Which way does the air flow?

[RCAV8R13]

Hi guys,
This should be an interesting discussion.
Which way does the air flow through the motor fairing in an EDF? The fairings I'm talking about are like the one pictured. They have a hole in the back of the fairing.
Does air flow out (aft) through this hole or in (forward)?
Here, I'll plant the seed of controversy. I contend that the air flows in the hole in the back and forward through the motor and out the gap between the motor and the fan.
Kip

[JetPlaneFlyer]

I'd tend toward the flow being toward the rear but I'd suspect that there was not any great velocity in either direction... Why dont you glue a few fillements of fine thread near the exit and observe which way (if any) they blow when the fan runs?.. or parhaps if the thread method is not sensitive enough place some talcum powder or similar in the fairing and see where it comes out or if it just sits there?

[RCAV8R13]

I'd tend toward the flow being toward the rear but I'd suspect that there was not any great velocity in either direction... Why dont you glue a few fillements of fine thread near the exit and observe which way (if any) they blow when the fan runs?.. or parhaps if the thread method is not sensitive enough place some talcum powder or similar in the fairing and see where it comes out or if it just sits there?
Humm, tattle tail strings would do the trick. With a clear plastic tail pipe one could observe the effect.
Good idea.

Kip

[Brandano]

If the impeller hub is flush with the motor mount there is nothing forcing air through the motor from the front. However, the backplate of the impeller will act like a centrifugal blower, and effectively extract air away from the motor through its front venting holes (if present), which would be replaced by fresher air from the back venting holes of the motor. you could check this by running a brushed motor in place of the current brushless and checking how much graphite dust from the brushes gets on the motor windings through time. That said, this probably isn't optimal, since in the case of a brushed motor the graphite dust will cause a mess, and the cooling air the motor uses is at least in part mixed with the hot air it has already used for cooling.

[RCAV8R13]

Here's what I think is going on.

Kip

[JetPlaneFlyer]

You could well be right..
I did some experimenting with a motor fairing type shape (actually just a thick symmetrical airfoil at zero AoA) with Profilli and sure enough the pressure at the trailing edge is higher than the pressure at the mid chord location... Some testing is called for to find out the truth :) .

Steve

[bwalt822]

RCAV could be correct. But the area where the hole is will be a stagnation point that is also within separation. Separation areas have low pressure but possibly not lower then the low pressure in the upstream slot by the fan.

There are variables that could cause it to go either way. If you cant get tufts (the thin strings) to work then you could put a drop of oil just inside the hole and see which way it is blown when you power the fan up. Just be sure to hold the fan at an angle where gravity doesnt make the oil run

[Neil Stainton]

I agree with RCAV8R13, in a parallel duct the air will probably leak forward, BUT the duct shouldn't be parallel. It should try to maintain a constant CSA, in which case the air flow through the housing will be small.

To re-iterate, the pressure and thus the direction of flow depends upon the CSA of the duct at the two ends of the fairing.

Neil.

[RCAV8R13]

I agree with RCAV8R13, in a parallel duct the air will probably leak forward, BUT the duct shouldn't be parallel. It should try to maintain a constant CSA, in which case the air flow through the housing will be small.

To re-iterate, the pressure and thus the direction of flow depends upon the CSA of the duct at the two ends of the fairing.

Neil.
Well Neil, i agree with you. The duct usually does taper after the motor but I've never seen one that followed the same shape as any of the motor fairing. So, the pressure increase at the rear of the fairing probably would not be as big as what I drew (parallel duct) but the csa would get a bit bigger. So now the question is:
How can we increase the flow through the motor fairing?
Kip

[Brandano]

The contour shouldn't follow the same shape of the motor fairing, the keyword here is area. if the diameter at the fan is 50 mm and the radius of the fairing is 25 mm, the area of the duct is ((50/2)^2*pi) - ((25/2)^2*pi) = 1472.62 odd mm^2. If we remove the fairing, to get the same area we need a diameter of sqrt(1472.62/pi)*2= 43.3 mm, which is only marginally smaller than the original 50 mm, and yet is a much smaller area than a 50 mm diameter duct. This goes to show how important the duct size actually is, since small variations can cause large differences in overall area.

