's structure, possibly stuff with e.g. mold spores. My wife could never remember which way was which for our window fan; I could tell by smell when she had left it exhausting for more than an hour.
If you have only one, set it to intake and leave another window open a bit (maybe like half the fan size). You can confirm that air is exiting that way by hanging some strips of tissue paper in it.
Positive pressure (blowing in, from outdoors) is also the right thing to do if you are trying to keep smells from other rooms at bay. Whereas negative pressure (blowing out, to outdoors) is best to keep a smell confined to one room (e.g. this is how bathroom fans work).
Agreed. In my own experience, it's much more effective when cooling a room to pull the cool air in than to try to exhaust the hot air from the entire house so that cool air will eventually trickle into the room you're trying to cool. Changing from pushing the air out to pulling it in directly was an epiphany for me when I was younger and lead to many nights of better sleep.
Even better, now that I own my own house, is a high powered fan in one window pushing the air out and another (lower-powered) fan in a different room pulling the air in with an additional window open. You can stand in front of the open window and feel the additional cool air streaming in. It's a beautiful thing.
In some zones, whole-house ventilation fans (which are installed in the ceiling below the attic) are excellent and efficient ways to cool and/or ventilate a house quickly—by exhausting the air in the living area of the house up and out through the passive vents in the attic. Open some windows, turn on the whole-house fan, and the air in the house is fully exchanged with fresh air from outside very quickly.
It helps to reduce stale odors, bring in fresh air, reduce indoor allergens...also handy when you accidentally burn something in the oven. :)
I think the key there is to open the windows at a time when it’s cooler outside. Without sufficient openings to pull air in, the pressure differential is inefficient
Just to add, for computer cases, you want it to be a high pressure zone because intake fans generally have filters in front to catch the dust before it gets into the case. Negative pressure means dust gets sucked in from the cracks.
Same applies for house fans -- you want the intake fan to be slightly stronger than the exhaust fan. Else the negative pressure will draw outside air in through cracks and all the fun smells that come from those places.
(I have a combination energy recovery ventilator/HEPA filter that has has a restriction filter on its exhaust for this reason. Else it would just suck crappy air in from the basement instead of fresh air through the filter.)
If you have to choose one, choose an intake fan for this reason. Best if you can leave another window slightly ajar.
It's a window unit: http://www.purifresh.com/erv.html It's the only such kind I could find when we were in the market a few years ago. (Was renting a place with mold issues.) Works well for a single room. It's wonderful to pull fresh air in during the winter.
While this is technically true, the difference is very minor at the typical pressure ranges being used.
The highest performing PC fans create about 10mmH2O pressure while typically you can expect about 1mmH2O pressure compared to 10mH2O of the atmosphere. This translates to densening of the atmosphere around components by 0.1% or 0.01%, respectively.
Correct placement and design is by orders of magnitude more important than the increased efficiency due to higher density of air.
You could see much more effect by making your air more humid (humid air has higher heat capacity), which will be also much more pronuncable than the increased density due to pressure.
Thanks for mentioning filters. I'm just in the process of designing an ITX case specific to a board/CPU and plan to partition it so the CPU fan is also the case fan. I hadn't thought about dust. Good point.
Dryer sheets :) I don't remember where I found this tip, but recycling used dryer sheets as air filters works fairly well (definitely beat having nothing).
You never want matching airflow in a PC case, it creates pockets of heat, frying components, usually the top disc drive or sometimes the PSU.
Positive pressure is very good at assisting the PSU fan and any radiator exit fans you have, negative pressure is very good at cooling the edges of the MoBo and upper drives in a tower system. Negative is generally cheaper, so its usually negative.
Height is also important if you've got a multi-floor house. In upper stories, make the fans face outward and on the ground floor make the fans face inward and you'll set up a continuous air path throughout the whole house favoring cooler air.
Personally for a single fan in a smaller room where you're near the window, I'd say just point it in. That way you get fresh air right where you are sitting, you get the breeze, etc. In the other direction you won't get the cooling effect on your body from the breeze.
My 4rd grade teacher was convinced of the science of using two fans to cool off the classroom: one blowing cool air into the classroom from the hallway and one fan blowing hot air out of the classroom through a window. The kids that sat near the fan blowing in the room from the hallway were quite content, but the kids by the window were always complaining and they would secretly turn the fan in the window around so it was also blowing in. This would infuriate our teacher when he would discover it and he would claim that the reason the classroom is hot is because the fans were no longer arranged to create air movement through the classroom. But the window kids weren't buying it.
