Zero cooling drag experiment

[mikeje]

The big trouble spot for subaru or any liquid cooled aircraft engine is the cooling. I'm looking into the surface radiator idea and am designing an experiment to definitively answer the question for rv aircraft.

What I plan to do is build a wing section (I have a request posted in the classified section for "damaged" or donated wing parts to make up a wing: so far I have outboard leading edge skins and ribs for an rv-9 offered) and use spray tubes to spray hot coolant evenly on the top and bottom wing skins using evans npg coolant (boils at 375) in an open system. I'm going to mount this in the back of my pickup and use diverted coolant along with high accuracy temp sensors and a flow meter to determine the heat rejection capability at various speeds and angles of attack.

One key question is how much cooling is actually required, and will need volunteers to instrument their aircraft with a flow meter and temp sensors and collect the data to establish the actual cooling requirements. This will be relatively easy:
accurate input flow rate, input and output temps from the radiator in a steady state condition. (This data would be very valuable all on its own for designing standard cooling systems. Also, it would be interesting to see if ceramic coating the head and piston would reduce the cooling requirements with before and after measurements)

A few initial facts/ideas:

1. The heat transfer resistance is at the aluminum skin to air junction. All else is nearly negligible.
2. The heat transfer coefficient varies substantially with air velocity.
3. Radiant transfer can possibly be around 20% of total heat rejection. (near zero in a radiator: any radiant energy is reabsorbed)
4. The cooling ability of air goes down with altitude. If the air density is dropped in half, the heat coefficient for laminar flow (at 200 mph) drops by 30% as determined by this calculator:

http://www.efunda.com/formulae/heat_transfer/convection_forced/calc_lamflow_isothermalplate.cfm#calc

5. This system should be light weight: 8 lbs or so for the in-wing spray bars, plus pumps and additional coolant (4-5 gallons?)

6. Can the engine be safely operated in the 250-300 degree range? It seems that the obvious reason to keep it under 220 is localized boiling when using a water/glycol mix. If boiling is eliminated (using evans propylene glycol) why can't the engine run hotter, and therefore more efficiently and boost heat rejection capability? Any experience here?

This experiment is just the very first step to see if this will work from purely a heat load perspective. A lot of what I've read so far (mostly opinion) suggests that this is a stopping point, but I haven't found any experimental data one way or the other.

Based on my preliminary research, I think it might possibly work! Zero cooling drag anyone?



Mike Endersby

 

[rv6ejguy]

It would be an interesting experiment but just a lot of work in my opinion. Randy Crother's 7A has shown that cooling drag is probably no worse than a typical Lycoming installation and his installation is far from optimal.

The Evans coolant has poor heat transfer compared to conventional EGW mixes. You could gain some back my using much higher temps and therefore Deltas but the engine won't like operating at 300F for very long.

Surface conduction cooling will be heavier, more complicated and more maintenance intensive in my view. I thing you will need vast surface area to make it comparable to a good radiator setup and it will be very hard to incorporate on an RV airframe. Don't let me stop you from trying however, it's a fascinating approach.

 

[frankh]

My hat is off

to you Mike for giving this a go.

Certainly zero cooling drag would be a leap forwards, even better than an air cooled Lyc.

Another thought but could we use some kind of refridgerant..I.e as the liquid is sprayed from the tubes the reduction in pressure boils the coolant. That way you get the latenet heat of vaporisation working for you too.

Its even more complex than what you describe so taking your approach first is the way to go..But it it didn't work this might be a viable option.

Wouldn't be fun to have unlimited development money?..

Frank

 

[N91CZ]

Wings & Heat Exchangers

One key question is how much cooling is actually required, and will need volunteers to instrument their aircraft with a flow meter and temp sensors and collect the data to establish the actual cooling requirements.

Mike,
You'll need to dissipate approximately the same number of HP as are coming out the crankshaft. The energy of the fuel is roughly split in thirds. One third to the crankshaft, one third to the cooling system and one third in exhaust gases.
Heat exchangers and wings have opposite goals in terms of airflow characteristics. Heat exchangers strive to maintain turbulent flow for good heat transfer. On wings we attempt to do the opposite. The resulting lower heat transfer will require a very large surface area.
This concept was tried many years ago, but I forget which aircraft. It might be worth some time on Google to see where they fell short.

 

[Kevin Horton]

Interesting idea. Surface cooling has been tried many times, but I am not aware that anyone has ever been able to make it work well enough to use it successfully on a production aircraft. Several Schneider Trophy racers used it though.

