Liquid cooling

[Aesquire]

Perhaps the relevant part is that cooling drag is important to a surprising degree, as you make induced and parasite drag less, and speed increases.

And WW2 was where the power levels rose above 1000 hp. and speeds ran into compressability of the air that started to require a new paradigm of aerodynamic understanding. Cooling drag was nearly as important as the airfoils that were evolving to delay the supersonic shock wave, for real world speed.

That's why the Spitfire, with it's thin airfoil, was, with THE SAME ENGINE a faster plane than the Hurricane with it's thicker wing. You weren't getting near as close to the sonic limits with the earlier drag rise on the older, but still very good airfoil shape. And since it's a rare homebuilt aircraft that approaches the speed of sound, the lessons of the WW2 planes is still very important to grasp to improve, albeit incrementally, the performance we get out of a 10th the power.

 

[Aesquire]

The P-38 having the intercooler (radiator) INSIDE the wing, can be analogous to putting the radiator of an auto engined ( liquid cooled ) light plane tucked inside the fuselage. Imagine a RV-x with a V-6 or 8 in a tightly wrapped and low drag cowling, with the radiator, no ducting, behind the seats in the tail cone. Eventually the radiant & convective heat will transfer through the fuselage to hit equilibrium, but the total heat you can reject would, I bet, be far less than a few hundred horsepower V-x puts out. Power would be temperature limited.

We are used to designing to the engine manufacturer's RPM limits and using the prop choice to absorb that power to get the speeds that match airframe total drag. ( at mission cruise speed for example ) Then we modify if needed the cooling of that engine to keep it from overheating and reduce drag and are happy when the curves get pretty darn close... With a margin for anticipated operating conditions... because perfection can only happen at a single combination of speed, air temperature, density, and blessing of the lift daemons.

 

[Aesquire]

In the typical air-cooled engine in an HBA, cooling is determined by the airframe / FWF designer who typically is trying both to minimize drag and to adequately cool the engine. Mature designs, properly implemented, operate without “hot spots”.

Or, simply, "What he said".

 

[thjakits]

Hi all,

I learned most about liquid cooling for airplanes from this site:

Aircraft Rotary Engine News Letter

Paul Lamar was a true pioneer, not just for rotary engines but also for aerodynamics and cooling!

If you don't find what you need onthe above site, check this:

Forums - Rotary Engine Technology Forum

Rotary Engine Technology Forum

forum.rotaryeng.net

Most all Paul Lamar's stuff will be there somewhere![
[Paul passed away in 2017]


A few notes I learned over the years:
- Mustang duct was NEARLY perfect - one more improvement wasn't incorporated anymore...
- For most effective cooling you need to find your worst cooling case and start there.
- Next, you have a few variable to play with: airspeed, cooling intake location, outlet location, intake size, cooler area...
- In any case, you want the cooler to be rather thin and big inflow area, than smaller and thicker - in the usual prop powered speed ranges, a thick cooler will cost more cooling drag....
- intake and duct shape are extremly important to get right - you have great chance to get laminar flow all the way to the cooler!
- opening or divergence angles are VERY important
- slowing air down to the lowest possible cooler penetration speed is VERY important - duct needs to open like trombone ( I think that would be a parabel shape) - there various ways to shape the duct correctly if it can't be made optimal...
- You need a LOT of space behind the cooler where turbulent air can be dumped and then correctly accelerated out the outlet duct, which can create a slight thrust if so positioned....
- Besides Lamar's website he himself recommended to look at few pages of "Hoerner's aerodynamic drag" (BIG part is cooling!)
- You want to find out where on your airframe you have high pressure and low pressure areas!

I haven't visite the sites for way too long and would have to dig in just like you, so - I leave it to you!

Hoerner looks very intimitating - loads of text and formulas! Don't be afraid, start with Chapter index and just browse the diagrams and pictures - and things become clear shortly!

I have pdf copies of both Hoerner books - if you want me to share, pm me...

If you think of putting anything liquid cooled into a fast mover - Lancair, Glassair, Polen like, Midget Mustang or the II, NXT with a turboed V8, etc - your way to heaven is via Paul Lamar's site and the Hoerner books - verify it with CFD!!


Cheers,
thjakits

 

[Royal]

Thank you for all that. That's the confirmation I was looking for. Now I need to read it.

 

[pictsidhe]

Another book you may find useful for cooling ideas is "Road vehicle aerodynamic design". Race cars have space constraints too.

I haven't looked deeply at Paul Lamar's site and don't agree with everything I've seen there, but he references a lot of my better books...

If you get the diffuser wrong, you can easily get seperation and most of the air flowing through the middle of the radiator. This is bad for both cooling and drag.

