Questions part 2: Aerodynamics and cooling systems. (1 Viewer)

[AspiringHowardHughes]

It was suggested (by tomo pauk) that I create a separate post to discuss cooling systems.
Good idea.
Here it is.

Same format as my previous post, aerodynamic questions have been lumped in here as well owing to the common ground between a good cooling system and a fast plane.
For the sake of readability, I again ask that respondents number the question they are responding to as I have done while asking them.

1: Is there any reason it would not be feasible to cool all systems with one radiator? With the aim of reducing parasitic drag and preserving a clean flow of air around the plane, I have been tempted in my designing to use one large radiator, with air-water charge coolers and oil-water oil coolers. Am I correct in believing this would make for a faster (if not more internally complicated) aircraft? Is there precedent for oil coolers not being direct-to-air?

2: To what degree is the Meredith effect compromised when the radiator is positioned directly in the wake of the propeller (as in an annular installation)?

3: Does a leading-edge radiator (as on the mosquito) meaningfully avoid ingesting turbulent air? To what degree is produced thrust and airfoil performance compromised by the ejection of radiator air at the necessary angle?

4: To what degree does the slope of the forward glass influence speed? I have read that to be a reason for the speed of the FW 190, but I have also seen late-war prototype fighter sketches with the same slight slope as before the war.

5: This question is only peripherally related to aerodynamics. What is the difference in lost efficiency between a 90 degree bend in the air intake pipe and a 90 degree bevel gearset in the drive of said supercharger? Obviously both should be avoided if possible but I only ever read about the former being undesirable.

6: Is there any merit behind variably-sized air inlets? My instinct would be that the engine should consume air at the same rate the movement of the plane carries it into the air intake, to avoid turbulence, drag, or choking the engine.

7: The scale of the ductwork on turbocharged aircraft is usually justified by the need to cool the exhaust before it encounters the turbine blades. What is the temperature drop this actually creates? (In the P-47, for example)

8: Some German engine designs position a first supercharger stage co-axially with the propeller shaft, ingesting around a small propeller spinner. Does this installation have merit over the typical German elbow intake? The ram-air effect may be improved, but the propeller interfering with airflow could be a problem.

9: In hotter areas, was any attention given to actively cooling the pilot?

10: What was the most effective coolant composition available at the time of the second world war? What progress has been made since then?

Explanatory images available upon request.

 

[MIflyer]

Is there any reason it would not be feasible to cool all systems with one radiator?

Not all systems needing cooling (engine cooling, Supercharger Inter/aftercooler, lubricating oil, environmental control systems) require the same the same temperature be maintained. For example you want the engine cooling to be rather hotter than the oil cooling; to keep the oil from breaking down it should be no higher than maybe 220F but not cooler than about 180F to make sure the water is boiled off.

In some cases additional stages of cooling can be kicked in to take care of new demands, although I don't know how much that was done in WW2. For example the F-111 had it, which proved to be unfortunate on at least one occasion I am aware of.

I do not think that in WW2 that anyone ever gave any consideration to cooling the pilot except via ram air. Sometimes this was effective and other times it was disastrous. On the Gee Bee Z, for example, they put a large vent right in front of the pilot in the instrument panel but unfortunately located the intake for it right behind the exhaust.

The first airplane I know of that had some cooling for the pilot was the P-80A, which used an air cycle refrigeration system. Aside from being thoughtful the speed of the airplane probably made using a ram air cooling vent hard to do and cracking the canopy in flight probably was an invitation to disaster.

 

[Shortround6]

1. using separate systems allowed for better control of temperatures.
Most systems were 100% manual, that is under pilot control. Using a temperature control to adjust the flap opening was just starting to be used at the end of the war.
Worst condition for cooling was full power climb, used the most power with low forward speed so low airflow mass through the cowling and radiators/coolers. At times every cooling flap was wide open. Many planes could fly at top speed with the cooling flaps closed or nearly so.
Having to use both adjustable intakes and exits is a sign that somebody screwed up. The cooling systems work on a pressure deferential, closing the exit flap creates higher pressure through the system all the way to the intake. High pressure at the intake means less air enters than if there is lower pressure at the intake or across the opening. The excess air diverts around the opening.
If you can find any, look at photos from a wind tunnel using smoke around a radial engine. Air that is around 1/2 way out from the center is pushed out and around cowling even before it gets to prop. Turns out that spinners on most radial engines didn't actually do much of anything. The high pressure air build up in front of the engine and front of cowling was diverting air before the air got to the spinner. They sometimes added a few mph but most of the time change was not worth the extra maintenance problems.

28 cylinders over 3000 hp, no spinner most of the time. at a certain point aerodynamics cannot be done by eye.

