Cooling system drag/radiator efficiency

[M_a_t_t]

I've been reading Road Vehicle Aerodynamics by A.J. Scibor-Rylski (1st edition) on my lunch break at work recently. Today I read 'The idealised air cooling system' pg. 60-65. I have a few things I have been thinking about regarding this section but since it's 4am here I'm just gonna start with the simple question and get the rest out later.

This question might appear to be directed at JulianEdgar, but I don't intend it to. The more people looking at it the better.

He has suggested to measure cooling efficiency by taking pressure readings on both sides of the radiator (intercooler, etc) and comparing it to the pre-modified readings. If the difference is increased then the system is working better (based on memory, couldn't find the section in his book to cite it). The greater the pressure difference the higher the flow (and therefore speed for a given inlet opening). However, in Road vehicle aerodynamics it says you want slower flow for best heat dissipation. Which makes sense to me because of the longer time the air can absorb the heat from the radiator. So does the pressure measurement just simplify the required testing and provide adequate data? Or does this testing method need some revision? Or maybe I just need to read the section again.

I feel like I am missing something simple.

[me and my metro]

Other things to consider are cooling fans and coolant flow through the radiator. If you are strictly concerned with aero realize factory grill openings are designed for 120 degree air and stop and go traffic.

[Vman455]

I'm looking at the 2nd edition, so hopefully this material is in yours too--

I don't think you're missing anything. There are two competing aims here: first, to improve the radiator's capacity for dissipating heat and second, to minimize drag of the inlet/radiator/outlet. The dissipating capacity of the radiator is given by

Q = K1 * A(Tw - Ta) * V2^(4/5)

According to that, it isn't simply a case of slower air into the radiator improving its ability to cool; the dissipating capacity goes up with speed.

Greater pressure difference across the radiator should improve its cooling (because you're moving more air through it), but to get that you either need to slow down the air entering the radiator (losing momentum --> drag) or speed up the air exiting the radiator (accelerating --> drag). This is summed up in the equation S-R gives for the drag associated with the duct:

Dint = (Vi - Ve)(dm/dt) + (Pi - Pe)Ae

where the first term represents the momentum loss through the duct (change in velocity times mass flow rate) and the second term the "work required to overcome friction and internal obstructions" (change in pressure times outlet area)--for example, a radiator in the duct. The word "work" here is a bit misleading; pressure times area gives force, so this is the force required to do the work (energy) to overcome friction and obstructions in the duct.

Then you have to throw in the fact that all these equations are derived from considering an ideal duct, which is very much not the case on our cars (unless you've fully ducted the inlet and outlet and avoided radical direction changes!).

So, short answer: yes, measuring the difference in pressure across the radiator will be your best measure of its effectiveness (as JulianEdgar has shown in Autospeed articles), but if you're looking to optimize both cooling and drag from the radiator, you'll want to measure overall drag as well to see if there are any measurable changes using throttle-stop testing or another method.

My two cents, anyway, and hopefully my summary of S-R is correct.

[JulianEdgar]

So you're asking about two things, pressure difference across the radiator, and speed of airflow.

First, regarding pressure difference. The radiator core is a resistance to flow, and if there was no pressure difference across the core, there would be no airflow. And, the higher the pressure difference, the higher the flow. It is just the same as an electrical circuit with a resistor. If there is no voltage difference across the resistor, there will be no current flow, and the higher the voltage difference, the greater the current flow.

Second, regarding airflow speed. If we think of the airflow approaching the car, the more we can slow it, the higher the pressure. So the highest pressure recorded on a car is the stagnation pressure, where the airflow has been brought to a halt. In other words, flow speed is traded for pressure, or pressure is traded for flow speed. So if we can slow the speed of the airflow to zero (or nearly zero) in front of the radiator, we will have the highest pressure (good) and the lowest speed (good).

Now, as Scibor-Rylski explains, let's look at an idealised cooling system. We let the air in at high speed and then progressively slow it by using a diverging (ie increasing cross-sectional area) duct. Pressure rises as speed falls, until at the core, speed is low and pressure high. There is a pressure drop through the core, but now we have heat energy added to the air. We then use a converging duct to increase airflow speed and lose pressure, until at the exit duct we have the airflow speed matching (or even exceeding) the free stream speed. Then we smoothly add this exit air to the airflow around the car, thus creating little drag - or maybe, as was done in some WWII aircraft, even a little thrust.

https://www.youtube.com/watch?v=P7Wh1XoystM

https://www.youtube.com/watch?v=tfw17CVb4M0

[aerohead]

Originally Posted by M_a_t_t

I've been reading Road Vehicle Aerodynamics by A.J. Scibor-Rylski (1st edition) on my lunch break at work recently. Today I read 'The idealised air cooling system' pg. 60-65. I have a few things I have been thinking about regarding this section but since it's 4am here I'm just gonna start with the simple question and get the rest out later.

