Cooling drag question

[chuck]

I've been following the cooling drag threads for some time and am fully indoctrinated with the following blue print of cooling drag optimization:

1) Inlets as far out as possible (circular preferable)
2) Hard plenum
    a) reduce leakage by baffling
    b) reduce air speed in laminar fashion
    c) create max pressure in upper cowl
3) Clean up flow in plenum (per Atkinson sp?)
4) Use outlet area ratio per Dave Anders
5) Use exhaust augmentation 
6) Clean up internal airflow (per Bob Ax.)

One of the core ideas is that the air coming is high speed and that it slows down through the plenum to pass the engine at low velocity and then it speeds up at the cowl outlet.

My problem with this is that when I look at the amount of cross sectional area available through the cylinder fins, it looks to me like the area is, if anything, less than the area of the inlets. If the area was exactly the same, then the velocity through the fins would be the same as the inlet velocity, (not necessarily aircraft air speed) and there would be no need for a plenum (to slow the air down to low V/high P). Even worse there is more drag through the fins because of the greater surface area, so equal area would actually look like less area.

This observation could explain:

1) Why we can reduce the inlet size with little effect on cooling (RV-4's)
2) Atkinson reports lots of ugly airflow under the plenum
3) Washers between cylinder and baffle improve cooling in a measureable way
4) Why Axsom saw a drag decrease by cleaning up the airflow in the supposedly low-velocity region of airflow

Is my eyeball estimate of the available cylinder thru area way off, or am I missing something else?

[gmcjetpilot]

Originally Posted by chuck

1) Inlets as far out as possible (circular preferable) (A)
2) Hard plenum
a) reduce leakage by baffling
b) reduce air speed in laminar fashion
c) create max pressure in upper cowl
3) Clean up flow in plenum (per Atkinson sp?) (F)
4) Use outlet area ratio per Dave Anders
5) Use exhaust augmentation (E)
6) Clean up internal airflow (per Bob Ax.)
[/code]


(B) My problem with this is that when I look at the amount of cross sectional area available through the cylinder fins, it looks to me like the area is, if anything, less than the area of the inlets. If the area was exactly the same, then the velocity through the fins would be the same as the inlet velocity, (not necessarily aircraft air speed) and there would be no need for a plenum (to slow the air down to low V/high P). Even worse there is more drag through the fins because of the greater surface area, so equal area would actually look like less area. (B)

This observation could explain:

1) Why we can reduce the inlet size with little effect on cooling (RV-4's) (A)
2) Atkinson reports lots of ugly airflow under the plenum (C)
3) Washers between cylinder and baffle improve cooling in a measurable way (D)
4) Why Axsom saw a drag decrease by cleaning up the airflow in the supposedly low-velocity region of airflow

(B) Is my eyeball estimate of the available cylinder thru area way off, or am I missing something else? (B)

Nice summary:

A) " Why we can reduce the inlet size with little effect on cooling (RV-4's)"

I say any plane can reduce the inlet size with little effect on cooling, but it must be accompanied with cooling efficency improvements. Cooling drag reduction is using as little air as possible to get the job done. When you say little effect on cooling, I assume you mean no increase in CHT. Have you heard of problems with a Sam James Cowl and high CHT, that would not have been there any way on other RV models? A RV-8 with a stock cowl can have high CHT, and it will not get better with a SJ cowl.

Since reducing the inlet area has a great effect on reducing cooling drag on any plane I will say any plane benefits. However you must have enough cooling. I don't think the RV-6/7/8/9 run hotter than they typically do with a SJ cowl, but they do run hotter than a stock RV-4. The reason you can lose 40% inlet area is because you are using air more efficiently (see below)

The question is why is the RV-4 so efficient in its basic stock form? SEE Item -F- below. Here is one of the "secrets" of the round wide spaced inlet. Even the new stock Mooney Ad shown in the latest magazine shows a semi-round wide spaced inlets. See my diagram for explanation:



B) We can't do any thing about the fin's. First you NEED that many fins and surface area. Why? It has to do with heat transfer. You need enough surface area. You need that pressure differential between the high and low press plenum as a forcing function to get air thru those cyl fins. This pressure is the power of the cooling system, volumn the capacity. Volume is not an issue, there's plenty of that, but pressure is an issue. However in theory you're correct, wide spaced fins would allow less restriction and thus req less pressure diff (and cooling drag). The down side is the engine would burn up for lack of enough heat transfer. The engine engineers took care of the fin geom for us. You'll find different designs, Lyc, Cont, Frank, P&W and Wright but more similarities between them. We can can forget playing with fins, but it's a great question.