[ugly john]

RCAV,

I think you have it a bit off on your diagram. The higher velocity=lower pressure idea works with a wing in a free stream, but I do not think it applies here.

The area directly behind the fan will have high velocity, but it will also have the highest pressure. As you move aft over the fairing, the diameter and CSA increases, the velocity lowers, but the pressure will also lower.

IMHO, the pressure gradient will move the air flow through the motor in an aft direction.

UJ

[hul]

the 69mm Schuebeler has an annular gap behind the rotor that's likely meant to collect cooling air and guide it into the motor tube -> flow would be front to back.
Without this gap, I would say air flow is back to front like Kip (RCAV8R13) said in post #5.

Hans

[RCAV8R13]

RCAV,

I think you have it a bit off on your diagram. The higher velocity=lower pressure idea works with a wing in a free stream, but I do not think it applies here.

The area directly behind the fan will have high velocity, but it will also have the highest pressure. As you move aft over the fairing, the diameter and CSA increases, the velocity lowers, but the pressure will also lower.

IMHO, the pressure gradient will move the air flow through the motor in an aft direction.

UJ
Bernoulli's principal can be simply stated: You can either have velocity or pressure but never both. If the velocity increases the pressure must decrease and vise versa. Don't confuse pressure with force.
Kip

[Texas Buzzard]

A minute ago ran my 65mm ducted fan while I hand held it. I aimed the exhaust towad a table from about 3 ft. The table had several paper napkins on it's top ( brand name = Bounty)

All of the napkins blew off the table. When looking at the rear of the fan, my hair blew around. So what do you conclude from that experiment? :o :)

[RCAV8R13]

A minute ago ran my 65mm ducted fan while I hand held it. I aimed the exhaust towad a table from about 3 ft. The table had several paper napkins on it's top ( brand name = Bounty)

All of the napkins blew off the table. When looking at the rear of the fan, my hair blew around. So what do you conclude from that experiment? :o :)
Well, let's see. When the high speed low pressure hair hits the table, napkins or your hair, it slows down. The kinetic energy (inertia) is transferred to the table, napkins or hair. This transfer slows the air and the pressure goes up.
How's that?

Kip

[Neil Stainton]

> So what do you conclude from that experiment?

That fans blow, and you don't understand the question ;)

[Texas Buzzard]

Take a look: Here is copy & paste from the original post.

""Hi guys,
This should be an interesting discussion.
Which way does the air flow through the motor fairing in an EDF?""

Neil, if air doesn't "flow through" the fan unit then why did my tissue paper blow off the table??

Now it is possible that the way the question was posed to us, there can be more than one interpretation of it's meaning.

That question asked which way the air flowed throught the MOTOR FAIRING.
Perhaps in your mind the "motor fairing" is the "can" of the motor and NOT the outer-most part of the EDF, which I call the "fairing".

Maybe the question left some wiggle room. So that's why you said that I didn't understand the question.

Sometimes I am in error so please describe what you Think the Question Means. Be specific with your terms. Thanks

[ugly john]

Bernoulli's principal can be simply stated: You can either have velocity or pressure but never both. If the velocity increases the pressure must decrease and vise versa. Don't confuse pressure with force.
Kip

That only works for incompressible fluids.

UJ

[RCAV8R13]

> So what do you conclude from that experiment?

That fans blow, and you don't understand the question ;)
Oh I get it now. :D Of course it depends on one's perspective, some times they suck. ;)
Kip

[Neil Stainton]

That question asked which way the air flowed throught the MOTOR FAIRING.
Perhaps in your mind the "motor fairing" is the "can" of the motor and NOT the outer-most part of the EDF, which I call the "fairing".

I think the question means "what direction does the air flow within the motor fairing/shroud", not "what direction does air flow between the shroud and the duct". RCAV8R illustrated this in post #5.

Neil.

[Neil Stainton]

That only works for incompressible fluids.

Air is an incompressible fluid below Mach 0.8 or thereabouts. Ie at ducted fan speeds.

Neil.

[ugly john]

Nope. Air is considered compressible above about 300 mph. Doing a quicky calculation with a 50,000 rpm, 2.14 ID fan, the fan tip speed is about 317 mph - That is tip speed, not the air speed. Obviously we can't guess what kind of air speed that would produce, but I would bet money it is well within the compressible range.