 

[Shortround6]

2. It took quite a while for the Meredith effect to get close to working. It requires a very close balance between inlet drag, drag though the cooling medium (cylinders and baffles or radiator/s) expansion of the cooling air and compression in the exit duct to accelerate the air up to near (or over) airspeed.
The basic theory was simple and brilliant. Executing it took a lot of work or rather executing it well took a lot of work.

Problems
1. too much drag in the inlet, as in too small a change in the area of the duct or too steep turns in the duct create drag or eddies or turbulent flow. You may still get good cooling of the engine, you don't get the hoped for drag reduction.
2. Too much drag going through the fins of the engine or the radiator core. Since drag is proportional to the square of the speed, slowing the airflow down really helps reduce the drag as the air cools the hot surface/s. The slower speed also aids heat transfer, Air going through a radiator at 450fps doesn't pick up a lot of heat.
3. Using the taper of the exit duct to compress and speed up the air before it exits. Too short a duct created turbulence or high pressure areas which interfere with flow. The cross sectional area has to flow smoothly from large to small. Trying to have the flow change direction much makes things a lot harder, air flow in the inside of a curve is going to be slower than the airflow on the outside of the curve leading to turbulence and turbulence is the enemy.

Now try to get it work at different speeds and altitudes (air densities). The difference between 5,000ft and 22,000ft at the same speed is that at 5,000ft you have twice the weight (mass) of air flowing the cooling system at the same speed.
The Meredith effect was not all or nothing but there was a considerable difference between done well and not done well.

 

[don4331]

1. At least some models of the Jumo 213 in Fw190D used oil 2 water (well, not water but water/ethylene glycol mix). It actually a very good idea, liquid 2 liquid cooling is much more compact*, and you don't need to worry about "coring" in the oil cooler.

Coring is when the oil in outer portion of the cooler congeals, and the oil only flows through the inner (core) portion...which winds up being insufficient. You don't have the issue with oil 2 water as it is warmer.​

*The water radiator winds up being larger, but you only have one radiator, one set of cooling lines.​

​

It doesn't work as well for charge cooling, as you want always want maximum temperature delta in the Intercooler.

2. As noted just getting Meredith effect to work was challenging enough.

3. The lower surface of the wing is already a high pressure area, so not impact nearly as adversely by turbulent air. Not increasing the frontal area when using leading-edge radiators probably has greater effect. Fastest Tempests were those with leading-edge radiators.

4. How important is seeing what is right in front of the airplane? FAA Skua traded speed, for ease at landing on small carrier.

5. Bevel gear set is >95% efficient (often 98+%); 90* bend in a pipe depends on the flow rate - at idle it won't make much difference; at full throttle a lot.

6. Violates the KISS principle: the complexity of a variable intake vs improved performance makes it of limited benefit.

7. IMHO the piping is more about balance/allowing for fitting of intercoolers than cool the incoming exhaust. Wasn't the XP-39 turbocharger right against the Allison?

8. Putting the 1st stage supercharger co-axial with the propeller shaft allowed taking the power off the front of engine which was designed for the load. As opposed to having to beef up the existing supercharger drive.

9. Already answered better than I can

10. 50/50 mix of ethylene glycol and water was the most effective coolant in WWII; Propylene glycol has some better/some worse properties at the temperatures used by liquid cooled engine. Some of the coolants in the last 10 years are worth investigating.

 

[AspiringHowardHughes]

Thank you for the responses, Don.
A couple follow-up questions, of course.

1: What planes used air-to-air charge coolers? From what I have seen, air-to-water is more common just because of the compactness benefits. (and because a lot of these systems were not initially present in the plane)

4: I should clarify that I mean for a fixed height of cockpit. The penalty that I can think of is increased weight of glass, maybe? It could be thinner for the same protection value, however.

7: I thought the temperature of gasoline exhaust was too high to allow for such a close mounting, with the technology of the time. If it really is right against the engine, what differs compared to the P-47 installation to allow for this?

8: Does placing all the load on one end of the engine not run the risk of extreme crankshaft stress? I thought it would be best to put everything but the propeller on the rear, to try and equalize where the resistance is.

Another question adjacent to this concept:
In a two-stage supercharger system, which is better?
A: Fixed-speed first stage feeding into turbocharger second stage.
B: Turbocharger first stage feeding into fixed-speed second stage.
Assume nothing else needs consideration.

Will try and get around to another batch of questions tomorrow morning.

 

[don4331]

Thank you for the responses, Don.
A couple follow-up questions, of course.

1: What planes used air-to-air charge coolers? From what I have seen, air-to-water is more common just because of the compactness benefits. (and because a lot of these systems were not initially present in the plane)

4: I should clarify that I mean for a fixed height of cockpit. The penalty that I can think of is increased weight of glass, maybe? It could be thinner for the same protection value, however.