C) Clearly there must be some turbulent air flow under the (high pressure) plenum, but I would like to see the data or supporting argument for this theory. From the NASA/MSU/T&AM report I am not seeing it, but clearly turbulence is likely and undesirable and should be minimized. The critical area and one that needs the most focus is the inlet and diffuser TO THE plenum. When it gets into the plenum, it's slowed down and not as critical, but I guess small improvements can be made. The inlet is an area of much research and analysis. I see people carve a round hole in their cowl and say SEE! The real magic of the inlet starts at the lip, through the inlet throat (laminar) and into the diffuser. There's much work there. The plenum is just a big reservoir. Here's an excerpt showing pressure survey in upper plenum of test aircraft:



D) The reason playing with baffle to fin gap is the asymmetric fin depth front to back on Lyc engines. Since all cylinders are identical, the fins against the #2 and # 3 cyl (fnt lft, rear rt) are choked off, since the fin depth is almost zero on the side. The reason Lyc did this? Save weight and allow closer cyl spacing (I guess). See my diagram:

(I guess using the air you have better comes under the heading of cooling drag reduction.)


E) "Use exhaust augmentation " - Well understood and exploited in piston planes for 6 decades, but hard to execute in a stock RV airframe. Van could design an exhaust ramp verses the flat bottom into the belly, but this makes it complicated, heavier and more expensive. Pipes sticking out in the breeze is not ideal. To make an augmentor you would need to go externally, below the belly well past the firewall in my opinion.

Heck the 1958 Piper Apache I owned had exhaust augmentor tubes, as did Cessna 310's and many planes since than. Twin engine planes have more room to develop an augmentor, firewall to wing trailing edge. I'm not sure I'll try it myself. The reason, it adds weight and surface area, whose drag might negate the benifit. A recessed tunnel in the belly of the plane allowing the pipe to be reaccessed is ideal, but structurally not practical.


F) "Clean up internal airflow (per Bob Ax.)" - I think Bob had an idea which seemed reasonable. Why let the air float around in the accessory area when you want it to go out the exit. The results where mixed but interesting.

WHY DO RV-4's cool WAY BETTER than any other RV? That's a good question. Does any one know? No one has ever answered my question. So I came up with a theory. Instead of making that massive sealed vertical bulkhead as Bob did, why not replicate the internal dimensions of a RV-4 cowl, especially the lower cowl. Look at a RV-4 cowl compared to the RV-6/7/8/9/10 "wide cowl". The lower "wide cowl" volumn is huge compared to a narrow RV-4 cowl (hint, hint). An easy mod to the cowl may be to bond internal step/fillers to reduce the volumn, a molded form verses just dumping air into a cavernous lower cowl. This lower shape guides the air from the front of the cowl all the way to the exit. Exhaust pipes, air box may get into the way, but work around them as best you can. This also partly applies to the upper plenum and upper cowl. Instead of a big flat upper plenum, make it follow the contour similar to the RV-4. The concept in summary is replicate the RV-4 cowl internally. hmmmm

The RV-4 works. When you think about it (rationalize it), getting the entire volumn or cross section of a massive lower cowl funneled out the small exit produces needless turbulence. Control the exit flow from the front of the lower cowl all the way to exit, not just at the exit.

"This is brain surgery, not rocket science... now hand me that Icecream scoop." Homer Simpson

[Bob Axsom]

The information about the reduction of volume in the lower cowl and devising a way to get the air directed to the outlet cleanly is very interesting and consistent with what I saw in my experiment. My first modification addressed the idea of getting the air to the outlet cleanly. Basically three surfaces: (1) a surface that starts straight down then curves back to overlap the outlet area of the lower fuselage skin (2) a verticle plane extending at and angle determined by the location of cylinder #3 from the right side of the outlet to the right side of the cowl (3) a verticle plane extending at and angle determined by the location of cylinder #4 from the left side of the outlet to the left side of the cowl. I saw a 2 kt loss in speed in a flight test.

