By: Daniel R. Nicoson (email: A6intruder [a] adelphia.net)
Well, the summer is over, the master bedroom is now twice as large and a different color, and I still haven’t finished the vertical stabilizer! Does this sound familiar? All summer I’ve pondered my cooling system and now I’m finally going to organize those thoughts.
My aircraft is in the detail design phase with a few initial parts under construction. It is a four seater with tail volume coefficients closer to that of a DC-3 than a P-51. A small block Ford V-8 will power it. I was inspired by two articles written by Charles Airsman, Jr. (Ref #1 & #2) to consider using exhaust augmenters to make my cooling system simple and EFFECTIVE under all operating conditions. The point of this article is to lay out the considerations of my cooling system and some design rules of thumb for the exhaust augmenter. I have referenced several technical papers and previous Contact! articles that discuss these issues and provide some test data with conclusions. Most of this information applies to both air cooled and liquid cooled engines. You will see that exhaust augmenters change the rules normally applied to a pressure cowl cooling system.
First of all, what are the goals for my aircraft’s cooling system? Without defining my goals I could get way off track in the design phase. I’ve listed them in order of priority:
1-Adequately cool the engine in all phases of operation.
2-Maximum reliability.
3-Reduce cooling drag.
4-Keep weight to a minimum.
Goal #1: Adequately cool the engine in all phases of flight - If the engine can overheat during any operating condition, all the other goals are pointless. Let’s discuss some operational conditions.
Ground idle and taxi: If you think it’s OK to shut down while waiting your turn to take off in heavy traffic because of engine overheat, you’ll make yourself nuts. Just try to leave the regional pancake breakfast at the same time as everyone else and you’ll see what I mean. Now add a cranky tower controller… Besides the obvious multiple starts being hard on your battery, it’s a major safety issue to keep breaking your cockpit/checklist flow just prior to takeoff. The engine HAS to cool properly on the ground for any length of time.
High power run-up and takeoff ground run: A very demanding condition, very high power and little to no useful airflow. Prop blast is very turbulent with no forward air speed. Don’t count on those beautifully sculpted NACA scoops you built to be real effective until you have some airspeed. Ever wonder why the cooling inlets on the good old C-172 are the size of an open barn door? It isn’t because of cruise conditions! Those large openings are the only way to get enough cooling air in during ground ops and very low airspeed flight on an aircraft with the conventional pressure cowl cooling system. The engine should be comfortable at full power, zero airspeed for several minutes. Don’t cut down the size of these inlets to reduce drag unless you have a way to pull cooling air through the engine/radiator. We’ll see how to do that in a bit…
Climb conditions: Actually things are now getting easier. Although most of our aircraft will climb at or near full power producing lots of heat, you now have some significant airflow going past the aircraft. This airflow makes the cooling job easier because now we have high-pressure areas and low-pressure areas somewhere on the aircraft. In fact, cooling the engine is now very simple, take in cooling air at a high pressure point and exhaust the cooling air at a low pressure point. Easy, right? We’ll get there…
Cruise conditions: Actually the easiest point to cool the engine. Steady state, mid to high power levels, LOTS of airflow and very well defined pressure areas around the aircraft. Inlet and cooling air exhaust areas needed are the smallest at cruise because of the airspeed. It is easy to focus on the cruise condition when your design is a mini-airliner like mine. Don’t optimize the cooling airflow for cruise at the expense of the first three operating conditions or you won’t make it to cruise!
Descent and landing conditions: Generally low powered flight at pretty good airspeeds until you slow down for final approach. Your big concern with an air-cooled engine is shock cooling. How will your cooling airflow be controlled to avoid this? A properly set up liquid cooled engine avoids shock cooling by using a thermostat to re-circulate coolant and avoids over-cooling the engine (don’t throw out the thermostat!).
Goal #2: Maximum reliability - An obvious requirement that really could be a whole article of its own. One route to reliability is simplicity of the system (KISS Principle), few moving parts and proper selection of quality components. You’ll see later in this article that my cooling system design has only five moving parts; an air-cooled engine would have no moving parts in the cooling system. I will probably use a custom built radiator for this design and I am planning to do some homework on hoses and fittings to ensure performance and reliability.
Consider the failure modes of the components. What happens if the water pump belt breaks? How likely is that type radiator hose to fail? Are hoses properly supported so they don’t cause the fitting on the radiator to fatigue? Will radiant heat from the exhaust header melt that beautiful fiberglass-cooling duct? If your thermostat spring does break, is it fail-safe or fail-seize? Hopefully you are considering the failure modes of components throughout your aircraft. Those above are only a sampling of the possibilities.