So if sweaty 4th graders are any indication, have the fans blow on you regardless of whatever air movement is supposed to be made.
If the goal is to make the room cooler, your teacher is right. If the goal is to make the students cooler, then the 4th graders are right. 4th graders sweat, and the fan will cause convective cooling of the 4th graders.
This is a concept often poorly understood in basic cooling design. Even without sweat, if your AC vent is blowing directly on you the temperature it is at needs to be vastly different than that of air being using to more actively cool the entire room.
This is often a key difference between Car and Room AC systems.
There’s also a surprising gradient based on how consistent the temperature is top to bottom (and you get into trouble with this on reverse cycle AC systems)
And lastly it’s easy to ignore radiant heat from both walls/ceilings as well as even large monitors.
I should add that my teacher would think we couldn't understand the science or concept behind airflow going through the classroom to cool us off. He'd explain it over and over and get frustrated that so many students didn't seem to "get it." But really they did get it, it just wasn't working for them. I wonder if the arguments about scientific things nowadays aren't similar. It may not be that someone is stupid or doesn't get it, or even that they disagree, just that it's not working for them. The window kids understood that the room was supposed to be getting cooler, but it wasn't for them. They weren't against air flow, but it didn't seem to be working for them. Perhaps in theory the 2 fans in the classroom should have worked, but maybe the fans weren't strong enough, there were other places for the air to go (like the ceiling), etc. so that in practice it didn't actually work like the teacher claimed it would. Maybe there's a lesson there for public debates.
Sounds like a bunch of anti-vaxxers coming together and saying vaxxing doesn't work for them. This is the same reason why pilots are being trained not to listen to their own senses when flying: because those senses work against their reason.
I can see how a single fan blowing right into someone's face could be more efficient at cooling that person than the same fan working for the whole room; but frankly you also forget about all of the other pupils who aren't under the direct "sunlight".
Sounds like a perfect opportunity for the teacher to demonstrate the scientific method by setting up thermometers throughout the room and recording them throughout the day.
The issue is that the teacher needs to actually change the pressure in the room in order to create enough airflow. The fan in the hall needs to create a greater pressure (push air into the room) and the fan in the window needs to remove the pressure. If the fan in the window is stronger than the fan in the hall, then there will be a negative pressure (suction) to pull air into the room.
The problem is that the room probably has leaks such that if the fan in the hall is creating pressure, then the air doesn't all escape through the window. It'll escape through the path of least resistance or maybe even just spin inside the room (hot air rises too). If there is two doors for the room, some of it may be escaping under the other door. In that case, then air is just circulating back into the hallway. People also forget to take into consideration the expansion of hot air, which I've seen quotes for being up to 6x the volume. This complicates the pressures in the room.
As mentioned elsewhere in the discussion, the kids probably wanted to feel cool via evaporation since neither of the fans was moving nearly enough air to cool the entire room enough.
I run a couple large crypto mining farms in a pretty hot country. The 'kids' generate the heat and we have to move large amounts of air through them in order to keep them relatively cool. I've had to learn a lot about airflow design and it sounds very similar to what your teacher was trying to achieve.
Speaking of school, and forgive me for not adding anything to discussion at hand, my biggest distraction was always sweat. I’d get so hot and sweaty most of the times that my butt would leave a sweat stain on the seat. I’d say I spent a good portion of total learning time just worrying about the embarrassment that would result from having a sweaty butt and leaving a stain. Mind you I was not overweight, just a very sweaty kid. I don’t know this for certain but if my school had ac I may have actually learned better by paying attention in class instead of worrying about my butt. Educators take notice.
If all the other teachers were doing this too, the hall might have been negative-pressure enough that the air blowing in was circulating right back out before reaching the rest of the room.
That would mean the kids near the outlet fan were getting air that was just pulled in from outside via the window seams.
If this is because you want to feel cooler, place the fan so it blows cool air inwards onto you. You'll get far more benefit from having air blown on you than you will from just dropping the ambient temperature around you.
Whole house fans are great. They cool the house, clear out the 150º air radiating from the attic, and also take a lot less energy than air-conditioners. I use both, but my air-conditioner runs a lot less since I installed the attic fan.