In a subsequent experiment I further modified the system by running a baffle a little below the edge of the lower cowl from the angled baffles added in the first modification. This reduced the volume in the post cylinder area and it reduced the flow options of the air leaving the cylinder fins. The speed increased 6 kts over the first mod and 4 kts over the stock configuration. Photos and test results were added to the builder mod area of this website. These results are exactly in line with some of the information in George's post.

Today I will finish a mod to the air box that converts the top to a flat plate. No test yet of course but the results could go either way. I seems that as the flow gets cleaner the volume of air (mass) increases for greater cooling but it also produces an increase in cooling drag. Because the cooling efficiency is greater you can cut down the air mass required to flow through the system for appropriate cooling. This can be done at the inlet or outlet (I'm just at the thinking stage on this but when I have more time I plan on restricting the outlet and testing it now that I have CHT probes in all cylinders).

Bob Axsom

[chaskuss]

Originally Posted by gmcjetpilot

Nice summary:

snipped
D) The reason playing with baffle to fin gap is the asymmetric fin depth front to back on Lyc engines. Since all cylinders are identical, the fins against the #2 and # 3 cyl (fnt lft, rear rt) are choked off, since the fin depth is almost zero on the side. The reason Lyc did this? Save weight and allow closer cyl spacing (I guess). See my diagram:

(I guess using the air you have better comes under the heading of cooling drag reduction.)
snipped
"This is brain surgery, not rocket science... now hand me that Icecream scoop." Homer Simpson

George probably already knows this, but I'd like to clarify it for the others reading this thread. Lycoming was definitely trying to save weight here. Note that the finning is thinner on the cylinder head nearest the INTAKE port. Since this area has nice cool air flowing internally through the port, it should require less finning. As George has pointed out, the problem comes when the airframe manufacturer mounts the local baffling to tightly. It has been reported on the Matronics RV List, that adding one or two standard AN960 washers between the rear baffling & the cylinder head on cylinder #3 will make a noticeable improvement in CHT. The washers allow an increase in local airflow between the head and baffle.
Charlie Kuss

[Mike S]

George's comments dug something out of my memory------take a look at the cowling outlet of the AR 6 in the attached link.

http://www.ar-5.com/

Scroll down towards the bottom, click on photos under AR 6.

In fact, as I have mentioned before, there is a lot to learn by looking at this guys planes.

Mike

[chuck]

In trying to be complete I gave too much information so I'll try a different approach to asking my question.

As I understand this is what we supposedly do with the cooling air.
1) Inlet is small cross section
2) plenum to slow down air and minimize losses
3) through fins (large cross section)
4) to outlet (small cross section) to increase speed to ambient.

This is not unlike what we do to reduce losses through an air filter.

My observation is that since the inlet cross section and the cross section through the fins is similar we really have a system that looks like a constant velocity (I'm using that term loosly) system until the point the air passes the fins, at which point the cross section gets big (entire lower cowling) and then small at the outlet.

If that is the case I would surmise that:

1) The plenum as resoivoir doesn't do what we think (e.g.Atkinson's tufted video) reverse flows, turbulence and other counter flows.
2) Increasing fin cross section (e.g. baffle washers) should increase flow.
3) Reducing the effective cross section in the lower cowl will eliminate a needless (and uncontrolled) expansion. (Bob A)
4) Cleaning up flow in the lower cowling (as opposed to reducing cross section changes) should help (e.g. rounded outlet lip on RV8's).

So I'll raise the question:

IS THE CROSS SECTION THROUGH THE COOLING FINS LARGER OR SMALLER THAN INLET CROSS SECTION?

If it is smaller then I think current plenum/cooling systems don't have the right conceptual model of the cooling flow.

Or I could be full of it

[RV8RIVETER]

To try answer. The area of the fin openings should not be the same as the inlets. As George alluded to the surface area of the fins produces alot of drag so even if the area was the same, the volume of air entering the inlets would be greater than what passes thru the fins.

You also have to take into account the change in direction the plenum air has to make to go thru the fins. We are not talking laminar flow here. However, turbulence is not all bad. You want the air passing thru the fins to be somewhat turbulent, because it will transfer heat better. Like stirring a pot of soup.