Goal #3: Reduce cooling drag - This is the really sexy goal for a cooling system. Most people want to put it first on the List of Goals. Depending on the aircraft, cooling drag can be 20-35% of total drag. Reducing cooling drag normally means limiting the amount of air used to cool the engine. If you can cut your cooling drag in half, that’s significant free speed! Be careful not to reduce cooling drag at the expense of Goal #1 & #2.
The standard C-172 meets all of our goals except low cooling drag. Its cooling system is SO simple it sacrifices cruise performance for all around cooling. Higher performance certified aircraft add the complexity of cowl flaps and pilot workload to preserve all our goals and reduce cruise cooling drag.
Cooling demand is actually a function of power. High power, high cooling demand. The traditional pressure cowl cooling systems regulate cooling air as a function of airspeed and cowl flaps and have no link to the actual cooling being demanded by the engine. The ideal cooling system would regulate cooling airflow as a function of engine power setting. A cooling system incorporating exhaust augmenters is independent of airspeed and fully dependent on power setting. This allows reduction of cooling drag without compromising Goal #1 and requires no additional pilot workload.
Goal #4: Keep weight to a minimum - Liquid cooling systems get accused of breaking this goal all the time. This doesn’t have to be the case. I once read an article on a gyrocopter running a Subaru engine that wouldn’t cool very well. The builder had a pretty large radiator but did not want to add the weight of any ductwork. His solution was to compromise his flight profile to avoid overheating. He didn’t understand that an un-ducted radiator almost never works (Ref #3). A very basic exhaust ejector and minimal ductwork would probably have allowed this builder to use a smaller radiator and cool properly in all conditions with no significant weight gain. Don’t save weight at the expense of proper cooling!
My cooling system: Ok, we’ve set the stage. Let’s talk about the actual cooling system I intend to put together. As stated earlier, my design will use a small block Ford V-8 turning pretty good power for takeoff (400+ HP) and depending on my financial status for a given flight (fuel flow-cost), I expect to use 200+ Hp during cruise. I have designed a fairly wide fuselage to allow my oversize, non-FAA-average body to be comfortable with my three non-FAA-average passengers during flight. The added benefit of this wide fuselage is that I have useful room on each side of my engine to run a small radiator and associated cooling ducts. After the cooling air has gone through the radiators it will head downstairs and enter the exhaust augmenters that are faired into the belly contour of the aircraft. I will use two augmenters, one for each bank of cylinders.
The first hard numbers I had to calculate was to decide the basic dimensions of my radiators to cool my engine. In ref #4, Turbo Tom gives two rules of thumb on sizing the radiator. He says .9-1.2 square inches of radiator face area for each cubic inch displacement. Also, 1.85 to 2.1 cubic inches of radiator volume for each horse power. In Ref #1, Charles uses a radiator that calculates out to .9 square inches of radiator face area per cubic inch and 2.48 cubic inches of radiator volume for each horsepower. Since my system copies the concept Charles uses I will calculate my radiator needs using his numbers.
My radiator volume needed would be: 400 HP x 2.48 = 992 cubic inches of radiator volume. Now if we decide the radiator should be no thicker than 2.75” (ref #1) we could divide the 992 cubic inches of radiator volume by 2.75 = 360.72 square inches of radiator face area. We check this against the rule of .9 square inches of radiator face area per cubic inch: 351 CID x .9 = 315.9 square inches. It looks like radiator volume is the driving rule in this case; I will accept a higher radiator face area (360 square inches) in order to handle the high horsepower. So the starting point for my radiator is a 2.75” thick radiator with 360 square inches of frontal area. I’m comfortable with this starting point since both of these builders have successfully FLOWN their installation.
I just took several hours to roughly lay out this setup on my CAD drawing. It looks like I have room to fit a 9.5” x 25” x 2.75” radiator on each side of the engine. Doing the math this would give me a frontal area of 475 square inches, and 1,306 cubic inches of radiator volume. This would in theory support 526 HP! This is probably going to be a good match because that 9.5” x 25” radiator still needs end tanks and such which will decrease the effective radiator area.