Same. Well, it's an evaporative air conditioning unit but 90% of the time we don't turn on the evap function (which is only really useful on days which are very dry and not too hot) but just run it as a fan in the evenings. Blows all the hot air out of the house and replaces it with cool night air in minutes.
I've walked around music festivals for days in the sun with a water bottle, and used more water for soaking the hair on my head than for drinking. Works wonders :)
The cooling effect of moving air is mainly because of evaporation of sweat. So it is more important how humid the air is, than how warm. Warm dry air blowing on you will cool you, warm humid air won’t.
There is a far bigger problem than thermodynamics in play, namely: dust. Not just in computer cases: it also applies to places like photography darkrooms. The usual pattern of placing a single blowing-out fan in a ventilation duct is all wrong: it pulls air in from the cracks in doors and windows. The dust-proof way is to make the fan pump the air into the room, but also install a HEPA filter on the air intake pipe/fan.
This isn’t exactly wrong, but ignores the fact that the primary purpose of a fan in a darkroom isn’t to clear dust, but extract chemical fumes. You probably want fumes pumped directly outside, not back through the gaps in the door into the general building.
Everyone seems to be assuming that you have a big fan which actually changes the air flow in the whole room in a major way. If that's the case, great, put it any way you like - it doesn't matter.
For people like me, who use a cheap fan in the summer. Put it in front of you facing towards you. You will feel cooler.
If you are still interested in CFD I can give you a hand, but this is answer we are going to get from it anyway.
I agree with you - it's similar to the principle of a space heater in the winter. You point it at yourself and heat only a small zone. In the summer, point the fan at yourself and use evaporative cooling to cool just a small zone.
Fans big enough to significantly change air flow can be cheap. Window size box fans were $20 at the drug store last I checked. Build quality can't be great, but I had a fan like that last 5 years or so through heavy use.
Assumes small room with only one window and can't open anything else: Place the fan in the lower part of the window blowing down inwards. Sucks in air in bottom half, warm air goes out above the fan.
You are answering a different question. To put it in your terms, do you think your solution is more efficient than putting fan in upper part sucking the air out, and what's your reasoning?
Well, it's more efficient to place something on the windowsill than to hang a rotating fan from the upper part of the window, that will just move around and most likely blow downwards, obstructing the airflow.
When I had an apartment with several rooms and non-working air conditioning, what I did was to open a window near the front of the apartment, in the dining room, and set up a large fan to blow out the window. Then I would open a window in my bedroom, at the opposite end of the apartment, to provide inflow. This was very effective at reducing the summer inside temperature to that of the outside at night, where the purported "heat pump" built in did essentially nothing.
The only issue was that it also sucked rain and pollen in as well.
I assume you want to move the cold air in and the warm air out. How to best achieve that completely depends on the shape of your room and the holes to the outside. You should probably do some computational fluid dynamics on a model of your place to see which blowing direction works better. It probably works best if you have separate inlets and outlets for the air that are maximally far apart, like on the ceiling and in the floor.
Exactly what I do, use a standard temperature sensor, load your system, make the CPU and GPU cry :-) and see how it goes with cooling the inside of the system.
If you look at the design of windows on old houses from about 1700 to 1940 they facilitate exactly this. They open at the top and/or bottom to allow the circulation of air. Hot air will naturally rise and exit from the high up window. When air conditioning took off in the 1950s this became less important.
I agree on "maximally far apart", but more in the horizontal than the vertical direction.
Extract hot air in one location, preferably up high, and let cold air in with a window that is as far away as possible - it doesn't really matter high or low.
Can you recommend any specific software? I'd finally have a reason for the 3d-scanner in my phone... (getting the model at a suitable resolution of <3cm voxel-cube size to not have to model it all from hand)
To cool summer bedrooms, blow the air out the kitchen window, and open the bedroom window. This will draw cool night air in the bedroom window. You may be sweating at 7 PM but you might need a blanket at 1 AM. This works brilliantly. And putting the fan out the kitchen window means the bedroom is quieter.
To heat winter rooms that have an underpowered radiator, blow the fan onto the radiator. The job of heating the room is the same as cooling the radiator. This works very well but the room downstream (on the steam line) will be getting less heat. In practice, that was the neighbor's apartment. The freezing neighbors complained to the landlord who eventually tracked it down to my place. Vigorous discussions ensued.
