Understanding Stagnation Pressure in Cooling Fans

[Saladsamurai]

Okay. So I am slowly being introduced to some CFD modelling and stagnation pressure keeps coming up as one of the variables.

We are modelling coolers, which of course have a fan(s) that distribute the cool air.

What exactly are we saying if a fan has high or low stag pressure?

I looked it up on Wiki and others, but the definition is not all that clear to me. It says that P_stag is the pressure at a point of zero velocity.

So with that definition in mind, how does P_stag translate to fans?

What does it say qualitatively about the fan and its affect on cooling?

Thanks,
Casey

 

[Cyrus]

Cooling depends on temperature difference. The only way this would make sense is that higher stagnation pressure would result in a higher pressure jump across the actuator disk (plane of the fan) causes a higher wake velocity. This would translate into a higher Re when calculating the Nusselt number for heat transfer via convection.

 

[minger]

Stagnation or total quantities are often times used for boundary conditions because they describe the incoming "energy" of the flow.

For example, let's say that we're using CFD to analyze a simple 1D pipe flow that exits into the atmosphere. For well-posedness, we must place no more and no less than 2 boundary conditions at the inlet and 1 at the outlet.

Ask yourself what do we know? We _know_ the outlet static pressure, it's going to be ambient. Do we know the inlet static pressure? Well no, we don't. The inlet static pressure will be a result of velocities, losses, etc, which are to be computed. However, we can use the total pressure, which is a sort of energy that is contained within the flow. I say energy because as per your definition it is the static pressure in a fluid if the fluid is brought isentropically (no losses) to zero velocity, i.e. all of the kinetic energy is converted into internal energy.

Today's commercial codes allow the use of mass flow and other variables to be used as boundary conditions. However, realizing that the flow solver is computing derivatives on the conserved variables, specifying pressures and temperatures is usually the best way to go.

 

[Nick Bruno]

Where is the stagnation pressure taken on a fan? The blade tips?

 

[FredGarvin]

Pressures are usually a linear group that cover equal flow areas that spread from the ID to the OD. Similar to a pitot traverse.

 

[Nick Bruno]

So basically you are just measuring the velocity of the flow?

 

[Nick Bruno]

this is an interesting site you may like. If you click tutorials on the side bar it takes you to a balancing tutorial and talks about different kinds of coolers etc.

http://images.google.com/imgres?img...&sa=N&start=20&um=1&ei=_rF4SoKeOoiHlAey4JyaBQ

 

[Mike_In_Plano]

As I recall from working in R&D at Trane, fans would be fed into a duct that was roughly the area of the fan. Some distance down the duct, it was assumed that turbulence from the fan would die down and the back pressure could be measured.

Pressure was measured by placing flush taps around the periphery of the duct. The taps were all plumbed together to a common point where the pressure was compared to the inlet pressure of the fan - which was always atmospheric pressure.

Downstream, restrictions or variable-speed fans would be used to select the flow rate. I can't remember how we decided flow. Perhaps with anemometers?

In any case, each type of fan had a family of curves relating speed (measured with reflective tape and a tachometer) and flow. When flow was 0 CFM you reached your ultimate pressure or stagnation pressure.

In addition - Simply knowing a fan's curves, or the predicted pressure drop is insufficient to ensure good results. Turbulent flow against the part to be cooled is MUCH more effective than laminar flow.

Also, between two fans that have apparently the same performance, one may markedly outperform the other due to the formation of rotating air currents. This is particularly true if the fan is somewhat obstructed. Rotating air currents act normal to the direction of desired air flow, setting up a dynamic pressure drop (like an air curtain).

Aside from magazine articles, I've never seen a cooling problem attacked by CFD. I have seen an entire department of engineers fail miserably only to have the problem solved in about an hour by a contracted aerospace engineer. I'm curious how the results of the CFD turn out.

I wish you well,

- Mike

 

[Nick Bruno]

I have seen a cooling problem solved in COMSOL which, I think is a form of CFD and the results were accurate and neat to look at too!

I have a quesiton, why do you say "Turbulent flow against the part to be cooled is MUCH more effective than laminar flow." I don't understand these physics.

 

[Su Solberg]

Okay. So I am slowly being introduced to some CFD modelling and stagnation pressure keeps coming up as one of the variables.

We are modelling coolers, which of course have a fan(s) that distribute the cool air.

What exactly are we saying if a fan has high or low stag pressure?

I looked it up on Wiki and others, but the definition is not all that clear to me. It says that P_stag is the pressure at a point of zero velocity.

So with that definition in mind, how does P_stag translate to fans?

What does it say qualitatively about the fan and its affect on cooling?

Thanks,
Casey

I am working on Air Compressor acceptance test's review.

Consider air flow in a pipe, base on terms on ISO standard,

Stagnation Pressure is very similar to Static pressure in that case.
Both of them are talking about the pressure of a air stream brought to rest.
However, Stagnation make a requirement that the method to make it rest should be isentropic (KE->internal energy).

If there is a test pipe with no heat lost, you can treat stagnation P = static P.

p.s. I am a newbie in engineering. Please feel free to point out my misunderstanding and thanks for your understanding.

 

[Mike_In_Plano]

Re: Turbulent flow vs laminar flow for cooling

As I was taught, the flow velocity normal to a surface must always be zero.
.
The flow tangential to the surface varies according to the type of flow. With laminar flow, the viscosity of the gas drags against the surface, the air forms into streamlines, and the tangental velocity drops to zero at the surface.
Heat transport is accomplished by molecular diffusion between the stagnant, adjacent layer out to the outer layers, which transport the heat as they have mean flow velocity.
.
With turbulent flow, the mass overcomes viscosity within the air and the flow becomes chaotic. Tangential streams which attempt to form along the surface are broken as high and low pressure / velocity fluctuations prompt rapid exchange of this boundary air. Thus turbulent flow has a "scrubbing" action that it breaks up the boundary and replaces it with fresh air. This is a much more efficient action than simply allowing the air to flow by.
.
I was surprised to read about a recent "discovery" in which a micro machined block transported water through a plurality of small channels which discharged it against a heat sinking surface only to return the water back through a great number of adjacent channels. It seems the contribution of this invention was that it greatly reduced to heat exchanger size requirement. To me, this nothing more than what old power supply designers have been doing for decades by placing their critical heat spreaders close to the fan.

 

[minger]

If there is a test pipe with no heat lost, you can treat stagnation P = static P.

If the flow is not moving then you can treat P_0 = P otherwise they can and will be different.