 

[rv7charlie]

Induced drag goes down with speed; parasite drag goes up.
Researchers learned even more about diffusers (cooling inlet ducts) after WW2; there are more efficient diffuser shapes than what the P51 used.
Diffusers (cooling ducts):
Aerodynamics of Propulsion, by Kuchemann and Webber
(You might want to check with your state engineering school about borrowing a copy, rather than purchasing.)
https://www.amazon.com/Aerodynamics-Propulsion-Dietrich-Joanna-Kucheman/dp/B0000CIKRE
and
Compact Heat Exchangers, by Kays and London
A bit more affordable:
https://www.amazon.com/Compact-Heat-Exchangers-W-Kays/dp/1575240602

Diffuser design is *not* intuitive (at least it wasn't for me...). Getting it 'right', with max possible cooling from minimum drag is not easy; even the P51 could have been better.

There are numerous documents sourced from other authors on Lamar's site (you need to use the Wayback Machine to get to the site now) that are quite useful. He was a pretty talented inventer. But don't accept everything he wrote as gospel.

(Looks like pictsidhe types faster than me....)

Charlie

 

[skydawg]

Trying to minimize coolant drag coefficients has been a delima since the first car with a lot of designs. For example, large jet aircraft brute oils through the fuel tanks to not only cool the oil but also warm the fuel, creating minimal of any coolant drag. They also use fuselage skin to radiate heat. The Boeing 787 I fly uses multiple heat exchange elements to eliminate or reduce ducting protruding into the airstream.Jets have the advantage of a wide temperature differential between the inside and the outside skin temperature, normally about 40° below zero at cruise altitude.

My experimental Cessna 172, however, requires sufficient ducting to two radiators on its V8 engine. During test and development, we tried multiple ways to increase air flow through the two radiators, each mounted on different side of the engine, using prop wash only. However, we couldn't get enough air flow from the inner roots of the prop, and actually redesign the prop to have more airfoil towards the center to increase the air flow to the ducting opening that was about 12 in from the prop center. We tried different size duct openings and shapes, but each further increase drag at cruise speeds where the duct size did not have to be as big due to the airspeed. After about 3 months of testing, we figured out the smallest duck sizes and radiator size we could use with addition of a electric radiator fan that was mostly used during taxi and slow air speed ops. This allowed us to use the smallest size radiators, saving weight of the aluminum radiator as well as the extra coolant, which considerably decreased our drag coefficients at cruise. The fan draws only 6 amps and is mostly running on the ground or during climb and warmer ambient temperatures. The fan is controlled by the ECM, which turns it on and shuts it off as needed to maintain temp

There's a lot more to consider as we discovered, such as single vs dual pass radiators, pressure cap psi, air dwell time, maintaining pressure differentials each side of the radiator, ECT.. but there is always room for improvement and you won't know if it works until testing.

What you propose is similar to conventional cow flaps, where they are open during taxi and climb where extra cooling is needed, decreasing pressure behind the cooling fins that allows greater air flow. Your proposed design of a ducted door seems it would do the same function of decreasing drag when extra cooling is not needed. There are temperature switches you could use to automate the process, that could drive a servo to open and close the door even proportionally based on engine temperature.

 

[thjakits]

Hi all,
...there were quite a few rotary pilots on PL's newsletter.

RV7Charlie, aren't you one of those pioneers??

I seem to remember a weeks long discussion about cooler configuration - Tracy would insist and try a fairly small but thick cooler, PL promoting thing but a big area. Eventually Tracy would also mount a big thin one and find out that it worked better....

PL did also aerodynamics on early race cars, I think he was the essential guy on the Chaparral 2F??

I think he made it clear that you need to calculate cooling ducts precisely - especially opening angles - and the trombone before the cooler. Also sealing the duct to the cooler to avoid air passing by...
[Personally I would try to calculate the ducts, but reduce the opening angles 1° on each surface just to have a little cushion...]

Not sure outright, but IIRC - the P-51 didn't have that trombone in front of the cooler - that was the last improvement not put into the assembly line - "good enough as it is!"...

Skydawg - do you have your Experimental-172 in this forum??


Cheers,

thjakits

 

[Royal]

Back in the day when I was turboing Hondas....lol we had all sorts of intercoolers between our friends. People using the 300zx intercoolers always had heat soak and high intake temps. It was basically a box. I bought an intercooler from a guy that cut down diesel truck intercoolers and welded on end tanks. My buddy used mine first which was 5" high and he was having some issues with the boost he was running. He then used a 6" high one and put down 230hp to the wheels. ITs definitely about getting the right speed of air through the cooler and the longer the better allowing new air to continually hit the fins. On my prelude I have now I sealed the radiator and condenser with foam tape to the support and fan shroud. Cools very well now.