Now, density doesn't change until approaching mach 1.

And that is about as technical as I can get tonight. My aero and fluids books are at work, it's saturday and it's miller time! :D

[Brandano]

Bernoulli's principle operates on a principle of conservation of energy. if you maintain the same amount of energy in an air volume, you can convert pressure in speed or speed in pressure but can't have both increasing or decreasing at the same time. When you start to put in more energy into the system, as it happens behind an impeller, things get a bit more complex. The air accelerates, but the pressure doesn't necessarily decrease, or axial flow compressors wouldn't work. However, a lot depends on the shape of the ducting. I think that the proof of the pudding would be to add a few pressure taps onto a ducted fan ducting, attach them to a few lengths of PVC tubing and dunk these in a container of colored water to see where the water column raises the most. My bet is that high pressure behind the blades is quickly converted in speed with a sudden pressure drop as soon as the flow becomes less turbulent.

[Texas Buzzard]

I think the question means "what direction does the air flow within the motor fairing/shroud", not "what direction does air flow between the shroud and the duct". RCAV8R illustrated this in post #5.

Neil.

It does make a difference. "Inside the motor" and inside the EDF ( inside the shroud) just may mean two different zones....yes they are two different zones......Thanx for the help Neil.
I will modify my responce to that queston:
It is likely that the air pressure had dropped at the forward end of that motor due to the sucking away of the air. So the pressure differential will cause the air to enter the motor from what we consider the aft end.

[graupman]

I think the previously stated idea that this flow will depend on the change in area as the air moves through the duct is correct. Another factor that might play into this is the shape and "direction" of the gap. Sort of like the difference between a pitot probe and a static port - dynamic pressure matters.

I've attached a very simple axisymmetric cfd simulation to show what might be happening. I am not sure how accurate this is, but this simulation shows a low speed movement toward the tail of the plane.

[bwalt822]

Do you have the pressure plots from that CFD. Also its hard to tell but it doesnt look like the velocity increases by much when it goes into the reduction. Also, what is the recirculation region in front of the fan looking thing?

[graupman]

Static pressure contour plot attached.

The "bubble" at the edge of the fan is due to the fairly large gap I left in my mesh between the edge of the fan and the shroud. I think the backflow is probably exaggerated by inaccuracies in the cfd. This is an inviscid simulation - That backflow would probably be reduced with viscosity and a finer mesh. At least that's my guess.

[bwalt822]

im guessing that red means high pressure. If so, shouldnt static pressure be higher towards the back where velocity decreases?

Im not sure you need the jump where the fan is to simulate this effect. Try removing the fan and the pressure differential and see what comes out. Inviscid may or may not be the right way to do this because you might need the separation at the end of the hole in back.

[JetPlaneFlyer]

shouldnt static pressure be higher towards the back where velocity decreases?


Why would velocity decrease when the duct area is reducing? To get the same mass airflow the velocity must increase as the duct area reduces.

Even with a constant area duct very close to the fan (where the upstream bleed connection is modelled) there will be a little bit of compression from the fan, which can be seen in the simulation.

Based on this it looks like I may have been right with my gut feeling that the air would bleed in the same direction as main flow.

[JetPlaneFlyer]

Bernoulli's principal can be simply stated: You can either have velocity or pressure but never both. If the velocity increases the pressure must decrease and vise versa.
And therein lays the fundamental misunderstanding that many have with the Bernoulli principal. If this were true how come a compressor can take static air as atmospheric pressure and send it whistling along a pipe at high velocity AND at pressures many times atmosphere?

As stated earlier Bernoulli principal is all about conservation of energy so it only applies where no energy is added or taken away from the system... In the real world there is no such condition because friction is constantly 'removing' energy from the system in the form of heat. However in the case of a ducted fan massive energy is also added by the fan so it's perfectly possible that the air downstream of the fan will have both higher veloctity AND higher pressure than atmosphere.

Steve

[JetPlaneFlyer]

Nope. Air is considered compressible above about 300 mph. Doing a quicky calculation with a 50,000 rpm, 2.14 ID fan, the fan tip speed is about 317 mph - That is tip speed, not the air speed. Obviously we can't guess what kind of air speed that would produce, but I would bet money it is well within the compressible range.