The critical conditions are likely to be:

• Extended ground ops on a hot day, especially stationary with no wind, and taxiing with a tail wind, and
• extended max power climb at Vy on a hot day.

If you can handle these two cases, the max speed and cruise cases should be a piece of cake.

What sealant material will you use to seal the rivet lines, etc? It will need to withstand the coolant, at very high temperature.

It seems there would be a risk of the return line in the wing sucking air momentarily during transient negative g conditions, as may happen in turbulence. Will that cause any grief?

How will coolant leaks be detected, so you can get on the ground before the coolant level gets to a critical low quantity?

I wish you much success.

 

[N941WR]

...What sealant material will you use to seal the rivet lines, etc? It will need to withstand the coolant, at very high temperature.

Don't forget that as altitude increases, the boiling point decreases. Thus in an open system, you might lose cooling efficiencies as the coolant will boil at much lower temps. That is one reason why cars use a closed system.

...I wish you much success.

Ditto!

 

[Rotary10-RV]

The hard way

Mike,
I hope you have good luck with your experiment. Sadly I do not think you will find there is enough cooling area to do the job. There were several Schenider Cup racers that did this, but they were basicly cost no object aircraft. Remember they were seaplanes, and even the entire surface of the floats was used as a cooling surface. Using the floats turned out to be a good idea since while not flying the heat transfer to the water helped to keep the engine cool until takeoff. Peter Garrison of Flying magazine has done several wing-surface heat transfer articles. One was cooling water cooled engines, and the second was using the engine exhaust channeled through the leading edges to provide a deicing capability. Both seemed like good ideas, but both lacked good enough heat transfer to do the job.
The idea that has more promise is to figure the ducting needed to use a conventional modern aluminum radiator, (an excellent heat exchanger), to minimise drag or through Meridith effect have a net zero cooling drag at crusing speed. The P-51 was supposed to have ended up with a near zero cooling drag. The effect wasn't planned though they "got lucky" with the design. The designers found later when they improved the ducting for the D and later models that were going a bit faster than they expected. Designing a better belly mounted scoop with the best possible inlet and exit ramps is more likely to give you the best possible result. It might also be possible with leading edge inlets to a conventional radiator. Good luck in any case.
Bill Jepson

 

[mikeje]

More than likely, Ross is right in that it may not be a viable method. He has done more development than I ever hope to, but I just want to find out. And yes, it will be a lot of work!

The questions are:

1. How many horsepower actually go out the cooling system (especially during climb). Are there ways to reduce it?

2. Is there enough surface area to dissipate that heat in the climb phase.

3. Can a subaru ej25 turbo be run at say 250 degrees without significant engine wear or failure risk. Is oil temp breakdown the limiting factor? If it lasted 800 hrs, that is good enough, I think. Also, you get better sfc at higher temps. A cold engine is an enemy of efficiency.

4. I would really like to know how much drag a cooling system introduces, or at cruise, how much power is required to overcome that drag. Has this ever been measured? ballpark guesses? If its large enough on average, a zero cooling drag system could use a smaller engine, which puts out less heat, and weighs less.

5. In current installations, how much effective surface area does the radiator have. Its a simple mater of measuring and counting fins. Please post your findings. The flow through a radiator is mostly turbulent, but obviously not at near the velocities of surface cooling. Higher velocity cools better. Period. The aluminum surface doesn't know whether its on the wing or in the radiator! Also, a radiator loses effectiveness at altitude as well (lower air density)

I did some calculations based on Ross's data from his turbo 2.2 rv-6 (an amazing accomplishment!) and a "guestimate" of 23 gpm coolant flow and came up with about 84 k-watts of heat dissipation. Thats 112.5 hp.

Figuring a heat transfer coefficient of 100 w/m^2-c (I think very much within the possible range: this experiment will reveal it) and a temperature difference of 75 c, it takes 11.3 m^2 of surface area! (122 ft^2). Is there that much area available? You builders tell me!

Those calculations are based on laminar flow, which is much less effective than turbulent flow, obviously. How much of the wing are on an rv-9 is turbulent exclusive of the control surfaces? I would guess more than 50%.

So if you consider that my initial calculations were base on the thermal calculator at the link in my original posting (try playing around with it), which is for laminar flow, the turbulent portion will significantly reduce the area required. And don't forget about the radiant dissipation, which is essentially zero in a radiator.

If it can be show that it is impossible to have enough surface area for adequate cooling, no experiment is necessary. If you combine all the unknowns:

1. heat transfer coefficient of air anywhere from 10w/m^2-c to 300 w/m^2-c depending on air velocity, density, viscosity, type of flow, etc.

2. turbulent flow: how turbulent is it? The more the better for cooling.

3 the radiant component: everything radiates and absorbs heat. This setup will also absorb the sun's energy at a rate dependant on the surface

Also, the choice of coolant is open: a high boiling point is the major requirement in order to operate the system at zero pressure. If this whole thing ultimately is feasible, altitude will be the limiting factor.

Mike

 

[Rivethead]

Gotta love it!

I have always loved the concept of something for nothing. After your successful at your cooling experiment I'd like to talk to about my perpetual motion machine. I feel I'm very close to a major breakthrough but just need a little push over the top, so to speak.

 

[mikeje]

Kevin:

Great insights! I think you are exactly right about the 2 critical phases of operation. The ground issues might be solved this way: radiant dissipation, low cooling requirements and prop wash. Possibly a small on board radiator as a supplement with a high volume fan. The climb performance and altitude will be the limiting factor in flight, and climb operation would likely dominate.

Will proseal work? It's tried and true on the tanks, but I don't know if it will stand up to heat and the coolant.

There would need to be fail safe check valves, and a header tank. I envision using the rear inside corner of the wing as a sump on both sides with vent lines between the 2 wings and possibly 2 lightweight return pumps.

Also needed is a high pressure, high volume pump for the spray side. The mechanical pump could be a backup "limp home" mode. What kind of pressure will the oem pump put out on an ej25 turbo.

I would run the spray tubes parallel span wise through the lightning holes, probably with spray nozzles installed for uniform coolant contact.

It seems like wing leaks would be minor and a minimal amount of leakage could be tolerated (like oil burn). Still, a monitor would be necessary to catch a failure.

Mike

 

[rv6ejguy]

On a 200hp engine, I think you'd see a minimum of 30% of the rated hp going into the coolant, call it 60hp or around 45,000 watts.

A rough guess is that you'd need at least 200 square feet of skin area to do the job in climb on a high OAT day given the unknowns in flow condition on various parts of the airframe. This would vary widely with Delta. You don't have enough wing area excluding flaps, ailerons, tips and fuel tanks for this so you'd have to convert some of the fuselage areas as well. Looking very complicated now.

I wouldn't run a Sube at 250F. I think you'd see distress in the exhaust seat area and you'd have to watch piston to wall clearance, ring end gaps and resultant high oil temps as the two are so interrelated with the aluminum block.

I've seen figures of 15-25% for cooling drag estimates but never seen the supporting data. There is some interesting data from German and American sources from the WW2 era on speeds with rad doors open and closed (not fully). I think this was as high as 24 mph difference in the case of FW190, ME109 and P51 aircraft. On a 350-400 mph aircraft, this is a significant amount of drag. As Kevin stated, climb is your worst scenario so you must have a lot of extra capacity compared to cruise.

Sube news had some data on water flow rates and pump presures on various Sube engines. Some of the 6 cylinder engines were pumping upwards of 84GPM with pump pressures over 40 psi if I recall.

I typically see 10-15C drop from inlet to outlet in cruise. This is mirrored in some of the German data I found.

I don't think ground running would be such a big deal as you are only making 10-15 hp probably. This assumes some of the radiating area is located within the prop arc.

The Evans is messy stuff, more like oil. I'd just wonder how you'd effectively scavenge the sprayed coolant from all the nooks and crannies in various flight attitudes.

I ran some bench tests a couple months ago on various rads. The standing radiation rates (still air) were not totally insignificant. I did this to baseline the data comparing various rad sizes. The info was published on Sube news again.

 

[mikeje]

Ross,

My calculations are based on your temperature differentials across the radiator inlet and outlet, and a guess of 23 gpm flow rate. That worked out to 84kw. It would be great to know the actual flow rate, so precise heat rejection at various flight phases could be known

What is the effective surface area of your radiator setup? I know you said that you had 429 cubic inches of cooling, but how much total effective surface area?

In your opinion, how would this area compare to wing surface area as a transfer surface? What I have come across so far is that higher velocity on a flat plate is substantially better. Turbulent flow is substantially better than laminar. As near as I can tell so far, higher velocity trumps turbulent flow.

The coolant choice is up for grabs, as all of this is really! I put forth evans npg to just express the idea; there are probably better solutions to that question, but a very high boiling point is essential to be able to operate the system open to the atmosphere.