I also discovered that a HEPA filtration unit, even sitting on the floor, helped my allergies a lot. It was not cheap, but when I find my nose running again, it means that something had turned off the filter. For years I had thought that HEPA was a scam. Not so.
Fans are more efficient at pulling air out than pushing air in. So if you don't want to waste energy put all the obstacles - filters, rooms, etc - on the 'in' side and keep the exhaust as clear as possible
I like to think of it like emptying a bottle - if there's only a small hole it goes glug-glug, as occasionally air forces its way into the bottle.
It's the same with air - there has to be a good way to get cold air in and hot air out with minimal turbulence from the two flows.
I think you're better off blowing cold air in (there's no such thing really as 'sucking' in physics), as attempting to blow hot air out doesn't necessarily mean it's going to be replaced with cold air from the outside, rather than warmer air elsewhere in the building.
Very technically, it is better to have the fan IN the room blowing OUTSIDE through a large opening. The air accelerated by the fan accelerates more air and that air goes outside. It mixes with the outside air so even if some of it goes back to the room it will already be cooler.
If you have fan in the window blowing into the room it has less potential to multiply the air it takes, it will typically be also already IN the room itself which will greatly decrease its efficiency.
On the other hand if your goal is to cool yourself when sitting at the des it is generally better to place your fan blowing in, this means you can direct the flow of air to wherever you want. This means you can direct it at your desk to cool yourself by just the air movement alone. When you place your fan blowing outside you loose that ability. Blowing the air at you means you can cool yourself even if the air temperature outside is higher than inside!
If your goal is to move as much air through your room then
you need to take into consideration the following:
- do you have other openings in the room (like doors... huh)?
- can you have flow of air through the building -- typically there will be pressure difference between two sides of building unless the air is completely still
If you can have your door and the window open, observe the direction of air already naturally flowing. Then put your fan so that it blows with already moving air, for maximum efficiency.
We have a 2 story house where the bottom story is a daylight basement. When it gets hot, it usually stays pretty cool downstairs, so we hang down there during the day. In the evening, as soon as it's cooler outside than upstairs inside, we open the back door, both downstairs and upstairs, and turn on a high powered fan blowing air up the stairs... It is really effective at cooling the house.
I'm convinced that outwards is the most effective in this scenario:
Open your bedroom windows on a cool spring/summer night. On the opposite end of the house, put a fan in a window blowing outwards. This cools the entire house, giving your bedroom the freshest air. In the morning, close all the windows. This leaves the house cooled and drastically reduces your A/C usage.
Blowing inward doesn't have the same effect of pulling air through the whole house.
I usually place the fan so it faces outside. I figure out which way the wind is blowing, open two windows on opposite ends of the room/house that correspond with the wind direction, and then place the fan, facing out, on the downwind end of the room/house. My thinking is that the wind will push itself into the house through the fan-less open window, but that it will then lose momentum. Then, the fan on the downwind side of the house will help pull the air back out of the house, creating circulation that doesn't have to fight against the wind.
Recently, I bought a whole-house attic fan. Now, I open a window on the upwind side of the house and turn on the attic fan, which is enough to suck a toupee off someone's head. The attic fan both circulates cooler air and also clears out the hot air built up in the attic, which helps a lot.
Let's further assume that I'd rather have the room cooler than 30C. Also, regardless of which direction the fan will blow, there also needs to be another place for air to move into and out of the room. So let's also assume that the window is the only opening, but the fan doesn't fill the opening produced by raising the window. One more assumption: I can point the fan with freedom in most any direction once it's placed.
I believe I'd pull the cooler air into the room by having the air blow in. I might also point the fan up and away from the other side of the window (up because I want to disturb/displace the warm air higher in the room; away from the non-fan side because I don't want my cool air to be pushed right back out.)
I don't want to cool the warmer air, I want to replace it with cooler air.
The direction doesn't matter nearly as much as blowing the fan through a large-ish opening.
You want the air from the fan to accelerate other air on the way out/in. This is the same principle as turbofan engines, in which a higher bypass ratio (bigger fan) attached to the same fuel burning jet core is more efficient.
Why? Because the relationship between velocity and volume is not linear. Takes more energy to move the same amount of air the faster you accelerate it.