 

[rv7charlie]

I certainly wouldn't call myself a pioneer. I'm trying to be part of the 2nd (or 3rd) wave of rotary adopters, learning from Tracy, Ed Anderson, and several others who really were pioneers.

Paul Lamar is no longer here to defend himself, but I'm afraid that you've only repeated Paul's version of the thick/thin discussion. I won't try to speak for Tracy, but I can tell you that his Renesis powered RV4 still has the GM heater cores that are about 3 1/2" thick. I can also tell you that his 20B powered RV8 has a radiator core that, while it has a lot more face area than the pair of heater cores, it is the thickest available 'Scirocco' style racing radiator, at ~2 1/2" thick. You'd need to ask Tracy, but I suspect that the choice had more to do with getting enough core volume for the 20B and fitting it within the RV8 cowl, than with accepting Lamar's theory.

I'm dealing with the same issues fitting the rad into my Renesis RV7 cowl. I chose an automotive racing radiator as well. But mainly because I didn't want to hack up the stock cowl to remove the 'nostrils', and fit a chin scoop for a streamline diffuser feeding the heat exchangers. The wedge diffuser feeding it isn't optimal, but 'everything's a compromise', and it would have been impossible to fit proper streamline diffusers to heater cores in the stock inlets. Even the Scirocco style rad is a compromise; it fits both inlet and outlet on the same end (ease of packaging/running coolant lines), at the expense of thermo efficiency. It's a 2-pass, top>bottom flow, which means the thermo efficiency is lower on the bottom half of the rad. But less expensive to build than back>front flow.

Here is my understanding of the issue; I could be wrong. Consider some data points from success stories in other areas. For decades, aircraft oil coolers have been from 3 to 4 inches thick. NASCAR and high HP open-wheel racers use radiators that are about 4" thick. Production cars, on the other hand, do use very thin radiators; often only 1" thick. The difference is in application. Aircraft and racers are rarely idling, and have virtually no load on the engine when they are (prop load goes up exponentially with rpm). Cars, on the other hand, spend a *lot* of time idling, and running around town at relatively low speeds, but they have significant loads even when idling due to air conditioning, supplying the current for the cooling fan, etc, and typically have torturous air flow paths to & from the radiator core.

While a 1" thick core might make the most efficient heat exchanger from a thermodynamic standpoint, fitting one with enough volume to cool 150-300 HP into an aircraft is a tall (and wide) order. Then there's the issue of *aerodynamic* efficiency. The faster the air (and water) move through the heat exchanger, the higher the thermo efficiency, because delta-T will increase. But...everything's a compromise. At some point, the energy required to force the air (and water) through the core will begin to negate the efficiency gains. There is parasite drag in a radiator core, just like on an airframe. Hanging the core out in the airstream works for a Pietenpol; not so much for an RVx or Lancair. And even if you could find a place within the airframe to fit that 1" core, how do you get the air to it, and ensure equal flow across the entire core (otherwise the core volume without airflow is wasted), without converting the dynamic energy of the air to drag, instead of cooling? I'm not smart enough to figure out the optimum thickness vs face area vs packaging, but a lot of a/c designers and race car designers who are a lot smarter than me, seem to think that 3-4" thick, with minimum required face area/volume for the HP, is a pretty good place to be when going over a hundred mph.

Again, I could be wrong...

Charlie

 

[Dan Thomas]

I adapted an RAF Subaru 2.2L to a Glastar, and used the radiator from the car the engine came from. I mounted the rad behind the engine with the top of the rad against the firewall and the bottom about six or seven or eight inches (been a long time since, airplane long gone) forward of the firewall, and closed off the sides and sealed the opening at the bottom against the cowl so that all the air that entered the cowl through the air inlets (standard Glastar cowl) had to exit through that thin but large radiator. It worked just fine even at full power in extended climbs and even though the cooling air had already absorbed heat from around the engine itself and the exhaust mufflers before it went through the rad. A shallow, slanted deflector at the front of the outlet generated a bit of low pressure in the slipstream to enhance the pressure differential. The top of the rad was lower than the top of the engine, so a vapor release/filler point had to be created in the line above the engine.

 

[rv7charlie]

No doubt it works to adequately cool the engine; your technique has been used many times in the past with other engines. IIRC, BeltedAirPower's V-6 had a similar configuration. The question is, how much of a cooling drag penalty is there in doing it that way vs using an optimized inlet diffuser and exit duct, and is that penalty acceptable for that airframe and mission.