Let’s talk ductwork for a minute. Reference #3 is an excellent refresher on basic airflow through ducts and radiators. Read it! The drawings I provided for this article simply give the basic path for my cooling airflow. I have not had time yet to make all the curves mathematically smooth but you get the idea. When I laid out these drawings I considered round inlets and the NACA style surface inlet. Reference #5 discusses the use of NACA ducts and gives coordinates to achieve the proper shape. One point made was that ducts should not diverge by more than 5-7 degrees to avoid airflow separation and allow the air to slow down and build pressure. I didn’t have enough distance between the prop and the radiator to use a NACA inlet and ductwork that diverged at 7 degrees or less. I choose a rectangular duct opening just behind the prop. This insured the ducts diverged less than 7 degrees in all three dimensions. The inlet area is 15% that of the radiator frontal area per reference #1. This inlet uses a width to depth ratio of 5 as suggested by ref #5.
You’ll note that I tilted the radiator 15 degrees. I did this primarily to give the radiator discharge duct some room to make that horrible set of turns down to the augmenter. I figured horrible turns would be easiest at this point since the velocity here is the lowest in the system.
Exhaust Augmenter Theory: The augmenter is simply a jet pump that uses a smaller high-speed jet of gas to move a larger volume of air. In this case we are using the augmenter because it will help move cooling air through our radiator (or air cooled engine) based on the engine load regardless of aircraft speed. Reference #1 was able to run at full power on the ground continuously with no problems cooling the engine.
The ideal augmenter has three sections; the inlet shaped for smooth entry of the cooling air and positioning of the exhaust jet; the mixing section with a constant cross section and finally a diffuser section. References #7 & #8 discuss that it is actually the friction at the boundary of the high speed jet that creates the pumping action. These references specifically state that the exhaust pulse is not a factor that increases or decreases effectiveness. The expansion of the mixed flow in the diffuser adds 40-50% more vacuum at the inlet of the augmenter (Ref #6).
From reference #6 I came up with some basic rules of thumb for the design of my exhaust augmenter:
- A piston engine will create approximately .094 pounds-mass of exhaust gas per horse power
- The ideal exhaust augmenter can move 6 pounds-mass of cooling airflow for every pound-mass of exhaust gas (mass flow ratio)
- The mixing section cross section area should be 10 times the area of the exhaust jet area
- The mixing section length should be 6 or 7 diameters long
- The diffuser outlet/inlet area ratio should be 1.87
- Diffuser should not expand with any more than a 12 degree included angle
- Up to 6 inches of water pressure rise was realized between the inlet and the outlet of the augmenter.
- Operation without the diffuser section decreases the pressure rise by about 40%
In the case of my cooling system the augmenter begins with the exhaust pipe. You see from drawing #1 that the “shorty headers” feeds into a short 2’ long length of exhaust pipe that ends up at 2.15” in diameter. There is a slip joint and a ball joint along this exhaust pipe to allow for some movement of the engine. The discharge of this exhaust pipe creates the high-speed jet at the inlet of the augmenter.
Rule of thumb #3 says the mixing cross section area of the augmenter should be 10 times that of the exhaust jet area. Doing the math gives a diameter of 6.8” for the mixing section. Rule #4 tells us the mixing section should be 6 or 7 diameters long: 7 x 6.8” =47.6” long.
Now, the theoretical augmenter would include a diffuser section. Rules 5# and 6# would be incorporated to size the diffuser. I am going to skip the diffuser in my design for several reasons. First, size. With my layout being external to the aircraft, the increasing bulge for the diffuser under my aircraft belly is really going to upset orderly airflow at all the wrong places and significantly add to drag. Second, in Ref #1 Charles had very good performance without a diffuser. It is a compromise from the ideal, but for good reason. In my case, if testing proves that I’m not getting enough augmenter effect, then I will have to find another solution. After all, this is an experimental aircraft, some trial and error is allowed!
If you are using an air cooled engine, size your augmenter based on the smallest diameter exhaust pipe that will still let you develop full power. Then your mixing section diameter will be 10 times the area of this pipe (Rule #3) and 6-7 diameters long (Rule #4). Your inlet area feeding your cooling plenum ahead of your cylinders should be just large enough to allow proper cooling during extended full power ground runs. Remember, the augmenter will pull a bunch of air through those inlets.
Other Concerns: What other concerns and considerations do I have for this augmenter based cooling system? Heat of the mixing section walls. Ref #1 stated that this section only became warm to the touch. Makes sense if you’re running 180 thermostat, the cooling air through the radiator has to be something less in order to transfer heat. My fiberglass and end grain balsa floor should be okay up to about 180F. Now if I wrap the mixing section with good insulation there should be no problems.