Now, density doesn't change until approaching mach 1.

And that is about as technical as I can get tonight. My aero and fluids books are at work, it's saturday and it's miller time! :D

Pressure and density are directly proportional to one another (assuming constant temperature)... If you increase pressure then density must increase because when you compress the gas volume decreases (this is the definition of compression) but mass of course remains constant (and density = mass/volume)

Given that the two are directly and inextricably linked it's folly to claim air is compressible at 300mph but it's density cant change until mach 1.
Truth is air is 'compressible' (and therefore it's density can change) at any velocity but in a freestream flow the compression effects are small enough to be ignored at lower speeds.. Where you decide to draw the line and call the effect significant is somewhat arbitrary.

[bwalt822]

For this CFD model behind the fan, area looks like it is increasing which means velocity decreases. Fans increase TOTAL pressure and usually dynamic pressure, not necessarily static pressure.

Think about it like this, the forward port is essentially a static pressure port while the rear is a total pressure port(if there is no massive separation). Total pressure is always greater than static pressure which means the air will flow forwards.

Just because someone makes some Color Filled Drawings does not mean they are correct. (no offense graupman, you are on the right track and can settle this if you mess around with some settings)

Graupman, is that fan looking thing a wall in your boundary settings? And how did you do that pressure jump? The backflow could be caused by the high pressure leaking forward which it should not do to the extent seen. Viscosity and a finer mesh probably wouldn't fix it.

[Neil Stainton]

the rear is a total pressure port(if there is no massive separation).

Isn't the rear port static pressure minus dynamic pressure?

Neil.

[JetPlaneFlyer]

Think about it like this, the forward port is essentially a static pressure port while the rear is a total pressure port

But the rear port is orientated so that the dynamic pressure is negative (i.e. the dynamic pressure 'pulls' an the air within the port).... If the port was forward facing then yes the dynamic pressure would be added to the static pressure to give total, but as it is I'd say dynamic would be subtracted.

So I'll turn you argument around: 'Static pressure' is always larger than 'static pressure minus dynamic'... therefore flow will be 'rearward' as the model indicates.

*** EDIT*****.. I see Neil spotted the same thing!

[graupman]

This is the point where I need to point out some things about my simulation. This was a VERY simple set up - it's not necessarily giving you "real world" conditions. The thing is, I don't thing that is necessary.

The fan disc is everything in this problem. Without it, it would just be a venturi tube like a carburetor - the flow would definitely move toward the front from the low pressure in the venturi throat.

When you add the fan, you get a pressure increase from one side of the fan to the other. It's possible that I am misunderstanding some things, but I was taught to think of a propeller or fan as an "actuator disc", where the pressure immediately in front of the disc is p, and the pressure immediately behind it is p+dp. In this simulation I set the boundary conditions to have a 0.5 psi pressure jump from one side of the fan to the other, and the intake and outlet to be atmospheric pressure (a static thrust condition). This pressure jump is what is pushing the air toward the back of the duct, and it will push the air through the housing just the same. The only time that it will move forward is if there is a "leak" from one side of the disc to the other (along the shroud).

It has been a long time since I learned this stuff, so if any of my assumptions are wrong, please let me know. I read these things because it's interesting and I want to learn!

[bwalt822]

Jetplaneflyer and niel, you have managed to confuse me and you might be right, but... Isn't the back port a stagnation point which means that the pressure there is dynamic plus static? I thought dynamic pressure always positive because density is positive and velocity is squared which is always positive?

if the port in the back is static minus dynamic then the air should still flow towards the back without the fan right?

Graupman, can you run your CFD with an inlet velocity and without the pressure jump and fan just to see what happens?

I think that depending on the inlet conditions and fan power/pressure jump the flow could go either way

[JetPlaneFlyer]

Jetplaneflyer and niel, you have managed to confuse me and you might be right, but... Isn't the back port a stagnation point which means that the pressure there is dynamic plus static? I thought dynamic pressure always positive because density is positive and velocity is squared which is always positive?

Think of a plain straight tube, open both ends in an airstream. If the total pressure at both the forward and rear ends was equal to dynamic pressure plus static pressure then the total pressure at both ends would be the same and there would be no flow through the tube. CFD is not required to prove this to be incorrect ;)

Steve

[graupman]

If an inlet velocity is added and the fan pressure jump is removed the flow will reverse.