7: I thought the temperature of gasoline exhaust was too high to allow for such a close mounting, with the technology of the time. If it really is right against the engine, what differs compared to the P-47 installation to allow for this?

8: Does placing all the load on one end of the engine not run the risk of extreme crankshaft stress? I thought it would be best to put everything but the propeller on the rear, to try and equalize where the resistance is.

Another question adjacent to this concept:
In a two-stage supercharger system, which is better?
A: Fixed-speed first stage feeding into turbocharger second stage.
B: Turbocharger first stage feeding into fixed-speed second stage.
Assume nothing else needs consideration.

Will try and get around to another batch of questions tomorrow morning.

1. Most (all?) US 2 stage air cooled engines: B-17, B-24, B-29, P-47, F4F, F4U, etc, etc.

7. IMHO, the P-38 and P-47 installations were determined by packaging, not needing to cool exhaust.

There may have been (probably were) metallurgy improvements from '39 to '45...​

​

8. Yes, but by the altitude where you need the additional boost, the engine isn't making as much power.

It actually was a huge issue with the Allison C series issue: The original requirement was for 1k hp at sea level*; which required ~25 hp from the supercharger drive. Putting it in conjunction with the speed reducer saved cutting a gear and some other benefits.​

But when the USAAC asked for 1.1k hp at altitude, the consolidated drive off the front of the crankshaft became a huge issue - requiring redesign to the F series.​

11. All examples use B. A variable speed 1st stage (can be turbo or mechanical supercharger) feeding into a single speed second stage.

 

[AspiringHowardHughes]

Interesting. I suppose it makes perfect sense that designers already committed to air-cooling the engine would not want to add a liquid cooling system just for the intake.
Your answer to question seven goes against what I have read, but it is possible all I was exposed to was a speculative reasoning.

While all examples use option B (probably because mounting the second stage on the back of the engine is about as simple as it gets), would option A perform worse or differently?
If anything, I could see why one would want the variable-speed stage first, to account for the air pressure changing with altitude.

 

[Shortround6]

Interesting. I suppose it makes perfect sense that designers already committed to air-cooling the engine would not want to add a liquid cooling system just for the intake.
Your answer to question seven goes against what I have read, but it is possible all I was exposed to was a speculative reasoning.

While all examples use option B (probably because mounting the second stage on the back of the engine is about as simple as it gets), would option A perform worse or differently?
If anything, I could see why one would want the variable-speed stage first, to account for the air pressure changing with altitude.

Pretty much the goal of the GE turbo charger system used by all production US aircraft with turbos was to provide sea level pressure air to the carburetor inlet of the engine at no more than 100 degrees F. up to 25,000ft. This a pressure ratio of 2.7:1 This was not always achieved depending on the intercooler installation. Also as the turbo's themselves got better and they had better fuel more boost could be applied for more power but at higher temperature.
Another goal, also not quite achieved, was that the exhaust back pressure in the system was to be at sea level pressure. A lot of turbos were running in the high 30in range instead of 29.92in in service. This was supposed to help keep the base engines the same as the non turbo versions as far as exhaust components went.
Turbos did not provide quite the "free" power promised. They do for ground level engines but an aircraft engine running at 20,000ft has much lower back pressure than a sea level engine and for a 1000hp engine picks up about 80HP if the exhaust passages remained open and if we assume that somehow the amount (mass) of air and gas stayed the same. The turbo was getting a lot more than 80hp out the exhaust to power the turbo but it is not 100% free.
Engine designers and airframe makers also got better at designing exhaust systems to produce thrust from the exhaust gasses and much of the energy in the exhaust was either used in the turbocharger or directed in a direction that gave no propulsive effect.

My own opinion to the P-38 and P-47 was that the P-38 needed the 7 feet or so of exhaust duct to cool the exhaust before it hit the turbo. The turbo and pipe were exposed to the air and exposed turbine was hardly aerodynamic.

P-47 hid the exhaust pipes all the way back and also hid the turbo, They ducted cooling air to the turbo but you can't see the turbo from outside the plane.
Since the P-47 was the fastest plane (production) perhaps they took more care in the streaming?

 

[AspiringHowardHughes]

Good to know!
Seeing the turbo as a device not to boost but to "restore" desirable characteristics of the ingested air has been eye-opening.
I take it this means my gut feeling about why this two-stage order is preferred was correct....
Back to the drawing board.

 

[z42]

Good to know!
Seeing the turbo as a device not to boost but to "restore" desirable characteristics of the ingested air has been eye-opening.
I take it this means my gut feeling about why this two-stage order is preferred was correct....
Back to the drawing board.

It is common these days in turbocharged piston aircraft engines for general aviation usage. They call it 'turbo-normalized', where the point of the turbo is not to provide more power at sea level, but to allow the engine to produce (roughly) the sea-level performance regardless of altitude.

 

[z42]

1: Is there any reason it would not be feasible to cool all systems with one radiator? With the aim of reducing parasitic drag and preserving a clean flow of air around the plane, I have been tempted in my designing to use one large radiator, with air-water charge coolers and oil-water oil coolers. Am I correct in believing this would make for a faster (if not more internally complicated) aircraft? Is there precedent for oil coolers not being direct-to-air?

As D don4331 mentioned, some(?) Jumo engines used oil-water radiators for the oil coolers. I'm not sure if it was only the 213, or some versions of it, like D don4331 says, or did the Jumo 211 also feature it? I have a vague recollection it might have been a standard feature on the 211, but I might be easily wrong. It might be told in TSHPR.

P-51H would be another example.

2: To what degree is the Meredith effect compromised when the radiator is positioned directly in the wake of the propeller (as in an annular installation)?

I'm sure there is some effect, but I have no hard numbers to give. OTOH, I'd speculate the bigger issue is the limited space in the annular installation, and required turning of the airflow that is probably hard to get right without modern CFD. And of course, the similar to the wing leading edge style radiators, you have the airflow exiting at an angle. Some simple trigonometry might tell you how much of the thrust is wasted in the vertical component of the velocity vectors.

As a slight aside, an IMHO neat approach that to the best of my knowledge was not used in any WWII aircraft, is a chin radiator with a slanted radiator. I made a thread about that a few years ago at Chin radiator with horizontal radiator matrix

3: Does a leading-edge radiator (as on the mosquito) meaningfully avoid ingesting turbulent air? To what degree is produced thrust and airfoil performance compromised by the ejection of radiator air at the necessary angle?

For the first question, quite obviously yes? What are you thinking would be in the way producing turbulent air before it hits the intake? Of course, just designing the intake is a science all to itself. This article is about jet intakes, but to some extent the same issues apply to designing intakes for piston engine radiators. https://eprints.soton.ac.uk/46202/1/AIAA-26830-529.pdf

As for the exit angle, see above. But moreso than the angle, I suspect the limiting factor might be more the limited space available.

4: To what degree does the slope of the forward glass influence speed? I have read that to be a reason for the speed of the FW 190, but I have also seen late-war prototype fighter sketches with the same slight slope as before the war.

I don't think you can boil down the speed of the FW190 to a single issue like that, but yes, too steep cockpit windscreen was a mistake that many planes made, including the Spitfire. See e.g. https://www.wwiiaircraftperformance.org/mustang/Lednicer_Fighter_Aerodynamics.pdf

6: Is there any merit behind variably-sized air inlets? My instinct would be that the engine should consume air at the same rate the movement of the plane carries it into the air intake, to avoid turbulence, drag, or choking the engine.

There is merit, but another question is whether the added weight and complexity is worth it. And to that the answer seems to be no. As mentioned in the Sobester article I linked to above, even in modern supersonic jets the trend has been to move away from variable geometry inlets (of course, that's a development made possible by sophisticated CFD analysis). And essentially no subsonic jets have done it, AFAICS.

10: What was the most effective coolant composition available at the time of the second world war? What progress has been made since then?

70-30 water - ethylene glycol. Which AFAIK is what's still used. Water is a surprisingly effective coolant; adding ethylene glycol both depresses the freezing point as well as increases the boiling point, allowing to run the engine hotter and thus have a smaller and less draggy radiator. But the more glycol you add, the less effective heat transfer properties does the fluid have (water is better than glycol), so it's a compromise.

 

[AspiringHowardHughes]

As for your response to question 3, I suppose I should be more specific. I am aware that the boundary-layer air that cloaks the skin of aircraft in flight is seriously undesirable for ingestion into engines and cooling systems alike, as it is slow-moving and turbulent compared to the surrounding airstream. It makes sense that a leading-edge intake would avoid interacting with this layer, but how much benefit is actually due to the ingestion of "clean" air? Whenever I read on the radiator arrangement of the Mosquito, the virtue I see extolled is that the inlet is in a zone of high pressure, allowing for it to be reduced in size. I imagine that in an airplane that somehow lead with the radiator, there could be some benefit reaped from the clean air entering it - but this would likely have to be a pusher-propeller design. Given that the radiators of the Mosquito are still in the wake of the propellers, is the flow through them actually "clean"?

It is also nice to hear some solid answers on the topic of cockpit aerodynamics and coolant composition.
Thank you for your feedback.