I guess I have a hunch that the very high velocities across the wing surface will translate into very high dissipation rates.

You can't compare those race planes, because of their high hp, and therefore high cooling requirements. It is interesting to see the amount of drag difference in the rad door open or partially closed. If cooling drag really is 15-25% of the total drag, then this would make a huge difference in power requirements.

I think there is a chance at these low powers (200hp and down), 130 ft^2 will do it! If that were true, is this idea viable? Are the other engineering problems worth it IF the cooling can be accomplished in 130 feet of area, about 33 ft^2 per wing surface ? That for me is the starting question.

Another related question is what is the true airspeed in a high performance climb for a 180hp rv? Since this is likely to be the limiting factor for this cooling system, would it be unreasonable to simply limit climb to a minimum airspeed or something like that? Would that defeat the point of a powerful engine or is it a reasonable compromise?

Mike E

 

[BobK]

air conditioning coil

Not sure on what the hps it takes to run an air conditioner, but instead of forcing air to cool you, maybe a coolant/air conditioner coil exchanger. I use to have a hot water heater/ air conditioner exchange unit on the house. During the Summer, the air conditioner was able to heat the water in the tank. Reverse that and maybe the air conditioner can cool the oil. trade off - HP less drag vs hp to run the conditioner, however, should not have a problem cooling it on the ground. Oh yeah, the extra weight.

 

[Kevin Horton]

I think you are exactly right about the 2 critical phases of operation. The ground issues might be solved this way: radiant dissipation, low cooling requirements and prop wash. Possibly a small on board radiator as a supplement with a high volume fan. The climb performance and altitude will be the limiting factor in flight, and climb operation would likely dominate.

I think the climb case will always be worse than the cruise at altitude case - during climb the power is at least as high as in cruise, and the airspeed in climb is lower than in cruise.

If we look at this from an efficiency point of view, the aircraft spends most of its time in cruise, so this is where drag reduction is most important.

If we ignore the cost in weight, complication and extra failure modes, the idea of a supplemental radiator makes sense. Use it to help out on the ground and in climb. Have moveable doors to close off the air entry and exit to this radiator - you close the doors in cruise, and get a big drag reduction.

 

[rv6ejguy]

I think the climb case will always be worse than the cruise at altitude case - during climb the power is at least as high as in cruise, and the airspeed in climb is lower than in cruise.

If we look at this from an efficiency point of view, the aircraft spends most of its time in cruise, so this is where drag reduction is most important.

If we ignore the cost in weight, complication and extra failure modes, the idea of a supplemental radiator makes sense. Use it to help out on the ground and in climb. Have moveable doors to close off the air entry and exit to this radiator - you close the doors in cruise, and get a big drag reduction.

Exactly what the Germans did on the He100 fighter which had surface conduction cooling plus a small retractable rad in the belly.

Yes, climb is always where we have to watch coolant temps the most given the high power settings and low speeds. Temps are never an issue in cruise even at 15,000 feet on a warm day with a turbocharged engine.

 

[rv6ejguy]

Pretty difficult to calculate effective surface area of radiators with all the louvers especially.

Cooling drag should be roughly proportional to installed hp and the resultant mass flow required.

The aircraft I was referring to were stock WW2 fighters, not Reno racers which depend to a large degree on spray bar cooling.

I think boundary layer thickness, especially aft of the spar would seriously reduce actual heat transfer.

For water flow rates, It looked as though the Subaru 4 cylinder pumps are in the 40-45 GPM rates (free) at around 4500 rpm. You'd have to make a wild guess with pumping restrictions on what actual flow would be or measured it with a flow meter on a running engine.

Either EG or PG could be used, their properties in pure form are actually pretty similar. Unfortunately they are quite inferior to water in heat transfer but you probably could not use aqueous solutions in this setup due to corrosion considerations either.

So I think you see that the engineering considerations make this route technically questionable however maybe a test section with instrumentation could be mounted to a liquid cooled car or truck to gather performance data and various speeds. That would validate or invalidate the basic premise. We might be surprised!

 

[TSwezey]

I have thought about cooling using the wings also. What I think you will need is a combination of two systems. The first, for ground and climb ops, is using the radiators and the second, your wing system, for inflight cooling. You can you some type of cowl flap for the radiator which could be shut during flight. The biggest penalty is the added weight of all the coolant and the system itself. But you could really cut down on drag. It would really provide a nice de-ice system.