If the fan is adding no energy (static pressure) to the flow, the static pressure will be lower in the constriction due to the smaller area, therefore pulling the flow in the housing toward that low pressure area. Once the fan starts running, or adding energy, it will increase to a point where the static pressure drop due to the venturi equals the static pressure increase due to the fan (where flow through the housing is zero). Then once the fan begins to add more pressure than the drop due to the venturi, the flow will move toward the back.

At least that's the way I've been visualizing it...

[bwalt822]

Think of a plain straight tube, open both ends in an airstream. If the total pressure at both the forward and rear ends was equal to dynamic pressure plus static pressure then the total pressure at both ends would be the same and there would be no flow through the tube. CFD is not required to prove this to be incorrect

Air flows through that straight tube because it is already moving and no work is being done on the flow, there are no pressure changes. Just because pressure does not change doesn't mean that there is no flow. A straight open tube does not have a stagnation point that matters in this case.

But now take that same front part of the tube and bend it 90 degrees, cut it in half and stick a pressure gauge in between the two halves. The wouldnt the pressure read less on the 90 degree bend side due to the loss of dynamic pressure at that front opening while the other opening would remain the same, now the air wants to flow the other direction.

graupman, you are probably right. The question is, does a normal EDF produce a large enough pressure change to push the air backwards. An axial flow compressor certainly does but where is the line?

[JetPlaneFlyer]

Air flows through that straight tube because it is already moving .

Dynamic pressure is the energy associated with a moving fluid. the fact that it's moving is what gives it dynamic pressure and like anything related to velocity its a vector quantity that has direction as well as magnitude.

[bwalt822]

Dynamic pressure is the energy associated with a moving fluid. the fact that it's moving is what gives it dynamic pressure and like anything related to velocity its a vector quantity that has direction as well as magnitude.

But dynamic pressure is velocity squared so direction does not matter, the squared term is always positive. If you turn a pitot tube 180 degrees, theoretically it should still work. I say theoretically because its hard to find clean air when its backwards and we are having a theory debate.

and for anyone else watching, the "flow" we are talking about is the flow in the small tube depicted in graupmans cfd images.

[JetPlaneFlyer]

But dynamic pressure is velocity squared so direction does not matter, the squared term is always positive.

Good point!...
But still... if you imagine that the open ended tube is full of stagnant air, there is a valve at either end preventing any air movement. Then someone opens the valves. If the total pressure is equal at both ends of tube then the static air sitting inside would just 'sit there' because there would be no force to cause it to move.
This is not what would happen in reality I think we both agree?

[bwalt822]

The air would start moving backwards but this is probably due to effects that can be ignored in this discussion such as friction. In an inviscid simulation, if this happened then yes, the air would stay there and not start moving. We are talking about a steady state operation here where the only forces that matter are pressure forces. We(I) are(am) also assuming that there is no separation at the rear port.

If you turn a pitot tube around it would probably measure a slightly slower velocity. This slight difference is not enough to cause the air to flow the other way im guessing.

[graupman]

A pitot tube will only work in the 180 degree position in inviscid flow. That matching stagnation point can only exist if drag is completely zero. Obviously this is never the case. Direction is the difference between a pitot port and a static port. On a full size aircraft, the air speed indicator is measuring the difference between the pressure at a stagnation point (front of the pitot tube) and the free stream pressure (the static port). The static port is usually 90 degrees to the flow.

To address another question, I think that as long as the fan is creating thrust, the flow through the housing must be rearward. The point where the flow is zero or near zero is where the fan is generating about the same amount of thrust as there are losses in the duct. If the thrust is positive, the air has to be pushed out the back, and as long as the “rear” opening of the housing is near the free stream pressure, it will be lower than the pressure just behind the fan.

[Julez]

I'd strew some styrofoam crumbs in the hole and see if they stay inside or not ;)

[Neil Stainton]

In real flow a rear facing hole will have a pressure close to static minus dynamic. That is how total energy probes used for variometers in gliders work - they report the static pressure less dynamic, which is approx constant if a glider is exchanging altitude for airspeed in still air.

Neil.

[JetPlaneFlyer]

I'm confused.. my head hurts :confused: