Intercooler theory 101

[red900ss]

Sydneykid et al may be able to answer this...

When measuring the pressure drop accross an intercooler how do you determine the drop from restriction vs. drop from thermal efficiency? i.e. as the air is cooled there is a reduction in volume -> less pressure.

Is it just a given that on the bench there is only a minor heat loss from radiation so it doesn't factor in?

Just wondering....

[Doctor]

So what you are saying is that there is a potential for psi and volume to decrease with the implementation of a better cooling core? Given all things are equal. But do you think that a few degrees could in fact affect volume and pressure?

[red900ss]

Density is a product of temperature and pressure. The whole idea of an intercooler is to remove heat from the air, making it more dense.

My rough calculation is for every 1 degree drop in temp. there is a 0.1 psi decrease in pressure. If there's anyone out there who has done fluid dynamics recently (it was a while ago for me ) they may know better. In this thread Sydneykid had tested a 36 degree temp. drop across a cooler. This would result in a calculated 3.6 psi drop. Using 8 psi = 5% power then 3.6 psi would be a whole 4 kw (std RB25DET). Doesn't seem right. :uhh:

Can someone enlighten me as to the relationship between density and power???

[Roy]

Based on my memory of thermodynamics (its been about 3 years) the equation is something like T2/T1 = P2/P1 ^(1-r/r) with r being a gas constant based on density, which should be easy for air.

I dont even think this equation will work as i suspect you will need the flow rate, and the equivelant Reynolds number (heat exchangers require turbulent flow for max effeciency, a reynolds number exceeding 2000 is generally considered turbulent. Hence the construction of the baffles in the IC, also to maximise surface area)

I think pressure drop from temp differential is negligible when looking at the total pressure drop across the heat exchanger. we dont consider it when sizing them at work, and we deal with some pretty large units that use phase change, ie steam to condensation to maximise effeciency.

... youve now ruined my weekend, as im now curious myself, and ill have to work it out.

[rev210]

Interesting...

The colder air 'leaving' the intercooler is definately 'slower', kinetically, than the air traveling from turbo to intercooler. As such the shortest pipe needs to be the interccoler to plenum one. It is therfore at least being 'slowed' at an atomic level' as heat energy is taken out/lost from the system.

I am not sure if the same principle can be applied to intercoolers but, in an exhaust situation, thermal shielding adn trapping the heat engery in the pipe increases substancially the speed of the exhaust gas exiting the engine.

the paradox would be that hot air is faster to get from 'a' to 'b' and yet cold air is better for combustion.

At some point there is less benifit from a colder intake charge then there is from keeping the air speed up.

Don't know if this is right but, someone want to comment?

[red900ss]

I guess what I'm really saying is that to get an accurate idea of the efficiency of an intercooler you would have to measure pressure drop both on the bench (no cooling air flow) and installed (with cooling air flow).

For example:

You bench test a cooler and get a 3 psi drop. This drop would be due to constriction of airflow.

Then say, you test it "in situ"and get a 3 psi drop. This would indicate poor cooling.

This is an extreme example I know. But from it we can see that you would want as big a split as possible between bench and installed.

Any comments?

[red900ss]

rev210, I would have thought that getting the air as cold as possible, i.e. dense, would overide how quickly the air gets through the pipe (let pressure differential get it there).

As I understand it power is directly proportional to density of the charge. 1% more dense=1% more power.

[Warpspeed]

Interesting discussion.

I have always wondered about the pressure drop across the intake pipework, compared to the pressure drop across the exhaust. Ignoring the slight extra mass due to the addition of fuel, the mass flow should very nearly be the same.

But the exhaust temperature will be a lot higher though.

Does a foot of intake pipe have the same pressure drop as a foot of exhaust pipe, if the diameters are the same ? The same number of air molecules are going down the pipe per second.

The same thing must happen in an intercooler, even though the air shrinks as it gets cooler, the same number of air molecules must come out that go in. So the air will slow down, but surely the denser air will have more drag.