So ideally you want to position the fan in a large opening far enough back that the fast moving air can get a bunch of other air moving as well. Trade the high velocity fan air for a much higher volume.
Growing up, we had no air conditioning. Instead, we had 2 fans upstairs, blowing out two different windows. In this way, 3 rooms all got air sucked in the windows. (We shut the downstairs windows, since no one slept there.)
Actually the way to do it is to have the upstairs fans blowing out the windows, then the downstairs window open. This is more effective at night when the outside temperature has dropped and the house is still hot. Hot air rises, so you will get cool air pulled in. This is how whole house/attic fans work.
In your scenario, inward. You are looking to achieve the windchill effect. Maximizing this with evaporative cooling would be ideal. You'll need a fluid cooling system that seeps liquid to the surface for this..
If you blow air in, you're sucking bugs in and blowing air out of your central air filters. If you blow air out the window, fewer bugs. That's enough to make me not care about the thermodynamic answer.
This assumes there are enough bugs to be a problem. Some places this isn't the case. Also more likely to have bugs at lower floors of buildings as they tend to hang around trees and plants.
I'd worry more about dust. In a big city, just leaving windows quarter-opened for a week results in everything covered with a visible layer of dust and finely-ground rubber
Assuming you're either near a big road, or your city has a major problem with air pollution. I'm near a sort-of important road in a German city and I don't observe anything of this magnitude even though my windows are fully opened quite regularly. (Air pollution is not that bad in my city. Many other German cities are considering driving bans.)
Maybe you should move to a country with stricter air quality laws. I leave my windows open all summer and I can't observe any ground rubber in my apartment.
So far I've managed without fans because my home has one side in the shade and one side in the sun so I just create a draft and pull all the blindes down to avoid direct sunlight.
Being a Swedish summer, maybe around july august I'll need to take the fan out of storage and then it's just to get a draft on myself. Not to move any air around. The draft between the two sides of my home does that better.
Assuming you want to cool down the room and there is 1 window:
Open the window in the room and open the door to the room. Ideally open a window somewhere else in the building and make sure there is a potential airflow from that other window to the one in your room.
Place the fan in the window of the room you want to cool down and let it blow outwards. Alternatively you can place the fan in the same way in that other window you opened.
I'm not the top floor in the corner room which is the hottest room. The entire unit is per floor is much warmer than any other level. The best non-AC cooling I've been able to achieve is to run the bathroom fans with the bathroom windows closed. It pushes out warm air near the ceiling and draws in cooler air from windows and lower parts.
Too many assumptions to answer precisely. Chances are you'll feel cooler if the fan is blowing inwards because that will help your body reduce its temperature via evaporation of perspiration.
There are much more complex cooling setups that can be achieved that utilise outward-blowing fans though, but a single fan can't will cool _you_ faster if it is blowing onto you.
In my computer setups, I place three intake fans each on both the front and bottom of the case and then place two exhaust fans each on the top and back. Again (see my other post), the idea is to maintain the natural airflow. Since heat rises, I want my exhaust fans to assist with that natural flow rather than fight against it.
For computer cases, take in account dust. You can use dust filters on the intake, and create positive pressure in the case. So you'd need more air blowing in, and less blowing out in order for dust not to accumulate inside (with negative pressure, dust will start actively sneaking in through holes in the case).
A bit of googling and a good dose of intuition leads me to think that the output area of a fan cools faster than the input. Reason being that air is more likely to be turbulent after a foil, and turbulent air transfers heat better.
On the other hand, if the purpose is to express, say, smell, blowing the smell out is the better solution.
If there's lots of humidity inside the room you'd probably prefer negative pressure in the room.
Reason: positive pressure can pressure moist air through walls and ceilings and when it hits the cold air on the other side of the insulation it condenses in thr construction and causes rot (or corrosion).
In to room, via nice HEPA filter, for cleanliness.
Ignoring that, In to the room. Fan can not create much vacuum but can create much pressure.
Let me give you my most handy example. Imagine a queue. A push can cascade from last person to first person, but a pull can only be apply to last person.
While we're on the subject, one thing that I never understood is why fans are directly on top of the heatsink instead of floating a centimeter above it, allowing the air to flow more easily, and also make less noise.
Can I pose a similar question? I have a cieling fan in my conservatory that can blow up or down. Which way is better? It is pretty close to the cieling if that matters.