Generally, the slower the airframe, the more you can get away with without seriously impacting performance. I realize that a Glastar isn't slow, but it's not an RV or Lancair, either. Most (but not all) RVs with water cooled alt engines tend to be slower at the same fuel burn than their Lyc powered siblings. Since extra weight has very little effect on cruise speed, the speed loss can be almost totally attributed to cooling drag increases.

When optimized, liquid cooling should be more aerodynamically efficient than air cooling. But it's pretty difficult to do it optimally in an airframe that was designed for air cooling.

Charlie

 

[pictsidhe]

The optimum radiator thickness is the thinnest one that will fit and not cause flow problems. Ideally, one of infinite area and infinitely thin. Obviously, while that would have an infinitely small pressure loss, it would be infinitely hard to do...
You need to look at the complete system. Available space, drag and effort to do all come in to the equation. Lowest drag of the complete system can mean both a suboptimal radiator and diffuser. A perfect diffuser feeding a small frontal area and very thick core is not optimum. Neither is a big area thin core and a dreadful diffuser, though this seems to be favoured for cars... In the middle are probably several options.
I'm currently working on improving the intercooling for my car. The volume available for a diffuser is smaller than the intercooler... But I can still make a big improvement. Compromises are often forced on us.

 

[Dan Thomas]

Generally, the slower the airframe, the more you can get away with without seriously impacting performance. I realize that a Glastar isn't slow, but it's not an RV or Lancair, either. Most (but not all) RVs with water cooled alt engines tend to be slower at the same fuel burn than their Lyc powered siblings. Since extra weight has very little effect on cruise speed, the speed loss can be almost totally attributed to cooling drag increases.

A 130-hp Subaru redlines at 5600 RPM. A 125-hp Lyc redlines at 2700 RPM. The stated cruise speed for the O-235-powered Glastar, 125 HP, at 2500 RPM was 130 MPH IIRC. To get the engine and prop combination to the same power level, I had to cruise that Soob at 5200 RPM, where I saw a TAS of 132 or so. Cooling drag didn't seem to be too bad, no worse than the stock Lyc setup.

But at 5200 the fuel burn and noise was pretty bad, so I usually cruised at 4600, where I saw about 110 MPH. One didn't dare lean it too much or you'd burn those tiny valves real quick. Did that on a max-performance takeoff once. It had EI but not EFI. Had a Holley carb with a manual mixture control modification.

 

[rv7charlie]

Lots of variables there, but the ultimate test for power required is fuel flow at the same airspeed on the same airframe, or virtually identical airframe.

 

[wrmiles]

I have never designed a liquid engine cooling system but have considerable experience on turboprop oil cooling systems.

The problem is that I had considerable constraints on space available for the cooler and ducting and management direction on type, size and location of the scoop, resulting such things as scrunched up diffuser and ducting and use of a NACA scoop when an external scoop would have been better. It didn't help that worst case was usually ground idle with minimal airflow velocity.

 

[Royal]

I snagged a transmission cooler from a dodge dart today. Its long and has very thin channels. Could use that for the radiator or an oil cooler as well if the heat transfer was better on that. I honestly would never use a Subaru engine in a plane. 130 hp could be had from a few other engines that are much more efficient and smoother. They do look the part of an airplane engine though.

 

[trimtab]

Lots of people use very thin oil coolers for their radiators. The oil cooler is operating at essentially zero pressure generally at the tail end before the sump. Cooling radiators generally operate at pressure, and you'll need a way to accommodate the issue. Water boils pretty cool up at high altitude, for one. You can use brines and other materials as the transfer medium to address some of the issues, or even fluorochemicals if you have deep pockets (as they do in certain gummint installations). So the bottom line is you limit yourself by using a radiator that can't take a few atmospheres (even though you will only really need a single atmosphere).

 

[thjakits]

I wonder from where you get that idea?
Oil coolers ".... at the tailend before the sump at zero pressure"?? How do you suggest the oil gets collected to go into the cooler - before going into the sump??

As far as I know - oil coolers are thermostat controlled right after the pump - at seriously high pressure - before the oil goes into the engine for lubrication duty. Literally in ALL applications I am aware of...

Also with the water boiling at altitude? Usually a cooling circuit in any car today runs under pressure, sos the water never boils below about 120°C - many a engine runs best today above 95°C cooling temp....
For higher altitude - one will have to change the pressure valve for a higher differential pressure and one should have a sealevel max, pressure valve too, but I doubt that you will have problems with "boiling water"....
I also doubt that specific liquidcooled aviation engines are cooled with much exotic coolants - loads of anti-freeze and anti-corrosion additives, but brine?? Never heard of that....