So maybe it ballances out, high velocity thin hot air will have the same drag (and pressure drop) as low velocity cool dense air ? ? ?

[red900ss]

hmm another can of worms.

Ignoring all other factors, if we consider that;

Total Pressure=Dynamic ("moving")+Static pressure (not "moving")

then I guess the air after the cooler must be going faster. How density affects this I don't know. I guess that's why the AFM is measuring mass as this is the constant.

[Roy]

Testing in situ will be difficult, as you would really want to know the mass flow rate of air passing thru the cooler as the cooling medium, and its upstream & downstream temperatures. I suppose you could log the ambient temp on the day and the speed of the car. Making sure the air hitting the face of the cooler passes thru the core instead of around would be another factor. It sure would be easier on a water to air cooler.

...and typically when talking about effeciency of heat exchangers you are talking about their ability to remove or add heat, so talking about pressure drops and effeciency as one may not really be correct. Pressure drop is one thing effeciency is another.

[meggala]

/meggala yawns you have all lost me

[Sydneykid]

Thanks guys, now my head really hurts. I'll try and cut to the chase, the only real way to test an intercooler is to put it on the car and check the temperature and pressure drop across it and the power it produces. A number of testers have tried bench methods without really getting a definitive answer.

A common attempt has been to put the intercooler on a flow bench (like you would a cylinder head). That doesn't really work because a flow bench creates a vacuum on one side (of the intercooler) and measures the resistance. Since a turbo compresses the air before it goes through the intercooler this test is not truely representative of how the intercooler behaves under pressure.

Intercooler heat transfer is very difficult to measure "in the lab" there are simply too many variables. I have seen one attempt to judge intercoolers by their weight, on the assumption that the more mass they have, the more heat they can soak. This of course ignores the heat transfer ability, or scrubbing, of the inlet charge as it flows though the intercooler. Not to mention the heat transfer ability from the core to the ambient air.

Without a simple and effective method of measuring the true worth of an intercooler I have resorted to experience and real world results. So we use a standard R33 GTST intercooler for outputs up to 200 rwkw (max), up to 300 rwkw we use a standard R33 GTR intercooler and over 300 rwkw we use a Trust tube and fin. Since most of our work is for circuit race or road rally (targa) conditions, we have the luxury of basically ignoring bar and plate intercoolers. This makes the choices much simpler.

Hope that adds to the discussion.

[red900ss]

Originally posted by Roy

.... Pressure drop is one thing effeciency is another.

I was wondering if this would be a way to measure the efficiency. i.e. comparing bench vs. installed pressure drop (...at a constant/given pressure on the "in" side of the cooler). Or is temperature an easier way?

I guess a BFO fan infront of the cooler on a dyno would be the way to go rather than on road.

[Roy]

... with the war in Iraq and escalating fuel prices, i cant afford petrol anymore, so in about two weeks ill be asking the best way to calculate the calories i burn as i skip/walk/run to work.

The answer is of course a lot.

As for the pressure drop, i can understand the concen if running OE turbo, if you want to be sure that your engine is seeing say 14 psi, just connect your boost controller to a line from the manifold, not the compressor housing. This will ensure your engine sees the desired boost level.

The 2psi pressure drop may result in more work for the turbo, but i dont think the effeciency will fall away that quickly.

Plus with the airflow the OE turbo puts out, i doubt you would even be getting more than max 1psi drop across a good quality intercooler. Daily variables such as humidity, temperature and latent heat in the engine bay may impact performance more.

With a data logger and some thermocouples and pressure transducers you could soon get an excellent idea of how you IC is working on the road and on the track. I need to win lotto so i can tinker in y garage.

[red900ss]

Deal Roy you win lotto and work on my car!

[kwazza11]

just alittle bit O/T but what bearing does this have on cooler piping sizes? most people tend to have smaller piping on the turbo to cooler piping and say this decreases lag. but using this theory wouldnt it increase lag as the larger volume in the larger cooler to plenum piping slow down the charge even more as it expands?

or is this done because the air expands within the cooler?

ie charge has less restriction through the cold side so flow is not interupted as much..

hope that makes sense.. i just woke up

but if anyone can enlighten me? i'm just about to install my cooler and am trying to decide what size piping i should use, i dont want a big increase in lag but dont want to restrict any power increases. (this is on an rb20det)

[Sydneykid]

Hi kwazza11, after having spent several hours with Frank Durney (ex Williams F1 Aerodynamicist) on a couple of occasions, I can confidently say "I know nothink" on this subject.

The way I look at is this, in a standard turbo Skyline GTST;

*the outlet from the turbo is small (around 45 mm)

*the adaptor pipe increases this diameter (to around 60 mm)

*the pipework from there to the intercooler is similar (63 mm)

*the outlet from the intercooler is the same (63 mm)

*the pipework increases in diameter (to around 75 mm)

*the throttle body has the largest diameter of them all (around 80 mm)

So I figure if Nissan does larger diameter the further you get from the turbo, then there must be logic in that. So I do the same. We use 63 mm from the turbo to the intercooler and 80 mm from the intercooler to the throttle body.

My "seat of the pants" comparison is that going 80 mm all the way, lowered the response of the engine to throttle inputs. As expected the dyno showed no real difference. So we stick to the Nissan "increasing diameter" theory.

I have never seen the reverse, ie; any decreases in diameter.

Hope that helps.

[Warpspeed]

The way I once had it explained to me by a Garrett engineer, many years ago, and it goes something like this:

The compressor wheel accelerates the air to high velocity, where the high speed air has high kinetic energy but little actual pressure. The purpose of the scroll is to smoothly decelerate the air, where the kinetic energy turns into static pressure and increased temperature. The air is braked (slowed down).

so the cross sectional area of the scroll smoothly increases towards the outlet. The ideal situation is to continue the expansion after the scroll until the desired velocity is achieved to feed the engine plenum, or intcooler.

So the turbo exit pipe can be fairly small maybe two inch to match the turbo outlet. This should then ideally increase in diameter to match the cross sectional area of the intercooler. Usually this is not practical for space reasons.

The best shape for accelerating or decelerating air is exponential, or shaped like a trumpet. Where the initial increase is gradual, and it then flares out at an ever increasing rate. In fact many well designed intercooler end tanks look exactly like a trumpet flare, but with a rectangular outlet to match the core shape.

The whole turbo to intercooler pipe should ideally look like a continuation of this, where the deceleration rate of the air is constant between the turbo and intercooler core. This will give minimum pressure drop, and minimum pipe volume (in theory at least).

At the intercooler outlet you want to again match the core flow area to the throttle body size. This might require another exponential pipe of decreasing area. Think in terms of velocity and flow area, and treat the air very gently (happy air).

What you do not want is rapid changes in air velocity through the system, (angry air) caused by step changes in pipe size or sudden sharp turns. Again probably imposible to achieve in practice, but worth a try if you have the space and inclination to do it.

[2rismo]

i know this might not answer any of the above questions, (maybe it will!?!?) but DAMN i found this a useful resource for understanding the differences and applications of intercoolers etc etc etc...

Also gave me a better understanding of end tank design and core finishing.

http://www.are.com.au/index.htm

[meggala]

funny this I actually pulled my front mount off my r32.

shock horror I hear you all say.

the proof is the car lost no powerand became a little more responsive.

no I'm not going insane the cooler is going on the cefiro and that is where the mods will occur. the r32 the driven around on .8 all teh time as my wife drives it more than me and she does not need the cooler.

I have looked at a number of light turne cars out of japan mainly drifters running extra boost may be a smal turbo up grade and the one thing they have in comon the stock cooler.

yes there might be pressure drop yes it not perfect for 40'o day at the track but for every day use its fine. the cefiro wil be coping the power up grades so the swap has occured If I told you how quick it was you wouldnt believe me I have dr drift to thank for it as well as he did most of the work and dropped the socket down the cooler pipe which lead to all this happening in the first place

now to sort the cooler out for the cefiro

shouldnt be to hard and d yes it will be painted black

meggala