I also agree with George that to make an optimum augmentor per report would be excessively long and complex. But, the key word is optimum. Perfect case gives 6 inches of pressure drop. I think a compromise can be made that while not optimum, will provide enough benefit to be worthwhile. I have talked with a Long Ezy builder who's set-up is far from optimum (8 inches), yet he can actually feel the exhaust inlet draw at idle. One good thing for pushers. I would be happy with 3 inches, which I think would be enough to further decrease inlet size and still cool idle and slow (because the augmentor is really only to make up for low cooling flow when on the ground, slow, or in climb). I am going to give it try anyway.

From another source:
"S.J. Miley (Miss St) actually recommended making the high pressure plenum's cross-sectional area equal to that of the inlet(s), so the the plenum is not so much a plenum as a duct."

How that squares with plenum resevoir ideas I don't know.

I have been really brainstorming these cooling ideas and induction pressure recovery mods. I am going to try and make a scaled prototype of the induction assy first and instrument it with a water manometer to see what I can find out. I have also been thinking about ways to test and have ruled out a high pressure blower, as they are very expensive. But I was thinking about 3-4 sources of pressurized air (compressors or bottles) hooked up to a ganged together set of nozzles to blast air into the test subject opening. That should provide both high volume and pressure. If that fails plan B is mount it to the car and drive 70mph. Wonder what kind of looks I will get on that one?

Any suggestions or other ideas?

[RV8RIVETER]

Almost forgot. For the working engineers out there, how difficult is it to set-up a fluid dynamics simulation of inlet shapes or plenum shapes to see what it says.

I found some of the cheaper desktop programs that were only $3,000.00 .

[gmcjetpilot]

Wade Lively: "From another source:
"S.J. Miley (Miss St) actually recommended making the high pressure plenum's cross-sectional area equal to that of the inlet(s), so the the plenum is not so much a plenum as a duct."

How that squares with plenum reservoir ideas I don't know."

Yea that is GREAT. I am not sure either. I guess practical aspects come into effect. Like room for the plugs and push rod tubes. I missed that quote from S.J. Miley, but tend to believe him. Hummm wounder if he is still around.


Chuck Bass "If that is the case I would surmise that:
1) The plenum as reservoir doesn't do what we think (e.g.Atkinson's tufted video) "

Do you have a copy of this video I have not seen this tuff test.

Chuck Bass - "IS THE CROSS SECTION THROUGH THE COOLING FINS LARGER OR SMALLER THAN INLET CROSS SECTION?" I see you want answers! I see you are frustrated but there is no one answer. 20 ingredients going into the aerodynamic soup of calculating inlet area. There's nothing like flight test. We have good data from NASA, Barnard, Sam James and Dave Anders. Here is an example of some old NACA stuff. This looked at fins spacing and position relative to airflow: http://naca.larc.nasa.gov/reports/19...report-674.pdf (Although for a radial, one conclusion was cooling improved with tighter fin spacing, until resistance increased too much. This 1939 data helped Lycoming optimized fin spacing for good cooling and reasonable air resistance. There's a usable range of fin spacing, depth, thickness and structural considerations. OLD but interesting news. If you have energy to read them, they are interesting but too technical for most non engineers. I have an engineering degree and know just enough to get in trouble.) Here is one more of 42 on cooling fins: http://naca.larc.nasa.gov/reports/19...report-676.pdf

What's the area thru the fins to start with? Do you know? I don't. We are stuck with that fin area we have to provide the needed cooling. To get air down thru the fins to do the "WORK" of cooling needs differential pressure across those fins, to produce mass airflow to reject the heat, other wise CHT's are too high. Cooling is a necessary evil.

The easy answer to your question is we copy what other RV'ers have done. The inlet area has been calculated for us already for our cruise speed and engine HP.

The area of the inlet for my 180HP RV w/ aluminum inlets are about 4.6" dia at the mouth and 4" at exit. They are 1.75" deep and nest into the cowl. The nozzle/diffuser uses the A-10 or -20 NASA laminar airfoil contour. They are custom inlets, not from Sam James, which does not have the same design or internal shape. My cowl is a modified stock cowl with these aluminum inlets to a custom plenum. My total inlet area is 2 * Pi/4 * (4")^2 = 25.2 in-sq, 33 sq based on mouth. How does that relate to fin area of the O360A1A? I don't know or care as long as I go fast and stays cool.