When we fitted a ceiling fan in our living room, the instructions stated that in summer, it was best to have it blowing upwards (sucking air up from floor level) because that air would be cooler, and in winter from the same principle, you could blow warmer air down onto the room occupants.
In practise it didn't work that well - the temperature difference in a couple of metres of air space (in a single average-height room) is pretty minimal. We left the fan blowing downwards to use the windchill effect in summer - that worked well enough.
Fans don't really cool, they just move air around. Thus you'd want to blow the 30 degree air toward the 20 degree air to decrease the temp inside the room.
If it’s placed under the ceiling - to blow outward because warmer air rises up and would be leaving the room through that opening if there were no fan.
Fans achieve two effects: ventilation (air exchange between two voids) and circulation (air distribution within a void). The first will achieve the goal of replacing a fluid volume with another (hotter/colder), but may not evenly mix the fluid within the conditioned volume.
Overpressures tend to prduce airstreamss or turbulence downstream, but underpressures don't induce much flow beyond the immediate orifice. Unless convection, ducting and venting can reliably create flow patterns, blowing air in to a conditioned space is a better option thansucking ot out, though if you want to control for emissions (say, filtering exhaust), you might want to combile both.
There's a classic physics experiment showing what happens when you run a water sprinkler, the kind with a spinning arm and one or more jets, backwards (sucking rather than blowing), and hhow it doesn't spin. Feynman discusses this in a story.
The fan(s) themselves contribut some additional heat to the space, but this is fairly negligible, especially relative to circulation effects.
If the heat source is external (e.g., incident sunlight), then natural airflow patterns which might draw air through a hotter chamber (say, an attic or other roof void) could have the opposit of the intended effect. It's key to check overall circulatory flow.
And if internal systems, or people, have significant evaaporative or similar cooling mechanisms, increased internal circulation will greatly improve cooling capacity.
Forced inflow may increase dust or other factors, if not filtered.
TL;DR: If you've only one option, I'd blow in rather than suck out. You might want to consider overall airflow, circulation, and heat distribution or souces (and sinks) within the void. Blowing onto rather than away from the primary heat source (in a cooling scenario) is likely best. If heat sources are external, exhaust from the top of the void is optimal.
Intakes should generally be low, exhausts high (for coolng).
For a large enough and complex enough (multiple chambers, zone, multiple heat sources, significant convection effects, dead zones), you would likely want a combination of intake, circulation, and exhaust fans, though I'd prioritise them in that order.
Disclaimer: I'm a space alien cat, not a fluid thermodynamics engineer.
Cooling down your room and warming up the outdoors is the same thing, if it's a purely mechanical problem, because the energy to warm up the outdoors has to come from somewhere and the room is the only other entity.
Most replies intelligently include open windows. Running an electric fan in a closed room with unopened or no windows may prove fatal and can lead to irreversible Fan Death. It's not a good way to go.
Well it depends on the heigth to floor - call that htf, in which direction the fan will work effeciently. Basically you have a flow of cold air inwards at ground level and a flow of hot air outwards at roof level.
So the best solution would be to support the hot outflow, in the mid-layer, and let thermodynamics at the windows do the work.
Let's say you have three fans in a computer case, one at the back, one at the front, and one on top of the CPU. If they all blow out of the case, then you've basically created a negative air pressure zone, and you will be sucking air through the vent holes elsewhere, and through any gaps in the case. This may have effects for where the dust is taken in.
If all of them blow into the case from the outside, you've created basically a high pressure zone, and air will be taken from the room and pushed out through the same cracks and vent holes elsewhere.
A fan if it's pushing against high pressure will inherently push less air because it's pushing against air that isn't moving as well. A fan that is trying to pull air out can also become inefficient if it is working hard to suck air in to eventually push out. Therefore it's best to have matching airflow - what you push out, you should try to also pull in somewhere else in more or less an equal CFM (cubic feet / minute).
For a room, it's best to have one area that is your intake and the other side of the room be the exhaust. This way you aren't just exhausting air you just pulled in. If you just have one fan, I suggest blowing out, which will draw air in from other spaces, such as under the door, because of the negative air pressure by pushing air out.
(I've worked these problems with indoor gardens and HVAC systems, where there's a lot higher CFM fans and spaces, but the theory is basically the same)