This question might appear to be directed at JulianEdgar, but I don't intend it to. The more people looking at it the better.

He has suggested to measure cooling efficiency by taking pressure readings on both sides of the radiator (intercooler, etc) and comparing it to the pre-modified readings. If the difference is increased then the system is working better (based on memory, couldn't find the section in his book to cite it). The greater the pressure difference the higher the flow (and therefore speed for a given inlet opening). However, in Road vehicle aerodynamics it says you want slower flow for best heat dissipation. Which makes sense to me because of the longer time the air can absorb the heat from the radiator. So does the pressure measurement just simplify the required testing and provide adequate data? Or does this testing method need some revision? Or maybe I just need to read the section again.

I feel like I am missing something simple.

Research conducted by FIAT in the mid-80s found that altering upper-body flow, could directly affect underbody flow, and that the entire vehicle required consideration.
Dr. Albert Morelli and Pininfarina conducted exhaustive cooling research with their famous CNR project of 1976-1978, and they also were measuring the 'whole-car' effects.
I don't consider this sort of thing 'simple'. It would be easy to drop $100,000 sorting it out in a wind tunnel, under precisely monitored conditions.

[JulianEdgar]

Originally Posted by aerohead

.
I don't consider this sort of thing 'simple'. It would be easy to drop $100,000 sorting it out in a wind tunnel, under precisely monitored conditions.

That's being unduly defeatist. I've achieved excellent on-road results in improving heat exchanger efficiency (in that case, an underbonnet intercooler) with simple on-road testing and modification.

[aerohead]

Originally Posted by JulianEdgar

That's being unduly defeatist. I've achieved excellent on-road results in improving heat exchanger efficiency (in that case, an underbonnet intercooler) with simple on-road testing and modification.

The context has to do with a certainty of the quanta being unimpeachable under strict SAE testing protocols, and all data reduced to standard SAE atmosphere.
Any drift in atmospheric pressure during the course of the exercise, for instance, would be enough to throw a caution flag on the play. And call into question, the accuracy of the observed numbers.
We're talking about small fractions of total drag, and some 'signals' could be lost in the 'background noise'.

[JulianEdgar]

Originally Posted by aerohead

The context has to do with a certainty of the quanta being unimpeachable under strict SAE testing protocols, and all data reduced to standard SAE atmosphere.
Any drift in atmospheric pressure during the course of the exercise, for instance, would be enough to throw a caution flag on the play. And call into question, the accuracy of the observed numbers.
We're talking about small fractions of total drag, and some 'signals' could be lost in the 'background noise'.

What you are talking about seems to have no relevance whatsoever to either the original poster's question nor the measured gains I made in intercooler efficiency by making aerodynamic changes.

A good example of someone talking from the experience of having actually improved heat exchanger efficiency versus someone just theorising.

[aerohead]

Originally Posted by JulianEdgar

What you are talking about seems to have no relevance whatsoever to either the original poster's question nor the measured gains I made in intercooler efficiency by making aerodynamic changes.

A good example of someone talking from the experience of having actually improved heat exchanger efficiency versus someone just theorising.

Atmospheric pressure has a natural variability which can skew data, unless all, before and after ( A-B ) measurements were captured in relatively rapid succession.
Altering upper body drag can alter underbody drag. The entire vehicles must be measured.
An apparent advantage at the intercooler could have introduced an effect somewhere else. The whole car needs to evaluated as an 'aerodynamic singularity.' Exactly what FIAT and Pininfarina did.

[JulianEdgar]

Originally Posted by aerohead

Atmospheric pressure has a natural variability which can skew data, unless all, before and after ( A-B ) measurements were captured in relatively rapid succession.
Altering upper body drag can alter underbody drag. The entire vehicles must be measured.
An apparent advantage at the intercooler could have introduced an effect somewhere else. The whole car needs to evaluated as an 'aerodynamic singularity.' Exactly what FIAT and Pininfarina did.

You seem to have gone onto a different topic. The OP asked about cooling efficiency, and measuring aero pressures is a simple and highly effective way of improving cooling efficiency through aerodynamic modification.