Another concern is noise. If this straight pipe set up is just too noisy, a muffler could be used. In my case I would turn the shorty headers around, discharge at the front of the engine and run a small muffler along the side of the motor prior to entering the augmenter. Hopefully a muffler won’t be needed. It would bring on a lot of packaging problems and take out some effectiveness of the augmenter.
What about idle performance? Of course there is a very low heat rejection demand at idle but how much velocity will the jet have out of a 2.15” diameter exhaust pipe? Testing will tell. If I use a smaller outlet for better augmenter jet speed at idle, will I have too much exhaust restriction at full power? Cam design will affect the sensitivity to back pressure. A cam with less “overlap” (intake and exhaust valve open at same time) will be less sensitive to back pressure. If you use a turbo charger minimum back pressure after the turbo is essential. The good news is that Ref #1 had no problems cooling at idle.
Reliability: If you just finished ground testing your air-cooled engine with exhaust augmenters, you should have a cooling system with no moving parts! Any internal ductwork should be properly secured so vibrations and air loads can’t crack or cause the duct to disconnect. Radiant heat from the exhaust shouldn’t be allowed to damage ductwork or other items. Do not wrap exhaust pipes with exhaust insulation! This stuff is fine for cars that only run full bore for short periods (15-30 seconds). Aircraft engines, Lycoming OR Ford, are commonly used at 60 % or more power for hours on end and transfer a lot of heat through the exhaust pipe wall. If you block the discharge of this heat energy on the outside of the pipe with wrap, you can cause the pipe material to overheat and burn a hole through the wall. Very dangerous! Plan for proper airflow around the exhaust and just route heat sensitive stuff away from the exhaust pipe and/or use proper heat shields.
My cooling system will have 5 moving parts, the thermostat, water pump, two water pump drive belts and the pressure relief cap. I will have to research the thermostat failure mode. I would want any failure to allow full circulation through the radiators. Water pumps do fail but usually not until very high miles and advanced age. Old seals and bearings are usually the cause of failure. A good quality OEM pump should be fine with two new belts. Consider a spring-loaded tensioner to keep belts happy. Belts are relatively cheap so change them each annual. Think OEM on the thermostat, pressure cap and water pump. Ford and GM hate warranty claims and they strive to make these components very reliable.
You might have noticed that I did not use all of the augmenter design rules of thumb that I distilled from the references. I left in the other rules of thumb for those of you with a more formal engineering approach that might want to do a more detailed analysis.
Please also note that this article reflects a written plan for my aircraft. I have not built or tested this design yet, but I think it is a well-founded plan based on the references cited.
I would encourage all of you that have been flying and developing engine installations to submit your thoughts and experiences on the failure modes of any components you have had problems with. I am thinking of mundane parts like radiator pressure caps, radiator hoses, water pumps, belts, thermostats and duct work etc. Write a note about what happened and why you think it happened. Send this info to Mick or I. I will be happy to compile this information in another article so we can all keep learning from each other and avoid repeating the same errors.
Meanwhile, fly safe!
References & Further Reading
- Charles D. Airesman, Jr., Drag Reduction Through Water Cooling and Exhaust Augmentation, Contact! #39, p. 11
- Charles D. Airesman, Jr., Subaru EJ-22 Installation in a Varieze, Contact! #47, p. 6
- S. J. Miley, Review of Liquid-Cooled Aircraft Engine Installation Aerodynamics, Contact! #17, p. 13
- Turbo Tom Wyatt III, Adequate Cooling of Liquid-Cooled Engines, Contact! #40, p. 13
- NACA Submerged Inlets, http://abweb.larc.nasa.gov:8080/~kleb/naca/inlets.html
- Eugene J. Manganiello and Donald Bogatsky: An Experimental Investigation of Rectangular Exhaust-Gas Ejectors Applicable for Engine Cooling, NACA Report No. 818, 1945 http://naca.larc.nasa.gov/reports/1945/naca-report-818
- Ir.J.J.G. Koopmans, A Theory for the Design of Multiple Exhausts for Steam Locomotives, The Ultimate Steam Page, http://www.trainweb.org/tusp/koopmans/koopmans.html
- Ing. Livio Dante Porta, Theory Of The Lempor Ejector As Applied To Produce Draught In Steam Locomotives, The Ultimate Steam Page, http://www.trainweb.org/tusp/lempor/lempor.html
- Jess Meyers, Experiences With Liquid Cooling and Radiators, Contact! #43, p. 14
Note: This article was first published in the Contact! magazine. The article in the magazine also has a very nice 3D drawing of the system.