The NASA research was NOT complete and somewhat left unfinished in the early 80's. Great data and observations where made, but they did not give design details. That's where our creativity comes in. When it comes to utilizing the NASA concepts, I think Barnard's Holy Cowl (now Sam James) interpretation is very good but not perfect. I choose to make my own cowl (from a stock one) and plenum to get it the way I wanted it.

There are some excellent books on engine cooling, but don't recall the title or author's. Than there is stacks of NASA reports, many from War Time on the subject of air cooled engines. Me personally, I copied from Dave Anders and Tracy Saylor, because they showed it was a reasonable area. Lazy but smart. Know when to copy. The racers get away with tiny inlets and fixed geometry because they are made to just go fast, not climb for 12,000 feet. So be careful who you copy and why.

The inlet is only one piece of the puzzle. The way you handle the air after that is critical.

Variable inlet geometry would be nice, but that's not really workable. Even Cessna has variable geometry in the form of cowl flaps. Why don't we? Well one it adds weight and design complexity, but it's doable. My exhaust tunnel and cowl flap would involve massive belly structure rework. It could be done, but not sure if it's practical or worth it.

Looking at some of the old classic GA planes of the past you might say, geee what where they thinking. In the last two decades the standard has changed. However they did not have 20 channel engine monitors, much less one CHT. This is where LoPresti made bread and butter in his latter years, making NASA (MSU) inspired cowl mods for older factory planes: http://www.speedmods.com/

Now the new Mooney looks like this: http://www.mooney.com/



Lycoming Cylinders
Absolutly the fins are thicker / deeper on the (hot) exhaust port side by design. Clearly this was engineered (well), but since they where designed in the 1940's / 50's something, all those guys are gone, so we can't ask. The (cool) intake side does not need the same "fin-age" (made up a word to day mom). It does save weight and allow a more compact design. I think the Lyc design is brilliant and well thought out. Also the fin area, height, thickness and spacing has been researched to death I recall, in the 30's/40's.

There's no free lunch, you will always have some cooling drag. I don't think redesign of the fins is the key. Hey that is why the Hawker Sea Fury is the fastest plane at at Reno with a BIG radial. At least the air is going in the same direction as the fin. As was pointed out the air has to turn and go down (draft) on our little Horz opposed engine.

Van's stock baffle if fitted too high and tight on #2 & #3, will choke the air off to the lower (deeper fins) on the bottom of the cylinder. You need all the air you can get there. Kent Paser ("Speed w/ Economy) played with these gaps and got dramatic results.

The washer trick, absolutly you can control the gap and increase the flow on the #2 and more important #3 jug. It is really a local airflow issue, that is all. You are bypassing air to get it where you want. However I suspect you are thinking bypassing the air around the fins (more baffle fin gap) air would speed up. It would reduce drag and cooling may increase to a point. A 1937 NACA report discusses baffle to fin gap: http://naca.larc.nasa.gov/reports/19...aca-tn-620.pdf
(be careful how you use this radial ring cowl data; also more cooling may be more drag as well.)

The washer trick works but is a little crude. It is a good way to experiment, but a better way is have baffles fit properly without shims and washers. When you add washers you tend to get leakage out the side of the lower wrapped around portion of the baffle. Some baffles have formed flanges on the edges to control the edge leakage. You can achieve the similar effect with silicone beads.

My theory is you need to have a good cooling system to start with before you can reduce cooling drag. Every little part has to work together. One small detail deficiency can ruin the efficency of the cooling system.

Mike S - "take a look at the cowling outlet of the AR 6 in the attached link." Isn't that crazy. This is a good use for fiberglass. Every detail on these formula racers are extream. Not sure how much we can incorporate in our "daily flyer's", but agree, there are things to learn. At some point there's a balance and sacrifice for speed. These are all out racers. Cooool! (I think most formula planes are trailered to the race)

[Mike S]

George, my suggestion about the AR 6 cowl was in support of the idea you put forth in a prior post about "Why rv 4's cool so well"

You suggested it might be that the inner cowl is tall and narrow compared to other RV's----------if I read your intent correctly.

No attempt to push fiberglass, but the SHAPE of the thing is what I find interesting. It sticks WAY DOWN below the fuse--------most folks would think "wow That must have lot of drag"--------In my case, I think Hummm, good idea, cause a low pressure area BEHIND the end of the cowl, and SUCK out the hot air.

Mike